Search

The ICES Workshop on ICES reference points (WKREF2) was tasked review the WKREF1 report and based on the outcome develop updated guidelines for the ICES reference points system and recommendations for ACOM consideration. The WKREF1 report has suggested 5 key recommendations to simplify and harmonise the ICES reference points framework representing a major change to the current guidelines. At WKREF2, we detailed discussions and four key concerns were raised about the proposed approach. The first related to the simplification of rules to define Blim. Around two thirds of category 1 stocks would end up as WKREF1 “Blim Type 2” where Blim would be set as a fraction of B0. The Allee effect or “depensation” maybe more important than previously thought and should be furthered explored for ICES stocks since it has important consequences for Blim. A number of challenges and issues around defining Blim using the current guidelines were documented. Some suggestions on improvement criteria were discussed including using classifiers to define spasmodic stocks and using change point algorithms to address non-stationary productivity regimes. However, further work is need to make these approaches operational and there was no consensus that the WKREF1 Blim types should replace the current guidelines.WKREF1 recommended that the FMSY proxy should be based on a biological proxies and should be less than the deterministic FMSY. It was pointed out that the stochastic FMSY estimated in EqSim for example, is lower than the deterministic FMSY and that the current guidelines ensure that the FMSY should not pose a more than 5% risk to Blim. A large amount of work described in WD 1 was carried out to develop an MSE framework to consistency and robustness test a candidate reference point system for North East Atlantic stocks. However, WKREF2 recommended that further work needs to be carried out to condition and test the simulation framework before the conclusions could be adopted by ICES and incorporated into the guidelines.A number of considerations for defining MSY related reference points were discussed including using model validation and prediction skill to ensure that ICES provide robust and credible ad-vice. There is evidence that density dependence (DD) is important in the majority of ICES stocks (68% in recruitment and 54% in growth). The correct prediction of the shape and strength of density-dependence in productivity is key to predicting future stock development and providing the best possible long-term fisheries management advice. A suggested approach to use surplus production models (SPMs) to account for DD in FMSY was suggested and discussed but there was no consensus on whether that approach was appropriate. There was consensus that the FECO approach as a means of adapting target fishing mortality to medium-term changes in productivity should be included in the guidelines subject to a benchmark and ACOM approval.While WKREF1 and 2 focused mainly on Category 1 stocks ToR c) called for a “simplified and harmonised set of guidelines for estimating MSY and precautionary reference points applicable in the advice framework across various ICES stock categories.” Ideally the ICES assessment categories should provide equivalent risk across all stocks. This issue was discussed but no recommendations emerged.There was no consensus a revised reference point framework was proposed at WKREF2. However, it was agreed that it should be presented here for further discussion at ACOM and other fora. The key feature of the suggested approach is that the stock status evaluation is treated independent of the Advice Rule (AR). The main feature of the system is that the biomass trigger is not linked to a stock status evaluation, it is linked to the expected biomass when fishing at the target fishing mortality, in contrast to the current ICES approach. It also entailed that FMSY would also become an upper limit of fishing mortality and that the advised fishing mortality would be set at or lower than that level. WKREF2 did not discuss what to do in situations where SSB< Blim or alternative forms of HCR for the advice rule. Building community understanding and consensus around simplified and harmonised guidelines has yet to be achieved. A further workshop WKREF3 will be required to achieve that aim. The report includes 6 recommendations for ACOM consideration.

The Working Group on Multispecies Assessment Methods (WGSAM) aims at enabling research on predator-prey interactions for developing advice on the ecosystem approach to fisheries management.This report details results related to WGSAM term of reference B: “Update of key-runs (standardized model runs updated with recent data) of multispecies and ecosystem models for different ICES regions”. Multispecies model key-runs are used in ICES advice processes, and WGSAM provides critical expert review of these key-runs to recommend appropriate use of results.Although key-run reviews have been conducted in the past, requests for reviews are increasing. Therefore, WGSAM first formalized a consistent set of review criteria to conduct key-run reviews. These are outlined in section 2 of this report and are posted online (https://iceseg.github.io/wg_WGSAM/ReviewCriteria.html). WGSAM then applied these review criteria to one key-run for the North Sea ecosystem. The review is detailed in section 3 of this report. As the review criteria were applied, WGSAM also noted any difficulties with the review process in order to further refine the review criteria and to make future key-run reviews more efficient and effective.WGSAM recommends the use of natural mortality estimates from the North Sea SMS key-run for use in single species stock assessment models of North Sea cod, haddock, herring, Norway pout, southern North Sea sandeel, northern North Sea sandeel, sprat, and whiting.

