WKPELA - Report of the Benchmark Workshop on Pelagic Stocks

The Benchmark Workshop on Pelagic Stocks (WKPELA) meeting took place in Lisbon, Portugal, on 6–10 February 2017, hosted by IPMA (at its head quarter in Alges). It was chaired by Andrés Uriarte (ICES Chair), Spain, and by Dankert Skagen (External Chair), Norway, with the assistance of the invited external experts Yi-Jay Chang, Tai-wan, and Martin Dorn, USA, (and ICES Professional secretary: David Miller/ Support secretary:  Sarah Millar). The meeting was attended entirely or part time by 20 scientists from France, Spain and Portugal and 3 stake holders from the Netherlands, Spain and Portugal (Annex 1 list of participants).  Three stocks passed the benchmarking process: Sardine (Sardina pilchardus) in Divisions 8c and 9a (Cantabrian Sea, Atlantic Iberian Waters); Sardine in Divisions 8a,b,d and Subarea 7 (Bay of Biscay, Southern Celtic Seas and English Channel) and Horse mackerel (Trachurus trachurus) in Division 9a (Atlantic Iberian waters) (Southern stock). Decisions on the assessments were undertaken after technical discussions and agreement among the ICES members and invited external experts for the Workshop.  
For the southern Sardine (Divisions 8c and 9a), no need for a revision of the current stock boundaries was evidenced. Some immigration of sardines from the north (8abd, Bay of Biscay) to the Cantabrian sea (8c) may happen but this was perceived to be lim-ited (SARDYN project 2002–2005) and recent studies on spatial cohort dynamics and growth patterns confirms the limited nature of such straying. Within the southern stock there are signs however of some regional sub structuring and potentially differ-ent population dynamics, especially concerning the Gulf of Cadiz area (9aSouth), as shown by analysis of otolith shape, morphometrics and recruitment dynamics. How-ever, they were not considered enough as to change current stock definition.  
For this southern sardine stock, detailed reviews of the data inputs were prepared in advance including, among others, revised estimates of the mean weights at age in the catches and the stock, small revision of the abundance of age 1 estimates in the acous-tics surveys and major revisions of the SSB estimates from the DEPM surveys, the latter supported by ICES WGACEGG. Assessment was carried with Stock Synthesis 3 and starts in 1978. Regarding the fishery and population modelling in the assessment, the main adopted modifications refer to the methods to estimate the initial population (now starting at equilibrium since 1972), the stock-recruitment relationship (lognormal deviations from a Beverton-Holt stock recruitment model with steepness fixed at 0.71), the acoustic survey selectivity-at-age (fixed over the assessment period at 1 for all ages 1-6+) and the fishery selectivity-at-age (fixed for ages 3–5 for all the three periods con-sidered: 1978–1987, 1988–2005 and 2006–2016). Furthermore, as a result of a systematic review of the fitting achieved for a combination of various scenarios of survey selectiv-ity and fishery selectivity and for seven M scenarios, it was found that the assessment was optimised for a vector of natural mortality slightly higher than previously as-sumed. Accordingly, M at age was set equal the expected Gislason et al., values at age multiplied by a factor of 0.7. The Final adopted assessment model shows a better fit to the data available and provides more precise estimates of biomass, recruitment and fishing mortality than the last sardine assessment (WGHANSA 2016). The new assess-ment produces rather consistent estimates with former ICES assessments though some decrease of the biomass in 2016 (by about 25%) is obtained. Definition of Biological reference points could not be covered during the WKPELA meeting and were left for a posterior work. For the Northern sardine (Divisions 8a,b,d and Subarea 7), the basis for dealing to-gether all these regions were reviewed. The workshop concluded that in the absence of firm evidences of connectivity between the Bay of Biscay and subarea 7 sardine pop-ulations, and taking into account the indications of shelf sustained populations in each area (whereby all stages are found in substantial amounts in both regions – eggs, larval, juveniles and adults), it would be preferable to deal with the Bay of Biscay and subarea 7 separately as different stock units. The conclusion of WKPELA was that currently the limited and poor quality of the data available for the English Channel and Celtic sea region prevents producing an assessment of subarea 7. Furthermore, a joint assessment of Divisions 8abd and subarea 7 require the strong assumptions of similar population trends and age composition of catches in both areas, which is currently perceived as highly unlikely (according to the recent survey estimates and indications of different growth patterns). Therefore the conclusion, which the reviewers supported, was that it could be justified to explore assessing the Bay of Biscay part (Divisions 8a,b,d) as a separate stock unit.
An isolated assessment for the sardine in the Bay of Biscay (8abd) was prepared based on Stock Synthesis 3 and starting in 2000. Assessments with a4a were also tried in par-allel, mainly for verification of results. Input data were: catches at age, acoustic esti-mates of population biomass and age structured information, a series of Egg Abundance as indicator of spawning biomass and two SSB estimates of DEPM surveys in 2011 and 2014. A complete systematic review of the fitting achieved for a combina-tion of various scenarios of survey and fishery selectivity and for different patterns at age of natural mortality led to optimise the assessment for a vector of natural mortality as proposed by Gislason et. al, but multiplied by factor of 0.9., together with assuming a fixed fishery survey catchability for ages 35 and flat survey catchability for ages 2–5 over the entire assessment period. Such setting of the assessment was accepted as the best model for providing scientific advice. This assessment supposes an improvement over the previous survey trend-based assessment applied by ICES. However biomass estimates from SS3 are lower than the biomass indices derived from the surveys which intend to be provide absolute estimates of biomass (estimated catchability for Pelgas biomass index, 2.4, and for DEPM 1.8, and these values are perceived to be too high). The WK recommended to use the results of the assessment as indicative of trends only (category 2 stocks). Lack of time left the definition of biological reference points for a later stage after the workshop.  
For southern Horse mackerel (in Division 9a), recent studies support the current as-sumption of stock identity. Little review of the data input was required, this only af-fected to horse mackerel bottom trawl survey data from the Portuguese and Spanish IBTS, which was analysed from 1983–2015. The expert group decided that the “stand-ard” stratified mean was an acceptable method to deal with the non-normal abundance distribution of horse mackerel and with the variability in the survey data. Therefore, this estimator replaced the simple average method used before as the abundance index for tuning the assessment. The stock assessment was performed with the AMISH (As-sessment Method for the Ibero-Atlantic Southern Horse mackerel), as in the last bench-mark of the stock. Major Model assumptions were: Initial recruitment (age 0) is governed by the Beverton and Holt S-R relationship, assuming a steepness of h=0.8; One selectivity block for the survey abundance index, and three fishery selectivity blocks for the fitting of catch-at-age (1992–1997, 1998–2011, 2012 onwards) with Selec-tivity-at-age constant for ages 7+. The three time-blocks for the catch selectivity accom-modates the recent changes in the fishery due to the strong year-classes of 2011 and 2012 and the increase of horse mackerel catches by purse-seiners. The outputs from the adopted assessment were very consistent with the former ICES assessment and so were the associated biological reference points (which did not require to be changed).  
External experts endorsed the decisions made for setting up the assessments for the provision of advice, as reflected in the stock annexes and justified along the main text of the report. Finalising the work to address ToRs f) and g) outlined above, on the Eval-uation of the compliance of the southern Sardine management plan with the precau-tionary approach, was delayed to a second WKPELA meeting foreseen on 29–31 May 2017.