Report of the Benchmark Workshop on Pelagic Stocks (WKPELA)

The Benchmark Workshop on Pelagic stocks (WKPELA) 2016 took place in two parts. A data compilation workshop, 24–26 November 2015, and the main meeting, 29 February–4 March 2016, both at ICES HQ.
There were 23 participants from 7 countries. Out of the 23 participants, 6 were on behalf of stakeholders. Additionally, three external reviewers, two from outside the ICES community, took part in the meeting and reviewed what was presented. The complete participants list can be found in Annex 1.
The purpose of the meeting was to examine the assessment of the Norwegian spring-spawning herring stock (NSSH), both in terms of data and methods. This benchmark meeting was the second, the first benchmark took place in 2008. In 2008 the VPA module in the TASACS toolbox was chosen to be the standard assessment tool in the next years. The assessments showed retrospective problems around 2010–14, but have been quite consistent in the last couple of years. An analysis showed that this problem was mainly caused by ages 13–15 in one survey. The TASACS assessment, however, did not perform well when the plusgroup was lowered to age 12.
WKPELA accepted the findings that there were no indications of increased natural mortality in a period after around 2009. In this respect, biological data from 626 749 individuals of NSSH were analysed. These had been collected internationally in the years 1994–2015 over the full distribution area and over the whole year.   
Some of the surveys time-series have now been re-evaluated. The revised data seemed to improve the assessment, and WKPELA accepted these new survey indices. The re-vision of the survey data leads to higher spawning-stock biomass (SSB) in the years after 2004, compared to the final TASACS assessment in 2015. The upward revision is in the range of 500–800 thousand tonnes.
WKPELA accepted updated values for maturity-at-age in the years 2005–2011. The ma-turity data were last revised in 2010. WKPELA suggests updating the maturity values each year according to the procedure in WKHERMAT 2010. The revision of the ma-turity data alters the perception of the spawning-stock biomass in these years, lowering the peak in 2009, but raising the estimate in 2007.
Three assessment models were explored, TASACS, XSAM and one separable model. XSAM is a model template based on a state space model and structural time-series models for fish stock assessments (note that XSAM is not the same as SAM). The main new achievement in this framework is to utilize prior knowledge of sampling errors to a greater extent to improve inference than what has been possible earlier. WKPELA accepted XSAM as the standard assessment tool for the NSSH. However, TASACS, with the same settings as in 2015, shall be used as an alternative assessment, if for some reasons XSAM should fail. XSAM is still under development, so WKPELA suggests allowing for small adjustments if needed. A short-term forecast module is not yet ready, but will be before WGWIDE 2016. The code, a documentation, and an example of how to use the framework will be made available before WGWIDE 2016, at its Shar-Point site. The benchmarked assessment for NSSH uses the same years as the TASACS assessment, but age range is 3–12+ instead of 0–15+. Two surveys on the adult population are used in XSAM, instead of eight surveys on all life stages in TASACS. Broadly speaking, SSB trends estimated by XSAM and TASACS are similar and the SSB estimated by TA-SACS lies within the 95% confidence interval from XSAM.
The reference point Blim was re-examined. The evaluation did not change it, Blim = 2.5 million tonnes is still considered appropriate. The F in the management plan was also re-examined. The current value of 0.125 is considered adequate and the management plan can still be considered consistent with the precautionary approach, but it has not been evaluated whether this will lead to maximum long-term yield.
The stock annex was updated according to the decided adjustment to the assessment input data and methodology. A description of the short-term forecast module in XSAM, and a detailed description of how to run XSAM are still outstanding, but will be ready before WGWIDE 2016 when the new assessment model will be used for the first time to provide catch advice.