The Working Group on Multispecies Assessment Methods (WGSAM) aims to advance the operational use of knowledge on predator-prey interactions for advice on fisheries and ecosystem management. This report summarises the achievements of a 3-year cycle during which the group consolidated criteria to evaluate key-runs and more in general the skills assessment of multispecies models, released key-runs for the Baltic Sea, North Sea and Irish Sea all evaluated with those criteria, progressed in the areas of multiple models comparison, ensemble modelling and on the estimation of biological reference points in the context of multispecies interactions. The updated keyruns for the North Sea and the Baltic Sea provided the best available estimates of predation mortality for a number of key commercial stocks in these two ecoregions which have been already integrated into the stock assessments throughout benchmarks and inter-benchmarks. Analyses accumulate showing that ignoring strong trophic interactions may lead to bias in the perception of stocks status and in the calculation of reference points. Evaluations show advantages of using multi-model ensembles to capture the dynamics of the main stocks and the system overall. Results accumulated so far suggest that the benefits of ensemble modelling exist for both simple models, i.e. multispecies production models, as well as more complex ecosystem models. Various approaches are available to the practice of ensemble modelling, including a fully Bayesian ensemble framework suitable also for multi-model forecasts. The report includes also progresses with software developments to enhance accessibility of some complex routines, including ensemble modelling beyond “just a simple average approach” and computation of multispecies reference points, to a broader group of modellers and users. The group sees these developments as a great opportunity to work more towards cross-platform comparisons and further on multispecies skill assessment which will remain important themes for continuation of the work. To further progress the use of multispecies and ecosystem models, collection of ecosystem data remains highly relevant, with priority on stomach data and other information on processes affecting trophic interactions and trophodynamics of ecosystems (i.e., predator-prey overlap, temperature-dependent consumption, availability of other food).

WKRPChange was tasked with examining how ICES handles the estimation of target and limit reference points in the face of changing environmental conditions. In particular, the meeting participants were asked to review the robustness of the current ICES procedures and to suggest specific improvements that could be made, especially relating to changes in stock productivity arising from environmental conditions, species interactions, and density-dependent effects. Part of the work involved reviewing the basis of the ICES reference points and contrasting the ICES procedures with those in the USA and Canada, and part on providing specific guidance for future reference point estimation within ICES. One common approach to changing environmental conditions is to truncate data series. WKRPChange agreed that this may be necessary in some cases, but several studies were presented showing that the estimation of reference points becomes unreliable (both noisy and potentially biased) as the time series is reduced, and therefore recommended that modelling the specific process involved is generally a better approach than truncation. The meeting noted several examples within the current ICES management system for which reference points are allowed to vary (e.g. F in the case of NEA cod, Blim in the case of Iberian Sardine) according the prevailing conditions. WKRPChange noted that this was only required if conditions were expected to change significantly over the lifespan of the reference points, and that where it was implemented the status determination (the “traffic lights”) should be made accordingly.The key recommendation of WKRPChange is consistent with the conclusions of WKGMSE2, namely that a scoping exercise should be undertaken for each stock to identify any key drivers. Where there is good evidence for ecosystem-driven changes in stock productivity that process should be accounted for in setting reference points. The meeting highlighted that reference points have a finite lifespan, generally related to the benchmark cycle, and the estimation of the reference point should predominantly take into account processes likely to be important over that lifespan. WKRChange noted that many ICES stocks are managed by Harvest Controls which are evaluated through a MSE process. In this case there is considerable scope for including such environmentally driven processes in the Operating Model. However, many stocks are managed through the standard ICES HCR with reference points derived through the EqSim program. There is therefore a specific recommendation that the possibility to include density-dependentgrowth be incorporated into EqSim, to allow more realism to be included in the estimation of reference points where the evidence indicates that this is important within the reference point life span. The meeting also highlighted the recent work at WKRISH 6, which gave scope to “fine tune” the Ftarget to account for small changes in environmental drivers without requiring full reestimation of the reference points.

