REPORT OF THE 2013 MEETING OF THE ICCAT WORKING GROUP ON STOCK ASSESSMENT METHODS (WGSAM) (Madrid, Spain March 11 to 15, 2013)

Tamaño: px
Comenzar la demostración a partir de la página:

Download "REPORT OF THE 2013 MEETING OF THE ICCAT WORKING GROUP ON STOCK ASSESSMENT METHODS (WGSAM) (Madrid, Spain March 11 to 15, 2013)"

Transcripción

1 SCRS/2013/010 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) REPORT OF THE 2013 MEETING OF THE ICCAT WORKING GROUP ON STOCK ASSESSMENT METHODS (WGSAM) (Madrid, Spain March 11 to 15, 2013) SUMMARY The Meeting of the ICCAT Working Group on Stock Assessment Methods was held in Madrid, Spain, March The objective of this meeting was to develop representation models to establish assessment model diagnostics, the definition of limit reference points and management strategy evaluation and the revision of TORs for Peer review, which includes identification and selection of invited experts. RÉSUMÉ La Réunion du Groupe de travail de l ICCAT sur les méthodes d évaluation des stocks a eu lieu à Madrid (Espagne) du 11 au 15 mars L'objectif de cette réunion consistait à élaborer des modèles de représentation destinés à établir des diagnostics des modèles d'évaluation, la définition des points limites de référence et l'évaluation des stratégies de gestion ainsi que la révision des termes de référence du processus d'examen par des pairs, ce qui inclut l'identification et la sélection d'experts invités. RESUMEN La Reunión del Grupo de trabajo de ICCAT sobre métodos de evaluación de stock se celebró en Madrid, España, del 11 al 15 de marzo de El objetivo de la reunión era desarrollar modelos de representación para establecer diagnósticos de los modelos de evaluación, la definición de límites de referencia y evaluación de estrategias de gestión y la revisión de los términos de referencia para el proceso de revisión por pares, lo que incluye la identificación y selección de expertos invitados. 1. Opening, adoption of Agenda and meeting arrangements Dr. Pilar Pallares, ICCAT Assistant Executive Secretary, opened the meeting and welcomed participants. The meeting was chaired by Dr. Paul de Bruyn. Dr. de Bruyn welcomed the Working Group participants, reviewed the objectives of the meeting and proceeded to review the Agenda which was adopted with minor changes (Appendix 1). The List of Participants is attached as Appendix 2. The List of Documents presented at the meeting is attached as Appendix 3. The following participants served as Rapporteurs for various sections of the report: Section Rapporteurs 1 P. de Bruyn 2 R. Coelho and J. Ortiz de Urbina 3 S. Cass-Calay 4 H. Arrizabalaga, G. Merino, D. Die and C. Brown 5 V. Ortiz de Zarate 6 G. Scott 7 N. Abid and C. Palma 8-9 P. de Bruyn 2017

2 2. Model diagnostics Document SCRS/2013/025 summarizes common model diagnostics available for Stock Synthesis and describes their interpretation. Examples of model misspecification are described and the resulting diagnostics are illustrated. The authors also provide a framework to facilitate an efficient and complete evaluation of model diagnostics. The Group acknowledged the importance and usefulness of such diagnostics and noted that most of the tools and diagnostics presented would be similar and useful for most non-linear models, and not necessarily specific to SS3. The Group discussed the importance and some issues with the estimated parameters correlation matrix. One point raised was that in the correlations matrix there may be a dependency between parameters. Still, it was noted that the examination of the estimated correlations between the parameters is an important diagnostic and further work is recommended. A framework for model diagnostics in SS3 is presented in Appendix 4. Document SCRS/2013/027 presents a variety of methods for exploratory data analysis and evaluation of the goodness of fit and inspection of diagnostics for the use in stock assessment. The intention is to list a set of common methods that can be used for a range of stock assessment models (from simple to complex) and for different data requirements, (e.g., biomass, size and age). A range of methods, to show how they can be implemented in R then discuss the consequences for stock assessments and strategies resolving problems identified was listed. A general recommendation from the Group was that all stock assessment papers and documents presented to the ICCAT SCRS should provide appropriate diagnostics for the models being used, recognizing that even though different models and methods may have slightly different diagnostics, in general many diagnostics will be common to all methods. The Group noted the need to move towards having diagnostics in all models, but keeping in mind that the diagnostic tools may be very large and take much time to work on, there was a general recommendation that those tools should be general and flexible. It was suggested that those diagnostics could be developed and used experimentally during 2013 for the albacore and swordfish stock assessments. After receiving feedback from those working groups, the diagnostic tools and recommendations could, as necessary, be revised for next year. The Group noted the importance of having all these methods and tools well documented and freely available in open-source software. It was mentioned that these tools are currently being developed and implemented in R and made available through CRAN or other repositories (e.g., the package CPUE, R4SS, FLR, R4MFCL, etc). The CPUE package can read model outputs from any assessment method (e.g., ASPIC, VPA, MFCL, SS3) and then run a standard set of model diagnostic tools. Examples on the use of those tools were provided to the Group, and some graphical examples are provided in the paper. Document SCRS/2013/030 documents available model diagnostics for the virtual population analysis routine VPA-2BOX. The particular example used for this document is the 2012 assessment of western Atlantic bluefin tuna. Model diagnostics included model convergence statistics, bias and standard error of model results, correlation and co-variance matrices for model parameters, sensitivity analyses of abundance indices and life history assumptions, and bootstrap analysis to evaluate model robustness and estimate central tendency and variance of parameter estimates. The usefulness of these diagnostics in evaluating model performance during the stock assessment is discussed, and recommendations for future improvement are provided. The Group noted the importance of such tools and discussed, in general, the bootstrapping procedure, and its potential for diagnosing problems with violations of model assumptions. The Group also noted that preliminary data analysis and data cleaning procedures remain important procedures. 2018

3 3. Review of current ICCAT methods for estimating EFFDIS Document SCRS/2013/021 presents a review of the previous methods of estimation of the total longline effort. The most recent estimation adopted by the SCRS was made in 2009 used the nine major ICCAT tuna and tuna like species to obtain Task I global nominal catches (in weight) and CPUEs from partial catch and effort (Task II) statistics. In 2011, the Secretariat presented a proposal based on a similar methodology but splitting the base effort calculations in two main areas (ATL/MED) in order to reduce the side effects of poor catch and effort statistics available for the Mediterranean Sea. The overall longline (LL) effort estimations increased considerably (>10%) in most recent years. The SCRS considered this new approach an important improvement, but at the same time noted that, LL Task II catch and effort statistics of important Mediterranean Countries need to be revised. The current model basic assumption considers that catch rates are equivalent at partial and global level. Comparing the results with previous estimates (obtained during the inter-sessional meetings on ecosystems in 2007 and 2008) the global results do not show major differences. However, at more disaggregated levels the differences are larger for certain flags. In the majority of the cases, large relative variations are usually associated with various improvements and corrections made to some specific datasets. Global geographical distribution shows a small increase near the Venezuela waters due to various corrections made to spatial distributions of catch and effort statistics from Venezuela in several years. Potential areas for improvement of EffDIS estimation for the future are presented in order to stimulate discussion at the 2013 ICCAT WGSAM on ways to improve EffDIS information. The Group recognized that EffDIS is a valuable tool that allows visualization and quantification of the spatial and temporal distribution of longline effort. It is often used by Sub-Committee on Ecosystems, particularly for assessments of the impact of ICCAT fisheries on bycatch species, including seabirds. In the past there has been concern regarding the method used to calculate these estimations and improvements in the methodology are required. Recent improvements to the dataset were described including stratification by month and the addition of fleets which reduced the total effort of unclassified fleets ( Other ) from 17% to 13%. The Group also discussed recommendations for improving the EffDIS dataset, including where possible: That data series reported by quarter or year be resubmitted by month. That data series reported without spatial information, or by stratification larger than 5 by 5 be resubmitted using at least a 5 by 5 stratification. That data series reported without longline effort be revised to include the corresponding effort. Recover historical datasets, particularly prior to Expand the number of species used in the raising of effort using the ratio of T2CE to T1NC to include major shark species (blue, porbeagle, and mako). This action would improve estimation of total effort. Note: EffDIS estimates are sensitive to species composition. A more complete characterization of species composition could be helpful. Differentiating Atlantic and Mediterranean fishing effort. Revise/improve the most important Mediterranean Sea Task II catch and effort series. To investigate the possibility of integrating into the analysis metadata regarding fleet behaviour targeting, (aiming to identify fleet profiles) in order to achieve more accurate projections of relative effort of fishing fleets. Examine suitability of assumptions for raising fleets (e.g., U.S. and Japan assumed to report full information and are not currently raised). The Secretariat noted that improvements to EffDIS began in 2007 and are ongoing. Some improvements are largely complete (i.e., most catch-effort data is now by month and by 5 by 5 degree square) while others are pending. The Group agreed that the recommendations listed above are appropriate and important. In addition, the Group suggested the following: An effort should be made to develop similar EffDIS estimates for the BB and PS gears. This would be particularly useful for evaluations of time-area closures. 2019

4 The Working Group recognizes the importance of accounting for changes in fishing operations and characteristics of the main fleets from each CPC operating within the ICCAT Convention area as these affect the efficiencies of the fleets for catching target and by-catch species. Documentation of these technological and behavioral changes is particularly important to understand the national reports of catch and effort annually submitted (Task II-CE). Taking into consideration also that CPCs are required to report fleet composition data (Task I-FC), it is recommended that CPCs present an SCRS document with the details of the fleet composition, sampling, coverage, and statistical methodology to estimate total catch, catch and effort, catch-at-size for each of the main fleet components. This report should also communicate the potential limitations and or restrictions of the data and information provided to be taking into account within any further analysis by the SCRS or the Secretariat. The uncertainty of EffDIS estimates should be characterized. Substitutions, raising ratios and proportion of unclassified fleets ( other ) should be clearly described in documentation. Additional methods of raising data should be explored (e.g., Rubin 1976) to evaluate their statistical rigor. Methods to cross-validate EffDIS estimates should be explored (e.g., VMS data). 4. Limit reference points and management strategy evaluation At the beginning of the session, draft conclusions of a recently held ISSF Workshop (Harvest Control Rules and Reference Points for Tuna RFMOs, San Diego, California, USA, March 6-8, 2013) were presented to the Methods Group as the meeting pertained to Management Strategy Evaluations and Limit Reference Points agenda item. It was noted that the conclusions presented were considered draft since a final review of the Workshop report was still in progress. The ISSF Workshop participants indicated that management strategies include monitoring, stock assessment, harvest control rules, reference points and management actions. The workshop was convened to review the current status of the adoption of these elements into the decision-making process of five Tuna RFMOs and to make recommendations for harmonizing and facilitating the process among RFMOs. Of the five RFMOs, CCSBT has formally adopted a management strategy (management procedure) for decision-making. The other four RFMOs are making substantial progress to identify and test key elements of management strategies, such as reference points (limit and target) and harvest control rules. Some of this work is being done primarily by the RFMO science bodies, sometimes without formal Commission mandates. The workshop focused on key issues that should be kept in mind when developing and testing management strategies: data and models, the treatment of F MSY as a target or a limit, testing of the strategy, and implementation. The workshop noted that these management strategies have worked quite well elsewhere in fisheries, and that there is no technical constraint to advance them in the tuna RFMOs. This could be done in many cases with relatively simple, existing tools. The report contains specific recommendations on limit and target reference points, harvest control rules, and other considerations for management strategy evaluations. The finalized report of the workshop shall be posted on the ISSF web site (http://iss-foundation.org) in the near future. A summary of the history of development of reference points, harvest control rules and the implementation of Management Strategy Evaluations in ICCAT was presented. In the ICCAT Convention text MSY is the only reference point mentioned. The precautionary approach was developed after the creation of ICCAT and the SCRS has discussed the benefits of having precautionary target reference points. However the Convention text has not still incorporated other reference points and MSY has therefore functioned as the target reference point for ICCAT since the creation of the Commission. But recently the Commission adopted a Decision Framework (Rec ) which effectively treats F MSY as a limit and not a target reference point. The presentation was divided in two parts summarizing ICCAT and trfmo s activities: The first part summarised the implementation of the Kobe process in ICCAT. A range of stock assessment methods are used by the Standing Committee of Research Statistics (SCRS), e.g., ASPIC, Bayesian Surplus Production Models, Adapt, Multifan-CL, Stock Synthesis. The main management objective of ICCAT is to maintain the populations of tuna and tuna-like fishes at levels which will permit the maximum sustainable catch. Thus MSY was originally interpreted as a target. ICCAT was formed before the Precautionary Approach (PA) and the Ecosystem Approach to Fishery Management (EAFM), so neither are mentioned in the convention, in spite of this, the Commission has embarked on including both the PA and EAFM in its decisions. Stock assessments routinely consider a range of uncertainties and assessment are conducted for bycatch species (e.g., seabirds, turtles) and sharks. 2020

5 Recovery plans are in place for both the eastern and western bluefin (as well as for other stocks like North Atlantic swordfish), and work on developing an Operating Model under the GBYP is commencing this year. Harvest Control Rules (HCR) including Limit Reference Points (LRPs) are being developed for North Atlantic swordfish and North Atlantic albacore this year, by conducting a Management Strategy Evaluation (MSE) to evaluate the performance of reference points as part of HCR. The second part summarized the Third Joint Tuna RFMOs meeting (Kobe III; It was recognised that an MSE process needs to be widely implemented in the trfmos in order to implement a PA for tuna fisheries management. Kobe III recommended that a Joint MSE Technical Working Group be created and that this joint working group work electronically, in the first instance, in order to minimize the cost of its work. Three activities are currently being conducted i.e., a review of the Kobe Advice framework, MSE tools and the use of parallel and cloud computing. ICCAT is developing HCRs which include limit reference points (LRP) for North Atlantic swordfish (see Rec ) and North Atlantic albacore. The SCRS plans to develop these HCRs by evaluating the performance of alternative limit reference points when they are incorporated in an HCR. This evaluation can be done through simulations conducted with an MSE framework which incorporates a range of quantifiable uncertainties. The Group discussed that incorporation of such range of uncertainties can be complicated when there are multiple assessment scenarios or assessment models used. In such cases it is advantageous to select HCRs and LRPs which are robust across models and/or scenario results. The Group also discussed the fact that reference points may not remain constant in time. The ICCAT WGSAM has discussed in the past how to develop an HCR (Anon. 2011). The Commission has recommended that stocks be managed with a high probability (Rec ) of being in the green quadrant of the Kobe phase plot, although the Commission did not specify that probability level. The Group discussed that the probability of various management strategies, including alternative HCRs, to comply with this Recommendation can be evaluated with MSEs. The Group noted the importance of clarifying and harmonizing terminology. When revising the generic HCR described in WGSAM (Anon. 2011), the Group noted that, although the WGSAM focused on B LIM (the biomass level below which the HCR line is set at F=0; i.e., the point where the fisheries should be closed), whereas managers may be more interested in B thres (the hinge-point of the HCR occurring at a level of biomass reduction below which some reductions in F should start). This generic HCR can be parameterized for each stock (i.e., setting up concrete values for the so called B LIM, B thres and the target value of F). An example of a harvest control rule and limit reference points to be considered by the Swordfish assessment Group in its June meeting is shown in Figure 1. The Group agreed that the data availability will drive the decision of whether designing generic or species specific harvest control rules. For some data poor stocks, it may be necessary to develop an HCR based only on F because there is little data available for estimating biomass. The Group discussed the need to define a threshold level of relative biomass below which, the HCR indicates additional management is triggered. MSEs may be useful in defining the appropriate threshold level that minimizes unnecessary management action as the stock fluctuates naturally around MSY levels, but still provides adequate protection for maintaining stock levels. Another possibility raised was to define nonlinear HCRs that would be triggered once biomass fell below that which would support MSY (e.g., the threshold level would be equal to B MSY ), but which would require negligible reductions when biomass was close to B MSY (e.g., exponential decline); this alternative would not require a separately define threshold level. A wide variety of HCRs, including the shapes of the reduction functions, can be evaluated through simulation. Rec for North Atlantic albacore requests a limit reference point that would trigger a rebuilding plan when biomass falls below it (B THRESHOLD ). The Group discussed plans to develop this for albacore (SCRS/2013/033, SCRS/2013/034 and SCRS/2013/035), using MSE, and then to draw upon this experience in order to prepare LRPs for the North Atlantic swordfish assessment. It was noted that, in past assessments, swordfish management advice was based on ASPIC results, whereas the initial work for albacore would be based on Multifan. In addition to considerations of the time involved to develop swordfish-specific MSEs in time for this year s assessment, the Group considered that the applicability may depend on the type of model(s) used for the management advice. In other words, that the Assessment Model to use in the MSE should not be more complex than the Operating Model used (Multifan for albacore and ASPIC for swordfish). If the full range of uncertainties is going to be incorporated alternative fleet/fishery mixes (e.g., relative effort of longline vs. purse seine vs. baitboat, or otherwise between fleets with different age selectivities, which is highly involved for Multifan) may have to be considered. This has a high price in terms of time for multiple runs. As alternative mixes of fisheries would likely results in changes in selectivity and thereby changes to benchmarks, careful consideration must be given to how to interpret results across multiple scenarios. 2021

6 The WGSAM recognized that this is an ongoing process, and that short-term and long-term objectives should be clarified. The Group recognized the importance of a firm commitment to this process, including clear objectives and commensurate funding, in order to meet management needs. For further developments in this task, WGSAM indicated a need of new projects and international collaborations to move towards a common MSE framework. The progress on the Albacore case study was presented as an example of MSE framework that could be extended to other stocks. In addition, WGSAM recognized the benefit of the work presented as a result of collaborations between ICCAT s Secretariat and national organisations and encourages further interactive cooperation. 5. Identification of key research needs and components of the SCRS Science Strategic Plan as well as identification of capacity limitations and gaps and how these can be addressed Strategic Planning is recommended as a structured approach to guide the future workings of the SCRS (2011 SCRS Report and responsive to Res on Best Available Science). Document SCRS/2013/024 outlined an approach for identifying key research needs and components of and a roadmap for developing the SCRS Strategic Plan. SCRS/2013/024 points out that Strategic Planning deals with three basic constructs: What do we do?, For whom do we do it? and How do we excel? Furthermore, the key components of strategic planning include an understanding of the SCRS mission (our purpose), our vision for the future, values we shall apply in conduct of our work, our goals and strategies to achieve them. It was pointed out that Strategic Planning also provides a methodology to identify critical capacity and data gaps and prioritize research activities to address them. A roadmap and time-frame for developing the SCRS Strategic Plan was proposed in SCRS/2013/024 which includes contracting a consultant to provide a framework for the specific methodology to be applied in developing the Strategic Plan and regular consultation and review by SCRS officers and SCRS Plenary prior to review and acceptance by the Commission. The Methods Working Group endorsed the plan and recommended plan development be initiated as outlined in SCRS/2013/ A discussion and amendment of the current ICCAT peer review TORs, as well as the agreement of a protocol for invited expert/reviewer selection In its Resolution on Best Available Science (Res ), the Commission called for strengthening the peer review mechanisms within the SCRS, including the participation of outside experts. This section provides some background on the evolution the peer review process within the SCRS, as well as discussions by the Group on how to strengthen the peer review process and enhance the participation of external experts, including the development of terms of reference. When considering these recommendations, it is important to note that the Commission also called on its members (within the same Resolution, Res ) to consider broadening financial support and mechanisms for the purpose of the implementation of this peer review/external expert participation strengthening, as well as for the support of the other objectives identified therein. In order to successfully implement the steps described within this report, and achieve the objectives called for in Res , it is critical that this effort be fully funded. External peer reviews of the work of ICCAT SCRS working groups have been conducted (Santiago et. al. 2013). At the 2010 Kobe II meeting it was concluded that peer review should be included in all trfmo scientific assessments. The Report of the Independent Performance Review of ICCAT (Hurry et.al., 2008) indicated the analyses used by the SCRS to formulate its advice are peer reviewed through a rigorous three stage process working/assessment groups to species groups to SCRS plenary). The structure of the process, the diversity of the participants/analysts and the large number of people involved does not guarantee that errors will not be made, but it provides a reasonable assurance that if errors are made, they will be discovered, admitted, discussed and corrected. Peer reviews of assessments in ICCAT currently in place follow the process adopted by the SCRS in Recently, the SCRS recommended conducting at least 2 in situ reviews per year. The purpose of the reviews is to provide additional peer review to SCRS and its species groups to guide improvements in stock assessments. 2022

7 During the 2012 meeting of WGSAM, the issue of peer review in ICCAT was again discussed and Terms of Reference (TOR) for external experts as peer reviewers in the SCRS stock assessment meetings were drafted (Anon. 2013). Terminology regarding both invited experts and external reviewers attending the WGs have been used somewhat interchangeably in ICCAT. Therefore, the Group reviewed documents presented at the meeting that dealt with these topics. The WGSAM discussed document SCRS/2013/23 which presented potential TORs to distinguish between invited experts and external reviewers. SCRS/2013/23 identified three different levels of peer review functioning. First, there is internal peer review, which is usually facilitated by working groups of diverse national scientists who in real time review and provide critical input/advice to the scientific process; on occasion this is supplemented by the participation of an external expert. Second, external peer review is facilitated by sending the assessment outcomes to contracted external experts for review and quality control or having the experts attend working group meetings as observers who then report on the meeting outcomes. External peer review may also be conducted through a joint meeting with an advisory expert panel. Lastly, scientific peer review may be facilitated by publishing scientific outputs in peer reviewed journals or presenting them at international conferences. Invited experts take part in the assessment process, providing information and advice on how the stock assessment can be conducted/improved/streamlined within the assessment process. An external reviewer should in theory play no active part in the assessment. In this context, the Secretariat provides a potential transparent process for selection of experts from a list of experts kept by ICCAT. Another document (SCRS/2013/028) was presented to WGSAM. It described aspects of the Center for Independent Experts (CIE), a NOAA-funded process that provides peer reviews for the U.S. National Marine Fisheries Service (NMFS), in the United States. The process of selecting reviewers consists of matching the skills required for the review, with the ability of suitable experts subject to constraints set up to avoid picking candidates with potential conflicts of interest. This process is conducted by the CIE independently of the client, NMFS. This paper presented some of the lessons learned by the CIE to inform the discussions on the TOR for the 2013 peer review of the upcoming albacore assessment in 2013 and future peer reviews. Furthermore, the paper explicitly highlights some problems that are pending in the ICCAT review process (Table 1). After reviewing all the information provided, discussion took place to clarify the separation of invited expert versus reviewer roles within SCRS performance. Consequently a text with TORs was accomplished for both invited expert and reviewer assistance to the peer review SCRS process. Revised TORs for an invited expert are included in Appendix 5 and revised TORs for an external reviewer in Appendix 6. Likewise, it was recommended that the expert participation in a given species stock assessment will be proposed by the Working Group Chair and the SCRS Chair. In addition, a list of CIE and other RFMO reviewers shall be kept available at ICCAT. Alternatively, ICCAT could consider a common list of reviewers developed with other RFMOs as suggested in SCRS/2013/028. The Group discussed the possibility of using the salary rates and time frames for CIE reviews as guidelines for ICCAT to contract peer reviews. Currently CIE reviewer is paid a consulting fee of US $ per day and an average of 14 days for stock assessment review (2 days of travel, 5 days of meeting, 4 days of preparation and 3 days of report writing) (SCRS/2013/028). 7. Other matters 7.1 ICCAT cloud computing The cloud is an emerging paradigm in the way we use and share computing resources (hardware, software,) and information (documents and data). Virtually unlimited (budget dependent) and elastic (use what you need) computing resources, can be deployed in minutes instead of weeks. Computing power, security, concurrency and shared work are fully optimized on the cloud. Provision of advice for the Commission is increasingly dependent upon the use of computer intensive methods such as, Monte Carlo simulations, bootstrapping used when building Kobe II strategy matrices, MCMC runs, stochastic projections used on stock assessments and for example the evaluation of limit reference points using MSE. Often in the last few years it has not always been possible to conduct such analyses during working groups, making it difficult for working groups to finalise reports. Therefore the SCRS (guided by the Methods Working Group and the trfmo MSE WG) recommended the ICCAT Secretariat to study the use of the Cloud Infrastructure (virtual servers, distributed and parallel computing, concurrent services, etc.). 2023

8 The ICCAT Secretariat presented a work plan to the Group with the guidelines for the development of ICCAT Cloud Computing infrastructure (document to be added to This work plan describes the preliminary studies made, the cloud topology/model proposed, and the requirements (hardware, and cloud services) necessary to a first year development phase (Phase 1). An estimation of the expected costs was also presented. In addition, the Secretariat also presented to the Group the 2013 deployment plan (Phase 1) and, finally, the current development status. In summary, it was described how the cloud servers are administered and used, how services are deployed and configured (Apache web server, R-CRAN, RStudio server, etc.) and the documentation (short tutorials: cloud administration, Studio administration, remote access User s guide) already available. All these documents should be published on the under-development ICCAT cloud website (http://tunalab.iccat.int). The Secretariat has planned to start some important tests of the cloud infrastructure already deployed, starting with the 2013 Albacore Data Preparatory Meeting (Madrid, Spain April). 7.2 Future work plan The Working Group discussed the future work plan and retained mainly the following actions: WGSAM recommends reviewing the protocols and algorithms for estimating Effort distribution (5x5) for longline (EFFDIS), and extended to purse-seine, and bait boat gears, currently prepared by the Secretariat. The WG should also include estimates of uncertainty on these products. It is suggested that published estimates in the ICCAT Web page, include also detailed description of the estimate assumptions and uncertainty related to these products to make aware the potential users of their limitations. The Commission expects risk-based advice on management measures as prescribed in the Kobe II Strategy Matrix and as embedded in its Decision Framework (Rec ). An important aspect of providing such scientific advice is adequate quantification of uncertainty in stock condition and future prospects under future management option scenarios. With the advent of more commonly applied, highly parameterized stock assessment models, the computational investment in quantifying uncertainty in stock status and future prospects is quite heavy. This is also the experience at other trfmos and a number of approximations for quantifying both process and observational uncertainty are being applied to develop risk-based management advice. Guidance on the evolution of and possibility of harmonizing methods to apply for uncertainty characterization across species groups should be provided by WGSAM. Including during the agenda items of 2014 some of the Horizontal Themes identified during the process of elaborating the SCRS Strategic Plan in 2013, particularly those related to participation and capacity building and quality control of the stock assessments and management advice. WGSAM recognized that there is a trend in recent assessments conducted by the SCRS to use multiple modeling methods to estimate the status of the stock relative to ICCAT conservation benchmarks. While WGSAM agrees the use of multiple approaches is a good practice, situations have arisen where the different methods give results that are not consistent yet equally plausible. Having guidance from the WGSAM on best practices to reconcile or combine such results would be very helpful (see, for example, ICES 2007). 7.3 Collaboration The World Conference on Stock Assessment Methods for Sustainable Fisheries (WCSAM) will take place in Boston, USA, July The conference will provide a forum for presentations on the application and future of stock assessment methods. It will consider single stock approaches for data rich and poor stocks, and also multispecies and ecosystem based approaches. It is being organised by researchers from a range of scientific institutions and RFMOs across the world. The conference will be preceded by a two-day workshop (15-16 July 2013) where studies on the application of stock assessment methods to predefined data sets will be reviewed. ICCAT is actively participating in WCSAM. Continued collaboration through participation in other Regional Fisheries Bodies (RFBs) meetings regarding implementation of MSE is needed to enhance utilization of this important tool for addressing uncertainties and risks associated with stock assessment models and providing better scientific advice. 2024

9 Continued collaboration with ICES working groups and, in particular for assessment of sharks species is viewed as an important work area. WGSAM recognized the benefits of the work presented as a result of collaborations between ICCAT s Secretariat and national organisations and encourages further interactive cooperation. 8. Recommendations 1) WGSAM recommends SCRS and the Secretariat work with other trfmos to develop common protocols for peer reviews when applicable, particularly with regard to identification of suitable experts. 2) Diagnostics should be evaluated for assessment models. Suitable diagnostics may vary between assessment models, but model appropriate diagnostics should be presented to help evaluate the quality of management advice arising from the assessments. 3) Peer review reports on stock assessment working groups should be in the form of an SCRS document, with a summary of the peer review included as part of the detailed report of the assessment meeting. The recommendations in the peer review summary are to be included and taken into consideration in future assessment sessions and may not necessarily be addressed in the same year as the assessment. 4) Long term work plans should be drafted for the joint trfmo MSE Working Group in order to ensure the Group functions effectively and made available on the Group s website. (http://code.google.com/p/trfmomse/) 5) WGSAM endorsed and recommended the plan outlined in SCRS/2013/024. As the SCRS Strategic Plan is currently being developed, Species Groups should include an item in the agendas of their meeting in 2013 to assess data gaps and needs and identify goals and strategies in advance of SCRS Plenary to permit the 2013 SCRS to validate goals and strategies and agree on mission, vision, and values components for the Strategic Plan. 6) Reimbursement for invited experts and external reviewers could be based on the standard time frames and rates developed by the CIE. Invited external experts and peer reviewers should follow the TORS prescribed by the WGSAM in ) WGSAM has described specific recommendations to improve estimation of the temporal and spatial distribution of longline effort (EffDIS), and recommends that these efforts continue. In addition, WGSAM recommends that efforts be made to develop similar EffDIS estimates for the BB and PS gears. 8) For years in which stock assessments are to be conducted, in order to enhance quality assurance of scientific advice, working groups are required to prepare detailed work plans in order to provide guidance for the meeting preparations and to ensure complete and timely availability of required data and model inputs, as well as to facilitate the coordination of responsibilities within the working group as and/or with the Secretariat. 9) WGSAM recognizes the importance of accounting for changes in fishing operations and characteristics of the main fleets from each CPC operating within the ICCAT area of competence, as these affect the efficiencies of the fleets for catching target and by-catch species. Documentation of these technological and behavioral changes is particularly important to understand the national reports of catch and effort annually submitted (Task II-CE). Taking into consideration also that CPCs are required to report fleet composition data (Task I- FC), it is recommended that CPCs present an SCRS document with the details of the fleet composition, sampling, coverage, and statistical methodology to estimate total catch, catch and effort, catch-at-size for each of the main fleet components. This report should also communicate the potential limitations and or restrictions of the data and information provided to be taking into account within any further analysis by the SCRS or the Secretariat. 9. Adoption of the report and closure The report was adopted during the meeting. The Chairman thanked the participants and the Convener of WGSAM for their efficiency and hard work. The meeting was adjourned 2025

10 References Anon Report of the 2010 ICCAT Working Group on Stock Assessment Methods (Madrid, Spain, April 21 to 23, 2010). Collect. Vol. Sci. Pap. ICCAT, 66(3): Anon Report of the 2012 ICCAT Working Group on Stock Assessment Methods (Madrid, Spain, April 16 to 20, 2012). Collect. Vol. Sci. Pap. ICCAT, 69 (3): Hurry, G.D., Hayashi, M. and Maguire, J.J Report of the Independent Review. International Commission for the Conservation of Atlantic Tunas. Unpublished ICCAT report, ICCAT, Madrid. 105pp ICES, Report of the Study Group on Risk Assessment and Management Advice. ICES CM 2007/RMC:02. Neilson, J., Arocha, F., Calay, S., Mejuto, M., Ortiz, M., Scott. G., Smith, C., Travassos, P., Tserpes, G. and Andrushchenko, I The Recovery of Atlantic Swordfish: The Comparative Roles of the Regional Fisheries Management Organization and Species Biology, Reviews in Fisheries Science, 21:2, Rubin, D.B Inference and Missing Data. Biometrika 63: Santiago, J., Scott, G.P., Pereira, J.G Implementation of best science in the SCRS. ICCAT Collect. Vol. Sci. Pap. ICCAT, 69 (5):

11 RAPPORT DE LA RÉUNION DE 2013 DU GROUPE DE TRAVAIL ICCAT SUR LES MÉTHODES D ÉVALUATION DES STOCKS (Madrid, Espagne, mars 2013) 1 Ouverture, adoption de l ordre du jour et organisation des sessions Le Dr Pilar Pallarés, Secrétaire exécutive adjointe de l ICCAT, a ouvert la réunion et a souhaité la bienvenue aux participants. La réunion a été présidée par le Dr Paul de Bruyn. Après avoir souhaité la bienvenue aux participants au Groupe de travail, Dr de Bruyn a passé en revue les objectifs de la réunion et a procédé à l examen de l ordre du jour qui a été adopté avec quelques changements mineurs (Appendice 1). La liste des participants figure à l Appendice 2. La liste des documents présentés à la réunion est jointe à l'appendice 3. Les participants suivants ont assumé la tâche de rapporteurs pour les diverses sections du rapport : Section Rapporteurs 1 P. de Bruyn 2 R. Coelho et J. Ortiz de Urbina 3 S. Cass-Calay 4 H. Arrizabalaga, G. Merino, D. Die et C. Brown 5 V. Ortiz de Zarate 6 G. Scott 7 N. Abid et C. Palma 8-9 P. de Bruyn 2. Diagnostics du modèle Le document SCRS/2013/025 récapitule les diagnostics du modèle communs disponibles pour Stock Synthèse et décrit leur interprétation. Des exemples d'erreurs de spécification du modèle sont décrits et les diagnostics résultants sont illustrés. Les auteurs fournissent également un cadre visant à faciliter une évaluation efficace et complète des diagnostics du modèle. Le Groupe a reconnu l'importance et l'utilité de ces diagnostics et a fait remarquer que la plupart des outils et des diagnostics présentés seraient similaires et utiles pour la plupart des modèles non-linéaires, et pas nécessairement spécifiques à SS3. Le Groupe a discuté de l'importance de la matrice de corrélation des paramètres estimés et de quelques questions y relatives. On a soulevé le point selon lequel, dans la matrice de corrélation, une dépendance pourrait exister entre les paramètres. Il a toutefois été fait remarquer que l'examen des corrélations estimées entre les paramètres constitue un diagnostic important et que des travaux supplémentaires sont recommandés. L'Appendice 4 fournit un cadre pour les diagnostics du modèle dans SS3. Le document SCRS/2013/027 présente une gamme de méthodes pour l'analyse exploratoire des données et l'évaluation de la qualité de l'ajustement et pour l'inspection des diagnostics aux fins de leur emploi dans l'évaluation des stocks. Il s'agit d'énumérer un ensemble de méthodes communes pouvant être utilisées pour une gamme de modèles d'évaluation des stocks (du simple au plus complexe) et pour différentes exigences en matière de données (p.ex. biomasse, taille et âge). Une gamme de méthodes a été présentée, afin de démontrer comment elles peuvent être mises en œuvre dans R, et examiner ensuite les conséquences pour les évaluations de stocks et les stratégies pour résoudre les problèmes identifiés. Le Groupe a formulé une recommandation générale, à savoir que tous les documents sur les évaluations des stocks présentés au SCRS de l'iccat devraient fournir des diagnostics appropriés pour les modèles qui sont utilisés, tout en reconnaissant que même si différents modèles et méthodes peuvent avoir des diagnostics légèrement différents, en général, de nombreux diagnostics seront communs à toutes les méthodes. 2027

12 Le Groupe a évoqué la nécessité de disposer de diagnostics dans tous les modèles, tout en gardant à l'esprit que les outils des diagnostics peuvent être très volumineux et que leur utilisation peut prendre un temps considérable ; le Groupe a recommandé en règle générale que ces outils soient génériques et souples. Il a été suggéré que ces diagnostics soient mis au point et utilisés en 2013 à titre expérimental pour les évaluations des stocks de germon et d'espadon. Après avoir reçu les réactions de ces Groupes d'espèces, les outils de diagnostic et les recommandations pourraient, si nécessaire, être révisés pour l'année suivante. Le Groupe a constaté qu'il était important que ces méthodes et outils soient bien documentés et librement disponibles dans un logiciel open-source. On a mentionné que ces outils étaient actuellement élaborés et mis en œuvre dans le langage R et qu'ils étaient disponibles à travers CRAN ou d'autres plateformes (p.ex. progiciel CPUE, R4SS, FLR, R4MFCL, etc.). Le progiciel "CPUE" peut lire les sorties du modèle à partir de n'importe quelle méthode d'évaluation (p.ex. ASPIC, VPA, MFCL, SS3) et il exécute un jeu standard d'outils de diagnostic du modèle. On a fourni au Groupe des exemples sur l'emploi de ces outils et des exemples graphiques sont présentés dans le document. Le document SCRS/2013/030 documente les diagnostics du modèle disponibles pour l'analyse de population virtuelle de routine VPA-2BOX. L'exemple particulier utilisé pour ce document est l'évaluation du thon rouge de l Atlantique Ouest de Les diagnostics du modèle incluaient les statistiques de convergence du modèle, les biais et erreur standard des résultats du modèle, la corrélation et les matrices de covariance pour les paramètres du modèle, les analyses de sensibilité des indices d'abondance et les postulats sur le cycle vital et les analyses par bootstrap pour évaluer la solidité du modèle et estimer la tendance centrale, ainsi que la variance des estimations des paramètres. L'utilité de ces diagnostics dans l'évaluation des performances du modèle pendant l'évaluation des stocks est discutée et des recommandations sont formulées aux fins d'amélioration future. Le Groupe a constaté l'importance de ces outils et a discuté, en termes généraux, de la procédure par bootstrap et de son potentiel pour diagnostiquer des problèmes d'infraction des postulats du modèle. Le Groupe a également constaté que les analyses de données préliminaires et les procédures de nettoyage des données demeurent des procédures importantes. 3. Examen des méthodes actuelles de l'iccat visant à estimer EFFDIS Le document SCRS/2013/021 présente un examen des méthodes d estimation antérieures de l'effort palangrier total. La plus récente estimation adoptée par le SCRS remonte à 2009 ; elle utilisait les neuf principales espèces de thonidés et d espèces apparentées de l'iccat pour obtenir les prises nominales globales de la Tâche I (en poids) et les CPUE à partir des statistiques partielles de prise et d'effort (Tâche II). En 2011, le Secrétariat a présenté une proposition reposant sur une méthodologie similaire, mais divisant les calculs de l'effort de base en deux zones principales (ATL/MED) dans le but de réduire les effets secondaires des statistiques de prise et d'effort insuffisantes dont on disposait pour la Méditerranée. Les estimations de l'effort palangrier global (LL) se sont considérablement accrues (>10%) au cours de ces dernières années. Le SCRS a considéré que cette nouvelle approche constituait une amélioration importante, mais il a constaté dans le même temps que les statistiques de prise et d'effort palangriers de la Tâche II d'importants pays de la Méditerranée devaient être révisées. Le postulat fondamental du modèle actuel considère que les taux de capture sont équivalents au niveau partiel et global. Si l'on compare les résultats avec des estimations antérieures (obtenus pendant les réunions intersessions du Sous-comité des Écosystèmes en 2007 et 2008), les résultats globaux ne font pas apparaître de grandes différences. Or, à des niveaux plus désagrégés, les différences sont importantes pour certains pavillons. Dans la majorité des cas, de grandes variations relatives sont habituellement associées à diverses améliorations et corrections effectuées à certains jeux de données spécifiques. La distribution géographique globale fait apparaître une faible augmentation près des eaux vénézuéliennes en raison de diverses corrections apportées aux distributions spatiales des statistiques de prise et d'effort du Venezuela au cours de plusieurs années. Des domaines potentiels d'amélioration de l'estimation d'effdis pour l'avenir sont présentés afin de stimuler les débats du Groupe de travail sur les méthodes d'évaluation des stocks de 2013 en ce qui concerne les façons d'améliorer les informations EffDIS. 2028

13 Le Groupe a reconnu que EffDIS est un outil efficace qui permet de visualiser et de quantifier la distribution spatio-temporelle de l'effort palangrier. Le Sous-comité des Écosystèmes l'utilise fréquemment, notamment pour les évaluations de l'impact des pêcheries de l'iccat sur les espèces accessoires, y compris les oiseaux de mer. Par le passé, la méthode employée pour calculer ces estimations a suscité des préoccupations et la méthodologie doit être améliorée. Les récentes améliorations au jeu de données ont été décrites, notamment la stratification par mois et l'ajout de flottilles qui ont ramené l'effort total des flottilles non-classifiées ("Autres") de 17% à 13%. Le Groupe a également discuté des recommandations visant à améliorer le jeu de données d'effdis, y compris, si possible : Re-soumettre par mois les séries de données déclarées par trimestre ou par année. Re-soumettre les séries de données déclarées sans information spatiale, ou par stratification supérieure à 5 x 5 en utilisant une stratification d'au moins 5 x 5. Réviser les séries de données déclarées sans effort palangrier afin d'inclure l'effort correspondant. Récupérer les jeux de données historiques, notamment antérieurs à Élargir le nombre d'espèces utilisées dans l'extrapolation de l'effort en utilisant le ratio de T2CE à T1NC afin d'inclure les principales espèces de requins (requin peau bleue, requin-taupe commun et requin-taupe bleu). Cette action améliorerait l'estimation de l'effort total. Note : Les estimations d'effdis sont sensibles à la composition spécifique. Il serait utile de disposer d'une description plus complète de la composition spécifique. Établir la différence entre l'effort de pêche dans l'atlantique et dans la Méditerranée. Réviser/améliorer les principales séries de prise et d'effort de la Tâche II de la mer Méditerranée. Examiner la possibilité d'intégrer dans l'analyse des métadonnées concernant le comportement de ciblage des flottilles (dans le but d'identifier les profils des flottilles) afin de parvenir à des projections plus précises de l'effort relatif des flottilles de pêche. Examiner le caractère pertinent des postulats d'extrapolation des flottilles (p.ex. on postule que les États- Unis et le Japon déclarent la totalité des informations et leurs données ne sont pas actuellement extrapolées). Le Secrétariat a observé que les améliorations apportées à EffDIS ont démarré en 2007 et se poursuivent. Certaines améliorations sont largement complètes (c.-à-d. que la plupart des données de prise-effort sont désormais par mois et par carré de 5ºx5º) tandis que d'autres sont en suspens. Le Groupe a convenu que les recommandations énumérées ci-dessus sont appropriées et importantes. En outre, le Groupe a suggéré ce qui suit : Un effort devrait être déployé pour élaborer des estimations d'effdis similaires pour les engins de palangre et de senne. Ceci serait particulièrement utile pour les évaluations des fermetures spatiotemporelles. Le Groupe de travail reconnaît l'importance de tenir compte des changements dans les opérations de pêche et les caractéristiques des principales flottilles de chaque CPC qui opère dans la zone de la Convention de l ICCAT, étant donné que ceux-ci affectent l'efficacité des flottilles pour capturer les espèces cibles et les espèces accessoires. Il est important de documenter ces changements technologiques et comportementaux afin de comprendre les rapports nationaux de prise et d'effort présentés tous les ans (Tâche II-CE). Comme les CPC sont également tenues de déclarer les données sur la composition des flottilles (Tâche I-FC), il est recommandé que les CPC présentent un document du SCRS contenant le détail de la composition des flottilles, de l'échantillonnage, de la couverture, ainsi que de la méthodologie statistique visant à estimer la prise totale, la prise et l'effort et la prise par taille pour chacune des principales composantes de la flottille. Ce rapport devrait aussi signaler les limites et/ou les restrictions potentielles des données et des informations fournies que le SCRS ou le Secrétariat devrait prendre en compte dans toute nouvelle analyse qu'il entreprendra. L'incertitude des estimations d'effdis devrait être décrite. La documentation devrait clairement décrire les substitutions, les ratios d'extrapolation et la proportion des flottilles non-classifiées ("autres"). Il conviendrait d'explorer des méthodes additionnelles d'extrapolation des données (p. ex. Rubin 1976) en vue d'évaluer leur rigueur statistique. 2029

14 Il faudrait explorer des méthodes permettant de valider par croisement les estimations d'effdis (p. ex. données VMS). 4. Points limites de référence (LRP) et évaluation de la stratégie de gestion (MSE) Au début de la session, le Groupe a été saisi des conclusions provisoires d'un atelier récemment tenu par l'issf (normes de contrôle de la ponction et points de référence pour les ORGP thonières, San Diego, Californie (États- Unis), 6-8 mars 2013), étant donné que la réunion avait trait au point de l'ordre du jour consacré aux évaluations de la stratégie de gestion et aux points limites de référence. Il a été fait remarquer que les conclusions présentées étaient considérées provisoires étant donné que la version finale du rapport de l'atelier était toujours en cours d'élaboration. Les participants à l'atelier de l'issf ont indiqué que les stratégies de gestion incluaient le suivi, l'évaluation des stocks, les normes de contrôle de la ponction, les points de référence et les actions de gestion. L'atelier a été convoqué afin d'examiner l'état actuel de l'adoption de ces éléments dans le processus de prise de décision de cinq ORGP thonières et de formuler des recommandations en vue d'harmoniser et de faciliter le processus parmi les ORGP. Sur les cinq ORGP, la CCSBT a officiellement adopté une stratégie de gestion (procédure de gestion) aux fins de la prise de décision. Les autres quatre ORGP ont considérablement avancé pour identifier et tester les principaux éléments des stratégies de gestion, tels que les points de référence (limites et cibles) et les normes de contrôle de la ponction. Une partie de ce travail est essentiellement réalisé par les organes scientifiques des ORGP, parfois sans mandat officiel de leur Commission. L'atelier s'est concentré sur des questions fondamentales qu'il fallait garder à l'esprit lorsqu'on développait et testait des stratégies de gestion : données et modèles, traitement de F PME comme cible ou limite, test de la stratégie et mise en œuvre. L'atelier a fait remarquer que ces stratégies de gestion ont bien fonctionné dans d'autres instances au sein des pêcheries et qu'aucune contrainte technique ne s'oppose à leur introduction au sein des ORGP thonières. Cela pourrait se faire dans de nombreux cas avec des outils existants relativement simples. Le rapport contient des recommandations spécifiques sur les points de référence limites et cibles, les normes de contrôle de la ponction et d'autres considérations pour les évaluations des stratégies de gestion. Le rapport finalisé de l'atelier sera publié dans un proche avenir sur le site web de l'issf (http://iss-foundation.org). On a présenté un résumé de l'historique du développement des points de référence, des normes de contrôle de la ponction et de la mise en œuvre des évaluations de stratégies de gestion au sein de l'iccat. Dans la Convention de l ICCAT, la PME est le seul point de référence mentionné. L'approche de précaution a été élaborée après la création de l'iccat et le SCRS a discuté des avantages de disposer de points de référence cibles de précaution. Or, le texte de la Convention n'a pas encore incorporé d'autres points de référence et la PME a donc fonctionné comme le point de référence cible pour l'iccat depuis la création de la Commission. Mais, récemment, la Commission a adopté un cadre de prise de décision (Rec ) qui traite effectivement F PME comme une limite et non comme un point de référence cible. La présentation a été divisée en deux parties qui résumaient les activités de l'iccat et des ORGP thonières. La première partie récapitulait la mise en œuvre du processus de Kobe au sein de l'iccat. Le Comité permanent pour la recherche et les statistiques (SCRS) utilise une gamme de méthodes d évaluation des stocks, p.ex. ASPIC, modèles bayésiens de production excédentaire, Adapt, Multifan-CL, Stock Synthèse. Le principal objectif de gestion de l'iccat vise à maintenir les populations de thonidés et d'espèces apparentées à des niveaux qui permettront la prise maximale équilibrée. C'est pourquoi on a interprété, à l'origine, que la PME était une cible. L'ICCAT a été créée avant que n'apparaissent l'approche de précaution (PA) et l'approche écosystémique à la gestion des pêcheries (EAFM) ; c'est pourquoi aucune d'entre elles n'est citée dans la Convention. Malgré cela, la Commission a inclus les deux approches dans ses décisions. Les évaluations de stocks envisagent habituellement une gamme d'incertitudes et les évaluations sont réalisées pour les espèces accessoires (p.ex. oiseaux de mer, tortues marines) et les requins. Les plans de rétablissement sont en place pour le thon rouge de l Atlantique Est et Ouest (ainsi que pour d'autres stocks comme l'espadon de l'atlantique Nord) et les travaux axés sur le développement d'un modèle opérationnel dans le cadre du GBYP débuteront cette année. Cette année, des normes de contrôle de la ponction (HCR) incluant des points limites de référence (LRP) sont en cours d'élaboration pour l'espadon de l'atlantique Nord et le germon de l Atlantique Nord en réalisant une évaluation de la stratégie de gestion (MSE) en vue d'évaluer les performances des points de référence dans le cadre de HCR. 2030

15 La deuxième partie récapitulait la troisième réunion conjointe des ORGP thonières (Kobe III; Il a été reconnu qu'un processus de MSE devait être largement mis en œuvre au sein des ORGP thonières afin de mettre en œuvre une approche de précaution dans la gestion des pêcheries de thonidés. Kobe III a recommandé la création d'un Groupe de travail technique conjoint sur la MSE qui travaille par voie électronique, en premier lieu, afin de minimiser le coût de ses travaux. Trois activités sont en cours de réalisation, soit l'examen du cadre d'avis de Kobe, les outils de MSE et le recours à l'informatique parallèle et dématérialisée. L'ICCAT est en train de mettre au point des normes de contrôle de la ponction (HCR) qui incluent des points limites de référence (LRP) pour l'espadon de l'atlantique Nord (cf. Rec ) et le germon de l Atlantique Nord. Le SCRS a l'intention de développer ces HCR en évaluant les performances de points limites de référence alternatifs lorsque ceux-ci seront incorporés dans une HCR. Cette évaluation peut se faire par le biais de simulations effectuées avec un cadre MSE qui incorpore une gamme d incertitudes quantifiables. Le Groupe a signalé que l'incorporation d'une telle gamme d'incertitudes peut s'avérer compliquée lorsque plusieurs scénarios d'évaluation ou modèles d'évaluation sont utilisés. Dans de tels cas, il est avantageux de sélectionner des HCR et LRP qui sont solides face à tous les différents résultats des modèles et/ou des scénarios. Le Groupe a également discuté du fait que les points de référence pourraient ne pas rester constants dans le temps. Par le passé, le Groupe de travail sur les méthodes d'évaluation de l'iccat s'est penché sur la façon d'élaborer une HCR (Anon. 2011). La Commission a recommandé que les stocks soient gérés avec une "forte probabilité" (Rec ) de se trouver dans le quadrant vert du diagramme de phases de Kobe, même si la Commission n'a pas précisé ce niveau de probabilité. Le Groupe a indiqué que la probabilité que diverses stratégies de gestion, y compris des HCR alternatives, respectent cette Recommandation peut être évaluée avec des MSE. Le Groupe a constaté l'importance de clarifier et d'harmoniser la terminologie. Lors de la révision de la HCR "générique" décrite dans le WGSAM (Anon. 2011), le Groupe a fait remarquer que, même si le WGSAM s'est concentré sur B LIM (le niveau de biomasse en-dessous duquel la ligne de la HCR est établie à F=0 ; c.-à-d. le point où les pêcheries devraient être fermées), les gestionnaires pourraient être plus intéressés par le B thres (le point d'articulation de la HCR survenant à un niveau de réduction de la biomasse en-dessous duquel quelques réductions de F devraient s'amorcer). Cette HCR générique peut être paramétrée pour chaque stock (c.-à-d. en établissant des valeurs concrètes pour B LIM, B thres et la valeur cible de F). La Figure 1 fournit un exemple de la norme de contrôle de la ponction et des points limites de référence dont la réunion d'évaluation de l'espadon devra tenir compte au mois de juin. Le Groupe a convenu que la disponibilité des données orientera la décision, à savoir s'il convient de concevoir des normes de contrôle de la ponction "génériques" ou "spécifiques aux espèces". Pour les stocks dont les données sont insuffisantes, il pourrait s'avérer nécessaire de mettre au point une HCR basée uniquement sur F car il existe peu d'informations disponibles pour estimer la biomasse. Le Groupe a examiné la nécessité de définir un niveauseuil de la biomasse relative en-dessous duquel la HCR prévoit le déclenchement de mesures de gestion supplémentaires. Les MSE pourraient être utiles pour définir le niveau seuil approprié qui réduit les actions de gestion non nécessaires étant donné que le stock fluctue naturellement autour des niveaux de la PME, mais qui fournit cependant une protection adéquate pour maintenir les niveaux des stocks. On a avancé une autre possibilité : définir des HCR non-linéaires qui seraient déclenchées une fois que la biomasse chuterait en-dessous de celle qui permettrait la PME (p.ex. le niveau seuil serait égal à B PME ), mais qui nécessiteraient des réductions négligeables quand la biomasse se rapprocherait de B PME (p.ex. chute exponentielle) ; cette alternative ne nécessiterait pas un niveau seuil défini séparément. Une vaste gamme de HCR, y compris les formes des fonctions de réduction, peuvent être évaluées par la simulation. La Rec , en ce qui concerne le germon de l'atlantique Nord, prévoit un point limite de référence qui déclencherait un programme de rétablissement lorsque la biomasse chuterait en-dessous de celui-ci (B THRESHOLD ). Le Groupe a discuté de plans visant à mettre au point ce point pour le germon (SCRS/2013/033, SCRS/2013/034 et SCRS/2013/035), en ayant recours à la MSE, et de tirer ensuite parti de cette expérience pour élaborer des points limites de référence pour l'évaluation de l'espadon de l Atlantique Nord. Il a été fait remarquer que, dans des évaluations antérieures, l'avis de gestion sur l'espadon reposait sur les résultats d'aspic, alors que les travaux initiaux sur le germon se fonderaient sur Multifan. Outre le fait de considérer le temps nécessaire à l'élaboration de MSE spécifiques à l'espadon suffisamment à l'avance pour l'évaluation de cette année, le Groupe a considéré que l'applicabilité pourrait dépendre du type de modèle(s) utilisé pour l'avis de gestion. En d'autres termes, le modèle d'évaluation à utiliser dans la MSE ne devrait pas être plus complexe que le modèle opérationnel utilisé (Multifan pour le germon et ASPIC pour l'espadon). Si toute la gamme d'incertitudes va être incorporée, il faudra peut-être envisager des mélanges alternatifs flottille/pêcherie (p.ex. effort relatif de la palangre par opposition à la senne par opposition à la canne, ou autrement entre les flottilles dotées de différentes 2031

16 sélectivités d'âges, ce qui a une grande importance pour Multifan). Ceci a un prix élevé en termes de temps pour les scénarios multiples. Comme des mélanges alternatifs de pêcheries entraîneraient probablement des changements de sélectivité et par conséquent des changements de paramètres, il faut envisager prudemment la façon d'interpréter les résultats de multiples scénarios. Le WGSAM a reconnu qu'il s'agit d'un processus en cours et qu'il conviendrait de clarifier les objectifs à court terme et à long terme. Le Groupe a reconnu l'importance d'un engagement ferme envers ce processus, en incluant notamment des objectifs clairs et un financement proportionnel, afin de répondre aux besoins de gestion. Afin que cette tâche puisse être développée plus avant, le WGSAM a indiqué que des nouveaux projets étaient nécessaires ainsi qu'une collaboration internationale afin de s'avancer vers un cadre MSE commun. Les progrès obtenus dans l'étude de cas sur le germon ont été présentés comme un exemple du cadre MSE qui pourrait être élargi à d'autres stocks. En outre, le WGSAM a reconnu les mérites des travaux présentés, fruit de la collaboration entre le Secrétariat de l'iccat et des organisations nationales et il a encouragé une plus grande coopération interactive. 5. Identification des besoins en matière de recherche et des composantes clefs pour le plan stratégique de la science du SCRS, ainsi qu'identification des limitations en matière de capacité et des lacunes et de la façon dont celles-ci peuvent être résolues La planification stratégique est recommandée comme étant une approche structurée visant à orienter les travaux futurs du SCRS (rapport de 2011 du SCRS et réponses à la Rés de l'iccat sur la meilleure science disponible). Le document SCRS/2013/024 esquissait une approche pour identifier les principaux besoins et éléments en matière de recherche et une feuille de route pour mettre au point le plan stratégique du SCRS au titre de Le SCRS/2013/024 souligne que la planification stratégique traite de trois concepts fondamentaux : "Que faisons-nous?", "Pour qui le faisons-nous?" et "Comment excellons-nous?". En outre, les éléments clefs de la planification stratégique incluent une compréhension de la mission du SCRS (notre objectif), notre vision du futur, les valeurs que nous appliquerons à la conduite de notre travail, nos buts et stratégies pour y parvenir. Il a été souligné que la planification stratégique prévoit aussi une méthodologie visant à identifier la capacité critique et les lacunes en matière de données et à établir une priorité dans les activités de recherche pour les aborder. Le SCRS/2013/024 proposait une feuille de route et un calendrier pour l'élaboration du plan stratégique du SCRS , lequel prévoit de recruter un consultant qui serait chargé de fournir un cadre pour la méthodologie spécifique à appliquer lors de l'élaboration du plan stratégique et de consulter régulièrement les mandataires du SCRS aux fins de son examen en séance plénière du SCRS avant son examen et approbation par la Commission. Le Groupe de travail sur les méthodes a entériné le plan et a recommandé que son développement soit lancé, tel que décrit dans le SCRS/2013/ Discussion et amendement des termes de référence actuels de l'examen par les pairs de l'iccat, et accord sur un protocole de sélection des experts/examinateurs invités Dans sa Résolution sur la meilleure science disponible (Rés ), la Commission a préconisé le renforcement des mécanismes d'examen par les pairs au sein du SCRS, notamment la participation d'experts externes. Cette section présente un aperçu de l'évolution du processus d'examen par les pairs au sein du SCRS, ainsi que des discussions tenues par le Groupe sur la façon de renforcer le processus d'examen par les pairs et d'accroître la participation des experts externes, y compris l'élaboration de termes de référence. Lorsque l'on examine ces recommandations, il est important de noter que la Commission a également exhorté ses membres (dans la même Résolution, Rés ) à envisager d'élargir l'appui et les mécanismes financiers afin de mettre en œuvre le renforcement de l'examen par les pairs/de la participation d'experts externes, et d'appuyer les autres objectifs qui ont été identifiés dans cette Résolution. Afin de mettre en œuvre avec succès les démarches décrites dans ce rapport et atteindre les objectifs énoncés dans la Rés , il est essentiel que cet effort soit intégralement financé. Des examens par les pairs externes des travaux réalisés par les Groupes de travail du SCRS de l'iccat ont été menés à bien (Santiago et. al. 2013). À la réunion de 2010 de Kobe II, il a été conclu que l'examen par les pairs devrait être inclus dans toutes les évaluations scientifiques des ORGP thonières. Le rapport de l évaluation indépendante des performances de l ICCAT (Hurry et.al., 2008) indiquait : "Les analyses utilisées par le SCRS 2032

17 pour formuler un avis font l objet d un examen par des pairs qui comporte un processus rigoureux en trois étapes (groupes de travail/d évaluation, groupes d espèces, plénières du SCRS). La structure du processus, la diversité des participants/analyses et le grand nombre de personnes impliquées ne garantissent pas que des erreurs ne soient pas commises, mais ils donnent la garantie raisonnable que si des erreurs sont faites, elles seront découvertes, acceptées et corrigées." Les examens par les pairs des évaluations actuellement réalisées au sein de l'iccat suivent le processus adopté par le SCRS en Le SCRS a récemment recommandé de réaliser au moins deux examens par an in situ. Les examens visent à fournir un examen par les pairs additionnel au SCRS et à ses Groupes d'espèces afin de les orienter sur des améliorations à apporter aux évaluations de stocks. Pendant la réunion de 2012 du WGSAM, la question de l'examen par les pairs au sein de l'iccat a été une nouvelle fois discutée et l'on a rédigé des termes de référence pour la participation des experts externes en tant qu'examinateurs par les pairs dans les sessions d évaluation des stocks du SCRS (Anon. 2013). La terminologie concernant à la fois les experts invités et les examinateurs externes assistant aux Groupes de travail a été utilisée à l'iccat de façon à peu près interchangeable. C'est pourquoi le Groupe a examiné des documents présentés à la réunion qui traitaient de ces thèmes. Le WGSAM s'est penché sur le document SCRS/2013/23 qui présentait des termes de référence potentiels visant à faire la distinction entre les experts invités et les examinateurs externes. Le SCRS/2013/23 identifiait trois niveaux différents de fonctionnement de l'examen par les pairs. Premièrement, un examen par les pairs interne a lieu ; celui-ci est habituellement réalisé par les groupes de travail constitués de divers scientifiques nationaux qui examinent en "temps réel" et fournissent un avis critique sur le processus scientifique ; cet exercice est occasionnellement complété par la participation d'un expert externe. Deuxièmement, l'examen par les pairs externes se produit : les résultats de l'évaluation sont envoyés à des experts externes recrutés pour les examiner et "contrôler la qualité" ou bien les experts peuvent assister en qualité d'observateurs aux réunions du Groupe de travail et faire un rapport sur les résultats de la réunion. L'examen par les pairs externe peut aussi être réalisé par le biais d'une réunion conjointe avec un comité consultatif d'experts. Dernièrement, l'examen scientifique par les pairs peut s'effectuer en publiant les résultats scientifiques dans des revues examinées par des pairs ou en les présentant à des conférences internationales. Les experts invités prennent part au processus d évaluation en fournissant des informations et en formulant des avis sur la façon dont l'évaluation du stock peut être menée/améliorée/simplifiée au sein du processus d'évaluation. Un examinateur externe ne devrait pas, en théorie, prendre une part active à l'évaluation. Dans ce contexte, le Secrétariat fournit un processus potentiellement transparent pour sélectionner des experts à partir d'une liste d'experts maintenue par l'iccat. Un autre document (SCRS/2013/028) a été présenté au WGSAM. Celui-ci décrit des aspects du Centre pour les experts indépendants (CIE), processus financé par NOAA pour fournir des examens par les pairs au National Marine Fisheries Service (NMFS) des États-Unis. Le processus de sélection des examinateurs consiste à faire coïncider les compétences requises pour l'examen avec la capacité d'experts adéquats soumis à des contraintes établies afin d'éviter de sélectionner des candidats qui pourraient avoir des conflits d'intérêts. Ce processus est réalisé par le CIE indépendamment du client NMFS. Ce document présentait quelques-unes des leçons apprises par le CIE pour informer les discussions sur les termes de référence pour l'examen par les pairs de l'évaluation sur le germon qui se déroulera en 2013 et pour les futurs examens par les pairs. En outre, le document souligne explicitement quelques problèmes en suspens dans le processus d examen de l'iccat (Tableau 1). Après avoir examiné toute l'information fournie, des discussions ont eu lieu en vue de clarifier la séparation entre expert invité et rôles de l'examinateur au sein des performances du SCRS. En conséquence, des termes de référence ont été élaborés pour l'assistance de l'expert invité et de l'examinateur au processus d examen par les pairs du SCRS. Les termes de référence révisés pour un expert invité figurent à l'appendice 5 et les termes de référence révisés pour un examinateur externe se trouvent à l'appendice 6. Pareillement, il a été recommandé que le Président du Groupe de travail et le Président du SCRS proposent que des experts participent à l'évaluation de stock d'une espèce donnée. En outre, une liste d'examinateurs du CIE et d'autres ORGP thonières devra être disponible à l'iccat. Alternativement, l'iccat pourrait envisager une liste commune d'examinateurs élaborée avec d'autres ORGP thonières, comme il est suggéré dans le SCRS/2013/028. Le Groupe a examiné la possibilité d'utiliser les tarifs salariaux et les cadres temporels employés pour les examens du CIE afin d'orienter l'iccat lorsqu'elle fera appel à des examinateurs par les pairs. Actuellement, un examinateur du CIE reçoit un honoraire de $800,00 par jour et en moyenne 14 jours sont requis pour l'examen d'une évaluation de stock (deux jours de voyage, cinq jours de réunion, quatre jours de préparation et trois jours de rédaction du rapport) (SCRS/2013/028). 2033

18 7. Autres questions 7.1 Informatique en nuages à l'iccat Le "nuage" est un paradigme émergent dans la façon dont nous utilisons et mettons en commun les ressources informatiques (matériel, logiciel) et les informations (documents et données). Des ressources informatiques virtuellement illimitées (dépendantes du budget) et élastiques (on emploie ce dont on a besoin) peuvent être développées en quelques minutes au lieu de plusieurs semaines. La puissance informatique, la sécurité, la concurrence et le travail en commun sont pleinement optimisés sur le nuage. La formulation d'avis à la Commission dépend de plus en plus de l'emploi de méthodes informatiques intensives, telles que les simulations de Monte Carlo, le bootstrap utilisé lors de la création des matrices de stratégie de Kobe II, les scénarios de MCMC, les projections stochastiques employées dans les évaluations de stocks et, par exemple, l'évaluation des points limites de référence utilisant la MSE. Au cours de ces dernières années, il a souvent été impossible de réaliser ces analyses pendant les réunions des groupes de travail, qui ont en conséquence eu du mal à finaliser les rapports. C'est pourquoi le SCRS (guidé par le Groupe de travail sur les méthodes et le GT MSE des ORGP thonières) a recommandé au Secrétariat de l'iccat d'étudier l'utilisation de l'infrastructure du nuage (serveurs virtuels, informatique diffusée et parallèle, services simultanés, etc.). Le Secrétariat de l'iccat a présenté un plan de travail au Groupe qui contenait les directives de développement de l'infrastructure informatique en nuage de l'iccat (document à ajouter à Ce plan de travail décrit les études préliminaires réalisées, la topologie/le modèle de nuage proposé et les exigences (matériel et services du nuage) nécessaires à la première année de phase de développement (Phase 1). Une estimation des coûts escomptés a également été fournie. En outre, le Secrétariat a présenté au Groupe le plan de développement de 2013 (Phase 1) et finalement l'état de développement actuel. En résumé, il a été donné une description de la façon dont les serveurs du nuage sont administrés et utilisés, comment les services sont déployés et configurés (serveur web Apache, R-CRAN, RStudio server, etc.) et de la documentation déjà disponible (brefs didacticiels : administration du nuage, administration de Studio, guide d'utilisateur d'accès à distance). Tous ces documents devraient être publiés sur le site web en nuage de l'iccat qui est en cours de construction (http://tunalab.iccat.int). Le Secrétariat a prévu de lancer des tests importants sur l'infrastructure en nuage déjà développée, à partir de la réunion de préparation des données sur le germon de 2013 (Madrid, Espagne, avril). 7.2 Plan de travail futur Le Groupe de travail a discuté du futur plan de travail et a essentiellement retenu les actions suivantes : Le Groupe de travail sur les méthodes d'évaluation recommande d'examiner les protocoles et les algorithmes pour estimer la distribution de l'effort (5x5) pour la palangre (EFFDIS) et de les élargir à la senne et à la canne, que le Secrétariat est actuellement en train d'élaborer. Le Groupe de travail devrait également inclure des estimations de l'incertitude entourant ces produits. Il a été suggéré que les estimations publiées sur le site web de l'iccat incluent aussi une description détaillée des postulats estimés et de l'incertitude entourant ces produits afin que les utilisateurs potentiels se rendent compte de leurs limitations. La Commission espère un avis sur les mesures de gestion, lequel serait basé sur les risques, tel qu'énoncé dans la matrice de stratégie de Kobe II et inscrit dans son cadre de décision (Rec ). Un aspect important lié à la formulation de cet avis scientifique est la quantification adéquate de l'incertitude entourant l'état des stocks et les perspectives futures selon les scénarios d'options de gestion futures. Avec l'arrivée de modèles d'évaluation des stocks appliqués plus communément et hautement paramétrés, l'investissement informatique que représente la quantification de l'incertitude entourant l'état des stocks et les perspectives futures est assez lourd. Les autres ORGP thonières font le même constat et un certain nombre d'approximations pour quantifier les deux processus et l'incertitude par observation sont appliqués afin de formuler un avis de gestion basé sur les risques. Le Groupe de travail sur les méthodes d'évaluation devrait fournir une orientation sur l'évolution des méthodes et sur leur éventuelle harmonisation afin de solliciter la caractérisation de l'incertitude parmi les groupes d'espèces. 2034

19 Il conviendrait d'inclure dans l'ordre du jour de 2014 quelques-uns des thèmes horizontaux identifiés pendant le processus d'élaboration du plan stratégique du SCRS de 2013, notamment ceux liés à la participation et au renforcement des capacités ainsi qu'au contrôle de la qualité des évaluations des stocks et à l'avis de gestion. Le Groupe de travail sur les méthodes d'évaluation a reconnu que lors des récentes évaluations menées par le SCRS, on a eu tendance à utiliser de nombreuses méthodes de modélisation pour estimer l'état des stocks par rapport aux paramètres de conservation de l ICCAT. Même si le Groupe de travail sur les méthodes d'évaluation est d'avis que l'emploi de nombreuses méthodes constitue une bonne pratique, il s'est avéré que parfois des méthodes différentes ont donné des résultats incohérents, même s'ils étaient tout à fait plausibles. Il serait très utile de recevoir une orientation de la part du Groupe de travail sur les méthodes d'évaluation sur les meilleures pratiques permettant de réconcilier ou de combiner ces résultats (cf. p.ex. CIEM 2007). 7.3 Collaboration La Conférence mondiale sur les méthodes d évaluation des stocks pour les pêcheries durables (WCSAM) aura lieu à Boston (États-Unis) du 15 au 19 juillet La conférence fournira une enceinte pour les présentations sur l'application et l'avenir des méthodes d'évaluation des stocks. Elle envisagera des approches de stocks uniques pour les stocks riches et pauvres en données, ainsi que des approches plurispécifiques et basées sur l'écosystème. Celle-ci est organisée par des chercheurs issus d'une gamme d'institutions scientifiques et d'orgp de par le monde. La conférence sera précédée par un atelier de deux jours (15-16 juillet 2013) où des études sur l'application de méthodes d'évaluation des stocks à des jeux de données prédéfinis seront examinées. L'ICCAT participera activement à la WCSAM. Il est nécessaire de poursuivre la collaboration en participant aux réunions d'autres organisations régionales des pêcheries en ce qui concerne la mise en œuvre de la MSE en vue de renforcer l'utilisation de cet important outil permettant d'aborder les incertitudes et les risques associés aux modèles d'évaluation des stocks et de fournir un meilleur avis scientifique. La poursuite de la collaboration avec les groupes de travail de la CIEM, notamment aux fins de l'évaluation des espèces de requins, est considérée comme un domaine de travail important. Le WGSAM a reconnu les mérites des travaux présentés, fruit de la collaboration entre le Secrétariat de l'iccat et des organisations nationales et il a encouragé une plus grande coopération interactive. 8. Recommandations 1) Le WGSAM recommande que le SCRS et le Secrétariat collaborent avec d'autres ORGP thonières afin de mettre au point des protocoles communs pour les examens par les pairs, le cas échéant, notamment en ce qui concerne l'identification d'experts adéquats. 2) Il conviendrait d'évaluer des diagnostics pour les modèles d'évaluation. Les diagnostics appropriés pourraient varier entre les modèles d'évaluation, mais des diagnostics du modèle appropriés devraient être présentés pour contribuer à évaluer la qualité de l'avis de gestion découlant des évaluations. 3) Les rapports des examens par les pairs réalisés par les groupes de travail d'évaluation des stocks devraient être présentés sous la forme d'un document du SCRS, un résumé de l'examen par les pairs devant faire partie du rapport détaillé de la réunion d'évaluation. Les recommandations formulées dans le résumé de l'examen par les pairs doivent être incluses et prises en compte dans les futures sessions d'évaluation et pourraient ne pas être nécessairement abordées au cours de la même année que l'évaluation. 4) Les plans de travail à long terme devraient être rédigés pour le Groupe de travail conjoint sur la MSE des ORGP thonières afin de garantir le bon fonctionnement du Groupe et les plans devraient être disponibles sur la page web du Groupe (http://code.google.com/p/trfmo-mse/) 2035

20 5) Le WGSAM a entériné et recommandé le plan décrit dans le document SCRS/2013/024. Comme le plan stratégique pour la science du SCRS est en cours d'élaboration, les Groupes d espèces devraient inclure un point aux ordres du jour de leur réunion en 2013 pour évaluer les lacunes et les nécessités en matière de données et identifier les objectifs et les stratégies avant la plénière du SCRS afin de permettre au SCRS en 2013 de valider les objectifs et les stratégies et de se mettre d'accord sur sa mission, sa vision et les composantes des valeurs pour le plan stratégique de ) La rémunération des experts invités et des examinateurs externes pourrait se fonder sur les cadres temporels et les taux standard mis au point par le CIE. Les experts externes et les examinateurs invités devraient suivre les termes de référence stipulés par le WGSAM en ) Le WGSAM a formulé des recommandations spécifiques visant à améliorer l'estimation de la distribution spatiotemporelle de l'effort palangrier (EffDIS) et il recommande que ces efforts soient poursuivis. En outre, le WGSAM a recommandé que des efforts soient déployés pour élaborer des estimations d'effdis similaires pour les engins de canne et moulinet et de senne. 8) Pour les années où des évaluations de stocks doivent être réalisées, afin de renforcer l'assurance de qualité de l'avis scientifique, les groupes de travail seront tenus de préparer des plans de travail détaillés afin de fournir une orientation pour les préparatifs de la réunion et garantir une disponibilité complète et dans les délais des données requises et des entrées du modèle, ainsi que pour faciliter la coordination des responsabilités au sein du groupe de travail et/ou avec le Secrétariat. 9) Le WGSAM reconnaît l'importance de tenir compte des changements dans les opérations de pêche et les caractéristiques des principales flottilles de chaque CPC qui opère dans la zone de la Convention de l ICCAT, étant donné que ceux-ci affectent l'efficacité des flottilles pour capturer les espèces cibles et les espèces accessoires. Il est important de documenter ces changements technologiques et comportementaux afin de comprendre les rapports nationaux de prise et d'effort présentés tous les ans (Tâche II-CE). Comme les CPC sont également tenues de déclarer les données sur la composition des flottilles (Tâche I-FC), il est recommandé que les CPC présentent un document du SCRS contenant le détail de la composition des flottilles, de l'échantillonnage, de la couverture, ainsi que de la méthodologie statistique visant à estimer la prise totale, la prise et l'effort et la prise par taille pour chacune des principales composantes de la flottille. Ce rapport devrait aussi signaler les limites et/ou les restrictions potentielles des données et des informations fournies que le SCRS ou le Secrétariat devrait prendre en compte dans toute nouvelle analyse qu'il entreprendra. 9. Adoption du rapport et clôture Le rapport a été adopté pendant la réunion. Le Président a remercié les participants et le Président du WGSAM pour le travail intense accompli. La réunion a été levée. Références Anon Report of the 2010 ICCAT Working Group on Stock Assessment Methods (Madrid, Spain, April 21 to 23, 2010). Collect. Vol. Sci. Pap. ICCAT, 66(3): Anon Report of the 2012 ICCAT Working Group on Stock Assessment Methods (Madrid, Spain, April 16 to 20, 2012). Collect. Vol. Sci. Pap. ICCAT, 69 (in press). Hurry, G.D., Hayashi, M. and Maguire, J.J Report of the Independent Review. International Commission for the Conservation of Atlantic Tunas. Unpublished ICCAT report, ICCAT, Madrid. 105pp ICES, Report of the Study Group on Risk Assessment and Management Advice. ICES CM 2007/RMC:02. Neilson, J., Arocha, F., Calay, S., Mejuto, M., Ortiz, M., Scott. G., Smith, C., Travassos, P., Tserpes, G. and Andrushchenko, I The Recovery of Atlantic Swordfish: The Comparative Roles of the Regional Fisheries Management Organization and Species Biology, Reviews in Fisheries Science, 21:2, Rubin, D.B Inference and Missing Data. Biometrika 63: Santiago, J., Scott, G.P., Pereira, J.G Implementation of best science in the SCRS. ICCAT Collect. Vol. Sci. Pap. ICCAT, 69 (in press). 2036

21 INFORME DE LA REUNIÓN DE 2013 DEL GRUPO DE TRABAJO SOBRE MÉTODOS DE EVALUACIÓN DE STOCKS (Madrid, España 11 a 15 de marzo de 2013) 1 Apertura, adopción del orden del día y disposiciones para la reunión La Dra. Pillar Pallarés, Secretaria Ejecutiva Adjunta de ICCAT, inauguró la reunión y dio la bienvenida a los participantes. La reunión estuvo presidida por el Dr. Paul de Bruyn. El Dr. de Bruyn dio la bienvenida a los participantes en el Grupo de trabajo, examinó los objetivos de la reunión y procedió a revisar el Orden del día que fue adoptado con pequeños cambios (Apéndice 1). La lista de participantes se adjunta como Apéndice 2. La lista de documentos presentados a la reunión se adjunta como Apéndice 3. Los siguientes participantes actuaron como relatores de las diversas secciones del informe: Sección Relatores 1 P. de Bruyn 2 R. Coelho and J. Ortiz de Urbina 3 S. Cass-Calay 4 H. Arrizabalaga, G. Merino, D. Die and C. Brown 5 V. Ortiz de Zarate 6 G. Scott 7 N. Abid and C. Palma 8-9 P. de Bruyn 2 Diagnósticos del modelo En el documento SCRS/2013/025 se resumen los diagnósticos del modelo comunes disponibles para Stock Synthesis y se describe su interpretación. Se describen ejemplos de especificaciones erróneas del modelo y se ilustran los diagnósticos resultantes. Los autores también proporcionan un marco para facilitar una evaluación completa y eficaz de los diagnósticos del modelo. El Grupo reconoció la importancia y utilidad de dichos diagnósticos, e indicó que la mayor parte de las herramientas y diagnósticos presentados serían similares y útiles para la mayoría de los modelos no lineales y no necesariamente específicos para SS3. El Grupo debatió la importancia de la matriz de correlación de parámetros estimados y algunas cuestiones relacionadas. Se planteó que en la matriz de correlación podría haber una dependencia entre parámetros. Aún así, se indicó que el examen de las correlaciones estimadas entre parámetros supone un importante diagnóstico y se recomendaron más trabajos en este sentido. En el Apéndice 4 se presenta un marco para diagnósticos del modelo en SS3. En el documento SCRS/2013/027 se presentaban una variedad de métodos para los análisis exploratorios de datos y la evaluación de la bondad del ajuste, así como para la inspección de diagnósticos con miras a su utilización en las evaluaciones de stocks. La intención es establecer una lista de un conjunto de métodos comunes que puedan utilizarse para una gama de modelos de evaluación de stock (desde modelos simples a modelos complejos) y para diferentes requisitos de datos (por ejemplo, biomasa, talla y edad). Se estableció un listado con una gama de métodos para mostrar cómo podían implementarse en R, para posteriormente debatir las consecuencias para las evaluaciones de stock y las estrategias que puedan resolver los problemas identificados. Una recomendación general del Grupo fue que todos los documentos de evaluación de stock presentados al SCRS de ICCAT deberían proporcionar diagnósticos adecuados para los modelos utilizados, reconociendo que aunque diferentes métodos y modelos podrían tener diagnósticos ligeramente diferentes, en general numerosos diagnósticos diferentes podrían ser comunes a todos los métodos. 2037

22 El Grupo indicó la necesidad de que se disponga de diagnósticos en todos los modelos pero, teniendo en cuenta en todo momento que las herramientas de diagnóstico pueden ser muy voluminosas y que trabajar con ellas puede requerir una cantidad considerable de tiempo, se recomendó que, a nivel general, dichas herramientas deberían ser genéricas y flexibles. Se sugirió que dichos diagnósticos podrían desarrollarse y utilizarse de forma experimental en 2013 durante las evaluaciones de los stocks de atún blanco y pez espada. Tras recibir una respuesta de los Grupos de estas especies, las herramientas de diagnóstico y las recomendaciones podrían, si fuera necesario, revisarse para el año siguiente. El Grupo constató la importancia de que estos métodos y herramientas estén bien documentados y estén disponibles en un programa informático open source de acceso libre. Se mencionó que estas herramientas se están desarrollando e implementado actualmente en R, y que están disponibles a través de CRAN y otras plataformas (por ejemplo, el paquete cpue, R4SS, FLR, R4MFCL, etc.). El paquete "cpue" puede leer resultados de modelo de cualquier método de evaluación (por ejemplo, ASPIC, VPA, MFCL, SS3) y después ejecutar un conjunto estándar de herramientas de diagnóstico del modelo. Se proporcionaron al Grupo ejemplos de la utilización de estas herramientas, y en el documento se incluían algunos ejemplos gráficos. En el documentos SCRS/2013/030 se documentan los diagnósticos de modelo disponibles para el análisis de población virtual de rutina VPA-2BOX. El ejemplo específico utilizado en dicho documento es la evaluación de 2012 del stock de atún rojo del Atlántico oeste. Los diagnósticos del modelo incluían estadísticas de convergencia del modelo, sesgo y error estándar de los resultados del modelo, correlación y matrices de covarianza para los parámetros del modelo, análisis de sensibilidad de índices de abundancia y supuestos de ciclo vital y análisis de bootstrap para evaluar la robustez del modelo y estimar la tendencia central y la varianza de estimaciones de parámetros. Se debatía la utilidad de los diagnósticos a la hora de evaluar el funcionamiento del modelo durante la evaluación de stocks y se facilitaban recomendaciones para futuras mejoras. El Grupo resaltó la importancia de dichas herramientas y debatió, en general, el procedimiento de muestreo repetitivo (bootstrapping) y su potencial para diagnosticar problemas con infracciones en los supuestos del modelo. El Grupo también indicó que los análisis de datos preliminares y los procesos de limpieza de datos siguen siendo procedimientos importantes. 3 Examen de los métodos actuales de ICCAT para estimar EffDIS En el documento SCRS/2013/021 se presentaba una revisión de los métodos anteriores de estimación del esfuerzo de palangre total. La estimación más reciente adoptada por el SCRS se realizó en 2009 y utilizó las nueve especies principales de túnidos y especies afines para obtener capturas nominales globales de Tarea I (en peso) y las CPUE a partir de estadísticas parciales de captura y esfuerzo (Tarea II). En 2011, la Secretaría presentó una propuesta basándose en una metodología similar, pero desglosando los cálculos del esfuerzo básico en dos zonas principales (ATL/MED) con el fin de reducir los efectos secundarios de las escasas estadísticas de captura y esfuerzo disponibles para el Mediterráneo. Las estimaciones de esfuerzo de palangre (LL) global se incrementaron considerablemente (>10%) en los años más recientes. El SCRS consideró que este nuevo enfoque era una importante mejora, pero al mismo tiempo indicó que tienen que revisarse las estadísticas de captura y esfuerzo de Tarea II de palangre de importantes países del Mediterráneo. El supuesto base del modelo actual considera que las tasas de captura son equivalentes a nivel parcial y global. Al comparar los resultados con estimaciones anteriores (obtenidas durante reuniones intersesiones del Subcomité de ecosistemas en 2007 y 2008), los resultados globales no mostraban diferencias importantes. Sin embargo, a niveles más disgregados las diferencias eran mayores para algunos pabellones. En la mayoría de los casos, las importantes variaciones relativas se suelen asociar con diferentes mejoras y correcciones de algunos conjuntos de datos específicos. La distribución geográfica global muestra un pequeño incremento cerca de las aguas de Venezuela debido a las diferentes correcciones realizadas en las distribuciones espaciales de estadísticas de captura y esfuerzo de Venezuela durante varios años. Se presentan las áreas potenciales de mejora de la estimación de EffDIS en el futuro con el fin de impulsar los debates en la reunión WGSAM de ICCAT de 2013 sobre modos de mejorar la información EffDIS. 2038

23 El Grupo reconoció que EffDIS es una herramienta valiosa que permite visualizar y cuantificar la distribución espacial y temporal del esfuerzo de palangre. Ha sido utilizado a menudo por el Subcomité de ecosistemas, sobre todo para evaluar el impacto de las pesquerías de ICCAT en especies capturadas de forma fortuita, entre ellas las aves marinas. En el pasado, ha generado preocupación el método utilizado para calcular estas estimaciones, y se requieren mejoras en la metodología. Se describen las mejoras recientes en los conjuntos de datos, lo que incluye la estratificación por mes y la inclusión de flotas que redujeron el esfuerzo total de flotas sin clasificar ( Otras ) desde el 17% al 13%. El Grupo también debatió las recomendaciones para mejorar el conjunto de datos EffDIS, lo que incluye, cuando sea posible que: Las series de datos comunicadas por trimestres o años se vuelvan a presentar por mes. Las series de datos comunicadas sin información espacial o con estratificaciones superiores a 5ºx5º se vuelvan a presentar utilizando al menos la estratificación 5ºx5º. Se revisen las series de datos comunicadas sin esfuerzo de palangre para incluir el esfuerzo correspondiente. Se recuperen conjuntos de datos históricos, sobre todo anteriores a Se amplíe el número de especies utilizadas para extrapolar el esfuerzo utilizando la ratio de T2CE a T1NC para incluir las principales especies de tiburones (tintorera, marrajo sardinero y mako). Esta acción mejoraría la estimación del esfuerzo total. Nota: Las estimaciones de EffDIS son sensibles a la composición por especies. Una caracterización más completa de la composición por especies podría ser útil. Se establezca una diferenciación entre el esfuerzo pesquero del Atlántico y del Mediterráneo. Se revisen y mejoren las series más importantes de captura y esfuerzo de Tarea II del Mediterráneo. Se investigue la posibilidad de integrar en el análisis metadatos relacionados con la estrategia de la flota en función de la especie objetivo (con el fin de identificar perfiles de flota) para conseguir proyecciones más precisas de esfuerzo relativo de flotas pesqueras. Se examine la idoneidad de supuestos para extrapolar flotas (por ejemplo, se asume que Estados Unidos y Japón comunican toda la información y actualmente y sus datos no han sido extrapolados). La Secretaría indicó que las mejoras a EffDIS comenzaron en 2007 y están realizándose. Se ha avanzado mucho en algunas de las mejoras (a saber, la mayor parte de los datos de captura y esfuerzo se presentan ya por mes y cuadrículas de 5ºx5º), aunque otras están pendientes. El Grupo convino en que las recomendaciones enumeradas antes son apropiadas e importantes. Además, el Grupo sugirió lo siguiente: Deben realizarse esfuerzos para desarrollar estimaciones EffDIS similares para los artes de cebo vivo y de cerco. Esto resultaría especialmente útil para las evaluaciones de vedas espacio-temporales. El Grupo de trabajo reconoce la importancia de tener en cuenta los cambios en las operaciones pesqueras y en las características de las flotas principales de cada CPC que opera en la zona del Convenio de ICCAT, ya que éstas afectan a la eficacia de las flotas a la hora de capturar especies objetivo y especies de captura fortuita. La documentación de estos cambios tecnológicos y de estrategia es especialmente importante para entender los informes nacionales de captura y esfuerzo que se presentan anualmente (Tarea II-CE). Teniendo en cuenta también que las CPC tienen que comunicar datos de composición de la flota (Tarea I-FC), se recomienda que las CPC presenten un documento SCRS con información detallada sobre la composición de la flota, muestreo, cobertura y la metodología estadística utilizada para estimar la captura total, captura y esfuerzo y captura por talla para cada uno de los principales componentes de la flota. En este informe deberían comunicarse también las limitaciones y/o restricciones potenciales en los datos y en la información que se prorporciona, para que se tengan en cuenta en cualquier análisis ulterior realizado por la Secretaría o por el SCRS. Debería caracterizarse la incertidumbre asociada con las estimaciones EffDIS. En la documentación deberían describirse claramente las sustituciones, las ratios de extrapolación y la proporción de flotas sin clasificar ( Otras ). Deberían explorarse métodos adicionales para extrapolar los datos (por ejemplo, Rubin 1976) con el fin de evaluar su rigor estadístico. Deberían explorarse métodos para realizar verificaciones cruzadas de las estimaciones de EffDIS (por ejemplo, datos VMS). 2039

24 4 Puntos de referencia límite (LRP) y evaluación de estrategias de ordenación (MSE) Al comienzo de la sesión, se presentaron al Grupo de métodos las conclusiones provisionales de las Jornadas de trabajo ISSF que celebraron recientemente (Normas de control de la captura y puntos de referencia para las OROP de túnidos, San Diego, California, Estados Unidos, 6-8 de marzo de 2013), ya que la reunión estaba relacionada con el punto del orden del día Puntos de referencia límite (LRP) y evaluación de estrategias de ordenación (MSE). Se constató que las conclusiones presentadas se consideraban provisionales, ya que se estaba procediendo a una revisión del informe de las Jornadas. Los participantes en las Jornadas ISSF indicaron que las estrategias de ordenación incluían seguimiento, evaluación de stock, normas de control de la captura, puntos de referencia y acciones de ordenación. Las Jornadas se convocaron para revisar el estado actual de adopción de estos elementos en el proceso de toma de decisiones de cinco OROP de túnidos y para formular recomendaciones con miras a armonizar y facilitar el proceso entre las OROP. De las cinco OROP, la CCSBT ha adoptado oficialmente una estrategia de ordenación (procedimiento de ordenación) para la toma de decisiones. Las otras cuatro OROP de túnidos están progresando mucho en la identificación y prueba de elementos clave de estrategias de ordenación, como puntos de referencia (límite y objetivo) y normas de control de la captura. Parte de este trabajo lo están desarrollando sobre todo los organismos científicos de las OROP, a veces sin el mandato formal de las Comisiones. Las Jornadas se centraron en temas clave que deben tenerse en cuenta a la hora de desarrollar y probar estrategias de ordenación: datos y modelos, el tratamiento de F RMS como objetivo o como límite, la prueba de la estrategia y la implementación. En las Jornadas se indicó que estas estrategias de ordenación habían funcionado bastante bien en otras instancias al aplicarlas a las pesquerías, y que no existe ningún tipo de restricción técnica para desarrollarlas en las OROP de túnidos. En muchos casos podría hacerse de un modo relativamente sencillo, con las herramientas existentes. En el informe se incluían recomendaciones específicas sobre puntos de referencia límite y objetivo, normas de control de la captura y consideraciones para las evaluaciones de estrategias de ordenación. El informe finalizado de las Jornadas se publicará en breve en la página web de ISSF (http://iss-foundation.org). Se presentó un resumen de la historia del desarrollo de los puntos de referencia, normas de control de la captura e implementación de evaluaciones de estrategias de ordenación en ICCAT. En el texto del Convenio de ICCAT el único punto de referencia mencionado es el rendimiento máximo sostenible (RMS). El enfoque precautorio se desarrolló tras la creación de ICCAT, y el SCRS ha debatido los beneficios de aplicar puntos de referencia objetivo precautorios. Sin embargo, el texto del Convenio no ha incorporado todavía otros puntos de referencia y, por tanto, el RMS ha funcionado como punto de referencia objetivo para ICCAT desde la creación de la Comisión. Pero recientemente la Comisión ha adoptado un marco de toma de decisiones (Rec ) que considera F RMS como un punto de referencia límite y no como punto de referencia objetivo. La presentación se dividía en dos partes que resumían las actividades de ICCAT y las de las OROP de túnidos. La primera parte resumía la implementación del proceso de Kobe en ICCAT. El Comité Permanente de Estadísticas e Investigación de ICCAT (SCRS) utiliza una gama de métodos de evaluación de stock, por ejemplo, ASPIC, modelos de producción excedente bayesianos, Adapt, Multifan-CL y Stock Shynthesis. El principal objetivo de ordenación de ICCAT es mantener las poblaciones de túnidos y especies afines en niveles que permitan capturas máximas sostenibles. Por lo cual, el RMS se interpretó originalmente como un objetivo. ICCAT se creó antes de que surgieran los conceptos de enfoque precautorio (PA) y enfoque de ordenación pesquera basada en el ecosistema (EAFM), por lo cual ninguno de ellos se menciona en el Convenio. Pero, a pesar de ello, la Comisión se ha embarcado en el proceso de incluir tanto el PA como la EAFM en sus decisiones. Las evaluaciones de stock consideran de forma rutinaria una serie de incertidumbres y se llevan a cabo evaluaciones de especies de captura fortuita (por ejemplo, aves marinas y tortugas marinas) y de tiburones. Se han establecido planes de recuperación para el atún rojo del Atlántico este y del Atlántico oeste (así como para otros stocks, como el pez espada del Atlántico norte) y este año han comenzado los trabajos para desarrollar un modelo operativo en el marco del GBYP. Este año se van a desarrollar normas de control de la captura que incluyen puntos de referencia límite (LRP) para el pez espada del Atlántico norte y el atún blanco del Atlántico norte, mediante evaluaciones de estrategias de ordenación (MSE) con el fin de evaluar el funcionamiento de los puntos de referencia como parte de las HCR. 2040

25 En la segunda parte se resumió la Tercera reunión conjunta de OROP de túnidos (Kobe III; Se reconoció que es necesario implementar ampliamente un proceso MSE en las OROP de túnidos para implementar el enfoque precautorio en la ordenación de pesquerías de túnidos. En Kobe III se recomendó la creación de un Grupo de trabajo técnico conjunto sobre MSE y que dicho Grupo trabajase electrónicamente, en un primer momento, para minimizar los costes de sus trabajos. Actualmente, se están realizando tres actividades, a saber, una revisión del marco de asesoramiento de Kobe, de herramientas MSE y la utilización de una computación en paralelo y en nube. ICCAT está desarrollando HCR que incluyen puntos de referencia límite (LRP) para el pez espada del Atlántico norte (véase la Rec ) y para el atún blanco del Atlántico norte. El SCRS prevé desarrollar estas HCR mediante la evaluación del funcionamiento de puntos de referencia límite alternativos al incorporarlos a una HCR. Esta evaluación puede realizarse mediante simulaciones realizadas con un marco de MSE que incorpore una gama de incertidumbres cuantificables. El Grupo debatió el hecho de que la incorporación de dicha gama de incertidumbres podría complicarse cuando existen múltiples escenarios de evaluación o cuando se utilizan varios modelos de evaluación. En estos casos, sería conveniente seleccionar normas de control de la captura y puntos de referencia límite que sean robustos frente a diferentes resultados de escenarios y/o modelos. El Grupo también debatió el hecho de que los puntos de referencia no permanecen constantes en el tiempo. El WGSAM de ICCAT debatió en el pasado el modo de desarrollar una HCR (Anon. 2011). La Comisión ha recomendado que los stocks se gestionen de tal modo que exista una elevada probabilidad (Rec ) de que se sitúen en el cuadrante verde del gráfico de fase de Kobe, aunque la Comisión no especificó el nivel de probabilidad. El Grupo indicó que las probabilidades de que diferentes estrategias de ordenación, lo que incluye HCR alternativas, puedan cumplir esta recomendación pueden evaluarse mediante evaluaciones de estrategias de ordenación (MSE). El Grupo indicó la importancia de aclarar y armonizar la terminología. Al revisar la HCR genérica descrita en WGSAM (Anon 2011), el Grupo constató que, aunque el WGSAM se centró en B LIM (el nivel de biomasa por debajo del cual la línea HCR se establece en F=0; a saber, el punto en el que deberían cerrarse las pesquerías), los gestores podrían estar más interesados en B thres (el punto de inflexión de la HCR que se produciría en un nivel de reducción de biomasa por debajo del cual deberían comenzar algunas reducciones de F). Esta HCR genérica puede parametrizarse para cada stock (a saber, establecer valores concretos para las denominadas B LIM, B thres y el valor objetivo de F). En la Figura 1 se muestran ejemplos de norma de control de la captura y puntos de referencia límite que deberían considerarse en la reunión de evaluación de pez espada en de junio. El Grupo acordó que la disponibilidad de datos regirá la decisión de diseñar normas de control de la captura genéricas o específicas de las especies. Para algunos stocks con pocos datos podría ser necesario desarrollar una HCR basada solo en F porque existen pocos datos disponibles para estimar la biomasa. El Grupo debatió la necesidad de definir un nivel de umbral de biomasa relativa por debajo del cual la HCR indique que tienen que iniciarse medidas de ordenación adicionales. Las evaluaciones de estrategias de ordenación podrían ser útiles a la hora de definir un nivel de umbral apropiado que minimice las acciones de ordenación innecesarias, ya que el stock fluctúa naturalmente en torno a los niveles de RMS, pero que sin embargo proporcione una protección adecuada para mantener los niveles del stock. Otra posiblidad planteada fue definir HCR no lineales que se desencadenarían cuando la biomasa descienda por debajo de niveles que permiten el RMS (por ejemplo, el nivel de umbral sería igual a B RMS ), pero que requerirían reducciones insignificantes cuando el nivel de biomasa se sitúe cerca de B RMS (por ejemplo, descenso exponencial); esta alternativa no requeriría un nivel de umbral que se defina de forma independiente. Mediante la simulación pueden evaluarse una variedad de HCR, lo que incluye las formas de funciones de reducción. La Rec para el atún blanco del Atlántico norte solicitaba un punto de referencia límite que active un plan de recuperación si la biomasa desciende por debajo de dicho punto (B THRESHOLD ). El Grupo debatió planes para desarrollar esta cuestión con respecto al atún blanco (SCRS/2013/033, SCRS/2013/034 y SCRS/2013/035), mediante la utilización de una MSE y, posteriormente, para basarse en esta experiencia y preparar LRP para la evaluación de pez espada del Atlántico norte. Se constató que en pasadas evaluaciones el asesoramiento de ordenación para el pez espada se había basado en los resultados de ASPIC, mientras que el trabajo inicial para el atún blanco se basaría en MULTIFAN. Además de las consideraciones del tiempo necesario para desarrollar las MSE específicas del pez espada a tiempo para la evaluación de este año, el Grupo consideró que la aplicabilidad podría depender del tipo(s) de modelo(s) utilizado(s) para el asesoramiento en materia de ordenación. En otras palabras, que el modelo de evaluación que se va a utilizar en la MSE no debe ser más complejo que el modelo operativo utilizado (MULTIFAN para el atún blanco y ASPIC para el pez espada). Si se va a incorporar toda la gama de incertidumbres quizá tendrían que considerarse mezclas alternativas de flotas/pesquerías (por ejemplo: esfuerzo relativo de palangre vs. cerco vs. cebo vivo, o sino entre flotas con diferentes selectividades de edad, lo 2041

26 que reviste gran importancia para MULTIFAN). Esto tiene un alto precio en términos de tiempo para los ensayos múltiples. Dado que las mezclas alternativas de pesquerías darían lugar a cambios en la selectividad y, por tanto, a cambios en los niveles de referencia, debe considerarse con precaución el modo de interpretar los resultados de los múltiples escenarios. El WGSAM reconoció que es un proceso en curso, y que deben aclararse los objetivos a corto plazo y a largo plazo. El Grupo reconoció la importancia de un firme compromiso con este proceso, lo que incluye unos objetivos claros y una financiación acorde con dichos objetivos, con el fin de dar respuesta a las necesidades de ordenación. Para seguir avanzando en esta tarea, el WGSAM indicó que son necesarios nuevos proyectos y colaboraciones a nivel internacional para avanzar hacia un marco de MSE común. Se presentaron los progresos en el estudio de caso del atún blanco como ejemplo de marco MSE que podría aplicarse a otros stocks. Además, el WGSAM reconoció los beneficios del trabajo presentado como resultado de la colaboración entre la Secretaría de ICCAT y organizaciones internacionales e instó a que continuase la cooperación interactiva. 5. Identificación de necesidades de investigación y elementos clave para el plan estratégico para la ciencia del SCRS, así como identificación de las limitaciones de capacidad y de las lagunas y del modo en que pueden abordarse Se recomienda la planificación estratégica como un enfoque estructurado para orientar los trabajos futuros del SCRS (Informe del SCRS de 2011 y respuesta a la Resolución de ICCAT sobre la mejor ciencia disponible [Res ]. En el documento SCRS/2013/024 se describía un enfoque para identificar necesidades clave de investigación y sus componentes, así como una hoja de ruta para desarrollar el plan estratégico de ciencia del SCRS En el documento SCRS/2013/024 se indicaba que el plan estratégico aborda tres pilares básicos: Qué hacemos?, Para quién lo hacemos? Cómo destacamos? Además, los componentes clave de la planificación estratégica incluyen una comprensión de la misión del SCRS (o su finalidad), nuestra visión de futuro, los valores que aplicaremos al realizar nuestro trabajo, nuestros objetivos y las estrategias para alcanzarlos. Se indicó que la planificación estratégica proporciona también una metodología para identificar la capacidad crítica y lagunas de datos, así como para establecer prioridades en las actividades de investigación con el fin de abordarlas. En el documento SCRS/2013/024 se proponía un hoja de ruta y un calendario para desarrollar el Plan estratégico del SCRS para la ciencia para , que incluía la contratación de un asesor que proporcione un marco para la metodología específica que se tiene que aplicar para desarrollar el plan estratégico y una consulta regular y revisión por parte de los cargos del SCRS y las plenarias del SCRS antes de su presentación para que sea revisado y aceptado por Comisión. El WGSAM aprobó el plan y recomendó que el desarrollo del plan se inicie siguiendo la propuesta del documento SCRS/2013/ Debate y enmienda de los términos de referencia actuales para la revisión por pares en ICCAT, así como acuerdo sobre un protocolo para la selección de revisores/expertos invitados En la Resolución de ICCAT sobre la mejor ciencia disponible [Res ], la Comisión solicitó un reforzamiento de los mecanismos de revisión por pares en el SCRS, mediante la participación de expertos externos. Esta sección proporciona información sobre la evaluación del proceso de revisión por pares dentro del SCRS, así como los debates del Grupo sobre modos de reforzar el proceso de revisión por pares y mejorar la participación de expertos externos, lo que incluye el desarrollo de términos de referencia. Al considerar estas recomendaciones, es importante indicar que la Comisión también pidió a sus miembros (en la misma Res ) que considerasen la ampliación del apoyo y los mecanismos financieros para la implementación de este reforzamiento de la revisión por pares/participación de expertos externos, así como para respaldar los demás objetivos identificados en dicha Resolución. Para implementar adecuadamente las fases descritas en este informe y alcanzar los objetivos de la Res , es fundamental que este esfuerzo cuente con una financiación completa. Se han realizado revisiones por pares externas del trabajo de los grupos del SCRS de ICCAT (Santiago et al. 2013). En la reunión de Kobe II de 2010 se concluyó que la revisión por pares debería incluirse en todas la evaluaciones científicas de las OROP de túnidos. El Informe de la revisión independiente del desempeño de ICCAT (Hurry et al., 2008) indicaba que: Los análisis utilizados por el SCRS para formular su asesoramiento 2042

27 son objeto de una revisión por pares mediante un riguroso proceso que consta de tres fases (grupos de evaluación/grupos de trabajo - Grupos de especies - sesiones plenarias). La estructura del proceso, la diversidad de participantes/analistas y el gran número de personas implicadas no garantiza que no se cometan errores, pero proporciona una garantía razonable en cuanto a que si se cometen, éstos podrán detectarse, admitirse y corregirse. Actualmente, y desde la adopción del proceso en el SCRS de 2002, se realizan revisiones por pares de las evaluaciones. Recientemente, el SCRS recomendó que se realicen al menos dos revisiones por pares in situ por año. La finalidad de estas revisiones es proporcionar revisiones por pares adicionales al SCRS y a sus grupos de especies para orientales sobre mejoras en las evaluaciones de stocks. Durante la reunión de 2012 del WGSAM, se volvió a debatir la cuestión de la revisión por pares en ICCAT y se redactaron los términos de referencia (TOR) para la participación de expertos externos en calidad de revisores por pares en las evaluaciones de stock del SCRS (Anon. 2013). La terminología relacionada con expertos invitados y revisores externos que asisten a los Grupos de trabajo se ha utilizado de un modo indistinto en ICCAT. Por consiguiente, el Grupo revisó los documentos presentados a la reunión que abordaban estos temas. El WGSAM debatió el documento SCRS/2013/023 que presentaba términos de referencia potenciales para distinguir entre expertos invitados y revisores externos. En el documento se identificaban tres niveles diferentes del funcionamiento de la revisión por pares. En primer lugar está la revisión por pares interna, que generalmente realizan los grupos de trabajo formados por científicos de diferentes nacionalidades que en tiempo real revisan y facilitan aportaciones críticas/asesoramiento al proceso científico; a veces este proceso se complementa con la participación de un experto externo. En segundo lugar, la revisión por pares externa se facilita mediante el envío de los resultados de la evaluación a expertos externos contratados para su revisión y control de calidad o mediante la asistencia de expertos a las reuniones de los grupos en calidad de observadores que después informan de los resultados de la reunión. La revisión por pares externa podría realizarse también mediante una reunión conjunta con un comité experto asesor. Finalmente, la revisión por pares científica puede proporcionarse mediante la publicación de los resultados científicos en revistas objeto de revisión por pares o mediante su presentación a conferencias internacionales. Los expertos invitados participan en el proceso de evaluación, facilitando información y asesoramiento sobre cómo puede realizarse/mejorarse/agilizarse la evaluación de stock dentro del proceso de evaluación. En teoría, un revisor externo no puede desempeñar un papel activo en la evaluación. En este contexto, la Secretaría proporciona un proceso transparente potencial para la selección de expertos de una lista de expertos que mantiene ICCAT. Se presentó otro documento (SCRS/2013/028) al WGSAM. En dicho documento se describían aspectos del Centro de Expertos Independientes (CIE), un proceso financiado por NOAA que facilita revisiones por pares para el Servicio Nacional de Pesquerías Marinas de Estados Unidos (NMFS), en Estados Unidos. El proceso de selección de revisores consiste en hacer coincidir las capacidades requeridas para el revisor con las capacidades de expertos adecuado, el proceso está sujeto a las restricciones establecidas para evitar la selección de candidatos con potenciales conflictos de intereses. Este proceso lo realiza el CIE de un modo independiente del cliente, a saber, el NMFS. Este documento presentaba algunas de las lecciones aprendidas por el CIE para aportar información al debate sobre los términos de referencia de la revisión por pares de 2013 con vistas a la próxima evaluación de atún blanco en 2013 y de procesos futuros de revisión por pares. Además, el documento resalta explícitamente algunos problemas que están pendientes en el proceso de revisión de ICCAT (Tabla 1). Tras revisar toda la información facilitada, se inició un debate para aclarar la diferenciación de los roles del experto invitado y del revisor dentro del funcionamiento del SCRS: Por tanto se desarrollaron textos con términos de referencia para el experto invitado y la asistencia del revisor en el proceso de revisión por pares del SCRS. Los términos de referencia revisados para un experto invitado se incluyen en el Apéndice 5 y los términos de referencia revisados para un revisor externo en el Apéndice 6. Del mismo modo, se recomendó que la participación de expertos en una evaluación de stock de una especie determinada fuese propuesta por el presidente del Grupo de dicha especie y por el Presidente del SCRS. Además, se mantendrá en ICCAT una lista de revisores del CIE y de otras OROP a la que se podrá tener acceso. Como alternativa, ICCAT podría considerar la posibilidad de desarrollar, junto a otras OROP, una lista común de revisores, tal y como se sugiere en el documento SCRS/2013/028. El Grupo debatió la posibilidad de utilizar las tasas de salarios y los marcos temporales de las revisiones del CIE como orientación para ICCAT en la contratación de revisiones por pares. Acualmente, los revisores de CIE perciben un canon de asesoría de 800,00 $ por día y suelen contratarse para un promedio de 14 días para la revisión de la evaluación de stock (dos días de viaje, cinco días de reunión, cuatro días de preparación y tres días para redactar el informe) (SCRS/2013/028). 2043

28 7. Otros asuntos 7.1 Computación en nube de ICCAT La nube es un paradigma emergente en el modo en que utilizamos y compartimos los recursos informáticos (hardware y software) y la información (documentos y datos). Prácticamente ilimitados (dependiendo del presupuesto) y flexibles (se utiliza lo que se necesita), los recursos de computación pueden desarrollarse en minutos en vez de en semanas. La potencia de computación, la seguridad, la concurrencia y el trabajo compartido se optimizan totalmente en la nube. La provisión de asesoramiento a la Comisión depende cada vez de la utilización de métodos intensivos de computación, como las simulaciones de Monte Carlo, el bootstrapping utilizado al crear las matrices de estrategia de Kobe II, los ensayos MCMC, las proyecciones estocásticas utilizadas en las evaluaciones de stocks y, por ejemplo, la evaluación de puntos de referencia límite mediante las evaluaciones de estrategias de ordenación. Durante los últimos años, en numerosas ocasiones no ha sido posible realizar dichos análisis durante las reuniones de los Grupos, por lo cual a éstos les resulta difícil finalizar los informes. Por tanto el SCRS (con la orientación del Grupo de trabajo sobre métodos y el Grupo de trabajo de MSE de las OROP de túnidos) recomendó que la Secretaría de ICCAT estudie la utilización de la infraestructura en nube (servidores virtuales, computación distribuida y en paralelo, servicios simultáneos, etc.). La Secretaría de ICCAT presentó al Grupo un plan de trabajo que incluía las directrices para el desarrollo de una infraestructura ICCAT de computación en nube (documento que tiene que añadirse a Este plan de trabajo describe los estudios preliminares realizados, la topología/modelo de nube propuesto, y los requisitos (hardware y servicios en nube) necesarios para la fase de desarrollo del primer años (Fase 1). También se presentó una estimación de los costes previstos. Además, la Secretaría también presentó al Grupo el plan de desarrollo de 2013 (Fase 1) y, finalmente, el estado actual. En resumen, se describía cómo se utilizan y administran los servidores en nube, cómo se despliegan y configuran los servicios (Apache web server, R-CRAN, RStudio server, etc.) y la documentación (breves tutoriales: administración en nube; administración de Studio, guías de usuario para el acceso remoto) ya disponible. Todos estos documentos deberían publicarse en la página web ICCAT en nube en construcción (http://tunalab.iccat.int). La Secretaría prevé iniciar pruebas importantes de la infraestructura en nube ya desarrollada, que comenzarán con la reunión de preparación de datos sobre atún blanco de 2013 (Madrid, España, 22 a 26 de abril de 2013). 7.2 Plan de trabajo futuro El Grupo debatió el plan de trabajo futuro y propuso sobre todo las siguientes acciones: El WGSAN recomienda que se revisen los protocolos y algoritmos para estimar la distribución del esfuerzo (5x5) para el palangre (EffDIS) actualmente preparados por la Secretaría, así como su aplicación a los artes de cerco y cebo vivo,. El Grupo debería incluir también estimaciones de la incertidumbre en relación con estos productos. Se sugirió que las estimaciones publicadas en la página web de ICCAT deberían incluir también descripciones detalladas de los supuestos estimados y de la incertidumbre relacionada con estos productos para que los potenciales usuarios sean conscientes de sus limitaciones. La Comisión espera un asesoramiento sobre medidas de ordenación basado en el riesgo, tal y como se establece en la matriz de estrategia de Kobe II y en su marco de toma de decisiones (Rec ). Un aspecto importante a la hora de proporcionar dicho asesoramiento es cuantificar adecuadamente la incertidumbre sobre el estado del stock y las perspectivas futuras en escenarios de futuras opciones de ordenación. Con la llegada de modelos de evaluación de stock altamente parametrizados y más comúnmente aplicados, la inversión en computación para cuantificar la incertidumbre sobre el estado del stock y las perspectivas futuras es bastante elevada. Otras OROP de túnidos han tenido experiencias similares y se están aplicando una serie de aproximaciones para cuantificar ambos procesos y la incertidumbre por observación para desarrollar asesoramiento en materia de ordenación basado en el riesgo. El WGSAM debería proporcionar orientación sobre la evolución y la posibilidad de armonizar métodos para caracterizar la incertidumbre en los diferentes grupos de especies. 2044

29 Incluir en los puntos del orden del día de 2013 algunos temas horizontales identificados durante el proceso de elaboración del Plan estratégico para la ciencia del SCRS en 2013, sobre todo los relacionados con la participación, creación de capacidad y control de calidad de las evaluaciones de stock y del asesoramiento en materia de ordenación. El WGSAM reconoció que en las evaluaciones recientes realizadas por el SCRS se observa una tendencia a la utilización métodos de modelación múltiples para estimar el estado del stock en relación con los niveles de referencia de conservación de ICCAT. Aunque el WGSAN está de acuerdo en que la utilización de enfoques múltiples es un buena práctica, han surgido situaciones en las que los diferentes métodos han producido resultados no coherentes aunque igualmente plausibles. Sería muy útil disponer de orientaciones proporcionadas por el WGSAM sobre las mejores prácticas para conciliar o combinar dichos resultados (por ejemplo, ICES 2007). 7.3 Colaboración La Conferencia mundial sobre métodos de evaluación de stock para pesquerías sostenibles (WCSAM) se celebrará en Boston, Estados Unidos, del 15 al 19 de julio de La Conferencia proporcionará un foro para las presentaciones sobre la aplicación y el futuro de los métodos de evaluación de stock. Se considerarán enfoques de stock únicos para stocks con muchos datos y para stocks con pocos datos, y también enfoques basados en el ecosistema y para múltiples especies. La Conferencia la organizan investigadores de varias instituciones científicas y OROP de todo el mundo y estará precedida por unas Jornadas de dos días (15-16 de julio de 2013), en las que se examinarán estudios sobre la aplicación de métodos de evaluación de stocks a conjuntos de datos predefinidos. ICCAT participará activamente en la WCSAM. Se requiere la continuación de la colaboración mediante la participación en las reuniones otras organizaciones regionales de pesca (ORP) relacionadas con la implementación de la MSE para mejorar la utilización de esta importante herramienta a la hora de abordar incertidumbres y riesgos asociados con los modelos de evaluación de stocks y proporcionar un mejor asesoramiento científico. Se considera que la continuación de la colaboración con los grupos de trabajo ICES, en particular en lo que concierne a la evaluación de especies de tiburones, es un área de trabajo importante. El WGSAM reconoció los beneficios del trabajo presentado como resultado de la colaboración entre la Secretaría de ICCAT y organizaciones internacionales e instó a que continuase la cooperación interactiva. 8. Recomendaciones 1) El WGSAM recomienda que el SCRS y la Secretaría colaboren con otras OROP de túnidos para desarrollar protocolos comunes para las revisiones por pares, cuando proceda, sobre todo en lo que concierne a la identificación de expertos adecuados. 2) Deberían evaluarse diagnósticos para los modelos de evaluación Los diagnósticos adecuados podrían variar entre los modelos de evaluación, pero deberían presentarse diagnósticos de modelo apropiados para contribuir a evaluar la calidad del asesoramiento de ordenación procedente de las evaluaciones. 3) Los informes de revisión por pares realizados por los grupos de trabajo de evaluación de stock deben adoptar la forma de un documento SCRS, y deben incluir un resumen de la revisión por pares como parte del informe detallado de la reunión de evaluación. En el resumen de la revisión por pares se tienen que incluir las recomendaciones y éstas tienen que tenerse en cuenta en futuras sesiones de evaluación y podrían no tener que abordarse necesariamente en el mismo año que la evaluación. 4) Deberían redactarse planes de trabajo a largo plazo para el Grupo de trabajo conjunto MSE trfmo con el fin de garantizar que el Grupo funciona con eficacia, y dichos planes deberían publicarse en la página web del Grupo. (http://code.google.com/p/trfmo-mse/). 2045

30 5) El WGSAM aprobó y recomendó el plan esbozado en el documento SCRS/2013/024. Dado que actualmente se está desarrollando el Plan estratégico para la ciencia del SCRS, los grupos de especies deberían incluir en los órdenes del día de sus reuniones de 2013 un punto para evaluar las necesidades y lagunas en los datos e identificar objetivos y estrategias antes de las sesiones plenarias del SCRS para que en su reunión de 2013 el SRCS pueda validar los objetivos y estrategias y llegar a un acuerdo sobre la misión, visión y componentes de valor del plan estratégico ) La remuneración de los expertos invitados y revisores externos podría basarse en los marcos temporales y honorarios desarrollados por el CIE. Los expertos externos invitados y los revisores por pares deberían seguir los términos de referencia establecidos por el WGSAN en ) El WGSAM ha formulado recomendaciones específicas para mejorar la estimación de la distribución espacial y temporal del esfuerzo del palangre (EffDIS) y recomienda que se continúe con esta tarea. Además, el WGSAM recomienda que se realicen esfuerzos para desarrollar estimaciones EffDIs similares para los artes de cebo vivo y de cerco. 8) Para los años en los que se van a realizar evaluaciones de stock y con el fin de mejorar la garantía de calidad del asesoramiento científico, los grupos de trabajo tienen que preparar planes de trabajo detallados para proporcionar orientaciones para la preparación de la reunión y para garantizar la disponibilidad completa y puntual de los datos requeridos y de las entradas del modelo, así como para facilitar la coordinación de responsabilidades en el seno del Grupo de trabajo y/o con la Secretaría. 9) El WGSAM reconoce la importancia de tener en cuenta los cambios en las operaciones pesqueras y en las características de las flotas principales de cada CPC que operan en la zona competencia de ICCAT, ya que éstos afectan a la eficacia de las flotas a la hora de capturar especies objetivo y especies de captura fortuita. La documentación de estos cambios tecnológicos y de estrategia es especialmente importante para entender los informes nacionales de captura y esfuerzo que se presentan anualmente (Tarea II-CE) Teniendo en cuenta también que las CPC tienen que comunicar datos de composición de la flota (Tarea I- FC), se recomienda que las CPC presenten un documento SCRS con información detallada sobre la composición de la flota, el muestreo, la cobertura y la metodología estadística utilizada para estimar captura total, la captura y esfuerzo y la captura por talla para cada uno de los principales componentes de la flota. En este informe deberían comunicarse también las limitaciones y/o restricciones potenciales en los datos y en la información que se presenta, con el fin de que se tengan en cuenta en cualquier análisis ulterior realizado por la Secretaría o por el SCRS. 9. Adopción del informe y clausura El informe fue adoptado durante la reunión. El Presidente dio las gracias a los participantes y al coordinador del WGSAM por su eficacia y el gran trabajo realizado. La reunión fue clausurada. Referencias Anon Report of the 2010 ICCAT Working Group on Stock Assessment Methods (Madrid, Spain, April 21 to 23, 2010). Collect. Vol. Sci. Pap. ICCAT, 66(3): Anon Report of the 2012 ICCAT Working Group on Stock Assessment Methods (Madrid, Spain, April 16 to 20, 2012). Collect. Vol. Sci. Pap. ICCAT, 69 (3): Hurry, G.D., Hayashi, M. and Maguire, J.J Report of the Independent Review. International Commission for the Conservation of Atlantic Tunas. Unpublished ICCAT report, ICCAT, Madrid. 105pp ICES, Report of the Study Group on Risk Assessment and Management Advice. ICES CM 2007/RMC:02. Neilson, J., Arocha, F., Calay, S., Mejuto, M., Ortiz, M., Scott. G., Smith, C., Travassos, P., Tserpes, G. and Andrushchenko, I The Recovery of Atlantic Swordfish: The Comparative Roles of the Regional Fisheries Management Organization and Species Biology, Reviews in Fisheries Science, 21:2, Rubin, D.B Inference and Missing Data. Biometrika 63: Santiago, J., Scott, G.P., Pereira, J.G Implementation of best science in the SCRS. ICCAT Collect. Vol. Sci. Pap. ICCAT, 69 (5):

31 TABLE Tableau 1. Questions non résolues au sujet du système d'examen par les pairs de l'iccat identifiées par le SCRS (ICCAT 2013). TABLA Tabla 1. Cuestiones no resueltas sobre el sistema ICCAT de revisión por pares identificadas por el SCRS (ICCAT 2013). FIGURE Figure 1. Approche possible pour une norme de contrôle de la ponction et des points limites de référence pour l'espadon de l'atlantique Nord qui utilise le format développé par le Groupe de travail sur les méthodes d'évaluation du stock (Anon. 2011). S'appuyant sur la méthode actuellement utilisée pour formuler un avis de gestion, le Groupe de travail pourrait utiliser les approches de la dynamique de la biomasse et les données d'entrée disponibles afin d'identifier un point limite de référence de la biomasse associé à la plus faible biomasse jamais observée dans la série, et à un seuil associé à 0,8*B PME. Dans cet exemple, le B THRESHOLD se basait sur les travaux de Neilson et al. (2013) qui a travaillé avec les résultats de la VPA présentés dans l'évaluation des stocks de 2009 afin de déterminer l'impact de la variation du recrutement observée sur les points de référence. Ceux-ci sont fournis à titre d'exemple uniquement et ces points de référence pourraient être davantage affinés pendant la réunion de préparation des données de juin Suite à la réunion du mois de juin et avant la session d évaluation de septembre, le Groupe de travail réalisera une MSE en vue d'évaluer le caractère pertinent de ces points limites de référence proposés. FIGURA Figura 1. Un posible enfoque para una Norma de Control de la Captura y puntos de referencia límite para el pez espada del Atlántico norte que utilizan la plantilla desarrollada por el Grupo de trabajo sobre métodos de evaluación de stocks (Anon. 2011). Basándose en el método utilizado actualmente para facilitar asesoramiento en materia de ordenación, el Grupo de evaluación podría utilizar enfoques de dinámica de biomasa y los datos de entrada disponibles para identificar un punto de referencia límite de biomasa asociado con la biomasa más baja observada en la serie y un umbral asociado con 0,8*B RMS. En este ejemplo, B THRESHOLD se basó en los trabajos de Neilson et al. (2013) que trabajó con los resultados del VPA presentados en la evaluación de stock de 2009 para determinar el impacto de la variación de reclutamiento observada en los puntos de referencia. Esto se presenta sólo a título de ejemplo, ya que dichos puntos de referencia podrían definirse mejor durante la reunión de datos/métodos de junio de Tras la reunión de junio y antes de la reunión de evaluación de septiembre, el Grupo de trabajo realizará una evaluación de estrategias de ordenación (MSE) para evaluar la idoneidad de estos puntos de referencia límite propuestos. Appendice 1. Ordre du jour. Appendice 2. Liste des participants. Appendice 3. Liste des documents. APPENDICES Appendice 4. Cadre visant à faciliter une évaluation efficace et exhaustive des performances du modèle Stock Synthèse. Appendice 5. Termes de référence révisés pour un expert invité. Appendice 6. Termes de référence révisés pour un examinateur externe. 2047

32 APÉNDICES Apéndice 1 Orden del día. Apéndice 2 Lista de participantes Apéndice 3. Lista de documentos. Apéndice 4. Un marco para facilitar una evaluación exhaustiva y eficaz del funcionamiento del modelo Stock Shynthesis. Apéndice 5. Términos de referencia revisados para un experto invitado. Apéndice 6. Términos de referencia revisados para un revisor externo. 2048

33 Table 1. Unresolved issues about ICCAT peer review system identified by the SCRS (ICCAT 2013). Issue Quotation from ICCAT 2013 Need for consistency in quality of the review inconsistence in advice: where you get different peer reviewers from one assessment to another. Dual role of experts as reviewers and analysts A potential problem was that when a reviewer actively participates in a meeting, he/she will also have part ownership of the results from the meeting For example by participating in a data prep meeting they will have had responsibility for inputs into the assessment. Reviewers becoming part authors of the outputs in a multistage assessment process It was agreed that if there is a capacity problem then we there is a need to strengthen the stock assessment teams and not rely upon a peer reviewer to provide missing expertise. Figure 1. A possible approach for a Harvest Control Rule and Limit Reference Points for North Atlantic swordfish that uses the template developed by the Stock Assessment Methods Working Group (Anon. 2011). Building upon the method currently used to provide management advice, the Assessment Group could use biomass dynamics approaches and available input data, to identify a biomass limit reference point associated with the lowest biomass observed in the series, and a threshold associated with 0.8*B MSY ). In this example, the B THRESHOLD was based on the work of Neilson et al. (2013) who worked with VPA results presented in the 2009 stock assessment to determine the impact of observed recruitment variation on reference points. These are provided as examples only, and such reference points could be refined further during the June 2013 Methods/Data meeting. Following the June meeting and prior to the September assessment meeting, the Working Group will conduct an MSE to evaluate the suitability of these proposed limit reference points. 2049

34 Appendix 1 AGENDA 1. Opening, adoption of agenda and meeting arrangements 2. Model diagnostic discussion - Basic review of assessment models used by ICCAT - Common assessment model diagnostics - Development of protocols for presenting model diagnostics 3. Review of current ICCAT method for estimating EFFDIS - Discussions and development of improved methods for calculating EFFDIS 4. Limit Reference points and Management Strategy Evaluation - Discussion on generic LRPs and use in HCRs - How LRPs are used in other fora - LRP developments for ICCAT 5. Examples of simulation testing of LRPs using MSE and including biological information 6. Identification of key research needs and components for the SCRS Science Strategic Plan as well as identification of capacity limitations and gaps and how these can be addressed. 7. A discussion and amendment of the current ICCAT peer review TORs, as well as the agreement of a protocol for invited expert/reviewer selection. 8. Other matters 9. Recommendations 10. Adoption of the report and closure 2050

35 Appendix 2 LIST OF PARTICIPANTS SCRS CHAIRMAN Santiago Burrutxaga, Josu SCRS Chairman, Head of Tuna Research Area, AZTI-Tecnalia, Txatxarramendi z/g, Sukarrieta Bizkaia, Spain Tel: (Ext. 497); , Fax: , CONTRACTING PARTIES EUROPEAN UNION Arrizabalaga, Haritz AZTI - Tecnalia /Itsas Ikerketa Saila, Herrera Kaia Portualde z/g, Pasaia Gipuzkoa, Spain Tel: , Fax: , Coelho, Rui Portuguese Sea & Atmosphere Institute, I.P. (IPMA), Avenida 5 de Outubro, s/n, Olhão, Portugal Tel: , Fax: , Gaertner, Daniel I.R.D. UR nº 109 Centre de Recherche Halieutique Méditerranéenne et Tropicale, Avenue Jean Monnet, B. P. 171, Sète Cedex, France Tel: , Fax: , Merino, Gorka AZTI - Tecnalia /Itsas Ikerketa Saila, Herrera Kaia Portualde z/g, Pasaia - Gipuzkoa, Spain Tel: , Fax: , Ortiz de Urbina, Jose María Ministerio de Economía y Competitividad, Instituto Español de Oceanografía, C.O de Málaga, Puerto Pesquero s/n, Fuengirola, Malaga, Spain Tel: , Fax: , Ortiz de Zárate Vidal, Victoria Ministerio de Economía y Competitividad, Instituto Español de Oceanografía, C.O. de Santander, Promontorio de San Martín s/n, Santander, Cantabria, Spain Tel: , Fax: , Scott, Gerald P. AZTI-Tecnalia, SW 50 th Ct., Cooper City, Florida 33330, United States Tel: , Soto Ruiz, María Ministerio de Economía y Competitividad, Instituto Español de Oceanografía, c/corazón de María, 8, Madrid, Spain Tel: , Fax: , EQUATORIAL GUINEA Burgos Bizantino, Rodolfo Director General, Gabinete del Ministro, Ministerio de Pesca y Medio Ambiente, Zona Malabo II, Edificio Ministerial, Bloko Norte, Malabo Tel: , Nguema Asagono, Mariano Ministerio de Pesca y Medio Ambiente, Dirección General de Recursos Pesqueros, Zona Malabo II, Edificio Ministerial, Bloko Norte, Malabo Tel: , MOROCCO Abid, Noureddine Center Regional de L'INRH á Tanger/M'dig, B.P. 5268, Drabed, Tanger Tel: , Fax: , 2051

36 PANAMA Kwai Ben, Franklin Director General de Investigación y Desarrollo, Autoridad de los Recursos Acuáticos de Panamá (ARAP), INPESCA- Instituto Nicaragüense de la Pesca y Acuicultura, Avenida Justo Arosemena, Calle 45, Buena Vista, Edificio La Ribiera Tel: , Fax: , UNITED STATES Brown, Craig A. NOAA Fisheries Southeast Fisheries Center, Sustainable Fisheries Division, 75 Virginia Beach Drive, Miami, Florida Tel: , Fax: , Cass-Calay, Shannon NOAA Fisheries, Southeast Fisheries Center, Sustainable Fisheries Division, 75 Virginia Beach Drive, Miami, Florida Tel: , Fax: , Die, David Cooperative Unit for Fisheries Education and Research University of Miami, 4600 Rickenbacker Causeway, Miami, Florida Tel: , Fax: , Hobbs, James A. Associate Researcher Scientist, University of California, 1 Shields Ave, 1066 Academic Surge, Davis, California Tel: , Fax: , Lauretta, Matthew NOAA Fisheries Southeast Fisheries Center, 75 Virginia Beach Drive, Miami, Florida Porch, Clarence E. Chief, Sustainable Fisheries Division, Southeast Fisheries Science Center, National Marine Fisheries Service, 75 Virginia Beach Drive, Miami, Florida Tel: , Fax: , OBSERVERS FROM NON-GOVERNMENTAL ORGANIZATIONS Federation of Maltese Aquaculture Producers - FMAP Deguara, Simeon Research and Development Coordinator, Federation of Maltese Aquaculture Producers-FMAP, 54, St. Christopher St., VLT 1462 Valletta, Malta Tel: , Fax: , ************* ICCAT SECRETARIAT c/ Corazón de María, 8-6th -7th fl Madrid, Spain Tel: , Fax: , Pallarés, Pilar Ortiz, Mauricio Kell, Laurence Palma, Carlos 2052

37 Appendix 3 LIST OF DOCUMENTS SCRS/2013/023 Suggested revisions and clarifications to the Peer review process in ICCAT. de Bruyn, P.A., Santiago, J. and Kell, L. SCRS/2013/024 A Plan for The Plan. Santiago, J., de Bruyn, P., Arrizabalaga, H., Murua, H. and Scott, G. SCRS/2013/025 Model diagnostics for Stock Synthesis 3: Examples from the 2012 assessment of Spanish mackerel and cobia in the U.S. Gulf of Mexico. Cass-Calay, S.L., Tetzloff, J.C., Cummings, N.J. and Jeffery Isely, J. SCRS/2013/027 Examples of Stock Assessment diagnostic. Kell, L. SCRS/2013/028 SCRS/2013/030 SCRS/2013/033 SCRS/2013/034 SCRS/2013/035 Some key issues to get right in peer reviews of stock assessments: Lessons from the U.S. Center of Independent Experts. Die. D.J. and Shivlani, M. VPA-2Box model diagnostics used in the 2012 assessment of western Atlantic bluefin tuna (Thunnus thynnus). Cass-Calay, S.L. and Lauretta, M. An example of a management procedure based on a biomass dynamic stock assessment model. Kell, L., Merino, M., de Bruyn, P., Arrizabalaga, H., Muara, H. and Santiago, J. An example of conditioning an operating model using Multifan-CL. Kell, L., Merino, M., de Bruyn, P., Arrizabalaga, H., Muara, H. and Santiago, J. An example of a management strategy valuation of a management procedure based on a biomass dynamic stock assessment model. Kell, L., Merino, M., de Bruyn, P., Arrizabalaga, H., Muara, H. and Santiago, J. 2053

38 Appendix 4 A FRAMEWORK TO FACILITATE AN EFFICIENT AND COMPREHENSIVE EVALUATION OF STOCK SYNTHESIS MODEL PERFORMANCE 1. Does the model run? a. No use echo input to debug b. Yes continue 2. Are there any parameters on bounds? a. No continue b. Yes change starting values/change bounds/add priors rerun 3. Plot model output. Anything obviously wrong? Examples: productivity way too low, selectivity patterns that don t make sense, drastic decrease/increase in biomass in a single year, abnormal recruitment patterns (boom/bust). a. No continue b. Yes go through report file to diagnose (depends on problem) 4. Does the hessian converge? a. No check warning file, check estimated parameters in report file b. Yes continue 5. Examine parameter estimates. Plot parameter distributions along with starting values, bounds, and priors. Do parameters appear well estimated? a. No check bounds, check priors, check phase of estimation b. Yes continue 6. Look at trace plots of parameter estimates relative to phase of estimation? Do model parameters change much in the final phase? a. No continue b. Yes try alternative phases: for example, important scaling parameters like mean recruitment and catchability might be estimated in the first phase, recruitment deviates estimated added in the second phase, and selectivity added in the final phase. 7. Look at mean and standard deviation of estimated parameters. Is CV of estimated parameters less than 1? a. No is there data to inform parameter? i. No change bounds/add informative prior/fix parameter ii. Yes check correlation matrix b. Yes continue 8. Are any of the parameters highly correlated? a. No continue b. Yes why? Does one of the parameters require an informative prior? 9. Plot model fits to data and diagnostics. Is model fitting data reasonably? a. No diagnose the problem. b. Yes continue 10. As appropriate, tune input variance levels, sigmar, bias adjustment. Is model performance improved? a. No diagnose the problem, retune as needed. b. Yes continue 2054

39 11. Check for model stability to initial starting parameters using Jitter analysis. Does model converge to a global solution? a. No identify why. i. look at which likelihood components are changing ii. Evaluate the phases of estimation iii. Plot distribution of estimated parameters over all model runs b. Yes continue (try again with larger deviation from starting values) 12. Profile leading model parameters such as stock-recruitment parameters (steepness/r0) or natural mortality. Was the profile smooth? a. No Plot estimated parameters as a function of profiled leading parameter 1. Do any of the parameters hit bounds across the runs? Do any of the parameters bounce between alternative solutions? Do some parameters show similar patterns? a. Yes may not have enough data to inform all estimated parameters: add informative priors/reduce the number of estimated parameters. b. Yes Does profile show leading parameter is well estimated? Do the different data components show similar signals? i. No parameter may require informative prior or need to be fixed ii. Yes profile at finer scale 1. Does profile remain smooth? a. Yes continue 13. Evaluate model sensitivity to key model assumptions, data weighting choices, and alternative data inputs. Was model highly sensitive to any key model assumptions or certain data sources? a. No continue b. Yes Is model specified correctly? Are assumptions appropriate? Is model overparameterized? Should data be re-weighted? c. 14. Evaluate model sensitivity to the most recent years of data using a retrospective analysis. Did the retrospective analysis reveal any inconsistencies in the data? a. No continue b. Yes identify source of the retrospective pattern 15. Evaluate model uncertainty using bootstrap approach. Plot distribution of parameter estimates and derived quantities from bootstrapped runs. Compare MLE of parameter estimates to mean of bootstrap results. Are parameters or derived quantities well estimated when data is resampled? a. No do distributions show multi-modality or high proportion of bounding? i. Yes may not have enough data to inform all estimated parameters: add informative priors/reduce the number of estimated parameters. b. Yes continue 16. Optional: Evaluate model convergence using MCMC approach. Use standard approaches to evaluating MCMC results: look at trace plots/plot posterior distributions/compare MLE to mean of posterior distribution. Does MCMC converge on a single solution? Are MLEs of parameters/derived quantities similar to mean of posterior distributions? Poor performance has been noted in some SS models that otherwise appear well parameterized. However, poorly performing MCMC simulations may indicate a need to further explore model parameterizations. 2055

40 Appendix 5 Introduction REVISED TORS FOR AN INVITED EXPERT These Terms of Reference (TORs) have been developed as a guide to assist working groups in defining the work to be conducted by invited experts. It is acknowledged that the requirements may differ for each individual assessment session and these differences can be reflected in modifications to these TORs. The chair and the working group should clearly define in the work plan for the group the final TORs as well as tasks required from the invited expert which should then be approved by the SCRS. In this way, the invited experts should have a clear definition of what is required of them. In order to provide quality control feedback on this process, it would be beneficial if the working group could provide feedback on the work conducted by the invited expert. The proposed TORs for an invited expert are slight modifications of the TORs proposed by the WGSAM in These therefore are: 1) Prior to the meeting, the invited external expert(s) will be given access to previous reports of the working group. To the extent possible, the expert should attend both the data preparatory and stock assessment sessions. 2) The external expert (s) will be provided with the official data to be used in the stock assessment, and will be bound by any applicable confidentiality agreements that apply to participating members of the assessment working group. These data will be made available to the expert(s) at the same time they are available to the working group in general. 3) Fully participate in the discussions of the appropriate analyses to be conducted at the meeting including, but not limited to: The data and software available for analysis and based on this information the selection of the assessment model(s) to be used which is appropriate for the data available, model assumptions, biological parameters, selection of model run(s). As necessary conduct specific tasks or analyses as requested by the assessment group. When appropriate, suggest alternative assessment methods that could better characterize the dynamics of the stock. As appropriate participate in the development of the main conclusions of the stock assessment and management recommendations from the meeting. Participate in the identification of specific research needs for the future. 4) The comments and suggestions of the external expert will be taken into consideration by the working group during the stock assessment process and in the preparation of the meeting report. If a specific body of work has been conducted by the expert, they should fully document this work in an SCRS document, and provide at least a summary document as to their contribution to the meeting as described in the work plan. 2056

41 Appendix 6 REVISED TORS FOR AN EXTERNAL REVIEWER Introduction These Terms of Reference (TORs) are intended to provide a guideline for working groups to define the role of external reviewers contracted to review the outputs of the assessment working group. These TORs can and should be modified as appropriate to cover the diverse requirements of the individual assessment sessions. It is intended that advice provided by external reviewers will be taken into account in future stock assessment sessions in order to improve the assessment process. It is envisioned that the reviewer will provide a detailed peer review report that should be submitted as an SCRS document, as well as a summary of the peer review report to be included in the detailed report of the stock assessment. The role of the external reviewer is to: 1) Evaluate the adequacy, appropriateness, and application of data used in the assessment. 2) Evaluate the adequacy, appropriateness, and application of methods used to assess the stock and if appropriate recommend alternative approaches to be accomplished in the future. 3) Evaluate the methods used to estimate population benchmarks and stock status (e.g., MSY, F MSY, B MSY, or their proxies). 4) Evaluate the adequacy, appropriateness, and application of the methods used to evaluate future population status, given the commissions objectives. 5) Evaluate the adequacy, appropriateness, and application of methods used to characterize the uncertainty in estimated parameters. Comment on whether the implications of uncertainty in technical conclusions are clearly stated. 6) Comment on whether the stock assessment results are clearly and accurately presented in the detailed report of the stock assessment. 7) Comment on potential improvements on the stock assessment SCRS process (CPC participation, transparency, objectivity, documentation, uncertainty characterization, etc.) as applied to the reviewed assessments. 8) Comment on the adequacy of the workplan for the assessment and whether it was adequately addressed by the data or assessment working groups. 9) Consider the research recommendations provided by the working group and suggest any additional recommendations or prioritizations warranted. Clearly denote research and monitoring needs that could improve the reliability of future assessments. Recommend an appropriate interval for the next assessment considering control rules or management strategy in effect. 10) Prepare a Peer Review Report which should specifically address each TOR. Complete and submit this Peer Review Report along with a summary no later than the two weeks after completion of the assessment meeting. 2057

42 SCRS/2013/023 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) SUGGESTED REVISIONS AND CLARIFICATIONS TO THE PEER REVIEW PROCESS IN ICCAT Paul de Bruyn 1, Josu Santiago 2 and Lawrence Kell 1 SUMMARY ICCAT has historically utilised a rigorous three stage internal review system to ensure the quality of its scientific advice to management. It has been noted, however, that the process would benefit from additional external review. Although ICCAT has put in place protocols to conduct these external reviews, the SCRS in 2012 requested clarification of the process. This document clarifies the terminology regarding the three separate forms of scientific peer review; clarifies and defines the roles of invited experts and external reviewers and proposes a transparent method for identifying and selecting external experts. It is the intention of this document to stimulate discussion surrounding these issues in order to improve and streamline the definitions and processes described here, to the benefit of the peer review system in ICCAT. RÉSUMÉ Historiquement, l'iccat utilise un système rigoureux de révision en interne en trois étapes afin de garantir la qualité de l'avis scientifique destiné à la gestion. On a cependant observé que le processus pourrait tirer profit d'un examen externe supplémentaire. Bien que l'iccat ait mis en place des protocoles pour réaliser ces examens externes, le SCRS a demandé en 2012 des éclaircissements concernant le processus. Le présent document précise la terminologie concernant les trois types d'examen scientifique externe par des pairs, précise et définit les rôles des experts invités et des examinateurs externes et propose une méthode transparente d'identification et de sélection des experts externes. Ce document vise à stimuler les discussions sur ces questions dans le but d'améliorer et de simplifier les définitions et les processus au profit du système d'examen par les pairs de l'iccat. RESUMEN ICCAT ha utilizado históricamente un riguroso sistema de revisión interna en tres fases para garantizar la calidad de su asesoramiento científico en materia de ordenación. Sin embargo, se ha constatado que el proceso se beneficiaría de una revisión externa adicional. Aunque ICCAT ha implementado protocolos para estas revisiones externas, en 2012 el SCRS solicitó aclaraciones para este proceso. En este documento se aclara la terminología relacionada con los tres tipos de revisión científica por pares, se aclaran y definen los papeles desempeñados por los expertos invitados y los revisores externos y se propone un método transparente para identificar y seleccionar a los expertos externos. La finalidad de este documento es fomentar el debate en torno a estos temas con miras a mejorar y racionalizar las definiciones y procesos descritos aquí, para beneficiar al sistema de revisión por pares de ICCAT. KEYWORDS Peer review, Invited expert, Reviewer, Terms of reference, Transparency, Best practice, Quality of scientific advice 1 ICCAT Secretariat, C/ Corazón de Maria, 8, 6th floor, Madrid 2 AZTI Tecnalia, Txatxarramendi ugartea z/g Sukarrieta (Bizkaia), Spain; 2058

43 Introduction A topic that has been discussed in several forums is the subject of peer review of RFMO scientific outputs. The 2010 KOBE II meeting in Barcelona concluded with a recommendation that peer review should be included in all RFMO scientific assessments. This topic was again debated by the expert panel invited by the EU funded TXOTX project partners to London in November The expert panel agreed that peer review of scientific outputs should be conducted. They further concluded that If funds are available, peer reviews should be conducted when possible as although there are drawbacks to the process (delays in finalisation of outputs, increase in resources) the advantages of having these reviews outweigh the negatives. This review should be done in a structured way rather than an ad hoc basis. Where expert panels are introduced, these should comprise at least three members to avoid a dominance of one viewpoint or an impasse that could arise from having opposing views. (de Bruyn et al. 2011). The Peer review process can imply different things to different people. For the purposes of this document, we define peer review as having three different levels. Firstly, there is internal peer review, usually facilitated by working groups of diverse national scientists who in real time review and provide critical input/advice to the scientific process; on occasion this is supplemented by the participation of an external expert. Secondly, external peer review is facilitated by sending scientific outputs to contracted external experts for review and quality control or having the experts attend WG meetings as observers who then report on the meeting outcomes. External peer review may also be conducted through a joint meeting with an advisory expert panel. Lastly, scientific peer review may be facilitated by publishing scientific outputs in peer reviewed journals or presenting them at international conferences. An additional form of peer review covering performances reviews of the processes being carried out within an organisation/rfmo could be considered, but that is not the intention here. This latter definition ensures that things are being done correctly within an organisation, and is more a management topic and less a scientific issue. It should be remembered that in terms of the peer review process, quality is more important than quantity, and peer reviews should be conducted thoroughly and professionally even if the requirements for this ensure that they are not conducted very regularly. According to the Report of the Independent Performance Review of ICCAT (Hurry et al., 2008) the analyses used by the SCRS to formulate its advice are peer reviewed through a rigorous three stage process (working/assessment groups to species groups to SCRS plenary). The structure of the process, the diversity of the participants/analysts and the large number of people involved does not guarantee that errors will not be made, but it provides a reasonable assurance that if errors are made, they will be discovered, admitted, discussed and corrected. Apart from this internal system, external peer reviews of the work of ICCAT SCRS working groups have been conducted in the past and a protocol to conduct such reviews has been in place for some time (Santiago et al., 2013). For example, in 2003 the Albacore Species Group (SCRS/2003/113) and the Methods Working Groups (SCRS/2003/039) were both peer-reviewed under the ICCAT Stock Assessment Peer Review program. Other types of peer reviews, such as participation of external experts in Standing Committee on Research and Statistics (SCRS) meetings, publication of SCRS works in peer review journals and at world conferences have also been used. It is thus important at this stage to clarify what is meant by peer review as sometimes the distinctions between forms of peer review have been lost at the working group level. Forms of peer review at ICCAT Omitting the broader scale review of the RFMO processes (which has been conducted for ICCAT; Hurry et al. 2008), the peer review of assessments in ICCAT currently in place is the one adopted by the SCRS in 2002 [Appendix 4 of the 2002 SCRS Report, (Anon., 2003)] following the recommendations of the SCRS in 2001 [2001 Report of the Ad Hoc Working Group on SCRS Organization (Appendix 4 of the 2001 SCRS report) and in the 2001 SCRS Report itself (Anon., 2002)] after discussing the proposed methods of conducting peer review detailed in Restrepo (2000). The SCRS recommended conducting at least 2 in situ reviews per year. The purpose of the reviews is to provide additional scientific peer reviewed advice to the SCRS and its species groups for improvements in their stock assessments. Conducting an in situ review, wherein the reviewer/invited expert may provide working papers in advance of the session, actively participates in analysis and in report drafting, permits an immediate feedback to the working group and SCRS and facilitates the development of suggestions for future research. This format is thus, in the short-run, the method of peer review viewed most practical for ICCAT. 2059

44 Until now, both invited experts/wg discussions and external reviewers attending the WGs have been used somewhat interchangeably in ICCAT working groups although they have separate distinct purposes. Invited experts take part in the assessment process. They provide information and advice on how the stock assessment can be conducted/improved/streamlined DURING the assessment process, introducing new modelling approaches or new perspectives during a part of or for the whole process. The result is an improved product which hopefully has benefitted from the expertise of the invited person. An external reviewer, on the other hand, should in theory play NO ACTIVE PART in the assessment process. The reviewer in this case could participate in the WG in a more passive observer role or could be brought in once an assessment has been conducted and provide a review of the assessment products. The review then points out the deficiencies of the assessment and suggests ways in which it can be improved either before presentation to management or in the future. It is important that these clear differences are noted. Current Terms of reference During the 2012 meeting of the ICCAT Working Group on Stock Assessment Methods (WGSAM), the issue of peer review in ICCAT was discussed and Terms of Reference for the participation of external experts as peer reviewers in the SCRS stock assessment meetings were drafted (Anon. 2012): 1. Prior to the meeting, the external reviewer(s) will be given access to previous reports of the working group. 2. Fully participate in the discussions of the appropriate analyses to be conducted at the meeting including, but not limited to: The selection of the assessment model(s) to be used, model assumptions, biological parameters, selection of model run(s). When appropriate, suggest alternative assessment methods that could better characterize the dynamics of the stock. Participate in the development of the main conclusions of the stock assessment and management recommendations from the meeting. Participate in the identification of specific research needs for the future. 3. The comments and suggestions of the external reviewer will be taken into consideration by the Working Group during the stock assessment process and in the preparation of the meeting report. The external reviewer will prepare an independent report with recommendations to improve the assessment and the review processes which will be added to the meeting report as an annex upon its completion. These terms of reference clearly apply to invited experts. Despite the drafting of these revised TORs by the WGSAM, at the 2012 meeting of the SCRS it was requested that the group revisit the current TORs in order to clarify the role of an invited expert and to draw a distinction between invited experts and external peer review. It is thus the intention of this document to stimulate discussion on this matter and provide suggestions for possible amendments to the current TORs for invited experts as well as the peer review system in general. Possible future revisions to Peer Review TORs As stated above, the first step when revising the protocols for peer review is to clarify the roll or position of the contacted person/s. Is the expert required to assist with an assessment (i.e. to increase scientific capacity at a stock assessment meeting) or is the person required to review the outputs of a stock assessment session. To this end potential separate TORs are presented for the two different roles. Possible revised TORs for an invited expert The proposed TORs for an invited expert are slight modifications of the TORs proposed by the WGSAM in These therefore are: 1. Prior to the meeting, the external expert(s) will be given access to previous reports of the working group. 2. The external expert (s) will be provided with the official data to be used in the stock assessment, and will be bound by the same confidentiality agreements that apply to participating members of the assessment working group. These data will be made available to the expert/s at the same time they are available to the working group in general. 2060

45 3. Fully participate in the discussions of the appropriate analyses to be conducted at the meeting including, but not limited to: The data available for analysis and based on this information the selection of the assessment model(s) to be used which is appropriate for the data available, model assumptions, biological parameters, selection of model run(s). When appropriate, suggest alternative assessment methods that could better characterize the dynamics of the stock. Participate in the development of the main conclusions of the stock assessment and management recommendations from the meeting. Participate in the identification of specific research needs for the future. 4. The comments and suggestions of the external expert will be taken into consideration by the Working Group during the stock assessment process and in the preparation of the meeting report. The external expert will prepare an independent report with recommendations to improve the assessment which will be added to the meeting report as an annex upon its completion. As part of this report, the external expert/s may: comment on the appropriateness of the discussions and analyses held at the meeting in terms of the meeting's objectives; as already defined above during the working group discussion, suggest alternative assessment methods where better methods exist and suggest specific research for the future; not provide an evaluation as to whether the assessment should be accepted or not, although full quantification of the uncertainty around the assessment results is part of the assessment process and encouraged. As the expert will function as a member of the working group, he will have input in the discussions on the soundness of the conclusions and recommendations without needing to specifically address them here unless expressly requested to do so. Possible TORs for an external Reviewer 1. An external reviewer shall be selected who meets the following criteria; expertise, working knowledge, recent experience in one or more of the subject areas involved in the review. The subject areas used as criteria for selection of the reviewer shall be determined by the chair of the working group along with the Chairman of the SCRS and the ICCAT Secretariat. 2. The external reviewer shall be required to; read background material provided by the working group on the assessment in progress as well as any previous assessments on the stock produce a review report addressing the review terms of reference. These terms of reference shall be determined by the chair of the working group along with the chairman of the SCRS and the ICCAT secretariat but will include issues such as, quality of data used in the assessment, appropriateness of model applied for the assessment, appropriateness and completeness of quantification of uncertainty, quality and appropriateness of model diagnostics, soundness of the main conclusions and recommendations, alternative assessment methods where better methods exist and specific research for the future. An example of this type of TOR is provided in Appendix The external reviewer shall not necessarily be present at the actual stock assessment session, and could be contacted remotely to review the outputs. Should the chair of the working group, the SCRS chair or the ICCAT secretariat determine that the external reviewer s presence is necessary at the assessment session, he will play little active role in the discussions taking place at that meeting but will act as an observer who may comment or offer suggestions as necessary. 2061

46 Selection of invited experts or external reviewers The secretariat will keep a list of experts who have been agreed to participate in the ICCAT review system and who have been judged to have the necessary experience and expertise. This will allow selection of candidates as soon as the SCRS calendar has been agreed. Nominations from this list should be made by the chair of the working group, the SCRS chair or the ICCAT secretariat at which stage they should be distributed to members of the appropriate working group for comment. Selection of the final candidate/s shall be communicated by the chair of the working group who along with the SCRS chair and ICCAT secretariat will have the final say on candidate selection. Conclusions It is clear that ICCAT has put considerable effort into developing and modifying its peer review system. It has been complimented on the thoroughness of its three stage internal review process as well as its flexibility to modify the process as required, but it is acknowledged that external expertise can greatly improve scientific capacity which in turn will improve the scientific advice to management. The external review process, however, requires clarification, particularly as to the role of the external expert, either as an active participant in the scientific processes or as a reviewer of the scientific outputs. It is hoped that the potential TORs presented in this document will stimulate discussions as to how this clarification can best be conveyed to the external experts. A clear and transparent way in which external experts can be identified and selected has also been proposed, which is in accordance with how ICCAT currently operates, with minor clarifications. Beneficial discussions on this matter are also anticipated. References Anon Report for the Biennial Period, , part II, Vol. 2. ICCAT, Madrid, Spain. Anon Report for the Biennial Period, , part I, Vol. 2. ICCAT, Madrid, Spain. Anon Report of the 2012 Meeting of the ICCAT Working Group on Stock Assessment Methods. Unpublished ICCAT report, ICCAT, Madrid. 23pp De Bruyn, P., M. Aranda, and H. Murua Best practice document based on, or extrapolated from, the collated material in WP 2-3; Strategic report on proposed future research topics, detailing approaches, mechanisms and funding; and Identified opportunities for co-operation within the proposed research topics, and recommendations for regional networks. EU FP7 project nº TXOTX, Deliverable 4, 100 pp. Hurry, G. D., Hayashi, M. and Maguire, J. J Report of the independent review. International Commission for the Conservation of Atlantic Tunas (ICCAT). Unpublished ICCAT report, ICCAT, Madrid. 105pp Restrepo, V. R Proposed Stock Assessment Quality Control Procedures for ICCAT. ICCAT Collect. Vol. Sci. Pap., 52(5) Santiago, J., G.P. Scott, J.G. Pereira, Implementation of best science in the SCRS. ICCAT Collect. Vol. Sci. Pap., SCRS/2012/

47 AN EXAMPLE OF TERMS OF REFERENCE FOR AN INDEPENDENT REVIEW Annex 2 : Tentative Terms of Reference For the Peer Review SEDAR 21 Highly Migratory Species (HMS) Sandbar, Dusky, and Blacknose sharks Review Workshop 1. Evaluate the adequacy, appropriateness, and application of data used in the assessment. 2. Evaluate the adequacy, appropriateness, and application of methods used to assess the stock. 3. Recommend appropriate esti mates of stock abundance, biomass, and exploitation. 4. Evaluate the methods used to estimate population benchmarks and stock status ( e.g., MSY, Fmsy, Bmsy, MSST, MFMT, or their proxies); recommend appropriate management benchmarks, provide estimated values f or management benchmarks, and declare stock status, consistent with the stock status determination criteria, benchmark, and biological reference points in the Consolidated HMS FMP, proposed FMPs and Amendments, other ongoing or proposed management programs, and National Standards. 5. Evaluate the adequacy, appropriateness, and application of the methods used to project future population status, rebuilding timeframe, and generation time; recommend appropriate estimates of future stock condition (e.g., exploita tion, abundance, biomass). 6. Evaluate the adequacy, appropriateness, and application of methods used to characterize the uncertainty in estimated parameters. Provide measures of uncertainty for estimated parameters. Ensure that the implications of uncertain ty in technical conclusions are clearly stated. 7. Ensure that stock assessment results are clearly and accurately presented in the Stock Assessment Report and that reported results are consistent with Review Panel recommendations. If there are differences b etween the AW and RW due to reviewer s requests for changes and/or additional model runs, etc., describe those reasons and results. 8. Evaluate the SEDAR Process as applied to the reviewed assessments and identify any Terms of Reference that were inadequately addressed by the Data or Assessment Workshops. 9. Consider the research recommendations provided by the Data and Assessment workshops and make any additional recommendations or prioritizations warranted. Clearly denote research and monitoring needs that coul d improve the reliability of future assessments. Recommend an appropriate interval for the next assessment, and whether a benchmark or update assessment is warranted. 10. Prepare a Peer Review Summary summarizing the Panel s evaluation of the stock assessment and addressing each Term of Reference. Provide a list of tasks that were not completed, who is responsible for completing each task, and when each task will be completed. Complete and submit the Final Summary Report within 3 weeks of workshop conclusion. Appendix

48 SCRS/2013/024 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) A PLAN FOR THE PLAN Josu Santiago 1, Paul de Bruyn 2, Haritz Arrizabalaga 3, Hilario Murua 3 and Gerald P. Scott 1 SUMMARY Considering the outcomes of the 2011 Working Group on the Organization of the SCRS, the Resolution on Best Available Science [Res ] and the necessity for provision of appropriate advice to present and future requests from the Commission, strategic planning is proposed to develop a structured approach to guide the future workings of the SCRS. This document is intended to motivate discussion within the SCRS around definition of components for and creating a roadmap for a SCRS Strategic Plan. RÉSUMÉ Compte tenu des résultats obtenus par le Groupe de travail sur l organisation du SCRS en 2011, la Résolution sur la meilleure science disponible (Rés ) et la nécessité de formuler un avis adéquat aux requêtes actuelles et futures de la Commission, un plan stratégique est proposé en vue d'élaborer une approche structurée pour orienter les futurs travaux du SCRS. Le présent document vise à encourager les débats au sein du SCRS autour de la définition des composantes d'un plan stratégique du SCRS et de la création d'une feuille de route pour ce plan. RESUMEN Considerando los resultados de la reunión de 2011 del Grupo de trabajo sobre organización del SCRS, la Resolución sobre la mejor ciencia disponible [Res ] y la necesidad de proporcionar un adecuado asesoramiento científico frente a solicitudes actuales y futuras de la Comisión, se propone el desarrollo de un plan estratégico para elaborar un enfoque estructurado para orientar el trabajo futuro del SCRS. Este documento pretende fomentar las discusiones dentro del SCRS sobre la definición de los componentes y la creación de una hoja de ruta para un plan estratégico del SCRS para KEYWORDS Best science, Strategic planning, Transparency, Best practice, Quality of scientific advice 1 AZTI Tecnalia, Txatxarramendi ugartea z/g Sukarrieta, Basque Country, Spain 2 ICCAT Secretariat, C/ Corazón de Maria, 8, 6 th floor, Madrid 3 AZTI Tecnalia, Herrera Kaia Portualdea z/g, Pasaia, Basque Country, Spain 2064

49 What is Strategic Planning? There are several definitions of Strategic Planning. According to Wikipedia, Strategic Planning is an organization's process of defining its strategy, or direction, and making decisions on allocating its resources to pursue this strategy. In order to determine the direction of the organization, it is necessary to understand its current position and the possible avenues through which it can pursue a particular course of action. Generally, strategic planning deals with at least one of three key questions: 1."What do we do?" 2."For whom do we do it?" 3."How do we excel?" The key components of 'strategic planning' include an understanding of the organization's mission (the organization s purpose), vision (the organization s future direction), values, goals and strategies. A strategic plan identifies the long-term or overall objective and outlines the steps to achieve the desired future outcomes of the organization. According to the experts working group of the EU 7th framework project TXOTX, all Regional Fishery Regional Organization, and those using the results of fisheries research, should develop strategic research plans which will allow them prioritizing research in line with clearly defined objectives. Each objective should be listed alongside a responsible party and potential collaborators including specified deliverables with deadlines, levels of funding and resources needed, and where possible identifying funding sources (De Bruyn et al., 2011). The plan thus identify objectives and provide the organization a management tool to prioritize research, focus the energy, resources, and to align all the efforts/time of the organization in the same direction. Does the SCRS need a Strategic Plan? The Rules of Procedure of the ICCAT Convention described the mandate of the SCRS, on which each member of the Commission may be represented, is responsible for developing and recommending to the Commission all policy and procedures for the collection, compilation, analysis and dissemination of fishery statistics. It is the SCRS' task to ensure that the Commission has available at all times the most complete and current statistics concerning fishing activities in the Convention area as well as biological information on the stocks that are fished. The SCRS also coordinates various national research activities, develops plans for special international cooperative research programs, carries out stock assessments, and advises the Commission on the need for specific conservation and management measures. The report of the independent review panel (Hurry et al., 2008), stated that the SCRS has regularly reviewed its mode of operation and adapted to meet the new challenges and the needs of the Commission. The current arrangement has evolved to meet the needs and resources of ICCAT (Hurry et al., 2008). In this respect, several SCRS initiatives to adapt to increased demands (in quantity and quality) of the Commission have been regularly put in place; but without a clear systematic development. Considering the increase in demand for more scientific advice from the Commission, the 2011 meeting of the Working Group on the Organization of the SCRS (Anon., 2012) identified a set of priorities regarding Research & Development investments, participation of CPC national scientists, capacity building, quality assurance and transparency. And during the 2011 ICCAT Commission meeting in Istanbul, reaffirming the necessity that any conservation and management measure is based on the best possible scientific advice, the Commission adopted the Resolution on Best Available Science [Res ]. The Resolution proposes a set of actions affecting to the different links in the chain of the development of the scientific advice (Santiago et al., 2013). Most of the actions are related to quality assurance, including aspects in the sphere of collection of data, research, participation and capacity building, dialog with the SCRS and, very particularly, quality control of the stock assessments and advice (Figure 1). The necessity for providing appropriate and best scientific advice on current and future requests, the development of a strategic research plan appears the best way to provide timely, appropriate and best responses in a structured manner. An agreed research plan will also allow effectively adapting and adjusting the SCRS activities to funding sources, anticipating changes and necessities as well as preparing for them. Planning also 2065

50 should contribute towards a more efficient functioning and a better utilization of the always limited existing assets, resources and capabilities of the SCRS and the Secretariat. In addition, at the KOBE II Joint tuna-rfmo meeting held in Barcelona in 2010, during the presentation of point 9 of the agenda Enhanced cooperation among tuna RFMOs the presenter recommended that all tuna RFMOs develop strategic research plans in order to identify research priorities, joint collaborative effort and funding resources globally (Anon 2010). Therefore the SCRS proposed, starting in 2013, to develop a draft Strategic Science Plan which will be considered at an ad hoc meeting of the SCRS during This will be peer reviewed before approval by SCRS and adoption by the Commission. It is the purpose of this document to stimulate discussion around the definition of the components and the creation of a roadmap for the elaboration of the SCRS Strategic Plan. Components of the SCRS Strategic Plan The purpose of the SCRS Strategic Plan is to provide the overall framework and goals for development and coordination of science and science-related activities needed to support provision of sound scientific advice as the centerpiece for the conservation and management of tuna and tuna-like species in the Atlantic and the Mediterranean. The Plan will contemplate the period and will include the following components: - Mission: Outline the purpose of the SCRS, in line with the Rules of Procedure defined in the ICCAT Convention, its values and the necessities of the Commission. In other words, define what business are we in? - Vision: A statement describing where the SCRS desires to be in 2020; the target around which we pursue to focus the attention and energies of the SCRS. - Values: The guiding principles of the SCRS, including the elaboration of a code of conduct. Some examples of values to promote and defend include: independence, objectivity transparency, honesty, and cooperation. - Goals: The goals are broad priorities for the SCRS. Most strategic plans include few carefully selected goals relevant to stakeholder needs and available resources. Goals should be SMART : Specific, Measurable, Attainable, Realistic, and Timely. Goals are not likely to change from year to year if they are correctly identified in the strategic plan (but the strategies to achieve goals will change with new ideas and initiatives). Goals could be classified within Thematic Areas: examples are Data Collection, Research Priorities, Stock Assessments and Advice, Participation and Capacity Building, and Dialog and Communication. For example a goal for the SCRS within the Theme Participation and Capacity Building could be defined as Improving the science capabilities of the SCRS - Strategies: Within each goal should be strategies, initiatives, projects and priorities that will advance the goal. For example, the goal Improving the science capabilities of the SCRS may include strategies such as Supporting visiting opportunities for developing economy scientists at national laboratories creating grants to support 4-5 pre- or post-doctoral stays, elaborating 3 curricula designs of stock assessment modeling approaches for different levels of knowledge,... Strategies will of course influence the allocation of resources, but also reflect available resources. - Tactics, Performance measures: To accomplish the goals applying strategies identified the plan should include accountability and performance measures. Identify who (i.e. SCRS, Secretariat, CPCs and Commission) is responsible for which elements, how success will be measured and by when (interim dates and deadlines). 2066

51 Roadmap for Elaboration of the SCRS Strategic Plan The SCRS Strategic Plan will be coordinated by the SCRS Chair and will be elaborated following the following phases (Figure 2): Phase What Who When 1 Assess the situation: gaps and needs, SCRS (rapporteurs, conveners identify goals and strategies and chair) & Secretariat 02-08/ Validate goals and strategies; agree on mission, vision and values SCRS plenary 10/ Elaborate 1st draft SCRS (rapporteurs, conveners and chair) & Secretariat 11/ / Completion of the Plan SCRS (ad-hoc meeting) 05/ External peer-review External experts 06-07/ Approval of the Plan ICCAT Commission 11/2014 As part of the first three planning phases, specialized technical assistance will be hired to help identifying and adapting the necessary methodology of strategic planning to ICCAT specificities. Based on the specific method agreed, they will stimulate and advise the SCRS in the development of the Strategic Plan. The first and third phase will be conducted mainly electronically. References Anon., Report of the Joint Tuna RFMOS Meeting of Experts to Share Best Practices on the Provision of Scientific Advice. KOBE2 WS1 Report No. TRFMO2_W1-020/2010. Barcelona, Spain. 19pp. Anon., Report of the 2011 meeting of the Working Group on the Organization of the SCRS, Madrid, Spain March 2 to 4, Collect. Vol. Sci. Pap. ICCAT, 68(5): (2012). De Bruyn, P., M. Aranda, and H. Murua, Best practice document based on, or extrapolated from, the collated material in WP 2-3; Strategic report on proposed future research topics, detailing approaches, mechanisms and funding; and Identified opportunities for co-operation within the proposed research topics, and recommendations for regional networks. EU FP7 project nº TXOTX, Deliverable 4, 100 pp. Hurry, G.D., Hayashi, M. and J.J. Maguire, Report of the independent review. International Commission for the Conservation of Atlantic tunas. Unpublished ICCAT report, ICCAT, Madrid. 105pp. Santiago, J., G.P. Scott, J.G. Pereira, Implementation of best science in the SCRS. ICCAT Collect. Vol. Sci. Pap., SCRS/2012/

52 CPCs DATA RESEARCH (Biology, methods, ) 1.2. Improve data collection (including bycatches 1.4. Facilitate participation of scientists 5. Meeting participation fund 1.3.Support research programs 2.1. Enhance participation of scientists SCRS 4. SCRS publish scientific findings in scientific peer review literature 3. Strengthen peer review mechanisms ASSESSMENT 1.5. Contribute training 2.2. SCRS: Code of conduct for scientists 2.4. Fully documentation of sources and history of revisions 2.7. Provide different opinions in the report SCIENTIFIC ADVICE 2.3. Independent and objective scientific input (peer reviewed) 2.5. Provide clear, transparent & standardized scientific findings COM 1.1. Improve communication 2.6. Provide well defined rules for efficient decision making DECISION MAKING Figure 1. Diagram showing the main actions identified under Resolution by ICCAT on Best Available Science [Res ]. 10 SCRS 1st discussion Assess the situation/ Identify objectives and strategies Complete the Plan Adoption by COM 4 SCRS ad hoc meeting 10 Approval by SCRS Complete the Plan Peer review Figure 2. Timeline of the elaboration of the SCRS Strategic Plan. 2068

53 SCRS/2013/025 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) MODEL DIAGNOSTICS FOR STOCK SYNTHESIS 3: EXAMPLES FROM THE 2012 ASSESSMENT OF COBIA IN THE U.S. GULF OF MEXICO Shannon L. Cass-Calay 1, Jakob C. Tetzlaff 1, Nancie J. Cummings 1 and J. Jeffery Isely 1 SUMMARY This document summarizes common model diagnostics available for Stock Synthesis 3 and describes their interpretation. Examples of model misspecification are described and the resulting diagnostics are illustrated. Solutions to improve model performance are also discussed. RÉSUMÉ Le présent document récapitule les diagnostics du modèle communs disponibles pour Stock Synthèse 3 et décrit leur interprétation. Des exemples d'erreurs de spécification du modèle sont décrits et les diagnostics résultants sont illustrés. Les solutions visant à améliorer les performances du modèle sont également discutées. RESUMEN Este documento resume los diagnósticos del modelo comunes disponibles para Stock Synthesis 3 y se describe su interpretación. Se describen ejemplos de especificaciones erróneas del modelo y se ilustran los diagnósticos resultantes. Se discuten también soluciones para mejorar el rendimiento del modelo. KEYWORDS Stock Synthesis, Stock Assessment, Model Diagnostics 1. Introduction Stock Synthesis (SS) is an integrated stock assessment model which is widely used in western and southeastern United States, and increasingly used throughout the world (i.e. ICES 2012). SS has a statistical framework that allows the user to calibrate a population dynamics model using a variety of fisheries and survey data (Methot 2011). SS is a flexible model which can be specified using multiple genders, growth morphs, and/or stocks within one or more areas, and can be parameterized using an age-based structure or size-based structure. SS allows the user to include ageing error, and estimate growth, the spawner-recruitment relationship, and movement between areas. Due to its flexibility, SS can be parameterized to mimic many commonly used assessment models such as ASPIC, age-structured production models, statistical catch-at-age models and virtual population analysis. SS also includes an integrated projection routine which allows the uncertainty in estimated parameters to be propagated to the projections and the calculation of management references, thus facilitating risk analyses, including projections of possible annual catch limits. A comprehensive collection of R functions (r4ss; is available to summarize and plot model results, manipulate files, and visualize model parameterizations. R4SS produces a variety of model diagnostics. Unfortunately, many preliminary SS model configurations, particularly complex ones, exhibit poor diagnostic behaviors which are difficult to interpret, and correct. This contributes to the extensive learning curve for SS. To facilitate the use of SS for ICCAT assessments, we illustrate example diagnostics that suggest improper model parameterization or structure, and describe the cause of the poor model behavior. When possible, we also include an improved diagnostic for contrast. We also include a checklist that describes the steps required to adequately, and efficiently evaluate the performance of SS model runs. 1 U.S. Department of Commerce, NOAA Fisheries, Southeast Fisheries Science Center, Miami Laboratory, 75 Virginia Beach Drive, Miami, Florida U.S.A

54 2. Materials and methods During , the Southeast Fisheries Science Center completed an assessment of cobia (Rachycentron canadum) in the U.S. Gulf of Mexico. The assessment was conducted using Stock Synthesis 3.24h (beta), a version of stock synthesis that allows discard-only fleets. The data inputs are briefly summarized below. A detailed description can be found in SEDAR (2013). Model results and diagnostics were developed using r4ss, a comprehensive set of functions in R for summarizing and plotting SS results (http://code.google.com/p/r4ss/). 2.1 Life history Life history data used in the assessment included natural mortality, growth, maturity, and fecundity. A single von Bertalanffy equation was used to model growth of cobia for both sexes. The von Bertalanffy parameters were estimated within the SS model. A fixed weight-length relationship (W=aFL b ) was used for both sexes combined. The age at 50% maturity was fixed at 2, and fish age-3+ fish were assumed to be fully mature. Fecundity was assumed to be directly proportion to female weight in the SS model. The sex ratio was fixed at 60% females for all ages. A point estimate of natural mortality (M=0.38 y -1 ) was used to scale the age-based estimates of natural mortality (Lorenzen 1996). 2.2 Landings and Discards Catch and discards series used in the assessment included: Commercial landings ( ) in metric tons. Recreational landings ( ) in 1000s of fish. Commercial discards ( ) in metric tons. Discard mortality rate = 5%. Recreational discards ( ) in 1000s of fish. Discard mortality rate = 5%. Shrimp bycatch ( ) was modeled as a discard only fleet (Discard Mortality = 100%). Shrimp Bycatch was assumed to be a function of the shrimp fishing fleet effort and was modeled using a superyear approach (Methot 2011). 2.3 Length Composition Due to a lack of direct observations of age composition, the cobia stock assessment model used a length-based structure. Sources of length composition included: Commercial length composition ( ). Reef fish observer program length composition ( ). Recreational length composition ( ). SEAMAP trawl survey length composition ( ) was assumed to be representative of the shrimp fishery. 2.4 Age conditioned on length (ALKs) Recreational age composition data were made conditional on length. Using these conditional age compositions has the advantage of linking age data directly to length data (essentially creating an age-length key). As a result, the data contain more detailed information about the relationship between size and age and so provides a stronger ability to estimate growth parameters, especially the variance of size-at-age. 2.5 Indices Two relative indices of abundance were used in the stock assessment. Both indices are fishery-dependent and both provide indices of abundance for the recreational fishery for cobia in the Gulf of Mexico. Marine Recreational Fishery Statistics Survey (MRFSS), an index of total catch. Headboat Survey, an index of landings. 2070

55 3. Results and Discussion The R script r4ss automatically plots model results and diagnostics for SS. A thorough examination of these figures may reveal improper model behavior resulting from model misspecification. For example, abnormal recruitment patterns may occur, stock productivity estimates may be unrealistically high or low, selectivity patterns may be nonsensical and drastic increases/decreases in stock biomass may occur in a single year. Examples of these problematic behaviors follow. A small number of cobia are captured each year as bycatch to the shrimp trawl fishery in the Gulf of Mexico. Observations of length composition in this fishery suggest that most cobia in shrimp trawls are small, but some large fish are also captured, apparently while riding then net. When the selectivity of the shrimp bycatch fleet is misspecified, allowing unrealistically high numbers of large fish to be intercepted, unusual model results occur. Upon the initiation of the shrimp bycatch fleet, an improbable depletion of the spawning stock biomass occurs (Figure 1) and the equilibrium SPR drops precipitously (Figure 2). These improbable results are also accompanied by unusual pattern in recruitment, specifically a boom and bust behavior that is unlikely given the biology of cobia (Figure 3). Examinations of the fit to the length (or age) composition can also be informative. A sudden lack of fit to length composition can be caused by the improper specification of a management regulation (Figure 4). In this example, the model was specified to re-estimate the retention function following the implementation of a minimum size limit in However, the length composition data suggests that the management regulation was ignored for a number of years, causing large residuals in the length composition from (Figure 4). The model was improved by retaining the pre-regulatory retention function for several years after the imposition of the minimum size regulation, allowing the change in retention to coincide with the observed shift in length composition. The covariance matrix may also contain evidence of improper model formulation. In particular, these tables should be scrutinized for evidence of high coefficients of variation and strong correlations between selectivity patterns and growth parameters. In this example, there was evidence of a high degree of confounding in estimates of shrimp selectivity (Shrimp_5) and the growth parameter (Lmin; Table 1). Also, there was high correlation between the shrimp selectivity parameters themselves (Shrimp_2, Shrimp_3, Shrimp_4; Table 2). As a result, this model could not reliably estimate the double-normal selectivity pattern of the shrimp bycatch fishery. Trace plots are a useful tool which show the parameter estimates relative to the phase of estimation. In a well behaved model, the parameters should not change a great deal after the final phase of estimation. In this example (Figure 5), it is clear that model parameters varied substantially following the final phase of estimation. This undesirable result was alleviated by reconfiguring the phases of estimation. A thorough evaluation of model sensitivity to key assumptions, data weighting choices and alternative data inputs may reveal an undesirable degree of model sensitivity. In this example (Figure 6) the model converged on alternative solutions when the weighting of the shrimp fishery length composition data was changed. This unstable model result warrants a thorough examination of the model components to more properly specify component weighting. A reiterate reweighting procedure, such as (Francis 2011) could be instructive. Stock synthesis includes several automated routines to evaluate model stability. One, the Jitter analysis allows the user to examine the effect of varying model input parameters on model results. A well behaved model should converge on a global solution across a reasonable range of input parameters (Table 3). Likelihood profiling is also an automated routine in Stock Synthesis. This tool allows the user to evaluate model performance across a range of values of an input parameter (e.g. steepness, R0, sigma-r). Ideally the profile should be a smooth functional shape. Abnormal model behavior is indicated by numerous spikes and sawtoothed profiles. One should take care to evaluate the likelihood profile with sufficient precision. In this example, a likelihood profile on steepness was run using the range At a precision of 0.05, the profile appears smooth (Figure 7). However, a finer precision, 0.01 reveals model instability (i.e. numerous peaks in the profile). In this case, the poor performance was caused by growth and selectivity parameter estimates which were highly correlated with each other. 2071

56 This improper model behavior is also evident in plots of the maximum likelihood estimates of various model parameters across a range of steepness (Figure 8). The performance of stock synthesis can also be evaluated by plotting the distribution of parameter estimates and derived quantities across bootstrapped replicates. In stock synthesis, the variances of all the estimated parameters are propagated to the bootstraps. In a well behaved model, the maximum likelihood estimate should be similar to the mean of the bootstraps. In this example (Figure 9 and Figure 10), 1000 bootstraps are summarized. Note the deviations between the maximum likelihood estimate (MLE) and the mean for several model parameters. This indicates that the model performance in not ideal. Stock synthesis models that estimate steepness merit additional caution because steepness is often inestimable. In this example (Figure 10), the mean steepness estimate of the bootstraps differs substantially from the maximum likelihood estimate. Furthermore, a large number of bootstrap runs estimate the steepness at the theoretical maximum, 1.0. The behavior of model runs at steepness = 1.0 often exhibit unrealistic behaviors (e.g. FSPR20% unsustainable) that may influence other parameter estimates. Our experiences with the stock assessment of Gulf of Mexico cobia encouraged us to create a checklist to facilitate an efficient and complete evaluation of model diagnostics. These are summarized in Table 4. Acknowledgments The authors would like to acknowledge the assistance of Dr. Rick Methot and Ian Taylor who graciously reviewed numerous preliminary SS runs of Gulf of Mexico cobia. We would also like to thank the SEDAR 28 participants including data providers, analysts and reviewers. References ICES Report of the Working Group on the Assessment of Southern Shelf Stocks of Hake, Monk and Megrim. ICES Document CM 2012/ACOM: 11. FRANCIS. R.I.C Data weighting in statistical fisheries stock assessment models. Can. J. of Fish. Aquat. Sci. 68: LORENZEN, K The relationship between body weight and natural mortality in juvenile and adult fish: a comparison of natural ecosystems and aquaculture. Journal of Fish Biology, 49: METHOT, R.D User manual for Stock Synthesis: model version 3.23b. Nov 7, NOAA Fisheries Service, Seattle, WA. 2072

57 Table 1. Mean, standard deviation, CV and correlation of model parameters. Table 2. Mean, standard deviation, CV and correlation of model selectivity parameters. 2073

58 Table 3. The results of a Jitter analysis evaluating the effect of varying input parameters on model results. Run Likelihood SSB unfished SSB 2011 Depletion

59 Table 4. A flowchart to facilitate an efficient and complete evaluation of Stock Synthesis model diagnostics SS Model Diagnostics 1. Does the model run? a. No use echo input to debug b. Yes continue 2. Does the hessian converge? a. No check warning file, check estimated parameters in report file b. Yes continue 3. Are there any parameters on bounds? a. No continue b. Yes change starting values/change bounds/add priors/simplify parameterization rerun 4. Plot model output. Anything obviously wrong? Examples: productivity way too low, selectivity patterns that don t make sense, drastic decrease/increase in biomass in a single year, abnormal recruitment patterns (boom/bust). a. No continue b. Yes go through report file to diagnose (depends on problem) 5. Examine parameter estimates. Plot parameter distributions along with starting values, bounds, and priors. Do parameters appear well estimated? a. No check bounds, check priors, check phase of estimation b. Yes continue 6. Look at trace plots of parameter estimates relative to phase of estimation? Do model parameters change considerably in the final phase? a. No continue b. Yes try alternative phases: for example, important scaling parameters like unfished recruitment and catchability might be estimated in the first phase, recruitment deviates estimated added in the second phase, and selectivity added in the final phase. 7. Look at mean and standard deviation of estimated parameters. Is CV of estimated parameters less than 1? a. No is there data to inform parameter? i. No change bounds/add informative prior/fix parameter ii. Yes check correlation matrix b. Yes continue 8. Are any of the parameters highly correlated? a. No continue b. Yes why? Does one of the parameters require an informative prior? 9. Plot model fits to data and diagnostics. Is model fitting data reasonably? a. No diagnose the problem. b. Yes continue 10. Check for model stability to initial starting parameters using Jitter analysis. Does model converge to a global solution? a. No identify why. i. look at which likelihood components are changing ii. Evaluate the phases of estimation iii. Plot distribution of estimated parameters over all model runs b. Yes continue (try again with larger deviation from starting values) 11. Profile leading model parameters such as stock-recruitment parameters (steepness/r0) or natural mortality. Was the profile smooth? a. No Plot estimated parameters as a function of profiled leading parameter i. Do any of the parameters hit bounds across the runs? Do any of the parameters bounce between alternative solutions? Do some parameters show similar patterns? 1. Yes may not have enough data to inform all estimated parameters: add informative priors/reduce the number of estimated parameters. b. Yes Does profile show leading parameter is well estimated? Do the different data components show similar signals? i. No parameter may require informative prior or need to be fixed ii. Yes profile at finer scale 1. Does profile remain smooth? a. Yes continue 2075

60 12. Evaluate model sensitivity to key model assumptions, data weighting choices, and alternative data inputs. Was model highly sensitive to any key model assumptions or certain data sources? a. No continue b. Yes Is model specified correctly? Are assumptions appropriate? Is model overparameterized? Should data be re-weighted? 13. Evaluate model sensitivity to the most recent years of data using a retrospective analysis. Did the retrospective analysis reveal any inconsistencies in the data? a. No continue b. Yes identify source of the retrospective pattern 14. Evaluate model uncertainty using bootstrap approach. Plot distribution of parameter estimates and derived quantities from bootstrapped runs. Compare MLE of parameter estimates to mean of bootstrap results. Are parameters or derived quantities well estimated when data is resampled? a. No do distributions show multi-modality or high proportion of bounding? i. Yes may not have enough data to inform all estimated parameters: add informative priors/reduce the number of estimated parameters. b. Yes continue 15. Evaluate model convergence using MCMC approach. Use standard approaches to evaluating MCMC results: look at trace plots/plot posterior distributions/compare MLE to mean of posterior distribution. Does MCMC converge on a single solution? Are MLEs of parameters/derived quantities similar to mean of posterior distributions? a. No b. Yes continue Model Misspecified Base Model Figure 1. An improbable drop in spawning stock biomass resulting from model misspecification of the shrimp bycatch selectivity (left) and the improved behavior of the corrected model (right). 2076

61 Model Misspecified Base Model Figure 2. An improbable drop in equilibrium SPR resulting from model misspecification (left) and the improved behavior of the corrected model (right). Model Misspecified Base Model Figure 3. Unusual boom and bust recruitment deviations caused by model misspecification (left) and the improved behavior of the corrected model (right). 2077

62 Model Misspecified Base Model Figure 4. Pearson residuals for the fit to the length composition resulting from model misspecification (left) and the improved behavior of the base model (right). Figure 5. Trace plots of model parameters relative to the phase of estimation (dashed vertical lines). Parameter estimates should not vary substantially following the final phase of estimation. 2078

63 Figure 6. This model converged on an alternative solution when length composition data was reweighted. Figure 7. Likelihood profile for steepness at intervals of 0.05 (left) and 0.01 (right). 2079

64 Figure 8. Plot of MLE of parameters (y-axis) across a range of values for steepness (x-axis). 2080

65 Figure 9. Distribution of estimated shrimp selectivity parameters from 1000 bootstrap replicates. Blue lines represent mean estimates from the bootstrap samples, red lines represent the point estimate of the parameters from the base model. Figure 10. Distribution of estimated equilibrium recruitment and steepness from 1000 bootstrap samples. Blue lines represent mean estimates from the bootstrap samples, red lines represent the point estimate of the parameters from the base model. 2081

66 SCRS/2013/033 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) AN EXAMPLE OF A MANAGEMENT PROCEDURE BASED ON A BIOMASS DYNAMIC STOCK ASSESSMENT MODEL L. Kell 1, G. Merino 2, P. de Bruyn 1, J.M. Ortiz de Urbina 3, H. Arrizabalaga 2, H. Murua 2 and J. Santiago 3 SUMMARY We describe a Management Procedure based on a biomass dynamic stock assessment model, which is intended for used to evaluate harvest strategies as part of Management Strategy Evaluation. RÉSUMÉ Le présent document décrit une procédure de gestion basée sur un modèle dynamique d évaluation des stocks de la biomasse qui est destinée à évaluer les stratégies de capture dans le cadre de l évaluation de de la stratégie de gestion. RESUMEN En este documento se describe el procedimiento de ordenación basado en un modelo dinámico de evaluación de la biomasa del stock, que se puede utilizar para evaluar estrategias de captura como parte de una evaluación de estrategias de ordenación. KEYWORDS Albacore, Harvest Control Rule, FLR, R, Management Strategy Evaluation, Management procedure 1. Introduction We describe a Management Procedure (MP) based on a biomass dynamic stock assessment model, which is intended to be used to evaluate harvest strategies as part of a process of Management Strategy Evaluation (MSE). An MP is the combination of pre-defined data, together with an algorithm to which such data are input to provide a value for a total allowable catch (TAC) or effort control measure, e.g. a stock assessment method including the estimation of reference points for use in a harvest control rule (HCR). A main objective of an MSE is to show through simulation trials whether a proposed MP or harvest strategies is robust to uncertainty (Rademeyer, 2007). A HCR is set of well-defined rules used for determining a management action. 2. Material and Methods 2.1 Management Procedure The MP is based on a biomass dynamic stock assessment, which assumes that the stock is homogenous with no immigration or emigration and that population processes are stationary. Stock dynamics are modelled by the Pella-Tomlinson surplus production function (Pella and Tomlinson, 1969). An example of a HCR is shown in Figure 1 as part of a phase plot. The orange line is the HCR where for a given stock biomass (x-axis) a harvest rate (y-axis) is set, i.e. management action is predetermined. The black line is the replacement line, i.e. for a given stock biomass any harvest rate above the black line will cause the stock to 1 ICCAT Secretariat, C/Corazón de María, Madrid, Spain; Phone: Fax: AZTI-Tecnalia, Herrera Kaia Portualdea, 20110, Pasaia, Spain; Phone: Fax: Instituto Español de Oceanografía IEO- CO Málaga, Pto. Pesquero s/n, Fuengirola (Málaga), Spain; Phone: Fax:

67 decline and any harvest rate below the line will cause the stock to increase. Therefore for the target harvest rate (the horizontal segment of the HCR) the target biomass is found at the intersection with the HCR. If the stock declines below the break point (i.e. a trigger biomass) then harvest rate is reduced progressively to the minimum harvest rate at a biomass level referred to as B lim. B lim is also referred to as a limit reference point (LRP). 2.2 Operating Model To demonstrate the use of a HCR we first created an Operating Model (OM) which simulated a stock that initially was at virgin and then declined due to increasing levels of fishing until the stock was over exploited. Fishing then decline to a level below F MSY after which the fishing level remained constant. Figure 2 shows time series of harvest rate, stock biomass and yield and Figure 3 is the corresponding Kobe phase plot of harvest rate relative to F MSY and stock biomass relative to B MSY. We then sample from the OM to create data for use in the MP. Two data series are generated, catch by year and an index of abundance with two levels of CV (20 and 40%) see Figure 4. This gives 2 levels of precision in stock and reference point estimates. 2.3 Software Software used was a biomass production model implemented as a package in R, this allows it to be used with a variety of other packages for plotting, summarizing results and to be simulation tested, e.g. as part of the FLR tools for management strategy evaluation (Kell et al., 2007). 3. Results In the first set of simulations we do not use the HCR to set catches; instead we conduct a retrospective analysis for every second year from year 60 to year 80. We then estimated stock status, reference points and stock status relative to reference points. We did this for each retrospective run and level of CPUE uncertainty. The estimates of F MSY, B MSY, F, stock biomass, F: F MSY and B: B MSY are shown in Figure 5 which shows the inter-quartile ranges. Two main points can be seen from this analysis, i.e. that estimates obtained from an CPUE index with a CV of 20% are more precise than those based on a CV of 40% and that in the later retrospectives (i.e. those that use more data) estimates are again more precise. Next three HCRs were run for a single realisation (a single time series) to demonstrate different options for use as part of an MSE. i.e. i) At Once the F based on harvest control is implemented as a TAC ii) 10% F Bound the F each year is not allowed to vary by more than 10% a year. iii) 10% TAC Bound the TAC each year is not allowed to vary by more than 10% a year; and iv) Multi-annual the TAC is set for three years based on the average value TAC as predicted by the HCR. Figures 6 and 7 show the time series and phase plots for the 4 HCRs. There is little difference between setting a TAC annually or for three year periods. However, there are large differences in the performance of the HCRs, when the inter-annual variability in TAC or F (i.e. capacity or effort) is restricted. This will have consequences both for the stock and the economic performance of fleets. 4. Discussion and Conclusions The paper is intended to demonstrate how a MP can be modelling, despite the examples being simple some important points are worth noting i.e. Unsurprisingly reference points estimated by a stock assessment include uncertainty, the level of which depend upon the quality of the data and knowledge of stock and fleet dynamics. For a given level of risk, e.g. of stock collapse, reference points used in a HCR will need to be based on the level of stock assessment uncertainty. This could be done by making reference points such as B trigger (a level of biomass that if the stock falls below results in a reduction in harvesting) a percentile of a probability distribution of a reference point such as B MSY rather than a multiplier of a point estimate. Different ways of estimating parameter uncertainty, e.g. normal approximations, bootstrapping and Bayesian approaches could be evaluated and their use as part of HCRs and MPs investigated. In practice a HCR will have economic as well as biological consequences and these should be considered when designing a HCR. 2083

68 Bibliography L. Kell, I. Mosqueira, P. Grosjean, J. Fromentin, D. Garcia, R. Hillary, E. Jardim, S. Mardle, M. Pastoors, J. Poos, et al. Flr: an open-source framework for the evaluation and development of management strategies. ICES Journal of Marine Science: Journal du Conseil, 64(4):640, J. J. Pella and P. K. Tomlinson. A generalized stock production model. Inter-American Tropical Tuna Commission, R. A. Rademeyer, _E. E. Plag_anyi, and D. S. Butterworth. Tips and tricks in designing management procedures. ICES Journal of Marine Science: Journal du Conseil, 64(4):618{625, Figure 1. Example of a Harvest Control Rule (orange) plotted on a phase plot of harvest rate relative to F MSY and stock biomass relative to B MSY. The black line is the replacement line. 2084

69 Figure 2. Simulated time series of harvest rate, stock biomass and yield. Figure 3. Phase plot of harvest rate relative to F MSY and stock biomass relative to B MSY. 2085

70 Figure 4. Simulated time series of index of abundance, assuming a measurement error with CVs of 20 and 40%. Figure 5. A comparison of the interquartile range with the true values (lines) for two levels of CPUE CV. Bars are derived from a retrospective analysis, i.e. the values estimated in the last year. 2086

71 Figure 6. Time series of harvest rate, stock biomass and yield for the four HCRs. Figure 7. Phase plots of harvest rate relative to F MSY and stock biomass relative to B MSY for the four HCRs. 2087

72 SCRS/2013/034 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) AN EXAMPLE OF CONDITIONING AN OPERATING MODEL USING MULTIFAN-CL Laurence T. Kell 1, Gorka Merino 2, Paul De Bruyn 1, Josetxu Ortiz de Urbina 3, Haritz Arrizabalaga 2, Josu Santiago 2 and Hilario Murua 2 SUMMARY We use Multifan-CL to develop an Operating Model (OM) for North Atlantic Albacore; conditioning the OM based on three sets of biological parameters. The work is a first step in developing a Management Strategy Evaluation framework for North Atlantic Albacore. RÉSUMÉ Nous utilisons Multifan-CL afin de développer un modèle opérationnel (OM) pour le germon de l'atlantique Nord; en conditionnant l'om sur la base de trois ensembles de paramètres biologiques. Le travail est une première étape dans l'élaboration d'un cadre d évaluation de la stratégie de gestion pour le germon de l'atlantique Nord. RESUMEN Utilizamos Multifan-CL para desarrollar un modelo operativo (OM) para el atún blanco del Atlántico norte; condicionando el OM basado en tres conjuntos de parámetros biológicos. El trabajo es un primer paso en el desarrollo de una evaluación de la estrategia de gestión marco para el atún blanco del Atlántico norte. KEYWORDS Albacore, Management Strategy Evaluation, Multifan-CL, Operating Model 1 ICCAT Secretariat, C/Corazón de María, Madrid, Spain; Phone: Fax: AZTI-Tecnalia, Herrera Kaia Portualdea, 20110, Pasaia, Spain; Phone: Fax: Instituto Español de Oceanografía IEO- CO Málaga, Pto. Pesquero s/n, Fuengirola (Málaga), Spain; Phone: Fax:

73 1. Introduction This paper describes the conditioning of an Operating Model (OM) as part of a Management Strategy Evaluation (MSE) for North Atlantic albacore. Where an OM is a mathematical statistical model used to describe resource dynamics in simulation trials and to generate resource monitoring data when projecting forward. MSE describes the process of testing generic MPs or harvest strategies (Rademeyer et al., 2007). Conditioning is undertaken to ensure that the simulations are representative of the actual problem of managing the stock(a mar et al., 2008). There are several ways to condition an OM (Kell et al., 2006). The simplest is when the OM is based on a stock assessment model, as in this paper where we use the Multifan-CL (Fournier et al., 1998) North Atlantic albacore assessment to condition the OM. As part of the conditioning process we also consider two hypotheses about biological parameters based on a review of North Atlantic albacore biology (Santiago and Arrizabalaga, 2005) and life history theory (Gislason et al., 2010). 2. Material and Methods 2.1 Multifan-CL Multifan-CL (Fournier et al., 1998) is an ADMB based stock assessment program that implements a statistical, size-based, age-structured and spatial-structured model for use in fisheries stock assessment. It is used routinely for tuna stock assessments by the Oceanic Fisheries Programme (OFP) of the Secretariat of the Pacific Community (SPC) in the western and central Pacific Ocean (WCPO), (Hampton, et al., 2002). Both the Indian Ocean Tuna Commission (IOTC) and ICCAT use Multifan-CL to conduct stock assessments. The model is fitted to time series of catch, effort and size composition data from fishing fleets. Size composition data may be in the form of either length or weight-frequency data, or both. The model may also be fit simultaneously to tagging data, if available. Other information is provided to the model in the form of prior information on estimates of various biological and fisheries parameters and their variability. 2.2 Scenarios In order to capture the range of uncertainty about the biology of the stock, three alternatives assessment data inputs were created, i.e. OM1) the working group base case; OM2) based on a review of North Atlantic albacore and OM3) based on life life history theory. Although there is also uncertainty about the catch, effort and size data these did not vary between scenarios in this example. OM1 The model adopted as the base case was an update or continuity run of the previous base case from OM2 Biological parameters are based on a review of Northern Albacore biology and fisheries (Santiago and Arrizabalaga, 2005). Parameters were chosen to be consistent, i.e. the natural mortality and maturity-at-age vectors and Von Bertalanffy growth function were chosen so that there were no contradictions between age based parameters and the selected growth curve. OM3 Life history relationships were used to parameterise an age-structured equilibrium model, where SSB-perrecruit, yield-per-recruit and stock-recruitment analyses are combined, using fishing mortality (F), natural mortality (M), proportion mature (Q) and mass (W) -at-age (Gislason et al., 2010). 2.3 Future Projections Projections for the different management strategies require assumptions about processes such as recruitment. The scenarios were projected from 2000 to 2030 for process error in recruitment, i.e. historic recruit from within a scenario was sampled with replacement to select a recruitment value from 2000 to The stock was then projected for a fishing mortality equal to FMSY. 2.4 Software Software used was a biomass production model implemented as a package in R, this allows it to be used with a variety of other packages for plotting, summarising results and to be simulation tested, e.g. as part of the FLR tools for management strategy evaluation (Kell et al., 2007). 2089

74 3. Results The parameters are shown in Figure 1. Selectivity is estimated while the other parameters are fixed. Assuming that M increases with age results in an increase in selectivity F at older ages, since a higher F is required to explain the catches of larger fish. There is less effect seen for assuming that M is higher at younger ages. In ADMB when fitting a model with multiple parameters it is possible to initially fix some parameters during the frst optimisation (of some other parameters), the fixed parameters are then being estimated in later phases. This helps in fitting since the most influential parameters can be addressed first. The use of phases also allows a range of model formulations to be explored, i.e. from simple to complex. Figures 2 and 3 show the values of the objective function and number of parameters by scenario and phase. The biggest change in the likelihood is seen between phase 1 and 2 when selectivity is estimated. In phase 6 when changes in catchability by year are estimated little change in seen in the likelihood. The estimated time series are shown in Figure 4, while the corresponding equilibrium values and reference points in Figure 5 and the kobe phase plots in Figure 6. Finally a projection for OM1 is shown in figure 7. The biggest difference in the time series is seen for recruitment for OM3 (i.e. when M at younger ages is highest) and F for a U-shaped M (as would have been predicted from Figure 1). Changes are seen in the reference points, particularly for OM2 when a U-shaped M is assumed. 4. Discussion and Conclusions This paper provides an example of conditioning an OM on an assessment model for use as part of an MSE. We only considered 3 OMs based on biology. Estimates of fishing mortality stock status and reference points are highly dependent upon the assumed biology. There is an interaction between the assumed biology (i.e. M) and selectivity, since the lack of older fish in catches can be assumed by either. Previously changes have been made in the biological assumptions in the final phase, i.e. as a sensitivity analyses. This has several potential pitfalls, i.e. i) in ADMB the use of phases is aim to help finding a global maximum, changing important parameters in the final phase may mean that convergence is only to a local minima, ii) it will also be inefficient and iii) the biological assumptions should be part of any formal design used to develop the OM. Other hypotheses could have been considered, for example some fleets are assumed to have the same selectivity because there was insufficient information in the data to estimate these independently. Selectivity was also assumed not to change over time. Both assumptions are unlikely to be true. The catch and size data are also not true observations but have been generated in a variety of ways. There is therefore considerable uncertainty in the fleet data and assumptions. The indices of abundance also show different trends implying different historic dynamics. Therefore many more scenarios could have been run, reflecting the true uncertainty. Multifan-CL could have been run in a hierarchical design, e.g. if for each set of biological assumptions choices were made about the fleet data and then each phase was run to convergence then a set of scenarios could be developed. These could then be weighted either by applying externally specified weights (e.g. based on the plausibility of the assumptions) or likelihood-based weights based on fits to the data. Choice of scenarios and weightings may require expert judgment, consensus amongst experts may be sought but this may be problematical if some experts may have links with interest groups or tend to be more vocal than others? An alternative approach is that of the IWC Scientific Committee and CCSBFT which avoided quantification, and instead categories hypotheses as of high, medium and low weight. In cases where there is no agreement, but a plausible case can be made by some for a high weight, a medium weight is assigned. To avoid expert s choices of weights being influenced by the management implications of an associated hypothesis, these weights should ideally be finalised on the basis of informed discussion concerning the hypotheses alone conducted before any computations related to management. However, pragmatically, some flexibility on this point may be entertained in the interests of time for example to identify at an early stage that some hypotheses, although of appreciable plausibility, make little difference in terms of management implications when compared to corresponding default hypotheses, and hence need not be considered further (ACE, 2007). A consideration of robustness could also be made when choosing scenarios, i.e. is a scenario likely to have any effect? if not then ignore it. 2090

75 Bibliography L. Kell, I. Mosqueira, P. Grosjean, J. Fromentin, D. Garcia, R. Hillary, E. Jardim, S. Mardle, M. Pastoors, J. Poos, et al. Flr: an open-source framework for the evaluation and development of management strategies. ICES Journal of Marine Science: Journal du Conseil, 64(4):640, ACE, A. (2007). Report of the study group on risk assessment and management advice (SGRAMA). A mar, Z. T., Punt, A. E., Dorn, M. W., et al. (2008). The management strategy evaluation approach and the fishery for walleye pollock in the Gulf of Alaska. Resiliency of Gadid Stocks to Fishing and Climate Change, pages Fournier, D. A., Hampton, J., and Sibert, J. R. (1998). MULTIFAN-CL: a length-based, age-structured model for fisheries stock assessment, with application to south pacific albacore, Thunnus alalunga. Canadian Journal of Fisheries and Aquatic Sciences, 55(9): Fournier, D. A., Skaug, H. J., Ancheta, J., Ianelli, J., Magnusson, A., Maunder, M. N., Nielsen, A., and Sibert,J. (2012). AD model builder: using automatic differentiation for statistical inference of highly parameterized complex nonlinear models. Optimization Methods and Software, 27(2): Gislason, H., Daan, N., Rice, J., and Pope, J. (2010). Does natural mortality depend on individual size. Fish and Fisheries, 11(2): Hampton, J., Kleiber, P., Langley, A., and Hiramatsu, K. (2002). Stock assessment of bigeye tuna in the western and central pacific ocean. In Sec. Pacif. Comm., Oceanic Fish. Prog., 15th meeting, Stand. Comm. Tuna Billfish, BET-1. Kell, L., De Oliveira, J. A., Punt, A. E., McAllister, M. K., and Kuikka, S. (2006). Operational management procedures: an introduction to the use of evaluation frameworks. Developments in Aquaculture and Fisheries Science, 36: L. Kell, I. Mosqueira, P. Grosjean, J. Fromentin, D. Garcia, R. Hillary, E. Jardim, S. Mardle, M. Pastoors, J. Poos, et al. Flr: an open-source framework for the evaluation and development of management strategies. ICES Journal of Marine Science: Journal du Conseil, 64(4):640, Rademeyer, R. A., Plagányi, É. E., and Butterworth, D. S. (2007). Tips and tricks in designing management procedures. ICES Journal of Marine Science: Journal du Conseil, 64(4): Santiago, J. and Arrizabalaga, H. (2005). An integrated growth study for north atlantic albacore (Thunnus alalunga bonn. 1788). ICES Journal of Marine Science: Journal du Conseil, 62(4):

76 Table 1. Multifan-CL Parameters by phase used to condition the Operating Model under the three scenarios considered. Phase Phase 1 Phase 2 Phase 3 Phase 4 Phase 5 Phase 6 Phase 7 Phase 8 Phase 9 Parameters Effort deviation penalties Selectivity shapes Selectivity groupings Catchability groupings Size data weightings Scaling parameter of initial to total population Seasonal component of catchability Selectivity Temporal changes in recruitment Length dependant S.D. Mean length of 1st age class Individual mean lengths for 1st 2 age classes Time series change in catchability Dummy phase. Will be removed in future runs Dummy phase. Will be removed in future runs Bevereton Holt SRR Parameters of beta distribution defining prior for steepness Figure 1. Selectivity, natural mortality (M), maturity (mat) and individual growth (Mass) patterns used to condition the OM under three scenarios. 2092

77 Figure 2. Likelihood by phase and scenario. Figure 3. Number of parameters by phase and scenario. 2093

78 Figure 4. Estimated historic time series of the Operating Models. Figure 5. Equilibrium values and reference points estimated by Operating Model scenario. 2094

79 Figure 6. Kobe phase plots for the three Operating Models. Figure 7. Stochastic projection of OM1 for F = F MSY. 2095

80 SCRS/2013/035 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) AN EXAMPLE MANAGEMENT STRATEGY EVALUATION OF A HARVEST CONTROL RULE Laurence T. Kell 1, Gorka Merino 2, Paul De Bruyn 1, Josetxu Ortiz de Urbina 3, Haritz Arrizabalaga 2, Josu Santiago 2 and Hilario Murua 2 SUMMARY We show for North Atlantic Albacore how Management Strategy Evaluation (MSE) can be used to evaluate the performance of a limit reference point (LRP) as part of a harvest control rule (HCR). RÉSUMÉ Nous montrons pour le germon de l'atlantique Nord comment l évaluation de la stratégie de gestion (MSE) peut être utilisée pour évaluer la performance d'un point de référence limite (PRL) dans le cadre d'une norme de contrôle de la ponction (HCR). RESUMEN Se muestra, para el atún blanco del norte, cómo se puede utilizar la Evaluación de la estrategia de ordenación (MSE) para evaluar el rendimiento de un punto de referencia límite (LRP) como parte de una norma de control de la captura (HCR). KEYWORDS Albacore, FLR, Harvest Control Rule, Management Procedure, Management Strategy Evaluation, OperatingModel, R. 1 ICCAT Secretariat, C/Corazón de María, Madrid, Spain; Phone: Fax: AZTI-Tecnalia, Herrera Kaia Portualdea, 20110, Pasaia, Spain; Phone: Fax: Instituto Español de Oceanografía IEO- CO Málaga, Pto. Pesquero s/n, Fuengirola (Málaga), Spain; Phone: Fax:

81 1. Introduction In this paper we show how Management Strategy Evaluation (MSE) can be used to evaluate the performance of a limit reference point (LRP) as part of a harvest control rule (HCR). The Precautionary Approach Garcia [1996] requires stock status to be assessed relative to limits and targets, to predict outcomes of management alternatives for reaching the targets and avoiding the limits, and to characterize the uncertainty in both of these. These requirements impose specific needs for research, stock assessments, monitoring and management. Under the Precautionary Approach (PA) the use of a harvest control rule (HCR) is recommended to specify in advance what actions should be taken when limits are reached. However, although harvest control rules may include several precautionary elements, it does not necessarily follow that they will be precautionary in practice (Kirkwood and Smith, 1995). Since many harvest control rules are not evaluated formally to determine the extent to which they achieve the goals for which they were designed, given the uncertainty inherent in the system being managed (Punt 2008). For this reason Management Strategy Evaluation (MSE) using simulation modelling has increasingly been used to evaluate the impact of the main sources of uncertainty inherent in the system being managed (Cooke 1999, McAllister et al. 1999, Kell et al. 1999). 2. Material and Methods MSE is the process of testing generic harvest strategies or Management Procedures (MPs) using an Operating Model (OM). An OM is a mathematical statistical model that describes the resource dynamics in simulation trials and for generating resource monitoring data when projecting forward Rademeyer et al. (2007). A MP is the combination of pre-defined data, together with an algorithm to which such data are input to provide a value for a TAC or effort control measure. The intention is to demonstrate, through simulation trials, that the MP provides robust performance in the presence of uncertainties. In some cases the MP is ran as an auto-pilot, i.e. once the data are given to the MP and a TAC is then generated which is then implemented without negotiation. However, MSE has also been used to develop elements of management plans, e.g. to set inter-annual bounds on TACs (Kell et al., 2006). MSE involves a number of steps (Punt and Donovan, 2007) i.e. 1. Identification of management goals (and performance measures to quantify the extent to which those goals have been achieved). 2. Selection of hypotheses which impact on the risk of not achieving those goals, and development of Operating Models which represent those hypotheses. 3. Conditioning of the Operating Models on the available data and knowledge (and possible rejection of hypotheses [or combinations of hypotheses] which are not compatible with those data and knowledge). 4. Identification of candidate management strategies. 5. Simulation of the performance of the management strategies by projecting the Operating Model forward in which management is set using the management strategy. 6. The evaluation of the management strategies based on the performance measured. 2.1 Management Objectives To quantify how well management objectives are met by a particular management strategy is done using performance statistics. The precautionary approach (PA) recommends that F MSY be considered as an upper limit on harvest rate. However, the ICCAT convention was signed prior to the development of the PA and so there is no implicit mention of limit reference points in the convention. The original management objective of ICCAT is to provide the maximum continuing catch, interpreted as using maximum sustainable yield (MSY) as a target. Recently the tuna Regional Fisheries Management Organisations (trfmos) have developed a common framework for the provision of management advice, i.e. the Kobe Framework de Bruyn et al. (2012). 2097

82 Amongst other things this requires advice to be provided relative to F MSY and B MSY and for management options to be reported with respect to the resulting probability over time of B < B MSY and F > F MSY. This requires stock assessment advice to be presented as phase plot (Figure 1) where the x-axis corresponds to biomass relative to B MSY and the y-axis corresponds to harvest relative to F MSY. Quadrants are defined for the stock and fishing mortality relative to B MSY and F MSY ; i.e. red when B < B MSY and F > F MSY, green if B > B MSY and F < F MSY, and yellow otherwise. I.e. the red quadrant refers to an overfished stock subject to overfishing, green to a stock which is neither overfished or subject to overfishing and the yellow to a stock which is either overfished stock or subject to overfishing. Management advice from the ICCAT Commission is based on recommendation [Rec-11-13] i.e. 1. For stocks that are not overfished and not subject to overfishing (i.e., stocks in the green quadrant of the Kobe plot), management measures shall be designed to result in a high probability of maintaining the stock within this quadrant. 2. For stocks that are not overfished, but are subject to overfishing, (i.e., stocks in the upper right yellow quadrant of the Kobe plot), the Commission shall immediately adopt management measures, taking into account, inter alia, the biology of the stock and SCRS advice, designed to result in a high probability of ending overfishing in as short a period as possible. 3. For stocks that are overfished and subject to overfishing (i.e., stocks in the red quadrant of the Kobe plot), the Commission shall immediately adopt management measures, taking into account, inter alia, the biology of the stock and SCRS advice, designed to result in a high probability of ending overfishing in as short a period as possible. In addition, the Commission shall adopt a plan to rebuild these stocks taking into account, inter alia, the biology of the stock and SCRS advice. 4. For stocks that are overfished and not subject to overfishing (i.e. stocks in the lower left yellow quadrant of the Kobe plot), the Commission shall adopt management measures designed to rebuild these stocks in as short a period as possible, taking into account, inter alia, the biology of the stock and SCRS advice. This translates into the objectives summarized in Table 1. However, important elements are missing, e.g. probability and associated risk levels, time scales and the power of the scientific advice framework to actually predict the performance of management. There is also the original objective (O0) of achieving the maximum continuing catch and other objectives of keeping stakeholders happy. Within a body managing Areas Beyond National Jurisdiction (ABNJ) translates into potentially conflicting and not always explicitly expressed objectives. Particularly since there are implicit economic and human welfare objectives as well. For such objectives we look at discounted yield (O4) and effort (O5) as proxies for revenue or food supply and employment Performance Measures Performance measures are summarized in Table 2. We also ran two additional OM trials, where we assume perfect implementation of F based management, i.e. for F min and F MSY. The former provides a base line for stock recovery given the biology of the stocks and the later provides estimates of MSY and B MSY the variability to be expected when F equals F MSY. Any MP will perform below the base lines levels allowing relative as well as absolute comparisons to be made. 2.2 System Dynamics The major part of conducting an MSE is the selection of hypotheses which impact on the risk of not achieving management objectives, these relate to hypotheses about population dynamics and the behaviour of fleet including their response to management. Operating Models have to be developed which represent those hypotheses, there will always be alternative aims when evaluating MPs and different ways to condition OMs. 2.3 Conditioning of the Operating Model Conditioning of the Operating Models on the available data and knowledge (and possible rejection of hypotheses [or combinations of hypotheses] which are not compatible with those data and knowledge). There are four main ways to construct operating models Kell et al., (2006). Many evaluations of may use all four to some extent. The amount of knowledge, data requirements, and complexity of implementation differs quite markedly among these four types. Considerations below are expressed mostly in a Bayesian context, but there are other valid philosophies when constructing operating models. 2098

83 1. The operating model is the currently-used stock assessment model. Although use of the assessment model as the operating model seems to imply that assessment models describe nature almost perfectly, if a OMP cannot perform well when reality is as simple as implied by an assessment model, it is unlikely to perform adequately for more realistic representations of uncertainty. Basing an operating model on the current assessment model has arguably the lowest demands for knowledge and data. 2. The operating model is a model that can represent all of the available (and valid) data. The values for the parameters of the operating model are based only on the data for the fishery under consideration (i.e. in Bayesian models, priors would be non-informative, so that only data would speak). This approach is based on the idea that all relevant data sets are available and that only data matter when considering future events. The operating model need not be identical to the models underlying the assessments used as part of the OMPs. This approach assumes that, with no information to the contrary, the future will be similar to the past, which is a strong assumption. 3. As for (2) except that, in Bayesian models, priors would describe in a formal way the knowledge of scientists related to the validity of information sources. Probabilities, other than those available from data, may come from, for example, meta-analyses. This is still a data-orientated approach, but other data sources than those for the fishery under consideration have an impact when conditioning the operating model. 4. As for (3) except that the emphasis is on expert beliefs and other a priori information about the processes that may aspect the behaviour of management systems in the future (i.e. the focus is on the future, not on putting historical data). This is a less data-, and more hypothesis-orientated approach. For example, climatic change studies may show that a regime shift is possible (even though one has never been seen in the historical data sets) and should be taken into account when selecting ways to provide management advice. It is important therefore that operating models are flexible so that they can deal with such factors. In this example we use (1) to ensure continuity of advice and transparency to working group members, i.e. the OM was based on Multifan-CL. However simulation test using Multifan-CL as part of an MP would be computationally intensive we use a simpler stock assessment model for the MP i.e a biomass dynamic model. For conditioning of the OM we posed three hypotheses about the biological parameters and then ran Multifan- CL to convergence (see SCRS/2013/34 for full details). 2.4 Management Strategies Modelling a management strategy requires identification of candidate strategies. A generic MP based on a biomass dynamic model is described in SCRS/2103/33. This includes an estimator of stock parameters based on time series of catch and indices of abundance, calculation of reference points with associated uncertainty and an algorithm for projection based on a hockey stick harvest control rule (HCR). The HCR also includes constraints on changes in F (and hence implicitly on capacity and effort) and catches (hence revenue). A HCR is also shown as part of the phase plot in Figure 1, the orange line is the HCR and where a harvest rate (y-axis) is set for a given stock biomass (x-axis). The black line is the replacement line, i.e. for a given stock biomass any harvest rate above the black line will cause the stock to decline and any harvest rate below the line will cause the stock to increase. Therefore for the target harvest rate (i.e. the horizontal segment of the HCR) the target biomass is given by the intersection of the two lines. If the stock declines below the break point (i.e. a trigger biomass) then harvest rate is reduced progressively to a minimum level of harvest rate at a biomass level referred to as B lim. 2.5 Simulations Simulation are used to evaluate the performance of the MP by projecting the OM forward and setting quotas using the management strategy implicitly coded in the MP, i.e. using feedback between the MP and the OM. We only considered 3 OMs based on biology. Other hypotheses could have been considered for example some fleets are assumed to have the same selectivity as there was insufficient information in the data to estimate these independently selectivity was also assumed not to change over time. Both assumptions are unlikely to be true. The catch and size data are also not true observations have been generated in a variety of ways. There is therefore considerable uncertainty in the fleet data and assumptions. 2099

84 2.5.1 Uncertainty It is important to consider appropriate sources of uncertainty; traditional stock assessments mainly consider only uncertainty in observations and process (e.g. recruitment). However, uncertainty about the actual dynamics (i.e. model uncertainty) has a larger impact on achieving management objectives (Punt 2008). Therefore when providing management advice it is important to consider appropriate sources of uncertainty. Rosenberg and Restrepo (1994) catagorised uncertainties in fish stock assessment and management as being related to: fi Processes; caused by disregarding variability, temporal and spatial, in dynamic population and fisheries processes; Observations; sampling error and measurement error; Estimation; arising when estimating parameters of the models used in the assessment procedure; Models; related to the ability of the model structure to capture the core of the system dynamics; Implementation; where the effects of management actions may differ from those intended. Uncertainty due to implementation is often ignored but it is recognised to be an important aspect related to the success of management plans (Peterman 2004). However, little actual work has been done on including implementation error in MSE, to do so requires a better understanding of fishers response to management (e.g. Tidd, Palmer et al. in prep.). Particularly since the effects of a HCR can be quite different from those intended because of the response of fishers to economic incentives and HCRs are generally poorly equipped to represent human welfare and MSEs tend not to represent implementation error well Milner-Gulland and Rowclifie [2011]. Sources of uncertainty related to Models include: structural uncertainty; due to inadequate models, incomplete or competing conceptual frameworks, or where significant processes or relationships are wrongly specified or not considered. Such situations tend to be underestimated by experts (Morgan and Henrion, 1990) and, value uncertainty; due to missing or inaccurate data or poorly known parameters Here we consider uncertainty due to Process error in future recruitment Measurement error in historic catches Estimates of stock status and reference points obtained from a biomass dynamic model Structural uncertainty, i.e. about the true biological processes. 2.6 Evaluation Is based on comparing the performance measures and hence management objectives and the trade-ofis between them. 2.7 Software Software used was a biomass production model implemented as a package in R, this allows it to be used with a variety of other packages for plotting, summarising results and to be simulation tested, e.g. as part of the FLR tools for management strategy evaluation (Kell et al., 2007). 3. Examples Figures 10, 11 and 12 plot a number of possible realisations of the simulated projections from the OM (coloured lines) along with the medians and inter-quartiles (black lines), for F : F MSY, B : B MSY and catch. Rows are HCR options and columns the scenario. The first HCR is for reference, i.e. what would the performance be if we could actually fish at F MSY? Within the Monte Carlo trials (i.e. each choice of scenario and HCR) random numbers were used reused, e.g. so that realisation (i.e. iteration) n is comparable across trials. These plots show that basing judgment on the expected outcomes (e.g. the medians) is very misleading since yield and F (and hence effort) will be very different. I.e. it will show a large amount of variability and 50% of the time the actual outcomes will be outside the confidence range. Therefore evaluation of the HCRs should be based on appropriate performance measures see Figures 14 and

85 However, first we look at the recovery rate in Figure 14, this shows the probability of the stock recovering to be greater than B MSY while F is maintained below F MSY, i.e. the stock is in the green Kobe quadrant. Within a OM scenario the choice of B Trigger (i.e. 0.6 or 0.8) has no effect, it is the choice of the F Target that matters. The stock recovers quickest if the dynamics correspond to OM1 and slowest if they correspond to OM2, i.e. the mortality in the plusgroup determines recovery rate. Performance measures related to sustainability are presented in Figure 16, these show the probability after recovery of staying in the Green Quadrant after recovery. F being less than F MSY SSB being greater than > B MSY SSB being greater than B Loss, i.e. the smallest spawning biomass observed in the series of annual values of the spawning biomass. In this case B Loss, is the same as the Minimum Biological Acceptable Level (MBAL). A spawning biomass level below which, observed spawning biomasses over a period of years, are considered unsatisfactory and the associated recruitments are smaller than the mean or median recruitment (Serchuk and Grainger, 1992). Performance measures related to economics are presented in Figure 15, these show the inter-quartiles (bars) and medians (dots) for median yield after recovery. Total Yield Discounted (at 5%) Total Yield median Effort after recovery. Total Effort Discounted (at 5%) Total Effort Values have been scaled by the maximum value of that measure across all trials (i.e. so all values lie between 0 and 1). If the main objective is sustainability then any HCR that fails to meet the evaluation criteria in Figure 14 should not be selected for implementation. I.e. an FT arget of 0.7 times F MSY has a less than 0.5 probability of the stock staying in the Green Quadrant after recovery. There is also a high probability of SSB falling to a lower level than has been seen in the entire historical time series. Therefore it appears that the target F level is more important than the B Triggers considered. The choice of HCR can be refined by considering the Economic objectives, i.e. it may be possible to select a preferred option that satisfies the sustainability criteria. 4. Discussion The study presented in this paper is not intended to be used for management advice, rather it is intended to demonstrate how MSE can be used to develop a LRP that meets management objectives given uncertainty consistent with the Precautionary Approach. I.e. to provide an example of a framework that can be used to perform an actual MSE. As noted above the stages of conducting an MSE include Identification of management goals was done for sustainability objectives based on recovery of a stock to the Green Kobe Quadrant, ensuring it then stays there. For economic objectives differences in long-term yield and effort (e.g. capacity) levels were compared. However, to do a full economic impact assessment analyses should be done by fleet. These objectives were mapped to Performance Measures that can allow the benefits and tradeoffs of the different Management Strategies to be compared. However, important issues such as probability and risk levels and time scales have not been explicitly stated by the Commission these need to be agreed before any evaluation. Hypotheses about the stock dynamics were based on the biological assumptions, then Multifan-CL was used to estimate stock parameters using the WG data. However, there is considerable uncertainty about the catch data, fishing behaviour and stock structure for example (Fonteneau 2007). All of which may have a major impact on the risk of not achieving the management goals. It doesn t mean that it is necessary to create Operating Models to represent all possible hypotheses but ways of ensuring that advice is robust to what is known should include in the MSE. For example by using stock assessment models in the MP that do not require catch-at-age data to be produced before and assessment is conducted. 2101

86 Conditioning of the Operating Models is an important step and more thought needs to be conducted in order to determine what hypotheses to consider and how to weight them (SCRS ). How to include appropriate data and knowledge and possible rejection of hypotheses (or combinations of hypotheses) which are not compatible with those data and knowledge. For efficiency robustness test may help, i.e. to allow hypotheses to be discounted as they have little effect on the Performance Measures. This will require uncertainty to be considered in a systematic way, i.e. for selecting scenarios agreed a priori. Management Strategies considered were based on a biomass dynamic model, with a hockey stock harvest control rule, which set a TAC each year. There are a wide range of options that could have been considered e.g. limiting inter-annual variability in TACs or effort or setting TACs for three years. This is relatively easy to do in the framework presented but requires a dialogue with managers. However, as there has been no agreed ASPIC/Biomass Dynamic stock assessment and it will be necessary for the Albacore WG to agree on this before any LRP can be developed. Uncertainty A major impact on the OM are the assumptions about the catch data, the CV of the residuals to the catch-at-age is >130%, as well as noise there will be considerable bias. i.e. if the basic data were bootstrapped then very differences outcomes may be generated. Stock recruitment assumptions may have a big effect on outcomes and could be explored further, i.e. as robustness trials. 5. Conclusion A full MSE requires appropriate uncertainty to be considered, particularly in the choice of OM scenarios. One way of doing this is base the choice of scenarios are based on a factorial design. A full factorial experiment is one whose design consists of two or more factors, each with discrete possible values or levels, and where experimental units take on all possible combinations of these levels across all factors. Such a design is better able to represents the complexity of the real world and allows an evaluation of whether the effect of one factor depends on the level of another factor. The potentially large number of combinations in a full factorial design may mean that it is not possible to run them all in the time available in a stock assessment working group. Therefore a fractional factorial design in which some of the possible combinations are omitted may be preferred. However, when conducting a Management Strategy Evaluation a large number of scenarios need to be considered to evaluate the main sources of uncertainties, i.e. the Operating Models (OM) will need to be conditioned on a wider range of data and knowledge that routinely considered within a stock assessment. In other words while only a few scenarios are routinely be considered within an ICCAT stock assessment, many more scenarios will need to be run as part of an MSE. This presents a potential problem if the Scenarios from an MSE result in different conclusions from those ran in a stock assessment. It is proposed that a base case be proposed and then factors with levels that represent the main uncertainties. In the stock assessment WG the main effects can be evaluated by varying 1 factor at a time. Hopefully this will allow the stock assessment to bracket the main uncertainty and act as a simple screening experiment, to determine the factors have the greatest influence on the perception of stock dynamics. Based on the identification of the most important factors, a multilevel designed experiment can then be developed for the MSE that includes interactions between factors. Bibliography J. Cooke. Improvement of fishery-management advice through simulation testing of harvest algorithms. ICES Journal of Marine Science: Journal du Conseil, 56(6):797{810, P. de Bruyn, H. Murua, and M. Aranda. The precautionary approach to fisheries management: How this is taken into account by tuna regional fisheries management organisations (RFMOs). Marine Policy, S. Garcia. The precautionary approach to fisheries and its implications for fishery research, technology and management: an updated review. FAO Fisheries Technical Paper, pages 1{76, L. Kell, C. O'Brien, M. Smith, T. Stokes, and B. Rackham. An evaluation of management procedures for implementing a precautionary approach in the ICES context for North Sea plaice (pleuronectes platessa l.). ICES Journal of Marine Science: Journal du Conseil, 56(6): , L. T. Kell, J. A. De Oliveira, A. E. Punt, M. K. McAllister, and S. Kuikka. Operational management procedures: an introduction to the use of evaluation frameworks. Developments in Aquaculture and Fisheries Science, 36:379{407,

87 L. Kell, I. Mosqueira, P. Grosjean, J. Fromentin, D. Garcia, R. Hillary, E. Jardim, S. Mardle, M. Pastoors, J. Poos, et al. Flr: an open-source framework for the evaluation and development of management strategies. ICES Journal of Marine Science: Journal du Conseil, 64(4):640, G. Kirkwood and A. Smith. Assessing the precautionary nature of fishery management strategies. Fisheries and Agriculture Organization. Precautionary approach to fisheries. Part, 2, M. McAllister, P. Starr, V. Restrepo, and G. Kirkwood. Formulating quantitative methods to evaluate fisherymanagement systems: what fishery processes should be modelled and what trade-ofis should be made? ICES Journal of Marine Science: Journal du Conseil, 56(6):900{916, E. J. Milner-Gulland and J. M. Rowclifie. Conservation and sustainable use: a handbook of techniques. OUP Catalogue, R. A. Rademeyer, fie. E. Plagfianyi, and D. S. Butterworth. Tips and tricks in designing management procedures. ICES Journal of Marine Science: Journal du Conseil, 64(4):618{625, Figure 1. A phase plot of harvest rate relative to F MSY and stock biomass relative to B MSY, with an example of a Harvest Control Rule (orange) and replacement line (black). 2103

88 Figure 2. Observed size measurements (histograms) and fitted size distributions (lines). Figure 3. Catch-at-age residuals from Multifan-CL fits. 2104

89 Figure 4. Comparison of surplus production functions (dots) with OM age based production curves (lines). Figure 5. Comparison of time series of harvest rate and stock biomass derived from a Monte Carlo simulation based on the variance covariance matrix. 2105

90 Figure 6. Comparison of biomass dynamic parameters derived from a Monte Carlo simulation based on the variance covariance matrix. Figure 7. Comparison of reference points derived from a Monte Carlo simulation based on the variance covariance matrix. 2106

91 Figure 8. Comparison of harvest rate and harvest rate relative to F MSY derived from a Monte Carlo simulation based on the variance covariance matrix. Figure 9. Comparison of stock biomass and stock relative to B MSY derived from a Monte Carlo simulation based on the variance covariance matrix. 2107

92 Figure 10. Harvest rate relative to F MSY worm plots (colour) with inter quartiles and medians (black). Figure 11. Stock biomass relative to B MSY worm plots (colour) with inter quartiles and medians (black). 2108

93 Figure 12. Yield worm plots (colour) with inter quartiles and medians (black). Figure 13. Probability of stock recovery, i.e. being in the green quadrant of the Kobe Phase plot. 2109

94 Figure 14. Summary of sustainability performance measures. Figure 15. Summary of economic performance measures. 2110

95 SCRS/2013/117 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) AN EVALUATION OF DIFFERENT APPROACHES FOR MODELLING UNCERTAINTY IN ASPIC AND BIOMASS DYNAMIC MODELS Laurence T. Kell 1, Josetxu Ortiz de Urbina 2, Paul De Bruyn 1, Iago Mosqueira 3, Arni Magnusson 4 SUMMARY In this paper we compare the estimates of uncertainty obtained for a biomass dynamic stock assessment model. We do this for approaches based on simulation (i.e. the bootstrap, jack knife and Monte Carlo Markov Chain), likelihood profiling and the delta method. RÉSUMÉ Le présent document compare les estimations de l'incertitude obtenues pour un modèle d'évaluation des stocks dynamique de la biomasse. Nous réalisons ceci pour des approches basées sur la simulation (c.-à-d. bootstrap, "jack knife" et MCMC-Monte Carlo Markov Chain), profilage des vraisemblances et méthode delta). RESUMEN En este documento comparamos las estimaciones de incertidumbre obtenidas para un modelo de evaluación de stock de dinámica de biomasa. Esto se realiza para enfoques basados en simulación (es decir, bootstrap, jack knife y cadena Monte Carlo Markov), perfiles de verosimilitud y el método delta. KEYWORDS Albacore, Assessment, Biomass Dynamic, Bootstrap, Bayes, Likelihood Profile, Delta Method, Uncertainty 1. Introduction A main management objective of ICCAT is to maintain the populations of tuna and tuna-like fishes at levels which will permit the maximum sustainable catch. Scientific advice within ICCAT is therefore based on MSYbased reference points. In common with other tuna Regional Fisheries Management Organisations (trfmos) advice is presented showing the probabilities of a stock being greater than B MSY and fishing mortality being less than F MSY in the historical assessment and for different management options projected into the future. To estimate these probabilities a variety of methods have been used by assessment working groups, e.g. bootstrapping and Bayesian simulation. Magnusson et al. [2012] compared three methods for estimating uncertainty in aged based stock assessment (i.e. the bootstrap, delta method and Monte Carlo Markov Chain simulation); they showed through simulation that all three methods generated too narrow confidence intervals, underestimating the true uncertainty. In this paper we compare estimates of uncertainty from biomass dynamic models, e.g. ASPIC Prager (1992), and discuss the consequences for advice provided by the SCRS. 2. Material and Methods There are three main ways to estimate uncertainty within stock assessment models, frequentist, Bayes and likelihood. The frequentist approach treats a parameter as an unknown that poses a true value and estimates confidence intervals rather than probabilities. 1 ICCAT Secretariat, C/Corazón de María, Madrid, Spain; Phone: Instituto Español de Oceanografía IEO- CO Málaga, Pto. Pesquero s/n, Fuengirola (Málaga), Spain; Phone: Fax: European Commission, Joint Research Center, IPSC/Maritime Affairs Unit, Via E. Fermi 2749, Ispra VA, Italy 4 Marine Research Institute, Skulagata 4, PO Box 1390, 121 Reykjavik, Iceland 2111

96 In the likelihood and Bayes approaches parameters are considered to be random variables (Wade, 2000). Bayes methods provide a formal mechanism for updating belief as new data becomes available. Bayesian methods also require a formal description of prior beliefs that reflect current knowledge. Likelihood methods do not need such priors, though as for all methods they are dependent on belief about the appropriate representation of processes and form of error distributions Patterson et al. [2001]. Six methods for estimating parameter uncertainty were compared, i.e. the bootstrap, jack knife, Monte Carlo simulation of the input data, Bayesian estimation using Markov Chain Monte Carlo (MCMC) simulation, the delta method and likelihood profiling. 2.1 Data The data used were time series of the total catch biomass and a single index of abundance as used in the last assessment for North Atlantic swordfish (SCRS/2009/016). The index with the 10th and 90th is plotted in in Figure Methods Dynamics were modelled by a biomass dynamic model, where the biomass of a stock next year (B t+1 ) is equal to the biomass this year B t less the catch (C t ) plus the surplus production (P t ) where P is given by the Pella- Tomlinson surplus production function. 2.3 Estimation Estimation of parameters was conducted using the biodyn package based on ADMB and implemented in R. It was assumed that the index was a proxy for stock biomass Bootstrap Bootstrapping is a method for providing estimates of a sampling distribution for a statistic. It is routinely used by the SCRS to provide estimates of uncertainty from assessment packages such as ASPIC or VPA2box using the residuals from the original fit Jackknife The jackknife or leave one out procedure is a cross-validation technique to estimate the bias and variance of an estimator. It is similar to bootstrapping but the statistic is recomputed by leaving out one or more observations at a time from the sample set. The new set of replicates of the statistic then allow estimates of the variance and the bias of a statistic to be calculated Delta Method The delta method can provide estimates of the standard deviations for derived quantities in many (but not all) situations. It is based on finding approximations based on Taylor series expansions to the variance of functions of random variables. As part of the fitting process the Hessian (i.e. the matrix of second-order partial derivatives) is calculated by ADMB. The inverse of the Hessian matrix approximates the variance/covariance matrix of the parameter estimates. The standard errors of derived parameter, i.e. statistics that are not actual parameters in the model but derived from them, can also be estimated Likelihood Profile The delta method can be used to calculate standard deviations for derived quantities in many (but not all) situations. It is based on finding approximations based on Taylor series expansions to the variance of functions of random variables MCMC Markov chain Monte Carlo (MCMC) is a method for simulating a probability distribution for a statistic. It is used to approximate the posterior distribution of estimated parameters. One of the main difficulties with MCMC methods is ensuring that simulations have converged to a stationary distribution. The equilibrium distribution of 2112

97 the chain is the required posterior distribution but how do we know that the chain has reached equilibrium? A burn-in period where initial values are discarded helps. However, in complicated cases, e.g. where there is more than one local maximum or the posterior distribution is in the form of a ridge a chain, it can take a long time to move around the parameter space and a very long burn-in period may be required. While very large sample may have to be taken to ensure that the chain has not just become temporarily stuck in one part of the parameter space. For these reasons a variety of diagnostics are used to check convergence; i.e. that a stationary distribution has been reached e.g. Autocorrelation Plots measure the correlation between fit and fit+1 variable in a chain Correlation Plots can show if parameters are confounded Gelman-Rubin Diagnostic tests that the burn-in is adequate and requires that multiple starting points be used. Geweke Diagnostic, if burn-in is adequate, then the mean of the posterior distribution of from the first half of the chain should equal the mean from the second half of the chain 3. Results First the results and residual diagnostics from the assessment are presented and then we compare the different estimates of uncertainty for B : B MSY and F : F MSY. 3.1 Assessment Time series of estimated harvest rate and stock biomass compare and the assumed catch are presented in Figure 2, for both the aspic and biodyn R packages. The results were identical. 3.2 Residual Diagnostics Inspection of residual diagnostics are important to check that the assumptions are met both when fitting a model and conducting simulations. A main assumption of the assessment model used is that the CPUE series is a proportional to stock biomass, therefore we plot the observations against the fitted values in Figure 3. If the assumption is correct then the points should be distributed about the y = x line. However, it can be seen that the 95% CI of a linear regression (blue shaded area) fitted to the data do not contain the y = x line and that there appears to be autocorrelation between points. Figure 4 plots the residuals against year to evaluate whether there is any systematic pattern that may suggest the index is not a proxy for stock biomass. It was also assumed that the residuals are normally distributed and there is no autocorrelation. Therefore the Q-Q plots in Figures 5 compare a sample of data on the vertical axis to a statistical population on the horizontal axis, in this case a normal distribution. If the points follow a strongly nonlinear pattern this will suggest that the data are not distributed as a standard normal i.e. X N(0; 1). Any systematic departure from a straight line may indicate skewness or over or under dispersion. In Figure 6 the residuals are plotted against each other with a lag of 1 to identify autocorrelation. There are significant autocorrelations. There appears to be significant positive autocorrelation. It is also assumed that variance does not vary with the mean, i.e. there is no heteroscedasticity, this assumption is evaluated in Figure 7 where the residuals are plotted against the fitted values. 3.3 Densities The densities of stock relative to B MSY and of harvest rate relative to F MSY are plotted for each method of estimating parameter uncertainty in Figures 8 and 9. The Kobe phase plots, with points showing stock relative to B MSY and harvest rate relative to F MSY, are then shown in Figure 10 for the various method of estimating parameter uncertainty. 2113

98 4. Discussion Estimates of stock relative to B MSY and harvest rate relative to F MSY were compared for 6 different procedures. Estimates of uncertainty obtained from the same data and stock assessment model vary depending on the method used to estimate the uncertainty, i.e. their confidence intervals or probability distributions. The densities obtained for B : B MSY varied, with Monte Carlo simulation of the CPUE Index giving the narrowest estimate of uncertainty and MCMC and profiling giving the broadest. Similar results were seen for F : F MSY..Violation of the assumptions with respect to the indices used for fitting may result in biased estimates of estimated parameters, and hencereference points and stock trends. For variance estimates obtained from bootstrapping often assume that residuals are Independently and Identically Distributed (i.i.d.). Although uncertainty estimates was compared for a variety of methods, unlike Magnusson et al. (2012) we did not compare the estimates to known values through simulation, this should be done using Management Strategy Evaluation (MSE) to evaluate how to use estimates of uncertainty from stock assessment as part of a harvest control rule (HCR). Bibliography L. Kell, I. Mosqueira, P. Grosjean, J. Fromentin, D. Garcia, R. Hillary, E. Jardim, S. Mardle, M. Pastoors, J. Poos, et al. Flr: an open-source framework for the evaluation and development of management strategies. ICES Journal of Marine Science: Journal du Conseil, 64(4):640, A. Magnusson, A. E. Punt, and R. Hilborn. Measuring uncertainty in fisheries stock assessment: the delta method, bootstrap, and mcmc. Fish and Fisheries, K. Patterson, R. Cook, C. Darby, S. Gavaris, L. Kell, P. Lewy, B. Mesnil, A. Punt, V. Restrepo, D. W. Skagen, et al. Estimating uncertainty in fish stock assessment and forecasting. Fish and Fisheries, 2(2):125{157, M. Prager. Aspic-a surplus-production model incorporating covariates. Coll. Vol. Sci. Pap., Int. Comm.Conserv. Atl. Tunas (ICCAT), 28:218{229, P. R. Wade. Bayesian methods in conservation biology. Conservation Biology, 14(5):1308{1316,

99 Figure 1. Catch per unit effort index with 10th and 90th confidence interval. Figure 2. Time series of estimates of stock biomass, harvest rate and yield from ASPIC and biodyn assessments 2115

100 Figure 3. Observed CPUE verses fitted, blue line is a linear regression fitted to points, black the y=x line. Figure 4. Residuals by year, with lowess smoother and SEs. 2116

101 Figure 5. Quantile-quantile plot to compare residual distribution with the normal distribution. Figure 6. Plot of autocorrelation, i.e. residual t+1 verses residual t. 2117

102 Figure 7. Plot of residuals against fitted value, to check variance relationship. Figure 8. Densities of stock relative to B MSY for each method of estimating parameter uncertainty. 2118

103 Figure 9. Densities of harvest rate relative to F MSY for each method of estimating parameter uncertainty. Figure 10. Kobe phase plot with points showing stock relative to B MSY and harvest rate relative to F MSY for each method of estimating parameter uncertainty. 2119

104 SCRS/2013/162 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) LIKELIHOOD PROFILES BY DATA COMPONENTS TO EVALUATE INFORMATION CONTENT OF INDICES OF ABUNDANCE FOR NORTH ATLANTIC SWORDFISH Laurence T. Kell 1, Josetxu Ortiz de Urbina 2, Paul De Bruyn 1 SUMMARY We use likelihood profiling by data component, i.e. for each catch per unit effort (CPUE) series as a data exploratory tool. The approach allows the information on key parameters in each time series to be evaluated. RÉSUMÉ Nous avons recours au profilage des vraisemblances par élément des données, c.-à-d. pour chaque série de capture par unité d'effort (CPUE) comme outil exploratoire des données. L'approche permet d'évaluer l'information sur les paramètres fondamentaux dans chaque série temporelle. RESUMEN Se utilizaron perfiles de verosimilitud por componente de datos, a saber, por cada serie de captura por unidad de esfuerzo (CPUE) como herramienta exploratoria de datos. El enfoque permite evaluar la información sobre los parámetros clave en cada serie temporal. KEYWORDS Biomass Dynamic, Diagnostics, Stock Assessment 1. Introduction Biomass dynamic models are widely used in ICCAT for stock assessment and advice parameters are estimated by setting to time series of total catch and standardised catch per unit effort (CPUE) from fisheries. The latter are assumed to track stock abundance. However it is not uncommon for such indices to contain sufficient information to estimate both parameters. Also indices may be connecting and fitting therefore may involve weighted averages of contradictory CPUE data. This generally produces parameter estimates intermediate than would be obtained from the data sets individually Schnute and Hilborn (1993), who point out that the most likely parameter values are not intermediary to conflicting values; instead, they occur at one of the apparent extremes. We therefore use the ASPIC biomass model to explore uncertainty due to contradictory trends in time series of catch per unit effort (CPUE). We do this by calculating likelihood profiles for the parameter K (carrying capacity or unfished biomass B0) and MSY (maximum sustainable yield). Parameterising the assessment model in terms of MSY and K is preferred to r and K since providing management advice requires management target and limit reference points (Martell et al., 2007) and this way uncertainty in the reference points can be evaluated directly. 2. Materials and methods We use Piner Plots (ISC/11/BILLWG-3/01) which show the likelihoods of the different data components for a profiled parameter. This allows an evaluation of what data series are affecting the parameter. 1 ICCAT Secretariat, C/Corazón de María, Madrid, Spain; Phone: Fax: Instituto Español de Oceanografía IEO- CO Málaga, Pto. Pesquero s/n, Fuengirola (Málaga), Spain; Phone: Fax:

105 2.1 Data The data are from the North Atlantic Albacore assessment (SCRS2013-xxx), and comprise 4 long line CPUE time series one from Chinese-Taipei and the Japanese longline split into 3 periods to reflect changes in targetting. 2.2 Software Software used was a biomass production model implemented as a package in R, this allows it to be used with a variety of other packages for plotting, summarising results and to be simulation tested, e.g. as part of the FLR tools for management strategy evaluation (Kell et al., 2007). 3. Results The CPUE series from the combined index are plotted in Figure 1; as estimated in 2009 (red) and 2013 (blue), error bars are the 10th and 90th Confidence Intervals. The indices are replotted from 1980 onwards in Figure 2 to remove the large early values which otherwise make it difficult to compare recent trends. In Figure 3 the three indices as used in the 2009 (top panel) and 2013 (bottom panel) assessments are presented. These are scaled so the mean is 1 as used in the stock assessment, note that the CIs cross. In Figure 4 the variability over time is compared by plotted the 10th and 90th CIs divided by the mean; two effects are evident, i.e. that the variance in 2013 is much greater than in 2009 and that there appears to be a stepwise change in variability before the 1980s. Figure 5 compares the standardised combined index (red) with the standardised index by flag; error bars are the 10th and 90th Confidence Intervals. This is repeated for the recent period in Figure 6. Figure 7 plots the standardised indices of abundance as performed independently by flag. For each series the Residual sum of square profiles for K by each index, i.e. data components are plotted for K and MSY in Figures 8 and 9 for K and MSY. For the profiles of K only the Combined, Canada I & II and Portuguese indices gave a minimum. The indices from Morocco, Spain and USA imply that K is potentially very large. In comparison the Japanese index implies a very low value of K. For MSY the picture is more complicated, as there appears to low minima. For example in the case of Morocco. This implies MSY is low or very high. The estimated stock biomass is compared with the biomass predicted by the combined index (i.e. U=q) is shown in Figure 10. While the trajectory of yield and biomass is compared to the surplus production function in Figure Discussion Common actions to address connecting trends are to down weight data which in the opinion of the stock assessors are not representative of stock trends or to run several scenarios then combine these in the Kobe advice plots. Both approaches can also be used with stock assessment methods that use other data, e.g. Stock Synthesis that can use size composition data. However, the approach would be similar to identify what sources of data are influencing parameter estimates and derived quantities such as the stock relative to BMSY benchmarks. As pointed out by Schnute and Hilborn (1993) stock assessments sometimes, in retrospect, prove to be wrong, e.g. due to poor model assumptions or to data that do not reflect the biological process. Schnute and Hilborn (1993) demonstrated that when model or data errors are considered the most likely parameter values are not intermediary to conflicting values; instead, they occur at one of the apparent extremes. This could be because important processes may be acting (e.g. SCRS ) which could result in changes in distribution of the swordfish stock, catchability by the fisheries and potential population parameters such as virgin biomass (i.e. K) which are estimated as part of a stock assessment and used to provide reference points. Using all the standardised indices gave similar results to the combined index. However, using all actual indices as separate series allow a range of sensitivity analyses and hypotheses to be evaluated as in this study. 2121

106 Rather than using a combined index an alternative is to integrate the standardisation of catch-per-unit-of-effort into stock assessment models. The typical stock assessment includes two steps, i.e. where a GLM is used to analyse the raw catch and effort data to estimate a year-effect and then a population dynamics model is fitted to the year-effect. However, Maunder (1998) suggested that this two-step approach has several disadvantages such as including loss of information, difficulty in appropriately representing the error structure of the CPUE data, inadequate transfer of uncertainty, and reduced diagnostic ability. Using a combined index will make this even worse. Bibliography L. Kell, I. Mosqueira, P. Grosjean, J. Fromentin, D. Garcia, R. Hillary, E. Jardim, S. Mardle, M. Pastoors, J. Poos, et al. Flr: an open-source framework for the evaluation and development of management strategies. ICES Journal of Marine Science: Journal du Conseil, 64(4):640, S. Martell, W. Pine III, and C. Walters. A management oriented approach to age-structured stock assessments. Can. J. Fish. Aquat. Sci, 20, J. T. Schnute and R. Hilborn. Analysis of contradictory data sources in fish stock assessment. Canadian Journal of Fisheries and Aquatic Sciences, 50(9):1916{1923, Maunder, M.N., Integration of tagging and population dynamics models in fisheries stock assessment. Ph.D. Dissertation. University of Washington, Seattle, W. 2122

107 Figure 1. Combined index of abundance as estimated in 2009 (red) and 2013 (blue), error bars are the 10th and 90th Confidence Intervals. Figure 2. Combined index of abundance from 1980 onwards as estimated in 2009 (red) and 2013 (blue), error bars are the 10th and 90th Confidence Intervals. 2123

108 Figure 3. The three indices of abundance as used in the 2009 (top panel) and 2013 (bottom panel) assessments, scaled so the mean is 1 as used in the stock assessment; 10th CI (green), 90th CI (blue) and standardised index (red). Figure 4. 10th CI (green) and 90th CI divided by the standardised index; 2009 indices (red), 2013 indices (black). 2124

109 Figure 5. Comparison of the standardised combined index (red) with the standardised index by flag; error bars are the 10th and 90th Confidence Intervals. Figure 6. Comparison for the recent period of the standardised combined index (red) with the standardised index by flag; error bars are the 10th and 90th Confidence Intervals. 2125

110 7. Standardised indices of abundance as performed independently by flag. Figure 8. Residual sum of square profiles for K by each index, i.e. data components. 2126

111 Figure 9. Residual sum of square profiles for MSY by each index, i.e. data components. Figure 10. A comparison of the estimated stock biomass (red) with the biomass predicted by the combined index (black). 2127

112 Figure 11. Plot of the surplus production function (red) with the estimated trajectory of yield and biomass (black). 2128

113 SCRS/2013/017 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) REPORT OF THE 2013 INTER-SESSIONAL MEETING OF THE SUB-COMMITTEE ON ECOSYSTEMS (Madrid, Spain July 1 to 5, 2013) SUMMARY The intersessional meeting of the Sub-Committee on Ecosystems was held in Madrid, Spain, 1-5 July The main objective of the meeting was to advance in the assessment of the impact of tuna fisheries on sea turtle populations. RÉSUMÉ La réunion intersession du Sous-comité des écosystèmes a eu lieu à Madrid (Espagne) du 1er au 5 juillet L'objectif principal de cette réunion consistait à faire progresser l'évaluation de l'impact des pêcheries de thonidés sur les populations de tortues marines. RESUMEN La reunión intersesiones del Subcomité de ecosistemas se celebró en Madrid, España, del 1 al 5 de julio de El principal objetivo de la reunión era avanzar en la evaluación del impacto de las pesquerías de túnidos sobre las poblaciones de tortugas marinas. 1. Opening, adoption of Agenda and meeting arrangements The Meeting was held at the offices of the ICCAT Secretariat in Madrid, Spain from July 1-5, On behalf of the ICCAT Secretariat, Dr. Paul de Bruyn welcomed the Sub-Committee. The Sub-Committee on Ecosystems Co-Conveners, Dr. Shannon L. Cass-Calay (USA) and Dr. Alex Hanke (Canada) then described the objectives and logistics of the meeting. The Agenda was adopted with changes (Appendix 1). The List of Participants is included in Appendix 2. The List of Documents presented at the meeting is attached as Appendix 3. The following participants served as rapporteurs: Section Items 1, 10-11, 14 Items 2-5 Item 6 Item 7 Item 8 Item 9 Item 12 Item 13 Rapporteurs S. Cass-Calay A. Hanke A. Hanke, P. de Bruyn, P. de Bruyn A. Domingo, B. Giffoni T. Nalovic, P. de Bruyn T. Trott J. Pereira Part I: Ecosystem The Agenda as adopted by the Sub-Committee included three sections considered significant to the process of implementing an Ecosystem-based Fisheries Management (EBFM) approach for which no new information was provided. These sections were 2, 3 and 4 and were entitled respectively: (a) Review of new information on the implementation of ecosystem effects in enhanced stock assessments; (b) Review of new information on models that quantify ecosystem dynamics and predict the impact of perturbations on its components. 2129

114 It is not clear why there was a lack of participation in this section but the Sub-Committee felt it may have something to do with the limited capacity by CPCs to participate in this Sub-Committee. Alternatively, we may lack the expertise in the SCRS to fully participate in this branch of science. The Sub-Committee recommends that this be integrated in the strategic plan of the SCRS in the future so that we might increase the capacity to deal with the demands of developing the tools that will allow the implementation of the EBFM approach. Section 5. Review new information on the implementation of ecosystem based management principles Document SCRS/2013/137 covered the Ecological risk assessment (ERA) as a method for tracking the implementation of an ecosystem approach to fisheries (EAF) management in southern Africa (the Namibian experience). The ERA method, which is qualitative, has the support of southern African countries (Namibia, South Africa and Angola) through the Benguela Current Commission (BCC), with the assistance of WWF-South Africa and FAO EAF Nansen project. The method was chosen as it is simple and has structured guidelines for tracking EAF implementations. ERA helped in identifying issues and prioritizing them. The method provided clear objectives and issues for some of the fisheries in the Benguela Current Region. Steps required towards addressing issues/risk are recommended and regular reviews are needed to assess whether progress is being made in addressing the issues. One advantage of the ERA approach is that the method can aid consensus on different issues from a diversity of stakeholders. A total of six (6) Namibian fisheries were assessed using this methodology and 4 ERA -reviews were conducted. The Sub-Committee was interested in the new approach and recognized that the author could provide valuable guidance with respect to the unpacking exercise initiated during the meeting and reported on in section 7. Continued collaboration with the representatives of the Ministry of Fisheries and Marine Resources of Namibia is recommended. 6. Review progress on the development of a test case for implementing ecosystem based fishery management Document SCRS/2013/132 provided information on the biology and ecology of a total of 18 different fish species whose distributions include the Sargasso Sea. These species are divided into four groups that correspond with ICCAT species groupings: Group 1 Principal tuna species including yellowfin tuna, albacore tuna, bigeye tuna, bluefin tuna and skipjack tuna. Group 2 Swordfish and billfishes including blue marlin, white marlin and sailfish, Group 3 Small tunas including wahoo, blackfin tuna, Atlantic black skipjack tuna (little tunny) and dolphinfish, and Group 4 Sharks including shortfin mako, blue shark, porbeagle, bigeye thresher and basking shark. For each species, information and data was provided on distribution, fishery landings, migration and movement patterns, reproduction, age and growth, food and feeding habits and ecology in relation to oceanographic parameters, primarily water temperature. The importance of Sargassum as essential fish habitat was discussed and was linked to the feeding habits of tunas and other pelagic predators. Flying fishes are an important prey species in the diet of tunas and billfishes and, as they are largely dependent on Sargassum mats as spawning habitat, the Sargasso Sea plays a fundamental role in the trophic web of highly migratory, pelagic species in the northwest Atlantic. An evaluation of existing information suggests that the importance of the Sargasso Sea to various pelagic species rests mainly with its status in relation to one or more of the following: migratory route, spawning area, nursery area, feeding area, overwintering ground or pupping area. The Sub-Committee noted that ICCAT Resolution on the Sargasso Sea states: 1) The SCRS will examine the available data and information concerning the Sargasso Sea and its ecological importance to tuna and tuna-like species and ecologically associated species. 2) The SCRS will provide an update on the progress of this work in 2014 and report back to the Commission with its findings in Regarding item 1, a preliminary review of available data and information from the Sargasso Sea habitat including tuna and other ICCAT species catches has been provided and so studies on the importance of this region can be initiated. The Sub-Committee noted that a more extensive research of this ecosystem and its importance as essential habitat for tunas and tuna-like species would require an integrated collaboration of the SCRS with scientific groups specializing in the Sargasso Sea. The Sub-Committee concluded that in order to 2130

115 accomplish objective 2, it would require a work plan, collaborative research and meetings to properly assess the importance of the Sargasso Sea as essential habitat for ICCAT tuna species. The Sub-Committee felt that this task would not be accomplished before The Sub-Committee recommends continuing the contact with Sargasso Sea research teams and the UK-Bermuda scientists to develop a scientific collaborative plan to accomplish objective 2 of Res. [12-12]. Additionally, the Sub-Committee recognized that the above work is providing a useful foundation for adopting this region as a basis for a case study in implementing the Ecosystem Based Fisheries Management (EBFM) approach within ICCAT and this collaboration should continue to be supported. The Sub-Committee was requested by the Rapporteur to determine how we might address item 1 of Resolution Working paper SCRS/2013/132 provided a 46 page inventory and ecology of fish species of interest to ICCAT in the Sargasso Sea. With a view to being able to assess the relative significance of this ecosystem to ICCAT species, the Sub-Committee asked that the detail of the report be reflected in a table that relates important life history parameters to their dependence on the Sargasso Sea ecosystem. As an exercise the Sub- Committee determined the relevant parameters and evaluated each species dependencies using a presence/absence scoring system. Table 1 represents a preliminary assessment of the importance of the ecosystem and is also intended to be the basis for a more quantitative evaluation of the data. It was recommended by the Sub-Committee that those future evaluations: 1) Reflect the absence of information. 2) Clearly define the component of each species (population or stock management unit) the assessment applies to. 3) Define a rational scoring system for each of the parameters and map them to a common scale. 4) Characterize each of the life history parameters used. 5) Show catch of each species within the area relative to its catch in the management unit that encompasses it. 7. Discuss ways of including ecosystem values in the standardization and assessment of ICCAT assessed stocks The Rapporteur of the Ecosystem Based Fisheries Management (EBFM) sub-group of the Sub Committee on Ecosystems provided the Sub-Committee with a presentation that outlined a methodology for developing a sustainable development reporting system. The reporting system is part of an ecosystem based fisheries management framework and it explicitly links the conceptual objectives of management to operational objectives. The methodology translates conceptual objectives into operational objectives through a process termed unpacking. Unpacking involves subdividing higher order conceptual objectives into its components and further subdividing the components until an operational objective can be defined. The operational objective is then associated with a measurable indicator and reference point to facilitate reporting and provoking management action. It was shown how these status indicators could be part of an indicator system that is more explicit about the nature of the management action. The Sub-Committee observed that it would be beneficial for the Sub-Committee on Ecosystems to obtain some guidance on operational objectives from the Commission. It was noted that the SCRS strategic plan is a form of unpacking procedure. This more detailed unpacking could be framed within that process including strategies of how to accomplish the various tasks identified to operationalise EBFM. The SCRS needs to ascertain what information is currently available and how it can be used in an EBFM context. The potential use of the Sargasso Sea as a case study is promising as it provides an example on how to structure the work in order to inform the Commission. It was again noted that dialogue with the Commission is fundamental to the process although the process can be initiated in the interim, with the SCRS explaining its proposed methodology and plan to the Commission (including data needs). The Sub-Committee decided to conduct an example in unpacking using a template provided by the Rapporteur of the Ecosystem Based Fisheries Management (EBFM) sub-group. The template was re-organised and modified to suite the specific needs of ICCAT regarding the unpacking procedure. It was stressed that this was an exercise to familiarise the SC-ECO with the process. Much more in depth analysis and modification of the template would be necessary to obtain a finished product suitable for informing the Commission. This initial unpacking exercise is included in Appendix

116 There was some discussion following this exercise as to whether it should be conducted generically for the whole ICCAT area, or using a case study (Sargasso Sea). For this initial exercise it was decided that it should be generic for ICCAT species rather than the Sargasso Sea only, as although this may mask the complicated nature of the study, it could provide a generic framework for considering EBFM in the ICCAT context. On conducting the exercise, the Sub-Committee on Ecosystems made several observations regarding the process. 1) Resolving community level components is fairly complicated. In this example the Sub-Committee decided to move on to the species level components. 2) Populating the species list should be prioritised based on ICCAT species of interest. This will be an ICCAT management issue after all. Although non-iccat species must be considered, they can be brought in at a lower level. 3) It must be considered that other RFMOs are also involved in the management of fish stocks in the region (e.g. NAFO). 4) The difficulty/complexity and time required in conducting this unpacking exercise was noted and it was acknowledged that if a workshop or working group was needed to conduct this work for management purposes, it could require significant time and effort. 5) It was noted that several terms used in the template, such as Maintain population diversity need to be rephrased to better fit the ICCAT situation. The term unpacking was also noted as being new to many participants. Rather than using terms such as maintain or conserve it was suggested that it may be better to try to understand or monitor these factors. This is especially relevant for factors over which management would have little or no control, such as environmental factors. Due to the difficulty in running through the unpacking exercise in plenary, the Sub-Committee was later asked to send comments on the unpacking exercise to the rapporteur of the Sub-Committee who incorporated these comments into a table. The following additional observations were made on the exercise: The unpacking can t be developed in a general way. It needs to be separated on a regional basis. The exercise needs to focus on EBFM. Integrated management is too broad a starting point and the scope needs to be limited so that it does not stray too far from EBFM. Link local issues to national and international vision to have the support of the stakeholders and public. Identified missing aspects related to fisheries not accounted for in the exercise. o Collection of bait o Loss of gear o Garbage The word usage does not capture the dynamic nature of the ecosystem. It applies static terms to describe an objective when the system is dynamic. The list of indicators must be concise, justified and include reference points to guide management. The framework must be easy for managers to deal with. Missing Components; o Biophysical o Socioeconomic Framework must be adaptive. The framework must deal with the uncertainty in the management of the objectives. It must recognize that we control the human impacts not the ecosystem. Involvement of appropriate experts where expertise is lacking is crucial in developing the framework. To this end, it was suggested that a small task force could put together ideas to present to commission. E.g. what is scope, what are problems, what are solutions? The Sargasso Sea Alliance expressed their willingness and readiness to contribute to the process of conducting a case study building on the wealth of information they are already collecting on ecosystems. They clearly stated that there is no expectation for the SCRS to conduct all the work required, but rather proposed a collaboration to conduct the study building on the information presented in SCRS/2013/132. The discussion was then focused on a case study using the Sargasso Sea. The components of this study are presented in Table

117 Part II: By-catch 8. Review of the inputs used for the preliminary Ecological Risk Assessment (ERA) Information from the observer programs of Brazil and Uruguay ( ), as well as other efforts related to different aspects considered in the analysis of productivity and susceptibility (PSA) of sea turtles is summarized in document SCRS/2013/130. Five species of sea turtles are found in the western South Atlantic, loggerhead (Caretta caretta), leatherback (Dermochelys coriacea), hawksbill (Eretmochelys imbricata), olive ridley (Lepidochelys olivacea) and green turtle (Chelonia myda). All of them nest in Brazil, in either continental or insular areas. A growing number of nests and nesting females has been observed for loggerhead, leatherback and olive ridley turtles in recent years. Authors explored the spatial information regarding incidental catches of different species of sea turtles, relative to the areas covered by the PSA. The effort of both fleets ranged from 0.02% to a 6.75% of the total ICCAT effort for the same area. These areas (Regional Management Units or RMUs), in some cases are smaller than those defined with catches. On the other hand some RMUs, did not consider mixed stocks which have been identified by some authors. Information from satellite telemetry for species that interact with longlines, and existing regional information about gear selectivity for loggerhead and leatherback turtles is also summarized. This could be useful to inform the susceptibility component of the ERA. The authors of the document suggested the possibility of using CPUE, or total catch of sea turtles for replacing the RMU, as these had a greater distribution. Moreover they expressed concern that the RMU did not, in some cases, consider the diversity on nesting beaches, which comprise of multi-species nesting areas. It was also discussed that the possibility of separating the longline fleet into shallow and deep components could be done, and the use of satellite telemetry data could be included, both of which would introduce into the ERA the "encounterability" of the species and gear. Another aspect discussed was the possibility of using existing selectivity data as this would improve aspects of susceptibility within the PSA. 9. Ecological Risk Assessment (ERA) 9.1 Presentation of preliminary ERA Document SCRS/2013/134 (Appendix ERA) presented a review of the ERA conducted for sea turtles impacted by ICCAT fisheries. Sea turtles are highly philopatric to natal beaches. This has led to the development of genetically distinctive populations within most species, defined by broad geographic zones. These are generally referred to as Regional Management Units (RMUs). It is appropriate to manage threatening processes at the level of the RMU. However, many turtle species migrate or disperse widely at sea. There are thus large areas where RMUs from a single species overlap. Assigning a turtle encountered at sea to an RMU is currently virtually impossible, because they are indistinguishable. We have conducted this assessment at the level of the RMU, including attempting to assign bycatch information to RMUs whenever possible. There are 22 RMUs from six turtle species in the ICCAT region; nomenclature used here for RMUs is as follows: the species binomial initials identify each species (e.g. Caretta caretta = Cc, Chelonia mydas = Cm, etc.). The region of provenance follows (e.g. Atlantic = At, Indian = In), and geographic region identified the broad areas where the RMU occurs = NW, SE. Thus Cc-AtNW is the loggerhead population from the Atlantic Northwest. The objective of this ERA is to assess the risk to turtles from the impacts of tuna fishing in ICCAT region. It is a Level 2 (semi-quantitative) assessment, conducted within a Productivity-Susceptibility Assessment (or PSA) framework, at the level of the RMU. A PSA uses turtle life-history traits (to provide an indication of productivity (and therefore resilience to fishing impacts), and fishing parameters to account for susceptibility of different groups to capture by different gear types. In this method, all parameters are scored low, medium or high. Parameters can be up-or down-weighted; values are then summed and the resultant score rescaled to 1-3 for both productivity (P) and susceptibility (S). The vulnerability of each RMU to each gear type is then calculated according to the formula: The methodology calls for data deficiencies to be scored according to the precautionary principle. There were many productivity parameters for which no data were available, and these were scored as 1 (low productivity). 2133

118 For the susceptibility assessment, nine parameters were originally identified. However only two of these could be adequately addressed RMU overlap with the ICCAT region, and an estimate of annual turtle bycatch mortality. Mortality estimates were available for longlining and purse seining, but no data were available for gillnetting, and this is not considered further. 9.2 Discussion of preliminary ERA To clarify the process of the ERA it was explained that the information that comes out of this exercise is the property of the Sub-Committee and will not be published at this stage. It was acknowledged that this work is ongoing and it is anticipated that many collaborators will be included in future formulations of the ERA (and other models). Only once this collaboration has been finalized and all authors/contributors fully included and acknowledged, could this work be published by consent of the Secretariat and the SCRS (such was done with the sharks ERA). The Sub-Committee acknowledged the importance of this analysis as a preliminary step towards addressing the Commission request to assess the impact of ICCAT fisheries on sea turtle populations. This is the first time an ERA for sea turtles has been conducted within the SC-ECO. It was acknowledged that it is a difficult exercise because of the very different life history of sea turtles when compared to sharks, sea birds and teleost fish for which ERAs have been conducted by the SCRS. Turtles are impacted by many factors including land based activities and not just ICCAT fisheries. The initial work of the SC-ECO is to provide information and guide revisions of the ERA in the short term, and later to take ownership of the work and improve the model through expert collaboration and input. Discussion of the ERA presentation occurred and can be loosely categorized under the following headings: Quality of data The calls for data on sea turtles in 2012 and 2013 had a relatively limited response, with fewer than 20 countries submitting information on sea turtle interactions. All data was provided to the sea turtle ERA contractors, with confidentiality maintained. The Sub-Committee agreed that in order to fully address the Commission s request, CPCs need to submit data on sea turtle interactions where available as the ERA had to draw inferences from other oceans as well as make extrapolations based on the few data that were provided, which may provide a distorted picture of what is actually occurring and may bias the provision of advice on the model. The Sub- Committee expressed concern that areas being highlighted as being at risk are in fact areas for which data is available, whereas data poor regions may not be receiving the attention they require. Although the ERA did try and account for this problem, it could be resolved by expert collaboration which could fill many data gaps. During the meeting several scientists presented updated and more complete data (catch rates, observer programme summaries, and satellite tracking data) that highlight the importance of incorporating this additional data in the future analysis (ref. presentations which provided updated information). The estimation of total bycatch numbers needs to be revisited using the more complete by-catch information as well as revised estimation procedures. Use of Regional management units (RMU) There was concern that the use of RMUs may not be the most appropriate way to structure the model and that the risk should rather be evaluated for key regions in which the fleets interact with sea turtle populations. Other methods of structuring the model could be investigated. There was especially concern over RMUs which encompass large regions but have relatively few nesting sites. The RMU approach does not always account for disparities in the sizes of the breeding populations within each region. Seasonality The Sub-Committee on Ecosystems noted the importance of the seasonal characteristics of longline fishing effort and that this should be accounted for in the analysis. Tracking data, disaggregated by season could address this issue, in connection with available observer data on interactions. A potential problem is that the CPUE studies aren t always related to seasonal data. Lack of data makes this work difficult. Most of the data used in the study comes from the document created in 2012 by Coelho (2012). Associated data on turtle bycatch is sometimes missing. 2134

119 Longline fishing characteristics The Sub-Committee noted that shallow longline sets were more likely to interact with turtles than deep sets. This was due to the overlap in the sea turtle vertical distribution and the depth of the fishing gear. It was suggested to separate the longline deep sets from the longline shallow sets to account for the difference in sea turtle mortality between these two fleet sectors. This information is available from the sharks ERA conducted in Turtle tracking data The ERA contractors noted the problem that although they had maps of sea turtle tracks, they did not have access to the actual data regarding these movements and thus were not able to include much of this information in the current ERA. Also, the available tracking data is mostly for post nesting females from nesting beaches, not juveniles and males. The Sub-Committee debated how many tags are needed to have confidence in the turtle movement and discussed whether confidence in the tracks could be weighted. Turtle productivity and mortality (natural, fishing, post release) The suitability of the proxies for natural mortality and fishing mortality were discussed. The issue was raised that the use of female breeding numbers in the proxy for Z (total mortality) may not be appropriate and that other sources of mortality should be investigated including from available literature and direct observations of M (natural mortality) based on nesting site fidelity and natal beach tagging of females conducted over extended time periods. Regarding F (fishing mortality) estimates, post release survival is difficult to determine for sea turtles. For all estimates of mortality, further investigation of the available literature is required to obtain additional estimates of these parameters. For example, a longline boat crew may remove a hook and release a turtle and the observer could record this as a live release when in fact the turtle does not survive. The value of post release mortality for different sea turtles in different fleets needs to be further investigated. Including coastal fisheries The Sub Committee on Ecosystems acknowledged the importance of including the impact of coastal fisheries in the ICCAT Convention area on sea turtles. The Committee recognised the necessity of incorporating different gears, and selectivity (size, species, etc.) for these fisheries. In particular the need to include gillnets in the analysis was noted. Currently there is not enough information on bycatch from fisheries in the ICCAT Convention area to address this sector, although further expert collaboration may provide information useful for this assessment. 9.3 List of changes/improvements to be made to final ERA After the initial discussion of the ecological risk assessment it was largely acknowledged that an ecological risk assessment could provide a tool for the SCRS to evaluate the relative risk of ICCAT fishing gears on sea turtle populations. All files used to conduct the ERA were provided to the Sub-Committee on Ecosystems in order to provide a transparent tool the Sub-Committee could use to explore the input data, model structure and parameterization of the ERA, and make recommendations for its improvement. The Committee recognized that this approach was a valuable first step in addressing the Commission s request, although further work was necessary to modify and improve the assessment. It was acknowledged that this ERA is the start of a process which will require collaboration and refinement in order to provide management advice for sea turtle populations affected by ICCAT fisheries. To that end, the Sub-Committee discussed and outlined recommendations to be evaluated, and revisions to be undertaken in the short (before 9/2013), medium ( ) and long-term (2015). Short term recommendations are intended to be incorporated in an updated ERA produced by the ERA contractors prior to the SCRS meeting in September, whilst medium and long term suggestions will be incorporated in future iterations of the ERA by a collaborative effort within the Sub-Committee. The ERA, as currently structured, examines the risk by gear (LL-PS.) X Species X RMU. 2135

120 In order to improve the ERA the Sub-Committee made the following recommendations: 1) Short-term (2013, prior to the SCRS Plenary) a) Participating CPC experts to pass comments to the authors for consideration and summarization. The authors would prepare a short response in September. b) The longline fleet should be further classified into LL-deep and LL-shallow The EFFDIS estimates developed for the Shark ERA will be used for this purpose. c) Stratify the fleet and by-catch information by season. Explore the loss of data and the effect on the analysis. This will likely demonstrate the need for additional seasonal information. d) Conduct a GAP analysis to identify fleets/places where the data is lacking and where additional research efforts are required. It is important to identify CPCs that reported zero turtle bycatch or who did not make any report at all and do not have national observer programs, as they may in fact encounter sea turtles but have no data collection mechanisms to report them. e) The proxies for mortality (natural and fishery induced) need to be revisited. The risk levels should also be reclassified (i.e. what is low, medium and high risk). Alternative other proxies of mortality should be explored, including estimates from available literature. It is recommended to use standardized indices of catch per effort from major ICCAT longline fleets where available. 2) Mid-term ( ) a) Identify important data inputs (e.g. bycatch per unit effort BPUE, bycatch data by species including release mortality estimates) needed to improve the ERA analysis and assemble for use if necessary, data could be aggregated at the spatial scale of the RMU or other spatial unit determined by expert consultation. b) Obtain detailed satellite data. Determine vertical overlap of longline fisheries and turtles. c) Determine night vs. daytime fishing effort. This information is already integrated into the Shark ERA and will be provided by the authors of that study. d) Use available information on the size of turtles caught by the different fleets in the susceptibility component of the ERA analysis. e) Include information about large areas with low rates of capture by sea turtle species. f) Include information about confidence in BPUE estimates (e.g. CV, level of observer coverage) and determine how representative the observer coverage is of the fishing operations. g) Assemble additional satellite tracking information where possible (such as from seaturtle.org). h) Consider the inclusion of number of nesting beaches and the number/density of nests on those beaches as a component in the productivity portion of the ERA. i) Include an analysis of where turtles spend the most time (e.g. Kernel Density) using sat tracking information where possible. 3) Long Term > 2015 a) Obtain information about coastal observer programs, and explore other ways to estimate artisanal fishing effort and bycatch (e.g., contact other groups that conduct similar work NGOs/research institutes etc.). This is especially relevant for gillnet fisheries which may have a significant impact on sea turtle populations. b) Risk (as considered within the ERA) is defined the product of the magnitude of the adverse consequences of the hazard and the likelihood of the effect. The identification of hazard, the likelihood and the consequences are all terms characterised by, described with and measured with various types and degrees of uncertainty. Although it is impossible to identify all uncertainties. The scientific credibility, accuracy and integrity of a risk assessment hinges on the quality of its uncertainty analysis (Burgman, 2005). Therefore appropriate methods to identify, describe and subsequently address the various forms uncertainty should be developed in the future. 10. Review sea turtle available bycatch mitigation and safe-release protocols measures, and make recommendations as necessary This topic was evaluated at the 2013 meeting of the Sub-Committee on Ecosystems, and previously at the 2011 and 2012 meetings. The relevant documents and reports discussed at previous meetings are summarized in Table 3. The 2011 SC-ECO meeting also reviewed a summary (Anon. 2012) of the available literature pertaining to bycatch mitigation for sea turtles. 2136

121 New documents were also considered on by-catch mitigation and safe-release protocols. SCRS/2013/129 presented a circle hook experiment conducted on a Chinese Taipei longline fishing vessel in the tropical Atlantic Ocean. The experiment was conducted from September 2012 to May 2013 between 2ºS and 12 ºS and 17ºW and 26ºW. Traditional tuna and circle hooks were sequentially alternated throughout the experimental portion of the set with a 1:1 ratio. Fisheries observers monitored 200 sets, 36 turtles were hooked and an additional 19 were entangled in the mainline, branch line or buoy line. Randomization tests showed there were no significant differences for sea turtle catch rates by hook type (p=1.000), but there was a significantly higher catch rate of bigeye tuna (p=0.0002), yellowfin tuna (p=0.0045), swordfish (p=0.0001), and blue sharks (p=0.0209) on circle hooks as compared to traditional tuna hooks. Significantly higher catch rates were observed for albacore (p=0.0010) and spearfish (p=0.0097) caught on tuna hooks as compared to circle hooks. The Sub-Committee noted that release mortality during this study was high relative to the bycatch rates of the Chinese Taipei fleet as a whole, but the author stated that this study occurred in an area and time where bycatch rates are higher than average. The study indicated that catch rates of circle-like hooks and J-like hooks were equal, but that survival was higher on circle hooks. Document SCRS/2013/131 analyzed the data obtained by the Spanish Institute of Oceanography (IEO) sampling program and logbooks from the Spanish longline fleet fishing swordfish in the Mediterranean Sea for 2011 and This fleet has three components: surface longline, semi-pelagic longline and deep longline with different by-catch rate and consequently with different impact in the sea turtle populations. Swordfish is the target species with more than the 80% of the total catch (range 82%-93%). In 2004 a pilot study was conducted using 10 different types of hooks (J and circle), four baits and different longline depth. The results of this study showed that the impact of longlines on sea turtle is mainly dependent on the depth of the gear, although area and seasonal effects are also significant. Results also showed that the impact of this fishery on sea turtles can be considered low over the year. The Sub-Committee noted that the catches of turtles are low for sets at all depths and inquired about the type of hook being used. The author indicated that #1 J hooks are being used in the fishery. It was also stated that experiments had been conducted with circle hooks but were not continued due to the decrease in catches of swordfish and the difficulty of the fishermen in releasing the turtles from this gear. The Sub-Committee also considered work described previously (Sales et al., 2010) regarding the effectiveness of circle hooks for the mitigation of sea turtle bycatch and capture of target species in Brazilian pelagic longline fishery. Between 2004 and 2008 the performance of 18/0 10º offset non ring circle hooks was compared with 9/0 J non ring hooks (control) in the Brazilian pelagic longline fishery targeting swordfish, tuna and sharks. During this experiment 26 trips, 229 sets and 145,828 hooks were set alternating circle and J hooks and using mackerel (Scomber japonicus) as bait. A total of 60 different species, including turtles and seabirds, were caught. Statistical analyses (Mantel-Haenszel χ² tests) of species with at least 20 individuals caught were performed. Circle hooks resulted in a significant catch decrease for loggerhead Caretta caretta (55%) and leatherback Dermochelys coriacea (65%) sea turtles. Use of circle hooks resulted in increased capture rates of tunas (bigeye Thunnus obesus and albacore Thunnus. alalunga), and sharks (blue Prionace glauca and requiem sharks of the genus Carcharinus). There was no difference in the capture of yellowfin tuna (Thunnus albacares), shortfin mako shark (Isurus oxyrinchus), hammerhead sharks (Sphyrna lewini and S. zygaena). On the other hand, the capture rate of swordfish (Xiphias gladius) decreased significantly when using circle hooks. Additionally, use of circle hooks significantly decreased capture rates of bycatch species, such as pelagic stingrays (Pteroplatytrygon violacea) and white marlin (Tetrapturus albidus). Circle hooks performed similar to J hooks with respect to many species, and increased captures of marketable species such Atlantic pomfret (Brama brama), escolar (Lepidocybium flavobrunneum), and mackeral shark (Lamna nasus). Results demonstrate the effectiveness of circle hooks for the conservation of loggerhead and leatherback sea turtles, improving the capture rates of most target species, and significantly reducing the bycatch of the most common species, the pelagic stingray, thus economically improving this fishery. The Sub-Committee inquired about the cause of differences in catch rates on circle-hooks versus J-hooks observed in this study, but not in SCRS/2013/129. The author noted that there was a difference in the gear configuration used during the studies, the rings on hooks. These are used in the Chinese Taipei fishery, but not in Brazil. The Sub-Committee also noted that although the studies did not agree that circle hooks reduced the catch rates of sea turtles, they did agree that the use of circle hooks reduced bycatch mortality, likely because sea turtles more frequently swallow J-hooks. A second difference in gear configuration was also noted; the distance between floats and branchlines was shorter in the Brazilian longline fishery than in the Chinese Taipei fishery. It was not clear how these influenced catch rates, but the author noted that shorter branchlines imply a shallower set, which could improve survival of sea turtles. 2137

122 Document SCRS/2013/128 presented the preliminary incidental catch rates of sea turtle of Chinese Taipei longline fleets in the Atlantic Ocean. The data collected from 16,352 observed sets and 46.0 million hooks. Seven hundred and sixty-one turtles were caught. The major incidental catch species was leatherback (59.9%), olive Ridley turtle (26.9%) and loggerhead turtles (8.1%). Most sea turtles were hooked (58.9%). Most leatherback and olive Ridley turtles were caught in tropical areas, especially in the Gulf of Guinea. Loggerhead was caught not only in tropical areas but also in southwest Atlantic Ocean. The nominal incidental catch rate was per 1,000 hooks in average and ranged from per 1000 hooks by area. The Sub-Committee noted that the data used in the analyses were from the whole Atlantic Ocean and suggested that analyses should be further refined using the area and seasonal components. It was also noted that the proportion of observed sets that reported bycatch was very low, and it was suggested that a negative binomial or Poisson distribution should be used for the data instead of a normal distribution. The author also stated that the gear used in tropical areas for these vessels is different than that used in the northern and southern areas as the target species is bigeye tuna. The Sub-committee inquired about the location of hooking on the turtles. While the author indicated that this was not always recorded, it was also stated that leatherbacks are often hooked on the flipper rather than the mouth because they do not eat bait. 11. Response to the Commission regarding Rec (Recommendation by ICCAT on the By-Catch of Sea Turtles in ICCAT Fisheries) In 2010, the Commission recommended that: 1) SCRS initiate an assessment of the impact of the incidental catch of sea turtles resulting from ICCAT fisheries as soon as possible and no later than 2013 [Rec ; Paragraph 5]. 2) After the initial assessment is complete and the results presented to the Commission, SCRS shall advise the Commission on the timing of future assessments [Rec ; Paragraph 5] 3) The SCRS shall also provide advice to the Commission on approaches for mitigating sea turtle by-catch in ICCAT fisheries, including reducing the number of interactions and/or the mortality associated with those interactions [Rec ; Paragraph 4]. 4) As appropriate, the Commission and its CPCs should, individually and collectively, engage in capacity building efforts and other cooperative activities to support the effective implementation of this recommendation, including entering into cooperative arrangements with other appropriate international bodies. With regard to the mitigation of sea turtle bycatch in ICCAT fisheries, the SCRS recommends the following: 1) The SCRS reiterates the previous Commission recommendations [10-09] that: a) Purse seine vessels operating in the ICCAT Convention area avoid encircling sea turtles to the extent practicable, release encircled or entangled sea turtles, including on FADs, when feasible, and report interactions between purse seines and/or FADs and sea turtles. b) Pelagic longline vessels operating in the ICCAT Convention area carry on board safe-handling, disentanglement and release equipment capable of releasing sea turtles in a manner that maximizes the probability of their survival. c) Fishermen on pelagic longline vessels flagged to that CPC operating under their flag use the equipment specified in item b (above) to maximize the probability of sea turtle survival and are trained in safe-handling and release techniques. d) CPCs include in their Annual Reports other relevant actions taken to implement FAO s Guidelines to Reduce Sea Turtle Mortality in Fishing Operations with respect to ICCAT fisheries. 2) Furthermore, to reduce by-catch mortality of sea turtles, the SCRS specifically recommends that: a) Regarding safe-handling practices: i) When a turtle is to be removed from the water, an appropriate basket lift or dip-net be used to bring aboard sea turtles that are hooked or entangled in gear. No turtle should be hauled from the water by a fishing line attached to, or entangled upon the body of a turtle. 2138

123 ii) Vessel operators or crew assess the condition of sea turtles that are caught or entangled prior to release. Those turtles that are not able to swim, unconscious or unresponsive should be brought/maintained onboard and assisted in a manner consistent with maximizing their survival prior to release. These practices are described further in the FAO s Guidelines to Reduce Sea Turtle Mortality in Fishing Operations. iii) That turtles handled in fishing operations or by national observer programs (e.g. tagging activities) be handled in a manner consistent with the FAO s Guidelines to Reduce Sea Turtle Mortality in Fishing Operations. b) Regarding the use of line cutters: i) Longline vessels carry on board line-cutters and use these when safe de-hooking is not possible to release sea turtles. ii) Other types of vessels that use gear that may entangle sea turtles should carry on board line-cutters and use these tools to safely remove gear, and release sea turtles. c) Regarding the use of de-hooking devices: i) Longline vessels carry on board de-hooking devices to effectively and safely remove hooks from sea turtles. The Sub-Committee also recommends that when a hook is swallowed, no attempt be made to remove the hook. Instead, the line should be cut as close to the hook as possible. With regard to the assessment of impact of ICCAT fisheries on sea turtles, the SCRS initiated an Ecological Risk Assessment (ERA) for sea turtles in Progress to date includes: 1) In 2013 ICCAT provided a short-term contract that supported the development of a preliminary ERA for sea turtles species encountered by ICCAT fisheries. The ERA used data provided to the Secretariat by CPCs in 2011 and 2012, and as collated under a short-term contract funded by ICCAT in 2012 and other data sources compiled by the contractor. 2) At its Inter-sessional meeting in 2013, the Sub-Committee on Ecosystems reviewed the ERA progress to date and made important recommendations to improve the assessment over the short (before 10/2013), medium ( ) and long-term (2015+), including a request for updated/additional data from the CPCs 3) The SCRS will continue to improve the ERA and will advise the Commission on its plan for future sea turtle impact analyses at the 2014 meeting. 12. Other matters 12.1 Presentations A presentation was given on bycatch of turtles by the Moroccan fleet. Interviews conducted with Moroccan tuna vessel fishermen from 2008 to 2011 revealed that one turtle was encountered every 90 to 100 fishing days. Fishing days per trip range from 1 to 3 days but seeing the technical characteristics of the boats, they are generally less than 24 hr. Out of a sample of 100 boats specializing in tuna fishing south of Agadir within the Moroccan EEZ, it was determined that five in every six boats that fishes observes a turtle. The catch rates of turtles in this fishery are one turtle for every 20 fishing days. In this zone, turtles captured are immediately released. The results from the national observer program on shark captures onboard longline vessels have indicated that turtles are not present or their interactions are insignificant in this fishery. Morocco proposes to lead an awareness campaign with the different operators in the region for the protection of marine turtles. Document SCRS/2013/133 presented results on the Trans-Atlantic Leatherback INitiative (TALCIN). The second phase of this work provided the first ocean-scale analysis of spatio-temporal distribution of the leatherback turtle, based on electronic tagging, as well as ascertaining overlap with longline fishing effort. Data suggest that the Atlantic likely consists of two regional management units: northern and southern (the latter including turtles from South Africa). Although turtles and fisheries show highly diverse distributions, the authors highlight nine areas of high susceptibility to bycatch (four in the North Atlantic and five in the South/Equatorial Atlantic) that are worthy of further targeted investigation. These are reinforced by reports of leatherback bycatch at eight of the sites (Document SCRS/2013/133). 2139

124 The Sub-Committee noted that an analysis was conducted on the overlap of high fishing-pressure areas with leatherback habitat by quarter, and it was suggested that the Ecological Risk Assessment for sea turtles could use a similar approach. The Sub-Committee also noted that leatherback habitat use could overlap with high and low fisheries production areas. In this regard, it was suggested that this information should be compared to observer data to determine whether there is any correlation. The Sub-Committee recommended that further comparisons could be conducted of leatherback high-use areas with shallow water longline sets and deep water longline sets, using analyses as developed in the shark Ecological Risk Assessment. Document SCRS/2013/135 (Sightings and abundance of marine turtles in Azores) presented an analysis of the marine turtle sightings and observation effort by the observers onboard the Azores tuna baitboat fleet. A total of 1,823 trips were observed from 2001 to 2012, and 25,903 surveys with a mean duration of 15 minutes were performed. Occasional turtle sightings, from 1998 to 2000, are also given. For the loggerhead sightings and observation effort in the Azores area for 2001 to 2012 data, a relative index of abundance was estimated by a General Linear Model approach (GLM). The analysis shows that after the low sightings per unit of effort in 2001, the abundance index reached its highest level in 2002 and decreased in the following years and remained stable after 2008 to present. Information on leatherback and green turtle observations in the Azores area were also given. The information collected in Madeira area, for some years and months, included the number of surveys and loggerheads sighted. Nominal indices (all species and areas) and Standardized CPUE (Sightings per unit effort) for loggerheads in the Azores are also included. This is the only fishery independent index on the high seas. The Sub-Committee noted that the ability to sight turtles may be affected by factors such as sea state, cloud cover and the angle from which observers are searching, and that this should be incorporated into the analysis. It was also noted that there has been a documented increase in the numbers of turtles on the nesting beaches in the USA, and as it is known that these turtles can migrate to the Azores, the Committee thought that it would be interesting to see if sightings in the Azores increase in future years. The presenter shared further information on the results of genetic analysis conducted, which confirms that 100% of the turtles around the Azores come from the USA and Mexico. It was noted that this information did not correspond to the RMU CcATLNE area according to Wallace et al. (2010). The Sub-Committee recommended that an additional useful analysis would be to determine turtle growth rate using tag recovery data. SCRS/2013/138 describes an approach to collaborative research in fisheries science capacity building. The author listed the benefits of collaborative fisheries research: (1) enhanced credibility and legitimacy of the scientific findings in the minds of stakeholders, with potential to increase acceptability of management actions; (2) innovative and adopted fishing gear and practices that reduce bycatch and minimize habitat impact from fishing; (3) greater mutual understanding and trust among partners; and (4) opportunities to integrate diverse sources of knowledge about the coastal and marine environment. The author stated that the collaborative research approach has been increasingly adopted by government, industry and associations and conservation organizations, such as The Nature Conservancy and in WWF s Smart Gear program. The author elaborated on a new collaborative fisheries research fellowship program initiated at the Virginia Institute of Marine Science. The author noted that the fellow could benefit from collaboration with an expert in tuna-fisheries and bycatch and invited the Sub-committee to consider collaborating through this program. The Sub-Committee thought that this proposal was an interesting approach, and a good opportunity to build external support and collaborative research capacities with scientists, industry, managers and NGOs. The Sub-Committee discussed that the VIMS fellow should contact the head of delegation of the appropriate CPC to participate in future meetings of the SCRS Progress toward harmonization of data collection forms for longline observer programs During its 2012 meeting, the SC-ECO recommended that ICCAT coordinate with the trfmo By-catch Joint Technical Working Group (BJTWG) in order to assume a leadership role in developing minimum standards for harmonised longline observer data collection for the trfmos. The ICCAT Secretariat has since contacted Simon Nicol, the Chair of that group, and a process has begun to collate the forms for data collection activities of the Longline observer programs of ICCAT, WCPFC, IATTC, IOTC and CCSBT. Since ICCAT does not directly coordinate large-scale observer programs, it is necessary to communicate and cooperate with national observer programs. To that end, the Secretariat has contacted the coordinators of the national longline observer programs that operate in the ICCAT Convention area in order to obtain their data collection forms. These will then be used to identify and recommend minimum data collection standards. 2140

125 13. Recommendations The Sub-Committee recognized the excellent work conducted by Drs. Andrea Angel, Ross Wanless and Ronel Nel in compiling a preliminary Ecological Risk Assessment (ERA) for sea turtles. Their preparatory work has expedited the work of the Sub-Committee and has provided an excellent foundation for the ongoing impact assessment of sea turtle by-catch in ICCAT fisheries. The Sub-Committee acknowledged the value of this ICCAT initiative to provide financial support to hire experts to contribute to the SCRS's work and strongly recommended continuing with these productive activities. Ecosystems The Sub-Committee recognized the value of the unpacking exercise to define SCRS ecosystem objectives. It is recommended that the Sub-Committee Co-Convener develop a survey to populate a list of conceptual EBFM objectives to be distributed to the SCRS Officers. The Co-Convener will collate the results prior to SCRS Plenary. It is recommended that travel funds be allocated to support the participation of external experts to help develop the scientific tools required to implement EBFM approaches. It is recommended that the working group continue its collaboration with the Sargasso Sea Alliance with regard to the analysis of the ecological importance of the Sargasso Sea for tuna and tuna-like species and ecologically associated species. By-catch The Sub-Committee recommends that the SCRS contribute to the collaborative fisheries research approaches through the involvement of stakeholders in initiatives that address potential ICCAT fisheries interactions with bycatch species. ERA for Sea Turtles CPCs should provide by-catch data according to Task II standards. If that is not possible, the Sub- Committee recommends to CPC s that data concerning sea turtle bycatch be provided by species, with spatial and seasonal information (e.g. 5X5) that would allow assignment to the RMU and quarter. The Sub-Committee recognizes the need to include information on artisanal fisheries that operate within the ICCAT Convention area and encourages CPCs to submit relevant information. The Sub-Committee recommends that supplemental tagging (including electronic and conventional) of sea turtles be conducted and information on those experiments be made available to the Sub-committee. The Sub-Committee recommends that genetic studies on sea turtles be conducted to better understand which populations are being impacted by ICCAT fisheries, and that information on those experiments be made available to the Sub-Committee. Bycatch Mitigation for Sea Turtles The Sub-Committee reiterates the Commission recommendations as adopted in Rec Furthermore, to reduce by-catch mortality of sea turtles, the SCRS specifically recommends that: o Regarding safe-handling practices: When a turtle is to be removed from the water, an appropriate basket lift or dip-net be used to bring aboard sea turtles that are hooked or entangled in gear. No turtle should be hauled from the water by a fishing line attached to, or entangled upon the body of a turtle. Vessel operators or crew assess the condition of sea turtles that are caught or entangled prior to release. Those turtles that are not able to swim, unconscious or unresponsive should be brought/maintained onboard and assisted in a manner consistent with maximizing their survival prior to release. These practices are described further in the FAO s Guidelines to Reduce Sea Turtle Mortality in Fishing Operations. That turtles handled in fishing operations or by national observer programs (e.g. tagging activities) be handled in a manner consistent with the FAO s Guidelines to Reduce Sea Turtle Mortality in Fishing Operations. 2141

126 o Regarding the use of line cutters: Longline vessels carry on board line-cutters and use these when safe de-hooking is not possible to release sea turtles. Other types of vessels that use gear that may entangle sea turtles should carry on board line-cutters and use these tools to safely remove gear, and release sea turtles. o Regarding the use of de-hooking devices: Longline vessels carry on board de-hooking devices to effectively and safely remove hooks from sea turtles. The Sub-Committee also recommends that when a hook is swallowed, no attempt be made to remove the hook. Instead, the line should be cut as close to the hook as possible. 13. Adoption of the report and closure The report was adopted during the meeting as well as the work plan of the Sub-Committee for The Co- Conveners thanked the Secretariat and the participants for their hard work and dedication. The Co-Conveners also thanked Andrea Angel, Ross Wanless and Ronel Nel for assisting the Sub-Committee with important progress towards an Ecological Risk Assessment of sea turtles that interact with ICCAT fisheries, and also the Secretariat for funding this vital work. The meeting was adjourned. Following the Sub-Committee meeting, the Inter-American Convention for the Conservation and Protection of Sea Turtles expressed their readiness to contribute to the improvement of the ERA for sea turtles providing information on the number of nesting beaches and the number/density of nests on those beaches as a component in the productivity portion of the ERA. This information is currently available for all IAC countries from and can be found on the IAC Annual reports on the IAC website Literature cited Burgman MA, Risks and decisions for conservation and environmental management. Cambridge University Press, Cambridge, UK,

127 RAPPORT DE LA RÉUNION INTERSESSION 2013 DU SOUS-COMITÉ DES ÉCOSYSTÈMES (Madrid, Espagne, 1-5 juillet 2013) 1. Ouverture, adoption de l ordre du jour et organisation des sessions La réunion s'est tenue dans les bureaux du Secrétariat de l ICCAT à Madrid (Espagne) du 1er au 5 juillet Au nom du Secrétariat de l'iccat, le Dr Paul de Bruyn a souhaité la bienvenue au Sous-comité. Les coordinateurs du Sous-comité des écosystèmes, la Dr Shannon Cass-Calay (États-Unis) et le Dr Alex Hanke ont ensuite décrit les objectifs et l'organisation de la réunion. L'ordre du jour a été adopté sans modification (Appendice 1). La liste des participants se trouve à l Appendice 2. La liste des documents présentés à la réunion est jointe à l Appendice 3. Les participants suivants ont assumé les fonctions de rapporteur : Point Points 1, 10-11, 14 Points 2-5 Point 6 Point 7 Point 8 Point 9 Point 12 Point 13 Rapporteur(s) S. Cass-Calay A. Hanke A. Hanke, P. de Bruyn P. de Bruyn A. Domingo, B. Giffoni T. Nalovic, P. de Bruyn T. Trott J. Pereira Ière partie : Écosystème L ordre du jour, tel qu adopté par le Sous-comité, incluait trois sections considérées importantes pour le processus de mise en œuvre de l approche EBFM pour laquelle aucune nouvelle information n a été fournie. Il s'agissait des sections 2, 3 et 4 dénommées respectivement : a) Examen des nouvelles informations concernant la mise en œuvre des effets écosystémiques dans les évaluations des stocks améliorées et b) Examen des nouvelles informations sur les modèles qui quantifient les dynamiques des écosystèmes et prédisent l'impact des perturbations sur ses composants. On ne sait pas au juste pourquoi il y avait une absence de participation à cette section, mais le Sous-comité a estimé que ceci pourrait être dû à la capacité limitée des CPC à participer à ce Sous-comité. D'autre part, l expertise pourrait faire défaut au sein du SCRS pour participer pleinement à cette branche de la science. Le Sous-comité recommande que ceci soit intégré au plan stratégique du SCRS à l avenir de façon à ce que nous puissions peut-être accroître la capacité de faire face aux nécessités de développer des outils qui permettront la mise en œuvre de l approche EBFM. 5. Examiner les nouvelles informations sur la mise en œuvre des principes de gestion fondée sur l'écosystème Le document SCRS/2013/137 présentait l'évaluation des risques écologiques (ERA) comme une méthode de suivi de la mise en œuvre d'une approche écosystémique à la gestion des pêcheries (EAF) de l'afrique australe (l'expérience namibienne). La méthode ERA, qui est de nature qualitative, reçoit le soutien des pays de l'afrique australe (Namibie, Afrique du Sud et Angola) par l'intermédiaire de la Commission du courant de Benguela (BCC), avec l'aide du WWF- Afrique du Sud et du projet EAF-Nansen de la FAO. La méthode a été choisie car elle est simple et a des directives structurées permettant d'assurer le suivi des mises en œuvre de l'eaf. L'ERA a contribué à cerner les enjeux et à les classer par ordre de priorité. La méthode a établi des objectifs et des enjeux clairs pour certaines pêcheries de la région du courant de Benguela. 2143

128 Les étapes requises pour résoudre les problèmes et traiter les risques sont recommandées et des examens réguliers sont nécessaires pour évaluer si des progrès ont été accomplis pour résoudre les problèmes. Le fait que cette méthode puisse donner lieu à un consensus sur différentes questions entre diverses parties prenantes est l'un des avantages de l'approche ERA. Six pêcheries namibiennes ont été évaluées au moyen de cette méthodologie et quatre examens ERA ont été réalisés. Le Sous-comité était intéressé par la nouvelle approche et a reconnu que l'auteur pourrait fournir une orientation utile en ce qui concerne l exercice de dégroupage entamé pendant la réunion et décrite au point 7. Il est recommandé de poursuivre la collaboration avec les représentants du ministère des Pêches et des Ressources marines de la Namibie. 6. Examiner les progrès réalisés concernant l'élaboration d'un cas d'essai pour mettre en œuvre la gestion des pêcheries fondée sur l'écosystème Le document SCRS/2013/132 fournit des informations sur la biologie et l'écologie de 18 espèces de poissons dont les aires de distribution incluent la mer des Sargasses. Ces espèces sont divisées en quatre groupes qui correspondent à la classification des espèces relevant de l'iccat à savoir le groupe 1 comprend les principales espèces de thonidés, comprenant l'albacore, le germon, le thon obèse, le thon rouge et le listao, le groupe 2 comprend l'espadon et les istiophoridés, incluant le makaire bleu, le makaire blanc et le voilier, le groupe 3 comprend les thonidés mineurs incluant le thazard-bâtard, le thon à nageoires noires, la thonine commune et la coryphène commune et le groupe 4 comprend les requins, incluant le requin-taupe bleu, le requin peau bleue, le requin-taupe commun, le renard à gros yeux et le requin pèlerin. Pour chaque espèce, des informations et des données ont été fournies en ce qui concerne l'aire de distribution, les débarquements, les schémas de déplacement et de migration, la reproduction, l'âge et la croissance, l'alimentation et les habitudes alimentaires ainsi que l'écologie par rapport aux paramètres océanographiques, principalement la température de l'eau. L'importance des Sargasses comme habitat de prédilection des poissons a été abordée, ce qui a été expliqué par les habitudes alimentaires des thonidés et d'autres prédateurs pélagiques. Les poissons volants sont une proie importante du régime alimentaire des thonidés et des istiophoridés, et sachant qu'ils dépendent en grande mesure des amas des Sargasses qui constituent leurs habitats de frai, la mer des Sargasses joue un rôle fondamental de la chaîne trophique des poissons grands migrateurs pélagiques de l'atlantique du Nord-Ouest. Une évaluation des informations existantes donne à penser que l'importance de la mer des Sargasses pour diverses espèces pélagiques s'explique principalement par un ou plusieurs des éléments suivants : trajet migratoire, zone de frai, zone de nourricerie, zone d'alimentation, zone d'hivernage ou zone de mise bas. Le Sous-comité a observé que la Résolution de l'iccat sur la mer des Sargasses stipule ce qui suit : 1. Le SCRS examinera les données et informations disponibles sur la mer des Sargasses et son importance écologique pour les thonidés et espèces apparentées et pour les espèces apparentées écologiquement 2. Le SCRS communiquera en 2014 une actualisation du travail réalisé dans ce domaine et adressera en 2015 ses conclusions à la Commission. En ce qui concerne le point 1, un examen préliminaire des données et des informations disponibles sur l'habitat de la mer des Sargasses, y compris des captures des thonidés et d'autres espèces relevant de l'iccat, a été réalisé et des études sur l'importance que revêt cette région peuvent être entreprises. Le Sous-comité a noté qu'une recherche plus élargie de cet écosystème et de son importance comme habitat essentiel pour les thonidés et les espèces apparentées passerait par une collaboration intégrée du SCRS avec des groupes scientifiques spécialisés dans la mer des Sargasses. Le Sous-comité a indiqué que, pour accomplir l objectif 2, il conviendrait de disposer d un plan de travail, d une recherche collaborative et organiser des réunions afin d évaluer adéquatement l importance de la mer des Sargasses comme habitat essentiel pour les espèces de thonidés de l ICCAT. Le Sous-comité a signalé que cette tâche ne serait pas accomplie avant Le Sous-comité recommande de maintenir les contacts avec les équipes de recherche de la mer des Sargasses et les scientifiques du RU-Bermudes afin de développer un plan de collaboration scientifique pour accomplir l objectif 2 de la Rés De plus, le Sous-comité a reconnu que les travaux susmentionnés constituent un fondement des plus utiles pour adopter cette région comme base pour une étude de cas mettant en œuvre l approche de gestion des pêcheries basée sur l'écosystème (EBFM) au sein de l ICCAT et cette collaboration devrait continuer à être appuyée. 2144

129 Le rapporteur a demandé au Sous-comité de déterminer la réponse pouvant être apportée au point 1 de la Résolution Le document de travail SCRS/2013/132 fournit un inventaire de 46 pages et aborde l'écologie des espèces de poissons d'intérêt pour l'iccat dans la mer des Sargasses. Afin de pouvoir évaluer l importance relative de cet écosystème pour les espèces de l ICCAT, le Sous-comité a demandé que les détails du rapport soient reflétés dans un tableau reliant les paramètres importants du cycle vital et leur dépendance de l écosystème dans la mer des Sargasses. À titre d'exercice, le Sous-comité a déterminé les paramètres pertinents et a évalué les dépendances de chaque espèce au moyen d'un système de notation présence/absence. Le Tableau 1 présente une évaluation préliminaire de l importance de l écosystème et vise également à servir de base à une évaluation plus quantitative des données. Le Sous-comité a recommandé que ces futures évaluations : 1. rendent compte de l'absence d information, 2. définissent clairement la composante de chaque espèce (unité de gestion du stock ou de la population) faisant l'objet de l'évaluation, 3. définissent un système de notation rationnel de tous les paramètres et les échelonnent de la même façon, 4. décrivent tous les paramètres du cycle vital utilisés et 5. présentent la capture de chaque espèce dans la zone relative à sa capture dans l'unité de gestion qui l'englobe. 7. Discuter des moyens d'incorporer les valeurs écosystémiques dans la standardisation et l'évaluation des stocks évalués par l'iccat Le rapporteur du sous-groupe consacré à la gestion des pêcheries reposant sur l écosystème (EBFM) du Souscomité des écosystèmes a fourni au Sous-comité une présentation qui décrit une méthodologie de développement d'un système de déclaration sur le développement durable. Le système de déclaration fait partie d un cadre de gestion des pêcheries basé sur l écosystème et il relie explicitement les objectifs conceptuels de la gestion avec les objectifs opérationnels. La méthodologie traduit les objectifs conceptuels en objectifs opérationnels selon un processus dénommé «dégroupage». Le dégroupage subdivise les objectifs conceptuels de plus haut rang en composantes et subdivise à nouveau ces composantes jusqu'à ce qu'un objectif opérationnel puisse être défini. L'objectif opérationnel est ensuite associé à un indicateur mesurable et un point de référence afin de faciliter la déclaration et à déclencher une action de gestion. On a expliqué comment ces indicateurs d'état pourraient faire partie d'un système d'indicateurs qui est plus explicite quant à la nature de l'action de gestion. Le Sous-comité a fait observer qu'il serait utile que la Commission fournisse une orientation concernant les objectifs opérationnels au Sous-comité des écosystèmes. Il a été fait remarquer que le plan stratégique du SCRS est une forme de procédure de dégroupage. Ce dégroupage plus détaillé pourrait s'inscrire dans le cadre de ce processus, comprenant des stratégies sur la façon d'accomplir les diverses tâches identifiées pour mettre l'ebfm en pratique. Le SCRS doit déterminer quelles sont les informations disponibles actuellement et la façon de les utiliser dans un contexte EBFM. L'utilisation potentielle de la mer des Sargasses comme une étude de cas est prometteuse, car elle sert d'exemple sur la façon de structurer les travaux afin d'en informer la Commission. À nouveau, il a été fait remarquer que le dialogue avec la Commission est crucial pour le processus, même si le processus peut être entamé entre-temps et le SCRS peut expliquer à la Commission la méthodologie et le plan proposés (y compris les besoins en matière de données). Le Sous-comité a décidé de procéder à un exemple de dégroupage à partir d'un modèle fourni par le rapporteur du sous-groupe consacré à la gestion des pêcheries reposant sur l écosystème (EBFM). Le modèle a été réorganisé et modifié compte tenu des besoins spécifiques de l'iccat concernant la procédure de dégroupage. Il a été souligné qu'il s'agissait d'un exercice visant à familiariser le Sous-comité des écosystèmes avec le processus. Il sera nécessaire de procéder à une analyse plus poussée et de modifier le modèle en vue d'obtenir un produit fini adapté pour informer la Commission. L'exercice de dégroupage initial est présenté à l Appendice 4. Après l'exercice, une discussion a été tenue sur la question de savoir si celui-ci devrait être réalisé selon une approche générale pour l'ensemble de la zone de l'iccat ou selon une étude de cas (mer des Sargasses). Pour ce premier exercice, il a été décidé qu'il devrait être générique pour les espèces relevant de l'iccat plutôt que de se limiter à la mer des Sargasses. Cela pourrait toutefois masquer la nature complexe de l'étude, mais cela pourrait fournir un cadre générique pour placer l'ebfm dans le contexte de l'iccat. Lors de la réalisation de l'exercice, le Sous-comité des écosystèmes a formulé plusieurs observations concernant le processus. 2145

130 1) Il est assez difficile de résoudre les composantes à un niveau multi-spécifique. Dans cet exemple, le Sous-comité a décidé de passer aux composantes à un niveau spécifique. 2) Remplir la liste des espèces, sur la base des espèces présentant un intérêt particulier pour l'iccat, devrait être une tâche prioritaire. Après tout, il s'agira d'une question de gestion de l ICCAT. Même si les espèces ne relevant pas de l'iccat doivent être considérées, elles peuvent être placées à un niveau inférieur. 3) Il convient de tenir compte du fait que d'autres ORGP sont également impliquées dans la gestion des stocks de poissons dans la région (par ex. OPANO). 4) On a souligné la difficulté, la complexité et le temps requis pour mener à bien cet exercice de dégroupage et il a été reconnu que s'il s'avérait nécessaire d'organiser un atelier ou de réunir un groupe de travail pour réaliser ce travail à des fins de gestion, il faudrait disposer de temps et déployer des efforts pour ce faire. 5) Il a été noté que plusieurs termes utilisés dans le modèle, tels que «maintenir la diversité de la population», devraient être reformulés pour mieux s'adapter au contexte de l'iccat. De nombreux participants ont fait remarquer que le terme «dégroupage» ne leur était pas familier. Au lieu d'avoir recours à des termes tels que «maintenir» ou «conserver», il a été suggéré qu'il serait peut-être plus opportun d'essayer de «comprendre» ou de «suivre» ces facteurs. Cela concerne surtout les facteurs sur lesquels la gestion n'aurait que peu ou pas de contrôle, tels que les facteurs environnementaux. En raison de la difficulté d'exécuter l'exercice de dégroupage en plénière, il a été demandé au Sous-comité d'envoyer ultérieurement des commentaires sur l'exercice de dégroupage au rapporteur du Sous-comité qui a réuni ces observations dans un tableau. Les observations complémentaires suivantes ont été formulées sur l'exercice : Le dégroupage ne peut pas être développé de manière globale. Il doit être séparé par région. L'exercice doit se concentrer sur l'ebfm. La gestion intégrée est trop large. Le point de départ et le champ d'application doivent être limités de manière à ne pas trop s'éloigner de l'ebfm. Relier les questions locales à une vision nationale et internationale pour avoir l'appui des parties intéressées et du public. Identifier les aspects manquants liés à la pêche non inclus dans l'exercice : o Collecte d'appâts o Perte d'engins o Déchets Le terme «utilisation» ne reflète pas la nature dynamique de l'écosystème. Il utilise des termes statiques pour décrire un objectif alors que le système est dynamique. La liste des indicateurs doit être concise et justifiée, et inclure des points de référence visant à orienter la gestion. Les responsables doivent pouvoir utiliser facilement le cadre. Éléments faisant défaut : o biophysique o socio-économique Le cadre doit être souple. Le cadre doit tenir compte de l'incertitude dans la gestion des objectifs. Il faut admettre que nous contrôlons l'impact de l'homme, et non pas celui de l'écosystème. Le concours d'experts compétents lorsque l'expertise fait défaut est crucial pour développer le cadre. À cette fin, il a été suggéré qu'un petit groupe de travail pourrait rassembler des idées à présenter à la Commission. Par exemple : quel est le champ d'application? quels sont les problèmes? quelles sont les solutions? La Sargasso Sea Alliance a fait part de sa volonté de contribuer à la réalisation d'une étude de cas reposant sur un grand volume d'informations qu'ils ont déjà recueillies sur les écosystèmes. Ils ont clairement indiqué que l'on ne s'attend pas à ce que le SCRS réalise tous les travaux requis, mais ont plutôt proposé de collaborer pour mener à bien cette étude s'appuyant sur les informations présentées dans le document SCRS/2013/

131 La discussion a ensuite porté sur une étude de cas ayant recours à la mer des Sargasses. Le Tableau 2 présente les composantes de cette étude. IIème partie : Prises accessoires 8. Examen des données utilisées dans l évaluation préliminaire des risques écologiques (ERA) Le document SCRS/ 2013/130 présente un résumé des informations provenant des programmes d'observateurs du Brésil et de l'uruguay ( ), ainsi que d'autres efforts liés à différents aspects considérés dans l'analyse de la productivité et de la susceptibilité (PSA) des tortues marines. Cinq espèces de tortues marines sont présentes dans le Sud-Est de l'atlantique : la tortue caouanne (Caretta caretta), la tortue luth (Dermochelys coriacea ), la tortue caret (Eretmochelys imbricata), la tortue olivâtre (Lepidochelys olivacea) et la tortue verte (Chelonia myda). Toutes ces espèces font leur nid au Brésil, dans des zones continentales ou insulaires. Ces dernières années, on a observé un nombre croissant de nids et de femelles nidifiantes de tortues caouannes, de tortues luths et de tortues olivâtres. Les auteurs ont exploré les informations spatiales en ce qui concerne les prises accidentelles de différentes espèces de tortues marines, par rapport aux zones couvertes par la PSA. L'effort des deux flottilles oscillait entre 0,02 % et 6,75 % du total de l'effort de l'iccat pour la même région. Dans certains cas, ces zones (unités régionales de gestion ou RMU, selon les sigles anglais) sont plus petites que celles définies avec les captures. D'autre part, certaines RMU ne tenaient pas compte des stocks mixtes qui ont été identifiés par quelques auteurs. Les données de télémétrie satellitaire concernant les espèces qui interagissent avec les palangres ainsi que les informations régionales existantes sur la sélectivité des engins des tortues caouannes et des tortues luths sont également résumées. Ces informations pourraient servir à documenter la composante de susceptibilité de l'era. Les auteurs du document ont suggéré la possibilité d'utiliser la CPUE, ou le total des prises de tortues marines, pour remplacer la RMU, étant donné que ces données avaient une plus grande distribution. En outre, ils craignent que les RMU ne tiennent pas compte dans certains cas, de la diversité sur les plages de ponte, qui comprennent des zones de ponte plurispécifiques. On a également discuté de la possibilité de séparer la flottille palangrière en composantes profondes et peu profondes et d'utiliser des données de télémétrie satellitaire, ce qui pourrait permettre d'ajouter dans l'era le facteur de «possibilité de rencontre» entre l'espèce et l'engin. On a également discuté de la possibilité d'utiliser les données existantes sur la sélectivité, car cela permettrait d'améliorer certains aspects de la susceptibilité de la PSA 9. Évaluation des risques écologiques (ERA) 9.1 Présentation de l ERA préliminaire Le document SCRS/2013/134 (appendice de l'era) présente un examen de l'era concernant les tortues affectées par les pêcheries de l ICCAT. Les tortues marines présentent un niveau élevé de philopatrie envers leurs plages de naissance. Cela a donné lieu au développement de populations génétiquement différentes pour la plupart des espèces, définies par de larges zones géographiques. Ces zones sont généralement appelées unités régionales de gestion (RMU). Il convient de gérer les processus menaçants au niveau de l'unité RMU. Cependant, de nombreuses espèces de tortues migrent ou se dispersent en mer. Par conséquent, des RMU d'une seule espèce se chevauchent sur de grandes zones. Attribuer une RMU à une tortue localisée en mer est actuellement pratiquement impossible, car elles sont indiscernables. Nous avons réalisé cette évaluation au niveau de l'unité RMU, en essayant également d'attribuer les informations sur les prises accessoires aux RMU dans la mesure du possible. Il existe 22 RMU de six espèces de tortues dans la zone relevant de l'iccat. La nomenclature utilisée dans le cas présent pour les RMU est la suivante: deux initiales correspondant à l'espèce (p. ex. Caretta caretta = Cc, Chelonia mydas = Cm, etc.), suivies de la région de provenance (par ex. Atlantique = AT, océan Indien = In) et de la région géographique identifiant les zones où la RMU est appliquée= NW, SE. «Cc-AtNW» correspond donc à la population de tortues caouannes de l'atlantique Nord-Ouest. L'objectif de cette ERA consiste à évaluer les impacts des pêcheries de thonidés dans la zone relevant de l'iccat sur les tortues marines. Il s'agit d'une évaluation de niveau 2 (semi-quantitative), réalisée dans le cadre d'une évaluation de productivité-susceptibilité (PSA), au niveau de la RMU. Une PSA utilise les caractéristiques du cycle vital de la tortue (afin de fournir une indication de la productivité, et par conséquent de la résistance aux effets de la pêche) et les paramètres de la pêche pour tenir compte de la susceptibilité des différents groupes de capture par différents types d'engins. Dans cette méthode, tous les paramètres sont classés (bas, moyen ou élevé). 2147

132 Les paramètres peuvent être augmentés ou réduits. Les valeurs sont ensuite additionnées et le résultat obtenu est rééchelonné de 1 à 3 pour la productivité (P) et la susceptibilité (S). La vulnérabilité de chaque RMU à chaque type d'engin est ensuite calculée au moyen de la formule suivante : La méthodologie vise à classer les déficiences des données selon le principe de précaution. Aucune donnée n'était disponible concernant de nombreux paramètres sur la productivité et ceux-ci ont été classés dans le niveau 1 (faible productivité). En ce qui concerne l'évaluation de la susceptibilité, neuf paramètres ont été identifiés dans un premier temps. Toutefois, seuls deux de ces aspects ont pu être traités correctement, à savoir le chevauchement de la RMU avec la zone relevant de l'iccat et l estimation de la mortalité annuelle des tortues capturées en tant que prise accessoire. Les estimations de la mortalité étaient disponibles pour les pêcheries des palangriers et des senneurs, mais aucune donnée n'était disponible pour les pêcheries au filet maillant, et ce point n'est pas examiné plus avant. 9.2 Débat sur l ERA préliminaire Afin de clarifier le processus de l'era, il a été expliqué que l'information qui provient de cet exercice est la propriété du Sous-comité et ne sera pas publiée à ce stade. Il a été reconnu que ce travail est en cours et il est prévu que de nombreux collaborateurs seront inclus dans les futures formulations de l'era (et les autres modèles). Ce n'est que lorsque cette collaboration aura été conclue et que tous les auteurs/collaborateurs auront été pleinement inclus et reconnus, que ces informations pourront être publiées avec le consentement du Secrétariat et du SCRS (comme cela a été fait avec l'era sur les requins). Le Sous-comité a reconnu l importance de cette analyse comme démarche préliminaire en vue de répondre à la demande de la Commission d évaluer l impact des pêcheries de l ICCAT sur les populations de tortues marines. Il s'agit de la première fois qu'une ERA sur les tortues marines a été menée au sein du Sous-comité des écosystèmes. Il a été reconnu qu'il s'agit d'un exercice difficile car les tortues marines présentent des cycles de vie très différents par rapport aux requins, aux oiseaux de mer et poissons téléostéens au sujet desquels le SCRS a réalisé des ERA. Les tortues sont affectées par de nombreux facteurs, y compris par des activités terrestres, et pas uniquement par les pêcheries de l'iccat. Le travail initial du Sous-comité des écosystèmes consiste à fournir des informations et orienter les révisions de l'era dans le court terme, et ultérieurement à se charger des travaux et améliorer le modèle grâce à la collaboration et la contribution des experts. On a discuté de la présentation de l'era qui peut être classée dans les grandes lignes sous les rubriques suivantes : Qualité des données Les demandes de données sur les tortues marines, lancées en 2012 et 2013, ont reçu une réponse relativement limitée, moins de 20 pays ayant transmis des informations sur les interactions avec les tortues marines. Toutes les données ont été fournies aux experts externes dans le respect de la confidentialité des données. Le Souscomité a convenu que pour répondre pleinement à la demande de la Commission, les CPC devaient transmettre des données sur les interactions avec les tortues marines, si disponibles, étant donné que l'era avait dû tirer des conclusions d'autres océans et faire aussi des extrapolations sur la base des quelques données qui avaient été fournies, ce qui pourrait donner une image déformée de la réalité et risquerait de biaiser la formulation d'avis. Le Sous-comité s'est montré préoccupé par le fait que des zones réputées à risque sont en fait des zones pour lesquelles des données sont disponibles, alors que les zones pour lesquelles on dispose de peu de données pourraient ne pas recevoir l'attention qu'elles méritent. Même si l'era a tenu compte de ce problème, il pourrait être résolu par la collaboration d'experts qui pourraient combler de nombreuses lacunes dans les données. Au cours de la réunion, plusieurs scientifiques ont présenté des données mises à jour et plus complètes (taux de capture, résumés du programme d'observateurs et données de suivi par satellite) qui mettent en évidence l'importance d'intégrer ces données supplémentaires dans la prochaine analyse (cf. présentations qui ont fourni des informations mises à jour). L'estimation du nombre total de prises accessoires doit être réexaminée en utilisant les informations de capture les plus complètes ainsi que des procédures d'estimation révisées. Utilisation des unités régionales de gestion (RMU) Des inquiétudes ont été exprimées au sujet du fait que l'utilisation d'unités RMU peut ne pas être la façon la plus appropriée de structurer le modèle et que le risque devrait plutôt être évalué pour les principales régions dans lesquelles les flottilles interagissent avec les populations de tortues marines. D'autres méthodes de structuration 2148

133 du modèle pourraient être explorées. Le Sous-comité était particulièrement préoccupé par les RMU qui englobent de vastes zones, mais qui présentent relativement peu de sites de nidification. L'approche RMU ne tient pas toujours compte des disparités des tailles des populations reproductrices au sein de chaque région. Caractère saisonnier Le Sous-comité a relevé l'importance des caractéristiques saisonnières de l'effort de pêche à la palangre et a indiqué que cela devrait être pris en compte dans l'analyse. Les données de suivi, ventilées par saison pourrait résoudre cette question, en les reliant aux données d'observation sur les interactions. Un problème pourrait se poser car les études sur la CPUE ne sont pas toujours liées aux données saisonnières. Le manque de données complique cette tâche. La plupart des données utilisées dans l'étude proviennent du document rédigé en 2012 par Coelho (2012). Les données associées sur les prises accessoires de tortues marines font parfois défaut. Caractéristiques de la pêche palangrière Le Sous-comité a noté que les palangres mouillées à faible profondeur étaient plus susceptibles d'interagir avec les tortues que les palangres de profondeur. Cela s'explique par le chevauchement entre la distribution verticale des tortues marines et la profondeur de l'engin de pêche. Il a été suggéré de séparer les palangres mouillées à faible profondeur des palangres de profondeur afin de tenir compte de la différence de la mortalité des tortues marines entre ces deux composantes. Ces informations proviennent de l'era sur les requins réalisée en Données de suivi des tortues Les collaborateurs de l'era ont fait part d'un problème concernant le fait que, même s'ils disposent de cartes du suivi des mouvements des tortues marines, ils n'ont pas accès aux données réelles concernant ces mouvements et, par conséquent, ils ne sont pas en mesure d'inclure un grand nombre de ces informations dans l'era actuelle. En outre, les données de suivi concernent principalement les femelles nidifiantes provenant des plages de ponte et ne concernent pas les juvéniles ou les mâles. Le Sous-comité s'est demandé combien de marques étaient nécessaires pour atteindre un niveau fiable des mouvements des tortues marines et a débattu de la question de savoir si la fiabilité des mouvements suivis pouvait être pondérée. Productivité et mortalité des tortues marines (naturelle, par pêche, suivant la remise à l eau) La pertinence des indices approchants de la mortalité naturelle et de la mortalité par pêche a été examinée. On a abordé le fait que l'utilisation du nombre de femelles reproductrices dans l'indice approchant de Z (mortalité totale) pourrait ne pas être appropriée et qu'il conviendrait d'explorer d'autres sources de mortalité, telles que la documentation disponible et les observations directes de M (mortalité naturelle) reposant sur la fidélité au site de nidification et le marquage des femelles sur la plage natale sur de longues périodes. En ce qui concerne les estimations de F (mortalité par pêche), la survie des tortues marines après la remise à l eau est difficile à déterminer. Pour toutes les estimations de la mortalité, une étude plus poussée de la littérature disponible est nécessaire pour obtenir des estimations complémentaires de ces paramètres. À titre d exemple, lorsqu'un membre de l'équipage d'un palangrier retire un hameçon et relâche une tortue, l'observateur pourrait consigner ce cas comme une remise à l'eau à l'état vivant alors qu'en réalité la tortue ne survit pas. La valeur de la mortalité suivant la remise à l eau de diverses tortues marines dans différentes flottilles doit être examinée de manière plus approfondie. Inclusion des pêcheries côtières Le Sous-comité a reconnu qu'il était important d'inclure l'impact sur les tortues marines de la pêche côtière dans la zone de la Convention de l'iccat. Le Sous-comité a reconnu qu'il était nécessaire d'intégrer différents engins ainsi que la sélectivité (taille, espèces, etc.) de ces pêcheries. Il a été noté tout particulièrement qu'il était nécessaire d'inclure les filets maillants dans l'analyse. Même si actuellement la quantité d'informations sur les prises accessoires des pêcheries de la zone de la Convention de l'iccat est trop limitée afin de pouvoir donner suite à ce secteur, une collaboration plus intense d'experts pourrait fournir des informations utiles pour cette évaluation. 9.3 Liste des changements/améliorations à apporter à l'era finale Au terme du débat initial sur l'évaluation des risques écologiques, il a été largement reconnu que l'évaluation des risques écologiques pourrait fournir au SCRS un outil permettant d'évaluer le risque relatif des engins de pêche 2149

134 de l'iccat sur les populations de tortues marines. Tous les fichiers utilisés pour réaliser l'era ont été fournis au Sous-comité afin de fournir un outil transparent que le Sous-comité pourrait utiliser pour explorer les données d'entrée, la structure du modèle et le paramétrage de l'era et pour formuler des recommandations visant à l'améliorer. Le Sous-comité a reconnu que cette approche était une première étape utile pour répondre à la demande de la Commission, même si davantage de travaux étaient nécessaires pour modifier et améliorer l'évaluation. Il a été reconnu que cette ERA amorce un processus qui appellera la collaboration et le perfectionnement en vue de formuler un avis de gestion des populations de tortues marines affectées par les pêcheries de l'iccat. À cette fin, le Sous-comité a discuté et formulé des recommandations devant être évaluées, et des révisions à entreprendre à court (avant septembre 2013), moyen ( ) et à long terme (2015). Les recommandations à court terme sont destinées à être incorporées dans une mise à jour de l'era réalisée par les collaborateurs de l'era avant la réunion du SCRS en septembre, tandis que les suggestions à moyen et long terme seront incorporées dans les futures itérations de l'era grâce à la collaboration au sein du Sous-comité. L'ERA, dans sa structure actuelle, examine le risque par engin (LL-PS.) X espèce X RMU. Afin d améliorer l'era, le Groupe a formulé les recommandations suivantes : 1) À court terme (2013, avant la plénière du SCRS) a) Les experts des CPC participantes transmettront leurs commentaires aux auteurs à des fins de considération et récapitulation. Les auteurs prépareront une brève réponse en septembre. b) La flottille palangrière (LL) devrait être classée en deux catégories : palangre profonde et palangre peu profonde. Les estimations d'effdis formulées pour l'era sur les requins seront utilisées à cette fin. c) Stratifier l'information sur la flottille et les prises accessoires par saison. Explorer la perte des données et l'effet sur l'analyse. Ceci démontrera vraisemblablement le besoin d'informations saisonnières additionnelles. d) Réaliser une analyse des lacunes afin d'identifier les flottilles/lieux où les données font défaut et où des efforts de recherche additionnels sont requis. Il est important d'identifier les CPC qui ont déclaré des prises accessoires nulles de tortues marines ou qui n'ont réalisé aucune déclaration et ne disposent pas de programmes nationaux d'observateurs, car il se pourrait qu'elles rencontrent des tortues marines mais qu'elles n'aient pas de mécanismes de collecte de données pour les déclarer. e) Les indices approchants de mortalité (naturelle et provoquée par la pêche) doivent être révisés. Les niveaux de risque devraient également faire l'objet d'une reclassification (c.-à-d. quand peuton parler de risque faible, moyen et élevé). Il conviendrait d'explorer d'autres indices approchants, y compris les estimations émanant de la documentation disponible. Il est recommandé d'utiliser des indices standardisés de capture par effort des principales flottilles palangrières de l ICCAT, si disponibles. 2) À moyen terme ( ) a) Identifier d'importantes données d'entrée (p.ex. prise accessoire par unité d'effort-bpue, données de prises accessoires par espèce incluant les estimations de mortalité à la remise à l'eau) requises pour améliorer l'analyse ERA et assembler à des fins d'utilisation - si nécessaire, les données pourraient être agrégées à l'échelle spatiale du RMU ou à une autre unité spatiale déterminée par une consultation d'experts. b) Obtenir des données détaillées des satellites. Déterminer le chevauchement vertical des pêcheries palangrières et des tortues. c) Déterminer l'effort de pêche nocturne par opposition à diurne. Cette information est déjà intégrée dans l'era sur les requins et sera fournie par les auteurs de cette étude. d) Utiliser les informations disponibles sur la taille des tortues capturées par les différentes flottilles dans la composante de susceptibilité de l'analyse ERA. e) Inclure des informations sur les vastes zones dotées de faibles taux de capture par espèce de tortue marine. f) Inclure des informations sur la confiance dans les estimations de BPUE (p.ex. CV, niveau de la couverture d observateur) et déterminer la mesure dans laquelle la couverture d observateurs est représentative des opérations de pêche. 2150

135 g) Assembler des informations additionnelles de suivi par satellite, si possible (à travers seaturtle.org par exemple). h) Envisager l'inclusion du «nombre de plages de nidification» et le nombre/densité de nids sur ces plages en tant qu'élément dans la portion de productivité de l'era. i) Prévoir une analyse du lieu où les tortues passent le plus de temps (p.ex. densité Kernel) à l'aide des informations de suivi par satellite, si possible. 3) À long terme > 2015 a) Obtenir des informations sur les programmes d'observateurs côtiers et explorer d'autres façons d'estimer l'effort de pêche artisanal et la prise accessoire (p.ex. contacter d'autres groupes qui mènent des travaux similaires - ONG/instituts de recherche, etc.). Ceci est spécialement important pour les pêcheries de filet maillant qui pourraient avoir un impact important sur les populations de tortues marines. b) Le risque (tel qu'il est considéré dans l'era) est défini comme étant la combinaison de l'ampleur des conséquences d'un danger, s'il se produit, et de la probabilité que ces conséquences se produisent. L'identification du danger, la probabilité et les conséquences sont tous des termes caractérisés par et décrits et mesurés avec divers types et degrés d'incertitude. Même s'il est impossible d'identifier toutes les incertitudes, la crédibilité, la précision et l'intégrité scientifiques d'une évaluation des risques dépendent de la qualité de son analyse de l'incertitude (Burgman, 2005). Il conviendrait donc de mettre au point à l'avenir des méthodes appropriées visant à identifier, décrire et ensuite aborder les diverses formes d'incertitude. 10. Examiner les mesures et les protocoles disponibles d'atténuation et de remise à l'eau des prises accessoires de tortues marines en toute sécurité, et formuler des recommandations, si nécessaire Cette question a été évaluée lors de la réunion du Sous-comité des écosystèmes ainsi qu'aux réunions de 2011 et Le Tableau 3 présente un résumé des documents et rapports pertinents examinés lors de réunions précédentes. Lors de la réunion du Sous-comité des écosystèmes de 2011, un document récapitulatif de la littérature disponible concernant l'atténuation des prises accessoires de tortues marines a également été examiné. On a également examiné de nouveaux documents sur l'atténuation des prises accessoires de tortues marines et les protocoles de remise à l'eau en toute sécurité. Le document SCRS/2013/129 fait état d'une expérience de l'utilisation d hameçons circulaires par un palangrier du Taipei chinois dans l'océan Atlantique tropical. L'expérience a été réalisée entre septembre 2012 et mai 2013 entre 2ºS et 12º S et 17º W et 26ºW. Les hameçons traditionnels à thons et les hameçons circulaires ont été placés en alternance tout au long de la partie expérimentale de la palangre avec un ratio de 1:1. Les observateurs des pêches ont suivi 200 opérations au cours desquelles 36 tortues se sont hameçonnées et 19 se sont enchevêtrées dans la ligne principale, l'avançon ou la ligne de flotteurs. Les essais de randomisation ont démontré qu'il n'existait pas de différences significatives des taux de capture des tortues marines en fonction du type d'hameçon (p=1,000). En revanche, le taux de capture du thon obèse était significativement supérieur (p = 0,0002), ainsi que de l'albacore (p= 0,0045), de l'espadon (p = 0,0001) et du requin peau bleue (p= 0,0209) capturés au moyen d hameçons circulaires par rapport aux hameçons traditionnels à thons. Des taux de capture significativement plus élevés ont été observés dans le cas du germon (p = 0,0010) et du Tetrapturus spp. (p = 0,0097) capturés au moyen d hameçons à thons par rapport aux hameçons circulaires. Le Sous-comité a noté que la mortalité suivant la remise à l eau observée pendant cette étude était élevée par rapport aux taux de prise accessoire de la flottille du Taipei chinois dans son ensemble, mais l'auteur a déclaré que cette étude a été réalisée à un endroit et à un moment où les taux de prise accessoire sont plus élevés que la moyenne. L'étude a fait apparaître que les taux de capture au moyen d hameçons circulaires et d hameçons en forme de J étaient identiques, mais que le taux de survie était plus élevé dans le cas des hameçons circulaires. Le document SCRS/2013/131 analyse les données obtenues dans le cadre du programme d échantillonnage de l'institut espagnol d'océanographie (IEO) ainsi que les carnets de pêche de la flottille palangrière espagnole ayant ciblé l'espadon de la Méditerranée en 2011 et Cette flottille présente trois composantes, à savoir la palangre de surface, la palangre semi-pélagique et la palangre profonde ainsi que différents taux de capture de prise accessoire, ce qui donne lieu par conséquent à différents impacts sur les populations de tortues marines. 2151

136 L'espadon, totalisant plus de 80% de la capture totale (fourchette de 82 à 93%), est l'espèce la plus ciblée. En 2004, une étude pilote a été réalisée en utilisant 10 types d hameçons (en forme de J et circulaires), quatre appâts et différentes profondeurs de mouillage. Les résultats de cette étude ont montré que l'impact des palangres sur les tortues marines dépend principalement de la profondeur de l'engin, même si la zone et les effets saisonniers sont également importants. Les résultats montrent également que l'impact de cette pêcherie sur les tortues marines peut être considéré comme faible au cours de l'année. Le Sous-comité a noté que le niveau de prises accessoires de tortues est faible pour les opérations à toutes les profondeurs et a souhaité connaître le type d'hameçon qui a été utilisé. L'auteur a indiqué que les pêcheurs utilisent des hameçons en forme de J #1. On a aussi déclaré que les expériences ont été réalisées avec des hameçons circulaires, mais qu'elles n'ont pas été poursuivies en raison de la diminution des captures d'espadon et de la difficulté qu éprouvent les pêcheurs pour libérer les tortues de cet engin. Le Sous-comité s'est également penché sur les travaux décrits précédemment (Sales et al, 2010) concernant l'efficacité des hameçons circulaires pour atténuer les prises accessoires des tortues marines et la capture des espèces cibles dans le cadre de la pêcherie pélagique palangrière brésilienne. Entre 2004 et 2008, on a comparé les résultats obtenus par des hameçons circulaires sans anneau (18/0, 10º d'alignement) et des hameçons droits en forme de J 9/0 (contrôle) dans la pêcherie palangrière brésilienne ciblant l'espadon, les thonidés et les requins. Au cours de cette expérience, 26 sorties et 229 opérations ont été réalisées au cours desquelles hameçons ont été placés, en alternant hameçons circulaires et hameçons en forme de J et en utilisant le maquereau espagnol (Scomber japonicus) comme appât. Un total de 60 espèces, comprenant des tortues et des oiseaux de mer, a été capturé. Des analyses statistiques (tests de Mantel-Haenszel χ ²) des espèces dont au moins 20 spécimens ont été capturés ont été effectuées. Les hameçons circulaires ont donné lieu à une diminution importante des prises de tortues caouannes (Caretta caretta) (55 %) et de tortues luth (Dermochelys coriacea) (65 %). L'utilisation d hameçons circulaires s'est traduite par une augmentation des taux capture de thonidés (thon obèse, Thunnus obesus, et germon Thunnus alalunga) et de requins (requin peau bleue, Prionace glauca, et requin requiem de la famille des Carcharhinidés). Aucune différence n'a été observée en ce qui concerne l'albacore (Thunnus albacares), le requin-taupe bleu (Isurus oxyrinchus) et les requins-marteau (Sphyrna lewini et S. zygaena). D'autre part, le taux de capture de l'espadon (Xiphias gladius) a diminué de manière significative lorsque des hameçons circulaires étaient utilisés. En outre, l'utilisation d hameçons circulaires a significativement réduit les taux de prise accessoire d'espèces telles que la pastenague violette (Pteroplatytrygon violacea) et le makaire blanc (Tetrapturus albidus). Les hameçons circulaires ont donné les mêmes résultats que les hameçons en forme de J en ce qui concerne de nombreuses espèces, et ont augmenté les captures d'espèces commercialisables telles que la castagnole (Brama brama), l'escolier noir (Lepidocybium flavobrunneum) et le requin-taupe commun (Lamna nasus). Les résultats démontrent l'efficacité des hameçons circulaires pour la conservation de la tortue caouanne et de la tortue luth, car ils permettent d'améliorer le taux de capture de la plupart des espèces cibles et de réduire significativement les prises accessoires des espèces les plus communes, telles que la pastenague violette, et par conséquent d'améliorer économiquement cette pêcherie. Le Sous-comité a souhaité connaître la cause expliquant les différences entre les taux de capture des hameçons circulaires et des hameçons en forme de J observés dans cette étude, mais pas dans le document SCRS/2013/129. L'auteur a noté que la configuration de l'engin utilisé était différente en ce qui concerne les anneaux des hameçons. Ces derniers sont utilisés dans la pêche du Taipei chinois, alors que ce n'est pas le cas au Brésil. Le Sous-comité a également noté que, si les études ne s'accordaient pas sur le fait que les hameçons circulaires réduisent le taux de capture des tortues marines, elles s'accordaient toutefois sur le fait que leur utilisation réduit la mortalité des prises accessoires, probablement car les tortues marines avalent plus fréquemment les hameçons en forme de J. Une deuxième différence de la configuration de l'engin a été également relevée; la distance entre les flotteurs et les avançons était plus courte dans le cas de la pêcherie palangrière brésilienne que dans la pêcherie du Taipei chinois. On ne sait pas précisément dans quelle mesure cela a influencé les taux de capture, mais l'auteur a fait remarquer que des avançons plus courts diminue la profondeur de mouillage, ce qui pourrait améliorer la survie des tortues marines. Le document SCRS/2013/128 fait état des taux de capture préliminaires des prises accessoires de tortues marines des flottilles palangrières du Taipei chinois dans l'océan Atlantique. Les données recueillies couvrent opérations observées et 46,0 millions d'hameçons. 761 tortues ont été capturées. Les principales espèces capturées accidentellement étaient la tortue luth (59,9 %), la tortue olivâtre (26,9 %) et la tortue caouanne (8,1 %). La plupart des tortues marines se sont hameçonnées (58,9%). La plupart des tortues luth et des tortues olivâtres ont été capturées dans les régions tropicales, en particulier dans le golfe de Guinée. La tortue caouanne a été capturée non seulement dans les zones tropicales mais également dans le sud-ouest de l'océan Atlantique. 2152

137 Le taux nominal de capture accidentelle s'élevait à 0,018 pour 1000 hameçons en moyenne, oscillant entre 0,0012 et 0,0219 pour 1000 hameçons par zone. Le Sous-comité a noté que les données utilisées dans les analyses concernaient l'ensemble de l'océan Atlantique et a suggéré que les analyses devraient être affinées en utilisant les composants saisonniers et spatiaux. Il a également été noté que la proportion des observations des opérations au cours desquelles des prises accessoires ont été signalées était très faible, et il a été suggéré qu'une distribution binomiale négative ou une distribution Poisson devrait être utilisée pour des données, au lieu d'une distribution normale. L'auteur a également déclaré que l'engin que les navires utilisent dans les régions tropicales est différent de celui utilisé dans les régions du nord et du sud, car ils ciblent le thon obèse. Le Sous-comité a souhaité connaître quelles parties du corps des tortues marines sont hameçonnées. L'auteur a indiqué que cette information n'était pas toujours consignée, mais il a également été déclaré que les hameçons sont souvent plantés dans les nageoires des tortues luth, plutôt que dans la bouche car elles ne mangent pas les appâts. 11. Réponse à la Commission concernant la Recommandation de l'iccat sur les prises accessoires de tortues marines dans les pêcheries de l ICCAT (Rec ) En 2010, la Commission a recommandé que : 1) le SCRS réalise une évaluation des effets de la prise accidentelle de tortues marines dans le cadre des pêcheries de l'iccat dans les meilleurs délais et en 2013 au plus tard [Rec ; paragraphe 5]. 2) Après la réalisation de l évaluation initiale et la présentation des résultats à la Commission, le SCRS devra formuler un avis à la Commission relatif à la planification des évaluations futures [Rec ; paragraphe 5]. 3) Le SCRS devra également formuler un avis à la Commission sur les approches d'atténuation des prises accessoires de tortues marines dans les pêcheries de l'iccat, y compris sur la réduction du nombre d'interactions et/ou de la mortalité associée à ces interactions [Rec ; paragraphe 4]. 4) S il y a lieu, la Commission et ses CPC devraient, de manière individuelle et collective, déployer des efforts de renforcement des capacités et se livrer à d autres activités de coopération afin de soutenir la mise en œuvre effective de ladite Recommandation, y compris en concluant des accords de coopération avec d autres organismes internationaux adéquats. En ce qui concerne l'atténuation des prises accessoires de tortues marines dans les pêcheries relevant de l ICCAT, le SCRS recommande ce qui suit : 1) Le SCRS réitère les recommandations formulées antérieurement à la Commission [Rec ], à savoir : a. Les senneurs opérant dans la zone de la Convention évitent dans la mesure du possible d encercler les tortues marines, relâchent les tortues marines encerclées ou emmêlées, y compris avec les DCP, autant que possible, et déclarent les interactions entre les sennes et/ou les DCP et les tortues marines. b. Les palangriers pélagiques opérant dans la zone de la Convention ont à leur bord du matériel permettant de manipuler, de démêler et de relâcher en toute sécurité les tortues marines afin de maximiser leurs probabilités de survie. c. Les pêcheurs à bord des palangriers pélagiques opérant sous leur pavillon utilisent le matériel spécifié au point b susmentionné afin de maximiser les probabilités de survie des tortues marines et sont formés aux techniques de manipulation et de remise en liberté en toute sécurité. d. Les CPC rendent compte dans leurs rapports annuels de toutes les autres actions pertinentes prises en vue de mettre en œuvre les Directives de la FAO visant à réduire la mortalité des tortues de mer dans les opérations de pêche en ce qui concerne les pêcheries de l ICCAT. 2) De surcroît, afin de réduire la mortalité des tortues marines capturées accidentellement, le SCRS recommande spécifiquement ce qui suit : a. En ce qui concerne les pratiques de manipulation en toute sécurité : i. Pour sortir de l'eau une tortue, il convient d'utiliser un panier approprié ou une épuisette pour hisser à bord la tortue qui s'est planté un hameçon ou qui s'est emmêlée dans une ligne. Pour hisser une tortue hors de l'eau, il ne faut pas tirer sur la ligne qui est fixée ou emmêlée autour de son corps. 2153

138 ii. Avant de les remettre à l'eau, les opérateurs du navire ou l'équipage évaluent l'état des tortues marines qui sont capturées ou emmêlées. Les tortues ne pouvant pas nager, étant inconscientes ou ne réagissant pas devraient être hissées/maintenues à bord et il convient de leur porter secours afin de maximiser leur chance de survie avant leur remise à l'eau. Ces pratiques sont décrites plus avant dans les Directives de la FAO visant à réduire la mortalité des tortues marines dans les opérations de pêche. iii. Les tortues manipulées dans les opérations de pêche ou dans le cadre de programmes nationaux d'observateurs (p.ex. activités de marquage) doivent être traitées conformément aux Directives de la FAO visant à réduire la mortalité des tortues marines dans les opérations de pêche. b. En ce qui concerne l'emploi de coupe-lignes : i. Les palangriers transportent à leur bord des coupe-lignes qu'ils utilisent lorsqu'il n'est pas possible de retirer l'hameçon en toute sécurité afin de remettre la tortue à l'eau. ii. Les autres types de navires qui utilisent des engins dans lesquels les tortues marines sont susceptibles de s'emmêler devraient avoir à leur bord des coupe-lignes et utiliser ces outils pour retirer l'engin en toute sécurité et remettre les tortues à l'eau. c. En ce qui concerne l'emploi de dispositifs de retrait de l'hameçon : i. Les palangriers transportent à bord des dispositifs de retrait de l'hameçon afin de décrocher efficacement et en toute sécurité l'hameçon de la tortue marine. Le Sous-comité recommande en outre de ne pas tenter de retirer un hameçon qui a été avalé. En revanche, il conviendra de couper la ligne le plus près possible de l'hameçon. En ce qui concerne l'évaluation de l'impact des pêcheries de l'iccat sur les tortues marines, le SCRS a entamé une évaluation des risques écologiques (ERA) pour les tortues marines en Les progrès réalisés à ce jour sont les suivants : 1) En 2013, l'iccat a conclu un contrat à court terme dans le but d'élaborer une ERA préliminaire s'appliquant aux espèces de tortues marines capturées dans les pêcheries de l ICCAT. L'ERA a eu recours aux données fournies au Secrétariat par les CPC en 2011 et 2012 et réunies dans le cadre d'un contrat à court terme financé par l'iccat en 2012 et à d'autres sources de données rassemblées par le prestataire. 2) À sa réunion intersession de 2013, le Sous-comité des écosystèmes a examiné les progrès de l ERA réalisés jusqu'à présent et a formulé des recommandations importantes en vue d'améliorer l'évaluation à court terme (avant octobre 2013), à moyen terme ( ) et à long terme (après 2015), dont une demande de données mises à jour/supplémentaires des CPC. 3) Le SCRS continuera à améliorer l'era et formulera un avis à la Commission en ce qui concerne son plan concernant les analyses de l'impact sur les tortues marines à la réunion de Autres questions 12.1 Présentations Un exposé a été présenté sur les prises accessoires de tortues marines de la flottille marocaine. Des entretiens avec des pêcheurs thoniers marocains entre 2008 et 2011 ont fait apparaître qu'ils rencontraient une tortue tous les 90 à 100 jours de pêche. Les jours de pêche par sortie oscillent entre 1 à 3 jours, mais compte tenu des caractéristiques techniques des navires, les opérations durent généralement moins de 24 heures. Sur la base d'un échantillon de 100 navires spécialisés dans la pêche au thon au Sud d'agadir, dans la ZEE marocaine, il a été déterminé que cinq navires de pêche sur six observent une tortue. Le taux de capture des tortues dans cette pêcherie s'élève à une tortue tous les 20 jours de pêche. Dans cette zone, les tortues capturées sont directement relâchées. Les informations provenant du programme d'observateurs nationaux sur les prises de requins à bord des palangriers ont indiqué que les tortues ne sont pas présentes ou que leurs interactions liées à cette pêcherie ne sont pas significatives. Le Maroc propose de diriger une campagne de sensibilisation avec les différents opérateurs de la région aux fins de la protection des tortues marines. Le document SCRS/2013/133 présentait les résultats de l'initiative sur la tortue luth transatlantique (TALCIN). La deuxième phase de ces travaux a fourni la première analyse à échelle océanique de la distribution spatiotemporelle de la tortue luth, sur la base du marquage électronique et a de plus déterminé le chevauchement de la 2154

139 distribution avec l'effort de pêche palangrier. Les données donnent à penser que l'atlantique est probablement composé de deux unités régionales de gestion: Nord et Sud (cette dernière incluant des tortues de l'afrique du Sud). Même si la distribution des tortues et des pêcheries est très différente, les auteurs mettent en évidence neuf zones présentant une susceptibilité élevée de prises accessoires (quatre dans l'atlantique Nord et cinq dans l'atlantique équatorial/sud) qui devraient faire l'objet de recherches ciblées plus poussées. Ceci est étayé par des déclarations de prises accessoires de tortues luth dans huit sites sur neuf (document SCRS/2013/133). Le Sous-comité a noté qu'une analyse avait été menée sur le chevauchement entre les zones où la pression exercée par la pêche est très élevée et les zones d'habitat des tortues luth par trimestre, et il a été suggéré que l'évaluation des risques écologiques de tortues marines pourrait utiliser une approche similaire. Le Sous-comité a également noté que l'habitat des tortues luth pourrait se chevaucher avec des zones où la pression de la pêche est élevée et faible. À cet égard, il a été suggéré que ces informations pourraient être comparées aux données des observateurs afin de déterminer s'il existe une corrélation. Le Sous-comité a recommandé de poursuivre les comparaisons entre les zones très fréquentées par les tortues luth et les zones où des opérations de pêche à la palangre peu profonde et à la palangre profonde sont réalisées au moyen d'analyses telles que celles développées dans le cadre de l'évaluation des risques écologiques sur les requins. Le document SCRS/2013/135 (observations et abondance des tortues marines aux Açores) présente une analyse des observations et de l'effort d'observation des tortues marines par les observateurs déployés à bord des canneurs thoniers des Açores. Un total de sorties a été observé de 2001 à 2012, et prospections d'une durée moyenne de 15 minutes ont été effectuées. Les observations occasionnelles de tortues, de 1998 à 2000, sont également présentées. En ce qui concerne les observations et l'effort d'observation de la tortue caouanne dans la région des Açores entre 2001 et 2012, un indice d'abondance relative a été estimé au moyen d'une approche de modélisation linéaire généralisée (GLM). L'analyse montre que, suite au faible niveau d'observation par unité d'effort en 2001, l'indice d'abondance a atteint son niveau le plus élevé en 2002, il a diminué au cours des années suivantes et est resté stable depuis 2008 jusqu'à présent. Les informations sur les observations de tortues luth et de tortues vertes dans la région des Açores ont également été fournies. Les informations recueillies dans la région de Madère, de certaines années et mois, comprennent le nombre de prospections et de tortues caouannes observées. Les indices nominaux (toutes espèces et zones confondues) et les SPUE standardisées (observation par unité d'effort) des tortues caouannes dans les Açores sont également inclus. Il s'agit du seul indice indépendant des pêcheries en haute mer. Le Sous-comité a fait remarquer que la possibilité d'observer des tortues peut être affectée par des facteurs tels que l'état de la mer, la nébulosité et l'angle de vision des observateurs, et que ces facteurs devraient être intégrés à l'analyse. Il a également été noté que le nombre de tortues marines a connu une augmentation avérée sur les plages de nidification des États-Unis, et étant donné qu'il est notoire que ces tortues peuvent migrer vers les Açores, le Sous-comité a estimé qu'il serait intéressant d'examiner si les observations dans les Açores augmentent dans les années à venir. L'expert a fourni davantage d'informations sur les résultats des analyses génétiques réalisées, qui confirment que la totalité des tortues présentes dans la région des Açores proviennent des États-Unis et du Mexique. Il a été noté que cette information ne correspond pas à la RMU CcATLNE d'après Wallace et al Le Sous-comité a conseillé qu'il serait très utile de déterminer le taux de croissance des tortues en utilisant les données de récupération des marques. Le document SCRS/2013/138 décrit une approche de recherche collaborative pour le renforcement des capacités dans la science halieutique. L'auteur a énuméré les avantages de la recherche collaborative sur les pêcheries : 1) renforcement de la crédibilité et de la légitimité des résultats scientifiques dans l'esprit des parties prenantes, ce qui peut accroître l'acceptation des mesures de gestion, 2) engins et pratiques de pêche innovants et adoptés qui réduisent les prises accessoires et minimisent l'impact de la pêche sur l'habitat, 3) accroissement de la compréhension mutuelle et de la confiance entre les partenaires et 4) possibilités d'intégrer diverses sources de connaissances sur l'environnement côtier et marin. L'auteur a déclaré que l'approche de recherche collaborative est de plus en plus adoptée par les gouvernements, le secteur industriel, les associations et les organisations de conservation de la nature, tels que The Nature Conservancy et le concours Smart Gear du WWF. L auteur a discuté d un nouveau programme collaboratif de bourses d études consacré à la recherche halieutique lancé au Virginia Institute of Marine Science (VIMS). L'auteur a noté que les participants pourraient tirer profit d'une collaboration avec un expert en pêcheries thonières et prises accessoires et a invité le Sous-comité à envisager de collaborer à ce programme. Le Sous-comité a estimé que cette proposition était une approche intéressante et qu'elle fournit l'occasion de construire un soutien extérieur et des capacités de recherche collaborative avec des scientifiques, le secteur industriel, les gestionnaires et les ONG. Le Sous-comité a discuté du fait que les participants du VIMS devraient prendre contact avec le chef de la délégation de la CPC appropriée afin de participer aux futures réunions du SCRS. 2155

140 12.2 Progrès accomplis en vue de l'harmonisation des formulaires de collecte de données des programmes d observateurs à bord de palangriers Lors de sa réunion de 2012, le Sous-comité des écosystèmes a recommandé que l'iccat se coordonne avec le Groupe de travail technique conjoint des prises accessoires des ORGP thonières (BJTWG, selon les sigles anglais) afin d'assumer un rôle de chef de file en développant des normes minimum de collecte de données d'observateurs palangriers harmonisées pour les ORGP thonières. Le Secrétariat de l ICCAT a depuis pris contact avec Simon Nicol, le Président de ce groupe, et a commencé à recueillir les formulaires pour les activités de collecte de données des programmes d observateurs palangriers de l ICCAT, WCPFC, IATTC, CTOI et CCSBT. Étant donné que l'iccat ne coordonne pas directement de programmes d'observation à grande échelle, il est nécessaire de communiquer et de coopérer avec les programmes d'observateurs nationaux. À cette fin, le Secrétariat a contacté les coordinateurs des programmes nationaux d observateurs palangriers qui opèrent dans la zone de la Convention ICCAT afin d obtenir leurs formulaires de collecte de données. Ceux-ci seront ensuite utilisés pour identifier et recommander des normes minimum de collecte des données. 13. Recommandations Le Sous-comité a reconnu l'excellent travail des Drs Andrea Angel, Ross Wanless et Ronel Nel dans la réalisation d'une évaluation préliminaire des risques écologiques (ERA) pour les tortues marines. Leur travail de préparation a accéléré les travaux du Sous-comité et fourni une excellente base à l'évaluation actuellement en cours des impacts des prises accessoires de tortues marines dans les pêcheries relevant de l ICCAT. Le Souscomité a reconnu la valeur de cette initiative de l'iccat visant à fournir un appui financier afin de recruter des experts pour contribuer aux travaux du SCRS, et il a fortement recommandé de poursuivre ces activités productives. Écosystèmes Le Sous-comité a reconnu la valeur de l'exercice de dégroupage visant à définir les objectifs écosystémiques du SCRS. Il est recommandé que le co-coordinateur du Sous-comité des écosystèmes élabore une prospection destinée à dresser la liste des objectifs conceptuels de la gestion des pêcheries fondée sur l'écosystème (EBFM) afin de la distribuer aux mandataires du SCRS. Le co-coordinateur rassemblera les résultats avant la séance plénière du SCRS. Il est recommandé que des fonds pour les voyages soient alloués en appui à la participation d'experts externes dans le but de contribuer à la mise au point des outils scientifiques requis pour mettre en œuvre les approches EBFM. Il est recommandé que le Groupe de travail poursuive sa collaboration avec Sargasso Sea Alliance en ce qui concerne l'analyse de l'importance écologique de la mer des Sargasses pour les thonidés et les espèces de thonidés apparentées ainsi que les espèces associées du point de vue écologique. Prises accessoires Le Sous-comité recommande que le SCRS contribue aux approches collaboratives de recherche sur les pêcheries en permettant aux parties prenantes de s'impliquer dans des initiatives visant à aborder les interactions potentielles des pêcheries de l'iccat avec les espèces accessoires. ERA pour les tortues marines Les CPC devraient fournir des données de prises accessoires selon les normes de la Tâche II. Si cela n'est pas possible, le Sous-comité recommande aux CPC que les données concernant les prises accessoires de tortues marines soient fournies par espèce, avec des informations spatio-temporelles (p.ex. 5ºx5º) qui permettraient d'assigner la RMU et le trimestre. Le Sous-comité reconnaît la nécessité d'inclure des informations sur les pêcheries artisanales qui opèrent dans la zone de la Convention de l ICCAT et il encourage les CPC à transmettre les informations pertinentes. Le Sous-comité recommande que le marquage (électronique et conventionnel) des tortues marines soit poursuivi et que les informations sur ces expériences soient mises à sa disposition. Le Sous-comité recommande la réalisation d'études génétiques sur les tortues marines afin de mieux appréhender quelles sont les populations qui sont le plus touchées par les pêcheries de l'iccat et il demande que cette information sur ces expériences soit mise à sa disposition. 2156

141 Atténuation des prises accessoires de tortues marines Le Sous-comité réitère les recommandations de la Commission, telles qu'adoptées dans la Rec De surcroît, afin de réduire la mortalité des tortues marines capturées accidentellement, le SCRS recommande spécifiquement ce qui suit : o En ce qui concerne les pratiques de manipulation en toute sécurité : Pour sortir de l'eau une tortue, il convient d'utiliser un panier approprié ou une épuisette pour hisser à bord la tortue qui s'est planté un hameçon ou qui s'est emmêlée dans un engin. Pour hisser une tortue hors de l'eau, il ne faut pas tirer sur la ligne qui est fixée ou emmêlée autour de son corps. Avant de les remettre à l'eau, les opérateurs du navire ou l'équipage évaluent l'état des tortues marines qui sont capturées ou emmêlées. Les tortues ne pouvant pas nager, étant inconscientes ou ne réagissant pas devraient être hissées/maintenues à bord et il convient de leur porter secours afin de maximiser leur chance de survie avant leur remise à l'eau. Ces pratiques sont décrites plus avant dans les Directives de la FAO visant à réduire la mortalité des tortues marines dans les opérations de pêche. Les tortues manipulées dans les opérations de pêche ou dans le cadre de programmes nationaux d'observateurs (p.ex. activités de marquage) doivent être traitées conformément aux Directives de la FAO visant à réduire la mortalité des tortues marines dans les opérations de pêche. o En ce qui concerne l'emploi de coupe-lignes : Les palangriers transportent à leur bord des coupe-lignes qu'ils utilisent lorsqu'il n'est pas possible de retirer l'hameçon en toute sécurité afin de remettre la tortue à l'eau. Les autres types de navires qui utilisent des engins dans lesquels les tortues marines sont susceptibles de s'emmêler devraient avoir à leur bord des coupe-lignes et utiliser ces outils pour retirer l'engin en toute sécurité et remettre les tortues à l'eau. o En ce qui concerne l'emploi de dispositifs de retrait de l'hameçon : Les palangriers transportent à bord des dispositifs de retrait de l'hameçon afin de décrocher efficacement et en toute sécurité l'hameçon de la tortue marine. Le Sous-comité recommande en outre de ne pas tenter de retirer un hameçon qui a été avalé. En revanche, il conviendra de couper la ligne le plus près possible de l'hameçon. 14. Adoption du rapport et clôture Le rapport a été adopté pendant la réunion, ainsi que le plan de travail du Sous-comité au titre de Les cocoordinateurs ont remercié le Secrétariat et les participants pour le travail assidu et le dévouement dont ils ont fait preuve. Les co-coordinateurs ont également remercié Andrea Angel, Ross Wanless et Ronel Nel d'avoir aidé le groupe à réaliser d'importantes avancées en matière de l'évaluation des risques écologiques des tortues marines qui interagissent avec les pêcheries de l'iccat, et également le Secrétariat pour le financement de ce travail crucial. La réunion a été levée. Au terme de la réunion du Sous-comité des écosystèmes, la Convention interaméricaine pour la protection et la conservation des tortues marines («IAC») a fait part de sa volonté de contribuer à l'amélioration de l'era sur les tortues marines en fournissant des informations sur le «nombre de plages de nidification» et le nombre et la densité des nids sur ces plages en tant qu'élément dans la portion de productivité de l'era. Actuellement, tous les pays membres de l'iac peuvent consulter ces informations de , également disponibles dans les rapports annuels de l'iac sur la page web de l'organisation (http://www.iacseaturtle.org/informes.htm). Références Burgman MA, Risks and decisions for conservation and environmental management. Cambridge University Press, Cambridge, UK,

142 REUNIÓN INTERSESIONES DE 2013 DEL SUBCOMITÉ DE ECOSISTEMAS Madrid, España, del 1 al 5 de julio de Apertura de la reunión, adopción del orden del día y disposiciones para la reunión La reunión se celebró en las oficinas de la Secretaría de ICCAT, en Madrid, España, del 1 al 5 de julio de En nombre de la Secretaría de ICCAT, el Dr. Paul de Bruyn deseo la bienvenida al grupo. Los co-coordinadores del Subcomité de Ecosistemas, la Dra. Shannon Cass-Calay (Estados Unidos) y el Dr. Alex Hanke (Canadá) describieron los objetivos y la logística de la reunión. El orden del día fue adoptado sin cambios (Apéndice 1). La lista de participantes se adjunta como Apéndice 2. La lista de documentos presentados a la reunión se adjunta como Apéndice 3. Los siguientes participantes actuaron como relatores: Sección Puntos 1, 10 y 14 Puntos 2-5 Punto 6 Punto 7 Punto 8 Punto 9 Punto 12 Punto 13 Relatores S. Cass-Calay A. Hanke A. Hanke, P. de Bruyn, P. de Bruyn A. Domingo, B. Giffoni T. Nalovic, P. de Bruyn T. Trott J. Pereria Parte I - Ecosistema El orden día, tal y como fue adoptado por el Grupo, incluía tres secciones consideradas importantes para el proceso de implementar un enfoque EBFM para los cuales no se ha proporcionado nueva información. Estas secciones fueron las secciones 2, 3 y 4, tituladas respectivamente a) Examen de la nueva información sobre la implementación de los efectos del ecosistema en las evaluaciones de stock mejoradas; b) examen de la nueva información sobre modelos que cuantifican la dinámica del ecosistema y predicen el impacto de perturbaciones sobre sus componentes y c) examen de la nueva información sobre modelos que cuantifican la dinámica del ecosistema y predicen el impacto de perturbaciones sobre sus componentes. No está claro por qué hubo una falta de participación en esta sección, pero el Subcomité consideró que podría tener algo que ver con la capacidad limitada de las CPC para participar en este Subcomité. Por otro lado, podría ser que se carezca de experiencia en el SCRS para participar plenamente en este campo de la ciencia. El Subcomité recomienda que esta cuestión se incluya en el plan estratégico del SCRS en el futuro, de tal modo que el SCRS pueda incrementar su capacidad de responder a las demandas de desarrollo de herramientas que permitan implementar el enfoque EBFM. Sección 5: Examen la nueva información sobre la implementación de los principios de la ordenación pesquera basada en el ecosistema El documento SCRS72013/137 cubría la evaluación del riesgo ecológico como un método para hacer un seguimiento de la implementación de un enfoque ecosistémico en la ordenación de las pesquerías en África meridional (la experiencia de Namibia). El método ERA, que es cualitativo, cuenta con respaldo de los países de África del Sur (Namibia, Sudáfrica y Angola) a través de la Comisión de la Corriente de Benguela (BCC), con la asistencia del proyecto Nansen, EAF, NAFO y WWF-África del Sur. Se escogió el método por su simplicidad y porque cuenta con directrices estructuradas para hacer un seguimiento de las implementaciones EAF. La ERA contribuyó a la identificación de cuestiones y a priorizarlas. El método proporcionó objetivos claros y temas para algunas pesquerías en la región de la corriente de Benguela. 2158

143 Se recomendaron acciones para abordar temas/riesgos y se requieren revisiones regulares para evaluar si se está progresando en la resolución de estas cuestiones. Una ventaja del enfoque ERA es que el método puede contribuir al consenso sobre diferentes temas entre una diversidad de partes interesadas. Se evaluaron un total de seis (6) pesquerías namibias utilizando esta metodología y se realizaron cuatro revisiones de la ERA. El Grupo se interesó por el nuevo enfoque y reconoció que los autores podrían proporcionar una orientación valiosa con respecto al ejercicio de "deconstrucción" iniciado durante la reunión y que se describe en la sección 7. Se recomendó que continuara la colaboración con los representantes del Ministerio de Pesca y Recursos Marinos de Namibia. 6 Examen del progreso alcanzado en el desarrollo de un caso de prueba para implementar la ordenación pesquera basada en el ecosistema. En el documento SCRS/2013/132 se presentaba información sobre la biología y ecología de un total de 18 especies de peces diferentes cuya distribución incluye el mar de Sargazos. Estas especies se dividen en cuatro grupos que se corresponden con las agrupaciones de especies de ICCAT: Grupo 1: especies de túnidos principales, entre ellas rabil, atún blanco, patudo, atún rojo y listado. Grupo 2: pez espada y marlines, entre ellos aguja azul, aguja blanca, y pez vela. Grupo 3: pequeños túnidos entre ellos peto, atún aleta negra, bacoreta y dorado y Grupo 4: tiburones, lo que incluye marrajo dientuso, tintorera, marrajo sardinero, zorro ojón y tiburón peregrino. Para cada especie, se proporcionaron informaciones y datos sobre distribución, desembarques de las pesquerías, patrones de migración y movimiento, edad y crecimiento, alimentos y hábitos alimentarios y ecología en relación con los parámetros oceanográficos, sobre todo temperatura del agua. Se debatió la importancia del Sargassum como hábitat esencial de los peces y se vinculó con los hábitos alimentarios de los túnidos y otros depredadores pelágicos. Los peces voladores son una especie presa importante en la dieta de los túnidos y marlines, dado que éstos dependen en gran medida de las malezas de Sargazos como hábitat de reproducción, el mar de los Sargazos desempeña un papel fundamental en la cadena trófica de especies pelágicas altamente migratorias en el Atlántico noroccidental. Una evaluación de la información existente sugiere que la importancia del mar de los Sargazos para varias especies pelágicas reside sobre todo por uno o más de los siguientes factores: ruta de migración, zona de reproducción, zona de cría, zona trófica, zona de invernada o zonas de parición. El Subcomité indicó que en la Resolución de ICCAT sobre el mar de los Sargazos se establece lo siguiente: 1 El SCRS examinará los datos y la información disponibles sobre el mar de los Sargazos y su importancia ecológica para los túnidos y especies afines y para las especies ecológicamente asociadas. 2 El SCRS presentará información actualizada sobre los progresos de su trabajo en 2014 y comunicará a la Comisión sus hallazgos en En cuanto al punto 1, se ha proporcionado una revisión preliminar de los datos y la información disponibles sobre el hábitat en el mar de los Sargazos lo que incluye las capturas de túnidos y de otras especies de ICCAT y, por tanto, pueden iniciarse estudios sobre la importancia de esta región. El Subcomité constató que una investigación más extensiva de este ecosistema y su importancia como hábitat esencial para los túnidos y especies afines requeriría una colaboración integrada del SCRS con grupos científicos especializados en el mar de los Sargazos. El Subcomité concluyó que para alcanzar el objetivo 2, se requeriría un plan de trabajo, trabajos de investigación en régimen de colaboración y la celebración de reuniones para evaluar adecuadamente la importancia del mar de los Sargazos como hábitat esencial de las especies de túnidos de ICCAT. El Subcomité consideró que esta tarea no podrá completarse antes de El Subcomité recomienda que continúen los contactos con los equipos de investigación del mar de los Sargazos y los científicos del Reino Unido-Bermudas para desarrollar un plan de colaboración científica con el fin de alcanzar el objetivo 2 de las Resolución Además, el Subcomité reconoció que este trabajo facilitará fundamentos útiles para adoptar esta región como base para el estudio de caso para implementar el enfoque de ordenación pesquera basada en el ecosistema (EBFM) dentro de ICCAT y que se debería seguir respaldando esta colaboración. El relator solicitó al Subcomité que determinase el modo de abordar el punto 1 de la Resolución En el documento SCRS/2013/132 se establecía un inventario de 46 páginas y la ecología de especies de peces de interés para ICCAT en el mar de los Sargazos. Con miras a poder evaluar la importancia relativa de este ecosistema para las especies de ICCAT, el Subcomité pidió que los detalles de este informe se reflejen en una 2159

144 tabla que relacione importantes parámetros del ciclo vital con su dependencia del ecosistema del mar de los Sargazos. A modo de ejercicio el Subcomité determinó los parámetros pertinentes y evaluó las dependencias de cada especie utilizando un sistema de puntuación de la presencia/ausencia. La Tabla 1 representa una evaluación preliminar de la importancia del ecosistema y también pretende servir de base para una evaluación más cuantitativa de los datos. El Subcomité recomendó que estas futuras evaluaciones: 1) Reflejen la ausencia de información. 2) Definan claramente el componente de cada especie (población o unidad de ordenación de stock) al que se aplica la evaluación. 3) Definan un sistema de puntuación racional para cada uno de los parámetros y los representen en una escala común. 4) Caractericen cada uno de los parámetros del ciclo vital utilizados. 5) Muestren las capturas de cada especie en la zona con respecto a su captura en la unidad de ordenación que la abarca. 7 Debate sobre formas de incluir valores ecosistémicos en la estandarización y evaluación de los stocks evaluados por ICCAT El relator del subgrupo sobre ordenación de pesquerías basada en el ecosistema (EBFM) del Subcomité de ecosistemas proporcionó una presentación que resumía una metodología para desarrollar un sistema de comunicación de desarrollo sostenible. El sistema de comunicación es parte de un marco de ordenación de pesquerías basado en el ecosistema y vincula explícitamente los objetivos conceptuales de la ordenación con los objetivos operativos. La metodología traduce objetivos conceptuales en objetivos operativos mediante un proceso denominado deconstrucción (unpacking). La deconstrucción implica subdividir objetivos conceptuales de un orden superior en sus componentes y volver a subdividir los componentes hasta poder definir un objetivo operativo. A continuación, se asocia el objetivo operativo con un indicador medible y un punto de referencia para facilitar la comunicación y provocar acciones de ordenación. Se mostró el modo en que estos indicadores del estado podrían formar parte de un sistema indicador más explícito en cuanto a la naturaleza de la acción de ordenación. El Subcomité constató que sería beneficioso para el Subcomité de ecosistemas obtener orientaciones sobre objetivos operativos de la Comisión. Se indicó que el plan estratégico del SCRS es una forma de procedimiento de deconstrucción. Esta deconstrucción más detallada podría inscribirse en este proceso incluyendo estrategias para realizar las diferentes tareas identificadas para poner en marcha la EBFM. El SCRS tiene que cerciorarse de qué información está disponible actualmente y de cómo utilizarla en un contexto EBFM. La utilización potencial del mar de los Sargazos como caso de estudio potencial es prometedora ya que proporciona un ejemplo de cómo estructurar el trabajo para informar a la Comisión. Se indicó una vez más que el diálogo con la Comisión es fundamental para el proceso, aunque el proceso puede iniciarse mientras tanto, con la explicación del SCRS a la Comisión de su metodología y plan propuestos (lo que incluye las necesidades de datos). El Subcomité decidió realizar un ejemplo de deconstrucción utilizando una plantilla facilitada por el relator del subgrupo de ordenación de pesquerías basadas en el ecosistema (EBFM). La plantilla se reorganizó y modificó para adaptarla a las necesidades específicas de ICCAT relacionadas con el proceso de deconstrucción. Se resaltó que se trataba de un ejercicio para que el Subcomité de ecosistemas se familiarizase con el proceso. Sería necesario una modificación y análisis más exhaustivo de la plantilla para obtener un producto final apropiado para informar a la Comisión. El ejercicio inicial de deconstrucción se incluye en el Apéndice 4. Tras este ejercicio se abrió un debate sobre si debería realizarse de forma genérica para toda la zona de ICCAT o si se debería utilizar un estudio de caso (mar de los Sargazos). Se decidió que este ejercicio inicial debería ser genérico para las especies de ICCAT en vez de solo para el mar de los Sargazos, aunque esto podría ocultar la naturaleza compleja del estudio, podría proporcionar un marco genérico para considerar la EBFM en el contexto de ICCAT. 2160

145 Al realizar el ejercicio, el Subcomité de ecosistemas realizó varios comentarios sobre el proceso. 1) Resolver los componentes a nivel multiespecífico resultaba bastante complicado. En este ejemplo el Subcomité decidió centrarse en los componentes a nivel de especies. 2) La elaboración de la lista de especies debería establecer prioridades basándose en las especies de ICCAT de interés. Finamente será una cuestión de ordenación de ICCAT. Aunque deben considerarse especies no ICCAT; éstas podrían considerarse en un nivel inferior. 3) Debe considerarse que hay otras OROP de túnidos que participan en la ordenación de stocks de peces en la región (por ejemplo, NAFO). 4) Se indicó dificultad/complejidad y el tiempo requerido para realizar este ejercicio de deconstrucción y se reconoció que era necesario celebrar unas Jornadas de trabajo o convocar un Grupo de trabajo para realizar este trabajo con fines de ordenación, lo que requerirá mucho tiempo y esfuerzo. 5) Se constató que varios de los términos utilizados en la plantilla como mantenimiento de la diversidad de la población deben volverse a redactar para que se ajusten mejor a la situación de ICCAT. También se constató que el término deconstrucción era nuevo para los participantes. Se sugirió que en vez de utilizar términos como mantener o conservar sería mejor tratar de entender o hacer un seguimiento de estos factores. Esto es especialmente pertinente para factores sobre los que la ordenación tiene muy poco o ningún control, como los factores medioambientales. Debido a la dificultada de realizar el ejercicio de deconstrucción en las plenarias, se pidió al Subcomité que enviase comentarios sobre dicho ejercicio al relator del Subcomité para que incorporase dichos comentarios en una tabla. Se realizaron los siguientes comentarios adicionales sobre el ejercicio: La deconstrucción no puede realizarse de un modo general Tiene que desglosarse a nivel regional El ejercicio tiene que centrarse en la EBFM. La ordenación integrada es demasiado amplia como punto de partida y tiene que limitarse su ámbito de actuación para que no se aleje demasiado de la EBFM. Vincular temas locales con la visión nacional e internacional para contar con el apoyo delas partes interesadas y del público. Identificar los aspectos relacionados con las pesquerías que no se tuvieron en cuenta para el ejercicio. o Recogida de cebo o Pérdida del arte o Desperdicios El término utilización no capta la naturaleza dinámica del ecosistema. Aplica términos estáticos para describir un objetivo cuando el sistema es dinámico. La lista de indicadores debe ser concisa, estar justificada e incluir puntos de referencia para orientar la ordenación. El marco debe ser sencillo para que los gestores puedan tratarlo. Componentes que faltan: o Biofísicos o Socioeconómicos El marco debe ser flexible. El marco debe abordar la incertidumbre en la gestión de los objetivos. Debe reconocer que se pueden controlar los impactos humanos no el ecosistema. La participación de expertos adecuados cuando no se cuenta con experiencia es fundamental a la hora de desarrollar el marco. A este efecto, se ha sugerido que una pequeña unidad de trabajo podía reunir ideas para su presentación a la Comisión. Por ejemplo, su ámbito de actuación, problemas, soluciones. La Alianza del mar de los Sargazos expresó su voluntad y disposición para contribuir al proceso de realizar un estudio de caso basado en la gran cantidad de información que ya han recopilado sobre los ecosistemas. Afirmaron claramente que no se espera que el SCRS realice todo el trabajo requerido, sino que proponían una colaboración para realizar un estudio basado en la información presentada en el documento SCRS/2013/132. Después la discusión se centró en el estudio de caso del mar de los Sargazos. En la Tabla 2 se presentan los componentes de este estudio. 2161

146 Parte II Captura fortuita 8 Examinar las entradas utilizadas en la evaluación preliminar del riesgo ecológico (ERA) En el documento SCRS/2013/130 se resume la información de los programas de observadores de Brasil y Uruguay ( ), sí como otros esfuerzos relacionados con los diferentes aspectos considerados en el análisis de productividad y susceptibilidad (PSA) de las tortugas marinas. En el Atlántico suroccidental hay cinco especies de tortugas marinas: tortuga boba (Caretta caretta), tortuga laúd (Dermochelys coriacea), tortuga de carey (Eretmochelys imbricata), tortuga golfina (Lepidochelys olivacea) y tortuga verde (Chelonia myda). Todas ellas anidan en Brasil, en zonas continentales o insulares. En los últimos años, se ha observado un número creciente de nidos y hembras anidadoras para la tortuga boba, la tortuga laúd y la tortuga golfina. Los autores exploraron la información espacial sobre las capturas incidentales de diferentes especies de tortugas marinas, en las zonas cubiertas por el PSA. El esfuerzo de ambas flotas osciló entre el 0,02% y 6,75% del esfuerzo total de ICCAT para la misma zona. Estas zonas (unidades regionales de ordenación -RMU) en algunos casos son más pequeñas que las definidas con las capturas. Además, algunas RMU no consideraban stocks mezclados que han sido identificados por algunos autores. También se resumen la información de telemetría de satélite de las especies que interactúan con el palangre y la información regional existente sobre la selectividad de los artes para la tortuga boba y la tortuga laúd. Esto podría resultar útil para aportar información al componente de susceptibilidad de la ERA. Los autores del documento sugirieron la posibilidad de utilizar la CPUE, o la captura total de tortugas marinas, para sustituir a la RMU, ya que su distribución es más amplia. Además manifestaron su inquietud por el hecho de que la RMU no considere, en algunos casos, la diversidad de las playas de anidación que comprende zonas de anidación para múltiples especies. También se debatió la posibilidad de separar la flota de palangre en componentes superficiales y de profundidad y de incluir los datos de telemetría por satélite, ya que ambos introducirían en la ERA la probabilidad de encuentro de las especies y los artes. Otro aspecto que se debatió fue la posibilidad de utilizar los datos de selectividad existentes, ya que esto mejoraría aspectos de la susceptibilidad en el PSA. 9 Evaluación de Riesgo Ecológico (ERA) 9.1 Presentación de la ERA preliminar En el documento SCRS/2013/134 (Apéndice de la ERA) se presentaba una revisión de la ERA realizada para las tortugas marinas que sufren el impacto de las pesquerías de ICCAT. Las tortugas marinas muestran un alto nivel de filopatría a las playas de anidación. Esto ha dado lugar al desarrollo de poblaciones genéticamente diferenciadas dentro de la mayor parte de las especies, definidas por zonas geográficas amplias. Esto se suele denominar unidades regionales de ordenación (RMU). Se considera apropiado gestionar los procesos de amenaza a nivel de RMU. Sin embargo, muchas especies de tortugas marinas migran o se dispersan ampliamente en el mar. Por tanto hay grandes zonas en las que las RMU de una sola especie se solapan. Asignar una tortuga encontrada en el mar a una RMU en la actualidad es prácticamente imposible, porque no se pueden discernir. Hemos realizado esta evaluación a nivel de RMU, lo que incluye tratar de asignar la información sobre captura fortuita a RMU siempre que sea posible. Hay 22 RMU de seis especies de tortugas en la región ICCAT. La nomenclatura utilizada aquí para las RMU es la siguiente: las iniciales binomiales que identificada a cada especie (por ejemplo: Caretta caretta = Cc, Chelonia mydas = Cm, etc.), seguidas de la región de procedencia (por ejemplo: Atlántico = At, Índico = In), la región geográfica que identifica las amplias zonas de la RMU, por ejemplo, NW, SE. De este modo Cc-AtNW es la población de tortuga boba del Atlántico noroccidental. El objetivo de esta ERA es evaluar el riesgo para las tortugas procedente del impacto de la pesca de túnidos en la región de ICCAT. Es una evaluación de nivel 2 (semi-cuantitativa), realizada con un marco de evaluación de productividad y susceptibilidad (o PSA), a nivel de RMU. Una PSA utiliza los rasgos del ciclo vital de las tortugas (para proporcionar una indicación de la productividad (y, por tanto, de la resistencia (elasticidad) frente al impacto de la pesca), y los parámetros de pesca para establecer la susceptibilidad de los diferentes grupos a la captura por los diferentes tipos de artes de pesca. En este método, se aplica a los parámetros una puntuación baja, media o alta. Los parámetros pueden tener valores altos o bajos, los valores se suman y la puntuación se reescala con respecto a 1-3 para la productividad (P) y la susceptibilidad (S). La vulnerabilidad de cada RMU a cada tipo de arte se calcula posteriormente en función de la siguiente fórmula: 2162

147 La metodología requiere que se califiquen las deficiencias en los datos de conformidad con el enfoque precautorio. Hay muchos parámetros de productividad para los que no se dispone de datos, y a estos se les asigna la puntuación 1 (baja productividad). Para la evaluación de la susceptibilidad, inicialmente se identificaron nueve parámetros. Sin embargo, solo se pudieron abordar dos de un modo adecuado: solapamiento de la RMU con la región de ICCAT y estimación de la mortalidad anual por captura fortuita de las tortugas marinas. Las estimaciones de mortalidad se obtuvieron de las pesquerías de palangre y cerco, pero no se dispuso de datos de las pesquerías de redes de enmalle, y esto no se consideró posteriormente. 9.2 Discusión de la ERA preliminar Para aclarar el proceso de la ERA se explicó que la información que procede de este ejercicio es propiedad del Subcomité y no se publicará en esta fase. Se reconoció que este trabajo está en curso y se prevé que se incluirán muchos colaboradores en las futuras formulaciones de la ERA (y de otros modelos). Sólo cuando esta colaboración haya finalizado y todos los autores/colaboradores estén plenamente incluidos y reconocidos, podrá publicarse este trabajo con el consentimiento de la Secretaría y del SCRS (tal y como se hizo con la ERA de los tiburones). El Subcomité reconoció la importancia de este análisis como un paso preliminar antes de abordar la solicitud de la Comisión de que se evalúe el impacto de las pesquerías de ICCAT en las tortugas marinas. Esta es la primera vez que se realiza una ERA para las tortugas marinas en el Subcomité de ecosistemas. Se reconoció que se trataba de un ejercicio difícil debido a que las tortugas marinas tienen un ciclo vital muy diverso en comparación con los tiburones, las aves marinas y los teleósteos, especies para las que el SCRS había realizado ERA. Las tortugas se ven afectadas por muchos otros factores, lo que incluye las actividades en tierra, además de por las pesquerías de ICCAT. El trabajo inicial del Subcomité de ecosistemas es proporcionar información y orientar las revisiones de la ERA a corto plazo, y más tarde hacerse cargo del trabajo y mejorar el modelo mediante la colaboración y las aportaciones de los expertos. Se inició un debate sobre la presentación de la ERA que puede desglosarse a grandes líneas en los siguientes temas: Calidad de los datos Las solicitudes de datos sobre tortugas marinas de 2012 y 2013 se encontraron con una respuesta limitada, ya que menos de 20 países presentaron información sobre interacciones con tortugas marinas. Se facilitaron todos los datos a los prestatarios de la ERA sobre tortugas marinas, cumpliendo los requisitos de confidencialidad. El Subcomité acordó que con el fin de responder plenamente a la solicitud de la Comisión, las CPC tienen que presentar datos sobre interacciones con tortugas marinas, cuando estén disponibles, ya que la ERA tiene que extraer conclusiones de otros océanos y realizar extrapolaciones basadas en los pocos datos facilitados, lo que podría generar una imagen distorsionada de lo que está ocurriendo realmente y sesgar el asesoramiento facilitado en el modelo. El Subcomité manifestó su inquietud de que zonas resaltadas como zonas de riesgo sean en realidad las zonas para las que se dispone de datos, mientras que las regiones con pocos datos podrían no recibir la atención que requieren. Aunque la ERA ha tratado de tener en cuenta este problema, éste podrá resolverse únicamente mediante la colaboración de expertos que podrían cubrir muchas lagunas en los datos. Durante la reunión, varios científicos presentaron datos actualizados y más completos (tasas de captura, resúmenes de programas de observadores y datos de seguimiento por satélite) que resaltan la importancia de la incorporación de estos datos adicionales en análisis futuros (consúltense las presentaciones que proporcionaron información actualizada). Volvió a analizarse la estimación de la captura fortuita total en número utilizando la información más completa sobre captura fortuita, así como unos procedimientos de estimación revisados. Utilización de las Unidades Regionales de Ordenación (RMU) Se manifestó la inquietud que generaba el hecho de la utilización de las RMU no fuera el modo más apropiado de estructurar el modelo y que sería mejor evaluar el riesgo para regiones clave en las que las flotas interactúan con tortugas marinas. Podrían investigarse otros métodos de estructuración del modelo. Se manifestó una preocupación especial por las RMU que abarcan amplias regiones pero que incluyen un número relativamente bajo de lugares de anidación. El enfoque RMU no siempre tiene en cuenta la disparidad de los tamaños de las poblaciones reproductoras dentro de cada región. 2163

148 Estacionalidad El Subcomité de ecosistemas constató la importancia de las características estacionales del esfuerzo pesquero de palangre y manifestó que esto debía tenerse en cuenta en el análisis. Los datos de seguimiento disgregados por temporada podrían solucionar este tema, junto con los datos sobre interacciones de los observadores que estén disponibles. Un problema potencial sería que los estudios de la CPUE no siempre están relacionados con los datos estacionales. La ausencia de datos dificulta el trabajo. La mayoría de los datos utilizados en el estudio proceden del documento creado en 2012 por Coelho (2012). A menudo faltan datos asociados sobre capturas fortuitas de tortugas marinas. Características de la pesca de palangre El Grupo constató que los calados de palangre superficiales tienen más probabilidades de interactuar con tortugas que los calados profundos. Esto se debe al solapamiento entre la distribución vertical de las tortugas marinas y la profundidad del arte de palangre. Se sugirió que se separasen los calados profundos de palangre de los calados superficiales de palangre para considerar la diferencia de la mortalidad de tortugas marinas en estos dos sectores de la flota. Esta información está disponible en la ERA de tiburones realizada en Datos de seguimiento de tortugas marinas Los prestatarios encargados de la ERA constataron un problema que consistía en que aunque tenían mapas de seguimiento de tortugas marinas, no tenían acceso a los datos reales de estos movimientos y, por tanto, no pudieron incluir gran parte de esta información en la ERA actual. Además, los datos disponibles de seguimiento están relacionados sobre todo con hembras anidadoras y proceden de las playas de anidación, y no de juveniles ni de machos. El Subcomité debatió cuántas marcas serían necesarias para adquirir seguridad en los que concierne a los movimientos de las tortugas marinas y debatieron si la confianza de los datos de seguimiento podría ponderarse. Productividad y mortalidad (natural, por pesca y tras la liberación) de las tortugas marinas Se debatió la idoneidad de las aproximaciones de la mortalidad natural y la mortalidad por pesca. Se planteó la cuestión de que la utilización del número de hembras reproductoras en la aproximación de Z (mortalidad total) podría no ser apropiada, y de que deberían investigarse otras fuentes de mortalidad, lo que incluye las procedentes de la bibliografía disponible y las observaciones directas de M (mortalidad natural) basadas en la fidelidad al lugar de anidación y en el marcado de las hembras en las playas natales realizado durante largos periodos de tiempo. En lo que concierne a las estimaciones de F (mortalidad por pesca), la supervivencia tras la liberación es difícil de determinar para las tortugas. Para todas las estimaciones de mortalidad, se requieren más investigaciones de la bibliografía disponible para obtener estimaciones adicionales de estos parámetros. Por ejemplo, cuando la tripulación de un palangrero extrae el anzuelo y libera a una tortuga marina y el observador consigna este suceso como una liberación de ejemplar vivo cuando, de hecho, la tortuga no sobrevive. El valor de la mortalidad tras la liberación para diferentes tortugas marinas en diferentes flotas tiene que ser objeto de más trabajos de investigación. Inclusión de pesquerías costeras El Subcomité de ecosistemas reconoció la importancia de incluir el impacto de las pesquerías costeras de la zona del Convenio de ICCAT en las tortugas marinas. El Subcomité de ecosistemas reconoció también la necesidad de incorporar diferentes artes y la selectividad (talla, especies, etc.) de dichas pesquerías. Se constató en particular la necesidad de incluir en el análisis las pesquerías de redes de enmalle. Actualmente no hay suficiente información sobre captura fortuita de pesquerías en la zona del Convenio para abordar este sector, aunque establecer colaboraciones adicionales con expertos podría proporcionar información útil para esta evaluación. 9.3 Lista de cambios/mejoras a realizar en la ERA final Tras el debate inicial sobre la evaluación del riesgo ecológico, se reconoció ampliamente que una evaluación del riesgo ecológico podría proporcionar una herramienta al SCRS para evaluar el riesgo relativo de los artes de pesca de ICCAT para las poblaciones de tortugas marinas. Se facilitaron al Subcomité de ecosistemas todos los archivos utilizados para realizar la ERA con el fin de proporcionar una herramienta transparente que el Subcomité pueda utilizar para explorar los datos de entrada, la estructura del modelo y la parametrización de la 2164

149 ERA y formular recomendaciones para su mejora. El Subcomité de ecosistemas reconoció que este enfoque es un primer paso valioso para dar respuesta a la solicitud de la Comisión, aunque hay que seguir trabajando para modificar y mejorar esta evaluación. Se reconoció que esta ERA es el inicio de un proceso que requerirá trabajos de colaboración y mejora para proporcionar asesoramiento de ordenación para las poblaciones de tortugas marinas afectadas por las pesquerías de ICCAT. Para ello, el Subcomité de ecosistemas debatió y redactó recomendaciones para que fuesen evaluadas, así como las revisiones que deberían realizarse a corto (antes de septiembre de 2013), medio ( ) y largo plazo (2015). Se pretende incorporar las recomendaciones a corto plazo en una ERA actualizada producida por los prestatarios que realizan la ERA antes de la reunión del SCRS de septiembre, mientras que las sugerencias a medio y largo plazo se incorporarán en futuras iteraciones de la ERA mediante un esfuerzo de colaboración en el seno del Subcomité de ecosistemas. La ERA, tal y como está estructurada, examina el riesgo por arte (LL-PS). X especies, X RMU. Con el fin de mejorar la ERA el Grupo formuló las siguientes recomendaciones: 1) Corto plazo (2013, antes de las sesiones plenarias del SCRS). a) Expertos de las CPC participantes presentarán comentarios a los autores para su consideración y resumen. Los autores deberán preparar una breve respuesta en septiembre. b) La flota de palangre debería clasificarse además en palangre de profundidad y palangre de superficie. A este efecto, se utilizarán las estimaciones EFFDIS desarrolladas para la ERA de tiburones. c) Estratificar la información sobre la flota y la captura fortuita por temporada. Explorar la pérdida de datos y su efecto en los análisis. Esto demostrará probablemente la necesidad de información estacional adicional. d) Realizar un análisis de lagunas para identificar flotas/lugares en los que faltan datos y en los que se requieren esfuerzos de investigaciones adicionales. Es importante identificar las CPC que comunicaron que no habían realizado ninguna captura fortuita de tortugas marinas de las que no comunicaron ninguna información y no tienen programas de observadores nacionales, ya que, de hecho, podrían haberse encontrado con tortugas marinas pero no disponen de ningún mecanismo de recopilación de datos para comunicarlo. e) Las aproximaciones de la mortalidad (natural e inducida por la pesca) tienen que volver a examinarse. También deberían reclasificarse los niveles de riesgo (a saber, bajo, medio o alto). Por otro lado, deberían explorarse otras aproximaciones de la mortalidad, lo que incluye las estimaciones disponibles en la bibliografía. Se recomendó que se utilicen índices de captura por esfuerzo de las principales flotas palangreras de ICCAT cuando estén disponibles. 2) Medio plazo ( ) a) Identificar importantes datos de entrada (por ejemplo, captura fortuita por unidad de esfuerzo BPUE, datos de captura fortuita por especies, lo que incluye estimaciones de la mortalidad tras la liberación) requeridos para mejorar el análisis y la recopilación de información de la ERA, para su utilización. Si es necesario, los datos podrían agregarse a escala espacial de RMU o de otra unidad espacial determinada mediante consultas con expertos. b) Obtener datos detallados de satélite. Determinar el solapamiento vertical de las pesquerías de palangre y las tortugas marinas. c) Determinar el esfuerzo pesquero de día frente al esfuerzo pesquero de noche. Esta información ya está integrada en la ERA de tiburones y será facilitada por los autores de dicho estudio. d) Utilizar la información disponible sobre tallas de tortugas marinas capturadas por diferentes flotas en el componente de susceptibilidad del análisis de la ERA. e) Incluir información sobre grandes zonas con bajas tasas de captura fortuita de especies de tortugas marinas. f) Incluir información sobre la confianza en las estimaciones de BPUE (por ejemplo, CV, nivel de cobertura de observadores) y determinar la representatividad de la cobertura de observadores con respecto a las operaciones de pesca. g) Recopilar la información adicional sobre seguimiento vía satélite cuando sea posible (por ejemplo, a partir de seaturtle.org). h) Considerar la inclusión del número de playas de anidación y del número/densidad de nidos en dichas playas como un componente de la porción de productividad de la ERA. 2165

150 i) Incluir análisis de los lugares en los que pasan más tiempo las tortugas (por ejemplo, densidad Kernel) utilizando la información del seguimiento vía satélite cuando sea posible. 3) Largo plazo >2015 a) Obtener información sobre programas de observadores costeros y explorar otros modos de estimar el esfuerzo de pesca artesanal y la captura fortuita (por ejemplo, contactar con otros grupos que realizan trabajos similares: ONG, institutos de investigación, etc.). Esto es especialmente importante para las pesquerías de redes de enmalle que tiene un impacto significativo en las poblaciones de tortugas marinas. b) Riesgo (tal y como se considera en la ERA) se define como el producto de la magnitud de las consecuencias adversas del peligro y la probabilidad de que se produzcan. La identificación del peligro, la probabilidad y las consecuencias son todos términos que se caracterizan, describen y miden con diferentes tipos y grados de incertidumbre. Aunque es imposible identificar todas las incertidumbres, la credibilidad, precisión e integridad científicas de una evaluación de riesgo depende de la calidad de su análisis de incertidumbre (Burgman, 2005). Por tanto, en el futuro, deberían desarrollarse métodos apropiados para identificar, describir y, posteriormente, abordar las diferentes formas de incertidumbre. 10. Examinar las medidas y protocolos disponibles para la liberación segura y la mitigación de la captura fortuita de tortugas marinas, y formular recomendaciones cuando se requiera Este tema se evaluó en la reunión de 2013 del Subcomité de Ecosistemas, así como en las reuniones anteriores de 2011 y Los documentos e informes pertinentes debatidos en reuniones anteriores se resumen en la Tabla 13. La reunión del Subcomité de ecosistemas de 2011 también revisó un resumen (Anon de la bibliografía disponible sobre mitigación de la captura fortuita de tortugas marinas. También se consideraron nuevos documentos relacionados con la mitigación de la captura fortuita y los protocolos de liberación segura. En el documento SCRS/2013/129 se presentaba un experimento con observadores llevado a cabo en palangrero de Taipei Chino en el océano Atlántico tropical. El experimento se realizó de septiembre de 2012 a mayo de 2013 entre 2ºS y 12 ºS y 17ºW y 26ºW. Se alternaron secuencialmente anzuelos circulares y tradicionales de atún durante la parte experimental del lance con una ratio de 1:1. Los observadores pesqueros hicieron un seguimiento de 200 lances, en los que se engancharon en los anzuelos 36 tortugas y 19 más se enredaron en la línea madre, en la brazolada o en la línea de la boya. Las pruebas de aleatoriedad mostraron que no había diferencias significativas para las tasas de captura de tortugas marinas por tipo de anzuelo (p=1,000), pero hubo una tasa de captura notablemente más elevada de patudo (p=0,0002), rabil (p= 0,0045), pez espada (p=0,0001), y tintorera (p= 0,0209) en los anzuelos circulares en comparación con los anzuelos tradicionales para el atún. Se observaron tasas de captura significativamente más elevadas para el atún blanco (p=0,0010) y marlín peto (p=0,0097) capturados en anzuelos tradicionales en comparación con los anzuelos circulares. El Subcomité constató que la mortalidad tras la liberación durante este estudio era elevada en relación con las tasas de captura fortuita de la flota de Taipei Chino en su conjunto, pero el autor afirmó que este estudio se había realizado en una zona y periodo en las que las tasas de captura fortuita eran más elevadas que la media. El estudio indicaba que las tasas de captura de anzuelos circulares y anzuelos n forma de J eran iguales, pero que la supervivencia era mayor en anzuelos circulares. El documento SCRS/2013/131 analizaba los datos obtenidos a partir del programa de muestreo del Instituto Español de Oceanografía (IEO) y de los cuadernos de pesca de la flota palangrera española que pescó pez espada en el Mediterráneo en 2011 y Esta flota tiene tres componentes: palangre de superficie, palangre semipelágico y palangre profundo con diferentes tasas de captura fortuita y, por consiguiente, con un impacto diferente en las poblaciones de tortugas marinas. El pez espada es la especie objetivo y responde de más del 80% de la captura total (gama de 82-93%). En 2004 se realizó un estudio piloto con 10 tipos diferentes de anzuelos (en forma de J y circulares), cuatro cebos y diferentes profundidades del palangre. Los resultados de este estudio mostraban que el impacto del palangre en las tortugas marinas depende sobre todo de la profundidad del arte, aunque los efectos zona y temporada también fueron importantes. Los resultados mostraban también que el impacto de esta pesquería en las tortugas marinas puede considerarse bajo a lo largo del año. 2166

151 El Subcomité indicó que las capturas de tortugas son menores para los lances en todas las profundidades y preguntó sobre el tipo de anzuelo utilizado. El autor indicó que en la pesquería se utilizaron anzuelos J #1. Se afirmó también que se habían realizado experimentos con anzuelos circulares, pero que no continuaron debido al descenso en las capturas de pez espada y a la dificultades que se encontraban los pescadores al liberar a las tortugas de dicho arte. El Grupo consideró también el trabajo descrito anteriormente (Sales et al, 2010) sobre la eficacia de los anzuelos circulares en la mitigación de la captura fortuita de tortugas marinas y para capturar especies objetivo en las pesquerías brasileñas de palangre pelágico. Entre 2004 y 2008, se compararon los resultados de los anzuelos circulares alineados 18/0 10º sin anillo con los anzuelos en J sin anillo 9/0 (control) en la pesquería palangrera pelágica brasileña que se dirige al pez espada, a los túnidos y a los tiburones. Durante este experimento se realizaron 26 mareas y 229 operaciones de pesca en las que se calaron anzuelos alternando anzuelos circulares y anzuelos en J, utilizando caballa como cebo. Se capturaron un total de 60 especies diferentes, entre ellas tortugas marinas y aves marinas. Se realizaron análisis estadísticos (pruebas χ² Mantel-Haenszel) de las especies para al menos 20 ejemplares capturados. Con los anzuelos circulares se produjo un importante descenso de la captura de tortuga boba Caretta caretta (55%) y tortuga laúd Dermochelys coriacea (65%) La utilización de anzuelos circulares dio lugar a un incremento en las tasas de capturas de túnidos (patudo - Thunnus obesus y atún blanco -Thunnus alalunga), y de tiburones (tintorera - Prionace glauca y carcarinidos del género Carcharinus). No hubo diferencias en la captura de rabil (Thunnus albacares), marrajo dientuso (Isurus oxyrinchus) y peces martillo (Sphyrna lewini and S. zygaena). Por otro lado, la tasa de captura de pez espada (Xiphias gladius) descendió significativamente con el uso de anzuelos circulares. Además, con la utilización de anzuelos circulares descendieron significativamente las tasas de captura de especies de captura fortuita, como la raya látigo-violeta (Pteroplatytrygon violacea) y la aguja blanca (Tetrapturus albidus). Los anzuelos circulares tuvieron resultados similares a los de los anzuelos en J para muchas especies e incrementaron las capturas de especies comercializables como Los resultados demuestran la eficacia de los anzuelos circulares en la conservación de la tortuga boba y la tortuga laúd, mejorando las tasas de captura de la mayor parte de las especies objetivo, y reduciendo notablemente la captura fortuita de la especie más común, la raya látigo-violeta, por lo que suponen una mejora económica para la pesquería. El grupo preguntó por la causa de las diferencias en las tasas de captura entre los anzuelos circulares y los anzuelos en J observadas en este estudio pero no en el documento SCRS/2013/129. El autor indicó que había una diferencia en la configuración del arte utilizada durante los estudios, los anillos en los anzuelos. Estos se utilizaron en la pesquería de Taipei Chino pero no en Brasil. El Grupo también indicó que aunque los estudios no coincidían en que los anzuelos circulares reducen las tasas de captura de tortugas marinas, si coincidieron en que la utilización de anzuelos circulares reducía la mortalidad por captura fortuita, probablemente porque las tortugas marinas suelen tragarse con más frecuencia los anzuelos en forma de J. También se indicó una segunda diferencia en la configuración del arte; la distancia entre los flotadores y las brazoladas era inferior en la pesquería brasileña de palangre que en la pesquería de Taipei Chino. No está claro si esto influye en las tasas de captura, pero el autor indicó que unas brazoladas más cortas implican un calado más superficial, lo que podría incrementar la supervivencia de las tortugas marinas. El SCRS/2013/128 se presentaban las tasas de capturas incidentales preliminares de tortugas marinas de las flotas de palangre de Taipei Chino en el océano Atlántico. Los datos se recopilaron a partir de calados observados y 46,0 millones de anzuelos. Se capturaron 761 tortugas marinas. La captura incidental más importante de especies de tortugas marinas correspondía a la tortuga laúd (59,9%), a la tortuga golfina (26,9%) y a la tortuga boba (8,1%). La mayoría de las tortugas se enganchó en los anzuelos (58,9%). La mayoría de las tortugas laúd y tortugas golfina se capturó en zonas tropicales, especialmente en el Golfo de Guinea. La tortuga boba se capturó no sólo en aguas tropicales sino también en el océano Atlántico suroccidental. La tasa de captura incidental nominal se situó en un promedio de 0,018 por anzuelos, y osciló entre 0,0012 y 0,0219 por anzuelos por zona. El Subcomité constató que los datos utilizados en los análisis procedían del Atlántico total y sugirieron que se profundizara en los análisis utilizando los componentes de zona y periodo. También se indico que la proporción de lances observados que comunicaron captura fortuita era muy baja, y se sugirió que debería utilizarse una distribución Poisson o binomial negativa para los datos, en vez de una distribución normal. El autor también afirmó que el arte utilizado en zonas tropicales para estos buques es diferente al utilizado en las zonas septentrional y meridional debido a que la especie objetivo es el patudo. El Subcomité preguntó sobre la parte del cuerpo de las tortugas marinas que se engancha en los anzuelos. Aunque el autor indicó que esta información no siempre se registraba, también afirmó que las tortugas laúd se enganchan a menudo en las aletas, ya que no comen este tipo de cebo. 2167

152 11. Preparar la respuesta a la Comisión sobre la Rec (Recomendación de ICCAT sobre captura fortuita de tortugas marinas en las pesquerías de ICCAT) En 2010, la Comisión recomendó que: 1) El SCRS realizara una evaluación del impacto de la captura incidental de tortugas marinas resultante de las pesquerías de ICCAT lo antes posible y a más tardar en 2013 [Rec , párrafo 5]. 2) Tras finalizar la evaluación inicial y presentar sus resultados a la Comisión, el SCRS facilitará asesoramiento a la Comisión sobre la planificación de futuras evaluaciones [Rec , párrafo 5]. 3) El SCRS facilitará también a la Comisión asesoramiento sobre los enfoques para mitigar la captura fortuita de tortugas marinas en las pesquerías de ICCAT, lo que incluye la reducción del número de interacciones y/o de la mortalidad asociada con dichas interacciones [Rec , párrafo 4]. 4) Cuando proceda, la Comisión y sus CPC deberían, de forma individual y colectiva, emprender esfuerzos de creación de capacidad y otras actividades de cooperación para contribuir a la implementación eficaz de esta Recomendación, lo que incluye acuerdos de cooperación con otros organismos internacionales apropiados. En lo que concierne a la mitigación de la captura fortuita de tortugas marinas en las pesquerías de ICCAT, el SCRS recomienda lo siguiente: 1) El SCRS reitera las recomendaciones anteriores de la Comisión [10-09] en lo que concierne a que: a) Los cerqueros que operen en la zona del Convenio eviten cercar tortugas marinas en la medida de lo posible, liberen las tortugas cercadas o enredadas, también en los DCP, cuando sea viable, y comuniquen las interacciones entre el cerco y/o los DCP con las tortugas marinas. b) Los palangreros pelágicos que operen en la zona del Convenio lleven a bordo equipos para liberar, desenredar y manipular de forma segura las tortugas marinas de tal modo que se maximicen sus probabilidades de supervivencia. c) Los pescadores a bordo de los palangreros pelágicos que operan bajo el pabellón de CPC usen los equipos especificados en el punto b anterior para maximizar la probabilidad de supervivencia de la tortuga marina y estén formados en técnicas de manipulación segura y liberación. d) Las CPC deberían incluir en sus informes anuales información sobre otras acciones pertinentes emprendidas para implementar las Directrices de FAO para reducir la mortalidad de tortugas marinas en las operaciones pesqueras en lo que concierne a las pesquerías de ICCAT. 2) Además, para reducir la mortalidad por captura fortuita de tortugas marinas, el SCRS recomienda específicamente lo siguiente: a) En lo que concierne a las prácticas de manipulación segura: i. Cuando se va a sacar a una tortuga del agua, debe utilizarse un montacargas de canasto o un salabardo para izar a bordo las tortugas marinas enganchadas a los anzuelos o enredadas en el aparejo. No debe izarse a bordo desde el agua ninguna tortuga tirando de la liña de pesca a la que se ha enganchado o en la que se ha enredado el cuerpo de la tortuga. ii. Los operadores del buque o la tripulación deben valorar el estado de las tortugas marinas capturadas o enganchadas antes de liberarlas. Lar tortugas marinas que no puedan nadar, que estén inconscientes o no den señales de respuesta deben izarse/mantenerse a bordo y se le debe auxiliar de tal modo que se maximice sus posibilidades de supervivencia antes de liberarla. Estas prácticas se describen en las Directrices de la FAO para reducir la mortalidad de las tortugas marinas en las operaciones pesqueras. iii. Cuando se manipulen tortugas marinas en operaciones de pesca o en el marco de programas de observadores nacionales (actividades de marcado), dicha manipulación debe realizarse de un modo acorde con las Directrices de la FAO para reducir la mortalidad de las tortugas marinas en las operaciones pesqueras b) En lo que concierne a la utilización de corta líneas: i. Los palangreros tienen que llevar a bordo corta líneas y deben utilizarlos cuando no se puede extraer el anzuelo de forma segura para liberar a las tortugas marinas. ii. Otros tipos de buques que utilizan artes en los que puedan engancharse tortugas marinas deben llevar a bordo corta líneas y utilizar estas herramientas para extraer el arte de forma segura y liberar a las tortugas marinas. 2168

153 c) En lo que concierne a la utilización de dispositivos desanzueladores: i. Los palangreros deben llevar a bordo dispositivos desanzueladores para extraer de forma eficaz y segura los anzuelos de las tortugas marinas. El Subcomité recomienda también que en los casos en los que la tortuga marina se haya tragado el anzuelo, no se intente extraerlo. En este caso debe intentarse cortar la línea lo más cerca posible del anzuelo. En lo que concierne a la evaluación del impacto de las pesquerías en las tortugas marinas, el SCRS inició la evaluación del riesgo ecológico (ERA) para las tortugas marinas en Los progresos hasta la fecha son los siguientes: 1) En 2013, ICCAT proporcionó un contrato a corto plazo para respaldar una ERA preliminar para las especies de tortugas marinas que se encuentran en las pesquerías de ICCAT. La ERA utilizó los datos proporcionados a la Secretaría por las CPC en 2011 y 2012, y tal y como fueron recopilados en el marco de un contrato a corto plazo financiado por ICCAT en 2012, y otras fuentes de datos compiladas por el prestatario. 2) En su reunión intersesiones de 2013, el Subcomité de ecosistemas revisó los progresos de la ERA hasta la fecha y formuló importantes recomendaciones para mejorar la evaluación a corto (antes de octubre de 2013), medio ( ) y largo plazo (2015 +), lo que incluía una solicitud de datos adicionales/actualizados de las CPC. 3) El SCRS continuará mejorando la ERA y asesorará a la Comisión sobre su plan para futuros análisis del impacto de las tortugas marinas en la reunión de Otros asuntos 12.1 Presentaciones Se proporcionó una presentación sobre capturas fortuitas de tortugas marinas realizadas por la flota marroquí. Las entrevistas con pescadores de atuneros marroquíes realizadas desde 2008 hasta 2011 revelaron que se encontraba una tortuga Los días de pesca por marea oscilan entre 1 a 3 días, pero dadas las características técnicas de los buques, suelen durar menos de 24 h. A partir de una muestra de 100 barcos especializados en la pesca de túnidos al sur de Agadir, dentro de la ZEE marroquí, se determinó que cinco de cada seis buques que pescan observaron una tortuga. La tasa de captura de tortugas marinas de esta pesquería es una tortuga cada 20 días de pesca. En esta zona, las tortugas marinas capturadas se liberan inmediatamente. Los resultados del programa nacional de observadores de capturas de tiburones a bordo de los palangreros indicaban que no había presencia de tortugas o que sus interacciones con esta pesquería son insignificantes. Marruecos propone lanzar una campaña de sensibilización dirigida a diferentes operadores en la región para la protección de las tortugas marinas. En el documento SCRS/2013/133 se presentaban los resultados de la Iniciativa Trasatlántica para la Tortuga Laúd (TALCIN). La segunda fase de estos trabajos proporcionó el primer análisis a escala oceánica de la distribución espacio-temporal de la tortuga laúd, basado en marcado electrónico, y también determinó su solapamiento con el esfuerzo pesquero del palangre. Los datos sugieren que existen dos unidades regionales de ordenación en el Atlántico: septentrional y meridional (la segunda incluye tortugas de Sudáfrica). Aunque las tortugas marinas y las pesquerías muestran distribuciones muy diferentes, los autores señalaron nueve zonas en las que existe una gran susceptibilidad de captura fortuita (cuatro en el Atlántico norte y cinco en el Atlántico sur/ecuatorial), en las que se deben centrar más las investigaciones. Esto se ve reforzado por los informes de captura de tortuga laúd en ocho de estos lugares (documento SCRS/2013/133). El Subcomité constató que se realizó un análisis del solapamiento de zonas de alta presión pesquera con el hábitat de la tortuga laúd por trimestre, y se sugirió que la evaluación del riesgo ecológico para las tortugas marinas podría utilizar un enfoque similar. El Subcomité también indicó que la utilización del hábitat por parte de la tortuga laúd podría solaparse con zonas de elevada y baja producción pesquera. En este sentido, se sugirió que esta información debería compararse con los datos de observadores para determinar si existe alguna correlación. El Subcomité recomendó que se realicen comparaciones adicionales de las zonas muy utilizadas por la tortuga laúd con las operaciones de palangre en aguas superficiales y en aguas profundas, utilizando análisis similares a los desarrollados en la evaluación del riesgo ecológico de los tiburones. 2169

154 En el documento SCRS/2013/135 (avistamientos y abundancia de tortugas marinas en Azores) presentaba un análisis de los avistamientos y esfuerzos de observación de tortugas marinas realizados por observadores a bordo de la flota atunera de cebo vivo de Azores. Se observó un total de mareas desde 2001 hasta 2012, y se llevaron a cabo prospecciones con una duración media de 15 minutos. También se presentaban los avistamientos ocasionales de tortugas marinas realizados entre 1998 y A partir de los datos de avistamientos y esfuerzo de observación de tortuga boba en la zona de Azores de 2001 a 2012, se estimó un índice de abundancia relativa mediante un enfoque de modelo lineal generalizado (GLM). El análisis muestra que tras los escasos avistamientos por unidad de esfuerzo en 2001, el índice de abundancia alcanzó su nivel más elevado en 2002, y descendió en los años siguientes, manteniéndose estable desde 2008 hasta la actualidad. En el documento también se presentaba información sobre observaciones de tortuga laúd y la tortuga verde en la zona de Azores. La información recopilada en la zona de Madeira, para algunos años y meses, incluía el número de prospecciones y de tortugas bobas avistadas. También se incluyen los índices nominales (para todas las especies y zonas) y las SPUE estandarizadas (avistamientos por unidad de esfuerzo) para las tortugas bobas en Azores. Este es el único índice independiente de la pesquería en alta mar. El Subcomité indicó que la capacidad de avistar tortugas marinas puede verse afectada por factores tales como el estado del mar, las nubes y el ángulo desde el cuál los observadores realizan la búsqueda, y que dichos factores deberían incorporarse en el análisis. Se indicó que ha habido un incremento documentado en el número de tortugas en las playas de anidación de Estados Unidos, y dado que se sabe que estas tortugas marinas pueden migrar a Azores, el Comité consideró que sería interesante observar si se incrementan los avistamientos en Azores en años futuros. El ponente compartió información adicional sobre los resultados de los análisis genéticos realizados, que confirman que el 100% de las tortugas marinas halladas en torno a las Azores proceden de Estados Unidos y México. Se constató que esta información no correspondía a la zona RMU CcATLNE, según Wallace et al El Subcomité recomendó que un análisis adicional útil sería determinar la tasa de crecimiento de las tortugas utilizando los datos de recuperación de marcas. En el documento SCRS/2013/138 se describe un enfoque para la investigación en régimen de colaboración en la creación de capacidad en ciencias pesqueras. El autor estableció una relación de los beneficios de la colaboración en la investigación pesquera: 1) incrementar la credibilidad y legitimidad de los hallazgos científicos en las mentes de las partes interesadas, con el potencial de incrementar la aceptación de las acciones de ordenación; 2) adopción de prácticas y aparejos de pesca innovadores que reducen la captura fortuita y minimizan el impacto de la pesca en el hábitat; 3) una mayor comprensión y confianza entre los socios y 4) oportunidades de integrar diferentes fuentes de conocimientos sobre el medio ambiente marino y costero. El autor afirmó que el enfoque de colaboración en la investigación está siendo adoptado cada vez más por el gobierno, la industria y las asociaciones y organizaciones de conservación, como Nature Conservancy y el programa "Smart Gear" de WWF. El autor describió un nuevo programa de colaboración de becas investigación pesquera que se inició en el Instituto de Ciencia Marina de Virginia. El autor indicó que esta investigación se podría beneficiar de la colaboración con un experto en pesquerías de túnidos y captura fortuita, e invitó al Subcomité a que considere colaborar a través de este programa. El Subcomité consideró que esta propuesta constituía un enfoque interesante y una buena oportunidad para conseguir apoyo externo y capacidades de investigación en colaboración con los científicos, la industria, los gestores y las ONG. El Subcomité debatió que los miembros de VIMS deberían ponerse en contacto con el jefe de delegación de la CPC apropiada para participar en reuniones futuras del SCRS Progresos en la armonización de los formularios de recopilación de datos para los programas de observadores de palangre Durante su reunión de 2012, el Subcomité de ecosistemas recomendó que ICCAT trabaje en coordinación con el Grupo de trabajo técnico conjunto sobre captura fortuita de las OROP de túnidos (BJTWG) con el fin de liderar el desarrollo de normas mínimas para la recopilación armonizada de datos de observadores del palangre para las OROP de túnidos. Desde entonces, la Secretaría se ha puesto en contacto con el Sr. Simon Nicol, presidente de dicho Grupo, y se ha iniciado un proceso para cotejar los formularios sobre actividades de recopilación de datos de programas de observadores de palangre de ICCAT, WCPFC, IATTC, IOTC y CCSBT. Dado que ICCAT no coordina directamente programas de observadores de gran escala, es necesario ponerse en contacto y cooperar con los programas nacionales de observadores. En este sentido, la Secretaría se ha puesto en contacto con los coordinadores de los programas nacionales de observadores de palangre que operan en la zona del Convenio de ICCAT con el fin de obtener sus formularios de recopilación de datos. Estos se utilizarán para identificar y recomendar normas mínimas de recopilación de datos. 2170

155 13. Recomendaciones El Subcomité reconoció el excelente trabajo realizado por los Drs. Andrea Angel, Ross Wanless y Ronel Nel en la compilación de una evaluación del riesgo ecológico preliminar (ERA) para las tortugas marinas. Su trabajo preparatorio ha agilizado el trabajo del Subcomité y ha proporcionado una excelente base para la evaluación en curso del impacto de la captura fortuita de tortugas marinas en las pesquerías de ICCAT. El Subcomité reconoció el valor de la iniciativa de ICCAT de proporcionar apoyo financiero para contratar expertos que contribuyan a los trabajos del SCRS, y recomendó encarecidamente que continúen estas actividades productivas. Ecosistemas: El Subcomité reconoció el valor del ejercicio de deconstrucción para definir los objetivos ecosistémicos del SCRS. Se recomendó que el co-coordinador del Subcomité de ecosistemas desarrolle un estudio para elaborar una lista de objetivos EBFM conceptuales que se distribuirá entre los cargos del SCRS. El co-coordinador compilará los resultados antes de las sesiones plenarias del SCRS Se recomienda que se asignen fondos para viaje para respaldar la participación de expertos externos que contribuyan a desarrollar las herramientas científicas requeridas para implementar enfoques EBFM. Se recomienda que el Subcomité siga colaborando con la Alianza del mar de los Sargazos en los análisis de la importancia ecológica del mar de los Sargazos para los túnidos y especies afines y especies ecológicamente asociadas. Captura fortuita El Subcomité recomienda que el SCRS contribuya a los enfoques de investigación pesquera en régimen de colaboración mediante la participación de las partes interesadas en iniciativas que aborden las potenciales interacciones de las pesquerías de ICCAT con especies de captura fortuita. ERA para las tortugas marinas Las CPC deberían facilitar datos de captura fortuita siguiendo las normas de Tarea II. Si es posible, el Subcomité recomienda a las CPC que los datos sobre captura fortuita de tortugas marinas se proporcionen desglosados por especies, con información espacial y temporal (por ejemplo 5 x 5) lo que permitiría la asignación a la RMU y al trimestre. El Subcomité reconoce la necesidad de incluir información sobre pesquerías artesanales que operan dentro de la zona del Convenio ICCAT e insta a las CPC a presentar la información pertinente. El Subcomité recomienda que se lleve a cabo un marcado suplementario (con marcas convencionales y electrónicas) de tortugas marinas y que la información de estos experimentos se ponga a disposición del Subcomité. El Subcomité recomienda que se realicen estudios genéticos sobre tortugas marinas para entender mejor que poblaciones se ven afectadas por las pesquerías de ICCAT, y que la información sobre estos experimentos se ponga a disposición del Comité. Mitigación de la captura fortuita de tortugas marinas El Subcomité reitera a la Comisión las recomendaciones adoptadas mediante la Rec Además, para reducir la mortalidad por captura fortuita de tortugas marinas, el SCRS recomienda específicamente lo siguiente: o En lo que concierne a las prácticas de manipulación segura: Cuando se va a sacar a una tortuga del agua, debe utilizarse un montacargas de canasto o un salabardo para izar a bordo las tortugas marinas enganchadas a los anzuelos o enredadas en el aparejo. No debe izarse a bordo desde el agua ninguna tortuga tirando de la liña de pesca a la que se ha enganchado o en la que se ha enredado el cuerpo de la tortuga. 2171

156 Los operadores del buque o la tripulación deben valorar el estado de las tortugas marinas capturadas o enganchadas antes de liberarlas. Lar tortugas marinas que no puedan nadar, que estén inconscientes o no den señales de respuesta deben izarse/mantenerse a bordo y debe prestársele asistencia de tal modo que se maximice sus posibilidades de supervivencia antes de liberarla. Estas prácticas se describen en las Directrices de la FAO para reducir la mortalidad de las tortugas marinas en las operaciones pesqueras Cuando se manipulen tortugas marinas en operaciones de pesca o en el marco de programas de observadores nacionales (actividades de marcado), dicha manipulación debe realizarse de un modo acorde con las Directrices de la FAO para reducir la mortalidad de las tortugas marinas en las operaciones pesqueras o En lo que concierne a la utilización de corta líneas: Los palangreros tienen que llevar a bordo corta líneas y deben utilizarlos cuando no se puede extraer el anzuelo de forma segura para liberar a las tortugas marinas. Otros tipos de buques que utilizan artes en los que puedan engancharse tortugas marinas deben llevar a bordo corta líneas y utilizar estas herramientas para extraer el arte de forma segura y liberar a las tortugas marinas. o En lo que concierne a la utilización de dispositivos desanzueladores: Los palangreros deben llevar a bordo dispositivos desanzueladores para extraer de forma eficaz y segura los anzuelos de las tortugas marinas. El Subcomité recomienda también que en los casos en los que la tortuga marina se haya tragado el anzuelo, no se intente extraerlo. En este caso debe intentarse cortar la línea lo más cerca posible del anzuelo. 14. Adopción del informe y clausura El informe fue adoptado durante la reunión así como el plan de trabajo del Subcomité para El coordinaror dio las gracias a la Secretaría y a los participantes por el gran esfuerzo realizado. Los co-coordinadores del Subcomité también expresaron su agradecimiento a Andrea Angel, Ross Wanless y Ronel Nel por ayudar al Grupo con importantes progresos en la evaluación del riesgo ecológico de tortugas marinas que interactúan con las pesquerías de ICCAT, y también a la Secretaría por financiar este trabajo de vital importancia. La reunión fue clausurada. Tras la reunión del Subcomité de ecosistemas, la Convención Interamericana para la Protección de tortugas marinas expresó su disposición para contribuir a la mejora de la ERA para las tortugas marinas proporcionando información sobre el "número de playas de anidación" y el número/densidad de nidos en dichas playas como un componente de la porción de productividad de la ERA. Esta información está disponible actualmente para los países CIT desde 2005 hasta 2013 y puede consultarse en los informes de la CIT en la página web de la IAC: Referencias Burgman MA, Risks and decisions for conservation and environmental management. Cambridge University Press, Cambridge, UK,

157 TABLEAUX Tableau 1. Évaluation préliminaire de l'importance écologique de la mer des Sargasses pour les thonidés et les espèces de thonidés apparentées ainsi que les espèces associées du point de vue écologique. Reproduction : 1) preuve directe de reproduction dans la région et 2) zone de frai présumée. Migration : déplacement vers ou à travers la région lors des migrations ou des mouvements annuels. Régime alimentaire : activités de recherche de nourriture dans la région. Cycle vital : utilisation de la région, une ou plusieurs fois, au cours de son cycle vital. Capture : Mer des Sargasses et océan Atlantique. Année de référence : AP = analyse en cours. NF = aucune pêcherie Tableau 2. Éléments de l'écosystème de la mer des Sargasses. Tableau 3. Documents SCRS concernant l'atténuation des prises accessoires et les protocoles de remise à l eau en toute sécurité présentés aux réunions du Sous-comité des écosystèmes de 2011 et TABLAS Tabla 1. Evaluación preliminar de la importancia ecológica del mar de los Sargazos para los túnidos y especies afines, así como para especies asociadas desde el punto de vista ecológico. Reproducción: 1) prueba evidente de reproducción en la zona, 2) Presunta reproducción en la zona. Migración: Desplazamientos dentro o a través de la zona durante las migraciones o movimientos anuales. Alimentación: Utilización de las zonas para forraje. Ciclo vital: Utilización de la zona en una o más fases del ciclo vital. Captura: Mar de los Sargazos y Atlántico, año de referencia 2009 AP = Análisis pendiente; NF = no pesquería. Tabla 2. Elementos del ecosistema del mar de los Sargazos. Tabla 3. Documentos del SCRS relacionados con la mitigación de la captura fortuita y los protocolos de liberación segura presentados a las reuniones del Subcomité de ecosistemas de 2011 y APPENDICES Appendice 1. Ordre du jour. Appendice 2. Liste des participants. Appendice 3. Liste des documents. Appendice 4. Exercice de dégroupage. APÉNDICES Apéndice 1. Orden Del Día Apéndice 2. Lista de participantes. Apéndice 3. Lista de documentos. Apéndice 4. Ejercicio de deconstrucción. 2173

158 Table 1. Preliminary assessment of the ecological importance of the Sargasso Sea to tunas and tuna-like species and to ecologically associated species. Reproduction: 1) Direct evidence of spawning in area 2) Presumed spawning in area. Migration: Moves into or through area during annual movements or migrations. Feeding: Uses area for foraging. Life History Cycle: Uses area in one or more phases of its life history cycle. Catch: Sargasso Sea and Atlantic, reference year AP = analysis pending; NF = no fishery. Species Direct Reproduction Presumed Migration Feeding Life history cycle Sargasso Sea Catch t (2009) Atlantic Catch (t) by stock (2009) W. Bluefin No Unknown Yes Yes Yes Yellowfin No Yes Yes Yes Yes Albacore No Yes Yes Yes Yes Bigeye No Yes Yes Yes Yes Skipjack No Probably Probably Probably Probably Swordfish No Yes Yes Yes Yes Blue marlin Yes N/A Yes Yes Yes White marlin No Yes Yes Yes Yes Sailfish No Probably Probably Probably Probably Wahoo Yes N/A Yes Yes Yes AP Blackfin tuna No Yes Yes Yes Yes AP Atl. black skipjack No Yes Yes Yes Yes AP tuna Dolphinfish No Yes Yes Yes Yes AP Shortfin mako No Yes Yes Yes Yes AP Blue shark No Yes Yes Yes Yes AP Porbeagle shark No Unknown Yes Yes Yes AP Bigeye thresher shark No Unknown Probably Probably Probably AP Basking shark No Unknown Yes Probably Unknown NF % of total Atlantic Catch 2174

159 Table 2. Elements of the Sargasso Sea Ecosystem. 1. Abiotic elements 1.1 Oceanographic Gyre/s Eddies Currents (1) Gulf Stream (GS North Wall) Temperature (1) Sea Surface (a) Atlantic Warm pool or WHWP (b) Atlantic Multi decadal oscillation (2) Water Column ph Dissolved Oxygen (1) Horizontal (2) Vertical Salinity Turbidity 1.2 Climatic Wind velocity ENSO (El Nino Southern Oscillation) NAO Hurricane frequency 1.3 Geomorphology (1) Eg. Seamounts and ridges causing upwelling etc. 2. Biotic elements 2.1 Fishery species Retained species (1) Target species (a) (i) Species i Stock j 1. Abundance 2. Distribution 3. All removals (including discards) 4. IUU (catch by illegal, unreported and unregulated fisheries) (ii) Stock k (iii) Stock l (2) Non target species (a) Species m (b) Species n (3) Bait collection Non retained species (Those species caught or directly impacted by the fishery but not used) (1) Protected or special species (a) Mammals (b) Seabirds (c) Turtles (d) Some sharks (2) General discarded species (a) Species o (b) Species p 2.2 Other biota Macro algae (1) Eg. Sargassum mats Phytoplankton Zooplankton Ichthyoplankton Other fish species (forage species) 2175

160 2.3 Habitat Pelagic (1) Sargassum mats Benthic 2.4 Biodiversity and Interactions Species richness (1) Species abundance (i.e. the abundance of each species) Impacts on environment (1) Ecosystem structure (a) Ghost fishing (b) Discarding/Provisioning (c) Habitat (FADs) (2) General environment (a) Waste disposal (b) Water quality 3. Human elements 3.1 Impacts on environment 3.2 Community Well Being Cultural Heritage Industry (1) Tuna fishing (a) Income (b) Capacity (c) Employment (2) Tuna processing (a) Income (b) Employment General Community (1) Food (2) Infrastructure (3) Fees 3.3 Administration 3.4 Management Research Institutions NGOs RFMOs Consultation 4. Economic elements 5. Elements not necessarily applicable to the Sargasso Sea Rainfall amount 2176

161 Table 3. SCRS documents pertaining to by-catch mitigation and safe-release protocols presented to the 2011 and 2012 Sub-Committee on Ecosystems meetings. SCRS/2011/057 Movements of Juvenile Loggerhead Turtles in the Southwestern Atlantic Ocean. C. Barceló, A. Domingo, P. Miller, L. Ortega, B. Giffoni, G.Sales, L.McNaughton, M. Marcovaldi, S.S. Heppell and Y. Swimmer. SCRS/2011/058 Setting deeper, catching fewer? Sea turtle by-catch on deep set pelagic longlines in Uruguayan Waters. P. Miller, M. Pons, A. Domingo SCRS/2011/059 J hooks vs. C hooks in the Uruguayan monofilament longline fishery. A. Domingo, M. Pons, P. Miller, C. Barceló, S. Jiménez and Y.Swimmer. SCRS/2011/060 Spatio-temporal correlation between leatherback turtles and industrial fisheries in the Atlantic Ocean. Fossette S, Coyne MS, Augowet E, Broderick AC, Chacon D, Domingo A, Eckert SA, vans D, Felix ML, Formia A, Godley BJ, Hays GC, Kelle L, López- Mendilaharsu M, Luschi P, Miller P, Nalovic MA, Nougessono S, NSafou M, Parnell RJ, Prosdocimi L, Sounguet GP, Turny A, Verhage B, Witt MJ, Georges J-Y. SCRS/2011/067 Testing new designs of fish drifting fish aggregating device (DFAD) in Eastern Atlantic to reduce turtle and shark mortality. Franco, J., Moreno, G., Lopez. J., and Sancristobal, I. SCRS/2011/068 Review of actions by Brazil in meeting the Rec and FAO Guidelines to reduce sea turtle mortality in fishing operations. N. de Oliveira L. Júnior, B. Giffoni, F. Niemeyer Fiedler and G. Sales SCRS/2012/051 A review of sea turtle mitigation measures across the five trfmos and other fisheries management organizations. Coelho, R., Fernández-Carvalho, J. and N. Santos, M. SCRS/2012/085 Groupe Tortues Marines France- bycatch group activities in French waters. Claro, F., Sacchi,J. and Poisson, F. SCRS/2012/089 Effects of hook and bait on turtles bycatch in a southern Atlantic pelagic longline fishery. Neves Santos, M., Coelho, R. and Fernández Carvalho, J. SCRS/2012/090 Preliminary results on the use of 17/0 circle hooks and mackerel on sea turtles bycatch in a north-eastern tropical Atlantic swordfish longline fishery. Neves Santos, M., Coelho, R. and Amorim, S.J. 2177

162 Appendix 1 AGENDA 1. Opening, adoption of Agenda and meeting arrangements Part I: Ecosystem 2. Review of new information on the implementation of ecosystem effects in enhanced stock assessments. 3. Review of new information on models that quantify ecosystem dynamics and predict the impact of perturbations on its components. 4. Review of new information on the use of indicators to quantitatively or qualitatively reflect the health of an ecosystem and its fishery, ecological, economic and social components. 5. Review new information on the implementation of ecosystem based fishery management principles. 6. Review progress on the development of a test case for implementing ecosystem based fishery management 7. Discuss ways of including ecosystem values in the standardization and assessment of ICCAT assessed stocks. Part II: By-catch 8. Review of the inputs used for the preliminary Ecological Risk Assessment (ERA) 8.1 Productivity 8.2 Susceptibility 9. Ecological Risk Assessment (ERA) 9.1 Presentation of preliminary ERA 9.2 Discussion of preliminary ERA 9.3 List of changes/improvements to be made to final ERA 10. Review sea turtle available bycatch mitigation and safe-release protocols measures, and make recommendations as necessary. 11. Prepare response to the Commission regarding Rec (Recommendation by ICCAT on the By-Catch of Sea Turtles in ICCAT Fisheries). 12. Other matters 13. Recommendations 14. Adoption of the report and closure 2178

163 Appendix 2 LIST OF PARTICIPANTS SCRS CHAIRMAN Santiago Burrutxaga, Josu Head of Tuna Research Area, AZTI-Tecnalia, Txatxarramendi z/g, Sukarrieta (Bizkaia), Spain Tel: (Ext. 497); , Fax: , CONTRACTING PARTIES ALGERIE Bouhadja, Mohamed Amene Centre National de Recherche et de Développement de la Pêche et de l'aquaculture (CNRDPA) Tel: , Fax: BRAZIL De Barros Giffoni, Bruno Rua Anotnio Athanazio, 273, Jardim Paula Nobre, Ubatuba, SP Tel: , Fax: , De OliveiraLeite, Nilamon Centro Nacional de Conservaçao e Manejo das Tartarugas Marinhas, Av. Paulino Muller, 1111, Jucutu Quara, Vitória/ES Tel: , CANADA Hanke, Alex Scientific, St. Andrews Biological Station/ Biological Station, Fisheries and Oceans Canada531 Brandy Cove Road, St. Andrews New Brunswick E5B 2L9 Tel: , Fax: , EUROPEAN UNION Carroceda Carballal, Aránzazu Ministerio de Economía y Competitividad, Instituto Español de Oceanografía - C.O. de A Coruña, Paseo Marítimo, Alcalde Francisco Vázquez, 10 - P.O. Box 130, A Coruña, Spain Tel: , Fax: , De la Serna Ernst, José Miguel Ministerio de Economía y Competitividad, Instituto Español de Oceanografía, C.O. de Málaga, Puerto Pesquero s/n, Fuengirola Málaga, Spain Tel: , Fax: , Fernández Costa, José Ramón Ministerio de Economía y Competitividad, Instituto Español de Oceanografía - C. Costero de A Coruña, Paseo Marítimo Alcalde Francisco Vázquez, 10 - P.O. Box 130, A Coruña, Spain Tel: , Fax: , Ortiz de Urbina, José María Ministerio de Economía y Competitividad, Instituto Español de Oceanografía, C.O de Málaga, Puerto Pesquero s/n, Fuengirola Málaga, Spain Tel: , Fax: , Pereira, Joao Gil Universidade dos Açores, Departamento de Oceanografia e Pescas9900 Horta, Portugal Tel: , Fax: , JAPAN Inoue, Yukiko Assistant Researcher, Ecologically Related Species Group, Tuna and Skipjack Resources Division, National Research Institute of Far Seas Fisheries5-7-1 Orido, Shimuzu-Ku, Shizuoka-City, Shizuoka Tel: , Fax: , Minami, Hiroshi Chef of Ecologically Related Species Group, Tuna and Skipjack Resources Division,, National Research Institute of Far Seas Fisheries, Orido, Shimizu-ku, Shizuoka-City, Shizuoka Tel: , Fax: ,

164 Ochi, Daisuke Ecologically Related Species Section, Tropical Tuna Resources Division, National Research Institute of Far Seas Fisheries, Orido, Shimuzu-Ku, Shizuoka Tel: , Fax: , MOROCCO El Marhoume, Ingénieur à la Division de la Protection des Ressources Halieutiques, Service de l'application de la Réglementation et de la Police Administrative, Département de la Pêche Maritime, BP 476, Agdal, Rabat Tel: , Fax: , NAMIBIA Iitembu, Johannes Angala Ministry of Fishery and Marine Resources, National Marine Research and information Centre, Strand Street, Swakopmund Tel: , Fax: , Uanivi, Uatjavi Minstry of Fishereis and Marine Resources, Directorete Resource Management, Strand Street, Swakopmund Tel: , Fax: , SOUTH AFRICA Angel, Andrea Private consultant, 65 Putney Rd., Kenilworth, Cape Town Tel: , UNITED KINGDOM (OVERSEAS TERRITORIES) Luckhurst, Brian Consultant, 2-4 Via della Chiesa, Acqualoreto (TR) Umbria, Italy Tel: , Trott, Tammy M. Senior Marine Resources Officer, Department of Environmental Protection, #3 Coney Island Road, CR04 St. George's, Bermuda Tel: , Fax: , UNITED STATES Cass-Calay, Shannon NOAA Fisheries, Southeast Fisheries Center, Sustainable Fisheries Division, 75 Virginia Beach Drive, Miami, Florida Tel: , Fax: , Nalovic, Michel A. Collaborative Research in Fisheries Science Fellow, Virginia Institute of Marine Science1208 Greste Rd., Gloucester Point, Virginia Tel: , Fax: URUGUAY Domingo, Andrés Dirección Nacional de Recursos Acuáticos - DINARA, Laboratorio de Recursos Pelágicos, Constituyente 1497, Montevideo Tel: , Fax: , OBSERVERS FROM COOPERATING NON-CONTRACTING PARTIES, ENTITIES, FISHING ENTITIES CHINESE TAIPEI Huang, Julia Hsiang-Wen Associate Professor, Institute of Marine Affairs and Resources Management, National Taiwan Ocean University, 2 Pei-Ning Road, Keelung Tel: , Fax: , ICCAT SECRETARIAT c/ Corazón de María 8 6 y 7 Planta, Madrid, Spain Tel: ; Fax: ; Pallarés, Pilar, Ortiz, Mauricio, Kell, Laurence, De Bruyn, Paul 2180

165 Appendix 3 SCRS/2013/128 SCRS/2013/129 LIST OF DOCUMENTS Understanding incidental catch of sea turtle of Taiwanese longline fleets in the Atlantic Ocean. Huang H. Circle hook effectiveness for catch of target species and incidental catch of sea turtles on a Taiwanese longline fishing vessel in the tropical Atlantic Ocean. Huang H., Swimme Y.r, Bigelow K.,Gutierrez A. and Foster D. SCRS/2013/130 Incidental captures of sea turtles by Brazilian and Uruguayan longline fishery ( ). Giffoni B., Leite N., Miller P., Pons M., Sales G. and Domingo A. SCRS/2013/131 SCRS/2013/132 SCRS/2013/133 SCRS/2013/134 Estimación de las interacciones de las pesquerías de grandes pelágicos en el Mediterráneo y Estrecho de Gibraltar con mamíferos, tortugas y aves marinas.proyecto de mitigación de estas interacciones. de la Serna J.M.; Macias D. and Ortiz J.M. Inventory and Ecology of Fish Species of Interest to ICCAT in the Sargasso Sea. Luckhurst B.E. Pan-oceanic analysis of the overlap of a large migrant with pelagic longline fisheries. Fossette S., Witt M.J., Miller P., Nalovic M.A., Albareda D., Almeida A.P., Broderick A.C., Chacón-Chaverri D., Coyne M.S., Domingo A., Eckert S., Evans D., Fallabrino A., Ferraroli S., Formia A., Giffoni B., Hays G.C., Hughes G., Kelle L., Leslie A., López- Mendilaharsu M., Luschi P., Prosdocimi L., Rodriguez-Heredia S., Turny A., Verhage B. and Godley B.J. Draft Ecological Risk Assessment of sea turtles to tuna fishing in the ICCAT región. Angel A., Nel R., Wanless R.M., Mellet B., Harris L. and Wilson I. SCRS/2013/135 Sightings and abundance of marine turtles in Azores. Pereira J. G. SCRS/2013/137 Tracking EAF/EBF implementation with Ecological Risk Assessment (ERA) : The Namibian experience. Iitembu J.A. and Uanivi U. SCRS/2013/138 TALCIN / VASG CFR Fellowship. Hartley T. 2181

166 Appendix 4 UNPACKING EXERCISE The results of an unpacking exercise conducted by the Sub-Committee on Ecosystems are shown in the table below. The table demonstrates how high order conceptual management objectives are translated into low order operational objectives. These outcomes are intended to form the basis of an ecosystem based fisheries management and reporting framework. Level Conceptual Objectives 1 Conserve Ecosystem Components 2 Conserve Community Diversity 3 Conserve Benthic Communities further unpacking required 3 Conserve Pelagic Community further unpacking required 2 Conserve Species Diversity 3 Maintain continued existence of all species within the management area 4 Manage exploitation of target species 5 Minimize ghost fishing 6 further definition required 5 Maintain ICCAT populations at MSY 6 Maintain species x above conservation limits 4 Minimize incidental catches of non-target species 5 Limit bycatch of species x managed by rfmo 5 eg Limit catch of porbeagle in NW Atlantic 5 Limit bycatch of species y not managed by rfmo 5 eg Limit catch of porbeagle in South Atlantic 5 eg Limit catch of forage species 4 Protect species at risk 5 Minimize bycatch of turtles 5 Minimize bycatch of seabirds 5 Minimize bycatch of marine mammals 5 Minimize bycatch of sharks 2 Maintain Population Integrity 3 Maintain continued existence of all populations within management area 4 Maintain spawning components 4 Ensure no spawning component is eliminated by fishing 5 Distribute fishing over spawning component 1 Conserve Components Role 2 Maintain Primary Productivity 3 Not relevant to fishing activity 2 Maintain Trophic Structure 3 Further unpacking required 2 Maintain Productivity of Populations 3 Manage exploitation of target species 4 Control overall exploitation rate 4 Ensure appropriate size selectivity of fishery 4 Prevent disturbance of fish when spawning 1 Monitor Components of Support System 2 Monitor Climatic Properties Monitor longterm properties Monitor short term properties 2 Monitor Oceanographic Properties Monitor longterm properties Monitor short term properties 2182

167 SCRS/2013/132 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) ELEMENTS OF THE ECOLOGY AND MOVEMENT PATTERNS OF HIGHLY MIGRATORY FISH SPECIES OF INTEREST TO ICCAT IN THE SARGASSO SEA Brian E. Luckhurst 1 SUMMARY This paper provides information on the ecology and movement patterns of a total of 16 different fish species whose distributions include the Sargasso Sea. These species are divided into four groups that broadly correspond with ICCAT species groupings: Group 1 Principal tuna species including yellowfin tuna, albacore tuna, bigeye tuna, bluefin tuna and skipjack tuna, Group 2 Swordfish and billfishes including blue marlin, white marlin and sailfish, Group 3 Small tunas including wahoo, blackfin tuna, Little Tunny (Atlantic black skipjack tuna) and dolphinfish, and Group 4 Pelagic sharks including shortfin mako, blue and porbeagle. For each species, information is provided on ecology and habitat use in relation to oceanographic parameters such as water temperature, depth preference and dissolved oxygen. In addition, movement and migration patterns are discussed in relation to conventional tag-recapture results and more recent PSAT (Pop-up Satellite Archival Tag) tagging. The importance of Sargassum as essential fish habitat is discussed and is linked to the feeding habits of tunas and other pelagic predators. Flyingfishes are an important prey species in the diet of tunas and billfishes and, as they are largely dependent on Sargassum mats as spawning habitat, the Sargasso Sea plays a fundamental role in the trophic web of these highly migratory, pelagic species. Recent findings from PSAT tagging of several pelagic shark species has revealed the importance of the Sargasso Sea in their life cycles. RÉSUMÉ Le présent document fournit des informations sur l'écologie et les schémas de déplacement de 16 espèces de poissons dont les aires de distribution incluent la mer des Sargasses. Ces espèces sont divisées en quatre groupes qui correspondent en gros à la classification des espèces relevant de l'iccat : Groupe 1: principales espèces de thonidés, comprenant l'albacore, le germon, le thon obèse, le thon rouge et le listao. Groupe 2: espadon et istiophoridés, incluant le makaire bleu, le makaire blanc et le voilier. Groupe 3: thonidés mineurs incluant le thazard-bâtard, le thon à nageoires noires, la thonine commune et la coryphène commune. Groupe 4 : requins pélagiques, incluant le requin-taupe bleu, le requin peau bleue et le requin-taupe commun. Pour chaque espèce, des informations sont fournies en ce qui concerne l'écologie et l'utilisation de l habitat par rapport aux paramètres océanographiques, tels que la température de l'eau, la préférence des profondeurs et l'oxygène dissous. En outre, les schémas de déplacement et de migration sont discutés par rapport aux résultats de marquage conventionnel/récupération et du marquage PSAT (marques pop-up reliées à des satellites) plus récent. L'importance des Sargasses comme habitat de prédilection des poissons est abordée, ce qui est expliqué par les habitudes alimentaires des thonidés et d'autres prédateurs pélagiques. Les poissons volants sont une proie importante du régime alimentaire des thonidés et des istiophoridés, et sachant qu'ils dépendent en grande mesure des amas des Sargasses qui constituent leurs habitats de frai, la mer des Sargasses joue un rôle fondamental dans la chaîne trophique de ces poissons grands migrateurs pélagiques. Les récents résultats du marquage PSAT de plusieurs espèces de requins pélagiques ont révélé l'importance de la mer des Sargasses dans leurs cycles vitaux. RESUMEN Este documento proporciona información sobre la ecología y los patrones de movimiento de un total de 16 especies diferentes de peces cuyas distribuciones incluyen el mar de los Sargazos. Estas especies se dividen en cuatro grupos que corresponden de manera general a las agrupaciones de especies de ICCAT: Grupo 1: especies de túnidos principales, entre ellas rabil, atún blanco, patudo, atún rojo y listado. Grupo 2: pez espada e istiofóridos, entre ellos aguja azul, aguja blanca y pez vela. Grupo 3: pequeños túnidos entre ellos peto, atún aleta negra, bacoreta y dorado y Grupo 4: tiburones pelágicos, lo que incluye marrajo dientuso, tintorera y marrajo sardinero. Para cada especie, se facilita información sobre la ecología y el 1 Study commissioned by Sargasso Sea Alliance, led by the Government of Bermuda. 2-4 Via della Chiesa, Acqualoreto (TR), Umbria, Italy, Retired, Senior Marine Resources Officer, Government of Bermuda,

168 uso del hábitat en relación con parámetros oceanográficos como la temperatura del agua, la preferencia de profundidad y el oxígeno disuelto. Además, se discuten los patrones de movimiento y migración en relación con los resultados de marcado-recaptura con marcas convencionales y del marcado más reciente con marcas PSAT (marcas archivo pop-up por satélite). Se debatió la importancia del Sargassum como hábitat esencial de los peces y se vinculó con los hábitos alimentarios de los túnidos y otros depredadores pelágicos. Los peces voladores son una especie presa importante en la dieta de los túnidos y marlines, y dado que éstos dependen en gran medida de las malezas de Sargazos como hábitat de reproducción, el mar de los Sargazos desempeña un papel fundamental en la cadena trófica de estas especies pelágicas altamente migratorias. Hallazgos recientes de marcado PSAT de varias especies de tiburones pelágicos han puesto de relieve la importancia del mar de los Sargazos en sus ciclos vitales. KEYWORDS Introduction Sargasso Sea, Sargassum, Tunas, Swordfish, Billfishes, Sharks, Ecology, Habitat, Oceanography, Movements The Sargasso Sea is located within the North Atlantic sub-tropical gyre and a series of currents define its boundaries with the most influential current being the Gulf Stream in the west. The importance of the Sargasso Sea derives from a combination of factors - oceanographic features, complex pelagic ecosystems, and its role in global ocean processes (Laffoley et al., 2012). The Sargasso Sea contains the majority of the world s only pelagic ecosystem based upon floating Sargassum which hosts a highly diverse community of associated organisms. Sargassum and the Sargasso Sea provides essential habitat for key life history stages of a wide variety of species, some of which are endangered or threatened e.g. four species of sea turtles and the European eel (Anguilla anguilla). A variety of oceanographic processes impact productivity and species diversity. Sargassum is known to drift through the Caribbean, into the Gulf of Mexico and up the eastern seaboard of the USA in the Gulf Stream. Eddies of water which break away from the southern edge of the Gulf Stream may then spin into the central gyre trapping a significant portion of Sargassum (Laffoley et al., 2012). Once it has become entrained by the clockwise movement of currents circulating around the gyre, it may remain for long periods. The importance of the Sargasso Sea as an Ecologically or Biologically Significant Marine Area (EBSA) was recognized by the 11 th Conference of the Parties to the Convention on Biological Diversity in The Sargasso Sea Alliance (SSA) has defined a large portion of the Sargasso Sea as its study area. The proposed study area extends from 22 o 38 o N and from 76 o 43 o W, centred on 30 o N and 60 o W (Figure 1). This comprises an area of ~ 4,163,499 km 2 (Laffoley et al., 2012). The Sargasso Sea plays an important role in the ecology and life history of a variety of pelagic fish species, many of which are highly migratory and of both commercial and recreational importance in many countries. The International Commission for the Conservation of Atlantic Tunas (ICCAT) has fisheries management responsibility for many of these large, apex predators which form the basis of significant fisheries in the Atlantic Ocean. These species include the five major tuna species of commercial importance in the North Atlantic as well as swordfish and three species of billfishes. In addition, a number of species (Small Tunas category in ICCAT) which could be categorized as pelagic mesopredators, are monitored but not actively managed. More recently, ICCAT has started to evaluate stocks of some pelagic sharks which can comprise a significant element of longline catches. The focus of this paper is to outline elements of the ecology, oceanography and movement patterns of a total of 16 species (eight of them from the family Scombridae) mentioned above with respect to the Sargasso Sea. Significance of the Sargasso Sea to pelagic fishes The overall importance of Sargassum for fish life histories has been recognised by the USA as essential fish habitat (SAFMC, 2002). In a study of the feeding ecology of pelagic predators, Ruderhausen et al (2010) examined the dominant prey of blue marlin (Makaira nigricans), dolphinfish (Coryphaena hippurus), yellowfin tuna (Thunnus albacares) and wahoo (Acanthocybium solandri) in the North Atlantic. They classified prey into three groups: 1) prey associated with floating Sargassum 2) Flying fish (Exocoetidae) - associated with Sargassum during spawning and 3) schooling prey, primarily Auxis spp. and cephalopods. The dominant prey of yellowfin tuna (>50cm FL) were flying fish, as well as scombrids (mackerels, tunas, wahoo) and cephalopods. Dolphinfish were seen to feed mostly on prey associated with floating structure, mainly Sargassum, and flying 2184

169 fish. Blue marlin and wahoo preyed predominantly on scombrids. Thus, flyingfish were a significant component of the diet of these pelagic predators. It is recognized that Sargassum is a critical component for reproduction of flying fish (Oxenford et al., 1995) and thus the association of flying fish with Sargassum during spawning indicates the significance of this habitat for one of the principal prey groups of tunas, dolphinfish and other pelagic predators. This represents a direct trophic link between Sargassum and the diets of tunas and tuna-like species thus again emphasizing the importance of the Sargasso Sea. ICCAT has recently recognized the importance of Sargassum as fish habitat and has requested that Contracting Parties assess the ecological status of Sargassum as habitat for tunas, billfish and sharks (ICCAT, 2012). In more general terms, Coston-Clements et al (1991) list a number of teleost species as being associated with pelagic Sargassum in the North Atlantic. The nature of the association is not always evident but is typically related to food and feeding habits. In the Gulf Stream off North Carolina, Casazza and Ross (2008) found that significantly more fishes (n = 18,799) representing at least 80 species, were collected from samples containing Sargassum, than from samples collected from open water habitat (60 species, 2,706 individuals). The majority (96%) of fishes collected in both habitats were juveniles. Underwater video recordings indicated a layered structure of fishes among and below Sargassum and that smaller fishes were more closely associated with the algae than larger fishes. Underwater video observations of schooling behaviors of dolphinfish and jacks (Carangidae) were also recorded. Wells and Rooker (2004) studied the distribution and abundance of fishes associated with Sargassum mats in the northwestern Gulf of Mexico during the summer months. A total of 36 species (17 families) was identified with seven species comprising over 97% of the catch. A large majority (over 95%) of the species collected were in early life history stages confirming the importance of pelagic Sargassum as nursery habitat for some species and suggesting that its presence may influence recruitment success. The early life history stages (primarily juveniles) of a number of pelagic predators associated with Sargassum are listed by Coston-Clements et al (1991) and they include dolphinfish, wahoo, swordfish (Xiphius gladius), blue marlin, white marlin (Tetrapterus albidus) and sailfish (Istiophorus albicans). The importance of the Sargasso Sea to various pelagic species is not always directly evident but an evaluation of existing information suggests that its importance rests mainly with its status in relation to one or more of the following: migratory route, spawning area, nursery area and feeding area. In the case of some species of pelagic sharks, it has been shown that the Sargasso Sea is of considerable importance as an overwintering ground or pupping area. In this paper, the migratory species with a known association with the Sargasso Sea have been divided into four groups, largely along ecological lines although many are apex predators. The groups are: 1) Principal tuna species 2) Swordfish and billfishes 3) Small tuna group as defined by ICCAT 4) Pelagic sharks. Group 1 Principal Tuna species Yellowfin tuna (Thunnus albacares) Ecology and oceanography The yellowfin tuna is a gregarious species, tending to form schools, either free-swimming or associated with FADs, underwater ridges and different marine animals. For example, the fishery for this species is associated with dolphins in the Pacific Ocean. Adults generally form shoals of specimens of the same size (ICCAT, 2010a). This behaviour also predominates in the juveniles which form shoals with specimens that do not necessarily come from the same breeding group in specific migration periods (ICCAT, 2010a). Free-swimming schools of yellowfin (i.e. not associated with FADs) tend to be made up of large individuals and to be monospecific. Yellowfin are found across a broad thermal range (18 31 C) and vertical distribution is determined by the thermal structure of the water column (Collette and Nauen,1983). In general, yellowfin limit their incursions into depths in which the water temperature does not fall more than 8º C with respect to the temperature of the surface layer. It spends more than 90% of its time in waters with a uniform temperature of around 22º C (ICCAT, 2010a). Although it is known to dive to depths of 350 m, adult and juvenile yellowfin spend most of their time in the surface layer, above 100 m. (ICCAT, 2010a). There are generally insignificant differences in depth distribution between day and night. The level of dissolved oxygen is a limiting factor for the depth distribution of yellowfin, as a concentration of 3.5 ml/ l appears to limit their depth distribution (ICCAT, 2010a). 2185

170 Migration and movements Yellowfin tuna is the species of tropical tuna that is considered to make the largest migrations, i.e. periodic and regular movements of a large part of the population (ICCAT, 2010a). As migratory behaviour varies with size (age), it is necessary to examine the migratory patterns of three size age categories: juveniles (50-65 cm), preadults ( cm) and adults ( cm) - in order to better understand the dynamics of this species in the Atlantic Ocean (ICCAT, 2010a). Juveniles (up to 50 cm FL) typically remain in coastal waters and undertake only modest movements. With increase in size, yellowfin movements become more extensive and by the time they reach sexual maturity, trans-atlantic migrations take place (ICCAT, 2010a). In general, adults make trophic migrations northwards in the summer months and then return to their spawning grounds in the winter months. Tag- recapture results of yellowfin tuna from tagging locations off the eastern seaboard of the USA indicate a strong west to east movement (Luckhurst, 2007). The majority of the straight-line tagging vectors suggest a transit across the Sargasso Sea in this migration (ICCAT, 2010a). Other data from the Bermuda Seamount (32 N, 64 W) suggest a seasonal migration in the North West Atlantic in the summer months (Luckhurst et al, 2001;Luckhurst, 2007) northward into the Sargasso Sea. Three recaptures of yellowfin tagged in Bermuda also indicate movement through the Sargasso Sea, one was recaptured off Cape Hatteras and the remaining two yellowfin near Puerto Rico (Figure 2). Albacore tuna (Thunnus alalunga) Ecology and oceanography Albacore is an epi- and mesopelagic oceanic species which seldom comes close to shore and prefers deep, open waters. Temperature is one of the dominant factors determining the distribution of albacore and they prefer cooler sea temperatures than more tropical species such as yellowfin tuna (ICCAT, 2010b). The preferred thermal range for albacore is 10-20ºC although temperatures outside this range can be tolerated for short periods. As a result of this thermal preference, the distribution of areas suitable for albacore in the North Atlantic includes the entire Sargasso Sea area and most of the north Atlantic. Albacore occur mainly in the temperature range of 14-20ºC off North-America and between 16-21ºC in the northeast Atlantic (ICCAT, 2010b). These thermal preferences are thought to act as barriers to movements of albacore across the equatorial zone from different regions and this has resulted in separate populations with the designation of north and south Atlantic stocks (ICCAT, 2010b). Albacore appear to be searching for an optimum temperature zone when they undertake periodic vertical migrations from warm surface waters to deep cooler waters. In a study in the Northeast Pacific, individuals 3-5 years old spent 80% of the time at 100 m, around the thermocline depth, and moved only occasionally to the mixed surface layer or to deeper waters (ICCAT, 2010b). It was also noted that albacore undertook vertical migrations with larger depth range during the day than during the night. Maximum depth distribution is reported to range from m in the Pacific Ocean (ICCAT, 2010b). The swim bladder is not fully developed in juvenile albacore and as a result, juveniles have less ability to undertake vertical migrations in the water column. Tunas have a high metabolic rate with consequent high oxygen demand. The minimum estimated dissolved oxygen concentration for albacore is 3.7 ml/l. and using this tolerance level, a large area of the eastern Atlantic, south of 20ºN latitude and extending westwards towards Brazil, is not suitable for albacore at depths greater than 100m (ICCAT, 2010b). Migration and movements Albacore have been documented to migrate from the north Atlantic to the Mediterranean and vice versa as well as making transatlantic migrations (ICCAT, 2010b). However, no migrations from the north to the south Atlantic have yet been recorded. Despite these tagging data, albacore migration routes are still uncertain. In the North Atlantic, both juveniles and adults apparently spend the winter in the central Atlantic area. When the water starts warming up in the spring, young albacore start a trophic migration, heading to highly productive waters in the northeast Atlantic (ICCAT, 2010b). In May, tuna start to concentrate in surface waters near the Azores at 38ºN latitude and begin to move north in waters of 17-20ºC. In the autumn, albacore start migrating back to the mid Atlantic. The trophic migration takes place for the first four years of their lifetime until they reach sexual maturity (ICCAT, 2010b). In contrast, adult albacore undertake reproductive migrations when summertime approaches. They migrate to their spawning grounds in the western part of the north Atlantic (including the Sargasso Sea) and also to an area offshore from Venezuela (ICCAT, 2010b). 2186

171 Bigeye tuna (Thunnus obesus) Ecology and oceanography Bigeye tuna is an epi- and mesopelagic species which generally prefers open waters. Like other tunas, they are gregarious and tend to form schools, either independently or in association with drifting objects, marine animals or seamounts. In the eastern Atlantic, bigeye tuna are frequently associated with a large variety of drifting objects, including dead whales, or with some living animals. Bigeye schools associated with floating objects are comprised primarily of small fish (under 5 kg.) although larger fish are also found. Free schools (not associated to any object) are typically formed by large individuals of the same species (ICCAT, 2010c). The main environmental factors affecting the vertical distribution of bigeye are the depth of the deep scattering layer and temperature (ICCAT, 2010c). The optimum temperature range for bigeye is 17º- 22ºC and they are not found in waters greater than approximately 29ºC (Collette and Nauen, 1983). However, bigeye have a broad thermal tolerance as they are exposed to temperatures down to about 5ºC when they dive to 500 m depth, i.e. up to 20ºC colder than surface water temperature. Bigeye exhibits a characteristic behavioural pattern with respect to depth. They remain within the surface layer, at a depth of about 50 m during the night but typically dive to depths of up to 500 m at sunrise (ICCAT, 2010c). Depths of over 1000 m were recorded in a study on bigeye tuna conducted using archival tags in the Coral Sea. The bigeye typically ascends swiftly to the temperate surface layer, probably in order to regulate body temperature or possibly to compensate for oxygen deficiency (ICCAT, 2010c). There appears to be a positive correlation between moonlight intensity and the depth at which bigeye tuna are found, the mean depth increasing as the intensity of lunar light increases. Dissolved oxygen concentration is also an important ecological factor for tunas because of their high metabolic rate. Bigeye tuna are able to withstand lower concentrations of dissolved oxygen than any other tuna species and, as a result, it is capable of inhabiting deeper waters (ICCAT, 2010c) where oxygen concentrations can be less than 1.5 ml/l. Migration and movements Bigeye tuna, in common with yellowfin, undertake very significant migrations. Tag-recapture data indicates that bigeye travel faster than yellowfin tuna. Also in common with yellowfin, bigeye seasonal movement patterns (trophic or spawning) may be characterised by size(age) groups (ICCAT, 2010c). In the eastern Atlantic spawning areas, young individuals (30-70 cm FL) tend to gravitate towards the equatorial area (Gulf of Guinea) forming mixed schools with young skipjack and yellowfin tuna. These fish are taken by purse-seine fleets. In contrast, adults (fish over 100 cm FL) are caught throughout the whole of the tropical and sub-tropical Atlantic with longline gear (ICCAT, 2010c). Tagging studies show trans-atlantic migrations westward from the Gulf of Guinea to the north of Brazil and movements from the Gulf of Guinea along the African coastline. As few tagging studies of bigeye have been conducted in the western Atlantic, data is sparse but trans-atlantic migrations have been recorded from the eastern seaboard of the US to the Gulf of Guinea as well as along the North-American coast with a few individuals reaching as far as 50ºN latitude (ICCAT, 2010c). As bigeye are found throughout the entire North Atlantic and have an optimum temperature range of 17-22ºC, the Sargasso Sea provides a suitable environment for spawning but due to a lack of tagging and research on bigeye in the western central Atlantic, there is little specific information about bigeye in the Sargasso Sea except ICCAT reported landings. Bluefin tuna (Thunnus thynnus) Ecology and oceanography Bluefin tuna has the widest geographical distribution of all of the tuna species in the North Atlantic and is the only large pelagic species living permanently in temperate Atlantic waters (Fromentin and Fonteneau, 2001). Bluefin occupy the surface and subsurface waters of both coastal and open-ocean areas, but have been recorded as diving to depths of 500m to 1000m with some frequency (ICCAT, 2010d). Archival tagging data have confirmed that bluefin can sustain a wide range of temperatures (down to 3 C, up to 30 C), while maintaining 2187

172 stable internal body temperatures (ICCAT, 2010d). Similar behaviour has also been reported for other large pelagic species (e.g. bigeye tuna, swordfish) and is generally associated with foraging in the deep scattering layer and possibly to physiological constraints to cool the body temperature (ICCAT, 2010d). The movement patterns of bluefin tuna and their spatial distribution have, until recently, been hypothesized to be controlled by preferential ranges and gradients of temperature, similar to other tuna species but more recently, it is believed that juvenile and adult bluefin tuna frequent and aggregate along ocean fronts (ICCAT, 2010d). This association appears to be related to foraging as bluefin feed on the concentrations of both vertebrate and invertebrate prey found in these areas. The types of ocean fronts known to be frequently visited by bluefin tuna are upwelling areas, such as the West coasts of Morocco and Portugal, and meso-scale oceanographic structures associated with the general circulation of the North Atlantic and adjacent seas (ICCAT, 2010d). Although there is some agreement amongst scientists about this association, bluefin tuna habitat appears to be more complex than can be explained by these oceanographic features alone. Migration and movements Bluefin tuna is a highly migratory species and is found throughout the pelagic ecosystem of the entire North Atlantic and its adjacent seas, mainly the Mediterranean Sea. It is known from the early results of conventional tagging that bluefin undertake transatlantic migrations and subsequent PSAT (Pop-up Satellite Archival Tag) tagging has helped determine the routes and the depths at which they migrate (Wilson and Block, 2009). They display homing behavior and spawning site fidelity in both the Gulf of Mexico and the Mediterranean Sea, which constitute the two main spawning areas (ICCAT, 2010d). Less is known about feeding migrations within the North Atlantic and the Mediterranean but PSAT tagging results indicate that bluefin tuna movement patterns vary considerably between individuals, years and areas (Block et al, 2005). Recent PSAT tagging of bluefin in the Gulf of Mexico in May (normally the peak spawning month) indicates that, after leaving the Gulf of Mexico, some fish migrate northward through the Sargasso Sea towards the Gulf of Maine (Eric Prince, pers. comm.) where they are often found during the summer months (Wilson et al, 2005). As there is no direct evidence of spawning in the Sargasso Sea, these fish are presumed to be feeding, possibly around seamounts, as they travel north. Migratory routes for bluefin tuna PSAT- tagged in the western Atlantic indicate that they transit the Sargasso Sea when migrating eastward to the Mediterranean Sea (Block et al, 2005). Other studies have placed spawning-sized bluefin in an area between Bermuda and the Azores in the central North Atlantic during the spawning period (Lutcavage et al, 1999). These fish were all located above 33 N in the northern sector of the Sargasso Sea and may have been using this area as a foraging ground. Skipjack tuna (Katsuwonus pelamis) Ecology and oceanography Skipjack tuna is an epipelagic species generally inhabiting open waters. In common with yellowfin and bigeye tunas, skipjack tend to form schools, either independently or in association with floating objects, marine animals or seamounts. Aggregations of this species tend to be associated with convergences, water mass boundaries, outcrops and other hydrographic discontinuities (Collette & Nauen, 1983). The distribution of skipjack is more tropical than other tunas as they are normally found in waters ranging from 20ºC to 30ºC although they can tolerate temperatures down to 15ºC. They generally dive only to depths where the water temperature is not more than 8ºC below the temperature in the surface layer (ICCAT, 2010e). Although skipjack remain close to the surface during the night (Collette & Nauen, 1983), they are capable of diving to a depth of 260m, a depth considerably less than other tuna species. The minimum values of dissolved oxygen in skipjack tuna habitat have been established at ml/l where temperature is not a limiting factor (ICCAT, 2010e). As a result, this generally restricts skipjack to waters above the thermocline making them more vulnerable to surface gear such as purse seines (ICCAT, 2010e). However, results of acoustic tagging indicate that skipjack can make brief dives down to 400 m with temperatures below 14ºC and an oxygen level close to 1.5 ml l/l. 2188

173 Migration and movements The movements of skipjack are influenced by environmental conditions (temperature, salinity, nutrients) and by their tendency to group around floating objects, which may attract other tuna species such as young yellowfin and bigeye (ICCAT, 2010e). The majority of tagging has been done in the equatorial waters of the eastern Atlantic. Results from this tagging effort show that migrations generally follow the coastline, moving both north and south but also with some westward movement. In the western Atlantic, there is very little information from tagging with the only migrations being along the Brazilian coast and minor movements in the Caribbean. For the entire Atlantic, there have been only two East-West transatlantic migrations recorded (ICCAT, 2010e). Thus there is no known specific association of skipjack with the Sargasso Sea. Skipjack are infrequently caught in Bermuda (located in the western Sargasso Sea) and then only in the summer months when the warmest SSTs occur (Luckhurst, pers. obs). Group 2 Swordfish and billfishes Swordfish (Xiphius gladius) Ecology and oceanography Swordfish are distributed throughout the Atlantic Ocean and Mediterranean Sea. They spawn mostly in the warm tropical and subtropical waters of the western Atlantic throughout the year but are found in the colder temperate waters during summer and fall months (ICCAT, 2010f). Although swordfish is an oceanic species, it is sometimes found in coastal waters, generally above the thermocline. The swordfish is the species of billfish with the greatest tolerance to temperature (5 to 27 C), but is usually found in surface waters at temperatures above 13 C (ICCAT, 2010f). Adult swordfish are generally solitary and are not known to form schools in the open ocean. Acoustic tagging had earlier shown that swordfish stay near the surface at night, but return to depths of up to 600 m during the day and are presumed to be feeding in the deep scattering layer (ICCAT, 2010f). A PSAT tag deployed on a small swordfish (59 kg) in the northwest Atlantic provided compelling evidence of this diurnal vertical migration behavior. Throughout the monitoring period, this fish made regular dives to m depth during daylight hours (Figure 3, inset) while during nocturnal hours, mean depth was much shallower with brief, regular periods spent at the surface (Luckhurst, 2007). This fish moved northward through the Sargasso Sea covering a distance of 2,629 km in 62 days (Figure 3). Migration and movements Swordfish are known to move through the Sargasso Sea as part of a seasonal migration from the tropical Atlantic to the temperate northwest Atlantic waters (Neilson et al., 2009). The data from conventional tagging based on almost 400 recaptures presented by Luckhurst (2007), indicates that the predominant movement pattern of swordfish in the North Atlantic appears to be north-south, transiting the Sargasso Sea (Figure 4), with some east-west movement These recaptures included several trans-atlantic movements. Although the reasons for this seasonal movement are unclear, it may well be associated with feeding and prey concentrations in thermal boundaries between water masses, suggesting that the Sargasso Sea may be a productive feeding ground (ICCAT, 2010f). In further support of this hypothesis, a swordfish conventionally-tagged from a longliner in the Northwest Atlantic in July 1997 moved in a southerly direction through the Sargasso Sea over 900 km before being recaptured off Bermuda in December, less than six months later (Luckhurst, pers. obs.; E. Prince, pers. comm., NMFS). Data indicating spawning in the north-central Gulf of Mexico and east of the Caribbean islands (ICCAT, 2010f) suggest that the north-south movements may constitute migrations to spawning grounds in warmer waters. Blue marlin (Makaira nigricans) Ecology and oceanography Blue marlin is an epipelagic, oceanic species typically found in wide-open, blue waters. Blue marlins are not schooling fish and are considered a rare and solitary species. Adults are mostly found in tropical waters and their distribution appears to be bounded by the 24 C isotherm (ICCAT, 2010g). As habitat preferences of billfish are poorly known compared with tuna, the advent of PSAT tagging has greatly enhanced our understanding of 2189

174 marlin habitat in recent years. Results from PSAT tagging in the western North Atlantic (Graves et al. 2002; Kerstetter et al. 2003) indicate that blue marlin typically remain where the SST range is C. Blue marlins are largely associated with the epipelagic zone and spend over 80% of their time in water temperatures ranging from C. They do, however, undertake frequent, short duration dives to depths where the temperature may be as much as 14 C below SST (ICCAT, 2010g).. A study which deployed 79 PSATs in several areas of the Atlantic (Goodyear et al. 2006) found that blue marlin showed a mean minimum temperature preference of 17.4 C while other studies suggest that the thermal preference for this species appears to be the warmest waters available in the open ocean (ICCAT, 2010g). Depth distribution data from PSAT tagging has indicated that blue marlin spend most of their time in warm near surface waters (<25 m) in the northwestern Atlantic (Kerstetter et al. 2003) but they make frequent, short duration vertical dives from the surface layer to depths >300 m. Goodyear et al. (2006) indicated that blue marlin made deep, short duration dives below 800 m but the mean dive depth for 48 tagged fish was m. Although dissolved oxygen requirements for marlins are poorly known, blue marlin are ram ventilators and require sufficient oxygen to support their high metabolic rates. Prince and Goodyear (2006) proposed that the minimal oxygen concentration for billfish is 3.5 ml/l, defining it as the hypoxic threshold for billfish species. PSAT tagging results from the eastern tropical Atlantic suggest that this minimum oxygen level is indeed a barrier for blue marlin (Prince et al, 2010). Migration and movements Blue marlin display extensive movements in the Atlantic Ocean. An examination of the databases of the Cooperative Tagging Center (CTC) and The Billfish Foundation (TBF) of conventional tag deployments indicates that the majority of tagging effort has taken place in the western Atlantic (Ortiz et al. 2003). There have been a total of 52,185 blue marlin releases and 769 recaptures as of the end of calendar 2005, with 18 of these recaptures demonstrating trans-atlantic movements (Luckhurst, 2007). Tag recapture rates for blue marlin are generally < 1% throughout the world s oceans (Ortiz et al. 2003). The dominant movement patterns for recaptured blue marlin are primarily from west to east and, as blue marlin prefer warm tropical waters, these movements are primarily in the tropical Atlantic. The few transatlantic and trans-equatorial movements represent about 5% of the documented blue marlin recaptures. However, despite this extensive tagging effort, blue marlin migration routes are still uncertain. The advent of PSAT tagging has provided many new, important insights into movement and migration as well as habitat use. The first PSAT tagging of blue marlin in the NW Atlantic took place in the recreational fishery in Bermuda in 1999 (Luckhurst, pers. obs.). Eight of the nine tagged blue marlin reported their positions after five days and moved distances ranging from km but in all compass directions (Graves et al. 2002). Longer deployments of PSAT tags on blue marlin from commercial longliners in the NW Atlantic demonstrated substantial movement distances (Kerstetter et al. 2003). Two blue marlin tagged with PSATs moved distances of 985 km and 1,968 km in 30 days. During the period , a total of 66 PSATs were deployed (E. Prince, pers. comm., NMFS), primarily from recreational fishing vessels in the wider Caribbean. These deployments resulted in long distance movements by a number of specimens (Figure 5). The longest movement vector of 4,606 km was of a blue marlin (68 kg) tagged in the Turks and Caicos Islands which moved this distance to the eastern tropical Atlantic in 91 days. Although blue marlin are principally a tropical species, they do migrate northward into the Sargasso Sea during the summer months and form the basis of a significant recreational fishery in Bermuda (Luckhurst, 1998). Sampling of female blue marlin, principally at tournaments held in July, confirmed that blue marlin were actively spawning in the Sargasso Sea (Luckhurst et al, 2006). This finding significantly extended the northern limit of known spawning areas in the northwest Atlantic into the northern half of the Sargasso Sea at Bermuda s latitude (32 N). White marlin (Tetrapterus albidus) Ecology and oceanography White marlin is an epipelagic species that is mostly solitary, although it is known to occasionally form small groups. They generally prefer water >100 m deep with surface temperatures above 22 C. They often associate themselves with ocean fronts, steep drop offs, submarine canyons, and other features where shoaling of prey 2190

175 species may occur (ICCAT, 2013). Much of the information about habitat use has been gathered in recent years from PSAT tagging. Deployments of these tags in the western Atlantic indicates that white marlin spend most of their time in the epipelagic zone in water C (ICCAT, 2013). More recent studies have shown that white marlin can expand the temperature range in which they live but they spend almost all their time in the warmest surface layer (97.2% of darkness hours and 80.3% of daylight hours) (ICCAT, 2013). They are capable of short duration deeper dives, which are probably for foraging, but their preference is to stay in the warmest water available. In common with most tropical tunas, billfishes generally do not venture into waters that are more than ~8 C below the surface temperature (ICCAT, 2013). Depth distributions from PSAT data indicate white marlin spend most of their time in warm near-surface temperatures (<25 m depth) in the western North Atlantic but they make frequent short- duration dives typically in the m depth range. As with other billfishes, dissolved oxygen requirements are poorly understood but PSAT tagging data suggests white marlin are limited by a minimum dissolved oxygen concentration of about 3.5 ml/1 (Prince and Goodyear, 2006). Migration and movements The majority of information on white marlin movements originates from conventional tag-recapture results mostly from the western North Atlantic. These results indicate a few transatlantic movements, but no transequatorial movements (ICCAT, 2013). The longest linear displacement recorded for a white marlin was 6,517 km (Ortiz et al., 2003). As with blue marlin, white marlin also migrate north in the summer months into the Sargasso Sea and are taken regularly in the recreational fishery in Bermuda but largely on a catch-andrelease basis (Luckhurst 1998). White marlin are also known to spawn in the Sargasso Sea (NMFS, 2007) Sailfish (Istiophorus albicans) Ecology and oceanography Sailfish is an epipelagic species but is the least oceanic of the Atlantic billfishes, as it shows a strong tendency to approach continental coasts, islands and reefs. Sailfish are found in schools during the winter months in the western central Atlantic, in Florida waters, offshore waters of the Gulf of Mexico and the Caribbean Sea (ICCAT, 2010h). It has been suggested that they form schools when the principal prey are abundant schooling species (e.g., clupeids) and they have been documented to hunt cooperatively to concentrate schools of prey species. Fish in Florida waters tend to disperse northward in the summer following the inside edge of the Gulf Stream up the east coast of the US (ICCAT, 2010h). Temperature preferences for sailfish appear to be associated with the seasonal movement of the 28 C isotherm and, as a consequence, they are often found above the thermocline in the surface layer. In the northwestern Atlantic, various studies have shown that sailfish spend most of the time in warm near-surface waters (10-20 m depth) but they also show that sailfish display frequent, short duration dives to depths of m, similar to white marlin (ICCAT, 2010h). Prince and Goodyear (2006) proposed that the minimal oxygen concentration for billfishes is 3.5 ml/l, defining it as the hypoxic threshold for these species. PSAT tagging results from the eastern tropical Atlantic tend to support the suggestion that this minimum oxygen level is a depth barrier for sailfish (Prince et al, 2010). Migration and movements Sailfish display more restricted movements in the Atlantic than blue or white marlin with no evidence of transatlantic or trans-equatorial movements (ICCAT, 2010h). However, based on minimum distance travelled in tag-recaptured fish, it has been suggested that sailfish in different areas make either cyclic annual movements, exhibit some degree of fidelity, or some combination of the two (Ortiz et al. 2003). Results from the western North Atlantic, where most of the tag and release of sailfish has taken place, indicate significant movements between the Straits of Florida and adjacent waters, and the Gulf of Mexico and the area near Cape Hatteras (35 N). In general, most of the tag- recaptures have occurred in the same general area as the point of release (ICCAT, 2010h). The longest movement recorded was from a sailfish tagged and released off the U.S. northeast coast, and recaptured off Suriname 11 months later having travelled a distance of 3,861 km (Ortiz et al. 2003). 2191

176 Group 3 Small tunas Wahoo (Acanthocybium solandri) Ecology and oceanography Wahoo is an epipelagic,oceanic species which is frequently solitary but may form small loose aggregations rather than compact schools (Collette and Nauen, 1983). Little was known about habitat use until the advent of electronic tagging. Thiesen and Baldwin (2012) deployed PSAT tags on four wahoo in the western North Atlantic which provided data over a total of 198 days. Wahoo spent >90% of their time in water less than 200m depth and > 90% of their time was spent in water temperatures ranging from 17.5 to 27.5 C. Three of the four fish made regular dives to depths greater than 200m (Thiesen and Baldwin, 2012). All four fish displayed significant differences in mean depth in daylight (50.7m) and during darkness (29.7m). Migration and movements Relatively little is known about wahoo movement patterns and migration in the western Atlantic and there has been very limited success using conventional tags to define migration patterns (Oxenford et al., 2003). In Bermuda, a small scale wahoo tagging program in 1998 succeeded in tagging only 15 wahoo before the program was forced to conclude. However, one of these tagged wahoo was recaptured 10 months later, 64 km away from the point of release (Oxenford et al., 2003). It is not possible to know if it remained in Bermuda waters during its time at liberty or returned after a seasonal migration. It is conceivable that it followed a migratory route into the western central Atlantic, with the Bermuda Seamount, located in the western Sargasso Sea as a seasonal feeding /spawning area, as has been postulated for yellowfin tuna and blackfin tuna in Bermuda (Luckhurst et al., 2001). Another form of evidence suggesting a migration pattern in the Sargasso Sea is marked seasonality in landings. In Bermuda, wahoo catches have a strong seasonal pattern with 60 70% of the annual landings consistently occurring in the second and third quarters of the year (April September) during the period of highest water temperatures (Luckhurst and Trott, 2000). Wahoo are known to actively spawn in Bermuda waters in the summer months (Oxenford et al, 2003). Historically, there are spring (April May) and fall (August September) runs of wahoo in Bermuda which vary inter-annually in magnitude and to a lesser degree in timing (Luckhurst and Trott, 2000). However, wahoo landings are consistently lowest (5 8% of annual landings) in the first quarter which coincides with the lowest annual SST (18 C). The advent of PSAT tagging of wahoo in recent years has provided important insights into movements and migration (Luckhurst, 2007). PSAT tags deployed on four wahoo in the western North Atlantic indicated that straight-line distances moved (deployment to pop-up) ranged from to 1,960 km (Thiesen and Baldwin, 2012). The movement patterns of these tagged fish appeared to be largely north-south movements in relation to the Gulf Stream. Blackfin tuna (Thunnus atlanticus) Ecology and oceanography Blackfin tuna is an epipelagic, oceanic species occurring in waters of at least 20 C but is most common in tropical waters. Their distribution is confined to the western Atlantic (Collette and Nauen, 1983). Blackfin frequently form large mixed schools with skipjack tuna. These mixed schools form the basis of an important fishery off the southwest coast of Cuba which uses live-baits and poles (Claro et al, 2001). In a study of blackfin around moored FADs in Martinique, Doray et al (2004) indicated that there is no evidence that moored FADs act as an ecological trap for blackfin tuna. There were a small number of juveniles found around the FADs and the authors postulated that these juveniles leave the moored FADs to undergo a trophic migration and then return in the following year as adults to spawn in the area of the Lesser Antilles. Migration and movements Very little is known about the migratory patterns of blackfin tuna in the western Atlantic but data from Bermuda suggest that an annual migration to the seamount may occur from the more southerly areas of their distribution, presumably for feeding. However, it is not known if blackfin spawn in Bermuda waters (Luckhurst, pers. obs.). An examination of commercial landings of blackfin in Bermuda between indicated that the 2192

177 proportion of annual landings taken in the third quarter (July-September) ranged from about 50% to over 70% (Luckhurst et al, 2001). This is the period of the warmest SST, a maximum of about 30 C. In contrast, first quarter (January-March) landings are less than 5% of the annual total when the SST is about 18 C (Luckhurst et al, 2001). Manooch (1984) reported that blackfin are only found off North Carolina during the warmest months (June- September). Thus, it appears that blackfin may be migrating north as SST warms in the summer months but move south in the fall to stay within their thermal preference zone (Collette and Nauen, 1983). Another line of evidence is derived from tag-recapture results in Bermuda. The long term tag-recapture rate of blackfin in Bermuda is 10.7% compared to 3.5% for the western Atlantic (Luckhurst et al, 2001). All Bermuda recaptures occurred during the summer months with the largest mode for time-at-liberty at approximately one year. Although the sample sizes are small, a second mode occurred at three years, with single recaptures at two and four years (Luckhurst et al, 2001). Taken together these findings tend to support the hypothesis of an annual summer migration through the Sargasso Sea to the Bermuda Seamount. In contrast, a conventional tagging program in the eastern Caribbean which tagged 787 blackfin had only 11 recaptures (1.4%). Despite times at liberty ranging to >3 years, all recaptures were near to or at the original release sites (Singh-Renton and Renton, 2007). Little Tunny (Euthynnus alleteratus) Ecology and oceanography Little Tunny is an epipelagic, neritic species which is typically found in coastal waters with swift currents, near shoals and around thermal fronts and upwelling areas. This species is more coastal than other tuna species and is most abundant in the tropical Atlantic where the water temperature is 24º- 30ºC (ICCAT, 2010i). This species lives in schools of similar-sized fish together with other scombrid species, but has a tendency to scatter during certain periods of the year (Collette and Nauen, 1983). In Bermuda, Little Tunny form monospecific schools inshore during the summer months (Luckhurst, pers. obs.) but as water temperatures start to fall they disappear and presumably migrate south. This observation confirms their presence in the western part of the Sargasso Sea but it is not known if their summer distribution extends eastward further into the Sargasso Sea. Migration and movements Very little is known about the migratory habits of Little Tunny in the western Atlantic. In Bermuda, schools are taken by seine nets during the summer months to be used for bait (Smith-Vaniz et al, 1999). They appear not to be present during the winter months suggesting that this may be another case of seasonal migration to the Bermuda Seamount. This is consistent with the observations of Manooch (1984) who reported that Little Tunny migrate northward in schools in the coastal waters of the US eastern seaboard in the spring and return southward in the fall. Dolphinfish (Coryphaena hippurus) Ecology and oceanography Dolphinfish is an epipelagic, oceanic species which spends most of its time in the surface layer. It is found in tropical and subtropical waters warmer than 20 C (Oxenford, 1999). Dolphinfish are attracted to floating objects of all kinds including windrows or mats of Sargassum around which they frequently congregate and feed. They are not known to orient to any particular oceanographic features but rather to floating objects which may be influenced by such features. Small dolphinfish often travel together in schools ranging from several individuals up to 50 fish. Larger adult fish are normally solitary or in pairs. Migration and movements Palko et al (1982) reported that there was little direct evidence of dolphinfish movements in the western central Atlantic. They believe that migrations and movements are likely to be affected by the movement of drifting objects (including mats of Sargassum) in oceanic waters with which dolphinfish are associated. Oxenford and Hunte (1986) proposed two migration circuits of dolphinfish in the northeast and southeast Caribbean, based largely on seasonality of fisheries by location and mean size-at-capture. They suggest a northeastern migration circuit incorporating the northern Caribbean islands, the southeastern US and Bermuda, and a southeast circuit incorporating the southeastern Caribbean islands and the north coast of Brazil. A part of the proposed northeastern circuit is supported by Beardsley (1967) who reported that dolphinfish probably move northward 2193

178 from Florida during spring and summer. Luckhurst and Trott (2000) reported that landings of dolphinfish in the Bermuda fishery from showed a high level of seasonality with third quarter (July-September) landings comprising 45-60% of annual landings. This is the period with the highest SST and this landings pattern, in common with other species already described, is consistent with a highly migratory species. Off the southeastern US, dolphinfish frequently congregate around Sargassum, which serves as both shelter and a source of food (Manooch, 1984). Many of the food types eaten by dolphinfish are found in floating mats of Sargassum and this alga is frequently found in their stomach contents. Group 4 Pelagic sharks Shortfin mako (Isurus oxyrinchus) Ecology and oceanography Shortfin mako is a common, offshore littoral and epipelagic species in coastal and oceanic waters that occurs from the surface down to at least 500 m depth (Kohler et al, 2002). The preferred water temperature range of shortfin mako in the North Atlantic appears to be in a narrow range from 17 to 22 C (Kohler et al, 2002). An analysis of surface temperature data from almost 2,800 sets of swordfish longline gear between the Gulf of Mexico and the Grand Banks found that the mean minimum and maximum temperatures in which makos were caught were 18.5 and 20.5 C, respectively. The highest catch rates for mako sharks in the Spanish swordfish longline fishery in the North Atlantic also occurred at 18 C and frequently along the margins of the Gulf Stream (Kohler et al, 2002). Other evidence on the preferred water temperatures is provided by acoustic tagging. A kg mako shark was followed for four days from Florida across the Gulf Stream and into the Sargasso Sea. The shark swam at depths ranging from the surface to 500 m but spent most of its time at depths where the temperature range was between 17 and 22 C (Kohler et al, 2002). A PSAT tag deployed on a shortfin mako off the southeastern US indicated only limited movement (72 km) over 60 days (Loefer et al, 2005). The depth range covered during this period was from the surface to 556 m and the water temperature range was C. This tagged mako demonstrated a diel pattern of vertical movement defined by greater mean depths during daylight hours (Loefer et al, 2005). The tagging data also suggested a seasonal behavioral change in vertical movements with increasing SST. Migration and movements In a long term shark tagging program ( ), a total of 5,333 shortfin mako sharks were tagged with 608 being recaptured (11.4%). Distances travelled ranged from no movement to a maximum of 5,310 km (Kohler et al, 2002). This maximum distance was a transatlantic migration, the tagged mako moving from the northeastern U.S. and being recaptured 1.4 years later off west Africa. In this tagging study, 75% of the makos travelled less than 926 km from their original tagging location with a mean distance of 737 km (Kohler et al, 2002). Times at liberty ranged from 1 day to 12.8 years. This longest-time-at-liberty was a male tagged off North Carolina that was recaptured only 457 km away, off South Carolina. Overall, 75% of tagged makos were at liberty for less than 2 years with a mean of 1.2 years (Kohler et al 2002). Kohler et al (2002) proposed a Sargasso Sea Hypothesis to explain shortfin mako migrations in the western North Atlantic. In January, shortfin makos are common along the western margin of the Gulf Stream, with at least one area of high abundance off Cape Hatteras, where the Gulf Stream flows near the continental shelf. In April and May, as inshore shelf waters warm, makos appear to move north following the Gulf Stream and begin moving onto the continental shelf between Cape Hatteras and the southern part of Georges Bank. From June through October, makos are caught on the continental shelf between Cape Hatteras and Cape Cod. The continental shelf south of Cape Cod may be the primary feeding grounds for a large part of the juvenile mako population in the western North Atlantic (Kohler et al, 2002). During late fall and early winter (November-December), makos move from the area between Cape Hatteras and the Grand Banks to offshore wintering grounds in the Gulf Stream and the Sargasso Sea. If it is assumed that 18 C Sargasso Sea water represents the preferred habitat for makos, then the core of their distribution in the western North Atlantic covers a latitudinal range between 20 and 40 N, bordered by the Mid-Atlantic Ridge on the east and the Gulf Stream on the west (similar in configuration to the SSA Area see Figure 1). Mako sharks do occur outside of these boundaries at different seasons, and they make trans-atlantic crossings. However, most of the recaptures can be explained based on a Sargasso Sea Hypothesis, including those returns from the Caribbean Sea. The distribution of recaptures suggests that the principal wintering grounds of juvenile makos are the western margin of the Gulf Stream and the northern part of the Sargasso Sea (Kohler et al, 2002). As shortfin mako have been designated as one of the two major shark species for management focus (ICCAT, 2008), the Sargasso Sea appears to be a very significant area for this species. 2194

179 Blue shark (Prionace glauca) Ecology and oceanography The blue shark is an oceanic-epipelagic and fringe-littoral shark that occurs from the surface to at least 600 m depth (Compagno, 1984). Blue shark movements are strongly influenced by water temperature and this species undergoes seasonal latitudinal migrations on both sides of the North Atlantic. It appears to have a wide thermal tolerance and is caught over a broad range of SST ( C) but seems to prefer water masses from C (Kohler et al, 2002). In general, larger fish of both sexes are caught over a wider temperature range than smaller sharks and blue sharks demonstrate tropical submergence to remain in the deep, cooler waters in the tropical and equatorial parts of their range (Compagno, 1984). In addition to temperature, reproductive condition and availability of prey strongly influence the distribution and movements of blue sharks (Kohler et al, 2002). Documented seasonal migrations to higher latitudes take place on both sides of the North Atlantic. Sizes of blue sharks generally decrease with increasing latitude. Migration and movements Based on tagging data, blue sharks in the North Atlantic make frequent trans-atlantic movements between the western and eastern regions of the Atlantic. They use the major North Atlantic current systems to accomplish these extensive movements. In addition, blue sharks are segregated by sex and size over large areas of the Atlantic with larger, mature fish of both sexes caught in the southern part of their range (Kohler et al, 2002). Immature males and females and sub-adult females dominate the northern regions (Henderson et al, 2001). Evidence from tagging studies and catch data suggest that there are distinct seasonal abundances and seasonal latitudinal migrations in blue sharks. In the western North Atlantic, the winter range of the blue shark is defined as eastward of the northern margin of the Gulf Stream (including the Sargasso Sea) where they can be found during all months of the year. Beginning in April-May, as the shelf waters warm, there is a shoreward movement from the Gulf Stream onto the continental shelf from North Carolina to Newfoundland (Kohler et al, 2002). Female blue sharks ( cm FL; 3-5 years old) arrive on the mating/feeding grounds of the continental shelf in late May where they interact with adult males (4-5 year olds). A summer breeding season for the blue shark is indicated by the high incidence of mating scars and the presence of sperm in the oviducal gland. Once the females are inseminated, they move offshore typically in November and the sperm is stored until the following spring when they fertilize their eggs (Kohler et al, 2002). From May through October, blue sharks are commonly caught on the continental shelf and are also found offshore from the outer edge of the shelf and into the Gulf Stream. According to Kohler et al (2002), most blue sharks along the eastern North American coast begin moving south (southeastern United States, Caribbean Sea) and offshore (central Atlantic) in the fall. Distances traveled for tagged blue shark have ranged up to 6,926 km. This maximum distance moved by a blue shark was from off the coast of New York State to east of Natal, Brazil in 1.4 years (Kohler et al, 2002). Over 75% of the tagged blue sharks traveled less than 1,852 km from their original tagging location with a mean distance of 857 km. Times at liberty ranged from 1 day to 9.1 years. Overall, 75% were at liberty for less than 2 years (Kohler et al, 2002). In a recent study of blue shark migration pathways in the northwest Atlantic using PSAT tags, Campana et al (2011) found that all blue sharks tagged on the continental shelf in the autumn (September-October), moved off the continental shelf to the south and/or east. Distances travelled ranged from 141 to 2,566 km (mean = 927 km). Blue shark movements appeared to be closely linked to the current and temperature structure of the water. After encountering the Gulf Stream and, in the subsequent months extending into the winter, the majority of tagged sharks remained in association with the warm waters of the Gulf Stream or its rings. However, nine of the 23 tagged sharks moved southward into the Sargasso Sea (within the SSA Area) with pop-up times mostly ranging from December to March. This appears to provide firm evidence that blue sharks are using the Sargasso Sea as an overwintering ground (Campana et al, 2011). Porbeagle shark (Lamna nasus) Ecology and oceanography The porbeagle is a coastal and oceanic shark that inhabits the cold temperate waters of the North and South Atlantic (Compagno, 1984). Porbeagle appear to occupy well-defined and relatively constant temperatures 2195

180 throughout the year (Campana et al, 2010a). Based on fishery data, porbeagle were caught at a mean temperature of 7.4 C, with 50% being caught between 5-10 C. There was no significant seasonal pattern of temperature preference indicating that porbeagle actively seek out their preferred temperature range (Campana et al, 2010a). In Canadian waters during the spring, porbeagle were caught most frequently in waters immediately adjacent to the frontal edge separating cool Shelf waters from warmer offshore waters but were not associated with fronts in the fall fishery. Porbeagles are physiologically adapted to live in cold water as they have the ability to conserve metabolic heat and maintain their bodies at considerably warmer internal temperatures than the surrounding water like other members of the family Lamnidae e.g. shortfin mako (Kohler et al, 2002). Body temperatures of 7-10 C above ambient have been recorded for the porbeagle. Migration and movements Porbeagle sharks are confined to the colder parts of the North Atlantic and the distances traveled by conventionally tagged porbeagles are considerably less than other pelagic sharks such as the blue shark and shortfin mako which have a broader thermal tolerance. Distances traveled by tagged porbeagles in the northwest Atlantic ranged from 7.4 to 1,859 km (Kohler et al, 2002). Over 90% of the porbeagles traveled less than 926 km from their original tagging location with a mean distance of 433 km. Times at liberty ranged up to 9.2 years but almost 75% of tagged porbeagles were at liberty for less than 4 years with a mean of 2.7 years (Kohler et al, 2002). On the basis of conventional tagging studies, Porbeagle sharks appeared to be restricted in their distribution to the cold temperate waters of the North Atlantic. However, the results of PSAT tagging of 21 porbeagles off eastern Canada has revealed that males, and immature sharks of both sexes, remained primarily on the continental shelf for periods of almost a year after tagging. However, mature females undertook lengthy migrations of up to 2,356 km through the winter at depths down to 1,360 m beneath the Gulf Stream to a subtropical pupping ground in the Sargasso Sea (Campana et al, 2010 b). These females migrated south from the Canadian Maritimes, mainly through the western part of the Sargasso Sea down to about 20 N (Campana et a,l 2010). In addition to this significant range extension for the species, the location of this critical life history stage in international waters has important implications for the NW porbeagle stock. These international waters are largely unregulated but porbeagle sharks fall under ICCAT management, thus making the protection of this southern portion of the Sargasso Sea of considerable importance for managing this stock. Conclusion This paper has served to illustrate the importance of the Sargasso Sea in different facets of the life history of a number of highly migratory species in the North Atlantic, many of which are commercially important. In addition to the 13 species of teleosts discussed here, there are also three species of pelagic sharks which have been shown to have significant associations with the Sargasso Sea. Given the broad taxonomic spectrum covered by these species groups, it is clear that the Sargasso Sea plays a fundamentally significant role in the life histories of many species in the North Atlantic Ocean. Acknowledgements I wish to acknowledge the support of the Sargasso Sea Alliance (SSA) in the production of this paper and to thank Eric Prince for providing unpublished tagging information. References Beardsley, G.L Age, growth and reproduction of the dolphin Coryphaena hippurus in the Straits of Florida. Copeia 1967: Block, B.A., S. Teo, A. Walli, A. Boustany, M.J. Stokesbury, C.J. Farwell, K.C. Weng.,H. Dewar and T.D. Williams Electronic tagging and population structure of Atlantic bluefin tuna. Nature 434: Campana, S.E, J.F. Gibson, M. Fowler, A. Dorey and W. Joyce a. Population dynamics of porbeagle in the Northwest Atlantic with an assessment of status to 2009 and projections for recovery. Collect.Vol. Sci. Pap. ICCAT, 65(6):

181 Campana, S.E., W. Joyce and M. Fowler b. Subtropical pupping ground for a cold-water shark. Can J Fish Aquat. Sci. 67: Campana, S.E., A. Dorey, M. Fowler,W. Joyce, Z. Wang, D. Wright and I. Yashayaev. 2011, Migration pathways, behavioural thermoregulation and overwintering grounds of blue sharks in the northwest Atlantic. PLoS ONE 6(2), e Casazza, T.L. and S.W. Ross Fishes associated with pelagic Sargassum and open water lacking Sargassum in the Gulf Stream off North Carolina. Fish. Bull. 106: Claro,R., K.C. Lindeman and L.R. Parenti (Eds) Ecology of the marine fishes of Cuba. Smithsonian Press, pp Collette, B.B. and. C.E. Nauen FAO Species Catalogue Vol. 2 Scombrids of the world. An annotated and illustrated catalogue of tunas, mackerels, bonitos and related species known to date, pp Compagno, L.J.V FAO species catalogue. Sharks of the world. An annotated and illustrated catalogue of shark species known to date. Part 1. Hexanchiformes to Lamniformes. FAO Fish Synop. No. 125, Vol. 4, pp Coston-Clements, L., L.R. Settle, D.E. Hoss and F.A. Cross Utilization of the Sargassum habitat by marine invertebrates and vertebrates a review. NOAA Tech. Memo. NMFS-SEFSC-296, pp. 24. Doray, M., B. Stequert and M.Taquet Age and growth of blackfin tuna (Thunnus atlanticus) caught under moored fish aggregating devices, around Martinique Island. Aquat. Living Resour. 17: Fromentin, J.M. and A. Fonteneau Fishing effects and life history traits: a case study comparing tropical versus temperate tunas. Fisheries Research 53: Goodyear, C.P., J. Luo, E.D. Prince and J.E. Serafy Temperature-depth habitat utilization of blue marlin monitored with PSAT tags in the context of simulation modeling of pelagic longline CPUE. Col. Vol. Sci.Pap. ICCAT 59: Graves, J.E., B.E. Luckhurst and E.D. Prince An evaluation of pop-up satellite tags for estimating postrelease survival of blue marlin (Makaira nigricans) from a recreational fishery. Fish. Bull. 100: Henderson, A.C., K. Flannery and J. Dunne Observations on the biology and ecology of blue shark in the North-east Atlantic. J. Fish Biol. 58: ICCAT Report of the ICCAT shark stock assessments meeting, SCRS/2008/017, 89 p. ICCAT. 2010a. ICCAT Manual, chapter yellowfin tuna, 26 p. ICCAT. 2010b. ICCAT Manual, chapter albacore tuna, 18 p. ICCAT. 2010c. ICCAT Manual, chapter bigeye tuna, 25 p. ICCAT. 2010d. ICCAT Manual, chapter bluefin tuna, 19 p. ICCAT. 2010e. ICCAT Manual, chapter skipjack tuna, 22 p. ICCAT. 2010f. ICCAT Manual, chapter swordfish, 18 p. ICCAT. 2010g. ICCAT Manual, chapter blue marlin, 15 p.. ICCAT. 2010h. ICCAT Manual, chapter sailfish, 14 p. ICCAT. 2010i. ICCAT Manual, chapter Atlantic Black Skipjack, 8 p. ICCAT, ICCAT Commission Resolution on the Sargasso Sea. 2197

182 ICCAT ICCAT Manual, chapter white marlin, 15 p. Kerstetter, D.W.,B.E. Luckhurst, E.D. Prince and J.E. Graves Use of pop-up satellite archival tags to demonstrate survival of blue marlin (Makaira nigricans) released from pelagic longline gear. Fish. Bull. 101: Kohler, N.E., P.A. Turner, J.J. Hoey, L.J. Natanson and R. Briggs Tag and recapture data for three pelagic shark species: Blue shark (Prionace glauca), Shortfin mako (Isurus oxyrinchus) and Porbeagle (Lamna nasus) in the North Atlantic Ocean. Col. Vol. Sci. Pap. ICCAT 54: Laffoley, D., H.S.J. Roe et al The protection and management of the Sargasso Sea: The golden rainforest of the Atlantic ocean. Summary science and supporting evidence case. Sargasso Sea Alliance, pp. 71. Loefer, J.K., G.R. Sedberry and J.C. McGovern Vertical movements of a shortfin mako in the western North Atlantic as determined by pop-up satellite tagging. Southeastern Naturalist 4: Luckhurst, B.E Analysis of Bermuda s Marlin Fishery - Catches of blue marlin (Makaira nigricans) and white marlin (Tetrapterus albidus) during the period with comments on tagging and tournament fishing effort. In: Report of Third ICCAT Billfish Workshop, Madrid, Spain, pp Luckhurst, B.E. and T.Trott Bermuda s commercial line fishery for wahoo and dolphinfish: landings, seasonality and catch per unit effort trends. Proc. Gulf Caribb. Fish. Inst. 51: Luckhurst, B.E.,T. Trott and S. Manuel Landings, seasonality, catch per unit effort and tag-recapture results of yellowfin tuna (Thunnus albacares) and blackfin tuna (T. atlanticus) at Bermuda. In: G.R. Sedberry, Ed. Island in the Stream: Oceanography and Fisheries of the Charleston Bump. American Fisheries Society Symposium 25: Luckhurst, B.E., E.D. Prince, J.K. Llopiz, D. Snodgrass and E.B. Brothers Evidence of blue marlin (Makaira nigricans) spawning in Bermuda waters and elevated mercury levels in large specimens. Bull. Mar. Sci. 79: Luckhurst, B.E Large pelagic fishes in the wider Caribbean and Northwest Atlantic Ocean: Movement patterns determined from conventional and electronic tagging. Gulf and Caribbean Research 19: Lutcavage, M.E., R.W. Brill, G.B. Skomal, B.C. Chase and P.W. Howey Results of pop-up satellite tagging of spawning size class fish in the Gulf of Maine: do North Atlantic bluefin tuna spawn in the mid- Atlantic? Can. J. Fish. Aquat. Sci.56: Manooch, III, C.S Fisherman s guide fishes of the southeastern United States.North Carolina State Museum of Natural History, Raleigh, pp Neilson, J.D., S. Smith, F. Royer, S.D. Paul, J.M. Porter and M.Lutcavage Investigations of horizontal movements of Atlantic swordfish using pop-up statellite archival tags. In: Neilson J.D., Arrizabalaga H., Fragoso N, Hobday A, Lutcavage M, Sibert J, Eds. Tagging and tracking of marine animals with electronic devices. New York, Springer, p NMFS Atlantic White Marlin Status Review. Report to National Marine Fisheries Service, SEFSC, Miami. Ortiz, M., E.D. Prince, J.E. Serafy, D.B. Holts, K.B. Davy, J.G. Pepperell, M.B. Lowry and J.C. Holdsworth A global overview of the major constituent-based billfish tagging programs and their results since Proceedings of the Third International Billfish Symposium. Marine and Freshwater Research 54: Oxenford, H.A. and W. Hunte. 1986, A preliminary investigation of the stock structure of the dolphin, Coryphaena hippurus, in the western central Atlantic. Fish. Bull. 84: Oxenford, H.A., R. Mahon and W.Hunte Distribution and relative abundance of flyingfish (Exocoetidae) in the eastern Caribbean. I. Adults. Mar. Ecol. Prog. Ser. 117:

183 Oxenford, H.A Biology of the dolphinfish (Coryphaena hippurus) in the western central Atlantic: a review. Scientia Marina 63: Oxenford, H.A., P.A. Murray and B.E. Luckhurst The biology of wahoo (Acanthocybium solandri) in the western central Atlantic. Gulf and Caribbean Research 15: Palko, B.J., G.L. Beardsley and W.J. Richards Synopsis of the biological data on dolphin fishes, Coryphaena hippurus Linnaeus and Coryphaena equisetis Linnaeus. NOAA Tech. Rep. NMFS Circ. 443, 38 p. Prince, E.D. and C.P. Goodyear Hypoxia-based habitat compression of tropical pelagic fishes. Fish. Oceanogr. 15: Prince, E.D., J. Luo, C.P. Goodyear, Hoolahan J.P., D. Snodgrass, E.S. Orbesen, J.E. Serafy,M. Ortiz and M.J. Schirripa Ocean scale hypoxia-based habitat compression of Atlantic istiophorid billfishes. Fish. Oceanogr. 19: Ruderhausen, P.J., J.A. Buckel, J. Edwards, D.P. Gannon, C.M. Butler and T.W. Averett Feeding ecology of blue marlins, dolphinfish, yellowfin tuna, and wahoos from the North Atlantic Ocean and comparisons with other oceans. Trans Am Fish Soc. 139: SAFMC Fishery Management Plan for Pelagic Sargassum Habitat of the South Atlantic Region. SAFMC, Charlston, SC. Singh-Renton, S. and J. Renton CFRAMP s large pelagic fish tagging program. Gulf and Caribbean Research 19: Smith-Vaniz, W.F., B.B. Collette and B.E. Luckhurst The Fishes of Bermuda: History, Zoogeography, Annotated Checklist and Identification Keys. American Society of Ichthyologists and Herpetologists, Special Publication No. 4, Allen Press, Lawrence, Kansas, pp Thiesen, T.C. and J.D. Baldwin Movements and depth/temperature distribution of the ecothermic scombrid, Acanthocybium solandri (wahoo), in the western North Atlantic. Mar. Biol. 159: Wells, R.J.D. and J.R. Rooker Spatial and temporal patterns of habitat use by fishes associated with Sargassum mats in the northwestern Gulf of Mexico. Bull. Mar. Sci.74: Wilson, S.G.,M.E. Lutcavage, R.W. Brill, M.P. Genovese, A.B.Cooper and A.W. Everly Movements of bluefin tuna (Thynnus thynnus) in the northwestern Atlantic Ocean recorded by pop-up satellite archival tags. Mar. Biol. 146: Wilson, S.G. and B.A. Block Habitat use in Atlantic bluefin tuna Thynnus thynnus inferred from diving behavior. Endang. Species Res. 10:

184 Table 1. Reported landings of five principal tuna species by ICCAT by gear type and total landings reported by FAO for Area 31 Western Central Atlantic (WCA) Area / Gear Yellowfin ATW - LL ATW -SURF WCA - FAO Albacore ATN - LL WCA - FAO Bigeye AT - LL WCA - FAO Bluefin ATW - LL ATW (Sport) WCA - FAO Skipjack ATW - LL ATW -SURF WCA - FAO Table 2. Reported landings of swordfish and principal billfishes by ICCAT by gear type and total landings reported by FAO for Area 31 Western Central Atlantic (WCA) Area / Gear Swordfish ATN - All Gear WCA - FAO Blue marlin ATN - All Gear WCA - FAO White marlin ATN - All Gear WCA - FAO Sailfish ATW - All Gear WCA - FAO Not available 2200

185 Table 3. Reported landings of four species of ICCAT Small Tunas category and total landings reported by FAO for Area 31 Western Central Atlantic (WCA) Area Wahoo WCA - FAO AT + MED NW AT Blackfin tuna ATW WCA - FAO Little tunny AT¹ NW AT WCA - FAO Dolphinfish WCA - FAO ¹ Majority of landings from West Africa Table 4. Reported landings of five species of sharks by ICCAT and total landings reported by FAO for Area 31 Western Central Atlantic (WCA) Area Shortfin mako WCA - FAO NW AT Blue shark WCA - FAO NW AT Porbeagle shark ¹ WCA - FAO ² NW AT Bigeye Thresher WCA - FAO N/A NW AT Basking shark WCA - FAO NW AT ¹ Fishery confined to cold, temperate waters; ² Canada suspended fishery in

186 Figure 1. Map of the proposed Sargasso Sea EBSA, including some of the major features that influence overall boundary definition and location (from Laffoley et al., 2012). 2202

187 Figure 2. Movement vectors of three conventionally-tagged Bermuda yellowfin tuna (Thunnus albacares) recaptured outside Bermuda s EEZ demonstrating demographic connectivity with the wider Caribbean and US eastern seaboard. Straight-line distances moved were between 1,000 and 1,300 km (from Luckhurst 2007). 2203

188 Figure 3. Movement vectors of conventionally-tagged swordfish (Xiphius gladius) in the North Atlantic from the CTC database until end of See text for details. Note predominantly north-south movements (from Luckhurst, 2007). 2204

189 Figure 4. PSAT track of swordfish movement across the Sargasso Sea, illustrating diurnal vertical migration to depths of 850 m (from Luckhurst, 2007). 2205

190 Figure 5. Movement vectors of blue marlin (Makaira nigricans) [dark vectors] and swordfish (Xiphius gladius) [pale vectors] tagged with PSATs in the western North Atlantic from (from Luckhurst, 2007). 2206

191 SCRS/2013/209 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) EJECUCIÓN DEL PROGRAMA NACIONAL DE OBSERVADORES A BORDO DE LA FLOTA INDUSTRIAL ATUNERA VENEZOLANA DEL MAR CARIBE Y OCÉANO ATLÁNTICO AÑO 2012 Juan Gassman 1, Carolina Laurent y Jesús H. Marcano SUMMARY While the INSOPESCA began to implement the National Programme of Onboard Observers as from 2011, the first deployments on vessels occurred in The purpose of the programme which is aimed at the Venezuelan industrial fishing float operating in the Caribbean Sea and Atlantic Ocean is to collect information so as to control and establish policies and regulations which ensure sustainable use of fishing resources. The fishing floats that monitor this programme focus on catches of tuna and other highly migratory bycatch species including billfish and sharks, in which different fishing gear is used such as purse seines, longlines, and pole and lines. In 2012, under the programme, 28 fishing trips took place with a total of 735 days onboard, representing a global total coverage of 5.28% of the total fishing trips and 6.02% of the total number of days at sea. Of them, 19 were by longline vessels which recorded a total of 253 sets in which 195,170 hooks were used and a total of 5,086 individuals were caught; the tuna group were the most representative with 3,042 specimens (59.81%), followed by the marlins group with 1,092 individuals (21.47%), 729 other fish (14.24%), 165 sharks (3.34%) and 58 specimens of swordfish (1.14%). 4 vessels of the purse seine float were boarded on which a total of 78 sets were recorded, 57 (73.08%) of which were successful sets (there was catch). The estimated catch of the purse seine fleet monitored was dominated almost exclusively by the tuna group (99.78%), among which yellowfin tuna (YFT) and skipjack (SKJ) were the most representative with (45.25%) and 184,17 (44.02%) t, respectively. 5 vessels of the baitboat fleet were boarded in which a total of 102 sets were observed involving 1,760 hooks, and a total catch of 14,224 individuals, among which the most signficant species were yellowfin tuna (YFT) with 7,016 specimens (49.33%), followed by skipjack (SKJ) with 5,307 individuals (37.31%), blackfin tuna (BLF) with 1,367 individuals (9.61%) and bigeye tuna (BET) with 492 individuals (3.46%). In each of these trips, information was collected on catches, effort, discards, bycatches of mammals, sea turtles and seabirds, fishing areas, as well as distribution and biological information of the different species caught, among other aspects. RÉSUMÉ A partir de 2011, INSOPESCA lance l'exécution du Programme national d'observateurs à bord mais ce n'est qu'en 2012 que démarrent les premiers arraisonnements d'embarcations. Le Programme cible la flottille de pêche industrielle du Venezuela qui opère dans la mer des Caraïbes et dans l'océan Atlantique afin de recueillir des informations qui permettent le contrôle et l'établissement de politiques et de normes qui garantissent la capture soutenable des ressources halieutiques. Les flottilles de pêche couvertes par ce programme ciblent les thonidés et réalisent des prises accessoires d'autres grands migrateurs, tels que les poissons porte épée, les requins, entre autres, en utilisant divers engins de pêche comme les filets de senne, les palangres, les appâts et les lignes. En 2012, le programme a couvert 28 sorties de pêche, avec un total de 735 jours d'embarquement, ce qui représente une couverture totale de 5,28% de toutes les sorties de pêche et 6,02% du nombre total de jours en mer. Sur ce nombre, on compte 19 palangriers qui ont enregistré un total de 253 opérations au cours desquelles hameçons ont été utilisés et spécimens ont été capturés, les thonidés étant les plus représentés avec spécimens (59,81%), suivis par les istiophoridés avec spécimens (21,47%), 729 autres poissons (14,24%), 165 requins (3,34%) et 58 exemplaires d'espadon (1,14%). Au sein de la flottille de senneurs, quatre senneurs ont été arraisonnés et l'on a enregistré un total de 78 coups de senne, dont 57 positifs (73,08%) (les opérations ont débouché sur des captures). La capture estimée de la flottille de senneurs faisant l'objet d'un suivi a été dominée quasi dans sa totalité par les thonidés (99,78%), ventilés comme suit : 1 Ministerio del Poder Popular para la Agricultura y Tierras, Viceministerio de Pesca y Acuicultura, Instituto Socialista de la Pesca y Acuicultura INSOPESCA, Correspondencia del autor: 2207

192 189,35 t d'albacore (45,25%) et 184,17 t de listao (44,02%). Au sein de la flottille de canneurs, cinq embarcations ont été arraisonnées et l'on a observé un total de 102 opérations avec hameçons et une capture totale de spécimens dont l'albacore était l'espèce la plus abondante (7.016 spécimens capturés, soit 49,33%), suivie du listao (5.307 spécimens capturés, soit 37,31%), du thon à nageoires noires (1.367 spécimens capturés, soit 9,61%) et du thon obèse (492 spécimens capturés, soit 3,46%). Dans chacune de ces sorties, on a recueilli l'information relative aux captures, à l'effort, aux rejets, aux captures accidentelles de mammifères, de tortues marines et d'oiseaux de mer, aux zones de pêche, tout comme à la distribution et l'information biologique des différentes espèces capturées, entre autres. RESUMEN A partir del año 2011, el INSOPESCA inicia la ejecución del Programa Nacional de Observadores a bordo, aunque es en el año 2012 que se inician los primeros abordajes de embarcaciones. El Programa está dirigido a la flota pesquera industrial venezolana que opera en el mar Caribe y océano Atlántico, a fin de recopilar información que permita el control y establecimiento de políticas y normas que garanticen el aprovechamiento sustentable de los recursos pesqueros. Las flotas pesqueras que monitorea este programa están enfocadas en la captura de túnidos y otros peces incidentales altamente migratorios como peces pico, tiburones, entre otros, con el uso de diferentes artes de pesca como redes de cerco, palangres, cañas y cordeles. Durante el año 2012 el programa abordó 28 cruceros de pesca con un total de 735 días a bordo, representando una cobertura total global del 5,28% del total de los viajes de pesca y el 6,02% del número de días totales de mar. De estos, fueron 19 embarcaciones palangreras, donde se registraron un total de 253 lances en los que se emplearon anzuelos y se capturaron en total individuos, siendo el grupo de túnidos el más representativo con ejemplares (59,81%), seguido del grupo de marlines con (21,47%) individuos, 729 otros peces (14,24%), 165 tiburones (3,34%) y 58 ejemplares de pez espada (1,14%). De la flota cerquera se abordaron 4 buques, donde se registró un total de 78 lances, de los cuales 57 (73,08%) fueron lances positivos (hubo captura). La captura estimada de la flota cerquera monitoreada estuvo dominada casi en su totalidad por el grupo de los túnidos (99,78%), entre los cuales el atún aleta amarilla (YFT) y el bonito (SKJ) fueron los más representativos con 189,35 (45,25%) y 184,17 (44,02%) toneladas, respectivamente. En la flota cañera se abordaron 5 embarcaciones donde se observaron un total de 102 lances con anzuelos involucrados, y una captura total de individuos, entre los cuales las especies más importantes fueron el atún aleta amarilla (YFT) con individuos (49,33%), seguido del bonito (SKJ) con individuos (37,31%), el atún aleta negra (BLF) con ejemplares (9,61%) y el atún ojo gordo (BET) con 492 individuos (3,46%). En cada uno de estos cruceros se recopiló información de capturas, esfuerzo, descartes, capturas incidentales de mamíferos, tortugas y aves marinas, áreas de pesca, así como distribución e información biológica de las diferentes especies capturadas, entre otros aspectos KEYWORDS Observadores, Túnidos, Pesca industrial Introducción Las flotas industriales venezolanas que faenan en el Mar Caribe y Océano Atlántico cuya pesca objetivo son los túnidos, se encuentra dividida en tres grandes grupos de acuerdo al tipo de arte de pesca que emplean: la cerquera, la palangrera y la cañera. Estas flotas también capturan de manera incidental, otros peces altamente migratorios como peces de pico, dorados, cabañas, carites, tiburones, entre otros. Por ello, a fin de dar seguimiento sobre éstas pesquerías, desde el año 1987 y bajo el auspicio del convenio ORSTOM (Oficina de Investigación Científica de los Territorios de Alta Mar), actual Instituto de Investigación para el Desarrollo (IRD) de Francia y en conjunción con la Universidad de Oriente, se implementó el Sistema de Bitácoras de la pesquería atunera venezolana (caña, cerco y palangre). Posteriormente, en 1989 y con la finalidad de implementar un programa de evaluación continua de la pesquería venezolana de atún en el Océano Atlántico, se crea la Comisión de pesca de Altura que agrupa varias instituciones: ORSTOM, UDO, SARPA y FONAIAP. Desde entonces además, inicia las actividades el Programa de Investigación Intensiva sobre Marlines de 2208

193 Venezuela (PIIMV) auspiciado por la Comisión Internacional para la Conservación del Atún del Atlántico (CICAA), el cual permanece hasta la actualidad, y en el que se han realizado importantes esfuerzos en el muestreo a bordo de embarcaciones artesanales palangreras, principalmente dirigidas a la captura de marlines, pez espada y atunes. En la actualidad, la información sobre las pesquerías venezolanas de túnidos relacionadas al Caribe - Atlántico, se obtiene bajo la responsabilidad del Instituto Socialista de la Pesca y Acuicultura (INSOPESCA), el Instituto Nacional de Investigaciones Agrícolas (INIA) y el Instituto Oceanográfico de Venezuela de la Universidad de Oriente. En el año 2010, la CICAA emitió la recomendación N 10/10 en la que se establecieron las condiciones mínimas que deben poseer los programas nacionales de observadores a bordo en el área de la Comisión, donde se resalta que debe existir una cobertura mínima del 5% del esfuerzo pesquero, en cada una de las pesquerías de palangre, cerco, cebo vivo y demás, basadas en lances, mareas o número de días. Por otra parte, la Organización de las Naciones Unidas en su resolución 63/112 del año 2009, referida al objeto de lograr la pesca sostenible, alienta a que las organizaciones regionales de ordenación pesquera y los Estados sigan elaborando programas de observación sólidos a fin de mejorar la colecta de datos, entre otras cosas, sobre las especies que se desea pescar y las que son objeto de captura incidental, que también podrían ser útiles para los instrumentos de supervisión, control y vigilancia. Por todo lo anterior, a partir del año 2011, el INSOPESCA, con la finalidad de realizar el seguimiento, monitoreo y evaluación de los recursos aprovechados por nuestras flotas industriales que pescan en aguas del mar Caribe y el océano Atlántico, inicia la ejecución del Proyecto denominado Programa Nacional de Observadores a Bordo de la flota industrial polivalente y atunera del Atlántico (PNOB-INSOPESCA). Las primeras acciones para la puesta en marcha del Programa, inician con el establecimiento de los alcances y el diseño de los instrumentos de colecta de la información, de la base de datos y la capacitación del personal técnico que participará en las campañas de pesca con el rol de Observadores a Bordo. Así, el objetivo del Programa Nacional de Observadores a Bordo de la flota industrial polivalente y atunera del Atlántico, es monitorear al menos un 5% de las campañas o esfuerzo de pesca de la flota y de esta forma colectar información necesaria para determinar el estado de los recursos pesqueros aprovechados, estimar la abundancia de los recursos, cuantificación de descartes, capturas incidentales de peces, mamíferos, aves y tortugas marinas, zonificación de áreas de pesca, rendimiento pesquero de las flotas, selectividad de los artes de pesca, entre otros aspectos, que permitirán la formulación de normas técnicas de ordenamiento que puedan contribuir al aprovechamiento sustentable de estos recursos. Funcionamiento del Programa Nacional de Observadores (PNOB-INSOPESCA) El proceso de abordaje de buques se realiza al azar, y en dependencia de la disponibilidad de observadores al momento del zarpe de las embarcaciones. El programa inició con 15 observadores a bordo, todos con formación en el área (biólogos marinos, ingenieros pesqueros y técnicos en oceanografía y pesca), los cuales fueron capacitados y evaluados antes de ser aceptados en el Programa. Se diseñaron ocho (8) instrumentos de recolección de información diferenciados para cada una de las flotas en base a las especificidades de cada uno de los tipos de artes de pesca, como a continuación se detallan: Registro de embarcación (RE) En esta sección se recopilan datos relacionados a la embarcación e identificación de un crucero específico. En el mismo, se registra información sobre el observador, características del barco, el arte de pesca y los equipos del barco. Cuaderno de pesca (CP) Se registra toda la actividad de navegación y pesca del barco. Se hace énfasis en la posición del buque, cambios en rumbo, velocidad, condiciones del estado del mar, temperatura, así como los avistamientos (aves, mamíferos y tortugas) y demás sucesos específicos de la actividad del barco. Registro de los lances (RLP; RLC; RLCC) Se usa el Registro Lance (sea palangrero, cerquero o cañero) para documentar información específica de los lances efectuados. Se centra principalmente determinar el esfuerzo y captura espacio temporal. 2209

194 Registro de captura de fauna marina (RCFM) El Registro de Captura de Fauna Marina es usado solo en buques cerqueros, para registrar toda la fauna que fue capturada y murió por las actividades pesqueras atunera, para los diferentes tipos de lances: sobre atunes, barcos cañeros, tiburones ballena, mamíferos y objetos flotantes. Muestreo morfométrico de especies (MME) Se usa la planilla de muestreo biológico de especies para anotar la captura de peces. Se considera captura toda aquella que es capturada por el barco, aun cuando ésta sea descartada. Las medidas morfométricas de las diferentes especies son tomadas siguiendo los patrones estándar del manual de campo de la CICAA. La longitud en peces de pico es la distancia desde la mandíbula inferior a la horquilla (MILH), mientras que en los túnidos se mide la longitud a la horquilla (LH). Registro de observaciones de mamíferos marinos (ROMM) Se usa esta sección para registrar información sobre todo mamífero marino (delfines y ballenas) visto durante un viaje, tanto en maniobras de pesca, como en actividades de navegación de la embarcación. Registro de tortugas marinas (RTM) En esta parte se registran datos biológicos de tortugas marinas observadas y también para registrar toda tortuga muerta avistada o capturada de forma incidental. Resumen de crucero (RC) La planilla Resumen del Crucero permitirá documentar información sobre todos los lances y el total de individuos por especie y el total en kilogramos obtenidos al finalizar la jornada de pesca. Actividades realizadas por el Programa en el año 2012 Cobertura Durante el año 2012 se realizó el abordaje y monitoreo de 28 cruceros de pesca, divididos de la siguiente forma: 19 palangreros, 4 cerqueros y 5 cañeros. El número total de días abordo del programa fue de 735, representando una cobertura total global del 5,28% del total de los viajes y el 6,02% del número de días totales de mar en el año 2012 (Tabla 1). - Flota palangrera (LL) La flota palangrera monitoreada estuvo compuesta por 19 embarcaciones industriales de entre 19 a 31 m de eslora, que tienen como objetivo la captura de túnidos. En éstas, se registraron un total de 253 lances (4,6% del total de lances de la flota en el año) en los que se emplearon anzuelos. Está flota utilizó anzuelos (atunero) de números variables entre 3,2 a 5,3, y emplearon como carnada, tanto en condición viva como muerta, sardina (Sardinella aurita), cataco (Trachuruslathami y Selar crumenophthalmus) y en menor medida cachorreta (Scomber japonicus). Las capturas de los cruceros abordados de la flota palangrera (Tabla 2), fueron en total individuos, siendo el grupo de túnidos el más representativo con ejemplares (59,81%), seguido del grupo de marlines con (21,47%) individuos, 729 otros peces (14,24%), 165 tiburones (3,34%) y 58 ejemplares de pez espada (1,14%). Dentro del grupo de los atunes, destaca el atún aleta amarilla (YFT) con un 35,37%, seguido del albacora (ALB) con un 19,33%, el atún aleta negra (BLF) con 2,52% y el ojo gordo (BET) con 1,59%. En los marlines, el pez vela (SAI) fue el más importante con un 14,96%, seguido de la aguja blanca (WHM) con 3,89%, luego el pez lanza (SPF) con 1,42% y las agujas azul (BUM) y corta (SPG) con 0,55% y 0,65%, respectivamente. Dentro del grupo de otros peces el dorado (DOL) con 7,49% y el peto (WHA) con 2,97% fueron los más importantes. Respecto a la captura de tiburones, el tiburón azul (BSH) con el 1,81% y el tiburón bobo (FAL) con 0,39%, fueron los más representativos. Es conveniente señalar que todos estos ejemplares capturados son medidos, y según disponibilidad de la faena, son examinados algunos de sus parámetros como el sexo y la madurez sexual. 2210

195 - Flota cerquera (PS) La flota cerquera monitoreada durante el año 2012 estuvo representada por 4 embarcaciones industriales de cerco de entre 35,9 a 51,5 m de eslora, dirigidas a la captura de atún aleta amarilla principalmente. En éstas, se registraron un total de 78 lances, de los cuales 57 (73,08%) fueron lances positivos en los que hubo captura. De estos lances positivos, la mayor cantidad fueron realizados sobre barcos cañeros (42) los cuales comparten parte de la captura obtenida con dicha embarcación, luego le siguieron los lances sobre tiburones ballena (11), sobre mamíferos (3) y finalmente, sobre palos (1) (Tabla 3). La captura estimada de la flota cerquera monitoreada estuvo dominada casi en su totalidad por el grupo de los túnidos (99,78%), entre los cuales el atún aleta amarilla (YFT) y el bonito (SKJ) fueron los más representativos con 189,35 (45,25%) y 184,17 (44,02%) toneladas respectivamente. Le siguieron el atún aleta negra (BLF) con 25,78 toneladas (6,16%), el atún ojo gordo (BET) con 11,4 toneladas (2,73%) y la cabaña negra (FRI) con 6,76 toneladas (1,61%). La captura incidental fue de 0,933 toneladas que representó apenas un 0,22% de la captura total estimada. La misma estuvo constituida por el pez vela (SAI) (0,007%), la aguja azul (BUM) (0,057%), el peto (WHA) (0,010%), el dorado (DOL) (0,030%), la picúa (GBA) (0,002%) y el tiburón bobo (FAL) (0,102%), además de otros peces cartilaginosos no identificados (0,014%) (Tabla 3). Por otro lado, se registraron los descartes de 0,20 toneladas de bonito (SKJ) en el mes de septiembre y 0,43 toneladas en octubre, por no tener una talla apta para el mercado, además de 0,10 toneladas de aleta amarilla (YFT) y 0,03 toneladas de tiburón bobo (FAL) también en octubre, por no estar en condiciones aptas. - Flota cañera (BB) Durante el año 2012 fueron monitoreados 5 cruceros de pesca de la flota cañera de entre 24,4 a 34,9 m de eslora. En estas embarcaciones controladas se observaron un total de 102 lances y 1760 anzuelos involucrados (Tabla 4). El total de la capturada fue de individuos, entre los cuales la especie más representativa fue el atún aleta amarilla (YFT) con individuos (49,33%), seguido del bonito (SKJ) con individuos (37,31%), el atún aleta negra (BLF) con ejemplares (9,61%), el atún ojo gordo (BET) con 492 individuos (3,46%) y finalmente la cabaña negra (FRI) y el dorado (DOL) con 25 y 17 individuos respectivamente. Distribuciones de tallas de las principales especies capturadas Las tallas de los ejemplares capturados de la especie atún aleta amarilla (YFT) y del ojo gordo (BET) mostraron importantes variaciones para las diferentes flotas monitoreadas, principalmente entre las de los buques palangreros, con las de los cerqueros y cañeros (Figura 1 y 2). Por el contrario, las tallas de individuos capturados de las especies atún aleta negra (BLF) y del bonito (SKJ) fueron muy similares en los buques de las diferentes flotas monitoreadas (Figura 3 y 4). Captura incidental de mamíferos, tortugas y aves marinas Durante los cruceros a bordados en el año 2012, la captura incidental de mamíferos, tortugas y aves, sólo fue observada en la flota palangrera como se detalla a continuación. - Mamíferos Fue registrado un caso de captura incidental de un delfín cuya especie no pudo ser identificada, durante el mes de junio de El mismo quedó enredado en una de las líneas del palangre, la cual fue cortada permitiendo la liberación del individuo vivo. - Tortugas La flota palangrera capturó de manera incidental durante los viajes con observador a bordo un total de seis tortugas marinas. De estas, tres fueron de la especie cardón (DKK), de las cuales una de ellas pudo ser liberada viva, mientras las otras dos fueron devueltas muertas. Además, se capturaron dos individuos de la especie carey (TEI) y un ejemplar de la tortuga verde (TCM), los cuales fueron devueltos vivas al mar. - Aves En cuanto a la interacción de aves con la flota palangrera, fueron capturados y murieron un total de 58 individuos de pardelas del género Puffinus al quedar enganchadas en los anzuelos. 2211

196 Tabla 1. Cobertura en número de días y viajes de pesca del Programa Nacional de Observadores a bordo en el año Tipo de flota Buques activos 2012 Total días de mar Total días abordo PNOB % Total viajes Total viajes PNOB % Palangrera (LL) , ,20 Cañera (BB) , ,20 Cerquera (PS) , ,79 Total , ,28 Tabla 2. Captura (Individuos) y esfuerzo de la flota palangrera monitoreada por el Programa Nacional de Observadores a bordo en el año ESPECIE MAY JUN JUL AGO SEP OCT NOV DIC TOTAL A. Aleta Amarilla (YFT) Albacora (ALB) A. Aleta Negra (BLF) A. Ojo gordo (BET) Bonito (SKJ) Pez vela (SAI) Aguja blanca (WHM) Aguja corta (SPG) Aguja Azul (BUM) Pez Lanza (SPF) Pez espada (SWO) Dorado (DOL) Peto (WHA) Escolar (ESC) Picúa (GBA) Carite ojón (LMA) Carite lucio (KGM) 1 1 Peje rata (RRU) 3 3 Peces Mar (MZZ) Tiburón azul (BSH) Tiburón bobo (FAL) Tiburón oceánico (OCS) Tiburón carite (SMA) 4 4 Tiburón cocodrilo (PCH) Zorro ojón (BTH) 1 1 Tintorera (TIG) 1 1 Cornuda (SPL) 1 1 Peces cart (CAR) TOTAL ESFUERZO N Lances N Anzuelos

197 Tabla 3. Captura estimada (Toneladas) y esfuerzo de la flota cerquera monitoreada por el Programa Nacional de Observadores a bordo en el año ESPECIE AGO SEP OCT NOV DIC TOTAL A. Aleta Amarilla (YFT) 3,40 76,05 67,80 4,10 38,00 189,35 A. Aleta Negra (BLF) 0,10 17,68 8,00 25,78 A. Ojo gordo (BET) 11,40 11,40 Bonito (SKJ) 5,70 24,65 110,32 8,00 35,50 184,17 Pez vela (SAI) 0,03 0,03 Cabaña negra (FRI) 2,55 0,11 1,10 3,00 6,76 Aguja Azul (BUM) 0,17 0,07 0,24 Dorado (DOL) 0,06 0,06 0,01 0,13 Peto (WHA) 0,03 0,02 0,04 Picúa (GBA) 0,01 0,01 Tiburón bobo (FAL) 0,22 0,06 0,08 0,07 0,43 Peces cart (CAR) 0,04 0,03 0,06 TOTAL 9,42 103,43 207,64 13,27 84,63 418,39 ESFUERZO N lances totales N lances positivos Lance cañero Lance tiburón ballena Lance mamíferos Lance sobre palos 1 1 N lances nulos Tabla 4. Captura (Individuos) y esfuerzo de la flota cañera monitoreada por el Programa Nacional de Observadores a bordo en el año ESPECIE AGO SEP OCT NOV TOTAL A. Aleta Amarilla (YFT) A. Aleta Negra (BLF) A. Ojo gordo (BET) Bonito (SKJ) Cabaña negra (FRI) Dorado (DOL) TOTAL ESFUERZO N Lances N Anzuelos

198 Figura 1. Distribución de frecuencias de tallas del atún aleta amarilla (YFT) capturado por embarcaciones palangreras, cerqueras y cañeras monitoreada por el Programa Nacional de Observadores a bordo en el año Figura 2. Distribución de frecuencias de tallas del atún ojo gordo (BET) capturado por embarcaciones palangreras, cerqueras y cañeras monitoreada por el Programa Nacional de Observadores a bordo en el año

199 Figura 3. Distribución de frecuencias de tallas del atún aleta negra (BLF) capturado por embarcaciones palangreras, cerqueras y cañeras monitoreada por el Programa Nacional de Observadores a bordo en el año Figura 4. Distribución de frecuencias de tallas del bonito (SKJ) capturado por embarcaciones palangreras, cerqueras y cañeras monitoreada por el Programa Nacional de Observadores a bordo en el año

200 Figura 5. Distribución de frecuencias de tallas de marlines capturados por embarcaciones palangreras monitoreadas por el Programa Nacional de Observadores a bordo en el año Figura 6. Distribución de los lances de las embarcaciones palangreras, cerqueras y cañeras monitoreadas por el Programa Nacional de Observadores a bordo en el año

201 SCRS/2013/130 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) CAPTURA INCIDENTAL DE TORTUGAS MARINAS POR LAS FLOTAS DE PALANGRE PELÁGICO DE BRASIL Y URUGUAY ( ) Bruno Giffoni 1 ; Nilamon Leite Jr 2 ; Philip Miller 3, 4 ; Maite Pons 3,4 ; Gilberto Sales²; Andrés Domingo 3 SUMMARY Sea turtles are highly migratory species and during their long life cycle they face different threats. One of the biggest threats is the incidental captures by fisheries, thus longline fisheries has been pointed out as a great concern due to the high capture rates. This work summarizes the incidental capture of sea turtles by Brazilian and Uruguayan longline fleet, taking account different aspects to be considered in sea turtles productivity and susceptibility analyses - PSA. The data analyzed were collected by observers of the Programa Nacional de Observadores a Bordo de la Flota Atunera Uruguaya (PNOFA) from Uruguay, and the Programa Nacional de Observadores de Bordo da Frota Pesqueira do Brasil (PROBORDO), Fundação Pró-TAMAR, Instituto ALBATROZ and Núcleo de Educação e Monitoramento Ambiental (NEMA) from Brazil. A total of 25, hooks were sampled by Brazil and Uruguay between 2008 and On average, the sampled effort represented 1,5% of total effort applied at the same area by ICCAT CPCs. Loggerhead was the species most captured (n = 6.594), followed by Leatherback (n = 1.379) and Olive ridley (n = 600). All species nest in Brazil. Regarding productivity aspects, the Brazilian population of loggerhead, leatherback and Olive ridley are increasing. For the susceptibility analyses and considering availability, seven different RMUs are being impacted by longline. As regards encounterability, the three turtle species spent most of the time at the same depth range as Brazilian and Uruguayan longline toward to catch swordfish. For selectivity, juveniles of the loggerhead turtle are the most captured, while for Olive ridley and leatherback, sub-adult and adult have been most captured by longline fleet for both countries. RÉSUMÉ Les tortues marines sont des espèces de grands migrateurs qui connaissent différentes menaces au cours de leur long cycle vital. Les captures accidentelles des pêcheries constituent l'une des plus grandes menaces ; c'est pourquoi, les pêcheries palangrières suscitent de fortes préoccupations en raison de leurs taux élevés de capture. Les présents travaux récapitulent la capture accidentelle des tortues marines réalisée par les flottilles palangrières brésiliennes et uruguayennes, en tenant compte des différents aspects à prendre en compte dans les analyses de productivité et de susceptibilité (PSA) des tortues marines. Les données analysées ont été recueillies par les observateurs du Programme national d'observateurs à bord de la flottille thonière uruguayenne (PNOFA) d'uruguay et du Programme national d'observateurs à bord de la flottille de pêche du Brésil (PROBORDO), la Fondation Pró-TAMAR, l'institut ALBATROZ et Núcleo de Educação e Monitoramento Ambiental (NEMA) du Brésil. Au total, ,089 hameçons ont été échantillonnés par le Brésil et l'uruguay entre 2008 et En moyenne, l'effort échantillonné représentait 1,5% de l'effort total appliqué à la même zone par les CPC de l ICCAT. La tortue caouanne était l'espèce la plus capturée (n = 6.594), suivie de la tortue luth (n = 1.379) et de la tortue olivâtre (n = 600). Toutes ces espèces font leur nid au Brésil. En ce qui concerne les aspects de productivité, la population brésilienne de tortue caouanne, de tortue luth et de tortue olivâtre est en augmentation. Pour les analyses de susceptibilité et compte tenu de la disponibilité, sept différentes unités régionales de gestion (RMU) sont impactées par la palangre. En ce qui concerne la probabilité de rencontre, les trois espèces de tortues ont passé la plupart du temps dans la même gamme de profondeur que la palangre brésilienne et uruguayenne qui ciblaient l'espadon. En ce qui concerne la sélectivité, les juvéniles de tortues couannes sont les plus capturés, tandis que pour la tortue olivâtre et la tortue luth, les sous-adultes et les adultes ont été les plus capturés par la flottille palangrière des deux pays. 1 Fundação Pró-TAMAR (Projeto TAMAR). Rua Antônio Athanazio, 273, Jardim Paula Nobre, Ubatuba SP, Brasil Centro Nacional de Proteção e Pesquisa das Tartarugas Marinhas (Projeto TAMAR), Instituto Chico Mendes de Conservação da Biodiversidade (ICMBIO), R. Andréia 01, Camaçari, BA , Brazil 3 Recursos Pelágicos, Dirección Nacional de Recursos Acuáticos (DINARA). Constituyente 1497, CP Montevideo, Uruguay l 4 CICMAR - Centro de Investigación y Conservación Marina, Uruguay 2217

202 RESUMEN Las tortugas marinas son especies altamente migratorias y durante su largo ciclo vital se enfrentan a diferentes amenazas. Una de las mayores amenazas es la captura incidental en las pesquerías, por ello se ha señalado a las pesquerías de palangre con inquietud debido a las altas tasas de captura. Este trabajo resume la captura incidental de tortugas marinas por parte de la flota de palangre brasileña y uruguaya, teniendo en cuenta diferentes aspectos a considerar en los análisis de productividad y susceptibilidad de tortugas marinas - PSA. Los datos analizados fueron recopilados por el Programa Nacional de Observadores a Bordo de la Flota Atunera Uruguaya (PNOFA) de Uruguay y el Programa Nacional de Observadores de Bordo da Frota Pesqueira do Brasil (PROBORDO), Fundação Pró-TAMAR, Instituto ALBATROZ y Núcleo de Educação e Monitoramento Ambiental (NEMA) de Brasil. Se muestrearon en total ,089 anzuelos por parte de Uruguay y Brasil entre 2008 y De media, el esfuerzo muestreado representaba el 1,5% del esfuerzo total aplicado en la misma zona por las CPC de ICCAT. La tortuga boba era la especie más capturada (n=6,594), seguida por la tortuga laúd (n=1,379) y la tortuga golfina (n=600). Todas las especies anidan en Brasil. Respecto a los aspectos relacionados con la productividad, la población brasileña de tortuga boba, tortuga laúd y tortuga golfina, está creciendo. Para los análisis de susceptibilidad y considerando la disponibilidad, siete RMU diferentes se ven afectadas por el palangre. Respecto a la encontrabilidad, las tres especies de tortugas pasan la mayor parte del tiempo en el mismo rango de profundidad donde la flota de palangre uruguaya y brasileña tratan de capturar pez espada. Respecto a la selectividad, los juveniles de tortuga boba son los más capturados, mientras que los adultos y subadultos de tortuga golfina y tortuga laúd han sido los más capturados por las flotas de palangre de ambos países. KEYWORDS Longline, Brazil, Uruguay, Productivity Susceptibility Analysis, Sea turtles, Southwestern Atlantic Ocean 1. Introducción Las tortugas marinas son especies altamente migratorias que durante su ciclo de vida están expuestas a diferentes amenazas. La captura incidental en las diferentes pesquerías, incluyendo la pesca con palangre de deriva, está considerada como una de las mayores amenazas para su conservación (Wallace et al, 2010a; 2011). En el océano Atlántico Sur occidental se encuentran cinco de las siete especies de tortugas existentes en el mundo, la tortuga cabezona (Caretta caretta), la laúd (Dermochelys coriacea), la verde (Chelonia mydas), la carey (Eretmochelys imbricata) y la olivácea (Lepidochelys olivacea). Todas ellas desovan en Brasil (Marcovaldi et al, 1999) y están amenazadas de extinción de acuerdo a las listas rojas de especies amenazadas de la Unión Internacional para la Conservación de la Naturaleza (IUCN, 2013). Desde que se creó en el año 2003 la Red de Especialistas de Tortugas Marinas del océano Atlántico Sur (Red ASO), investigadores de Brasil y Uruguay vienen trabajando en conjunto, adoptando un enfoque regional en el análisis de la información sobre la captura incidental de las tortugas marinas en las flotas de palangre de ambos países (Domingo et al, 2006; López-Mendilaharsu et al, 2007; Giffoni et al, 2008; Pons et al, 2010). El presente trabajo analiza 12 años de datos ( ) de captura incidental de tortugas marinas de estas flotas, tomando en consideración la información relevante necesaria para los Análisis de Productividad y Susceptibilidad (PSA), a efectos de aportar en el proceso de evaluación que viene desarrollando la Comisión Internacional para la Conservación del Atún Atlántico (CICAA). 2218

203 2. Métodos Los datos analizados fueron obtenidos del Programa Nacional de Observadores de Bordo de la Flota Atunera Uruguaya (PNOFA), del Programa Nacional de Observadores de Bordo da Frota Pesquera do Brasil (PROBORDO), de la Fundação Pró-TAMAR, del Instituto ALBATROZ y del Núcleo de Educação e Monitoramento Ambiental (NEMA). Todos los datos analizados en latitudes superiores a los 20º S son provenientes de las flotas de palangre de Brasil y Uruguay. Los datos obtenidos en latitudes menores a los 20º S provienen exclusivamente de la flota brasilera. Para cada lance de pesca se obtuvo la siguiente información: fecha, posición geográfica (Latitud y Longitud), número de anzuelos, número de tortugas capturadas incidentalmente por especie, entre otras informaciones. La CPUE nominal fue calculada como el número de tortugas marinas capturadas cada 1000 anzuelos. Para el análisis espacial del esfuerzo de pesca y de la captura incidental de tortugas se utilizó el software Arc Map 9.3 considerando cuadrículas de 5º de latitud por 5º de longitud. Para el esfuerzo total de pesca del palangre (nº de anzuelos) de todas las flotas que operan en el área, se utilizaron los datos reportados a la CICAA, por los países, que se encuentran disponibles en la base de datos de la misma (Task II Catch & Effort in Access Data Base [MS Access; version Nov 2012). Para determinar cuánto represento el esfuerzo de las flotas de Brasil y Uruguay en el esfuerzo total, se tomo en consideración los cuadrantes de 5º x 5º donde estas flotas pescaron y se dividió por el esfuerzo total reportado a la CICAA en los mismos cuadrantes y períodos. La información relacionada a la productividad se obtuvo de la literatura existente. Las informaciones relacionadas a susceptibilidad se obtuvieron de la literatura existente de la base de datos del PNOFA (Uruguay) y del Projeto TAMAR (Brasil). La susceptibilidad está relacionada al impacto que la pesca de palangre pelágico puede causar a las poblaciones de tortugas marinas considerando tres aspectos (Cortés et al, 2010): 1) Disponibilidad (Availability): Relaciona la distribución horizontal del esfuerzo de pesca con las áreas de distribución de las tortugas. En este trabajo, también se analiza el esfuerzo y las captura incidentales de tortugas marinas en relación a las Unidades Regionales de Manejo de tortugas (RMUs) identificadas por Wallace et al. (2010b) a nivel global. También se analizaron datos obtenidos de trasmisores satelitales utilizados en tortugas marinas en el ASO. 2) Encontrabilidad (Encounterability): Relaciona la distribución vertical del esfuerzo (profundidad) y el solapamiento con la distribución vertical de las tortugas marinas. Para este análisis se utilizó la información de los trasmisores satelitales y de la literatura existente. 3) Selectividad (Selectivity): Es la capacidad que tiene el arte de pesca de capturar un rango de tallas determinado de tortugas marinas. Para esto se analizo la literatura existente en la región del ASO. Las tortugas carey y verde no fueron consideradas en este trabajo, ya que no se observaron capturas (carey) o estas fueron muy bajas (verde, n=79). Es importante resaltar que los datos de este trabajo provienen de flotas de palangre pelágico con características diferentes y que las mismas no fueron consideradas. 3. Resultados y discusión Entre 1998 y 2010 el esfuerzo de pesca observado por Brasil y Uruguay fue de anzuelos, en tanto que el esfuerzo reportado por la CICAA ( ) para la misma área fue de anzuelos. Esto significa que el esfuerzo observado represento una media del 1,5% de todo el esfuerzo aplicado. El año 2005 fue el que tuvo mayor cobertura con un 6,5% del esfuerzo total, y el año 2000 represento apenas el 0,02% del esfuerzo total (tabla 1). Las flotas pescaron en una amplia zona desde los 7º 25 N hasta 44º 12 S y desde los 54º 33 W hasta los 9º 4 E (fig. 1). Fueron capturadas en el período un total de tortugas marinas. C. caretta fue la especie más capturada representando el 77% (n = 6.594) del total de tortugas capturadas. D. coriacea represento el 16,1% (n = 1.379) y L. olivacea el 6,9% (n = 600). 3.1 Produtividad Las tres principales áreas de desove de C. caretta en Brasil están en los estados de Bahía, Espíritu Santo y Río de Janeiro. Marcovaldi & Chaloupka (2007) analizando 16 años de datos de desove de las dos principales áreas (Bahía e Espíritu Santo) concluyeron que la población de hembras aumento cinco veces entre 1988 y 2003 y la tendencia continúa en aumento (Projeto TAMAR, datos no publicados). 2219

204 Al norte del estado de Espíritu Santo está la única área regular de desove de D. coriacea conocida en Brasil. Es una de las colonias reproductivas más pequeñas en el mundo, sin embargo su población viene aumentando considerablemente. Entre 1985 y 2005 el número de hembras que desovaron en el área paso de 4 a 16 y el número de nidos aumento de 20 a 80 (Thomé et al, 2007). La tendencia de aumento del número de nidos observada por Thomé y colaboradores (2007) se mantiene actualmente (Thomé comunicación personal). A pesar de la pequeña población de Brasil, en el Atlántico se encuentran las mayores poblaciones del mundo de esta especie, particularmente en Gabón, Suriname y Guyana Francesa (Fossette et al, 2008). Algunos trabajos presentan información sobre la captura de hembras provenientes de Gabón, por barcos palangreros brasileros en el área del ASO (Billes et al, 2005), así como información satelital sobre las migraciones de esas hembras al ASO (Fossette e al, 2010; Witt et al, 2011), solapándose la distribución con el área utilizada por las flotas de palangre de Brasil y Uruguay. La principal área de desove de la tortuga olivácea en Brasil es en el estado de Sergipe. Para esa población el número de nidos estimados por temporada de desove aumento de 252 en la temporada a en la temporada , un aumento de aproximadamente 10 veces en once años (Silva et al, 2007). A pesar de que las poblaciones brasileras de las tres especies de tortugas citadas presentan crecimientos muy importantes, hay que considerar que una tortuga marina puede demorar mas de 20 años para llegar a su madurez sexual (Spotila, 2004; Witherington, 2006; Scott et al, 2012). Esto significa que la evaluación del impacto de los individuos juveniles que son muertos hoy, como consecuencia de la captura incidental, se podrá percibir en las áreas de desove dentro de algunos años, cuando las hembras lleguen a la madurez sexual y retornen a las playas a desovar. Para mantener las poblaciones saludables es fundamental que las flotas palangreras de los países que pescan en el Atlántico, especialmente los países miembros de la CICAA, adopten cuanto antes, medidas de mitigación que permitan minimizar la mortalidad de las tortugas marinas capturadas incidentalmente. 3.2 Disponibilidad Wallace et al (2010b) afirman que es necesario conocer la distribución geográfica de las unidades de población o manejo como la de sus amenazas. En este sentido estos autores propusieron para las tortugas marinas 52 Unidades Regionales de Manejo (RMUs). En el presente trabajo verificamos que el esfuerzo de pesca de las flotas de Brasil y Uruguay se solapa a por lo menos siete diferentes RMUs (tabla 2). Para C. caretta se solapo con una RMU (fig 2), para D. coriacea con 4 RMUs (fig 3), incluyendo una al sudeste del océano Índico y para L. olivacea con 2 RMUs (fig 4). Según estos autores (Wallace et al, 2010b), todas las RMUs demuestran tendencias poblacionales crecientes o estables a corto plazo (tabla 2). Se comprobó que en algunos casos, las áreas geográficas definidas para las RMUs son menores que las áreas de distribución definidas por nuestro estudio. Particularmente para C. caretta (fig 2) y L. olivacea (fig 4). Para la tortuga cabezona también pudimos ver que la RMU que se solapa con las flotas brasilera y uruguaya no considero los stock mixtos, identificados anteriormente en trabajos de genética (Caraccio et al, 2008; Reis et al, 2009). Estos autores analizando ADN mitocondrial (DNAm) de individuos juveniles de tortuga cabezona capturados por las flotas de palangre de ambos países e identificaron individuos con haplótipos de: EUA, México, Turquía, además de dos haplótipos del océano Pacífico. Estos datos demuestran que otras RMUs, inclusive de áreas distantes del ASO, están siendo impactadas por las flotas palangreras que operan en esta región y que para algunas RMUs es necesario ampliar el área geográfica definida actualmente. Además para las tortugas cabezonas los estudios de telemetría satelital donde se analiza el uso de hábitat, han mostrado diferentes estrategias de dispersión para las hembras después del desove así como para los individuos juveniles capturados incidentalmente en los barcos palangreros. Las hembras que desovaron en Bahía se movieron hacia el norte, navegando próximas a la costa (Marcovaldi et al, 2010), pasando por aguas someras que no son utilizadas por las flotas de palangre de deriva, mientras que los juveniles utilizan áreas oceánicas las mismas que utilizan las flotas de palangre (Barceló et al, 2013). Esto explica la mayor interacción de los palangreros en la región del ASO con los juveniles de la tortuga cabezona (ver selectividad). 3.3 Encontrabilidad Entre las tortugas marinas la laúd es conocida por realizar las inmersiones más profundas y con mayor duración (Lutcavage & Lutz, 1997), llegando a alcanzar profundidades mayores a los 1000 metros en inmersiones de más de una hora de duración (López-Mendilaharsu et al, 2009a). A pesar de esto, en un estudio realizado en la región del ASO, se observó que la tortuga laúd ocupa la mayor parte de su tiempo en profundidades que varían entre 0 y 70 metros (López-Mendilaharsu et al. 2009b). Esta profundidad coincide con la de los anzuelos que se utilizan en el palangre brasilero y uruguayo que se dirigen a la pesca de pez espada (Xiphias gladius) (Kotas et al, 2005, Domingo et al. 2010). A pesar de que la tortuga laúd se alimenta principalmente de medusas y agua vivas 2220

205 (Eckert et al, 2012), por lo cual no es atraída por las carnadas utilizadas en los anzuelos (calamar y peces), es la segunda especie de tortuga marina más capturada por el palangre pelágico de deriva en el ASO (Domingo et al, 2006; Sales et al, 2008). Probablemente la tortuga laúd se aproxima al palangre a investigar y termina incidentalmente prendida del anzuelo en las aletas o enredada en la línea principal o en las secundarias. Individuos juveniles de tortuga cabezona (n = 27), que fueron capturados con el palangre en el ASO por las flotas uruguayas y brasileras a los cuales se les colocó trasmisores satelitales, pasaron la mayor parte de su tiempo entre los 10 y los 100 metros de profundidad (Barceló et al, 2013). Estudios semejantes realizados en la tortuga olivácea en el Atlántico Norte (Swimmer et al, 2006) demostraron que los individuos de esa especie (n = 14) pasaron la mayor parte del tiempo entre los 0 e 60 metros de profundidad. La franja de profundidad más utilizada por C. caretta y L. olivacea también coincide con las profundidades de actividad de los palangres utilizados por Brasil y Uruguay en las flotas dirigidas al pez espada. 3.4 Seletividad Diversos estudios realizados en el ASO han demostrado que las tortugas cabezonas mas capturadas por las flotas palangreras son las juveniles. Kotas et al (2004) analizando 58 capturas en 3 viajes de pesca obtuvieron un valor medio de la longitud curva del caparazón (LCC) de 58 cm. Ese mismo valor medio de LCC fue el que encontraron Pinedo & Polachek (2004). López-Mendilaharsu et al (2007) analizando el LCC de más de mil C. caretta capturadas incidentalmente por la flotas de Brasil y Uruguay, observaron un valor medio de 57,1 cm con longitudes que variaron entre 38cm y 98cm. Giffoni et al (2008), analizando 1730 tortugas cabezonas en las mismas flotas encontraron un valor medio de LCC ligeramente mayor, 58,9 cm. Sales et al (2008) analizando 614 tortugas cabezonas capturadas por la flota de palangre de Brasil encontraron valores de LCC que variaron entre 39 y 103cm con un valor medio de 58,1 cm. El tamaño medio de la LCC de las hembras de C. caretta que desovan en Brasil es de 103 cm(marcovaldi & Chaloupka, 2007) y la menor hembra registrada en actividad de desove en Brasil media 83cm (Kotas et al, 2004). Por lo tanto a pesar que las flotas de palangre de Brasil y Uruguay capturan individuos juveniles y adultos, los estudios realizados hasta el momento en la región del ASO muestran que los juveniles son el grupo más impactado por estas flotas. Debido al tamaño que alcanzan las tortugas laúd, generalmente estos animales, cuando son capturados incidentalmente, son liberados sin ser subidos a bordo. Algunos trabajos presentan información de talla de los pocos ejemplares que han sido subidos a bordo, capturados por barcos de palangre de deriva en el ASO. En el estudio realizado por Sales y colaboradores (2008), entre las 341 tortugas laúd capturadas incidentalmente, 52 fueron medidas y el LCC vario entre 50 y 194cm, con un valor medio de 127,2 cm. López-Mendilaharsu et al (2007) analizaron 35 individuos medidos entre 115 y 194cm, con un valor medio de LCC de 151,2 cm. Giffoni et al (2008) analizaron el LCC de 24 tortugas laúd, encontrando valores entre 80 y 170 cm con un valor medio de 132,5 cm. Las hembras de D. coriacea que desovan en Brasil tienen una talla de LCC que varía entre 139 y 182cm (n=24) con una media de 159,8 cm (Thomé et al, 2007). Estos estudios indican que los individuos subadultos y adultos son los mas capturados por las flotas de palangre de deriva de Brasil y Uruguay, aunque algunos individuos juveniles también son capturados, principalmente en la región N/NE de Brasil (Sales et al, 2008). Para la tortuga olviácea Sales y colaboradores (2008) analizaron la talla de LCC de 52 tortugas, encontrando un mínimo de 35 cm y un máximo de 80cm (media 52,6). Coluchi y colaboradores (2006) analizando el LCC de 26 individuos de tortugas oliváceas capturadas por la flota palangrera brasilera que opera en la región nordeste de Brasil, encontraron valores de LCC entre 35 y 65 cm con una media de 51,5 cm. Para la población que desova en Sergipe, principal zona de desove en Brasil, el valor medio de LCC es de 73,1 cm (min. = 62,5cm y máx. = 83cm). De acuerdo con estos estudios, los individuos sub- adultos y adultos son los más capturados por las flota de palangre de deriva de Brasil en el ASO. 4. Conclusiones Las flotas de palangre pelágico de Brasil y Uruguay dirigidos a atunes, pez espada y afines, representan una pequeña porción del esfuerzo total del palangre aplicado en el océano Atlántico sur. Para comprender mejor el impacto que las pesquerías de la CICAA pueden causar en las poblaciones de tortugas marinas es necesario que los países miembros (CPCs) envíen los datos de captura incidental de tortugas marinas a la CICAA. Las tres especies más capturadas por las flotas de palangre de deriva de Brasil y Uruguay son: C. caretta, D. coriacea y L. olivácea. 2221

206 Las poblaciones de esas especies que desovan en Brasil están aumentando. Sin embargo, si consideramos que las capturas se realizan no solo sobre individuos adultos, sino sobre sub- adultos y juveniles de las tres especies, es necesario adoptar, lo antes posible, medidas de mitigación que eviten la captura y mortalidad de las mismas. La continuidad de las tendencias poblacionales actuales en las playas de desove de Brasil, dependen en parte de los individuos sub-adultos y juveniles que están siendo capturados por las flotas de palangre en el ASO. Siete RMUs son impactadas por la pesca de palangre pelágico de Brasil y Uruguay, una para C. caretta, 4 para D. coriacea y 2 para L. olivacea. Las áreas geográficas definidas para algunas RMUs (C. caretta y L. olivacea) son menores a las áreas donde se registraron datos de captura incidental, por lo que es necesario revisar y ampliar los límites geográficos de las mismas. Trabajos de genética realizados en el ASO, con individuos juveniles de C. caretta capturados incidentalmente por las flotas de palangre pelágico de Brasil y Uruguay, han demostrado interacciones con otras RMUs, diferentes a las consideradas por Wallace y colaboradores (2010). Esto indica que se estarían impactando RMUs de otras zonas (Mediterráneo, Atlántico Norte, Índico-Pacífico) las cuales deberán ser consideradas en trabajos futuros. Se observó que existe un solapamiento entre la profundidad del arte de pesca de las flotas de Brasil y Uruguay dirigidas al pez espada y el área de preferencia de uso de las tres especies de tortugas marinas analizadas. El palangre de ambas flotas captura principalmente juveniles de C. Caretta y aduktos y sub-adultos de D. coriacea y L. olivacea. Referencias Barceló C., Domingo A., Miller P., Ortega L., Giffoni B., Sales G., McNaughthon L., Marcovaldi M., Heppell S.S. and Swimmer Y (2013). General movement patterns of tracked loggerhead sea turtle (Caretta caretta) in the southwestern Atlantic Ocean. Mar Ecol Prog Ser 479: Caraccio M.N., Domingo A., Márquez A., Naro-Maciel E., Miller P., Pereira A., 2008, Las aguas del Atlántico Sudoccidental y su importancia en el ciclo de vida de la tortuga cabezona (Caretta caretta): evidencias a través del análisis del adnmt. SCRS/2007/124 Col. Vol. Sci. Pap. ICCAT, 62, Coluchi,R (2006). Caracterização da captura incidental de tartarugas marinhas pela pesca de espinhel pelágico no nordeste do Brasil. M.Sc. Thesis, Universidade Federal do Rio Grande do Norte. p:66 Domingo, A., G. Sales, B. Giffoni, P. Miller, M. Laporta and G. Maurutto Captura incidental de tortugas con palangre pelagico en el Atlantico por las flotas de Brasil y Uruguay. Collective Volume of Scientific Papers International Commission for the Conservation of Atlantic Tunas 59: Domingo A., Forselledo R. Pons M. y Ortega L Análisis de la información del atún ojo grande (Thunnus obesus) obtenida por el programa nacional de observadores de Uruguay entre 1998 y Collect. Vol. Sci. Pap. ICCAT, 66(1): Eckert, K.L., B.P. Wallace, J.G. Frazier, S.A. Eckert, and P.C.H. Pritchard Synopsis of the biological data on the leatherback sea turtle (Dermochelys coriacea). U.S. Department of Interior, Fish and Wildlife Service, Biological Technical Publication BTP-R , Washington, D.C. Fossette S, Kelle L, Girondot M, Goverse E and others (2008) The world s largest leatherback rookeries: conservation and research in French Guiana and Gabon. J Exp Mar Biol Ecol 356: Fossette, S., C. Girard, M. López-Mendilaharsu, P. Miller, A. Domingo, D. Evans, L. Kelle, V. Plot, L. Prosdocimi, S. Verhage, P. Gaspar, & J.Y. Georges Atlantic Leatherback Migratory Paths and Temporary Residence Areas. PLoS ONE 5, e Giffoni B, Domingo A, Sales G, Niemeyer-Fiedler F, Miller P (2008) Interacción de tortugas marinas (Caretta caretta y Dermochelys coriacea) con la pesca de palangre pelágico en el Atlántico Sudoccidental: una perspectiva regional para la conservación. Collect. Vol. Sci. Pap. ICCAT, 62: IUCN IUCN Red List of Threatened Species. Version <www.iucnredlist.org>. Downloaded on 21 June Kotas, J.E; Petrere Júnior, M; Azevedo, V.G. de; Santos, S.dos (2005). A pesca de espinhel de superfície no sul do Brasil. In MMA, Programa REVIZEE, Série Documentos REVIZEE - Score Sul. São Paulo: Instituto Oceanográfico, USP. A pesca de emalhe e de espinhel de superfície na região Sudeste-Sul do Brasil. pp:

207 Kotas, J. E., S. dos Santos, V. G. de Azevedo, B. M. G. Gallo, and P. C. R. Barata Incidental capture of loggerhead (Caretta caretta) and Leatherback (Dermochelys coriacea) sea turtles by the pelagic longline fishery off southern Brazil. Fisheries Bulletin 102: López-Mendilaharsu, M., G. Sales, B. Giffoni, P. Miller, F. N. Fiedler, and A. Domingo Distribucion y composicion de tallas de las tortugas marinas (Caretta caretta y Dermochelys coriacea) que interactuan con el palangre pelagico en el Atlantico Sur. Collect. Vol. Sci. Pap. ICCAT, 60: López-Mendilaharsu, M., Rocha, C.F.D., Domingo, A., Wallace, B.P., Miller, P., 2009a. Prolonged deep dives by the leatherback turtle Dermochelys coriacea, pushing their aerobic dive limits. JMBA2-Biodiversity Records No. 6274, pp López-Mendilaharsu, M, C.F.D. Rocha, P. Miller, A. Domingo, & L. Prosdocimi. 2009b. Insights on leatherback turtle movements and high use areas in the Southwest Atlantic Ocean. Journal of Experimental Marine Biology and Ecology 378(1-2): Lutcavage M.E. and Lutz P.L. (1997) Diving physiology. In Lutz P.L. and Musick J.A. (eds) The biology of sea turtles, vol. 1. Boca Raton, FL: CRC Press, pp Marcovaldi MA, Marcovaldi GG Marine turtles of Brazil: the history and structure of Projeto TAMAR- IBAMA. Biological Conservation;91(1):35-41 Marcovaldi M, Chaloupka M (2007) Conservation status of the loggerhead sea turtle in Brazil: an encouraging outlook. Endang Species Res 3: Marcovaldi M, Lopez GG, Soares LS, Lima EHSM, Thomé JCA, Almeida AP (2010) Satellite-tracking of female loggerhead turtles highlights fidelity behavior in northeastern Brazil. Endang Species Res 12: Pinedo, M. C., and T. Polacheck Sea turtle bycatch in pelagic longline sets off southern Brazil. Biological Conservation 119: Pons M, Domingo A, Sales G, Fiedler FN, Miller P, Giffoni B, Ortiz M (2010) Standardization of CPUE of loggerhead sea turtle (Caretta caretta) caught by pelagic longliners in the Southwestern Atlantic Ocean. Aquat Living Resour 23: Reis, E. C.; Soares L. S.; Vargas S. M.; Santos F. R.; Young R. J.; Bjorndal K. A.; Bolten A. B.; Lôbo-Hajdu G. (2009). Genetic composition, population structure and phylogeography of the loggerhead sea turtle: colonization hypothesis for the Brazilian rookeries. Conserv Genet. DOI /s Sales, G., B. Giffoni, and P. C. R. Barata Incidental catch of sea turtles by the Brazilian pelagic longline fishery. Journal of the Marine Biological Association of the United Kingdom. 88: Scott, R; Marsh, R. and Hays, G.C (2012). Life in the really slow lane: loggerhead sea turtles mature late relative to other reptiles. Functional Ecology 2012, 26, Silva ACCD, Castilhos JC, Lopez GG, Barata PCR. Nesting biology and conservation of the olive ridley sea turtle (Lepidochelys olivacea) in Brazil, 1991/1992 to 2002/2003. Journal of the Marine Biological Association of the United Kingdom 2007;87(4):1047e56. Spotila, J.R (2004). Sea turtles: a complete guide to their biology, behavior, and conservation. The Johns Hopkins University Press. p:227 Swimmer, Y., R. Arauz, M. McCracken, L. McNaughton, J. Ballestero, M. Musyl, K. Bigelow, and R. Brill Diving behavior and delayed mortality of olive ridley sea turtles Lepidochelys olivacea after their release from longline fishing gear. Marine Ecology Progress Series 323: Thomé JCA, Baptistotte C, Moreira LMP, Scalfoni JT, Almeida AP, Rieth DB, Barata PCR (2007) Nesting biology and conservation of the leatherback sea turtle (Dermochelys coriacea) in the state of Espírito Santo, Brazil, to Chelonian Conserv Biol 6: Wallace, B. P., R. L. Lewison, S. McDonald; R. T; McDonald; R. K; Bjorkland; S. Kelez; C. Kot; E. M. Finkbeiner; S. Helmbrecht, and L. B. Crowder. 2010a. Global patterns of marine turtle bycatch in fisheries. Conservation Letters 3: Wallace, B. P. et al. 2010b. Regional management units for marine turtles: A novel framework for prioritizing conservation and research across multiple scales. PLoS ONE 5:e Wallace, B. P. et al Global conservation priorities for marine turtles. PLoS ONE 6:e Witt MJ, Bonguno EA, Broderick AC, Coyne MS, Formia A, Gibudi A, Mounguengui GAM, Moussounda C, Nsafou M, Nougessono S, Parnell RJ, Sounguet G-P, Verhage S, Godley BJ (2011). Tracking leatherback turtles from the world s largest rookery: assessing threats across the South Atlantic. Proc R Soc B 278: Witherington, B (2006). Sea turtles: an extraordinary natural history of some uncommon turtles. Voyageur press. p:

208 Tabla 1. Comparación entre el esfuerzo observado (nº de anzelos) y el esfuerzo total reportado (Task II / ICCAT) Year Effort (Br_Uy) Effort (Task II) % sampling , , , , , , , , , , , , TOTAL ,55 Tabla 2. RMUs impactadas por las flotas de palangre pelágico de Brasil y Uruguay. Turtle specie RMU ID Ocean Region Trend Short Trend Long Loggerhead 24 Atlantic Southwest Increasing Increasing Leatherback Olive ridley 51 Atlantic Northwest Increasing Increasing 52 Atlantic Southwest Increasing Increasing 53 Atlantic Southeast Stable Unknown 54 Indian Southeast Stable Stable 2 Atlantic West Increasing Increasing 5 Atlantic East Stable Unknown Figura 1. Distribución del esfuerzo de pesca observado 2224

209 Figura 2. RMU impactada, de las capturas y el área de desove de C. caretta. Figura 3. RMU impactada, distribución de las capturas y el área de desove de D. coriacea. Figura 4. RMU impactada, distribución de las capturas y el área de desove de L.olivacea. 2225

210 SCRS/2013/134 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) ECOLOGICAL RISK ASSESSMENT OF SEA TURTLES TO TUNA FISHING IN THE ICCAT REGION Andrea Angel 1, Ronel Nel 2, Ross M. Wanless 3, Bernice Mellet 2, Linda Harris 2, Ian Wilson 4 SUMMARY Marine turtles spend the majority of their lives at sea; therefore understanding anthropogenic sources of mortality at sea is essential to assess population viability. This Ecological Risk Assessment assesses the risk to turtles from the impacts of tuna fishing in the ICCAT region. We used a Level 2 (semi-quantitative) assessment, within a Productivity-Susceptibility Analysis framework, at the Regional Management Unit (RMU) level; the assessment was hampered by significant data gaps and highly variable bycatch rate estimates. Bycatch rates were scaled to mean annual fishing effort, per RMU. ICCAT longline fishing poses the greater threat to turtles than purse seining. Loggerhead and leatherback turtles potentially encounter the most longline fishing effort (~300 million and >650 million hooks/yr, respectively). The east Atlantic olive ridley, the south Caribbean green turtle and SW Atlantic leatherback turtle RMUs were consistently among the most vulnerable from both gear types. Conversely, the west Atlantic olive ridely turtles showed lowest risk. Regions where turtles are at highest risk included S Caribbean and tropics (20 N-15 S, both gear types), and loggerhead turtles in the Mediterranean (longline only). RÉSUMÉ Les tortues marines passent la plupart de leurs vies en mer ; c'est pourquoi il est fondamental d'appréhender les sources anthropogéniques de la mortalité en mer afin d'évaluer la viabilité des populations. Cette évaluation des risques écologiques (ERA) évalue le risque pour les tortues provenant des impacts des pêcheries de thonidés dans la zone relevant de l'iccat. Nous avons utilisé une évaluation de niveau 2 (semi-quantitative), dans le cadre d'une analyse de productivité-susceptibilité (PSA), au niveau de l'unité de gestion régionale (RMU) ; l'évaluation a été entravée par des lacunes considérables dans les données et par les estimations des taux de prises accessoires fort variables. Les taux des prises accessoires ont été échelonnés à la moyenne de l'effort de pêche annuel, par RMU. La pêche à la palangre au sein de l'iccat constitue une plus grande menace pour les tortues que la pêche à la senne. Les tortues couannes et les tortues luth sont potentiellement confrontées à l'effort de pêche palangrier le plus grand (~300 millions et >650 millions hameçons/an, respectivement). La RMU de la tortue olivâtre de l'atlantique Est, de la tortue verte du Sud des Caraïbes et de la tortue luth de l'atlantique Sud-Ouest se trouvait de manière constante parmi les plus vulnérables avec les deux types d'engins. En revanche, la tortue olivâtre de l'atlantique Ouest était exposée au risque le plus faible. Les régions où les tortues étaient les plus vulnérables incluaient le Sud des Caraïbes et les tropiques (20 N-15 S, deux types d'engins) et pour la tortue caouanne la Méditerranée (palangre seulement) Putney Rd., Kenilworth 7708, South Africa. Corresponding author: 2 Department of Zoology Nelson Mandela Metropolitan University, PO Box 7700, Port Elizabeth 6031, South Africa. 3 Percy FitzPatrick Institute, DST/NRF Centre of Excellence, University of Cape Town, Rondebosch 7701, South Africa Cambier Rd., Kreupelbosch 7945, South Africa

211 RESUMEN Las tortugas marinas pasan la mayor parte de sus vidas en el mar, por tanto, comprender las fuentes antropogénicas de la mortalidad en el mar resulta esencial para evaluar la viabilidad de la población. Esta evaluación del riesgo ecológico evalúa el riesgo para las tortugas procedente del impacto de la pesca de túnidos en la región de ICCAT. Se utilizó una evaluación de nivel 2 (semi-cuantitativa), dentro de un marco de análisis de susceptibilidad-productividad a nivel de unidad de ordenación regional (RMU); la evaluación se vio obstaculizada por las importantes lagunas de datos y por las estimaciones muy variables de tasa de captura fortuita. Se escalaron las tasas de captura fortuita al esfuerzo pesquero medio anual, por RMU. La pesca de palangre de ICCAT supone una amenaza mayor para las tortugas que la pesca de cerco. La tortuga boba y la tortuga laúd son potencialmente las que tienen más encuentros con el esfuerzo pesquero de palangre (~300 millones y >650 millones de anzuelos/año, respectivamente). Las RMU de las tortugas golfinas del Atlántico este, las tortuga verdes del Caribe sur y la tortugas laúd del Atlántico suroccidental fueron de forma continua las más vulnerables para ambos tipos de arte. Por el contrario, las tortugas golfinas del Atlántico oeste presentaban el riesgo más bajo. Las zonas en las que hay más peligro para los tortugas son: el Caribe meridional y tropical (20 N-15 S, ambos tipos de arte), y para las tortugas laúd el Mediterráneo (solo palangre). KEYWORDS By catch, ecological risk assessment (ERA), turtle, regional management unit (RMU), productivity-susceptibility analysis (PSA), vulnerability, gap analysis 1. Introduction There are many reasons behind decreasing numbers of sea turtles. Anthropogenic sources include, inter alia, marine pollution (particularly plastic bags/debris), nesting habitat loss/degradation through tourism and coastal development, boat collisions, the direct take of adults, collecting of eggs by humans, and predation of eggs and hatchlings by invasive mammals on islands (FAO 2004). But because marine turtles spend the majority of their lives in coastal or pelagic waters, any cause of mortality at sea is a critical determinant to population viability and the largest of these is fisheries bycatch (Lewison et al. 2004, Wallace et al. 2011). However, turtle captures are generally an unwanted and unwelcome byproduct of fishing activities (termed bycatch) as evidenced by the large body of work directed at reducing turtle bycatch, particularly in large-scale fisheries (e.g. Gillman et al. 2006, Pacheco et al. 2011). 1.1 An Ecosystem Approach to Fisheries This concept is also referred to as ecosystem-based fisheries management (Pikitch et al. 2004). The global trend towards implementing ecosystem-based approaches to fisheries management, rather than single species or target species-only approaches, has meant that responsibilities for assessing and managing non-target impacts of fishing fall increasingly to fishery management bodies. The Regional Fisheries Management Organisations (RFMOs) that manage tuna and tuna-like species began to acknowledge that responsibility, a move that was marked in the middle of the previous decade by the establishment of subsidiary bodies (known within ICCAT as Subcommittees) with a remit for assessing bycatch and other ecosystem considerations and making recommendations to higher RFMO bodies. 1.2 Tuna fishing and ICCAT Tunas are amongst the highest-value marine species exploited by commercial fisheries for human consumption. The most abundant tunas are also highly migratory, moving between territorial waters of many nations and the high seas. The International Commission for the Conservation of Atlantic Tunas (ICCAT) was established to address the concerns of overfishing and the clear need to manage tuna stocks sustainably. Although a diversity of gear types are used to capture tunas, the two commonest gear types are considered here: longlining and purse seining. Gillnetting and other set-net techniques are known to capture turtles, often at very high rates (e.g. Lewison et al. 2007, Alessandro & Antonello 2010); however, these gear types are seldom used for tunas in the ICCAT region, gillnet effort represents less than 2% and are characterised by being extremely data-poor. At the Subcommittee on Ecosystems and By-catch (SC-ECO) meeting in July 2013, it was agreed that these gear types be disregarded for this analysis. 2227

212 Longlining involves the deployment of a main line with buoys, radio beacons and other devices used to locate the lines at sea once deployed, keep the line at the correct setting depth, etc. As the main line is set, branchlines (typically m long) with baited hooks (and sometimes with weights, lights or other devices attached) are clipped onto the mainline. The lines sink to target fishing depths which can range from <10 m to >500 m. Turtles are captured primarily in two ways through becoming hooked when attempting to prey on baited hooks, or through entanglement when they swim across monofilament branchlines. Purse seining involves setting a vertical wall of nets, with floats on the surface and weights on the bottom of the net to ensure the net hangs vertically from the surface. The net is deployed in a circle around a school of fish, and when encirclement is complete, the weighted, underwater section is pulled closed (or pursed hence the name), and the net with the catch is hauled aboard. Setting is conducted on free-swimming schools and on fish aggregating devices (FADs), which can be anchored or drifting, natural or manmade. Turtles can become entangled in the netting used to construct FADs, where they may die of exposure or drown. They can also be accidentally captured alive during setting, although this seldom results in mortality (Clermont et al. 2012). Other fishing methods, such as pole-and-line, rod-and-reel, trolling, etc. are more highly selective and there is no soak time of gear in the water for extended periods, so any incidental captures can be dealt with immediately. These gear types are not as widely used, catch trivial biomass relative to the two major gear types, are not known to catch or interact with turtles to a significant level, and are not considered in this assessment. 1.3 Ecological Risk Assessment Ecosystem impacts of fishing can be objectively evaluated within an Ecological Risk Assessment (ERA) framework (Arrizabalaga et al. 2011). Determining vulnerability to fisheries is straightforward for data-rich taxa, but much more challenging for bycatch taxa, where data collection is seldom a priority (Ormseth and Spencer 2011). Patrick et al. (2009) modified a previous framework to assess the sensitivity for these data-poor conditions (a PSA, or Productivity-Susceptibility Analysis). The intent of a PSA is to express productivity P, in terms of parameters such as age to maturity and fecundity of the population under investigation, in relation to the likelihood of being caught in a particular fishery (i.e. susceptibility, S). This likelihood is often expressed as the degree of spatial overlap between the fishery and the population, fishing intensity and gear selectivity (Ormseth and Spencer 2011). The ERA with a PSA framework was developed for data poor environments, and can be conducted at one of three levels, depending on data quality; a qualitative Level 1 analysis, a semi-quantitative Level 2 analysis, and a fully quantitative, spatially explicit Level 3 analysis (Hobday et al. 2011, Tuck et al. 2011). A Level 3 ERA requires high-resolution spatio-temporal information on sea turtle bycatch in assessed fisheries (by national fleet or similar categorisation) and by gear type (longline, purse-seine etc.). It is also dependent on detailed population demographic information for sea turtles, preferably with annual census data, annual survival estimates for age/stage/sex classes, and knowledge on the spatio-temporal overlaps between fishing effort and turtle distribution by age/stage/sex classes. Previously, a Level 3 ERA was undertaken for seabirds under the auspices of ICCAT (Tuck et al. 2011). However, productivity and bycatch/susceptibility data for sea turtles in ICCAT (or elsewhere, e.g. Nel et al. 2013) are available only at the resolution or quality sufficient for a Level 2 ERA. Broadly speaking, a Level 2 ERA follows the methodology suggested by Milton (2001), modified by others (Arrizabalaga et al. 2011, Ormseth and Spencer 2011), using indicators of population productivity and susceptibility to capture in different fisheries. This assessment was requested by the SC-ECO of the ICCAT Standing Committee on Research and Statistics (SCRS) to address ICCAT recommendation which requires that the SCRS initiate an assessment of the impact of the incidental catch of sea turtles resulting from ICCAT fisheries. Terms of reference were provided by the ICCAT secretariat to conduct an ERA on the impact of ICCAT fisheries on sea turtles, at the level of Regional Management Units (RMUs) and based on information previously collated. It should be noted that this is a relative assessment of risk. Estimated values should not be treated as firm estimates of likely actual annual mortality or absolute scores of risk. The primary objective is to identify the populations most at risk and regions where threats are likely to be highest, a fundamentally comparative and relative objective. This approach was informed by the quality and quantity of data. Fishing data housed by ICCAT are, for the most, detailed, comprehensive and contain multiple attributes suitable for stratifying analyses or using in predictive models. To stratify impacts or risks to turtles, self-evidently some information of turtle capture rates/probabilities per stratum is required. In some regions turtle interactions with different gear types are documented, which would facilitate detailed analyses of risk relative to season, hook-type, bait-type, setting 2228

213 depth, etc. However, for the great majority of RMUs included in this assessment, and for the majority of fishing fleets, interaction data are simply not available or only at levels of gear type and broad geographical region. Further, for all turtles, the demographic information do not exist that would be required to conduct detailed analyses of risk which account for different catchability/exposure estimates by age, sex or colony within an RMU. As this analysis is the first attempt to assess the risk of all ICCAT fisheries to all turtles, we chose to retain full comparability between regions and RMUs, so as to highlight areas where more observer effort, research or measures to mitigate potentially significant capture can be directed. It will also, hopefully, stimulate further efforts, provision of additional data, and more detailed assessments in future. 1.4 Sea turtles and regional management units There are seven species of sea turtles most of which are distributed in tropical and sub-tropical waters. All are listed as threatened on the IUCN Red list (Table 1). Six species either nest or spend part of their life cycle in Atlantic Ocean. The concept of RMUs was developed by Wallace et al. (2010a), because resolving threats to widely distributed marine megafauna, such as sea turtles requires definition of the geographic distributions of both the threats as well as the population unit(s) of interest. The RMU framework is used as an approach to deal with the threats to sea turtles, which although are widely distributed, exhibit extremely high natal site fidelity and, as shown by genetic analysis, show very little reproductive interaction between regional populations. In total 58 RMUs have been identified globally (Wallace et al. 2010a). It provides a useful tool for inter alia identifying data gaps, evaluating conservation status of sea turtles and providing guidance to management initiatives. We recognized 22 RMUs within the Atlantic Ocean, belonging to six species (Table 2). The Atlantic Ocean is home to some of the largest and some of the smallest RMUs in the world (Wallace et al. 2010a). It hosts the largest populations (RMUs) of the critically endangered leatherback turtle (in Gabon) and the fastest growing RMUs (in the Northwest Atlantic). The east Atlantic populations of both hawksbill and olive ridley turtles, and the leatherback turtles in the south Atlantic (both east and west) are amongst the smallest RMUs in the world and are likely to be the most susceptible to fishing pressures; the South Atlantic leatherback turtles are at risk from high seas fisheries as they undertake trans-atlantic migrations (Marcovaldi et al. 1999) and the hawksbills and olive ridley turtles are at risk from coastal fisheries as they have a more coastal distribution (Wallace et al. 2010a). Kemp s ridley is endemic to the Atlantic Ocean, occurring in the Gulf of Mexico and the east coast of the U.S.A. Its population is moderate in size and its mostly coastal distribution will put it at most risk from coastal fisheries (such as gillnetting). The green and loggerhead turtles of the Atlantic and Mediterranean are generally moderate in size (i.e adult females nesting per annum). 1.5 Atlantic Ocean RMU s Summary Loggerhead turtle (Caretta caretta) 1. Atlantic northwest (Cc-AtNW) The loggerhead turtle nesting population in the northwest Atlantic Florida, North Carolina and Bahamas) consists of approximately nesting females (Ehrhart, Bagley and Redfoot 2003). Depending on the region, this population is considered stable or increasing. Increased human presence at nesting beaches threatens future nesting events mainly because of an increase in artificial lighting changed sand deposition regimes (Antworth et al. 2006). Furthermore, anthropogenic food provision has led to unnaturally high numbers of nest predators (particularly raccoons). Besides incidental catch in fisheries (also dredging), other offshore threats include pollution and power plant entrapment (Brazner and MacMillan 2008). 2. Atlantic northeast (Cc-AtNE) The Cape Verde Islands has a loggerhead turtle nesting population of approximately 1000 individuals (Ehrhart, Bagley and Redfoot 2003). Current trends in the number of nests suggest that this population is decreasing. This is mainly because of habitat destruction (coastal development) as well as high levels of illegal harvesting of eggs and nesting females (Marco et al. 2011). Incidental and targeted catch of sea turtles are the main offshore threats (Bolten et al. 2000, Mejuto 2008). 3. The Mediterranean (Cc-AtMed) The loggerhead turtle nesting population in the Mediterranean spans the coasts of Turkey, Greece and Cyprus and comprises individuals (Broderick et al. 2002). Currently this population is stable, but the destruction of nesting habitat is a looming threat (Margaritoulis 2005). Marine pollution is a major threat to this population. Loggerhead turtles are often caught incidentally in commercial fisheries that operate in this sea and they are vulnerable to boat strikes (Casale et al. 2008). 2229

214 4. Atlantic southwest (Cc-AtSW) The loggerhead turtle nesting population in Brazil is increasing with a mean annual number of 1237 nesting females (Marcovaldi & Chaloupka, 2007). Nesting beaches are under severe pressure from coastal development (Baptistotte et al. 2003). The greatest offshore threat is incidental capture in both coastal and pelagic fisheries (Kotas et al. 2004, Sales et al. 2008). 5. Indian southwest (Cc-InSW) The loggerhead turtle nesting population in the southwest Indian Ocean is shared with Mozambique. This population is showing an increasing trend with a mean annual 371 nesting females (Nel et al. 2013). Because the nesting grounds fall within a world heritage site, harvesting of eggs and nesting females have for the most part been eliminated. Major threats to this population occur offshore in the form of incidental fisheries bycatch (de Wet 2012) Green turtle (Chelonia mydas) 6. Atlantic northwest (Cm-AtNW) The green turtle nesting population in the northwest Atlantic Ocean is increasing. There are an estimated individuals that nest in this region annually (Troeng & Rankin 2005). The major land-based threat in Florida is coastal development whereas Costa Rica has the problem of direct harvesting of eggs and nesting females (Troeng & Rankin 2005). Legal harvest of eggs must remain controlled to ensure the continuous increase of this population. Offshore threats include incidental catch in commercial fisheries as well as various forms of pollution (Beerkircher et al. 2004). 7. Atlantic south Caribbean (Cm-AtScar) In Suriname, there are between 267 and 1816 green turtle females that nest annually (MTSG 2004). This population is considered to be increasing. 8. Atlantic south Central (Cm-AtSCen) The green turtle population is at Ascension Island ( nesting females per annum) is increasing whereas the population at Bioko is considered stable ( ; Radar et al. 2006). Main threats at Bioko include coastal development, destroying nesting habitat and increasing light pollution, as well as the illegal harvesting of eggs and females (Radar et al. 2006). Incidental catch in commercial fisheries and targeted catch by subsistence fishermen (Bioko) are of great concern (Tomas et al. 2010). Further offshore, threats include plastic ingestion and ghost fishing (entanglement in lost/discarded fishing gear) (Tomas et al. 2010)). 9. Atlantic southwest (Cm-AtSW) In the southwest Atlantic (Brazil and associated islands), there are between green turtle females that nest annually (Almeida et al. 2011). This population is considered to be stable, but incidental catch in commercial fisheries an immediate concern (Gallo et al. 2006). 10. Atlantic east (Cm-AtE) There are approximately 2000 females that nest on the beaches of Guinea-Bissau annually and this population is thought to be increasing (Catry et al. 2009). However poaching of eggs and nesting females is still a problem (Catry et al. 2002). Also, nests are vulnerable to density dependent mortality. The main offshore threat is incidental capture in fisheries. 11. Indian southwest (Cm-InSW) There is a very large population of green turtles that nest on islands in the south western Indian Ocean. The number of females that nest annually is increasing at both Europa and Tromelin islands. The main natural source of mortality is egg predation. The main non-natural threats are the fisheries throughout the Mozambique Channel (Bourjea et al. 2008). 12. Mediterranean (Cm-Med) The main nesting activity of green turtles in the Mediterranean occurs along the coastlines of Turkey. This population is declining, with approximately 400 individuals nesting annually. Major threats to this population include incidental catch in fisheries, boat strikes as well as entanglement in marine debris and ghost fishing (Casale 2011) Leatherback turtle (Dermochelys coriacea) 13. Atlantic Northwest (Dc-AtNW) The leatherback turtle nesting population throughout the Caribbean is increasing, with an estimated individuals nesting per annum (Fosette et al. 2008). Terrestrial threats include coastal development and the increase of artificial light pollution (Dutton et al. 2005). In addition to incidental catch in fisheries, other offshore threats include pollution and power plant entrapment. 2230

215 14. Atlantic south, divided into two sub-populations: 14a (Atlantic southeast) and 14b (Atlantic southwest) 14a. Atlantic southeast (Dc-AtSE) This population is considered to be the largest leatherback rookery in the world. It is estimated that there are between and leatherback turtle females that nest along the coast of west Central Africa (Fosette et al. 2008). The trend of this population is currently unknown. Nevertheless, the most significant beach-based threat is illegal egg harvesting (Verhage et al. 2006). Density-dependent nest destruction (where turtles inadvertently dig up other turtle nests because of space constraints) is common. Threats at sea include incidental catch in fisheries, ghost fishing and entanglement in marine debris (Weir et al. 2007, Mejuto et al. 2008). 14b. Atlantic southwest (Dc-AtSW) This is a very small nesting population of leatherback turtles and population trends are largely unknown. However, threats include nest erosion, fisheries bycatch as well as oil-related incidences (Pacheco et al. 2011). 15. Indian southwest (Dc-InSW) Approximately 80 leatherback females nest in the south western Indian Ocean annually (Nel et al. 2013). Although this population is small and stable, it is under the protection authority of a world heritage site. The main beach-based threat to this population is nest erosion, particularly in the face of climate change. Offshore threats largely consist of incidental capture in coastal and pelagic fisheries (de Wet, 2012) Hawksbill turtle (Eretmochelys imbricata) 16. Atlantic west Caribbean (Ei-AtWCar) Approximately 5000 hawksbill turtles nest throughout the Caribbean, including the beaches of Florida (Meylan 1999). Although this population is showing an increasing trend, habitat destruction because of coastal development is a growing concern (Beggs et al. 2007). Bycatch, ghost fishing and entanglement in marine debris are the main threats offshore (Meylan 1999). 17. Atlantic east (Ei-AtE) There is a very small population of hawksbill turtles at Bioko and Guinea-Bissau, averaging 10 nesting females per year (Tomas et al. 2010). This population is under severe pressure from egg collecting. At-sea threats include incidental catch in commercial fisheries as well as targeted catch by subsistence fishermen (Catry et al. 2009). Furthermore, this population is at risk from oil exploration activities (Tomas et al. 2010). 18. Atlantic southwest (Ei-AtSW) The hawksbill turtle population in Brazil and surrounds is increasing ( nesting females per annum; Mortimer and Donnelley 2007). However, coastal development together with light pollution continues to threaten the recovery of this population. The main offshore threats include incidental catch in fisheries and various forms of pollution (Mortimer and Donnelley 2007) Olive ridley turtle (Lepidochelys olivacea) 19. Atlantic west (Lo-AtW) The Olive ridley turtle population in the western Atlantic is increasing, thanks to controlled egg collection and conservation programs. Approximately females nest throughout Brazil, Suriname, Guinea and Venezuela (Kelle et al. 2009). The main terrestrial threat is increasing human presence at nesting beaches, with factors such as coastal development, artificial lighting and vehicles on beaches threatening this population (da Silva et al. 2007). Incidental bycatch in the shrimp trawl fishery is still a concern. 20. Atlantic east (Lo-AtE) The trends of the small (19 43) nesting population of Olive ridley turtles in West Africa are unknown. These turtles are exploited for eggs and meat and their nests are frequently raided by feral dogs (Weir et al. 2007). Offshore threats include incidental catch in commercial fisheries as well as targeted catch by subsistence fishermen (Weir et al. 2007) Kemp s ridley turtle (Lepidochelys kempii) 21. Atlantic northwest (Lk-AtNW) The Kemp s ridley turtle population in the northwest Atlantic is increasing and it is estimated that between 7000 and 8000 females nest here annually (Crowder and Heppel 2011). However, egg poaching is still a problem. Oil exploration, oil pollution and incidental bycatch in shrimp trawl fisheries are the main at-sea threats (Lewison et al. 2003). 2231

216 2. Methods The main aim of this Level 2 ERA for ICCAT fisheries was to estimate likely rates of turtle mortality from fishing operations, per RMU, relate those to each RMU s demographic parameters in a statistical model and estimate the likely relative scale of impacts to each turtle RMU by gear type. Broadly speaking, this approach involved several steps: 1. Identify turtle RMU at-sea distributions 2. Quantify turtle Bycatch Per Unit Effort (BPUE) rates per RMU and per gear type 3. Quantify the level of fishing effort to which each RMU is exposed 4. Extrapolate from BPUE and effort estimates to calculate the number of turtles killed per RMU and per gear type 5. Include mortality estimates in a PSA to estimate the relative risk faced by each RMU in the ICCAT region All data used for the generation of this assessment, including references, have been provided to the ICCAT secretariat as follows: Appendix 1 All raw BPUE data and calculations of estimated total bycatch and mortality per RMU Excel spreadsheet. Appendix 2 Productivity and susceptibility analysis matrix, including productivity and susceptibility parameters and criteria Excel spreadsheet. Appendix 3 Satellite metadata Excel spreadsheet. Folder GIS (Global Information System) containing all files and metadata used to create the effort and RMU overlay maps. 2.1 Turtle biology, bycatch and fisheries effort data An extensive literature review was conducted on turtle biology and demographics, and some 90 references were used to populate the productivity matrix (see Appendix 2 for references used). Where the published literature was older than 10 years, demographic values were confirmed using global databases, in particular the State of the Word's sea Turtles (SWOT), and the most recent IUCN red listing assessments. Extensive consultation was also carried out with turtle experts to refine and expand on published knowledge. Datasets from >2000 satellite tracked individuals in the Atlantic Ocean were obtained from Seaturtle.org (www.seaturtle.org). The RMU boundaries were obtained as shapefiles from SWOT (www.seamap.env.duke.edu/swot) which are freely available online. Data on turtle interactions and capture rates in ICCAT fisheries were provided to the SC-ECO group, in response to a specific call for such data. In addition to other data sourced, a comprehensive review of published and grey literature, on catch rates and related datasets, was undertaken for ICCAT fisheries (Coelho et al. 2012). This covered >170 publications and summarised the major findings in terms of turtle interactions in the ICCAT convention area, by gear type: longline, purse seine and drift nets. The authors also included other non-iccat fisheries in the Atlantic Ocean, such as artisanal surface longlining and shrimp trawl fishery. The bulk of the information, however, relates to longline fishing. Coelho et al. (2012) provide tabulated summaries for all reported turtle bycatch per unit effort (BPUE) data from the northwest and northeast Atlantic (including the mid- Atlantic continental shelf and western Azores), southwest and southeast Atlantic, and the Mediterranean Sea. For purse seine turtle interactions, we found no data other than that reported in Clermont et al. (2012), who reviewed EU purse seine fishery in the Atlantic Ocean (the main purse seine fishery in ICCAT). Clermont et al. (2012) also attempted a scaling up from observed catch to the total purse seine effort and we have therefore used their results in this analysis. We used longline and purse seine effort data provided by ICCAT, stratified by flag, month and 5 x5 and 1 x1 resolution, respectively. Some publications report BPUE by target species or target setting depth, but this was inconsistent and absent entirely from large areas of ICCAT longline effort. Longline effort was stratified by depth for use in the shark ERA done by ICCAT, but that dataset was not made available for this study. Furthermore, there were too few studies reporting turtle BPUE estimates with associated setting depth or target species. As a result we could not account for setting depth or exposure period for the whole of the convention area in this risk assessment. Appendix 1 lists all available BPUE data. 2232

217 Most papers report nominal BPUEs, but where standardized BPUE values were available they were included in preference to nominal data. We discarded all unknown or sea-turtle nei data from analyses. Observer, logbook and scientific experimental data were used in the analysis. The most recent distribution data available was used to assign reported turtle BPUE to a particular RMU. For instance it has been determined, using mtdna sequence analyses, that most juvenile loggerhead turtles caught off the Azores belong to the Northwest Atlantic RMU and probably not to the Northeast Atlantic RMU with which it overlaps, and BPUEs where thus assigned accordingly (Bolten et al. 1998, Bolten 2003). Only when this proved impossible, based on the catch information provided and defined turtle distributions, was a reported turtle BPUE assigned separately to both overlapping RMUs. Some of the reported BPUEs used in this assessment are drawn from studies with overlapping years, and therefore there is a small degree of non-independence; however without access to the raw data bycatch data it was impossible to control for in the analyses. 2.2 Productivity Productivity information was obtained for sea turtles species breeding on beaches facing the Atlantic Ocean (Table 3a). Ten productivity parameters were identified (nine in the original terms of reference, to which we added one), and sufficient information was available for nine of these (Table 3a). Parameters were assigned scores (1-3) with three as the highest productivity. The number of breeding females was used as a proxy for population size, which is standard procedure when dealing with sea turtles because little to no information is available on any other stage classes or male turtles; these values were categorized to assign scores according to Table 3b. The remigration interval refers to the number of years between breeding attempts. 2.3 Susceptibility The susceptibility analysis focussed on the horizontal overlap of turtles and fisheries operations, as well as variables that may influence interactions between turtles and different gear types. Nine parameters were identified for inclusion in the analysis, but data were available only for two of these: spatial overlap of each RMU with the ICCAT region and an estimate of the impact of mortality from bycatch (for two gear types longlining and purse seining). However, these are arguably the two most important parameters to assess risk directly (e.g. Small et al. 2013) Excluded parameters Although the ERA methodology is precautionary and calls for data gaps to be scored high (Hobday et al. 2011), where insufficient data exist for all RMUs, the effect of scoring everything high is the same as excluding it, as it contributes no discriminatory power to the assessment. Seven of the original parameters could not be assessed because data were either not available, or too few to sensibly include in this analysis. Also, the relevance of some parameters varies between gear types. A key deficiency is detailed spatio-temporal turtle distributions. Ideally, raw (or similar) data from individualbased tracking studies with robust sample sizes, from each RMU, preferably with good coverage of the rookeries within each RMU, would be used (c.f. seabird ERA for ICCAT, Tuck et al. (2011) and references therein). However, there is insufficient tracking data for most RMUs. Further, most tracking data are proprietary, and there were insufficient time and resources to establish the collaborative agreements with all (or most of the important) data owners such as would be required to access raw tracking data. This meant that the parameters geographical concentration and temporal overlap with fishing effort could not be assessed. Although some fine-scale information on turtle spatio-temporal habitat use in the ICCAT region is available (e.g. Pons et al. 2010), it is unlikely that sufficient data exist to assess temporal overlap at the scales required for this assessment. A second key deficiency is in the kinds of data collected by fisheries. The parameter vertical overlap relates to the depth at which gear is set, but ICCAT databases do not contain these data. Retention of turtles in gear, length of set (or soak time ) and depth of hooking (longlining, e.g. Chaloupka et al. (2004)) or damage from gear interactions (all gear types) are poorly quantified or poorly reported and we found no useful information to inform these parameters, either in collated data and papers provided by the ICCAT secretariat or from other sources. Soak time is likely to be an important discriminating factor in shallow- and deep-set longlines and although there are localized studies looking at these parameters (e.g. Gilman et al. 2006, Gardner et al. 2008, Lewinson et al. 2013) there is not enough bycatch information data for the ICCAT convention area to stratify our analysis at that level. However, the parameters vertical overlap, retention in gear, length of set and hooking depth/physical damage inform why a particular bycatch/mortality rate occurs, and provide insights into how that rate might be reduced. They could also be used as proxies if a bycatch rate was not available. But in a coarsescale ERA such as this they become less important than estimating likely annual mortality. 2233

218 A gap common to several of these non-assessed parameters is the rate or scale of mortality once a live turtle has been caught, disentangled/removed from gear and released. This is generally referred to as post-release mortality, and is essential to assess fishing impacts properly, but is virtually unknown (Hays 2003, Chaloupka et al. 2004, Swimmer et al. 2006) Leslie matrices and natural mortality A critical component of a PSA is assessing the impacts of fishing relative to natural (and other sources of) mortality. The original approach envisaged for this ERA was to construct RMU-specific Leslie population matrices, which would be used to estimate natural annual mortality and to compare this to estimated mortality from ICCAT fisheries. However, there are no age-structured demographic data for RMUs in the Atlantic Ocean (M. Chaloupka pers. comm., Bjorndal et al. 2011). Thus no age-structured demographic tables can be generated. Indeed very few Leslie matrices or similar stage-based demographic models have ever been produced for sea turtles (Chaloupka & Musick 1997, Crouse et al. 1987, Crowder et al. 1994, Heppell et al. 2005, National Research Council 2010). Several authors have attempted to estimate mortality in sea turtles (Chaloupka & Limpus 2005, Troëng & Chaloupka 2007) but existing models are still not capable of explaining adequately the known patterns from all life history stages (Hamann et al. 2010) Included parameters The overlap parameter was calculated as the number of 2.5 o squares covered by an RMU's distribution, as indicated from satellite tracking studies, as a proportion of all squares (2000) in the ICCAT region (for calculations Appendix 2). Datasets from >2000 tracked individuals were digitised, but due to proprietary data concerns, each individual s tracks were mapped as presence/absence in each square (see Appendix 3 for metadata). Squares were shaded more intensely where tracks from multiple individuals overlapped. Selfevidently the more tracked individuals per RMU, the greater the probability that the complete range of the RMU is described; there is likely to be greater heterogeneity in the distributions than indicated in this exercise (e.g. Pons et al. 2010). For example, areas within the RMU boundaries may not be used, and additional tracking studies are likely to both expand current boundaries and indicate areas where turtles avoid. This parameter was scored low if tracks covered <50 squares (equivalent to ~2.5% of the ICCAT region), medium if >50 but <100 and high if >100 squares (~5% of the ICCAT region; Table 4). The number of tracked individuals varied appreciably between RMUs, from To account for this we added a new parameter the number of tracks used to assess the overlap parameter. This provides a basic indication of the confidence in the score given to each RMU overlap estimate and this parameter was scored as low if the number of tags in the RMU was <5 tags, medium if >5 but < 30 and high if >30 tags for an RMU (Table 4). The parameter bycatch mortality relative to natural mortality requires two things, a bycatch mortality estimate and an estimate of the impact that this will have on an RMU. For the first component (bycatch mortality), all available data sources were mined in an effort to establish estimates of turtle capture rates per gear type and per RMU. This parameter was based largely on reported numbers of turtles captured alive, one of the original susceptibility parameters. Turtles captures alive was not retained as a separate parameter for two reasons. First because it forms a core component of the data used to estimate annual bycatch for both longlining and purse seining (most studies report catch, with relatively few reporting proportion of catch that was dead or died), and scoring a factor in multiple parameters should be avoided. Second because this is only of value when postcapture mortality rates are available, but they are not (Chaloupka et al. 2004). BPUE was calculated as numbers caught per 1000 hooks (longlining) and per set (purse seining) and assigned to RMUs. From this we estimated mean and standard deviations of reported BPUEs for each RMU. RMU boundaries were overlaid onto heat maps of ICCAT longline fishing effort to estimate the likely annual scale of effort to data which each RMU is exposed. To overcome inter-annual variability in fishing effort and to provide a general likelihood of the level of effort which a given RMU encounters annually, we used the 10-year mean of reported fishing effort data from Effort data were mapped at 5 x5 resolution for longline and at 1 x1 for purse seine, using mapping software (see Folder GIS for files used to create all maps). We then scaled the BPUE estimates to total effort to yield estimates of total annual catch per RMU and per gear type. Coelho et al. (2012) also reported data on catch:mortality ratios for longline. In most instances total catch is recorded, but the fate of captured turtles (dead or alive) is reported less frequently. Appendix 1 shows the available catch: mortality ratios for each RMU. Because so many RMUs were data-deficient for this aspect, we assumed that species-specific factors were most likely to influence the relationship between numbers captured and numbers that died. So as not to bias areas with few data we averaged catch: mortality ratios for each species and used those to convert captures into total annual mortalities per RMU. 2234

219 Without Leslie matrices or other models to estimate the relative impact of bycatch mortality on each RMU, we chose to scale the impact of bycatch mortality relative to another demographic parameter. Intra-specific catch was assumed to occur in proportion to abundance of age- and sex-classes, i.e. there is no a priori reason to suggest captures are biased towards males/females or certain age-classes (but see Crowder et al. 1994). Bycatch mortality (assumed to encompass all classes) was compared to the only turtle demographic parameter for which reasonably robust estimates exist numbers of breeding females. This is appropriate because turtles are polygynandrous, so females are the limiting resource for reproduction. Mortality per RMU was converted to an estimate of adult female mortality using the following heuristics: females constitute half the population, and adults constitute 10% of the population. Thus if total mortality was likely to include <5% of the female population, we scored the parameter as 1 (low) for the RMU. For simplicity, rather than perform those calculations for each RMU, they are equivalent to the following scale: if the numerical value of a mortality estimate (from all age, stage and sex classes) is 30% of the adult female population, then 5% of adult females killed. Therefore we scored this parameter as total bycatch relative to the adult female population as follows: 1 30%, 30%<2>100%, or 3>100%. As with Productivity parameters, Susceptibility scores were rated (1-3), however unlike the former, the greatest risk to each RMU was scored a three. Thus small, less-productive RMUs (most vulnerable to fishing impacts) scored lowest on the productivity scale, but the scale of susceptibility worked in the opposite direction. This was addressed when calculating relative vulnerability Weighting parameters Productivity parameters were weighted based on our confidence in the available data and in the degree to which they influence population resilience. Rookery size (either through direct female or track counts) is probably the most reliable and important parameter, and hence was weighted the heaviest (3). The trend data is similarly robust (many of the land-based monitoring programmes have been operating for 10 or more years) and if a population is decreasing, any bycatch mortality will be deleterious, so this was weighted next highest (2). Hatching and emergence success provide direct measures of productivity but are inconsistently reported, so were weighted down (0.5). All other metrics were unweighted. Susceptibility criteria were weighted equally (Table 4) for all gear types assessed Calculating relative vulnerability Both Productivity and Susceptibility scores were summed and rescaled to range from 1-3. Relative vulnerability V, was calculated as V = (P 3) 2 + (S 1) 2 where P is the productivity score for the RMU and S is the susceptibility score for the relevant RMU and gear type. In addition, Euclidean distances between RMU's V provided a quantitative measure of relative vulnerability (Ormseth and Spencer, 2011). 3. Results 3.1 Productivity P scores ranged from with a mean score of 2.0 (Table 5). The hawksbill turtle RMUs had the most extreme distribution within this analysis, with East Atlantic RMU (Ei-AtE) scoring lowest (score = 1.2) but the other two RMUs, from the Caribbean (Ei-AtWCar) and Southwest Atlantic (Ei-AtSW) ranked second with tied scores of Ei-AtE was also noticeable as the only RMU separated from other scores by 0.3 points. This was mostly due to the conservative scoring, due to very low data availability. The most productive (score = 2.6) was the green turtle RMU from the SW Indian (Cm-InSW), which has a distribution relatively marginal to the Atlantic Ocean. Aside from that, four of the six remaining (and Atlantic-specific) green turtle RMUs were in the lower half of the ranked scores = 1.7. Considering species-specific rather than RMU-specific scores, Kemp s ridley scored highest (2.3), but as it constitutes just one RMU it is not an average score. The loggerhead turtle mean score was the only species to average above 2 (mean score = 2.01). 2235

220 3.2 Susceptibility Longlining BPUEs were not reported for five of the 22 RMUs in the Atlantic Ocean, so for the bycatch risk parameter these RMUs were scored 3, the highest risk category. The mean for all RMUs which had at least one BPUE reported, was turtles/1000 hooks. All loggerhead turtle RMUs had BPUEs reported, with the highest mean catch rates from the southwest Atlantic (1.182 turtles/1000 hooks). Loggerheads also had, on average, BPUEs 2-3 times higher than reported means for green or leatherback turtles. The enormous geographic range of leatherback turtles meant that for the three RMUs endemic to the Atlantic Ocean, the mean annual effort to which they are exposed dwarfs that of any other Atlantic RMU. The highest ranked was the Atlantic Northwest RMU (Dc- AtNW), which encountered potentially as much as 270 million longline hooks annually; the other two encountered effort approaching 200 million hooks/year. It's noteworthy that two of three RMUs which straddle the Atlantic and Indian oceans (Cc-InSW and Dc-InSW), and which in the Atlantic Ocean are therefore confined to the SE Atlantic, had reported BPUEs that were an order of magnitude lower than their conspecifics elsewhere in the Atlantic. The third straddling RMU (Cm-InSW) has similarly very low BPUE (0.026 turtles/1000 hooks), which is less than half the next lowest green turtle RMU's BPUE; this pattern of significantly lower BPUEs in the SE Atlantic merits closer investigation. Some possible explanations include that turtle species occur at much lower densities in this region than elsewhere. This is supported by satellite tracking; to date none of the greet turtles tagged in the south Western Indian Ocean has entered the Atlantic Ocean. However, genetic analysis has confirmed a link with the Atlantic Ocean and hence this RMU is included in this analysis (Bourjea et al. 2007). Alternatively there may be differences in foraging behaviour rendering them less susceptible to bycatch, or there is under-reporting of turtle bycatch from this region. When the mean of reported BPUEs per RMU were scaled up to the longline effort to which each RMU is estimated to be exposed, seven of the 16 RMUs for which an estimate could be made had captures in the 10s of thousands annually. Excluding the outliers from the SE Atlantic, the three leatherback turtle RMUs totalled an estimated 165,000 individuals captured on longlines every year (mean ~55,000/RMU/year) and loggerhead turtle RMUs totalled ~200,000 captures/year (mean ~50,000/RMU/year) (Figures 1-6). Reported percentages of captured turtles that died immediately varied considerably but were generally around 5%, i.e. ~5% of reported captures were animals that were dead on hauling or died before being released. Notable exceptions were twofold. First, there were no ratios reported for hawksbill turtles. Second, the two ridley turtle species had much higher immediate mortality rates than other genera (18% for olive ridley turtles and 32% for Kemp s ridley turtles). We note that the reported conversion factors for both ridley species were from single studies with small sample sizes, so these results should be treated with due caution. Total estimated annual turtle mortality in longline fishing under ICCAT auspices, for the 13 RMUs for which estimates could be made, amounted to ~25,000 turtles killed per annum according to the approach and stratifications we adopted here. However, there is little confidence in this figure due to the application of a single BPUE estimate, usually from a range of reported BPUEs, to large aggregations of longline fishing effort data. The values should not be used as firm estimates or taken out of context; their primary value lies in the relative impacts that are likely to accrue to each RMU. Loggerhead turtles constitute almost half of that total (>9,000). These results illustrate the different susceptibilities of species to being caught in longlines. Loggerhead turtles are carnivorous and are likely to be attracted to baited hooks. Both these and leatherback turtles have relatively large populations compared to other turtle species in the Atlantic Ocean, have trans-atlantic crossings following the Gulf Stream, and are therefore extremely likely to interact with high seas longline fisheries. Hawksbill, ridley and green turtles have more coastal distributions, so although they may overlap to some degree with 5 x5 squares in which longline effort occurs, they probably overlap considerably less than the data available for this study suggests. On the other hand, their predominantly near-shore distributions means they are more likely to be impacted by coastal net fisheries, which are not covered in this assessment Purse seining Clermont et al. (2012) reported a total capture of 415 turtles in >9,000 observed purse seine sets between (representing 10% observer coverage). All six species occurring in the Atlantic Ocean were recorded as captured. Of these captures, 21 (5%) were reported as mortalities with another 18 (4%) where the fate was unknown, with no meaningful differences between sets on FADs versus free-schools (Table 6). They then scaled the BPUE to reported effort to estimate an annual bycatch of 218 turtles in the Atlantic Ocean from 1995 to Considering the low observed total bycatch (all species) and mortality we scored all but four RMUs Low 2236

221 (1) as impacts are likely negligible. However, because of uncertainties in RMU sizes, assigning captured turtles to overlapping RMUs, etc. we took a precautionary approach for four RMUs (Dc-AtSW, Dc-InSW, Ei-AtE and Lo-AtE) and raised their scores to 2. It is also worth noting that although Kemp s ridley turtles constituted a significant proportion of the Atlantic Ocean purse seine catch in the Clermont et al. (2012) dataset, the authors noted that the observed effort did not overlap with the known range of Kemp s ridley turtles. The most likely explanation is misidentification of the two very similar-looking Lepidochelys turtle species, with most, if not all, being olive ridely turtles, not Kemp s ridley turtles Vulnerability A plot of the all 22 RMUs impacted by longlining (Figure 7) combining the productivity and susceptibility scores gives a visual interpretation of how each RMUs productivity and susceptibility parameters are related. The RMUs in the top left hand corner are the most vulnerable while those towards the bottom right hand corner are the least vulnerable. The RMUs with the highest susceptibility and lowest productivity scores are the olive ridley turtles from the east Atlantic, green turtles from south Caribbean and Mediterranean loggerhead turtles. These rankings reflect the high levels of longline effort in the central tropical Atlantic Ocean and the fact that the olive ridley turtle population is one of the smallest and most data-deficient RMUs. The least susceptible RMUs are the loggerhead turtles in the Atlantic northwest, and the green and leatherback RMUs of the Indian southwest ocean. Despite the small population size of the latter RMU, overall it has a high productivity which contributes to its low vulnerability. Both the predominantly Indian Ocean RMUs are exposed to relatively low levels of longline effort where they occur in the Atlantic Ocean. Vulnerability scores for longline and purse seine were calculated and ranked (Table 7). Three of the six RMUs most vulnerable to longline fishing were also ranked in the top six for purse seine. The olive ridley turtles from east Atlantic ranked most vulnerable to both gear types; this result is driven primarily by two factors, namely conservative scoring of data-deficient productivity parameters and complete overlap with the highest levels of longline and purse seine effort. This combines with the high ranking in both vulnerability scores for both the green turtles of the south Caribbean the leatherback turtles from the Atlantic south west, to highlight the combined effects of high longlining and purse seine effort levels in the tropical Atlantic Ocean (Figures 8-10). There is also a smaller effort hotspot in the southern Caribbean, off Venezuela. Clearly, these two areas are of highest concern for reducing impacts on turtle species. The remaining two RMUs at high risk from longlining are the loggerhead turtles from the Mediterranean and Kemp s ridley turtles from the Atlantic northwest. The relative impacts of longlining and purse seining are difficult to discriminate in the vulnerability score tables, except through inspection of the mean scores for each analysis (Table 7). Although the mean vulnerability score for longlining was almost 20% higher than for purse seining, the difference in likely impacts is made more evident through an inspection of the mean scores for the highest-ranked RMUs, which was some 40% higher for longlining (mean = 2.26) than for purse seining (mean = 1.39). This relative assessment is, however, probably less powerful than comparing directly the estimated annual mortality from these two gear types, with purse seine fishing accounting for trivial levels of mortality compared to longlining. Considering species-level impacts, five of the six turtle species in the ICCAT region were recorded as being at highest vulnerability for both longline and purse seine PSAs. Thus there do not appear to be species-specific characteristics driving vulnerability. 3.3 Gap analysis This ecological risk assessment was limited by the quality and availability of data, such that only a semiquantitative level 2 analysis was possible. Even within this broad categorisation, many more factors that are believed to inform risk, encounter probabilities (or catchability) and survival probability after encounter, could not be quantified at all. This therefore represents a first step of what is likely to be an iterative ERA process. A recent paper by Lewison at al. (2013) reviewed the current state of knowledge in sea turtle bycatch presenting new ways forward for bycatch research and management. Here we detail data gaps relating to turtle productivity and susceptibility parameters, as well as fishing operational data and reporting, needed to improve future risk assessment for ICCAT fisheries. 2237

222 3.3.1 Productivity parameters Of the original parameters proposed for inclusion, two (generation length and numbers of adult males) were excluded because of a lack of adequate/robust data. However, because the assessment was conducted at the level of the RMU, there were data gaps for virtually all RMUs. Furthermore, many RMUs have a huge number of rookeries, ranging across biogeographic regions, yet for many population productivity parameters we had to rely on a single study from a single site. The biggest data gaps relate to estimates of annual survival for different age/stage classes, for which no data exist. This gap effectively prevents any numerical (as opposed to scenariobased) assessment of population demography and population dynamics. Without these, it is impossible to scale the impacts of fisheries mortality to natural mortality. Although spatio-temporal distribution patterns of turtles relates more to susceptibility to fishing or encounter probabilities with fishing gear, it is an area of research for which turtle biologists and conservationists should continue to focus efforts. The RMU boundaries used here are unlikely to be the true boundaries. More data, and better data availability, will reveal areas of high or low use by different species/rmus that would facilitate far more sensitive and believable estimates of annual fishing effort encountered by RMUs Susceptibility Assessing susceptibility of turtles to fishing effort requires paired information for each fishing effort stratum or parameter, a corresponding datum relating to turtle interactions is required. The data requirements increase rapidly with each additional stratum or parameter. For example, temporal patterns in risk to turtles from longline fishing could be divided by decade, by year, or by season (either summer/winter or quarter-year). While ICCAT catch and effort statistical databases contain data sufficient to create these strata, little is achieved if, in slicing the effort data in this way one merely creates data gaps because new layers have no corresponding turtle interaction data, or one is forced to assume that interaction data from one stratum can be applied to another. Some data were available to conduct temporal assessments of bycatch risk; these were insufficient to merit inclusion in the full PSA, but case studies of what is possible are shown in Box 1 for illustrative purposes. Assessing vertical overlap requires that the actual setting depth be recorded, and can be related to the probabilities that turtles are encountered at the same depths. Self-evidently any air-breathing, diving animal will have a vertical distribution in the water column that is a mix of time spent at the surface, time spent diving down to or surfacing from target/foraging depth, and time spent at target/foraging depths (which will vary considerably). What those depths are (mean, range, etc.) and how they vary by species, season, oceanographic conditions or other parameters is poorly known. How these turtle foraging ecology parameters interact with lines that straddle these categories, and how the catena on longlines increases or decreases encounter probabilities within a single set, or how long hooks take to reach target depths or be retrieved and are brought from depth into the zone where turtles are typically, found remain unknown. Equally, gear configurations and target depths vary tremendously between and within fleets, and data on target species are not recorded through ICCAT or housed in ICCAT s databases, so assigning risk profiles to effort stratified by target species was not possible. ICCAT observer programmes should routinely record the bait and hook types on which a turtle was hooked, if it was hooked in the mouth, foul-hooked or entangled with the line, and if hooked though biting or swallowing the hook, then where/how deep. Several authors have lamented the lack of data on post-release mortality (Hays 2003, Chaloupka et al. 2004, Swimmer et al. 2006) and this remains a critical data gap. Detailed information on operational aspects (preferably recorded through mandatory logbook schemes and reported to ICCAT) will also improve our ability to model fishery impacts on non-target species; metrics for parameters such as longline soak time, target fishing depth and setting/hauling data are obvious gaps. Another key gap is understanding data quality. Self-reported and voluntary data sources on bycatch numbers/rates are inherently less reliable than data from independent, scientific observer programmes. Furthermore, reporting turtle interactions without reference to the observed effort from which those interactions were drawn is of little value. To undertake a more robust level 2 ERA, or to attempt a level 3 ERA for some species/rmus would require turtle bycatch data linked to spatio-temporal effort observed, with corresponding parameters such as bait type, hook type, actual mean hook depth, start and end of fishing (soak time), soak period (day, night) and visibility of baited hooks for nocturnal sets (whether light sticks are used and cloud cover relative to moon phase). Further, turtle species identification, or RMU identity, as well as sex and size-class information of captured turtles is required, preferably based on a simple, reliable and robust measurements that observers can take reliably. In addition, turtle biology and spatio-temporal distribution patterns, by sex/age class, requires appreciably more information. 2238

223 BOX 1 - Case studies of temporal stratification of bycatch and fishing effort data The following two examples illustrate how more detailed bycatch data, such as those collected by observer programmes in the USA, Brazil and Uruguay can give a more detailed and accurate picture of fishing impacts on turtles in the ICCAT region. BPUE was calculated as the sum of all observed turtles caught by quarter, divided by the sum of the observed effort for that quarter, Below are two examples of seasonality in catch rates for turtles from the northwest and from the southwest Atlantic Ocean. We highlight the differences between these regions for loggerhead and leatherback turtles. 3.4 Case 1 - Brazil and Uruguay longline pelagic fleets The observer programmes in Brazil and Uruguay have collected detail information on effort, primarily for longline and turtle bycatch. Domingo et al. (2006) present tabulated data of all observed longline effort and turtle bycatch by species, from 1998 to 2004, grouped by month. We grouped these data into quarters and calculated the BPUE for four species of turtles. Table B1 below shows the aggregated data and highlights periods of highest BPUE for loggerhead and leatherback turtles. Table B1. Seasonal turtle bycatch from longline fisheries in Brazil and Uruguay (combined). Southwest Atlantic Quarter 1 Jan - Mar Quarter 2 Apr Jun Quarter 3 Jul - Sept Quarter 4 Oct - Dec Total years Tot. effort BPUE Tot. effort BPUE Tot. effort BPUE Tot. effort BPUE Tot. effort BPUE Species 422, ,859 1,240,600 1,032,218 3,149,638 Caretta caretta Dermochelys coreacea Chelonia mydas Lepidochelys olivacea Unknown Total bycath all sp Summary results: Table B1 shows that the highest BPUE for all species is in the second quarter (BPUE=1.1986), also the quarter of highest loggerhead and leatherback turtle catches. The lowest BPUE for loggerhead turtles occurred the in fourth quarter. The leatherback turtles are mostly caught during the second and third quarter, BPUE= and respectively, with the lowest BPUE in the fourth quarter (0.0262). Giffoni et al. (2012) analysed 13 years of turtle bycatch data from observer programmes in the Uruguayan and Brazilian pelagic longline fleets. They stratified the data by year and fishing areas, used by the fleet (Latitudes 19 S and 44 S) arriving at similar results for both species as are described here. 2239

224 Case 2 - U.S.A. North Atlantic pelagic longline fleet The observer programme for the U.S.A. pelagic longline fleet has been in place since It documents interaction rates between the fishery and non-target species, including sea turtles. A mandatory fisheries logbook system (FLS) has also been in place since 1992, wherein records of fishing effort, gear characteristics and commercial catch are captured. Table B2 shows results of seasonal turtle catches using observer and longline fishing effort (EffDIS) datasets provided by ICCAT, from 1999 to The effort and bycatch data were stratified by year, month and species. Additional information such hook type and fate of turtles was also available; only hook type data are presented. The data were grouped to allow for a seasonal comparison between the ocean regions (see Appendix 4 for full dataset and calculations). A change in gear from J hooks to circle hooks occurred in The original data includes some captures which were categorised as both circle and j hook captures, and some were not categorised; these are excluded here. Comparisons between hook type and bycatch are not discussed, and shown just for seasonal comparisons. Table B2. Seasonal turtle bycatch from the U.S.A. pelagic longline fleet, Northwest Atlantic Quarter 1 Jan - Mar Quarter 2 Apr - Jun Quarter 3 Jul - Sept Quarter 4 Oct - Dec Total years Hook style Hook style Hook style Hook style Circle J Total Circle J Total Circle J Total Circle J Total Caretta caretta BPUE Dermochelys coreacea BPUE Lepidochelys kempii BPUE Lepidochelys olivacea BPUE Grand Total BPUE (hooks/1000) Summary results: Table B2 shows BPUE for four species of turtles by hook type and quarter. The highest turtle BPUE for all species, irrespective of hook type, occurred in the second and third quarters (BPUE~0.015) with the lowest BPUE (0.001) in the first quarter. Olive and Kemp s ridley turtles where only caught during the second quarter and in the third quarter, respectively. Loggerhead turtles where mostly caught in the third quarter with circle hooks (BPUE =0.0073), and in the first quarter with J hooks (BPUE=0.0034). The highest loggerhead turtle BPUE, irrespective of hook type, was during the third quarter (BPUE=0.0096). The highest leatherback turtle bycatch was during the second quarter for both for circle and J hooks (BPUE= and , respectively). Comparison between data sets Loggerhead turtles in the southwest Atlantic are primarily caught during the second quarter compared to the northwest Atlantic where the majority of the bycatch (on circle hooks) occurs in the third quarter. The leatherback turtles are primarily caught during the second and third quarters in the southwest Atlantic and in the northwest Atlantic, which demonstrates that seasons are not necessarily good predictors for species-specific bycatch risks. Differences in migration periods, breeding seasons or foraging habits probably explain these patterns. This highlights the tremendous spatio-temporal variability in risk within and between species which, when combined with seasonal patterns in fishing effort, gear configurations, etc. can produce more nuanced assessments of risks or total mortality. However, to achieve these more fine-scaled analyses, detailed data collection (bycatch and fisheries information), and collaboration between observer programs, such as between Brazil and Uruguay, is required 2240

225 4. Discussion Fishing effort in ICCAT is dominated by two major gear types, reflected in proportional catches by gear type: purse seine (36%) and longline (33%). These two capture techniques have appreciable non-target capture risks (Wallace et al. 2010b). Further, an analysis of productivity and susceptibility for all bycatch species in the ICCAT convention area revealed that longline fisheries have the highest bycatch species diversity, followed by gillnets and purse seines (Arrizabalaga et al. 2011). Regional (intra-iccat) assessments and estimates of turtle bycatch from longlining have been conducted, including for the North American fleet operating in the North Atlantic (e.g. Levison & Crowder 2007, Gardner et al. 2008, Finkbeiner et al. 2011); the Brazilian and Uruguayan longline fleets in the Southwest Atlantic (e.g. Giffoni et al. 2012, Pons et al. 2012); and the Mediterranean fleets (e.g. Alessandro and Antonello 2010). Further, Clermont et al. (2012) reported on a largescale, multi-year programme assessing interactions between turtles and purse seine fishing in the ICCAT region. The results presented here are the first attempt at assessing the relative impact of the two major ICCAT fisheries on sea turtle populations across the entire Convention Area. A key outcome is that the direct impacts on turtles from purse seine fishing operations appears to be minor in comparison to the impacts from longline fishing. However, the overlapping areas of highly concentrated purse seine and longline effort in the tropical Atlantic Ocean combine to make the low-latitude Atlantic Ocean an area of high risk to turtles. Although significant areas of uncertainty remain, including turtle mortalities on lost FADs, and post-release mortality rate, it is unlikely that improved understanding of these will change the relative impacts by gear type or relative rankings of RMUs. At the species level, the distribution of scores through the ranked tables is largely similar for productivity, longline vulnerability and purse seine vulnerability scores. However, productivity scores were more noticeably grouped by species than either vulnerability scores (compare Tables 5 & 7). Of particular interest are the management implications from a lack of species-specific patterns in vulnerability. All turtle species appear vulnerable to being captured in both longline and purse seine gear, and management measures to mitigate capture will likely benefit all species occurring in the regions in which those mitigation measures are implemented. Turtle bycatch is dependent on many factors. Some are intrinsic to turtle biology such as foraging ecology and behavioural ecology, such as migration (Lewison et. al. 2013), while others relate directly to the method (e.g. type of gear) and strategy of fishing, which vary depending on the target species, oceanographic conditions (e.g. sea temperatures or depth of thermoclines), and fishing traditions and how those intersect with turtle ecology (Schofield et al. 2010). Pelagic longline fishing methods are very diverse but, in the context of turtle interactions, can be broadly divided into shallow vs deep sets depending on the target species. Turtle interactions have been shown to be approximately 10 times greater in longline shallow sets versus deep sets (Beverley et al. 2004, 2009). This is because turtles spend the majority of their time in the upper 100 m of the water column (Eckert et al. 1989; Polovina et al. 2003, 2004; Parker et al., 2005; Swimmer et al. 2006). Thus turtles are exposed to capture potentially for the entire soak time of a shallow longline set, whereas for deep sets they are at risk as the lines sink through the water column during deployment, and possibly again when gear is brought to the surface when it is retrieved (Watson et al. 2005). When targeting swordfish, longliners generally set between sunset and sunrise and are relatively shallow (20 m 30 m), because swordfish have nocturnal, near-surface feeding habits (Berkeley et al. 1981). When targeting tunas, lines are typically set in the morning and hauled in the evening, with hooks set at greater depths (between 10 >500 m). In addition, tuna-directed sets use bait only, whereas swordfish-directed sets use a combination of bait and light sticks. Light sticks have been shown to attract turtles (Witzell 1999, Wang et al. 2007). Leatherback turtles in particular are more often caught in nocturnal longline sets than in comparable diurnal sets (Gilman et al. 2006). The types of hooks used varies between fleets/target species. Broadly speaking there are three kinds of hooks used in pelagic longline fishing: Japanese-style (offset) tuna hooks, circle hooks, and J hooks (typically used when targeting swordfish (Beverly 2006)). J hooks are associated with higher turtle bycatch rates than the other types of hooks and switching from J hooks to circle hooks tends to decrease the severity of hooking. Also the use of larger hooks (> 51 mm) has been shown to reduce the chances of turtles, in particular loggerheads, from ingesting them (Watson et al. 2005, Beverley 2006). These changes have been implemented in several fisheries in particular because the they have resulted in little impact on catch rates (Watson et al. 2005, Gilman et al. 2006, Read 2007, Pacheco et al. 2011, Swimmer et al. 2011). 2241

226 This semi-quantitative, Level 2 ERA has demonstrated first that purse seine fishing poses negligible threats to turtles relative to longline fishing. Second, there are significant losses for several turtle populations from ICCAT longline fishing. Those losses cannot be fully quantified through this assessment, and there is no attempt to apportion any possible extinction risk to mortality from fishing. However there is a prima facie case for ICCAT CPCs to engage in more data collection on turtle interactions with their fishing operations. To be most useful, set-by-set turtle bycatch data and associated fishing operational data should be shared, so that more refined, indepth and detailed analyses of risk can be conducted at the scale of the ocean basin. Such analyses should be geared to account for risk factors more explicitly than was possible here, including seasonality of effort and bycatch rates and risks from deep versus shallow sets. A more detailed Level 2 analysis, or a Level 3 analysis for certain species, are likely required to assess the true likelihood that longline fishing is or is not a significant driver of extinction risk to certain turtle populations in the Atlantic Ocean. Equally, more fine-scale analyses of turtles RMUs (or other population delimitations) at-sea spatio-temporal distributions and densities (and therefore their exposure to longline fishing effort) are urgently required. While many of the datasets for assessing at-sea patterns do exist, they are not readily available for analysis such as this. References Achaval, F., Marin, Y. H., Barea, L. C Captura incidental de tortugas marinas con palangre pelágico oceánico en el Atlántico Sud-occidental. In Arena, G. & Rey, M. (eds). Captura de grandes peces pelágicos (pez espada y atunes) en el Atlántico Sud-occidental, y su interacción con otras poblaciones. INAPE PNUD, Montevideo, Uruguay. Pp Alessandro, L., and Antonello, S An overview of loggerhead sea turtle (Caretta caretta) bycatch and technical mitigation measures in the Mediterranean Sea. Reviews in Fish Biology and Fisheries, 20: Almeida, A., Moreira, L. M., Bruno, S. C., Thomé, J. C. A., Martins, A. S., Bolten, A. B., & Bjorndal, K. A Green turtle nesting on Trindade Island, Brazil: abundance, trends, and biometrics. Antworth, R. L., Pike, D. A., & Stiner, J. C Nesting ecology, current status, and conservation of sea turtles on an uninhabited beach in Florida, USA. Biological Conservation: 130: Arrizabalaga, H., de Bruyn, P., Diaz, G.A., Murua, H., Chavance, P., De Molina, A. D., Chavance, P., Molina, A. D., Gaertner, D., Ariz, J., Ruiz, J. and Kell L. T Productivity and susceptibility analysis for species caught in Atlantic tuna fisheries. Aquatic Living Resources 24, Baptistotte, C., Thome, J. C.A. and Bjorndal, K. A Reproductive Biology and Conservation Status of the Loggerhead Sea Turtle (Caretta caretta) in Espírito Santo State, Brazil. Chelonian Conservation and Biology 4(3): 1-7. Beerkircher, L. R., Lee, D. W., Brown, C. J., & Abercrombie, D. L SEFSC pelagic observer program data summary for US Department of Commerce, National Oceanic and Atmospheric Administration, National Marine Fisheries Service, Southeast Fisheries Science Center. Beggs, J. A., Horrocks, J. A., & Krueger, B. H Increase in hawksbill sea turtle Eretmochelys imbricata nesting in Barbados, West Indies. Endangered Species Research, 3(2), Berkeley, S. A., Irby, E. W., & Jolley, J. W Florida's commercial swordfish Fishery: Longline gear and methods. Marine Advisory Program, Florida Cooperative Extension Service, University of Florida. Florida Sea Grant Marine Advisory Bulletin. MAP-14. Pp 23 Beverly, S Hooks used in longline fishing. In South Pacific Commission (Ed). Fisheries Newsletter 117:45-48 Beverly, S., Robinson, E., & Itano, D Trial setting of deep longline techniques to reduce bycatch and increase targeting of deep-swimming tunas. In 17th Meeting of the Standing Committee on Tuna and Billfish, SCTB17, Majuro, Marshall Islands. Pp Beverly, S., Curran, D., Musyl, M., & Molony, B Effects of eliminating shallow hooks from tuna longline sets on target and non-target species in the Hawaii-based pelagic tuna fishery. Fisheries Research 96: Bjorndal, K., Bowen B., Chaloupka M., Crowder L., Heppell S., Jones C., Lutcavage M., Policansky D., Solow A., Witherington B Better science needed for restoration in the Gulf of Mexico. Science 331: Brazner, J. C., & McMillan, J Loggerhead turtle (Caretta caretta) bycatch in Canadian pelagic longline fisheries: Relative importance in the western North Atlantic and opportunities for mitigation. Fisheries Research 91:

227 Bolten, A.B., K.A. Bjorndal, H.R. Martins, T. Dellinger, M.J. Biscoito, S.E. Encalada, and B.W. Bowen Transatlantic developmental migrations of loggerhead sea turtles demonstrated by mtdna sequence analysis. Ecological Applications 8:1-7. Bolten, A. B., H. R. Martins, and K. A. Bjorndal Workshop to Design an Experiment to Determine the Effects of Longline Gear Modification on Sea Turtle Bycatch Rates. U.S. Dep. Commer., NOAA Tech. Memo. Bolten, A.B Active swimmers passive drifters: the oceanic juvenile stage of loggerheads in the Atlantic system. Pages in A.B. Bolten and B.E. Witherington (editors), Loggerhead Sea Turtles. Smithsonian Institution Press, Washington, D.C. Bourjea, J., Lapegue, S., Gagnevin, L., Broderick, D., Mortimer, J. A., Ciccione, S., Roos, D., Taquet, C., Grizel, H Phylogeography of the green turtle, Chelonia mydas, in the Southwest Indian Ocean. Molecular Ecology 16: Bourjea, J., Nel, R., Jiddawi, N. S., Koonjul, M. S. & Bianchi, G Sea turtle bycatch in the West Indian Ocean: Review, recommendations and research priorities. Western Indian Ocean Journal of Marine Science 7: Broderick, A. C., Glen, F., Godley, B. J., & Hays, G. C Estimating the number of green and loggerhead turtles nesting annually in the Mediterranean. Oryx 36: Casale, P., Abbate, G., Freggi, D., Conte, N., Oliverio, M., & Argano, R Foraging ecology of loggerhead sea turtles Caretta caretta in the central Mediterranean Sea: evidence for a relaxed life history model. Mar Ecol Prog Ser, 372, Casale, P., Mazaris, A. D., & Freggi, D Estimation of age at maturity of loggerhead sea turtles Caretta caretta in the Mediterranean using length-frequency data. Endangered Species Research, 13(2), Catry, P., Barbosa, C., Indjai, B., Almeida, A., Godley, B. J., & Vié, J. C First census of the green turtle at Poilão, Bijagós Archipelago, Guinea-Bissau: the most important nesting colony on the Atlantic coast of Africa. Oryx 36(04), Catry, P., Barbosa, C., Paris, B., Indjai, B., Almeida, A., Limoges, B., Silva, C. & Pereira, H Status, ecology, and conservation of sea turtles in Guinea-Bissau. Journal Information, 8(2). Chaloupka, M., Musick J Age, growth and population dynamics. Chapter 9, pp In Lutz P, Musick J (eds). The Biology of Sea Turtles. CRC Marine Science Series, CRC Press, USA Chaloupka, M., Parker, D., Balazs, G Modelling post-release mortality of loggerhead sea turtles exposed to the Hawaii-based pelagic longline fishery. Marine Ecology Progress Series 280: Chaloupka, M., & Limpus, C Estimates of sex-and age-class-specific survival probabilities for a southern Great Barrier Reef green sea turtle population. Marine Biology 146: Clermont, S., Chavance, P., Delgado De Molina, A., Murua, H., Ruiz, J., Ciccione, S., & Bourjea, J EU purse seine fishery interaction with marine turtles in the Atlantic and Indian Oceans: a 15 years analyses. Report: IOTC 2012-WPEB08-35 Rev_1. Pp 74 Coelho, R., Fernandez-Carvalho, J. and Santos, M.N A review of fisheries within the ICCAT convention area that interact with sea turtles. Unpublished report SCRS/2012/049 to ICCAT Intercessional Working Group on Ecosystems and Bycatch. 12p. Crouse, D.T., Crowder, L.B., Caswell, H A stage-based population model for loggerhead sea turtles and implications for conservation. Ecology 68: Crowder, L., & Heppell, S The decline and rise of a Sea Turtle: how Kemp s ridleys are recovering in the Gulf of Mexico. Solutions, 2(1), Crowder, L.B., Crouse, D.T., Heppell, S.S., Martin, T.H Predicting the impact of turtle excluder devices on loggerhead sea turtle populations. Ecological Applications 4: Da Silva, A. C. C., De Castilhos, J. C., Lopez, G. G., & Barata, P. C Nesting biology and conservation of the olive ridley sea turtle (Lepidochelys olivacea) in Brazil, 1991/1992 to 2002/2003. Journal of the Marine Biological Association of the United Kingdom 87: de Wet, A Factors affecting survivorship of loggerhead and leatherback sea turtles of South Africa. MSc thesis. Nelson Mandela Metropolitan University, Port Elizabeth. 188pp. Domingo, A., Sales, G., Giffoni, B., Miller, P., Laporta, M., & Maurutto, G Captura incidental de tortugas marinas con palangre pelágico en el Atlántico Sur por las flotas de Brasil y Uruguay. Collective Volume of Scientific Papers ICCAT 59:

228 Dutton, D. L., Dutton, P. H., Chaloupka, M., & Boulon, R. H Increase of a Caribbean leatherback turtle Dermochelys coriacea nesting population linked to long-term nest protection. Biological Conservation, 126(2), Eckert, S., Eckert, K., Pongamia, P. and Koopman, G Diving and foraging behaviour of leatherback sea turtles. Canadian Journal of Zoology 67: Ehrhart, L. M., Bagley, D. A., & Redfoot, W. E Loggerhead turtles in the Atlantic Ocean: geographic distribution, abundance, and population status.loggerhead Sea Turtles. Smithsonian Books, Washington DC.pp FAO Fisheries and Agriculture Department, Report of the Technical Consultation on Sea Turtles Conservation and Fisheries (2004. Bangkok, Thailand. FAO Fisheries Report. No Rome, FAO p. Finkbeiner, E. M., Wallace, B. P., Moore, J. E., Lewison, R. L., Crowder, L. B., & Read, A. J Cumulative estimates of sea turtle bycatch and mortality in USA fisheries between 1990 and Biological Conservation 144: Ferreira, R.L., Martins, H.R., Silva, A.A. and Bolten, A.B Impact of swordfish fisheries on sea turtles in the Azores. Arquipélago. Ponta Delgada. Life and Marine Sciences 18A: Fossette, S., Kelle, L., Girondot, M., Goverse, E., Hilterman, M. L., Verhage, B., Thoisy de B. & Georges, J. Y The world's largest leatherback rookeries: A review of conservation-oriented research in French Guiana/Suriname and Gabon. Journal of experimental marine biology and ecology 356: Gallo, B. M., Macedo, S., Giffoni, B. D. B., Becker, J. H., & Barata, P. C Sea turtle conservation in Ubatuba, southeastern Brazil, a feeding area with incidental capture in coastal fisheries. Journal Information, 5(1). Gardner, B., Sullivan, P. J., Epperly, S., & Morreale, S. J Hierarchical modelling of bycatch rates of sea turtles in the western North Atlantic. Endangered Species Research 5: Giffoni, B., Domingo, A., Sales, G., Pons, M., Miller, P., de O Junior, N., de K e Britto, M. y Maçaneiro, L. R Capture incidental de tortugas marinas (Caretta caretta y Dermochelys coriacea) por las flotas de palangre de Brazil y Uruguay. ICCAT - SCRS/2012/093 Gilman, E., Zollett, E., Beverly, S., Nakano, H., Davis, K., Shiode, D., Dalzell, P., and Kinan, I Reducing sea turtle by catch in pelagic longline fisheries. Fish and Fisheries 7: Hamann, M., Godfrey, M. H., Seminoff, J. A., Arthur, K., Barata, P. C. R., Bjorndal, K. A., Bolten, A. B., Broderick, A. C., Campbell, L. M., Carreras, C., Casale, P., Chaloupka, M., Chan, S. K.F., Coyne, M. S., Crowder, L. B., Diez, C. E., Dutton, P. H., Epperly, S. P., FitzSimmons, N. N., Formia, A., Girondot M., Hays, G. C., Cheng, I. J., Kaska, Y., Lewison, R., Mortimer, J. A., Nichols, W. J., Reina, R. D., Shanker, K., Spotila, J. R., Tomás, J., Wallace, B. P., Work, T. M., Zbinden, J. and Godley, B. J Global research priorities for sea turtles: informing management and conservation in the 21st century. Endangered Species Research 11: Hays, G.C., Broderick, A.C., Godley, B.J., Luschi, P., Nichols, W.J Satellite telemetry suggests high levels of fishing-induced mortality in marine turtles. Marine Ecology Progress Series 262: Heppell, S. S., Crouse, D. T., Crowder, L. B., Epperly, S. P., Gabriel, W., Henwood, T., Marquez R. and Thompson, N. B A population model to estimate recovery time, population size, and management impacts on Kemp s ridley sea turtles. Chelonian Conservation and Biology 4: Hobday, A. J., Smith, A. D. M., Stobutzki, I. C., Bulman, C., Daley, R., Dambacher, J. M., et al Ecological risk assessment for the effects of fishing. Fisheries Research 108: Honig, M. B., Petersen, S. L., & Duarte, A Turtle by-catch in longline fisheries operating within the Benguela current large marine ecosystem. Collect. Vol. Sci. Pap. ICCAT, 62(6), Kelle, L., Gratiot, N., & De Thoisy, B Olive ridley turtle Lepidochelys olivacea in French Guiana: back from the brink of regional extirpation?. Oryx,43(02), Kotas, J. E., dos Santos, S., de Azevedo, V. G., Gallo, B. M., & Barata, P. C Incidental capture of loggerhead (Caretta caretta) and leatherback (Dermochelys coriacea) sea turtles by the pelagic longline fishery off southern Brazil. Fishery Bulletin 102(2), Lewison, R. L., L. B. Crowder, A. J. Read, and S. A. Freeman Understanding impacts of fisheries bycatch on marine megafauna. Trends in Ecology and Evolution 19: Lewison, R.L., & Crowder, L.B Putting longline bycatch of sea turtles into perspective. Conservation Biology 21:

229 Lewison, R.L., Wallace, B.P., Alfaro Shigueto, J., Mangel, J.,Maxwell, S.,Hazen, E Fisheries bycatch of marine turtles: lessons learned from decades of research and conservation. In: J. Wyneken, J.A. Musick (eds). The Biology of Sea Turtles, Vol 3.CRC Press, Boca Raton, FL. Pp López, A. M., McClellan, D. B., Bertolino, A. R., & Lange, M. D The Japanese longline fishery in the Gulf of Mexico, 1978.Marine Fishery Review 41: Marco, A., Abella-Pérez, E., Monzón-Argüello, C., Martins, S., Araujo, S., & López-Jurado, L. F The international importance of the archipelago of Cape Verde for marine turtles, in particular the loggerhead turtle Caretta caretta. Zoologia Caboverdiana 2: 1-11 Marcovaldi, M. A Status and distribution of the olive ridley turtle, Lepidochelys olivacea, in the Western Atlantic Ocean. In Proceedings of the Regional Meeting, Marine Turtle Conservation in the Wider Caribbean Region: A Dialogue for Effective Regional Management, Santo Domingo. pp Marcovaldi, M. Â., & Chaloupka, M Conservation status of the loggerhead sea turtle in Brazil: an encouraging outlook. Endangered Species Research, 3(2), Margaritoulis, D Nesting activity and reproductive output of loggerhead sea turtles, Caretta caretta, over 19 seasons ( ) at Laganas Bay, Zakynthos, Greece: the largest rookery in the Mediterranean. Chelonian Conservation and Biology 4: Mejuto, J., García-Cortés, B., & Ramos-Cartelle, A Trials using different hook and bait types in the configuration of the surface longline gear used by the Spanish swordfish (Xiphias gladius) fishery in the Atlantic Ocean. Collection Vol. Sci. Pap. ICCAT, 62: Meylan, A. B Status of the hawksbill turtle (Eretmochelys imbricata) in the Caribbean region. Chelonian Conservation and Biology 3: Milton, D.A Assessing the susceptibility of fishing of populations of rare trawl bycatch: sea snakes caught by Australia's Northern Prawn Fishery. Biological Conservation 101: Mortimer, J. A., and M. Donnelly IUCN SSC Marine turtle specialist group. IUCN Red List status assessment hawksbill turtle (Eretmochelys imbricata). National Research Council 2010). Assessment of sea turtle status and trends: integrating demography and abundance. National Academies Press, Washington DC. Nel, R.,.Wanless, R.M., Angel, A., Mellet, B., Harris, L Ecological risk assessment and productivitysusceptibility analysis of sea turtles overlapping with fisheries in the IOTC region. Unpublished report to the IOTC and IOSEA marine turtle MoU, 44 pp. Nel, R., Punt, A.E., Hughes, G.R Are coastal protected areas always effective in achieving population recovery for nesting sea turtles? PLoS ONE 8, e Ormseth, O.A., Spencer, P.D An assessment of vulnerability in Alaska groundfish. Fisheries Research 112: Pacheco, J. C., Kerstetter, D. W., Hazin, F. H., Hazin, H., Segundo, R. S. S. L., Graves, J. E., Carvalho, F. and Travassos, P. E A comparison of circle hook and J hook performance in a western equatorial Atlantic Ocean pelagic longline fishery. Fisheries Research 107: Parker, D., Cooke, W. and Balazs, G Diet of oceanic loggerhead sea turtles (Caretta caretta) in the central Pacific Ocean. Fish Bulletin 103: Patrick, W.S., Spencer, P., Ormseth, O., Cope, J., Field, J., Kobayashi, D., Gedamke, T., Cortes, E., Bigelow, K., Overholtz, W., Link, J. and Lawson, P Use of productivity and susceptibility indices to determine stock vulnerability, with example applications to size U.S. Fisheries. Department of Commerce, NOAA Technical Memorandum NMFS-F/SPO pp. Pikitch, E.K., Santora, C., Babcock, E.A., et al Ecosystem-based fisheries management. Science 305: Pinedo, M. C., & Polacheck, T Sea turtle by-catch in pelagic longline sets off southern Brazil. Biological conservation, 119(3), Polovina, J., Howell, E., Parker, D., Balazs, G., Dive-depth distribution of loggerhead (Carretta carretta) and olive ridley (Lepidochelys olivacea) sea turtles in the Central North Pacific: might deep longline sets catch fewer turtles? Fish Bulletin 101: Polovina, J. J., Balazs, G. H., Howell, E. A., Parker, D. M., Seki, M. P., & Dutton, P. H Forage and migration habitat of loggerhead (Caretta caretta) and olive ridley (Lepidochelys olivacea) sea turtles in the central North Pacific Ocean. Fisheries Oceanography 13:

230 Pons, M., Domingo, A., Sales, G., Niemeyer Fiedler, F., Miller, P., Giffoni, B., Ortiz, M Standardization of BPUE of loggerhead sea turtle (Caretta caretta) caught by pelagic longliners in the Southwestern Atlantic Ocean. Aquatic Living Resources 23: Rader, H., Mba, M. A. E., Morra, W., & Hearn, G Marine turtles on the southern coast of Bioko Island (Gulf of Guinea, Africa), Marine Turtle Newsletter, 111, Read, A. J Do circle hooks reduce the mortality of sea turtles in pelagic longlines? A review of recent experiments. Biological Conservation 135: Sales, G., Giffoni, B. B., & Barata, R. C Incidental catch of sea turtles by the Brazilian pelagic longline fishery. J. of the Marine Association of the United Kingdom 88: Schofield, G., Hobson, V. J., Fossette, S., Lilley, M. K., Katselidis, K. A., & Hays, G. C Biodiversity Research: fidelity to foraging sites, consistency of migration routes and habitat modulation of home range by sea turtles. Diversity and Distributions, 16: Small, C., Waugh, S.M., Phillips R.A The justification, design and implementation of Ecological Risk Assessments of the effects of fishing on seabirds. Marine Policy 37: Swimmer, Y., Arauz, R., McCracken, M., McNaughton, L., Ballestero, J., Musyl, M., Bigelow, K., Brill, R Diving behavior and delayed mortality of olive ridley sea turtles Lepidochelys olivacea after their release from longline fishing gear. Marine Ecology Progress Series 323: Swimmer, Y., J. Suter, R. Arauz, K. Bigelow, A. Lopez, I. Zanela, A. Bolanos, J. Ballestero, R. Suarez, J. Wang, and C. Boggs Sustainable fishing gear: The case of modified circle hooks in a Costa Rican longline fishery. Marine Biology 158: Tomás, J., Godley, B. J., Castroviejo, J., & Raga, J. A Bioko: critically important nesting habitat for sea turtles of West Africa. Biodiversity and Conservation 19(9), Troëng, S., & Rankin, E Long-term conservation efforts contribute to positive green turtle Chelonia mydas nesting trend at Tortuguero, Costa Rica. Biological Conservation, 121(1), Troëng, S. and Chaloupka, M Variation in adult annual survival probability and remigration intervals of sea turtles. Marine Biology 151: Tuck, G.N, Phillips, R.A., Small, C., Thomson, R.B., Klaer, N.L., Taylor, F., Wanless, R.M. and Arrizabalaga, H An assessment of seabird fishery interactions in the Atlantic Ocean. ICES Journal of marine Science 68: Wallace, B.P., DiMatteo, A.D., Hurley, B.J., et al. 2010a. Regional management units for marine turtles: a novel framework for prioritizing conservation and research across multiple scales. PLoS ONE 5, e Wallace, B. P., Lewison, R. L., McDonald, S. L., McDonald, R. K., Kot, C. Y., Kelez, S., Bjorkland, R. K., Finkbeiner, E. M., and Crowder, L. B. 2010b. Global patterns of marine turtle bycatch. Conservation letters, 3: Wallace, B. P., DiMatteo, A. D., Bolten, A. B., Chaloupka, M. Y., Hutchinson, B. J., Abreu-Grobois, F. A., Mortimer, J. A., Seminoff, J. A., Amorocho, D., Bjorkland, R. K., Bourjea, J., Bowen, B. W., Dueñas, R., Casale, P., Choudhury, B. C., Costa, A., Dutton, P. H., Fallabrino, A., Elena M. Finkbeiner, E. M., Girard, A., Gorondot, m., Hamann, M., Hurley, B. J., López-Mendilaharsu, M., Marcovaldi, M. A., Musick, J. A., Nel, R., Pilcher, N.J., Troëng, S., Witherington, B. and Mast, R. B Global conservation priorities for marine turtles. PLoS One 6, e Wang, J. H., Boles, L. C., Higgins, B., & Lohmann, K. J Behavioral responses of sea turtles to lightsticks used in longline fisheries. Animal Conservation, 10: Watson, J. W., Epperly, S. P., Shah, A. K. and Foster, D. G Fishing methods to reduce sea turtle mortality associated with pelagic longlines. Canadian Journal of Fisheries and Aquatic Sciences 62: Weir, C. R., Ron, T., Morais, M., & Duarte, A. D. C Nesting and at-sea distribution of marine turtles in Angola, West Africa, : occurrence, threats and conservation implications. Oryx, 41(2), 224. Witzell, W. N The incidental capture of sea turtles in the Atlantic U. S. Fishery Conservation Zone by the Japanese tuna longline fleet, Marine Fisheries Review, 46(3), Witzell, W. N Distribution and relative abundance of sea turtles caught incidentally by the US pelagic longline fleet in the western North Atlantic Ocean, Fishery Bulletin 97:

231 Table 1. IUCN Red list threat status for all marine turtle species Common name Species name IUCN status Leatherback turtle Dermochelys coriacea Critically Endangered Hawksbill turtle Eretmochelys imbricata Critically Endangered Kemp's ridley turtle Lepidochelys kempii Critically Endangered Loggerhead turtle Caretta caretta Endangered Green turtle Chelonia mydas Endangered Olive ridley turtle Lepidochelys olivacea Vulnerable *Flatback turtle Natator depressus Data deficient *The flatback turtle is the only one that does not occur in the Atlantic Ocean Table 2. Turtle species regional management units in the Atlantic Ocean. Common name RMU RMU area Loggerhead turtle Cc-AtNW Atlantic northwest Cc-AtNE Cc-Med Cc-AtSW Cc-InSW Atlantic northeast Mediterranean Atlantic southwest Indian southwest Green turtle Cm-AtNW Atlantic northwest Cm-AtSCar Cm-AtSCen Cm-AtSW Cm-AtE Cm-InSW Cm-Med Atlantic south Caribbean Atlantic south central Atlantic southwest Atlantic east Indian southwest Mediterranean Leatherback turtle Dc-AtNW Atlantic northwest Dc-AtSE Dc-AtSW Dc-InSW Atlantic southeast* Atlantic southwest* Indian southwest Hawksbill turtle Ei-AtWCar Atlantic west Caribbean Ei-AtE Ei-AtSW Atlantic east Atlantic southwest Olive ridley turtle Lo-AtW Atlantic west Lo-AtE Atlantic east Kemp's ridley turtle Lk-AtNW Atlantic northwest *These two RMU areas are indistinguishable from each other and overlap completely at sea (drawn as one RMU on maps). Turtles caught in the South Atlantic area are assigned to one RMU only if caught close to the south eastern or south western coasts. Otherwise, they are assigned to both RMUs. 2247

232 Table 3a. Scoring and categorization for productivity parameters used in the ICCAT ecological risk assessment for turtles. Scores were divided into half-points and five categories were used when parameters scaled across orders of magnitude. Productivity score Number of breeding females 1 Population trend Generation length Age at maturity Hatching success Emergence success Mean clutch size Nests/ female/ season Remigration interval (Low) Very small, Small Declining/ Uncertain DD >30 years <50% <50% <90 eggs < 4 nests > 4 years (Medium) Medium, Large Stable DD years 50 75% 50 75% (high) Very large Increasing DD <16 years >75% >75% >120 >6 <2.6 1 Note that these were scored differently for each species, see Table 3b Table 3b. Categorisations used to assign scores for each sea turtle Regional Management Unit. Values relate to numbers of breeding females, following Wallace et al. (2010a) Species Very Small Small Medium Large Very Large Dermochlelys coriacea < >1000 Eretmochelys imbricata < >1000 Caretta caretta < ,000 >10,000 Chelonia mydas < ,000 >10,000 Lepidochelys olivacea < ,000 10, ,000 >100,000 Lepidochely kempii < ,000 10, ,000 >100,

233 Table 4. Scoring and categorization for susceptibility parameters used in the ICCAT ecological risk assessment for turtles. Susceptibility score Overlap with ICCAT region (no. squares) Confidence (no. of satellite tracks) Bycatch mortality relative to breeding females (%) 1 (Low) <50 <5 <30 2 (Medium) (high) >100 >30 >100 Table 5. Productivity scores and rank for each sea turtle regional management unit (RMU) overlapping with the ICCAT region. RMU Productivity score Rank Cm-InSW Ei-AtWCar Cm-AtNW Cc-AtNW Cc-AtSW Lk-AtNW Dc-InSW Cc-InSW Cm-AtE Lo-AtW Ei-AtSW 2 11 Dc-AtSE Cc-Med Cm-AtSCar Cm-AtSCen Cm-AtSW Cm-Med Dc-AtNW Lo-AtE Cc-AtNE Dc-AtSW Ei-AtE

234 Table 6. Total number of sea turtles captured (by species and fate) from >9,000 observed purse seine sets in the Atlantic Ocean ( ), taken from Clermont et al. (2012). Captures and fates are also presented by sets on Fish Aggregating Devices (FAD) and sets on free swimming-schools (FSC). Species Alive Dead Unknown Total caught Caretta caretta Chelonia mydas Dermochelys coriacea Eretmochelys imbricata Lepidochelys kempii Lepidochelys olivacea Unidentified Total Fate (%) 90.60% 5,06% 4,34% 100,00% FAD ,04% 3,48% 4,48% 100,00% FSC ,25% 6,54% 4,21% 100,00% Table 7. Ranked vulnerability scores for turtle RMUs in the ICCAT region, by gear type (LL = longlining, PS Purse seining).note RMUs that don t overlap with any PS effort are excluded from the table RMU LL Rank Lo-AtE Cm-AtSCar Cc-Med Dc-AtSW Lk-AtNW Cc-AtSW Cm-AtNW Ei-AtE Ei-AtWCar Dc-AtNW Cc-AtNE Cm-AtE Dc-AtSE Ei-AtSW Cm-Med Cm-AtSW Cm-AtSCen Lo-AtW Cc-AtNW Cm-InSW Cc-InSW Dc-InSW Mean 1.72 RMU PS Rank Lo-AtE Ei-AtE Cc-AtNE Dc-AtSW Cm-AtSCar Dc-AtNW Dc-AtSE Cm-AtSCen Cm-AtSW Dc-InSW Cc-AtSW Cc-AtNW Cm-AtNW Ei-AtWCar Cm-AtE Lo-AtW Ei-AtSW Mean

235 Figure 1. Mean longline fishing effort ( ) in the ICCAT region overlapped with Caretta caretta (loggerhead) turtle species Regional Management Units (shaded polygons). The highlighted numbers show the total estimated number of hooks to which each RMU is exposed. 2251

236 Figure 2. Mean longline fishing effort ( ) in the ICCAT region overlapped with Chelonia mydas (green) turtle species Regional Management Units (shaded polygons). The highlighted numbers show the total estimated number of hooks to which each RMU is exposed. 2252

237 Figure 3. Mean longline fishing effort ( ) in the ICCAT region overlapped with Dermochelys coriacea (leatherhead) turtle species Regional Management Units (shaded polygons). The highlighted numbers show the total estimated number of hooks to which each RMU is exposed. 2253

238 Figure 4. Mean longline fishing effort ( ) in the ICCAT region overlapped with Eretmochelys imbricata (hawksbill) turtle species Regional Management Units (shaded polygons). The highlighted numbers show the total estimated number of hooks to which each RMU is exposed. 2254

239 Figure 5. Mean longline fishing effort ( ) in the ICCAT region overlapped with Lepidochelys olivacea (olive ridley) turtle species Regional Management Units (shaded polygons). The highlighted numbers show the total estimated number of hooks to which each RMU is exposed. 2255

240 Figure 6. Mean longline fishing effort ( ) in the ICCAT region overlapped with Lepidochelys kempii (Kemp s ridley) turtle species Regional Management Units (shaded polygons). The highlighted numbers show the total estimated number of hooks to which each RMU is exposed. 2256

241 Figure 7. Productivity (x) and Susceptibility (y) scores for longline fishery in 22 sea turtle RMUs in the ICCAT region. Top-left corner = most vulnerable and bottom-right least vulnerable. Figure 8. Mean longline fishing effort ( ) in the ICCAT region, showing concentrated effort in the Mediterranean, and tropical Latitudes. 2257

242 Figure 9. Mean purse seine fishing effort ( ) in the ICCAT region, using fish aggregating devices (FADs). The effort is concentrated in the tropics, off West Africa between Namibia and Mauritania, and off Venezuela. 2258

243 Figure 10. Mean longline purse seine effort ( ) in the ICCAT region, sets over free swimming-schools (FSC). The effort is concentrated in the tropics, off West Africa between Namibia and Mauritania, and off Venezuela. 2259

244 SCRS/2013/012 Collect. Vol. Sci. Pap. ICCAT, 70(5): (2014) 2013 INTER-SESSIONAL MEETING OF THE SHARKS SPECIES GROUP (Mindelo, Cape Verde April 8 to 12, 2013) SUMMARY The Intersessional Meeting of the Sharks Species Group was held in Mindelo, Cape Verde, 8-12 April, The objective of the meeting was to develop the Special Research Programme on Sharks framed within the SCRS Science Strategic Plan foreseen for the period RÉSUMÉ La Réunion intersession du Groupe d'espèces sur les requins a eu lieu à Mindelo (Cap-Vert) du 8 au 12 avril L objectif de cette réunion consistait à élaborer le programme spécial de recherche sur les requins s'inscrivant dans le plan stratégique en matière de science du SCRS prévu pour la période RESUMEN La reunión intersesiones del Grupo de tiburones se celebró en Mindelo, Cabo Verde, del 8 al 12 de abril de El objetivo de la reunión era elaborar el Programa especial de investigación sobre tiburones enmarcado dentro del Plan estratégico de ciencia del SCRS previsto para el periodo Opening, adoption of Agenda and meeting arrangements Mr. Óscar David Fonseca Melício, President of the National Institute for Fisheries Development (INDP) of Cape Verde, welcomed the participants to Mindelo and was thanked by the Chairman of the SCRS, Dr. Josu Santiago for hosting the meeting at the Institute. Dr. Paul de Bruyn, on behalf of the ICCAT Executive Secretary, then opened the meeting. The meeting was chaired by Dr. Andrés Domingo, the Shark Species Group Rapporteur. Dr. Domingo welcomed Species Group participants and addressed the terms of reference for the meeting. After opening the meeting, the Agenda was reviewed and adopted with minor changes (Appendix 1). The List of Participants is included as Appendix 2. The List of Documents presented at the meeting is attached as Appendix 3. The following participants served as Rapporteurs for various sections of the report: Section Rapporteurs 1 P. de Bruyn 2 P. de Bruyn, A. Perry, A. Domingo 3 P. de Bruyn 4 E. Cortés, R. Coelho, G. Burgess, B. Seret 5 Species Group participants 6 J. Santiago, A. Domingo 7 J. Santiago, A. Domingo 8 P. de Bruyn 2. Review of the documents In SCRS/2013/044 it was identified that, to date, changes in target species have been incorporated in stock assessments at two different levels in the analysis. First, these changes are taken into account during the parameterization of generalized linear models used to compute the CPUE index standardization. Second, continuously time-varying catchabilities are directly incorporated during the fitting of the dynamic model used for the assessment. The latter step models the annual catchabilities as random draws from a stationary 2260

245 distribution of catchabilities. Empirical evidence, however, suggests that models in which large, one-time changes in catchabilities could very well describe the temporal changes in various fisheries. Here was presented a suite of Bayesian state-space production models fitted to the time series of South Atlantic blue shark (Prionace glauca) stock, in which a single change point in the stationary distribution of catchabilities is specified, with two catchability parameters being estimated, one before and another after the changing point. Despite the models introducing a single extra parameter, they resulted in an improved fit over the one-parameter catchability modelling approach. The models resulted in different estimates of reference points and harvest quotas. However, they all indicated that blue shark stock is above B MSY and that fishing mortality levels are still below F MSY. Although accounting for a single change point in catchability had no significant impact on the status of this particular fish population, it provides a robust way of taking into account changes in catchability as result of changing fisheries dynamics, and can be implemented to model other fish stocks. The Group discussed whether the assumption that the changes in catches over the study time period were due to changes in selectivity or catchability. Changes could also be attributed to the market demand. In the logbooks, there is information on landings not discards. It was noted that according to the logbooks provided by the fishermen, there has been changes in targeting over time, but these are hard to quantify. In SCRS/2013/045 it was noted that pelagic sharks are faced with complex movement decisions while residing in a relatively featureless and oligotrophic environment. They are also a common by-catch in pelagic fisheries, raising concerns about over-harvesting. Developing management plans and effective stock assessments requires understanding how these animals utilize entire ocean environments, as trans-oceanic movements are common. Here satellite telemetry and random mixed models were utilised to quantify the factors driving movement patterns in blue shark, Prionace glauca, across the South Atlantic Ocean. The majority of sharks showed residency to core areas, although there were individuals that made long distance movements, including two trans- Atlantic dispersal events. Habitat selection was primarily explained by sea surface temperature (SST) and the depth of the mixed layer (DML), but this varied by region. In areas hypothesized to be locations of gestation, adult female sharks selected shallower and warmer waters than males. The South Atlantic blue shark population should be treated as a single stock, although it is unlikely that they utilize a clockwise migration cycle across the Atlantic Ocean. The Group discussed the sensitivity of the model to assumptions of movement as other spatially explicit models are based on a large number of conventional tags, whereas this model uses a very limited number of satellite tags. It was explained that although there may be some differences in the precise location of the individuals tagged, they would always have been in the same zone as defined by the model. SCRS/2013/037 presented information for Portuguese longliners targeting swordfish in the Atlantic Ocean which regularly capture several elasmobranch species as by-catch, including currently protected species such as the bigeye thresher and the smooth hammerhead. This paper presents preliminary results from bigeye threshers and smooth hammerheads tagged with pop-up archive satellite transmitting tags during 2012 in the NE tropical region of the Atlantic. Strong diel vertical migration patterns were observed for the bigeye threshers with the most occupied depths being m during the day and m during the night, corresponding to water temperatures of 8-10ºC and 22-24ºC, respectively. For the smooth hammerhead no major differences were detected between the day/nighttime periods, with most of the time spent in the m depth range. While the data presented in this paper is still limited and part of ongoing projects, the preliminary results are useful to increase the knowledge on these species biology, ecology and habitat utilization patterns, and can serve as inputs for ongoing and future Ecological Risk Assessments analysis. The Group noted that the tagging study was able to collect information at depths greater than are normally exploited by the longline fishery and thus provides us with information that would not normally be obtained from the fishery. This information could be very important for the elaboration of the sharks research plan to be developed during the meeting. It was also noted that research is also ongoing for the oceanic whitetip shark Document SCRS/2013/038 discussed the bigeye thresher shark, Alopias supercilious, which is commonly caught as by-catch in pelagic longline fisheries targeting swordfish. As part of an ongoing program for fisheries and biological data collection, fishery observers have been placed onboard fishing vessels, collecting a set of information which includes size, sex and maturity stage, aiming to investigate the maturity of the bigeye thresher shark. A total of 1006 bigeye threshers were recorded throughout the Atlantic Ocean. Size of the specimens ranged from 94 to 264 cm FL (fork length). In the northern regions, there was a higher proportion of females (> 63%) and the observed modal size class was lower than that of the southern regions, where the largest specimens were found. Maturity ogives were fitted for 642 specimens with maturity data available. Size at first maturity was estimated at 208.6cm FL for females (corresponding to years) and cm FL for males (corresponding to 5-6 years). 2261

246 The Group noted that there are potentially different strategies between thresher shark species. SCRS/2013/042 identified that for the improvement of future stock assessment of shortfin mako (Isurus oxyrinchus) in the Atlantic Ocean, it is important to review biological parameters. In the last stock assessment meeting, the uncertainty about catch statistics, catchability and biological parameter was discussed regarding the poor fitting of estimated biomass trend to the observed trend on CPUE. Even granting that there may be unignorable amount of unreported catch, it cannot explain the increase of CPUE consistently observed in many fleets by itself. It is valuable to reassess the existing assumption that intrinsic rate of natural increase (r) of this species is quite low, rounding up existing knowledge on the biological parameter. This document provides information on the current status of biological studies for the populations in the North Pacific, focusing on the growth analysis, because we have latest study in this area and, needless to say, growth parameter plays an integral role in the population dynamics among various biological parameters. Important points to be taken into account in the preparation of future research plan are also discussed. SCRS/2013/040 provided a presentation of at-vessel mortality, post-release survival rate, and total mortality of silky sharks in the French tropical tuna purse seine fishery operating in the Indian Ocean. Currently, French tropical purse seiners in the Indian Ocean release all sharks and rays that are caught incidentally. Through participation in two commercial fishing trips and one chartered research cruise, we first recorded the number of sharks (primarily silky sharks, Carcharhinus falciformis) that were alive or dead, once they had been sorted by the crew on the upper and lower decks. More sharks were observed in the lower deck (73%) than in the upper deck. The silky sharks observed on the upper deck were significantly larger than the ones found in the lower deck. The immediate mortality (sharks that were dead at the time of observation) rates appeared to be linked with the location of the individual, as more sharks were found dead on the lower deck than the upper deck. The atvessel mortality rates also increased with the set size (tonnage). 20 silky sharks were tagged with MiniPATs (Wildlife Computers, Redmond, Washington, USA) to study their survival after release. In addition, 12 silky sharks were tagged with the same type of electronic tags during a scientific cruise. Of a sub sample of 32 silky sharks assessed alive upon retrieval and monitored for periods of up to days after release, 8 tags clearly showed mortality directly after release, while data from four tags suggested delayed mortality after 2 to 35 days and one in poor condition died after 3 days, eaten. In all, 16 tags showed that the sharks survived. Two tags failed to report data and one was incorrectly initiated. This document provides the first estimates, for silky sharks (length >85 cm TL) of at-vessel mortality and post-release mortality, respectively, of around 67% and 58%. The overall mortality rate of silky sharks by-caught by this fleet was concluded to be about 81%. A 'best practices' manual for fishers has been prepared to increase rates of survival of sharks caught by purse seine vessels. However, other methods prior to the sharks being brought onboard must also be investigated. The Group requested additional clarification as to how the sharks were selected for study. It was stated that, each shark was assessed according to the following scale: 1) Good: very active behaviour, biting, kicking; 2) Fair: little movement but still clear signs of life; 3) Poor: low response to external stimuli; 4) Dead. Then, 32 sharks that were showing signs of life (scale 1 and 2) were randomly selected. The high level of atvessel mortality was discussed and it was noted that the protocols for release currently in place had not yet been adopted at the time of the study. SCRS/2013/039 provided an overview of the elasmobranchs catch-at-size and sex-ratios on the Portuguese pelagic longline fishery in the Atlantic Ocean. The analysis was based on data collected from fishery observers, port sampling and from skippers logbooks (self-sampling), collected between 1997 and Data was analysed in terms of by-catch-at-size and compared between years, seasons (quarters), stocks (North and South, separated at 5ºN) and major fishing areas of operation for the Portuguese fleet (North, Tropical North, Equatorial and South). For the blue shark a general increasing trend on mean sizes was observed for both hemispheres with a decrease in the more recent years. For the shortfin mako the mean size has remained stable in the North and tended to decrease in the South. Some variability was noted in the seasonal and spatial comparisons. The sexratios proportions were compared between regions and seasons, and for the main species significant differences were found. The data presented in this working document is still preliminary, but provides new and important information on the catch-at-size trends and sex-ratios for the major pelagic sharks captured by the Portuguese pelagic longline fishery in the Atlantic Ocean. 2262

247 A brief explanation of the EU Portugal self-sampling was provided. The programme is based on a MS Excel spreadsheet, which allows skippers to calculate total catch weight from individual samples. It is useful to the skippers for compliance purposes and provides information for scientific purposes. It was noted that VMS information is difficult to obtain due to confidentiality issues. Efforts are being made to get this data from the fisheries management department in a form that is aggregated enough for distribution. SCRS/2013/046 reported length-length relationships between Fork Length, Precaudal Length and Total Length for the main six pelagic species (Prionace glauca, Carcharhinus brachyurus, Carcharhinus signatus, Sphyrna zygaena, Isurus oxyrinchus and Lamna nasus) captured by the Uruguayan pelagic longline fleet in the southwestern Atlantic Ocean between 1998 and The length-length relationships provided in this contribution covers an extended portion of the reported full size spectrum of each species considered, and represents the first length-length conversions ever reported for them species in the area. Document SCRS/2013/047 evaluated the catches of sharks in the artisanal driftnet fishery off Abidjan (Côte d Ivoire) for the period , using weight and size data collected for every shark species on 3 landing sites, and the proportion of sampled pirogues. During this period, the number of day trips decreased of half and the catches varied between 92 and 203 t. However, the proportion of sharks in the total catches varied from 2.1 % in 2008 to 31% in The most important species were the blue shark (Prionace glauca) and the shortfin mako (Isurus oxyrinchus), whose CPUEs (kg/day trip) were slightly increasing. The catches were composed of juveniles of cm TL for the blue shark and cm TL for the mako shark. The author clarified that the length measurement provided in the document was precaudal length. It was noted that the gear type described in the study was gillnet and that these were set 2 miles from the shore, potentially close to canyons or the continental drop-off. This could explain the relatively large number of sharks reported. SCRS/2013/041 noted that currently the reduction of by-catch mortality is an objective of the ecosystem approach to fisheries and a request made by consumers. The involvement and participation of resource users is necessary to develop efficient and practical mitigation techniques. Fishers handle animals as part of their job duties and it is essential to identify good practices that ensure the safety of the crews and optimize the survival of released animals. Combining scientific observations and empirical knowledge from fishers of the French purseseine fleet, handling and release guidelines are proposed for sharks and rays, including large ones, like whale sharks and manta rays incidentally caught by tropical tuna purse-seine fisheries. A good practices manual has been prepared to raise the fishers awareness of the preservation and conservation of biodiversity and encourage their participation in the sustainable management of marine resources. Bringing these best practices onto the decks of fishing vessels should contribute to the reduction of the fishing mortality of some vulnerable species. It would be positively viewed by consumers as an act that reduces fishing s footprint on the environment and promoting the animal welfare which would improve the image of fishing industry. New ideas emerging from exchanges between scientists and fishers are also proposed although not yet tested. Mitigation research is by definition an iterative process and different complementary methods must be carried out at different levels of the fishing process to significantly reduce the mortality of these animals. SCRS/2013/049 indicated that the lack of reliable fishery-dependent data and fundamental understanding of the biology of most shark species causes concern for the Sustainable management of shark populations in the Mediterranean Sea. The study aims at investigating on habitat occupancy, residency times and migratory pathways as well as providing behavioural data on temperature experience and swimming depth of the large pelagic shark mainly the blue shark (Prionace glauca). This study strives to also determine when and where sharks are most vulnerable and will assist in the conservation of the species. The use of satellite tag is proposed to investigate on the ecology of the large pelagic sharks. The preliminary results of the first SPOT (Smart position or temperature transmitting) tag deployed of a female blue shark are presented. SCRS/2013/048 noted that in 2010, the EC zero TAC for the porbeagle shark caused the closure of the seasonal targeted fishery traditionally performed by a small fleet of five long-liners of Yeu Island (Bay of Biscay). In order to improve knowledge on porbeagle, the French Ministry of Fisheries supported a scientific program aimed determining the movements of this shark in the NE Atlantic using pop-up satellite tags (PSAT). In summer 2011, three PSATs could be deployed on adult and sub-adult porbeagle females during a tagging cruise carried out in the Bay of Biscay with a longliner of Yeu Island. The three tags popped-up, one at 8 months and two at 12 months (i.e., original setting duration). Although the data transmitted by the tags need to be re-processed with various filters, preliminary analysis shows that the tagged sharks exhibited three different patterns of movements in the NE Atlantic. A mature female of 2.34m TL tagged off Quiberon Peninsula stayed a month in the vicinity, then moved north up to the Shetland shelf where it stayed about 2.5 months, to finally reach the Sea of Norway 2263

248 in November; then it moved to Iceland to return to Norway in February where the tag popped-up. During this migration, this shark did regular dives to 500 m depth, reaching a maximum of 1000 m depth. The second shark, a sub-adult female of 1.9 m TL was tagged off Noirmoutier Island. This shark did a large triangular trajectory in the Atlantic going north-west, reaching close to Greenland in November, then going straight south to the Azores in February-March, before coming back to almost the original tagging position, 12 months later. This shark also did regular dives down to about 1000 m depth. The third shark, a sub-adult female of 1.9 m TL was tagged off the Penmarch Peninsula, also moved north-west, did a return trip to the North Sea in October-November, before going back to the Bay of Biscay (off southern Ireland) in June with a jigsaw trajectory; it dived down to 800 m depth when it was off the continental shelf. Although limited, these observations show that the porbeagle shark uses large areas of the NE Atlantic and the water column down to 1000 m depth. A brief presentation of an on-going project carried out by EU Institutes was provided to the Group. The general objective of the project is to obtain scientific advice for the purpose of implementing the EU Plan of Action for the Conservation and Management of Sharks, as regards the facilitation of monitoring high seas fisheries and shark stock assessment on a species-specific level. The study is focused on 18 major elasmobranch species on a worldwide basis. In order to achieve the project goals, the team has been: collating and examining historical fisheries data, especially in terms of species composition, catches and effort; estimated global shark catches; identifying gaps in the current knowledge of fisheries, and also on the biology and ecology of sharks. In order to fill the gaps, and to support advice from RFMO on sustainable management of elasmobranch fisheries, a number of proposals are being prepared, namely in terms of designing observer programs, identifying scientific research priorities and the integration of information on t-rfmos. The Group welcomed the initiative and requested the authors to facilitate the outcomes of the project as soon as the information is available. 3. Presentation of Task I, Task II and tagging data The Secretariat presented a summary of the information on sharks submitted by the CPCs. Task I and Task II catch-effort and size samples were presented in the form of data catalogues for the purpose of identifying gaps in the available data. It was noted that although Task I data are available for many species of sharks, these data are extremely incomplete and in many cases, Task I data have not been accompanied by the corresponding Task II data. This is particularly true for species other than blue shark, shortfin mako and porbeagle for which there is generally more information available. It was also noted that there is more information available for the North Atlantic than for the South Atlantic, with very little data available for the Mediterranean. The Group requested that the data be presented in a format to easily identify gaps in order to address these deficiencies in the research plan (Appendices 4-6). The Secretariat also presented the available tagging information for blue shark, shortfin mako and porbeagle. The densities of tagging, recaptures as well as the tracks are provided in Figures 1-3. It was also suggested that an objective for the Group could be to develop a format for reporting the satellite tagging data to ICCAT. It was acknowledged that the dataset for each tag can be quite extensive and thus it is more likely feasible to report metadata for the electronic tags (such as the tagging and pop-up locations). 4. Current status of knowledge and research on pelagic sharks in the Atlantic and Mediterranean This information is dealt with extensively in the research plan detailed in Section 5 below. 5. Scientific research plan for sharks and the compilation of data A presentation was given on the plan for the SCRS Strategic Research Plan in order to put the current discussions into their greater context within the work of the SCRS. Strategic Planning is recommended as a structured approach to guide the future workings of the SCRS (2011 SCRS Report and responsive to Res on Best Available Science). Document SCRS/2013/024 outlined an approach for identifying key research needs and components of and a roadmap for developing the SCRS Strategic Plan. SCRS/2013/024 points out that Strategic Planning deals with three basic constructs: What do we do?, For whom do we do it? and How do we excel? Furthermore, the key components of strategic planning include an understanding of the SCRS mission (our purpose), our vision for the future, values we shall apply in conduct of our work, our goals 2264

249 and strategies to achieve them. It was pointed out that Strategic Planning also provides a methodology to identify critical capacity and data gaps and prioritize research activities to address them. A roadmap and time-frame for developing the SCRS Strategic Plan was proposed in SCRS/2013/024 which includes contracting a consultant to provide a framework for the specific methodology to be applied in developing the Strategic Plan and regular consultation and review by SCRS officers and SCRS Plenary prior to review and acceptance by the Commission Objectives and targets of the Shark Research and Data Collection Programme A presentation was given on the general framework for the Shark Research and Data Collection Programme providing a template for discussion and elaboration. The Chairman then requested that the participants provide comments on the structure of the plan, potential content as well as identification of sections in which they are willing to contribute. This was conducted and a template was agreed on by the Group. The separate sections were then elaborated on by the participants. 5.2 Development of the programme The proposed Shark Research and Data Collection Programme is presented in Appendix Other matters Cape Verde scientists showed a description of the fishing activities impacting shark species within their EEZ by both the national fleet and foreign fleets (European Union, China) operating under different fishing agreements. The Cape Verde fleet does not target elasmobranchs although they constitute a component of the by-catch when targeting other species; and there are no specific licenses for sharks in Cape Verde for any fleet. In the case of the foreign longline fleets operating in the Cape Verde EEZ, they report a high percentage of sharks representing over 75% of their catches and which are mostly composed by Prionace glauca and Isurus oxyrinchus. Considering the importance of shark species in the Cape Verde area, local scientists presented an initiative of developing a Data Collection Programme for their national fleet, for which technical assistance was required. Cape Verde again expressed its desire to obtain assistance to develop a Data Collection Programme, including sampling procedures and a data processing system on the shark species caught by its fleet. The Group acknowledged Cape Verde s initiative for the development of a Data Collection Programme for its national fleet with a special focus on shark species. Although sharks are not the target of the local fleet, these are an important component of their catch. The Group recommends that special funds from ICCAT be provided for this important initiative. 7. Recommendations The Species Group recommends that scientific observers be allowed to collect biological samples (vertebrae, tissues, reproductive tracts, stomachs, skin samples, spiral valves, jaws, whole and skeletonised specimens for taxonomic work and museum collections) from currently prohibited sharks species that are dead at haulback, provided that the samples are part of the research project approved by the SCRS. In order to obtain the approval, a detailed document outlining the purpose of the work, number and type of samples intended to be collected and the spatio-temporal distribution of the sampling work must be included in the proposal. Annual progress of the work and a final report on completion of the project shall be presented to the Sharks Species Group and the SCRS. Cape Verde expressed its desire to obtain assistance to develop a Data Collection Programme, including sampling procedures and a data processing system on the shark species caught by its fleet or landed in Cape Verde. Although sharks are not the target of the local fleet, these are an important component of their catch. The Group recommends that special funds from ICCAT be provided to this important initiative. The Group recommends that in 2014 a small group of SCRS scientist should be in charge of elaborating the biological sampling design for pelagic shark species in the Atlantic and Mediterranean. The expected budget of this action should be evaluated and proposed to SCRS for its approval. 2265

250 8. Adoption of the report and closure The Group expressed appreciation for all the arrangements and facilities provided by the INDP and its scientists for the more than satisfactory development of the meeting. The hospitality provided was extraordinary and the Species Group deeply acknowledged the unbelievable attention given to the participants by the Cape Verde scientists. SHARK RESEARCH AND DATA COLLECTION PROGRAM (Appendix 7) A. INTRODUCTION A great variety of shark species are found within the ICCAT Convention area, from coastal to oceanic species. Ninety-one species of sharks are currently present in the ICCAT databases. Biological strategies of these species are very diverse and they are highly adapted to their respective ecosystems and occupy a very high position in the trophic chain as active predators. Although diverse, the biological characteristics of these species share some general patterns that make them potentially more susceptible to overfishing. Even though elasmobranchs are currently impacted by commercial and recreational fisheries, there is still limited information about these species life cycles, biological parameters, movement patterns and habitat utilization, and in the general impact of fisheries in their populations in the ICCAT Convention area. Moreover, the current state of knowledge on ICCAT fisheries capturing sharks is causing concerns regarding their conservation status and management due to the gaps in the available catch, effort and discard data. And it is evident that the limited quantity and quality of information available affects the provision of scientific advice to the Commission. Numerous aspects of the biology of these species are still poorly understood or completely unknown, particularly for some regions, which contributes to increased uncertainty in quantitative and qualitative assessments. As regards information of fisheries activities of fleets capturing sharks (catch and by-catch), the reporting of Task I and Task II has improved in the recent years but this improvement is still insufficient to permit the Committee to provide quantitative advice on stock status with sufficient precision to guide fishery management toward optimal harvest levels for the majority of species. Therefore it is essential that the Committee advances in data collection and research on life history, together with describing the interactions with ICCAT fisheries, with the final objective of assessing the status of the stocks and provide adequate scientific advice for the sustainable management of elasmobranch fisheries in the ICCAT convention area. This step forward is critical for the evaluation of the efficacy of the management measures adopted by the Commission in recent years. During the 2012 Shark Species Group meeting, the Group recommended the development of a Shark Research and Data Collection Program (SRDCP) focused on the reduction of the main sources of uncertainty in the formulation of scientific advice, including the improvement of data collection and reporting procedures. Following this recommendation the 2013 Species Group has elaborated the general guidelines of the SRDCP containing the following aspects: (a) a general background of existing fishery and biological data for the main pelagic Atlantic and Mediterranean sharks, highlighting the main gaps of knowledge; (b) the main general objectives of the Program; (c) priorities in fisheries data collection; (d) research priorities on biological information; (e) research priorities on mitigation measures; and (f) other considerations for the SRDCP. The implementation of the SRDCP will be framed within the SCRS Strategic Plan which will provide the overall framework for development and coordination of science and science-related activities needed to support provision of sound scientific advice as the centrepiece for the conservation and management of tuna and tuna-like species in the Atlantic and the Mediterranean. In the case of data poor stocks, as is the case with shark species, a precautionary approach to fisheries management could implicitly account for the unknown uncertainty by being more conservative. And any investment in research will increase the potential benefits of ICCAT fisheries while reducing the risk to the resources. B. PELAGIC ATLANTIC AND MEDITERRANEAN SHARKS Ninety-one species of sharks (sharks and rays) have been reported to ICCAT. Understanding the need to limit the scope of the program, the Species Group considered the species caught (sixteen species represent 95% of the total reported catches) and other species with high susceptibility for which little biological information is available. Species to consider are: (blue (Prionace glauca; BSH), shortfin mako (Isurus oxyrinchus; SMA), 2266

251 longfin mako (Isurus paucus; LMA), bigeye thresher (Alopias superciliosus; BTH), common thresher (Alopias vulpinus; ALV), oceanic whitetip (Carcharhinus longimanus; OCS), silky (C. falciformis; FAL), porbeagle (Lamna nasus; POR), scalloped hammerhead (Sphyrna lewini; SPL), smooth hammerhead (Sphyrna zygaena; SPZ), great hammerhead (Sphyrna mokarran; SPK), sandbar (Carcharhinus plumbeus; CCP), dusky (Carcharhinus obscurus; DUS), night (Carcharhinus signatus; CCS), narrowtooth (Carcharhinus brachyurus, BRO), tiger (Galeocerdo cuvier; TIG), crocodile (Pseudocarcharias kamoharai; PSK), and white (Carcharodon carcharias; WSH) sharks, and the pelagic stingray (Pteroplatytrygon violacea; PLS) and manta rays (Mobulidae, MAN). a) Current biological knowledge Basic life history information required to assess the status of Atlantic shark stocks is most abundant for the North Atlantic area. There is considerably less information for the Equatorial and South Atlantic areas, and very little data for the Mediterranean. Thus, more than half of all studies on age and growth dynamics, reproduction, stock identification, and movement and migration patterns were conducted in the North Atlantic, with the majority corresponding to the northwest Atlantic. Similarly, most of the studies from the South Atlantic correspond to the Southwest Atlantic. Appendix 8-Table 1 summarizes studies conducted for all species combined in each of nine areas in the Atlantic Ocean and the Mediterranean Sea (Appendix 7-Figure 1). Appendix 8-Tables 2-17 show the same information on a species-specific basis for 16 species. The WGSHK will generate similar summary tables for additional species (narrowtooth shark, white shark, crocodile shark, and manta rays). Appendix 9 lists all the references used to generate Appendix 8-Tables Appendix 8 also provides additional references that were used to generate biological profiles for shark and ray species provided by the group. We collapsed all the life history and other parameters listed in the appendix tables into four data categories (reproduction, age and growth, stock ID, and movements and migratory patterns) most relevant for stock assessments and the ten geographical areas into four main areas (North Atlantic, South Atlantic, Equatorial Atlantic, and Mediterranean Sea) and examined that information on a species-specific basis. We used a traffic light approach to identify the degree of knowledge of those categories by general area and species, with: (1) red indicating no studies available at all; (2) yellow, 1 or 2 studies; (3) green, 3+ studies; and (4) white indicating that the species does not occur in a particular area (Appendix 7-Table 1). The following general conclusions can be drawn: the North Atlantic is the most data-rich area, but there are still 25% of cells with no information; the South and Equatorial Atlantic have almost identical levels of data availability, with over 75% of red cells; the Mediterranean Sea is the most data-poor region with about 90% of red cells. Individual species were classified according to the degree of data poorness (i.e., the number of red cells or with no information as a proportion of the total number of cells for that species as depicted in Appendix 7-Table 1) and data richness (i.e., the number of green cells or with 3+ studies as a proportion of the total number of cells for that species as depicted in Appendix 7-Table 1) (Appendix 7-Table 2). The most data-poor species was the longfin mako, followed by the great hammerhead, dusky, and tiger sharks and the pelagic stingray, whereas the least data-poor species was by far the blue shark. In contrast, blue shark, shortfin mako, and sandbar sharks were the most data-rich species and there were no occurrences of data richness for longfin mako, smooth and great hammerheads, and night shark. b) Fisheries information Pelagic sharks form an important part of the catch of the longline fisheries that target tuna, billfish and swordfish. The ICCAT SCRS Sub-Committee on By-catches began to assess pelagic sharks in Pelagic sharks are caught by various gears in the Atlantic Ocean, Gulf of Mexico, Mediterranean Sea and the Caribbean Sea, including longline, purse seine, gillnet, handline, rod and reel, trawl, troll, and harpoon, but they are mostly caught as by-catch in the pelagic longline fisheries or as target species. There are also important recreational fisheries in some countries. Several shark species, such as blue and shortfin mako, are captured and landed in large volumes by these fleets. During the period a total of and tonnes of blue shark and shortfin mako, respectively, were declared in the Atlantic Ocean with a maximum combined catch for both species in 2010 ( tonnes) and a minimum combined catch in 2011 ( tonnes) (Anonymous 2012). Others groups of pelagic sharks and rays are discarded, either due to ICCAT recommendations prohibiting retention (Recommendations 09-07, and 10-08, 11-08), or their low market value. 2267

252 Information on sharks has been submitted by CPCs since 1950, but only since 1982 has data been submitted for shark species other than BSH, SMA and POR. Data prior to 1990 is very limited for most species and so Task I data is only presented here after this date. Appendix 4 provides annual catch reported for all sharks and other elasmobranchs in the Task I database by flag (2012 data are preliminary) while Appendix 5 provides annual catch reported by species and area from Task I. Task II SZ data reporting has only occurred since In order to identify what data are available, this information is presented as a data catalogue in Appendix 6. The first shark assessment meeting was conducted in 2004 and only in 2007 was the independent Shark Species Group formalized. Except for 2010, every year to date there has been an inter-sessional Shark Species Group meeting, with a significant presence of scientists and work on these species. Appendix7-Figure 2 shows the evolution of the number of papers presented at the inter-sessional meetings. c) Species stock assessments The Shark Species Grouphas conducted stock assessments for three species to date: blue, shortfin mako, and porbeagle. Blue and shortfin mako sharks were first assessed in 2004 and subsequently in 2008, and 2012 (shortfin mako only). Porbeagle sharks were assessed cooperatively with ICES in In general, all these assessments are considered preliminary owing to limitations on quantity and quality of the information available and have focused only on Atlantic stocks; Mediterranean Sea stocks have not been assessed owing to lack of data. One important recommendation that consistently emerges from the Species Group meetings is that greater investments in monitoring and research directed at sharks are needed if improved advice on the status of these and other by-catch species is desired. Blue shark Based mostly on tagging information, three separate stocks of blue shark have been assumed to exist, but only two have been assessed (North and South Atlantic) because there was no information on the Mediterranean stock. For both North and South Atlantic stocks, although results continue to be considerably uncertain, biomass is believed to be above the biomass that would support MSY and current harvest levels below F MSY. Shortfin mako Because shortfin makos have a distribution similar to that of blue sharks, the same two hypothetical North and South Atlantic stocks have also been considered for this species. The 2012 assessment of the status of North and South Atlantic stocks included additional time series of relative abundance and increased coverage of Task I catch data with respect to the previous stock assessments conducted in 2008 and The available CPUE series showed increasing or flat trends for the finals years of each series (since the 2008 stock assessment) for both North and South stocks, hence the indications of potential overfishing shown in the previous stock assessment diminished and the current level of catches may be considered sustainable. For the North Atlantic stock, results of the two stock assessment model runs used indicated almost unanimously that stock abundance in 2011 was above B MSY and F was below F MSY. For the South Atlantic stock, all model runs indicated that the stock was not overfished and overfishing was not occurring. Although these results indicated that both the North and South Atlantic stocks are relatively healthy and the probability of overfishing is low, they also showed inconsistencies between estimated biomass trajectories and input CPUE trends, which resulted in wide confidence intervals in the estimated biomass and fishing mortality trajectories and other parameters. Particularly in the South Atlantic an increasing trend in the abundance indices since the 1970s was not consistent with the increasing catches. The high uncertainty in past catch estimates and deficiency of some important biological parameters, particularly for the southern stock, are still obstacles for obtaining reliable estimates of current status of the stocks. Porbeagle The Group attempted assessing the status of four porbeagle stocks (Northwest, Northeast, Southwest and Southeast) in conjunction with the ICES Working Group on Elasmobranch Fishes in In general, data for southern hemisphere porbeagle were too limited to provide a robust indication on the status of the stocks. For the Southwest, limited data indicated a decline in CPUE in the Uruguayan fleet, with models suggesting a potential decline in abundance to levels below MSY and fishing mortality rates above those producing MSY. But catch and other data were generally too limited to allow definition of sustainable harvest levels. For the Southeast, information and data were too limited to assess their status. 2268

253 The northeast Atlantic stock has the longest history of commercial exploitation, but a lack of CPUE data for the peak of the fishery added considerable uncertainty in identifying current status relative to virgin biomass. Exploratory assessments indicated that current biomass (for 2008) was below B MSY and that recent fishing mortality was near or above F MSY. Recovery of this stock to B MSY under no fishing mortality was estimated to take ca years. A Canadian assessment of the northwest Atlantic stock presented at the meeting indicated that biomass was depleted to well below B MSY, but recent fishing mortality was below F MSY and recent biomass appeared to be increasing. Additional surplus production modelling conducted at the meeting indicated a similar view of stock status, i.e., depletion to levels below B MSY and current fishing mortality rates also below F MSY. The Canadian assessment projected that with no fishing mortality, the stock could rebuild to B MSY level in approximately years, whereas surplus-production based projections indicated 20 years would suffice. Under the Canadian strategy of a 4% exploitation rate, the stock was expected to recover in 30 to 100+ years. Ecological Risk Assessment (ERA) Ecological Risk Assessments (ERAs) were conducted by the Shark Species Group in 2008 and The 2012 ERA included 16 species (20 stocks) and was generally believed to be more robust than the 2008 ERA. The ERA consisted of a risk analysis to evaluate the biological productivity of these stocks and a susceptibility analysis to assess their propensity to capture and mortality in Atlantic pelagic longline fisheries or ICCAT longline fisheries. Three metrics were used to calculate vulnerability (Euclidean distance, a multiplicative index, and the arithmetic mean of the productivity and susceptibility ranks). The five stocks with the lowest productivity were the bigeye thresher, sandbar, longfin mako, night, and South Atlantic silky shark. The highest susceptibility values corresponded to shortfin mako, North and South Atlantic blue sharks, porbeagle, and bigeye thresher. Based on the results, the bigeye thresher, longfin and shortfin makos, porbeagle, and night sharks were the most vulnerable stocks. In contrast, North and South Atlantic scalloped hammerheads, smooth hammerhead, and North and South Atlantic pelagic stingray had the lowest vulnerabilities. The information derived from the ERA allows identification of those species that are most vulnerable to prioritize research and management measures. It is apparent from the conclusions of the stock assessments summarized above that there is a lot of uncertainty surrounding the stock assessment results. The SRDCP will address some of the information deficits related to the biology, ecology, and fisheries of Atlantic sharks to reduce the uncertainties of stock assessments and improve the biological and ecological basis for managing and rebuilding some of the stocks. The research plan will also allow a more appropriate evaluation of the efficacy of the ICCAT management measures adopted in recent years. d) Current management ICCAT Recommendations and Resolutions There are currently 12 active ICCAT Recommendations and two active Resolutions that relate specifically to sharks (Appendix 7-Table 3). One additional Recommendation concerning sharks enters into force in May 2013 [Rec ]. Since 2009, four Recommendations have been adopted that prohibit the onboard retention, transhipment, and landing of some shark species that are considered to be vulnerable to overfishing: silky sharks (C. falciformis; [Rec ]), hammerhead sharks (family Sphyrnidae, with the exception of S. tiburo; [Rec ]), oceanic whitetip sharks (C. longimanus; [Rec ]), and bigeye thresher sharks (A. superciliosus; [Rec ]). CPCs are required to record releases and discards of these species, and to report these data to ICCAT. In the case of hammerheads, oceanic whitetips, and bigeye threshers, storing, selling, or offering for sale of any parts or whole carcasses is also prohibited with some exceptions for certain species. Specific exceptions to the above prohibitions apply to certain species. Recommendation also establishes that CPCs should endeavour that vessels flying their flag do not undertake a directed fishery for any thresher sharks (Alopias spp.). Several other ICCAT management measures are currently in place for sharks. CPCs are required to reduce fishing mortality levels for shortfin makos and porbeagles [Rec ; Rec ], to encourage the live release of incidentally caught sharks, particularly juveniles [Rec ], and to consider time and area closures and other measures for pelagic sharks in general [Rec ], and specifically for hammerheads [Rec ] and threshers [Rec ]. In 2013, the SCRS will assess potential management options for silky sharks [Rec ]. 2269

254 Shark finning is prohibited within ICCAT under Recommendation 04-10, which established that vessels should not have fins on-board weighing more than 5% of the weight of shark carcasses on-board, up to the first point of landing. CPCs are required to collect and submit Task I and Task II data for sharks, in accordance with ICCAT data reporting procedures, and also emphasised through multiple Recommendations [Rec ; Rec ; Rec ; Rec ; Rec ]. In the case of Atlantic shortfin mako (I. oxyrinchus), retention of the species became conditional upon the fulfilment of Task I data reporting obligations, from 2013 onwards [Rec ]. From 2014, this condition applies more broadly to other ICCAT species, including sharks [Rec ]. In addition to Task I and Task II data reporting, CPCs are required to report on steps taken to mitigate and reduce levels of by-catch and discards [Rec ]. In 2013, CPCs will also be required to report on their compliance with shark conservation and management measures [Rec ]. Additional Recommendations call upon CPCs to undertake research, where possible, to identify shark nursery areas, and determine ways to increase the selectivity of fishing gear [Rec ; Rec ; Rec ]. CPCs are also encouraged to fully implement national plans of action for sharks [Res ], in accordance with the FAO IPOA-Sharks. Other international measures Sharks and rays captured in association with ICCAT fisheries are subject to management and conservation measures under various international conventions and agreements. Below are examples of such measures, with the relevant species for the Shark Research Program listed in Appendix 7-Table 4. a) Convention for the Protection of the Mediterranean Sea Against Pollution (Barcelona Convention). Species listed under Annex II of the Barcelona Convention SPA/BD Protocol are to be granted maximum protection. Measures include controlling/prohibiting taking, possessing, killing, commercial trading, transporting, and commercial exhibition. Sharks and rays listed under Annex III must be maintained in a favourable state of conservation, through regulation of exploitation and other appropriate measures. b) Convention on the Conservation of Migratory Species of Wild Animals (CMS or Bonn Convention). Sharks and rays listed under CMS Appendix I are to be strictly protected, with retention prohibited, and efforts made to conserve or restore habitats, mitigate obstacles to migration, and control other threats. Signatories to a specific agreement known as the Memorandum of Understanding on the Conservation of Migratory Sharks have committed to implementing measures to conserve and sustainably manage migratory sharks and their habitat, including measures for fisheries research and management, c) Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES). Sharks and rays listed under Appendix II of CITES are subject to controls on their international trade. Export permits or re-export certificates are required, and may be issued only if specimens are legally obtained and if exports will not be detrimental to the survival of the species. For specimens introduced from the sea, export permits are issued by state into which the specimens are being brought. d) General Fisheries Commission for the Mediterranean (GFCM). Recommendation GFCM/36/2012/3 prohibits shark finning, and bans the retention, transhipment, landing, transfer, storage, sale, or display for sale of species listed under Annex II of the Barcelona Convention. The Recommendation also requires the recording and reporting of data about fishing activities, catches, by-catch, release, and discards for species listed under Annex II or Annex III of the Barcelona Convention. e) Past research recommendations by the Shark Species Group Over time, there has been a clear evolution in the scope of shark research recommendations put forward by the Sub-Committee on By-catch ( ), and later by the Species Group (2007-present). Early recommendations focused mainly on the need for better data on catches (particularly incidental catches) and landings, including data on discards. This need has been emphasized repeatedly, with similar recommendations made every year. Since 1997, the Group has also regularly highlighted the need for improved Task II data for sharks. 2270

255 Since the first ICCAT shark stock assessments in 2004, recommendations have also been made for research to improve the quality of output from these assessments. The group has noted the need for increased research into stock structure, life histories, population movements, and dynamics of all ICCAT fisheries that catch sharks, particularly to resolve inconsistent signals among CPUE series. The need for estimating historical catches and size frequencies has also been highlighted, as well as further analyses to assess the sensitivity of assessment outcomes to assumptions. The use of alternative methods for providing management advice has also been recommended, such as ERAs for vulnerable species for which fewer data are available. Since 2006, the Group has called for research to improve the data needed for ERAs. Following the shortfin mako assessment in 2012, the Group also recommended the development and evaluation of hierarchical models that can make use of information from multiple stocks or fleets. The Group has also recommended research to investigate the potential benefits of fishing gear modifications to reduce by-catch, measures for reducing discard mortality, restrictions on fishing areas and times, and minimum/maximum sizes for retention. Broadly, the Group has noted that in order to provide the advice requested of them, and particularly to provide quantitative advice on optimal harvest levels, there is a need for the Commission to make a larger research investment into improved data and facilitating better participation by national scientists and other experts in assessments. C) SHARK RESEARCH PROGRAM General objectives Although efforts are being made in recent years to improve shark data collection and research, the current knowledge on many fisheries and basic biology is still limited. These gaps in knowledge are responsible for much of the uncertainty in stock assessments, and have caused constraints to the provision of scientific advice. Therefore, the present proposal for a Shark Research and Data Collection Program (SRDCP) represents a further step to align with ICCAT Res on Best Available Science, to fill knowledge gaps on fisheries and biology issues by improving data collection, cooperation and capacity building. In order to achieve these goals, the SRDCP aims to provide guidance to SCRS researchers, by prioritizing those issues related to data collection and research lines on species biology/ecology, fisheries and mitigation measures. Finally, by promoting coordination between SCRS researchers, the SRDCP aims to improve the quality and reduce the uncertainty of the scientific advice on sharks provided to the Commission, and to better assess the impact of management measures on these species. 1. Fisheries data collection 1.1 Fleet and gear characterisation Accurate information about the gear characteristics and specifications at which species are captured is fundamental to understanding the impacts of fisheries. The fishing power, selectivity and catchability of fishing gear respond to several variables that must be analysed to understand the evolution of catches. Here are some of these variables: Longline Gear-fish interactions: Time-depth-temperature at depth of hooks (TDRs) Positions of fish regarding other neighbouring fish caught in the longline Fighting time of fish, once hooked (e.g., hook timers) Gear data: Number of baskets along mainline Number of hooks per basket Type and size of hook Presence or absence of lightsticks (also, color differences?) Location (Latitude and Longitude) of the longline set Time of set and haulback (e.g., daylight vs. nighttime) 2271

256 Use of lead weights on gangions Type of gangion Bait type: Live or dead bait Species (e.g., squid vs. mackerel) Purse seine The Minimum data requirements for PS were defined during the KOBE III By-catch Joint Technical Working Group Harmonisation of Purse-seine Data Collected by Tuna-RFMOs Observer Programmes. The main items are the following: vessel identification, vessel trip information, observer information, crew information, vessel and gear attributes, daily activities, school and set information, catch information, length information, species of special interest: Vessel and gear characteristics Fishing strategy Gear-fish interactions Gillnet Locations (Latitude and Longitude) and time at setting and hauling for each set Gillnet configuration Pelagic trawlers Locations (Latitude and Longitude) and time at setting and hauling for each tow Towing speed Characteristics of the net Recreational fisheries Gear type and characteristics Bait Chumming 1.2 Fleet dynamics As sharks are mostly caught as by-catch in ICCAT fisheries, a change in the dynamics of the fleets targeting those resources might have major implications on shark catches. Such changes are related to different issues, such as technological development (e.g., change from traditional longline to semi-automatic Florida style gear; use of high-tech FADs on the purse seine fisheries); shifts on target species as a result of their abundance; changes in the markets, management or piracy (e.g., some fleets change between deep setting for tunas and shallow setting for swordfish throughout the year, and consequently might change the fishing gear characteristics hook style, bait type, gangion material, etc., and fishing regime from day to night setting; while the purse seine fleets might impact differently sharks, when changing from free schools to FADs fishing); fleet movement between fishing areas throughout the year (e.g. due to the migratory behaviour of the target species, communication between skippers related with the occurrence of higher catches, exploitation costs related to bait or fuel prices, piracy, etc.). 1.3 Data necessary for assessment and management advice Catch (landings + discards) Effort Catch per effort (indices of relative abundance) Gear selectivity (if not fitted within the model) Size information 2272

257 Catches: Catch inputs for stock assessment can vary from highly aggregated information (e.g., catch of sharks ) to different levels of disaggregation and detail, ranging from nominal catches by species to species-specific catch series by gear, geographical area, and size. Dead discards: Estimation of dead discards can also be based on expansion to total numbers from a low number of observations to expansion to total numbers based on a high degree of observer coverage of the fleet and fine level of stratification (season or month, small areas of observation). Typically logbook and observer data are used to generate estimates of dead discards. Effort: Effort series by gear (e.g., number of hooks) and geographical area can also be used in several assessment methodologies. Indices of relative abundance: Indices of relative abundance can also vary from simple, nominal CPUE time series of short duration (few years) and with little contrast (one-way ticket) to (preferably fishery-independent) CPUE time series standardized through different statistical techniques (GLM, GLMM, GAM). Ideally these indices should be of long duration and wide geographical coverage and have good contrast (increasing and decreasing trend resulting from various levels of fishing). Selectivity: When sufficient length or age information is not available to estimate selectivity within the model, selectivity curves for the different abundance indices have to be generated based on auxiliary information externally to the model and then imputed as functional forms in age-structured models. Size information: No catch-at-age is available for sharks caught in ICCAT fisheries, but limited length-frequency information is available for some species. 2. Data poor assessment models Because of the lack of total catch information in some cases and some key biological information in other cases, traditional stock assessment models cannot be consistently applied to all species. There is a need for development of innovative methods of assessment of shark resources, particularly methods applicable to datapoor situations. Fortunately, a number of such methods that require different types and amounts of data have recently been developed (Appendix 7-Table 5). 2.1 Ecological Risk Assessment (ERA) Ecological Risk Assessments (ERAs), also known as Productivity and Susceptibility Analyses (PSAs), were originally developed to assess the vulnerability of stocks of species caught as by-catch in the Australian prawn fishery (Stobutzki et al. 2001a, b; Milton 2001), and although they only appeared about a little over a decade ago they have now been used rather extensively to assess vulnerability to fishing of elasmobranch fishes and other marine taxa. Ecological risk assessments are in fact a family of models that can range from purely qualitative analyses in their simplest form to more quantitative analyses, depending on data availability (Walker 2005b; Hobday et al. 2007). Most PSAs have been semi-quantitative approaches where the vulnerability of a stock to fishing is expressed as a function of its productivity, or capacity to recover after it has been depleted, and its susceptibility, or propensity to capture and mortality from fishing (Stobutzki et al. 2001a). Each of these two components, productivity and susceptibility, are in turn defined by a number of attributes which are given a score on a predetermined scale. Scores are then typically averaged for each index and displayed graphically on an X-Y plot (PSA plot). Additionally, vulnerability can be computed, for example, as the Euclidean distance of the productivity and susceptibility scores on the PSA plot. Applications to elasmobranch fishes have ranged from semi-quantitative PSAs (Stobutzki et al. 2002; Griffiths et al. 2006; Rosenberg et al. 2007; Patrick et al. 2010) to different degrees of quantitative analyses where the productivity component was estimated directly as r (maximum rate of population growth) in stochastic demographic models (Braccini et al. 2006; Zhou and Griffiths 2008; Simpfendorfer et al. 2008; Cortés et al. 2010; Tovar-Avila et al. 2010). The main advantages of PSAs can be summarized as: (1) being a practical tool to evaluate the vulnerability of a stock to becoming overfished based on its biological characteristics and susceptibility to the fishery or fisheries exploiting it, (2) they can be used to help management bodies identify which stocks are more vulnerable to overfishing so that they can monitor and adjust their management measures to protect the viability of these stocks, and (3) they can also be used to prioritize research efforts for species that are very susceptible but for which biological information is too sparse. 2273

258 2.2 Length-based models: SEINE (Survival Estimation in Non-Equilibrium Situations) One of the simplest data-poor methods is based on the premise that fishing pressure proportionally removes larger and older fish from the population and that increases (or decreases) in mortality rates are reflected by decreases (or increases) in mean length. These approaches generally have minimal data requirements and are therefore appealing for use in many elasmobranchs, but they have stringent assumptions which can sometimes be difficult to meet in long-lived species. The SEINE method (Gedamke and Hoenig 2006) is a reformulation of the widely used Beverton-Holt (1956, 1957) method, which only requires von Bertalanffy growth parameters, a size at full vulnerability, and mean length of fully vulnerable animals, and relaxes the assumptions that growth, recruitment, and mortality have been in equilibrium for a time period equal to at least the maximum age of the species of the Beverton-Holt method. This non-equilibrium formulation allows for trends to be inspected through a time series analysis of mean length data and provides the ability to estimate multiple mortality rates and the year(s) in which mortality changed. However, application of length-based approaches to relatively long-lived elasmobranchs should be done cautiously and model assumptions should be carefully considered prior to application an when interpreting results and producing management advice. 2.3 Age-structured Demographic Models (Life Tables/Euler-Lotka equation; Leslie Matrices) and Elasticity Analysis Demographic analyses of elasmobranch populations can be undertaken as (1) life tables based on a discrete implementation of the Euler-Lotka equation or (2) age-based Leslie matrix population models. These models are typically based on deterministic, density-independent population growth theory, whereby populations grow at an exponential rate r and converge to a stable age distribution. Data requirements include maximum age, survival from natural mortality, age-specific fecundity (the number of offspring produced per breeding female of age x), sex ratio at birth, frequency of parturition, proportion of mature or breeding females at age, and some associated information such as growth function parameters and a length-mass relationship. Elasticity analysis is an extension of age-based Leslie matrices or stage-based models that allows one to identify which vital rates influence population growth rate the most and thus which life stages (or ages) are more important for population growth. 2.4 Analytical Reference Points Methodology to analytically calculate reference points without an assessment model was first introduced in Brooks et al. (2006) and Brooks and Powers (2007), where it was demonstrated that reference points corresponding to maximum excess recruitment (MER; Goodyear 1980) could be derived simply from biological parameters and an assumption about the form of the stock recruit function. Brooks et al. (2010) re-derived those analytical solutions to calculate the Spawning Potential Ratio (SPR) at MER, then demonstrated how stock status could be determined given auxiliary information, and illustrated the method for 11 shark stocks. Although only vital rates are necessary to derive these analytical reference points, an estimate of current biomass or a time series of relative abundance is needed to evaluate the overfished criterion. Although this methodology has to be further tested, initial results are encouraging. Brooks et al. (2010) compared results for overfished status from stock assessments with predictions from the analytical method and found total agreement for the nine stocks of sharks for which an estimate derived from a more data-rich stock assessment method was available. 2.5 DCAC (Depletion-Corrected Average Catch) The DCAC is based on the potential yield formula of Alverson and Pereyra (1969) and Gulland (1970) where B MSY = 0.5B 0, F MSY = M, and Y pot = 0.5M B0. If abundance is reduced from B 0 to B MSY, a windfall harvest can be calculated as W = 0.5 B0 and Y pot can be considered a sustainable annual yield. The windfall ratio expresses the magnitude of the windfall harvest relative to a single year of potential yield. This windfall ratio forms the basis for a depletion correction of average catch. For a catch series of n years, the total cumulative catch consists of n years of sustainable production plus a windfall equivalent to W/Y pot years of potential yield. The DCAC ultimately provides an estimate of the yield that could have been sustained during a period of n years. 2274

259 2.6 AIM (An Index Method) The AIM (An Index Method, NOAA Fisheries Toolbox 2011) model is an analytical framework for interpreting abundance trends, which relates survey trends to fishery removals. The AIM model estimates a relative fishing mortality rate from a ratio of catch to a smoothed index of abundance. The second calculated quantity is the replacement ratio, which is obtained by taking the abundance index values divided by a moving average of the abundance index. The idea behind the replacement ratio is that values greater than one indicate that the population increased while values less than one suggest negative population growth. A regression of the natural logarithm of the replacement ratio against the natural logarithm of relative F can be solved for the relative F value that produces ln(replacement ratio)=0, i.e. stable population growth. The F producing stable growth can be considered as an F reference point, against which the relative F time series can be compared to evaluate overfishing. Implicit in this approach is that the catch and abundance index have the same selectivity. This methodology fundamentally assumes linear (density-independent) population growth. Furthermore, there is no age structure, thus biological parameters that have strong age trends or long time lags in population dynamics owing to late, protracted maturation and generation time are ignored. 2.7 Surplus Production Models Biomass dynamic models, also known as (surplus) production models, have been and still are fairly widely used in the assessment of teleost stocks. Use of these models in assessment of elasmobranch stocks, however, has been criticized because of violation of the underlying assumptions, notably the presupposition that r responds immediately to changes in stock density and that it is independent of the age structure of the stock (Holden 1977; Walker 1998). In general, production models trade biological realism for mathematical simplicity, combining growth, recruitment, and mortality into one single surplus production term. However, they are useful in situations where only catch and effort data on the stock are available and for practical stock assessments because they are easy to implement and provide management parameters, such as maximum sustainable yield (MSY) and virgin biomass (Meyer and Millar 1999a). The biomass dynamic models used in the last decade have characterized uncertainty through the use of either Bayesian inference or classical frequentist methods. Typically, in stock assessment work two stochastic components must be taken into consideration (Hilborn and Mangel 1997): natural variability affecting the annual change in population biomass (also known as process error) and uncertainty in the observed indices of relative abundance owing to sampling and measurement error (observation error). Bayesian surplus production models have been used by a number of researchers to assess the status of shark populations. The Bayesian Surplus Production model (BSP; McAllister et al. 2001; McAllister and Babcock 2006), a Schaefer production model that uses the SIR (Sampling Importance Resampling) algorithm for numerical integration, has now been used in numerous assessments of shark stocks in the Atlantic Ocean (McAllister et al. 2001, 2008; Cortés 2002b; Cortés et al. 2002, 2006 to cite a few). The BSP considers observation error only, which is integrated along with q (catchability coefficient) from the joint posterior distribution using the analytical approach described by Walters and Ludwig (1994). Both process and observation errors can be incorporated when using a dynamic state-space modelling framework of time series (Meyer and Millar 1999a). This approach relates observed states (CPUE observations) to unobserved states (biomasses) through a stochastic model. State-space models allow for stochasticity in population dynamics because they treat the annual biomasses as unknown states, which are a function of previous states, other unknown model parameters, and explanatory variables (e.g., catch). The observed states are in turn linked to the biomasses in a way that includes observation error by specifying the distribution of each observed CPUE index given the biomass of the stock in that year. A Bayesian approach to state-space modeling has only been applied fairly recently to fisheries (Meyer and Millar 1999a). One advantage of using a Bayesian approach is that it allows fitting nonlinear and highly parameterized models that are more likely to capture the complex dynamics of natural populations. Meyer and Millar (1999a, b) advocated the use of the Gibbs sampler, a special Markov chain Monte Carlo (MCMC) method, to compute posterior distributions in nonlinear statespace models. This Bayesian nonlinear state-space surplus production model has been adapted and applied in several assessments of Atlantic shark stocks (Cortés et al. 2002, 2006). Additionally, Jiao et al. (2009) compared hierarchical and non-hierarchical Bayesian production models applied to a complex of three hammerhead species (Sphyrna lewini, S. mokarran, and S. zygaena) to address the problem of assessing fish complexes for which there are no species-specific data. They found that the fit of the Bayesian hierarchical models was better than that of the traditional Bayesian models possibly due to the addition of multilevel prior distributions, among which was a multilevel prior of r intended to capture the variability of intrinsic rates of increase across species and populations of the hammerhead shark complex. 2275

260 3. Recovery of historical data Recently, The ICCAT Small Tunas Research Program was adopted by the SCRS and the first phase of that research program is to recover historical SMT datasets, available in various scientific institutions of ICCAT CPCs and currently not available in the ICCAT database. The data recovery includes: Task I nominal catch series by species, gear, area, flag and year Task II catch and effort statistics by species, month, 1ºx1º square or area Task II size (and/or weight) samples by species, gear, time strata and area In order to address this issue, a call for tenders was drafted with the specific objective of recovering historical time series from all fisheries in the ICCAT Convention area, either fisheries targeting small tunas or those catching them as by-catch. A similar procedure was followed in the context of the special research programme of Bluefin tuna (GBYP). These processes should be repeated for recovering historical datasets on shark species. The group can build on the experience obtained during those exercises in order to streamline and facilitate this important initiative. 4. Trade data Trade data are a potentially useful complementary source of information for the management and assessment of shark species caught in association with ICCAT fisheries. Identifying trends and changes in the trade of shark products (e.g., trade routes, volumes, values) may in turn help our understanding of the dynamics of fisheries capturing sharks. In the specific context of shark assessments, historical and current trade data may be used to identify potential gaps in reported catches and to develop proxy indices for estimating historical catches. Trade data have been used in one ICCAT shark assessment meeting. At the 2004 assessment of blue shark and shortfin mako, the group discussed an analysis of the Hong Kong shark fin trade that provided rough annual estimates of the capture of sharks in the Atlantic. In view of these estimates and the very incomplete nature of catch reporting to the Secretariat for blue sharks and shortfin makos, the assessment group explored the use of an alternative approach for estimating catch histories, based on the ratio of shark to tuna landings. Following the 2004 assessments, the group recommended that broader use be made of trade statistics, particularly for extending historical time series of catch estimates. 5. Observer Programmes As stated by FAO (1995) in order to have a responsible and sustainable management of fisheries, fishing countries need to assure the timely, complete and reliable collection of fishery statistics on catch and fishing effort. Such data needs to be updated regularly and submitted to the relevant RFMO to be used in the fishery assessment and for the provision of the scientific advice. The FAO Code of Conduct for Responsible Fisheries also states that fishing countries should implement effective fisheries monitoring, control, surveillance and law enforcement measures including, where appropriate, observer programmes, in order to collect basic fishery statistics. In the case of pelagic sharks, which are often caught as a by-catch (and discarded) within ICCAT fisheries, it is essential to implement Observer Programmes. In fact, whether fisheries management objectives include conservation issues, knowledge of shark fishing mortality is essential for any management framework, and observer programmes are the most reliable source of information for these species. Moreover, observer programmes are the only available method to accurately collect data on a number of important issues, such as: individual at-haulback mortality, fate and status when discarded; samples for less common or rare species; etc. ICCAT recommendations regarding the observer programs (GFCM adopts ICCAT resolutions in relation to sharks in the Mediterranean Sea, although the adoption by GFCM is usually carried out with a time lag) and current coverage are presented in Appendix 7-Table 6. When designing an observer program the level of coverage required is a key element. It depends on the objectives of the observer program (e.g. desired precision levels for by-catch rates, and the variability of the bycatch events, which depend on specific taxa and fishery combinations). In the case of the tuna fisheries impacting shark, the observer program should collect data aiming at the: (i) improvement of catch data collection for population assessments; (ii) estimation of by-catch and discards levels; (iii) collection of basic biological data; and (iv) gear and fisheries strategy. 2276

261 In most cases by-catch estimates are highly imprecise for observer coverage below 5-10%, therefore observer coverage rates above those levels will be required. By-catch estimates will remain highly imprecise for low occurrence species, for which a much higher level of coverage may be warranted. In general, the species composition of the sharks captures is similar amongst the different tuna fisheries in the convention area. However, the different fisheries may impact differently the shark species: Longline (sensu lato) impacts mainly blueshark (BSH) and shortfin mako (SMA), and to a minor extent hammerhead, threshers, silky and oceanic whitetip sharks; Gillnet (sensu lato) are impacting mainly silky (FAL), thresher (THR), oceanic whitetip (OCS), and shortfin mako (SMA) sharks; Purse seine are impacting mostly oceanic whitetip (OCS) and silky (FAL) sharks. Industrial fleets are amongst those that mostly impact shark stocks within the scope of tuna fisheries. The implementation of scientific observer programmes designed to improve shark data collection should focus on the two major fleets: pelagic longliners, namely those targeting swordfish or tropical tunas; and purse-seiners targeting tropical tunas. Although artisanal fleets may have considerable impact over some protected species, the small size of the vessels is an important constraint for an observer program. Therefore, other data collection schemes should be implemented for these fisheries. 6. Biological information 6.1 Stock structure To better understand the impact of fishing activities on elasmobranch populations and promote a more efficient management of their fisheries, it is first necessary to know whether elasmobranchs are migrating between regions that can be undergoing different types and levels of fishing activity. However, and even though those issues are of great importance, there is still very limited information on the stock structure of most pelagic elasmobranchs at an ocean wide level, and therefore promoting those types of studies is of utmost importance. Using incorrect assumptions about the stock structure and movements can lead to biased conclusions about the level of fishing that is sustainable in a given region, and thus information about these processes should be incorporated into stock assessments. Different approaches can be used in identifying and classifying stocks. However, given the difficulties and possible limitations of each of the techniques, and in order to provide the most accurate identification of stocks possible, scientific knowledge should gather different sources of information and consequently, a multidisciplinary approach using a combination of techniques is recommended Genetic studies Studying the genetic structure of a population can be a very useful tool for helping to determine whether there is migration between geographic areas. When the individuals of a species segregate into several reproductive stocks, the allele frequencies at neutral genetic markers diverge such that the variance in gene frequencies reflects the magnitude of reproductive isolation among these stocks (Heist, 2004). However, there also difficulties with the population genetics studies in the open ocean species as, for example, a small number of migrants per generation may be sufficient to render two populations genetically indistinguishable (Camhi et al., 2008). Several types of molecular markers have been used to estimate the stock structure in marine populations in the last decades (Utter, 1991). The choice of the technique to use depends on the research team capabilities, preferences, type of equipment available and quality of tissue available for analysis. In general, the molecular markers that have been used include allozymes, mitochondrial DNA and microsatellites, even though other techniques are also available. Each technique has its own distinct strengths and weaknesses, and reviews of those are presented in Heist (1999, 2004, 2008). A final consideration regarding genetic studies on pelagic sharks is that these species may undergo large scale seasonal migrations, and may segregate by sex and/or maturity stage. As such, careful planning of where and when to sample and collect tissues is very important. 2277

262 6.1.2 Biometric analysis The biometric analysis, including meristic and morphometric characters, provides a powerful complement to genetic stock identification approaches. Meristic characters generally include serially repeated measurements such as counts of vertebrae. Experimental work has shown that environmental factors such as temperature, salinity and oxygen tension can modify the expression of the genes responsible for meristic characters. In certain studies meristics have provided evidence of stock structure that is concordant with genetic information Population parameters Typical population parameters that are useful for population dynamics studies include age, growth and reproductive parameters that can then be used to estimate mortalities and intrinsic population growth rates. Different populations from one same species may show different biological parameters, and those should be taken into account during population dynamics studies and stock assessments. Further, as different populations from the same species may be subjected, through time, to different fishing pressures and mortalities, densitydependent mechanisms may also produce changes in the biological parameters and affect the dynamics of the populations. Those differences may be observed through comparative studies on the biological parameters across several populations of one species, and may serve as verifications from other stock structure methodologies. Some studies have used this approach for sharks, trying to determine possible stock separations based on life history parameters, but most have been carried out in coastal sharks. Examples are the works by Carlson and Parsons (1997), Yamaguchi et al. (2000) and Coelho et al. (2010). These comparative techniques have not commonly been applied to pelagic sharks, even though their importance is recognized and for stock assessment purposes (including ecological risk assessments) different biological parameters are used for each of the stocks (North Atlantic, South Atlantic and Mediterranean). In terms of methodologies, details on data collection and analysis for using such parameters for eventual comparison between regions are specified in Section 6.2 (life history information) of this research plan. This component of the plan may help to separate stocks, and may produce important biological parameters for using for each of the stocks Tagging The conventional approach of mark-recapture can be used. Recoveries through time provide ranges and patterns of movement, which can assist in inferring the degree of mixing among stocks. However, the success of such techniques depends largely on tagging and recapture efforts, and such studies are generally constrained by higher costs. The use of satellite tagging technology is encouraged as this type of tag transmits data on animal location without animal recapture, making them completely fishery-independent. Furthermore, these tags provide locations on intermediate positions and not only two observations in space-time (capture and terminal recapture) as with the conventional mark-recapture approach. A shortcoming of some types of satellite tags (e.g., pop-up tags) is that light-based location estimates can have substantial errors, and this can limit their advantage over conventional tags. These tags are electrically powered (by batteries, solar power, kinetic energy, etc.) leading to shorter times at-liberty on average than conventional tags Parasites (biological tags) Information on geographic distribution patterns, migrations and feeding habits of fish can be obtained through the study of parasites. The investigation of hosts and their parasites has improved the knowledge about the spatial distribution of the host s population (Abaunza et al., 2008). Lester and MacKenzie (2009) provide a guideline on how to use parasites as biological tags in fish population studies. In the Atlantic, for example, Garcia (2011) used parasites, as a complement to other techniques, to discriminate between stocks of swordfish (Xiphias gladius). 6.2 Life-history information Age and growth An understanding of the age structure and growth dynamics of a population is crucial for the application of biologically realistic stock assessment models and, ultimately, for effective conservation and management. Information on age and growth is also often used to estimate natural mortality or total mortality, which are crucial components of stock assessment models, and in the calculation of important population and demographic parameters, such as population growth rates and generation times. Successful fisheries management thus requires precise and accurate age information to make informed decisions, because inaccurate age estimates can lead to 2278

263 serious errors in stock assessments and possibly to overexploitation (Campana 2001). Despite their importance, published age and growth studies of sharks are still scarce and only a few have provided validation of the ageing method used (i.e., ratification through a direct method, such as injection of a chemical marker, that the growth bands on the structure being aged are deposited with a given periodicity, generally annually). Because sharks lack hard parts, such as large scales and otoliths, information on age and growth in sharks is usually derived from counts of opaque and translucent bands on vertebral centra or spines. Processing of samples is laborious and requires many hours in the laboratory. Preparation of vertebrae for ageing involves several steps. To enhance visibility of growth bands, vertebrae can be cut in half sagitally or sliced at varying thicknesses. Depending on the species, sections can be stained with various chemicals to enhance the growth bands (e.g. crystal violet, alizarin red). Opaque and translucent bands are counted by placing a section under a dissecting microscope interfaced with an image analysis system. Generally, two biologists read samples blindly (i.e., no knowledge of length or sex of specimen) and age estimates for which the readers agree are re-read using digitally stored images. Historically, the von Bertalanffy growth model (von Bertalanffy, 1938) has been the model applied to most elasmobranchs (Cailliet and Goldman, 2004), but alternative growth models have also been applied in recent years (Carlson and Baremore 2005; Neer et al. 2005, Coelho and Erzini, 2007, 2008). Many of these models still lack age validation and suffer from small sample sizes for some age groups. To resolve these issues, collaborations among scientists from several ICCAT CPCs and institutions are encouraged to develop more complete models. Another promising means of age validation for long-lived species is bomb radiocarbon dating. This technique focuses on the well-documented increase in radiocarbon (C 14 ) in the world s oceans, caused by the atmospheric testing of atomic bombs in the 1960s (Druffel and Linick 1978). The increase in atmospheric and oceanic radiocarbon was found to be synchronous with marine organisms containing carbonate, such as bivalves, corals, and fish bones (Kalish 1993, Weidman and Jones 1993, Campana 1997). This synchrony allows the period of increase to be used as a dated marker in calcified structures exhibiting growth bands, such as teleost otoliths and shark vertebrae (Campana et al. 2002a). This technique has been successfully used to validate the age estimation of the porbeagle shark (Lamna nasus), and has met with some success for a single shortfin mako (Isurus oxyrinchus, Campana et al. 2002b), and two great hammerheads (Sphyrna mokarran, Passerotti et al. 2010). Some previous work by Kerr at al. (2004) on the white shark (Carcharodon carcharias) also showed promise. This technique could further aid in the age validation and population assessment of many long-lived elasmobranch species. Funding would enable collaboration with colleagues that are experts in the application of this specific technique Reproductive biology Knowledge on the reproductive biology is essential for stock assessment models that attempt to accurately capture the biology of a species, such as age- and sex-structured models. Minimum size limits, for example, are usually set after consideration of the size at which most individuals become sexually mature. Female sharks tend to mature at a later age and larger size and reach a larger size and older age than their male counterparts. This pattern is reflected in the respective growth curves of each sex, and needs to be taken into account in stock assessments. Length of the reproductive cycle (specifically, how often females reproduce), the number of offspring per litter for females of different sizes or ages, and the proportion of mature and pregnant females at each size or length, are all needed to calculate fecundity, which is one of the main inputs to any demographic analysis or stock assessment. Incorrect estimation of any of these reproductive parameters will affect estimates of fecundity, biasing ensuing demographic analyses and stock assessments. In elasmobranch fishes reproductive patterns are commonly characterized by late sexual maturity, reproduction every one, two or even three years, long gestation periods, reduced fecundity, and well-developed, highly mobile offspring with relatively low natural mortality. But information on the reproductive biology of many species, even some commonly exploited, is still fragmentary. Funding would allow us to conduct studies on the reproductive biology of several important species in Atlantic waters, with the ultimate goal of providing information for stock assessments. Funds are needed to increase sampling efforts and expand the number of species currently being examined Mortality There are few direct estimates of instantaneous natural mortality rate (M) or instantaneous total mortality rate (Z) for elasmobranchs based on mark-recapture techniques or catch curves. Direct estimates of natural mortality 2279

264 have been obtained in mark-depletion experiments for juvenile lemon sharks only and estimates of M derived from Z or Z directly, in mark-recapture studies for a few species. The majority of population modelling studies for elasmobranchs has relied, however, on indirect estimates of mortality obtained through methods based on predictive equations of life history traits. Most of these methods make use of parameters estimated from the von Bertalanffy growth (VBG) equation, including those of Pauly (1980), Hoenig (1983), Chen and Watanabe (1989), and Jensen (1996) (see Roff 1992; Cortés 1998, 1999; Simpfendorfer, 1999a, 2005 for reviews of these methods) amongst others. These equations do not yield age-specific estimates of natural mortality, except for the Chen and Watanabe (1989) method. In contrast, methods proposed by Peterson and Wroblewski (1984) and Lorenzen (1996, 2000) allow estimation of size-specific natural mortality, which can then be transformed into age-specific estimates through the VBG function. The use of U-shaped curves (Walker 1998) has also been advocated to account for the fact that individuals must die off in their terminal year of life. A modified U-shape curve, the so-called bathtub curve (Chen and Watanabe 1989; Siegfried 2006) has been proposed for elasmobranch fishes because the initial decrease in natural mortality (M) at young ages is followed by a flatter profile, and M only increases sharply towards the oldest ages, possibly due to senescence. 6.3 Ecosystem-based approaches Shark Trophic Studies, Foraging Ecology and Bioenergetics Fisheries management bodies (FMBs) have, in recent years, stressed the need for an ecosystem approach to management. The current work carried out so far for sharks gives little consideration to ecosystem function because there are few quantitative species-specific data on competition, predator-prey interactions, and habitat requirements of sharks. To fully understand how sharks utilize ecosystems and interact with other species, more studies on diet, habitat use, and ecosystem modelling, are needed. To fully evaluate the impacts of sharks within the ecosystem, diet data incorporating published metabolic rate information (see review in Carlson et al. 2004) and excretion and egestion information (see review in Wetherbee and Cortés 2004) can be used to construct bioenergetic models for shark populations. Bioenergetic models can be used to assess shark predatory effects (i.e., consumption rates) on prey abundance, and the consequences of the reduction in predation rates through an increase in fishing mortality on shark populations. An example is the bioenergetics model constructed for cownose ray Rhinoptera bonasus, which was used to determine the relative effects of variation in different environmental variables on growth (Neer et al. 2004). Individual growth from the bioenergetics model can also be used for developing matrix projection models, which are designed to simulate the long-term population dynamics of, and examine how, various harvesting strategies would affect long-term stock status. Although it is commonly accepted that sharks are apex predators in many marine communities (Wetherbee and Cortés 2004), there are very few estimates of trophic levels (Cortés 1999). An alternative to estimating trophic level based on stomach contents is the use of stable isotopes of nitrogen and carbon from tissues of marine consumers. This approach is being increasingly used to estimate the trophic position of sharks in marine food webs, and potentially provides a viable alternative to diet-based estimation of trophic levels Habitat use Quantifying fish habitat use is important for management of fish populations and conservation planning. Habitat use studies are used to document habitat quality and its specificity to life history stages. Knowledge of movement patterns (i.e. use of space and activity patterns) is essential in understanding the behaviour of a species as well as defining essential habitat for that animal. An animal's movement patterns can have profound effects on its energetics, reproductive fitness, and survival (Matthews, 1990). Unlike animals in coastal marine environments, which may be able to utilize more definitive landmarks for navigation (e.g. bathymetry), pelagic predators have to rely on cues, which may be more difficult to define (e.g., geomagnetic gradients). Despite these limitations, there can still be predictable locations of abundant prey, such as within thermal fronts, and these have long been known as areas of high fish abundance (Block et al. 2011; Queiroz et al. 2012). Oceanographic conditions are likely to be strong drivers of the movements and distribution of pelagic sharks (Queiroz et al. 2012). 2280

265 To better understand the influence of the marine ecosystem on species habitat use, the collection of oceanographic information (e.g., sea surface temperature, chlorophyll concentration, current velocity, depth of the thermocline, oceanic fronts, and upwelling) is necessary. This information can be collected in situ or through remote sensing techniques. The ability to collect data on pelagic fish movement and its relationship with the environment has greatly increased with the latest advancements in technology, as is evidenced by a vast array of satellite telemetry and other types of research (Campana et al. 2011) Essential fish habitat and migratory patterns Better management of shark populations through habitat protection is the goal of the mandate to describe and identify essential fish habitat. This recognizes that all stages in a species life cycle are important, not just those stages vulnerable to exploitation. However, because of their migratory nature, identifying essential fish habitat (EFH) for pelagic sharks is very challenging. Using advanced technology can improve identification and quantification of EFH for sharks. This includes using acoustic listening stations to monitor the movements of some stages of sharks, even though the application of such techniques in the open ocean has severe limitations. Still, some work using this system may provide information on home range size and changes in habitat use through time, shark distribution in relation to prey density, timing of immigration and emigration, observation of philopatric behaviours (i.e., whether sharks return to their natal grounds), examination of intraspecific relationships (e.g. aggregation, competition, and group dynamics), and assessment of mortality rates within the population. Stable isotope analysis and microchemistry are also two expanding fields of research. While stable isotopes such as N 15 and C 15 have traditionally been used to study food web structure and estimate trophic level (see previous section), researchers are now using stable isotopes also to track movement of individual fish using these chemical signals as natural markers. Hardpart microchemistry of rare elements such as strontium can also be used to examine fish movements between natal and breeding grounds. Both techniques have shown promise for bony fishes, while research in elasmobranchs is still very preliminary. The Group can move to support investigation on these techniques and to gain insight into the migratory patterns, stock structure, and mixing rates of important shark species, all factors important to improve stock assessments Habitat and ecosystem-based modelling Several approaches have been used to predict potential fish distributions based on models of a species habitat use. For example, ecological niche modelling has been used to predict the potential ecological and geographic distribution to a variety of wildlife species. A niche is an ecological construct defining the optimum environment for growth, reproduction and survival of a species. One way to investigate species response to habitat is through examination of habitat preferences by constructing environmental niche models. Information on fish vertical movement in the water column collected by satellite tags can also be incorporated in habitat-based standardization (HBS) models (Bigelow et al. 1999). In HBS models effective effort is modelled as a function of the probability of interaction between the depth distribution of hooks and species in the water column. This model also requires information on gear configuration (e.g. hook depth). Ecosystem models are also being developed to provide some insight into the function of marine ecosystems and their potential responses to natural and anthropogenic disturbance. One particular important question is evaluating how the removal of apex predators through fishing and other sources of mortality will affect the overall ecosystem function. This question takes on increasing importance in light of the observation on the reductions of higher trophic level species and fishing down food webs proposed by Pauly et al. (1998). On the other hand, recent modelling work on a small scale coastal area found that reduction of abundance of certain sharks as a result of increases in fishing mortality did not cause considerable structural changes in the overall system (Carlson 2007). Some additional modelling work in the North Pacific Ocean also found that reducing one or a few shark groups does not cause top-down effects because of complementary increases in other apex predator groups, which were apparently filling empty niches (Kitchell et al. 2002). However, modelling work of a rocky reef system indicated that sharks might be strong shapers of that marine community and that considerable modifications might already have transpired due to removal of sharks by Galápagos fisheries (Okey et al. 2004). Such studies should also be investigated in the pelagic environment, aiming at the development of methods to further test similar hypotheses. In addition, a number of hypotheses related to the effectiveness, size and design of possible marine reserves in the open seas could be evaluated. 2281

266 7. By-catch mitigation Several research projects are being developed to mitigate by-catch, primarily for birds, turtles and mammals. Some of this research includes bird scaring (tori) lines, the use of dyed bait, testing underwater hook setting devices, devices to avoid entanglement of seabirds in trawl warp cables, the use of circular hooks, the use of equipment for the release of wildlife after capture, studies on habitat use, and possible application of TEDs in trawl fisheries. Also modifications of fishing gear for turtles, use of reflecting nets and acoustic alarms for mammals, and studies on the behaviour and habitat use of sharks. Conducting ecosystem-level studies on the collateral effects of fishing, such as the removal of species with high trophic value remains a priority. Research is also being conducted on ways to reduce shark by-catch (benefits of banning wireleaders, hooks that repel sharks, changing soak depth, hook type, bait type, etc.). A current practice on-board fishing vessels is to dump unwanted sharks and rays overboard in different ways. Indeed, sharks, and to a lesser extent rays, are usually considered by fishers as tough animals and they assume that they can easily survive when returned to the sea. Nevertheless, there are uncertainties about the post release fate of these individuals and survival rates of sharks and rays are likely to be variable among species within a fishery. Developing and promoting practices that maximize the health of sharks and rays when they are handled and released is fully justified. For the major gears impacting sharks and rays, good practices identified should be transferred to fishers and the implementation of these practices on-board monitored. 8. Other Considerations for the shark research programme 8.1 Capacity building One of the largest challenges facing enlightened fishery management is the procurement of accurate and robust catch, effort, landings, location and depth data. Although there have been improvements, in many areas of the Atlantic and Mediterranean such data collection is lacking or incompletely gathered. A concerted effort to enhance data gathering abilities in these regions should be a priority with the goal of bringing the quality and quantity of data up to currently accepted standards. Accurate identification and quantification at the species level is a fundamental imperative. Identification of species often is a difficult task as many species of elasmobranchs are similar in appearance and errors in identification are readily made, even by experts. Although identification guides have been produced for many key areas, learning to differentiate species is markedly enhanced by a hands-on learning experience. Led by recognised identification specialists, workshops for indigenous biologists held in their home areas using local biota are invaluable in producing quality control in this most basic of data gathering steps. An understanding of prioritized gathering of data categories is essential and details such as use of standardized length measures (TL, FL, PCL, DW), external sex determination, morphological signs of maturity, etc. must be established and recorded uniformly. The availability of pre-existing field-tested data-sheets and knowledge of when and where to modify fields as needed for local conditions can be shared, saving local scientists from going through trial-and-error periods which result in faulty or sub-par data collection. Archival of old data is to be encouraged along with notations describing the methodologies employed. As with species identification, a workshop setting involving well-trained instructors and local biologists has proven to be a profitable approach to build capacity in this arena. Knowledge of basic laboratory techniques is often poor or absent in many regions. Use of sectioned morphological hard parts and validation techniques to determine age and modern approaches to document reproductive biology must be employed as these life history parameters can vary locally and are essential in the assessment process. Hands-on training sessions involving veteran instructors and local biologists are important capacity builders for acquiring these key life history parameters. Modern analytical tools involving basic fishery assessment and management programs are infrequently employed in many areas and more advanced analyses built upon knowledge of the former are largely ignored. The more complex the analyses, the fewer number of individuals that have appropriate background training. Thus bringing people up to these levels requires a graded approach of insuring comfort at previous levels prior to initiating the learning curve for the next level. 2282

267 A discussion of context is always important. A lecture summarizing the activities of major players in regional and international management and conservation, such as ICCAT, CITES, GFCM, ICES, FAO, IUCN, provides a review of current conditions of Atlantic and Mediterranean stocks, what actions are in effect and planned, and an update on the quality of extant regional biodiversity. Major stumbling blocks can be identified, potentially leading to regional efforts that can be aimed at rectifying such targets. Periodic production of workshops focusing on the matter at hand will significantly increase the quality and quantity of data. Equally importantly, these activities will result in the training of one or more instructors who can carry back the knowledge and disseminate it in similar workshops in their home countries ( teach the teachers ), a strategy that keeps giving. Parenthetically, having the opportunity to give back is as satisfying to the instructors as it is to the students. Funding requirements for a multi-day workshop include travel, housing and food for all participants, including educators, and minimal amounts for in-country group travel (visits to fishery landing beaches, markets, etc.), acquisition of specimens for ID lab, and classroom/lab rental (latter perhaps can be used for free). Long-distance education and outreach are mechanisms employed to reach the non-scientist focus group, but it also can attract scientists unable to attend one of the workshops. Posting of Workshop activities and supplementary documents on an established internet site allows for a larger scientific audience than a workshop can physically and fiscally handle - however the hands-on learning approach is always the best way to go. The non-scientist group, including fishers, particularly will benefit from the development and posting of signs encouraging the safe return to water of endangered species, such as sea turtles (Chelonia), sawfishes (Pristidae), and other CITES listed elasmobranchs, as well as locally and regionally prohibited elasmobranchs and bony fishes. Lack of enforcement is a fundamental problem in virtually all regions, so development of an education campaign leading to the development and posting of signs and delivery of developed educational brochures at ports, fishing beaches, fish markets, etc., will promote self-enforcement by fishers. One-day training workshops bringing together local fishers, fishery observers, and scientists could be organized to review current mitigation methods and best fishing practices aimed at reducing shark mortality. Implementation of good handling/release guidelines could enhance crew safety and optimize survival of released animals. As noted in the 2011 Meeting of the Working Group on the Organization of the SCRS, the number of CPCs acceding to the ICCAT agreement has increased rapidly in the last decade. Unfortunately, the level of participation of scientists from CPCs in the work of SCRS has not kept pace. Particularly given the acknowledged data limitations for many shark species, the SRDCP should continue to build on the efforts of ICCAT to promote increased participation of CPC scientists in the work of SCRS (i.e., data collection, contribution to stock assessment, calculation of local fishery indicators, participation at working groups, etc.). 8.2 Collaboration and coordination Collaboration and cooperation are essential actions that build the base of any transnational research activity. In the case of pelagic sharks species occurring in the Atlantic and Mediterranean any research plan and efficient data collection focused on these widely distributed species requires the enforcement of mechanisms to strengthen relations between the scientific teams involved in the process. The areas of collaboration that should be reinforced within this collective action that were identified by the Group include: elaboration of common protocols for the collection and analysis of biological samples protocols for the storage and preservation of biological samples capacity building and training in data collection and analysis equitable distribution of the biological sampling effort framed in a predefined scientifically sampling scheme promotion of visiting opportunities and interchanges for scientists at national laboratories prioritize multilateral collaboration for specific projects to promote collaboration among scientific teams consistently involved in sharks research with in the SCRS. 2283

268 With regards to collaboration with other organisations, it is important for ICCAT to continue to interact with other RFMOs that conduct scientific studies and provide management for shark species of interest in this research plan (e.g., trfmos, GFCM, NAFO and ICES). The joint assessment of porbeagle with ICES in 2009 and the KOBE Joint Tuna RFMO By-catch Working Group provide good examples as to how this collaboration can be facilitated. On-going collaboration to improve the scientific advice necessary for management of these species is crucial. In terms of collaboration with other groups, a wide variety of seabirds, turtles, marine mammals and sharks (comprehensively including sharks, and batoids) are likely to be incidentally captured in various fisheries. These four taxa comprise top predators whose role in the ecosystem is believed to be of great importance. Several initiatives at a national and regional scale aimed at minimizing the effects of by-catch are being developed. Research associated with these efforts is the most relevant source of information about the affected species and has allowed for collection of valuable information on various aspects of their biology and behaviour, particularly as it relates to their interaction with fishing vessels. We have seen that some mitigation measures developed for some of these taxa could result in an increase in shark catches. In this context, it is important to identify and contact organizations and working groups that will address a multi taxon approach and analysis in order to optimize the results and benefits of research. 8.3 Funding The Group briefly discussed the potential sources of funding to support the SRDCP. It was agreed that at this stage of the definition of this ambitious research action it is not possible to estimate the required funds to accomplish the different elements identified in the program. The Group considered that the best approach to conduct an appropriate estimation of the required budget is through a group of SCRS scientists familiarized with elasmobranch fisheries that would be responsible to accomplish this task. Funding support for a short contract would be required for this purpose. Implementation of the SRDCP will be framed within the SCRS strategic plan which will provide the overall framework for the required coordination and for the development of the plan. In any case, in the interest of supporting its on-going activities, the Group concluded that there is an urgent need for combining efforts to build a joint coordinated biological sampling scheme for the whole Atlantic and Mediterranean. This aspect was considered critical to gain efficacy and synergies in the context of the multiple national observer programs currently in place. The definition of biological sampling protocols, time-area-size-sex strata for the different SHK species, and equitative distribution of sampling effort among different teams are aspects that need to be defined in the immediate future. Consequently, the Group recommends that a small group of SCRS scientists should be in charge of elaborating the biological sampling design; the Group also recommends that this task be conducted in 2014 and the corresponding costs funded by ICCAT. The expected budget of this action should be evaluated and proposed to SCRS for its approval. References Abaunza, P., Murta, A.G., Campbell, N., Cimmaruta, R., Comesana, A.S., Dahle, G.,Garcia-Santamaria, M.T., et al Stock identity of horse mackerel (Trachurus trachurus) in the northeast Atlantic and Mediterranean Sea: integrating the results from different stock identification approaches. Fisheries Research, 89: Bigelow, K.A., Hampton, J., Miyabe, N Application of a habitat-based model to estimate effective longline fishing effort and relative abundance of Pacific bigeye tuna (Thunnus obesus). Fisheries Oceanography, 11: Block B.A., and Jonsen I.D Tracking apex marine predator movements in a dynamic ocean. Nature. 475: Cailliet, G.M., and K.J. Goldman Age determination and validation in chondrichthyan fishes. In: Biology of sharks and their relatives. J.C. Carrier, J.A. Musick, and M.R. Heithaus, (eds.), pp CRC Press, Boca Raton, FL. Campana, S.E Accuracy, precision and quality control in age determination, including a review of the use and abuse of age validation methods. Journal of Fish Biology 59:

269 Campana, S.E., L.J. Natanson, S. Myklevoll. 2002a. Bomb dating and age determination of large pelagic sharks. Canadian Journal of Fisheries and Aquatic Sciences 59: Campana, S.E., W. Joyce, L. Marks, L.J. Natanson, N.E. Kohler, C.F. Jensen, J.J. Mello, H.L. Pratt Jr., and S. Myklevoll. 2002b. Population dynamics of the porbeagle shark in the northwest Atlantic. North American Journal of Fisheries Management 22: Campana S.E., Dorey A., Fowler M., Joyce W., Wang Z., Wright D., and Yashayaev I Migration pathways, behavioural thermoregulation and overwintering grounds of blue sharks in the Northwest Atlantic. PLoS One 6: e Camhi, M.D., Pikitch, E.K., Babcock, E.A (Eds) Sharks of the Open Ocean: Biology, Fisheries and Conservation. Blackwell Publishing. 502 pp. Campana, S.E Use of radiocarbon from nuclear fallout as a dated marker in the otoliths of haddock, Melanogrammus aeglefinus. Marine Ecology Progress Series, 150: Carlson, J.K., Parsons, G.R Age and growth of the bonnethead shark, Sphyrna tiburo, from northwest Florida, with comments on clinal variation. Environmental Biology of Fishes, 50: Carlson, J.K, K.J. Goldman, and C. Lowe Metabolism, energetic demand, and endothermy. In: Biology of Sharks and Their Relatives. J. Carrier, J. Musick, M. Heithaus (eds.), pp CRC Press, Boca Raton, FL. Carlson, J.K., and I.E. Baremore Growth dynamics of the spinner shark, Carcharhinus brevipinna, off the United States Southeast and Gulf of Mexico coasts: a comparison of methods. Fishery Bulletin 103: Chen SB, Watanabe S Age dependence of natural mortality coefficient in fish population dynamics. Nip Suisan Gak 55: Coelho, R., Erzini, K Population parameters of the smooth lanternshark, Etmopterus pusillus, in southern Portugal (NE Atlantic). Fisheries Research, 86: Coelho, R., Erzini, K Life history of a wide-ranging deepwater lantern shark in the north-east Atlantic, Etmopterus spinax (Chondrichthyes: Etmopteridae), with implications for conservation. Journal of Fish Biology, 73: Coelho, R., Rey, J., Gil-De-Sola, L., Fernandez-Carvalho, J., Erzini, K Comparing Atlantic and Mediterranean populations of the velvet belly lanternshark, Etmopterus spinax, with comments on the efficiency of density-dependent compensatory mechanisms. Marine Biology Research, 6: Cortés E (1998) Demographic analysis as an aid in shark stock assessment and management. Fish Res 39: Cortés E. 1999a. A stochastic stage-based population model of the sandbar shark in the western North Atlantic. In: Musick JA (ed) Life in the slow lane: ecology and conservation of long-lived marine animals. Symposium 23, American Fisheries Society, Bethesda, Maryland, pp Cortés, E. 1999b. Standardized diet compositions and trophic levels of sharks. ICES Journal of Marine Science 56: Druffel, E.M. and T.W. Linick Radiocarbon in annual coral rings of Florida. Geophysical Research Letters, 94: Garcia A., Mattiucci S., Damiano S., Santos M.N., Nascetti G Metazoan parasites of Xiphias gladius (Linnaeus, 1758) (Pisces: Xiphiidae) from Atlantic Ocean: implications for host stock identification. ICES Journal of Marine Science, 68: Heist, E.J A review of population genetics in sharks. American Fisheries Society Symposium, 23: Heist, E.J Genetics: Stock indetification. In: Elasmobranch Fisheries Management Techniques, Misick, J.A., Bonfil, R (Eds.). APEC Secretariat pp. Heist, E.J Molecular markers and genetic population structure of pelagic sharks. In: Sharks of the Open Ocean: Biology, Fisheries and Conservation, Camhi, M.D. 2285

270 Pikitch, E.K., Babcock, E.A (Eds). Blackwell Publishing pp. Hoenig JM Empirical use of longevity data to estimate mortality rates. Fish Bull 82: Jensen AL (1996) Beverton and Holt life history invariants result from optimal trade-off of reproduction and survival. Can J Fish Aquat Sci 53: Kalish, J.M Pre- and post-bomb radiocarbon in fish otoliths. Earth Planet Science Letters, 114: Kerr, L., T. Brown, A. Andrews, G. Cailliet, and K. Coale Investigations of radiocarbon in the white shark (Carcharodon carcharias) from the eastern Pacific. Published abstract. Third International Symposium on Fish Otolith Research and Application. Townsville, Queensland, Australia. Kitchell, J.F., T.E. Esssington, C. Boggs, D.E. Schindler, and C.J. Walters The role of sharks and longline fisheries in a pelagic ecosystem of the central Pacific. Ecosystems 5: Lester, R.J.G. and MacKenzie, K The use and abuse of parasites as stock markers for fish. Fisheries Research, 97: 1-2. Lorenzen K The relationship between body weight and natural mortality in juvenile and adult fish: a comparison of natural ecosystems and aquaculture. J Fish Biol 49: Lorenzen K Allometry of natural mortality as a basis for assessing optimal release size in fish-stocking programmes. Can J Fish Aquat Sci 57: Matthews, K.R An experimental study of the habitat preferences and movement patterns of copper, quillback, and brown rockfishes (Sebustes spp.). Env. Biol. Fish. 29: Neer, J.A., K.A. Rose, and B.A. Thompson Bioenergetics modeling of the cownose ray, Rhinoptera bonasus, in the northern Gulf of Mexico. Abstract: American Society of Ichthyologists and Herpetologists/American Elasmobranch Society Annual Meeting. Norman, OK. Neer, J.A., B.A. Thompson, and J.K. Carlson Age and growth of Carcharhinus leucas in the northern Gulf of Mexico: incorporating variability in size at birth. Journal of Fish Biology 66:1-14. Okey, T.A., S. Banks, A.F. Born, R.H. Bustamante, M. Calvopiña, G.J. Edgar, E. Espinoza, J.M. Fariña, L.E. Garske, G.K. Reck, S. Salazar, S. Shepherd, V. Toral-Granda, and P. Wallem A trophic model of a Galápagos subtidal rocky reef for evaluating fisheries and conservation strategies. Ecological Modeling 172: Pauly D On the interrelationship between natural mortality, growth parameters, and mean environmental temperature in 175 fish stocks. J Cons Int Explor Mer 39: Pauly, D., V. Christensen, J. Dalsgaard, R. Froese, and F. Torres Fishing down marine food webs. Science, 279: Passerotti, M.S., Carlson, J.K., Piercy, A.N., Campana, S.E Age validation of great hammerhead shark (Sphyrna mokarran), determined by bomb radiocarbon analysis. Fishery Bulletin, 108: Peterson I, Wroblewski JS (1984) Mortality rates of fishes in the pelagic ecosystem. Can J Fish Aquat Sci 41: Queiroz N., Humphries N.E., Noble L.R., Santos A.M., and Sims D.W Spatial dynamics and expanded vertical niche of blue sharks in oceanographic fronts reveal habitat targets for conservation. PLoS One 7: e Roff D.A The evolution of life histories: theory and analysis. Chapman and Hall, New York Shivji, M., Clarke, S., Pank, M., Natanson, L., Kohler, N., Stanhope, M Genetic identification of pelagic shark body parts for conservation and trade monitoring. Conservation Biology, 16: Siegfried K.I Fishery management in data-limited situations: applications to stock assessment, marine reserve design and fish by-catch policy. PhD Dissertation (University of California, Santa Cruz) Simpfendorfer C.A. 1999a. Mortality estimates and demographic analysis for the Australian sharpnose shark, Rhizoprionodon taylori, from northern Australia. Fish Bull 97:

271 Simpfendorfer C.A Demographic models, life tables, matrix models and rebound potential. In: Musick JA, Bonfil R (eds.) Elasmobranch fisheries management techniques. Asia-Pacific Economic Cooperation Secretariat, pp Utter, F.M Biochemical genetics and fishery management: an historical perspective. Journal of Fish Biology, 39: von Bertalanffy, L A quantitative theory of organic growth (inquiries on growth laws II.) Human Biology, 10: Walker PA, Hislop J.R.G Sensitive skates or resilient rays? Spatial and temporal shifts in ray species composition in the central and north-western North Sea between 1930 and the present day. ICES J Mar Sci 55: Ward P., Empirical estimates of historical variations in the catchability and fishing power of pelagic longline fishing gear. Rev Fish Biol Fisheries 18: Weidman, C.R. and G.A. Jones A shell-derived time history of bomb 14C on Georges Bank and its Labrador Sea implications. Journal of Geophysical Research, 98: Wetherbee, B.M. and E. Cortés Food consumption and feeding habits. In: Biology of sharks and their relatives. J.C. Carrier, J.A. Musick, and H.R. Heithaus (eds.), pp CRC Press, Boca Raton, FL. Yamaguchi, A., Taniuchi, T., Shimizu. M Geographic variations in reproductive parameters of the starspotted dogfish, Mustelus manazo, from five localities in Japan and in Taiwan. Environmental Biology of Fishes, 57:

272 Appendix 7-Table 1. Traffic light approach used to categorize the level of information (expressed as number of studies) available by topic in four geographical areas for 16 species of Atlantic sharks. Red = no studies available; yellow = 1 or 2 studies; green = 3+ studies; white = species does not occur in the area. Area NORTH ATLANTIC SOUTH ATLANTIC EQUATORIAL ATLANTIC MEDITERRANEAN Species Reproduction Age and growth Stock ID Movement and migration Reproduction Age and growth Stock ID Movement and migration Reproduction Age and growth Stock ID Movement and migration Reproduction Age and growth Stock ID Movement and migration BSH SMA LMA POR SPZ SPK SPL ALV BTH FAL OCS DUS CCP CCS TIG PLS 2288

273 Appendix 7-Table 2. Classification of species according to data poorness (proportion of red cells in Table 1, i.e. with no information) and data richness (proportion of green cells in Table 1, i.e., with 3+ studies). For data poorness, species are listed from worst to best; for data richness, species are listed from best to worst. Values indicate the number of (red or green) cells as a proportion of the total number of cells for each species. ranked red LMA 0.88 SPK 0.81 DUS 0.81 TIG 0.81 PLS 0.81 POR 0.75 SPZ 0.75 ALV 0.69 OCS 0.69 CCP 0.69 CCS 0.67 FAL 0.58 BTH 0.56 SMA 0.50 SPL 0.50 BSH 0.19 ranked green BSH 0.31 SMA 0.25 CCP 0.25 SPL 0.19 OCS 0.19 FAL 0.17 POR 0.13 BTH 0.13 DUS 0.13 ALV 0.06 TIG 0.06 PLS 0.06 LMA 0.00 SPZ 0.00 SPK 0.00 CCS

274 Appendix 7-Table 3. Recommendations and Resolutions adopted by ICCAT that relate specifically to sharks. Number Name (EN) Status Recommendation by ICCAT on compliance with existing measures on shark conservation and management Active May Recommendation by ICCAT on information collection and harmonization of data on by-catch and discards in ICCAT fisheries Active Recommendation by ICCAT on the conservation of silky sharks caught in association with ICCAT fisheries Active Recommendation by ICCAT on hammerhead sharks (family Sphyrnidae) caught in association with fisheries managed by ICCAT Active Recommendation by ICCAT on the conservation of oceanic whitetip shark caught in association with fisheries in the ICCAT convention area Active Recommendation by ICCAT on Atlantic shortfin mako sharks caught in association with ICCAT fisheries Active Recommendation by ICCAT on the conservation of thresher sharks caught in association with fisheries in the ICCAT convention area Active Resolution by ICCAT on porbeagle shark (Lamna nasus) Inactive Recommendation by ICCAT on the conservation of bigeye thresher sharks (Alopias superciliosus) caught in association with fisheries managed by ICCAT Inactive Supplemental Recommendation by ICCAT concerning sharks Active Supplementary Recommendation by ICCAT concerning the conservation of sharks caught in association with fisheries managed by ICCAT Active Recommendation by ICCAT to amend Recommendation concerning the conservation of sharks caught in association with fisheries managed by ICCAT Active Recommendation by ICCAT concerning the conservation of sharks caught in association with fisheries managed by ICCAT Active Resolution by ICCAT on the shark fishery Active Resolution by ICCAT on Atlantic sharks Inactive Resolution by ICCAT on cooperation with the Food & Agriculture Organization of the United Nations (FAO) with regard to study on the status of stocks and by-catches of shark species Active 2290

275 Appendix 7-Table 4. International measures that apply to elasmobranchs within the proposed ICCAT Shark Research and Data Collection Programme. Convention Measure Species Barcelona Convention Annex II Carcharodon carcharias Isurus oxyrinchus Lamna nasus Mobula mobular Sphyrna lewini Sphyrna mokarran Sphyrna zygaena Annex III Alopias vulpinus Carcharhinus plumbeus Prionace glauca CITES Appendix II Carcharodon carcharias Carcharhinus longimanus a Lamna nasus a Manta alfredi a Manta birostris a Sphyrna lewini a Sphyrna mokarran a Sphyrna zygaena a CMS Appendix I Carcharodon carcharias Manta birostris Memorandum of Understanding Carcharodon carcharias Isurus oxyrinchus Isurus paucus Lamna nasus GFCM Rec. GFCM/36/2012/3 Barcelona Convention Annex II and Annex III species (above) a Listing enters into effect September

276 Appendix 7-Table 5. Biological and fishery data requirements and output provided by a suite of data-poor methods that could potentially be used to assess the status of Atlantic sharks and generate management advice and research recommendations. METHOD DATA REQUIREMENTS REFERENCE POINTS MANAGEMENT ADVICE RESEARCH RECOMMENDATIONS Biology Fishery PSA level I, II qualitative qualitative No Qualitative Yes Length-based methods (SEINE) VBGF parameters mean recruitment length, time series of lengths Changes in Z Qualitative Yes PSA level III;Demographic models; Elasticity analysis age & growth, reproduction, M several (PSA only) No Mostly qualitative (e.g., size limits), but also F Yes Analytical benchmarks DCAC age & growth, reproduction, M M Index of relative abundance B/Bmsy Quantitative Yes catch, index of relative abundance Sustainable catch Quantitative Yes AIM catch, index of relative abundance F/Fmsy Quantitative (sustainable F) Yes Surplus production (ASPIC, BSP, others) r catch, index of relative abundance B/Bmsy and F/Fmsy Quantitative, projections Yes 2292

277 Appendix 7-Table 6. ICCAT recommendations regarding the observer programs Recommendation Objective % Coverage Rec Recommendation by ICCAT on Information Collection and Harmonization of Data on By-catch and Discards in ICCAT Fisheries Rec On a Multiannual conservation and management program for bigeye and yellowfin tunas Rec Recommendation amending the Rec. to establish a multiannual recovery plan for Bluefin tuna Voluntary PS implementation Shark data collection Current coverage By-catch and discard data Not defined By-catch/discards n/a The ICCAT Regional Observer Program shall be established in 2013 to ensure observer coverage of 100% of all surface fishing vessels 20 meters LOA or greater fishing bigeye and/or yellowfin tunas in the area/time closure. Bluefin catch compliance Tuna catch and by-catch data 100 % of PS 100 % PS, 100 % tranfers from PS, 100 % tranfers from traps to cages, 100 % farms, traps and towing vessels, 20 % active BB, LL and pelagic trawlers. 100 % from 2013 By-catch/discards By-catch/discards By-catch/discards n/a 100% Not yet evaluated 2293

278 Appendix 7-Figure 1. Map showing the geographical areas considered in the evaluation of the current state of biological knowledge as summarized in the appendix tables. 2294

279 Nº trabajos intersesiones Appendix 7-Figure 2. Evolution of number of documents presented at shark inter-sessional working group meetings. 2295

280 RÉUNION INTERSESSION DE 2013 DU GROUPE D ESPÈCES SUR LES REQUINS (Mindelo, Cap-Vert, 8 12 avril 2013) 1. Ouverture, adoption de l ordre du jour et organisation des sessions M. Óscar David Fonseca Melício, président de l Institut national pour le développement des pêches du Cap-Vert a souhaité la bienvenue aux participants à Mindelo. Le Président du SCRS, le Dr Josu Santiago, lui a adressé tous ses remerciements pour accueillir la réunion au sein de l Institut. Le Dr Paul de Bruyn, au nom du Secrétaire exécutif de l ICCAT, a alors ouvert la réunion. La réunion a été présidée par le Dr Andrés Domingo, rapporteur du Groupe d espèces sur les requins. Le Dr Domingo a souhaité la bienvenue aux participants du Groupe et a passé en revue les termes de référence de la réunion. Après l ouverture de la réunion, l ordre du jour a été examiné et adopté avec de légers changements (Appendice 1). La liste des participants est incluse à l Appendice 2. La liste des documents présentés à la réunion est jointe en tant qu Appendice 3. Les participants ci-après ont assumé la tâche de rapporteurs pour les divers points du rapport : Point Rapporteurs 9 P. de Bruyn 10 P. de Bruyn, A. Perry, A. Domingo 11 P. de Bruyn 12 E. Cortés, R. Coelho, G. Burgess, B. Seret 13 Participants du Groupe 14 J. Santiago, A. Domingo 15 J. Santiago, A. Domingo 8 P. de Bruyn 2. Présentation des documents Dans le document SCRS/2013/044, il a été identifié que jusqu à présent, les changements de l espèce ciblée ont été inclus dans les évaluations des stocks à deux niveaux différents dans l analyse. Tout d abord, ces changements sont pris en considération lors de la paramétrisation des modèles linéaires généralisés utilisés pour calculer la standardisation de l indice de CPUE. Ensuite, les capturabilités qui varient constamment dans le temps sont directement incluses lors de l ajustement du modèle dynamique utilisé aux fins de l évaluation. Cette dernière étape modèle les capturabilités annuelles comme des calculs aléatoires d après une répartition stationnaire des capturabilités. Des éléments de preuves empiriques suggèrent toutefois que les modèles incluant des changements importants des capturabilités survenant une seule fois pourraient très bien décrire les changements temporels de diverses pêcheries. Ce document présentait un ensemble de modèles bayésiens de production état-espace ajustés à la série temporelle du stock de requin peau bleue de l Atlantique sud (Prionace glauca), dans lesquels un seul point de changement de la répartition stationnaire des capturabilités était spécifié, avec deux paramètres de capturabilités estimés, un avant et un après le point de changement. Bien que les modèles n incluent qu un seul paramètre additionnel, ils donnaient lieu à un meilleur ajustement par rapport à l approche de modélisation de capturabilité à un seul paramètre. Les modèles donnaient lieu à des estimations différentes des points de référence et des quotas de capture. Ils indiquaient tous, cependant, que le stock de requin peau bleue se situe au-dessus de BPME et que les niveaux de mortalité par pêche sont toujours en-deçà de FPME. Même si le fait de ne tenir compte que d un seul point de changement de capturabilité n avait pas d impact significatif sur l état de cette population particulière de poisson, il constitue un moyen robuste de prendre en considération les changements de capturabilité résultant du changement de la dynamique des pêcheries et peut être mis en œuvre aux fins de la modélisation d autres stocks de poissons. Le Groupe a discuté du postulat selon lequel les changements survenant dans les prises dans la période à l étude étaient dus aux changements de sélectivité ou de capturabilité. Ces changements pourraient également procéder de la demande du marché. Des informations sur les débarquements non rejetés sont disponibles dans les livres de bord. Il a été signalé que, d après les livres de bord remis par les pêcheurs, des changements de ciblage sont survenus dans le temps mais que ceux-ci sont difficiles à quantifier. 2296

281 Le document SCRS/2013/045 notait que les requins pélagiques font face à des décisions de déplacements complexes tout en résidant dans un environnement relativement neutre et oligotrophique. Il existe aussi une prise accessoire habituelle dans les pêcheries pélagiques, suscitant des préoccupations quant à une surpêche. Le développement de programmes de gestion et d évaluations efficaces des stocks implique de comprendre comment ces poissons utilisent les environnements de tout l océan étant donné que les déplacements transocéaniques sont habituels. Des modèles mixtes aléatoires et de télémétrie par satellite ont été utilisés pour quantifier les facteurs qui provoquent des schémas migratoires chez le requin peau bleue, Prionace glauca, dans l Atlantique sud. La majorité des requins présentaient une résidence dans des zones centrales, même si certains spécimens effectuaient de longs déplacements, dont deux événements dispersés transatlantiques. La sélection de l habitat s expliquait surtout par la température de la mer en surface (SST) et la profondeur de la couche mixte (DML), mais ceci variait selon les régions. Dans les zones supposées être des zones de gestation, les requins adultes femelles choisissaient des eaux plus chaudes et moins profondes que les mâles. La population de requin peau bleue de l Atlantique sud devrait être considérée comme un seul stock, même s il est peu probable qu elle effectue un cycle migratoire dans le sens des aiguilles d une montre dans l océan Atlantique. Le Groupe a discuté de la sensibilité du modèle aux postulats de déplacements étant donné que d autres modèles spatialement explicites se basent sur un grand nombre de marques conventionnelles, alors que ce modèle utilise un nombre très restreint de marques satellite. Il a été expliqué que même s il existe des différences dans la localisation précise des spécimens porteurs de marques, ils se seraient toujours trouvés dans la même «zone», tel que défini par le modèle. Le document SCRS/2013/037 présentait des informations sur les palangriers portugais ciblant l espadon dans l Atlantique, qui capturent régulièrement plusieurs espèces d élasmobranches en tant que prise accessoire, notamment des espèces actuellement protégées, telles que le renard à gros yeux et le requin-marteau commun. Ce document présente les résultats préliminaires sur les renards à gros yeux et requins-marteau communs porteurs de marques pop-up archives par satellite ayant transmis des donnés en 2012 dans la région tropicale nord-est de l Atlantique. Des schémas migratoires journaliers verticaux très nets ont été observés pour le renard à gros yeux, les profondeurs les plus occupées étant m durant le jour et m durant la nuit, ce qui correspond à des températures de l eau de 8-10 ºC et ºC, respectivement. Pour le requin-marteau commun, aucune différence majeure n a été détectée entre le jour et la nuit, passant la plupart du temps dans la gamme de profondeur m. Même si les données présentées dans ce document sont toujours limitées et font partie de projets en cours, les résultats préliminaires sont utiles pour améliorer les connaissances sur la biologie, l écologie et les schémas d utilisation de l habitat de ces espèces et peuvent servir de valeurs d entrée pour les analyses actuelles et futures analyses d évaluations des risques écologiques. Le Groupe a constaté que l étude de marquage était en mesure de collecter des informations sur des profondeurs supérieures à celles normalement exploitées par la pêcherie palangrière et fournit donc des données qui ne seraient normalement pas obtenues de la pêcherie. Ces informations pourraient s avérer très importantes pour l élaboration d un programme de recherche sur les requins devant être développé pendant la réunion. Il a aussi été fait observer que les recherches sont toujours en cours concernant le requin océanique. Le document SCRS/2013/038 portait sur le renard à gros yeux, Alopias superciliosus, qui est habituellement capturé en tant que prise accessoire par les pêcheries palangrières pélagiques ciblant l espadon. Dans le cadre d un programme en cours visant à la collecte de données biologiques et sur les pêcheries, des observateurs des pêcheries ont été embarqués à bord de navires de pêche pour collecter un ensemble de données incluant la taille, le sexe et la phase de maturité afin de chercher à déterminer la maturité du renard à gros yeux. Au total, renards à gros yeux ont été enregistrés dans l Atlantique. La taille de ces spécimens oscillait entre 94 et 264 cm FL (longueur à la fourche). Dans les régions du nord, il y avait une plus forte proportion de femelles (> 63%) et la classe de taille modale était inférieure à celle des régions du sud, dans lesquelles les plus grands spécimens ont été observés. Des ogives de maturité ont été ajustées pour 642 spécimens pour lesquels des données de maturité étaient disponibles. La taille à première maturité a été estimée à 208,6 cm FL pour les femelles (correspondant à ans) et à 159,7 cm FL pour les mâles (correspondant à 5-6 ans). Le Groupe a noté qu il y avait des stratégies potentiellement différentes entre les espèces de renards de mer. Le document SCRS/2013/042 faisait observer que pour l amélioration de futures évaluations du stock de requintaupe bleu (Isurus oxyrinchus) dans l Atlantique, il était important d examiner les paramètres biologiques. Au cours de la dernière évaluation du stock, les incertitudes liées aux statistiques de capture, à la capturabilité et aux paramètres biologiques ont été discutées en ce qui concerne l ajustement médiocre de la tendance de la biomasse estimée à la tendance observée de la CPUE. Même en estimant qu il pourrait y avoir un volume patent de prises non déclarées, ceci n explique pas l augmentation de la CPUE constamment observée dans de nombreuses 2297

282 flottilles. Il est nécessaire de réévaluer le postulat existant, selon lequel le taux intrinsèque d augmentation naturelle (r) de cette espèce est assez faible, en regroupant les connaissances existantes sur les paramètres biologiques. Ce document fournit des informations sur l état actuel des études biologiques pour les populations du Pacifique nord, en se concentrant sur l analyse de la croissance, étant donné que des études ont dernièrement été menées dans cette zone et qu il est patent que le paramètre de croissance joue un rôle-clé dans la dynamique de la population parmi les divers paramètres biologiques. Des points importants à prendre en considération en vue du futur programme de recherche sont également discutés. Le document SCRS/2013/040 présentait la mortalité à bord du navire, le taux de survie après la remise à l eau et la mortalité totale des requins soyeux dans la pêcherie française de senneurs ciblant les thonidés tropicaux opérant dans l océan Indien. À l heure actuelle, les senneurs français de thonidés tropicaux opérant dans l océan Indien remettent à l eau tous les requins et toutes les raies qui sont capturés accidentellement. En participant à deux sorties de pêche commerciale et à une campagne de recherche affrétée, nous avons tout d abord enregistré le nombre de requins (essentiellement des requins soyeux, Carcharhinus falciformis) qui étaient vivants ou morts, une fois triés par l équipage sur les ponts supérieur et inférieur. Un plus grand nombre de requins a été constaté sur le pont inférieur (73%) que sur le pont supérieur. Les requins soyeux observés sur le pont supérieur étaient bien plus grands que ceux présents sur le pont inférieur. Le taux de mortalité immédiate (requins qui étaient morts au moment de l observation) semblait être lié à l emplacement des spécimens, étant donné qu un plus grand nombre de requins étaient trouvés morts sur le pont inférieur que sur le pont supérieur. Le taux de mortalité à bord du navire augmentait aussi avec la taille de l opération (jauge). Vingt requins soyeux ont été marqués à l aide de marques MiniPATs (Wildlife Computers, Redmond, WA, USA) afin d étudier leur survie après la remise à l eau. De surcroît, douze requins soyeux ont été marqués avec le même type de marques électroniques au cours d une campagne de pêche scientifique. Sur un sous-échantillon de 32 requins soyeux évalués en vie après recapture et suivis pendant une période allant de 100 à 150 jours après remise à l eau, 8 marques montraient clairement une mortalité directe après remise à l eau, alors que les données de quatre marques donnaient à penser à une mortalité retardée après 2 à 35 jours et un autre, en mauvais état, est mort dévoré après trois jours. Au total, 16 marques montraient que les requins survivaient. Deux marques n ont pas communiqué de données et une marque n a pas été correctement initialisée. Ce document fournit les premières estimations, pour les requins soyeux (longueur >85 cm TL), de la mortalité à bord du navire et de la mortalité après remise à l eau qui s élevaient, respectivement, à environ 67% et 58%. Le taux de mortalité global des requins soyeux capturés accidentellement de cette flottille a été estimé à environ 81%. Un manuel de «meilleures pratiques» a été élaboré à l attention des pêcheurs afin d accroître le taux de survie des requins capturés par les senneurs. Il convient toutefois de chercher à déterminer d autres méthodes avant que les requins ne soient hissés à bord. Le Groupe a sollicité des clarifications supplémentaires quant à savoir comment les requins avaient été sélectionnés pour cette étude. Il a été expliqué que chaque requin avait été évalué selon l échelle suivante : 1) Bonne comportement très actif, morsure, coups. 2) Moyenne Peu de mouvement mais signes de vie clairs. 3) Médiocre faible réponse aux stimuli externes. 4) Mort. Par la suite, 32 requins qui montraient des signes de vie (échelle 1 et 2) ont été sélectionnés aléatoirement. Le niveau élevé de mortalité à bord du navire a été discuté et il a été noté que les protocoles de remise à l eau actuellement en vigueur n avaient pas encore été adoptés au moment de cette étude. Le document SCRS/2013/039 présentait la prise par taille et les ratios des sexes des élasmobranches dans la pêcherie de palangre pélagique portugaise dans l Atlantique. Cette analyse se basait sur les données collectées par les observateurs des pêcheries, l échantillonnage au port et les livres de bords des capitaines (autoéchantillonnage), compilées entre 1997 et Les données ont été analysées en termes de prise accessoire par taille et comparées entre les années, saisons (trimestres), stocks (Nord et Sud, séparés à 5º N) et zones de pêche principales des opérations de la flottille portugaise (Nord, Tropicale Nord, Equatoriale et Sud). Pour le requin peau bleue, une tendance générale à la hausse pour les tailles moyennes a été observée pour les deux hémisphères, accompagnée d une diminution pour les années les plus récentes. Pour le requin-taupe bleu, la taille moyenne est restée stable dans le Nord et tendait à descendre au Sud. Une certaine variabilité a été constatée dans les comparaisons spatiales et saisonnières. Les proportions de ratios des sexes ont été comparées entre les régions et les saisons, et pour les principales espèces, d importantes différences ont été notées. Les données présentées dans ce document de travail sont encore préliminaires mais apportent de nouvelles informations importantes sur les tendances de la prise par taille et les ratios des sexes des principaux requins pélagiques capturés par la pêcherie palangrière pélagique portugaise dans l Atlantique. 2298

283 Une brève explication de l auto-échantillonnage réalisé par l UE-Portugal a été fournie. Ce programme se base sur une feuille de calcul MS Excel, qui permet aux capitaines de calculer le poids total de la capture d après des échantillons individuels. Il est utile aux capitaines à des fins d application et fournit des informations à des fins scientifiques. Il a été noté que les informations de VMS sont difficiles à obtenir en raison de questions de confidentialité. Des efforts sont actuellement déployés afin d obtenir ces données du Département de gestion des pêches sous une forme suffisamment regroupée pour permettre leur distribution. Le document SCRS/2013/046 faisait état des relations longueur-longueur entre la longueur à la fourche, la longueur précaudale et la longueur totale pour les six principales espèces pélagiques (Prionace glauca, Carcharhinus brachyurus, Carcharhinus signatus, Sphyrna zygaena, Isurus oxyrinchus et Lamna nasus) capturées par la flottille palangrière pélagique uruguayenne dans le sud-ouest de l Atlantique, entre 1998 et Les relations longueur-longueur soumises dans ce document couvrent une grande portion de toute la gamme de tailles déclarées de chaque espèce étudiée et sont les premières conversions longueur-longueur jamais déclarées pour ces espèces dans cette zone. Le document SCRS/2013/047 évaluait les prises de requins de la pêcherie artisanale de filets dérivants au large d Abidjan (Côte d Ivoire) pour la période , à l aide des données de poids et de taille collectées pour chaque espèce de requins sur trois sites de débarquement et de la proportion de pirogues échantillonnées. Au cours de cette période, le nombre de sorties de pêche journalières a diminué de moitié et les prises variaient entre 92 et 203 t. Cependant, la proportion de requins dans les prises totales variait de 2,1 % en 2008 à 31% en Les espèces les plus importantes étaient le requin peau bleue (Prionace glauca) et le requin-taupe bleu (Isurus oxyrinchus) dont les CPUE (kg/sortie journalière) étaient en légère augmentation. Les prises se composaient de juvéniles de cm LT pour le requin peau bleue et cm LT pour le requin-taupe bleu. L auteur a expliqué que la mesure de longueur soumise dans le document était la longueur précaudale. Il a été noté que le type d engin décrit dans l étude était le filet maillant et que cet engin était déployé à 2 miles de la côte, probablement proche des canyons ou de l escarpement du plateau continental. Ceci pourrait expliquer le grand nombre de requins signalés. Le document SCRS/2013/041 notait que la réduction de la mortalité des prises accessoires était actuellement un objectif de l approche écosystémique des pêcheries et une demande des consommateurs. L implication et la participation des utilisateurs de la ressource sont indispensables afin de développer des techniques d atténuation efficaces et pratiques. Les pêcheurs manipulent les animaux comme partie intégrante de leur travail et il est essentiel d identifier de bonnes pratiques garantissant la sécurité de l équipage et optimisant la survie des poissons remis à l eau. En combinant les observations scientifiques et les connaissances empiriques des pêcheurs de la flottille française de senneurs, des directives de manipulation et de remise à l eau sont proposées pour les requins et les raies, même pour les grands spécimens, tels que les requins baleines et les raies mantas accidentellement capturés par les pêcheries de senneurs ciblant les thonidés tropicaux. Un manuel de bonnes pratiques a été élaboré pour sensibiliser les pêcheurs à la préservation et conservation de la biodiversité et encourager leur participation à la gestion durable des ressources marines. La diffusion de ces meilleures pratiques sur les ponts des navires de pêches devrait contribuer à la réduction de la mortalité par pêche de certaines espèces vulnérables. Ceci serait positivement perçu par les consommateurs comme une mesure visant à réduire l empreinte de la pêche sur l environnement et à promouvoir le bien-être animal, ce qui améliorerait l image de l industrie halieutique. De nouvelles idées émergeant des échanges entre les scientifiques et les pêcheurs sont également proposées même si elles n ont pas encore été testées. La recherche sur l atténuation des prises accidentelles est par définition un processus itératif et différentes méthodes complémentaires doivent être menées à divers niveaux du processus de pêche pour réduire considérablement la mortalité de ces poissons. Le document SCRS/2013/049 affirmait que l absence de données fiables dépendant des pêcheries et de connaissances fondamentales sur la biologie de la plupart des espèces de requins suscite des préoccupations pour la gestion durable des populations de requins en Méditerranée. L étude vise à chercher à déterminer l occupation de l habitat, les temps de résidence et les routes migratoires et à fournir des données comportementales sur les températures affrontées et la profondeur de plongée des grands requins pélagiques, tels que le requin peau bleue (Prionace glauca). Cette étude s efforce également de déterminer le moment et le lieu où les requins sont les plus vulnérables et permettra de contribuer à la conservation de cette espèce. Il est proposé d utiliser des marques satellite pour étudier l écologie des grands requins pélagiques. Les résultats préliminaires de la première marque SPOT (Smart position or temperature transmitting) déployée sur un requin peau bleue femelle sont présentés. 2299

284 Le document SCRS/2013/048 notait que le TAC de zéro, instauré par la CE pour le requin-taupe commun, avait engendré la fermeture d une pêcherie saisonnière traditionnellement réalisée par une petite flottille de cinq palangriers de l île d Yeu (golfe de Gascogne). En vue d améliorer les connaissances sur le requin-taupe commun, le Ministère français des pêches a soutenu un programme scientifique visant à déterminer les déplacements de ce requin dans l Atlantique nord-est à l aide de marques pop-up par satellite (PSAT). Au cours de l été 2011, trois marques PSAT ont pu être apposées sur des requins-taupes communs femelles, adultes et sous-adultes, lors d une campagne de marquage menée dans le golfe de Gascogne avec un palangrier de l île d Yeu. Les trois marques se sont détachées, une à 8 mois les deux autres à 12 mois (soit la durée de déploiement originale). Même si les données transmises par les marques doivent être traitées une nouvelle fois avec divers filtres, l analyse préliminaire montre que les requins marqués présentent trois schémas migratoires différents dans l Atlantique nord-est. Une femelle mature de 2,34 m LT, marquée au large de la péninsule de Quiberon, est restée un mois à proximité puis s est déplacée vers la côte des Shetland où elle est restée environ 2,5 mois pour atteindre finalement la mer de Norvège en novembre ; elle s est déplacée ensuite en Islande pour revenir en Norvège en février où la marque s est détachée. Au cours de cette migration, ce requin a effectué des plongées régulières jusqu à 500 m de profondeur, atteignant un maximum de m de profondeur. Le second requin, une femelle sous-adulte de 1,9 m LT, a été marqué au large de l île de Noirmoutier. Ce requin a réalisé une grande trajectoire triangulaire dans l Atlantique se déplaçant nord-ouest, s approchant du Groenland en novembre puis allant directement au sud des Açores en février-mars avant de revenir presque à la position de marquage originale, 12 mois plus tard. Ce requin a également effectué des plongées jusqu à m de profondeur environ. Le troisième requin, une femelle sous-adulte de 1,9 m, a été marqué au large de la péninsule de Penmarch, s est également déplacé nord-ouest, est retourné dans la mer du Nord en octobre-novembre avant de revenir dans le golfe de Gascogne (au large du sud de l Irlande) au mois de juin en suivant une trajectoire erratique, il a plongé à des profondeurs allant jusqu à 800 m de profondeur lorsqu il se trouvait au large du plateau continental. Même si ces observations sont limitées, elles indiquent que le requin-taupe commun utilise de vaste zones de l Atlantique nord-est ainsi que la colonne d eau jusqu à m de profondeur. Une brève présentation d un projet actuellement mené par des instituts de l Union européenne a été présentée au Groupe. L objectif général de ce projet vise à obtenir un avis scientifique afin de mettre en œuvre le programme d action communautaire pour la conservation et la gestion des requins, pour faciliter le suivi des pêcheries en haute mer et l évaluation du stock de requins à un niveau spécifique à l espèce. Cette étude se concentre sur 18 élasmobranches principaux à un niveau mondial. En vue d atteindre les objectifs du projet, l équipe s est penchée sur : la collecte et l examen des données historiques des pêches, notamment en termes de composition par espèce, de prises et d effort ; l estimation des prises totales de requins ; l identification de lacunes dans les connaissances actuelles des pêcheries mais aussi la biologie et l écologie des requins. Afin de combler les lacunes et de soutenir l avis émanant des ORGP sur une gestion soutenable des pêcheries d élasmobranches, plusieurs propositions sont en cours d élaboration, notamment en termes de désignation de programmes d observateurs, d identification de priorités de recherche scientifique et d intégration de l information sur les ORGP thonières. Le Groupe a salué cette initiative et a demandé aux auteurs de soumettre les conclusions du projet dès que l information sera disponible. 3. Présentation des données de marquage et des données de Tâche I et de Tâche II Le Secrétariat a présenté un résumé de l information sur les requins soumise par les CPC. Les échantillons de taille et de prise-effort de Tâche I et Tâche II ont été présentés sous forme de catalogues de données afin d identifier des lacunes dans les données disponibles. Il a été fait observer que bien que les données de Tâche I soient disponibles pour de nombreuses espèces de requins, celles-ci sont extrêmement incomplètes et dans de nombreux cas la Tâche I n est pas accompagnée des données correspondantes de Tâche II. Ceci est particulièrement le cas pour les espèces autres que BSH, SMA et POR, pour lesquelles davantage d informations sont généralement disponibles. Il a aussi été noté qu il existe plus d informations pour l Atlantique nord que pour l Atlantique sud, et que très peu de données sont disponibles pour la Méditerranée. Le Groupe a sollicité la soumission des données dans un format permettant d identifier facilement les lacunes pour pallier à ces déficiences dans le programme de recherche (Appendices 4-6). Le Secrétariat a également présenté les données de marquage disponibles pour BSH, SMA et POR. Les densités des marquages, des récupérations et des déplacements sont illustrées aux Figures Il a aussi été suggéré que le Groupe pourrait se fixer l objectif de développer un format de déclaration des données de marquage par satellite à l ICCAT. Il a été reconnu que le jeu de données pour chaque marque peut être assez extensif et qu il serait plus viable de déclarer des métadonnées pour les marques électroniques (par exemple, marquages et localisations des marques pop-up). 2300

285 4. État actuel des connaissances et de la recherche sur les requins pélagiques de l Atlantique et de la Méditerranée Cette information est traitée de façon exhaustive dans le programme de recherche détaillé au point 5 ci-après. 5. Programme de recherche scientifique pour les requins et compilation des données Une présentation a été soumise sur le programme de recherche stratégique du SCRS afin d inclure les discussions actuelles dans un contexte plus vaste dans le cadre des travaux du SCRS. Un programme stratégique est recommandé comme approche structurée pour orienter les futurs travaux du SCRS (rapport du SCRS de 2011 et réponse à la Rés sur la meilleure science disponible). Le document SCRS/2013/024 décrivait une approche pour identifier les composantes et les besoins-clés en matière de recherche ainsi qu une feuille de route aux fins de l élaboration du Programme stratégique du SCRS pour Le SCRS/2013/024 signalait que le Programme stratégique traitait de trois axes principaux : «Que faisons-nous?», «Pour qui le faisons-nous?» et «Comment excellons-nous?». Par ailleurs, les composantes-clés du programme stratégique incluent la compréhension de la mission du SCRS (notre objectif), notre vision pour l avenir, les valeurs que nous devrons appliquer dans la conduite de nos travaux, nos objectifs et les stratégies pour les atteindre. Il a été indiqué que le Programme stratégique fournit aussi une méthodologie visant à identifier la capacité scientifique et les lacunes en matière de données et à établir l ordre de priorité des activités de recherche pour les résoudre. Une feuille de route et des délais pour le développement du Programme stratégique du SCRS pour ont été proposés dans le document SCRS/2013/024. Ceci inclut le recrutement d un consultant pour fournir un cadre pour la méthodologie spécifique à appliquer lors du développement du Programme stratégique, une consultation et examen réguliers par les mandataires du SCRS et la plénière du SCRS aux fins d examen et acceptation par la Commission Objectifs du programme de recherche Une présentation a été réalisée sur le cadre général du Programme de recherche sur les requins, constituant un modèle aux fins de discussion et d élaboration. Le Président a alors demandé aux participants de soumettre des commentaires sur la structure du programme, un contenu potentiel et l identification de points auxquels ils souhaitaient participer. Ceci a été effectué et un modèle a été convenu par le Groupe. Les divers points ont ensuite été élaborés par les participants Développement du Programme Le Programme de collecte de données et de recherche sur les requins est présenté à l Appendice Autres questions Les scientifiques du Cap-Vert ont décrit les activités de pêche menées par la flottille nationale et les flottilles étrangères (Union européenne, Chine), opérant dans le cadre de différents accords de pêche, qui ont un impact sur les espèces de requins dans leur ZEE. La flottille du Cap-Vert ne cible pas les élasmobranches même si ceuxci font partie de la prise accessoire lors du ciblage d autres espèces. Il n existe pas de licence spécifique pour les requins au Cap-Vert pour aucune des flottilles. Les flottilles palangrières étrangères opérant dans la ZEE du Cap- Vert enregistrant de forts pourcentages de requins, lesquels représentent plus de 75% de leurs captures, composées essentiellement de Prionace glauca et Isurus oxyrinchus. Compte tenu de l importance des espèces de requins dans la zone du Cap-Vert, des scientifiques locaux ont lancé une initiative visant au développement d un Programme de collecte de données pour la flottille nationale, pour lequel une assistance technique est requise. Le Cap-Vert a réitéré son souhait d obtenir une assistance pour développer un programme de collecte de données, incluant des procédures d échantillonnage et un système de traitement des données sur les espèces de requins capturées par sa flottille. Le Groupe a salué l initiative du Cap-Vert pour le développement d un Programme de collecte de données pour sa flottille nationale avec une attention particulière accordée aux espèces de requins. Bien que les requins ne soient pas ciblés par la flottille locale, ils sont une composante importante de sa capture. Le Groupe a recommandé d allouer des fonds spéciaux de l ICCAT pour cette importante initiative. 2301

286 7. Recommandations - Le Groupe a recommandé d autoriser les observateurs scientifiques à collecter des échantillons biologiques (vertèbres, tissus, organes de reproduction, contenus stomacaux, échantillons de peau, valves spirales, mâchoires, spécimens entiers ou squelettes pour des travaux taxonomiques ou collections de musées) d espèces de requins actuellement interdites qui sont morts à la remontée de l engin, sous réserve que les échantillons fassent partie du projet de recherche approuvé par le SCRS. Afin d obtenir cette autorisation, un document détaillé décrivant l objectif de ces travaux, le nombre et le type d échantillons devant être collectés et la répartition spatio-temporelle du travail d échantillonnage doivent être inclus dans la proposition. La progression annuelle des travaux et un rapport final sur l avancée du projet devront être présentés au Groupe d espèces sur les requins ainsi qu au SCRS. - Le Cap-Vert a fait part de son souhait d obtenir une assistance pour développer un programme de collecte de données, incluant des procédures d échantillonnage et un système de traitement des données sur les espèces de requins capturées par sa flottille ou débarquées au Cap-Vert. Bien que les requins ne soient pas ciblés par la flottille locale, ils sont une composante importante de sa capture. Le Groupe a recommandé d allouer des fonds spéciaux de l ICCAT pour cette importante initiative. - Le Groupe a recommandé qu en 2014 un petit groupe de scientifiques du SCRS soit chargé d élaborer un plan d échantillonnage biologique pour les espèces de requins pélagiques dans l Atlantique et en Méditerranée. Le budget prévu pour cette mesure devrait être évalué et proposé au SCRS aux fins de son approbation. 8. Adoption du rapport et clôture Le Groupe a adressé ses remerciements à l INDP et à ses scientifiques pour l organisation et le bon déroulement de la réunion. L hospitalité fournie a été extraordinaire et le Groupe a fortement apprécié l attention incroyable accordée aux participants par les scientifiques du Cap-Vert. PROGRAMME DE RECHERCHE ET DE COLLECTE DE DONNÉES SUR LES REQUINS (Appendice 7) A. INTRODUCTION La zone de la Convention de l ICCAT compte une grande variété de requins, aussi bien des espèces côtières que des espèces océaniques. 91 espèces de requins figurent actuellement dans les bases de données de l ICCAT. Les stratégies biologiques de ces espèces sont très diverses ; celles-ci sont fortement adaptées à leurs écosystèmes respectifs et occupent une position très élevée dans la chaîne trophique en tant que prédateurs actifs. Même si les caractéristiques biologiques de ces espèces sont diverses, elles partagent un schéma général qui les rend davantage susceptibles à la surpêche. Même si les pêcheries commerciales et sportives ont actuellement un impact sur les élasmobranches, il n existe que des informations limitées sur le cycle vital, les paramètres biologiques, les schémas migratoires et l utilisation de l habitat de ces espèces ainsi que sur l impact général des pêcheries sur leurs populations dans la zone de la Convention de l ICCAT. De plus, l état actuel des connaissances sur les pêcheries de l ICCAT qui capturent des requins suscite des préoccupations quant à leur état de conservation et leur gestion en raison de lacunes dans les données de prise, d effort et de rejets. En outre, il est patent que la quantité et qualité limitées des données disponibles influencent la soumission de l avis scientifique à la Commission. De nombreux aspects de la biologie de ces espèces sont encore méconnus ou totalement inconnus, notamment pour certaines régions, ce qui contribue à accroître les incertitudes dans les évaluations quantitatives et qualitatives. En ce qui concerne les données sur les activités halieutiques des flottilles capturant les requins (prise et prise accessoire), la soumission de la Tâche I et la Tâche II s est récemment améliorée mais cette amélioration n est toujours pas suffisante pour permettre au Comité de soumettre un avis quantitatif suffisamment précis sur l état du stock pour orienter la gestion des pêches vers des niveaux optimaux de capture pour la plupart des 2302

287 espèces. Il est donc essentiel que le Comité progresse en matière de recherche et de collecte de données sur le cycle vital, ainsi que sur la description des interactions avec les pêcheries de l ICCAT afin d évaluer l état des stocks et de soumettre un avis scientifique pertinent pour la gestion durable des pêcheries d élasmobranches dans la zone de la Convention de l ICCAT. Ce progrès est capital pour l évaluation de l efficacité des mesures de gestion adoptées par la Commission ces dernières années. Au cours de la réunion du Groupe d espèces sur les requins, tenue en 2012, le Groupe a recommandé d élaborer un Programme de recherche et de collecte de données sur les requins (SRDCP) portant sur la réduction des principales sources d incertitude dans la formulation de l avis scientifique, dont l amélioration des procédures de collecte et de déclaration des données. Faisant suite à cette recommandation, la réunion du Groupe d espèces de 2013 a élaboré les directives générales du SRDCP incluant les points ci-après : a) un contexte général des données biologiques et de pêche pour les principaux requins pélagiques de l Atlantique et de la Méditerranée, en mettant en évidence les principales lacunes dans les connaissances ; b) les principaux objectifs généraux du Programme ; c) les priorités en termes de collecte de données des pêcheries ; d) les priorités de recherche en termes de données biologiques; e) les priorité de recherche en termes de mesures d atténuation et f) d autres considérations pour le SRDCP. La mise en œuvre du SRDCP s inscrira dans le cadre du Programme stratégique du SCRS pour , lequel constituera un cadre général pour le développement et la coordination des activités scientifiques et liées à la science visant à étayer un avis scientifique robuste en tant qu élément central de la conservation et gestion des thonidés et espèces apparentées dans l Atlantique et en Méditerranée. Dans le cas des stocks pauvres en données, tels que les espèces de requins, une approche de précaution de la gestion des pêches pourrait implicitement tenir compte des incertitudes inconnues compte tenu de sa nature plus conservatrice. Et tout investissement en matière de recherche accroîtra les bénéfices potentiels des pêcheries relevant de l ICCAT tout en réduisant les risques présentés pour les ressources. B. REQUINS PÉLAGIQUES DE L ATLANTIQUE ET DE LA MÉDITERRANÉE 91 espèces de requins (requins et raies) ont été déclarées à l ICCAT. Conscient du besoin de limiter la portée du Programme, le Groupe d espèces sur les requins a pris en considération les espèces capturées (16 espèces représentent 95% des prises totales déclarées) ainsi que d autres espèces très susceptibles de l être et pour lesquelles peu d informations biologiques sont disponibles. Les espèces à prendre en considération sont les suivantes: (requin peau bleue (Prionace glauca; BSH), requin-taupe bleu (Isurus oxyrinchus; SMA), petite taupe (Isurus paucus; LMA), renard à gros yeux (Alopias superciliosus; BTH), renard (Alopias vulpinus; ALV), requin océanique (Carcharhinus longimanus; OCS), requin soyeux (C. falciformis; FAL), requin-taupe commun (Lamna nasus; POR), requin marteau halicorne (Sphyrna lewini; SPL), requin marteau commun (Sphyrna zygaena; SPZ), grand requin marteau (Sphyrna mokarran; SPK), requin gris (Carcharhinus plumbeus; CCP), requin de sable (Carcharhinus obscurus; DUS), requin de nuit (Carcharhinus signatus; CCS), requin cuivre (Carcharhinus brachyurus, BRO), requin tigre commun (Galeocerdo cuvier; TIG), requin crocodile (Pseudocarcharias kamoharai; PSK) et requin blanc (Carcharodon carcharias; WSH), ainsi que la pastenague violette (Pteroplatytrygon violacea; PLS) et raies mantas (Mobulidae, MAN). a) Connaissances biologiques actuelles Les informations de base sur le cycle vital requises pour évaluer l état des stocks de requins de l Atlantique sont plus abondantes pour la zone de l Atlantique Nord. Il existe considérablement moins de données pour les zones équatoriales et de l Atlantique Sud et très peu de données pour la Méditerranée. Par conséquent, plus de la moitié des études sur la dynamique de l âge et de la croissance, la reproduction, l identification des stocks et les schémas migratoires et de déplacement ont été réalisée dans l Atlantique Nord, la majorité correspondant à l Atlantique Nord-Ouest. De la même façon, la plupart des études de l Atlantique Sud correspondent à l Atlantique Sud-Ouest. Le Tableau 1 de l Appendice 8 récapitule les études menées pour toutes les espèces combinées dans chacune des neuf zones de l Atlantique et de la Méditerranée (Figure 1 de l Appendice 7). Les Tableaux 2-17 de l Appendice 8 présentent les mêmes informations spécifiques aux espèces pour 16 espèces. Le Groupe d espèces sur les requins générera des tableaux récapitulatifs analogues pour d autres espèces (requin cuivre, requin blanc, requins crocodiles et raies manta). L Appendice 9 répertorie toutes les références utilisées pour générer les Tableaux 2-17 de l Appendice 8. L Appendice 8 comporte également des références additionnelles qui ont été utilisées afin de générer les profils biologiques des espèces de requins et de raies soumis par le Groupe. 2303

288 Nous avons regroupé tous les paramètres du cycle vital ainsi que d autres paramètres répertoriés dans les tableaux de l Appendice en quatre catégories de données (reproduction, âge et croissance, identifiant du stock et schémas migratoires et de déplacement) les plus importantes aux fins des évaluations de stocks, et les dix zones géographiques en quatre zones principales (Atlantique Nord, Atlantique Sud, Atlantique équatoriale et Méditerranée) et nous avons examiné les informations spécifiques aux espèces. Nous avons utilisé une approche de feux lumineux pour identifier le degré de connaissances de ces catégories par zone et espèce générales : 1) rouge indiquant l absence totale d étude disponible, 2) jaune, 1 ou 2 études, 3) vert, 3+ études et 4) blanc, indiquant que l espèce n est pas présente dans une zone particulière (Tableau 1 de l Appendice 7). Les conclusions générales suivantes peuvent être tirées : l Atlantique Nord-Est la zone la plus riche en données mais il subsiste 25% des cellules sans information ; l Atlantique Sud et l Atlantique équatorial présentent des niveaux quasiment identiques de disponibilité des données, avec plus de 75% de cellules rouges; la Méditerranée est la région la plus pauvre en données avec près de 90% de cellules rouges. Les espèces individuelles ont été classées selon le degré de «pauvreté en termes de données» (c est-à-dire le nombre de cellules rouges ou sans information par rapport au nombre total de cellules pour cette espèce, comme illustré au Tableau 1 de l Appendice 7) et «richesse en termes de données» (c est-à-dire le nombre de cellules vertes ou avec 3+ études par rapport au nombre total de cellules pour cette espèce, comme illustré au Tableau 1 de l Appendice 7) (Tableau 2 de l Appendice 7). Les espèces les plus pauvres en données étaient la petite taupe, suivie du grand requin marteau, requin de sable, requins tigres et pastenague violette, alors que l espèce la moins pauvre en données était de loin le requin peau bleue. En revanche, le requin peau bleue, le requin-taupe bleu et le requin gris étaient les espèces les plus riches en données et il n y avait pas de «richesse en termes de données» pour la petite taupe, le requin marteau commun, le grand requin marteau et le requin de nuit. b) Information des pêcheries Les requins pélagiques constituent une grande partie des prises des pêcheries palangrières ciblant les thonidés, les istiophoridés et l espadon. Le Sous-comité sur les prises accessoires du SCRS de l ICCAT a commencé à évaluer les requins pélagiques en Les requins pélagiques sont capturés par divers engins dans l Atlantique, le golfe du Mexique, la Méditerranée et la mer des Caraïbes, notamment par la palangre, la senne, le filet maillant, la ligne à main, la canne et moulinet, le chalut, la ligne traînante et le harpon, mais ils sont surtout capturés en tant que prise accessoire dans les pêcheries palangrières pélagiques ou en tant qu espèce cible. Il y a aussi d importantes pêcheries récréatives dans certains pays. Plusieurs espèces de requins, telles que le requin peau bleue et le requin-taupe bleu, sont capturées et débarquées en grands volumes par ces flottilles. De 2001 à 2011, un total de et tonnes de requin peau bleue et de requin-taupe bleu, respectivement, ont été déclarées dans l Atlantique, avec une prise combinée maximum pour ces deux espèces enregistrée en 2010 ( tonnes) et une prise combinée minimum en 2011 ( tonnes) (Anonyme 2012). D autres groupes de requins pélagiques et de raies sont rejetées, soit en raison des recommandations de l ICCAT visant à l interdiction de leur rétention (Recommandations 09-07, et 10-08, 11-08), soit de leur faible valeur commerciale. Les CPC soumettent des données sur les requins depuis 1950, mais ce n est que depuis 1982 que les données sont soumises pour les espèces de requins autres que BSH, SMA et POR. Les données antérieures à 1990 sont très limitées pour la plupart des espèces et les données de Tâche I ne sont donc présentées ici qu après cette date. L Appendice 4 inclut la prise annuelle déclarée pour tous les requins et autres élasmobranches dans la base de données de Tâche I par pavillon (les données de 2012 sont préliminaires) et l Appendice 5 comporte la prise annuelle déclarée par espèce et zone de la Tâche I. Les données de taille de Tâche II ne sont présentées que depuis Afin d identifier les données disponibles, cette information est présentée sous forme de catalogue de données à l Appendice 6. La première réunion d évaluation des requins a été conduite en 2004 et ce n est qu en 2007 que le Groupe d espèces sur les requins indépendant a été formalisé. Hormis en 2010, chaque année, des réunions intersessions du Groupe d espèces sur les requins ont été tenues, avec une participation significative de scientifiques et d importants travaux sur ces espèces. La Figure 2 de l Appendice 7 illustre l évolution du nombre de documents présentés aux réunions intersessions. c) Évaluations des stocks des espèces Le Groupe d espèces sur les requins a mené à ce jour des évaluations des stocks de trois espèces: requin peau bleue, requin-taupe bleu et requin-taupe commun. Le requin peau bleue et taupe bleu ont tout d abord été évalués en 2004, puis en 2008 et 2012 (requin-taupe bleu uniquement). Le requin-taupe commun a été évalué en 2304

289 coopération avec le CIEM en En général, toutes ces évaluations sont considérées comme préliminaires en raison de la qualité et quantité limitées des informations disponibles et se sont concentrées sur les stocks de l Atlantique. Les stocks de la Méditerranée n ont pas été évalués faute de données. Une recommandation importante qui émane constamment des réunions du Groupe d espèces sur les requins est que de plus grands investissements en matière de suivi et de recherche sur les requins sont indispensables si l on souhaite améliorer l avis sur l état de ces espèces et d autres espèces de prises accessoires. - Requin peau bleue En se basant essentiellement sur les informations de marquage, l existence de trois stocks distincts de requin peau bleue a été postulée, mais seuls deux stocks ont été évalués (Atlantique Nord et Sud) en l absence d information sur le stock de la Méditerranée. Pour les stocks de l Atlantique Nord et Sud, bien que les résultats restent considérablement incertains, on estime que la biomasse se situe au-dessus de celle qui permettrait d atteindre la PME et que les niveaux de capture actuels se situent en-deçà de F PME. - Requin-taupe bleu Étant donné que le requin-taupe bleu a une répartition similaire à celle du requin peau bleue, les mêmes stocks hypothétiques de l Atlantique Nord et Sud ont été pris en considération pour cette espèce. L évaluation menée en 2012 de l état des stocks de l Atlantique Nord et Sud incluait des séries temporelles additionnelles de l abondance relative ainsi qu une couverture accrue des données de capture de Tâche I par rapport aux précédentes évaluations des stocks effectuées en 2008 et Les séries de CPUE disponibles montraient des tendances à la hausse ou planes pour les dernières années de chaque série (depuis l évaluation des stocks de 2008) pour les stocks Nord et Sud, indiquant donc que la surpêche potentielle affichée lors de la précédente évaluation des stocks a diminué et que le niveau de capture actuel pourrait être considéré soutenable. Pour le stock de l Atlantique Nord, les résultats des deux sorties des modèles des évaluations de stocks indiquaient presque unanimement que l abondance du stock en 2011 se situait au-dessus de B PME et que F était en deçà de F PME. Pour le stock de l Atlantique Sud, toutes les sorties des modèles indiquaient que le stock n était pas surpêché et qu il n y avait pas de surpêche. Même si les résultats indiquaient que les stocks de l Atlantique Nord et Sud sont relativement en bonne santé et que la probabilité d une surpêche est faible, ils présentent également des incohérences entre les trajectoires de la biomasse estimées et les tendances de la CPUE d entrée, ce qui donnait lieu à de vastes intervalles de confiance dans les trajectoires de la biomasse et de la mortalité par pêche estimées et d autres paramètres. Dans l Atlantique Sud, une tendance à la hausse des indices d abondance depuis les années 1970 ne correspondait notamment pas aux prises ascendantes. La grande incertitude quant aux estimations de capture passées et l absence de paramètres biologiques importants, en particulier pour le stock du Sud, constituent toujours un obstacle pour obtenir des estimations fiables de l état actuel des stocks. - Requin-taupe commun Le Groupe d espèces sur les requins a tenté d évaluer l état de quatre stocks de requins taupes communs (Nord- Ouest, Nord-Est, Sud-Ouest et Sud-Est) conjointement avec le Groupe de travail du CIEM sur les élasmobranches en En règle générale, les données sur le requin-taupe commun de l hémisphère Sud étaient trop limitées pour permettre une indication robuste de l état des stocks. Pour le Sud-Ouest, les données limitées indiquaient un déclin de la CPUE dans la flottille uruguayenne, les modèles suggérant un déclin potentiel de l abondance jusqu à des niveaux en-deçà de la PME et des taux de mortalité par pêche au-dessus de ceux permettant la PME. Toutefois, les données de capture et les autres données étaient généralement trop limitées pour permettre de définir des niveaux de capture soutenables. Pour le Sud-Est, les informations et données étaient trop limitées pour évaluer l état de ce stock. Le stock de l Atlantique Nord-Est compte le plus long historique d exploitation commerciale mais l absence de données de CPUE pour le pic de la pêcherie rajoutait de considérables incertitudes dans l identification de l état actuel par rapport à la biomasse vierge. Les évaluations exploratoires indiquaient que la biomasse actuelle (pour 2008) se situait en-deçà de B PME et que la récente mortalité par pêche était proche ou au-delà de F PME. Le rétablissement de ce stock à B PME sans aucune mortalité par pêche a été estimé prendre près de années. Une évaluation canadienne du stock de l Atlantique Nord-Ouest, présentée au cours de la réunion, indiquait que la biomasse était épuisée à des niveaux se situant bien en-deçà de B PME mais que la récente mortalité par pêche était en-dessous de F PME et que la récente biomasse semblait être en augmentation. Une modélisation de production excédentaire supplémentaire, conduite à la réunion, indiquait une vision similaire de l état du stock, c est-à-dire une raréfaction à des niveaux en-dessous de B PME et des taux de mortalité par pêche également en- 2305

290 dessous de F PME. L évaluation canadienne projetait que, sans mortalité par pêche, le stock pourrait se rétablir au niveau de B PME en ans environ, alors que les projections basées sur la production excédentaire indiquaient que 20 ans pourraient suffire. Dans le cadre de la stratégie canadienne d un taux d exploitation de 4%, il était prévu que le stock se rétablisse en 30 à 100+ ans. - Évaluation des risques écologiques (ERA) Le Groupe d espèces sur les requins a réalisé des évaluations des risques écologiques (ERA) en 2008 et L ERA de 2012 incluait 16 espèces (20 stocks) et a été généralement considérée plus fiable que l ERA de L ERA consistait en une analyse des risques pour évaluer la productivité biologique de ces stoc