The inter-benchmark IBPSandeel evaluated the use of density-dependent survey catchability to reduce retrospective patterns in the assessment of sandeel in area 2r and 3r. IBPSandeel con-cluded that the method proposed is appropriate and can be applied to provide advice on these two stocks.The sandeel advice on fishing opportunities is highly influenced by good incoming year classes. High uncertainty is usually associated with estimation of high recruitments, which justified the implementation of an Fcap strategy. However, sandeel assessments in 2r and 3r have also strong retrospective patterns in recruitment (i.e. the model has a tendency to overestimate recruitment in the terminal year) which are not properly accounted by the current Fcap.Sandeel recruitment in 2019 is estimated to be high throughout the entire North Sea, including areas 2r and 3r. For this reason, an adjustment to the assessment model was proposed to provide more reliable estimates of recruitment in the terminal year. A power function was implemented in the assessment of sandeel 2r and 3r to capture density-dependency catchability of age 0 fish in the survey. The adjusted models provided downward correction of the terminal year recruit-ment, which is considered more reliable as suggested by the reduced Mohn's rho statistic, while estimates of stock dynamics (SSB and R) remained highly consistent with the previous assess-ment.

For the past couple of decades, the fishing pressure induced by towed demersal gear has decreased and this decrease may have allowed the number of sensitive fish in the sea to increase. Here we use the fact that the sensitivity of fish to trawl induced mortality depends on the life history of the fish species as well as on its susceptibility to towed demersal gear to identify species sensitive to towed demersal gear. We find that the sensitive species thus identified correspond well with the species classified by IUCN as threatened. Using trawl surveys in European Atlantic waters from 36oN to 62oN and weighing the influence of each survey with its variability around the long-term development, we estimated combined indicators of the change in abundance of each species. The majority of the sensitive species increased in numbers after 2000, with the most notable exceptions being starry ray and wolffish. These two species both showed continued decline since their highest abundance in the mid 1980’s. The increase in most of the species was not coupled to limitations in allowed catches, as most of the species have no species specific catch limit.

The Workshop on the management strategy evaluation of the reference point, Fcap, for Sprat in Division 3.a and Subarea 4 (WKSpratMSE) was held in ICES headquarters in Copenhagen 11-12 December 2018, and continued to work by correspondence through February 2019. The aim of the workshop was to estimate an appropriate Fcap for sprat in Division 3.a and Subarea 4 (Skagerrak, Kattegat, and North Sea) and sprat in divisions 7.d and 7.e (English Channel). This was done in response to the outcomes of the Benchmark Workshop on Sprat (WKSPRAT 2018). Two external reviewers participated and reviewed the analytical work and conclusions of the workshop. See full participant list in Annex 1. Sprat in the English Channel (Divisions 7de) is a category 3 stock and is assessed based on the acoustic survey carried out in area 7.e. Previously, the “2-over-3” rule had been used to provide catch advice, but evaluations during the last WKSPRAT benchmark suggested that such rule is not appropriate for short, highly productive species. The workshop therefore compared through simulations the performances of a “1-over-2” rule (with and without an uncertainty cap) and of different fixed harvest rates. The simulations suggested that a 1-over-2 rule for short-lived highly productive species might cause the stock to fall below safe levels and eventually to collapse because the rule is not reactive enough to limit the catches when there is a recruitment failure. When removing the uncertainty cap, performances improve, but the risk still remains above safe limits. The use of a fixed harvest rate seems to prevent risk and maximize catch, depending on the proportion used. Simulations suggest that a 20% harvest rate is considered appropriate to maintain the stock at safe biomass levels and to produce relatively high yield. One external reviewer suggested that these rules continue to be explored at the relevant upcoming Workshop for Data-limited Short-lived Stocks (WKDLSSLS). Sprat in Skagerrak, Kattegat, and the North Sea (Division 3a, Subarea 4) is a category 1 stock and follows the ICES MSY approach where the advised yearly catches correspond to the estimated stock biomass in excess of the MSY Bescapement, but constrained so that the fishing mortality is no higher than a limit F (Fcap). A “full” closed loop management strategy evaluation (MSE) was carried out to test for a range of F’s and find an appropriate Fcap value that showed a less than 5% probability of SSB < Bescapement. The value of Fcap was conditioned on using Bpa as MSY Bescapement. A “short-cut” MSE featuring an assessment emulator was implemented to compare the two approaches. The stock and fleet dynamics were conditioned on the most recent assessment from WKSPRAT2018. Performance statistics between the “full” and “short-cut” MSEs were similar, suggesting that short-cut MSEs should continue to be explored, as they have the advantage of being unaffected by assessment numeric convergence. According to these analyses, Fcap of 0.69 is precautionary.

The ICES herring assessment working group (HAWG) met for seven days in March 2018 to assess the state of five herring stocks and four sprat stocks. HAWG also provided advice for seven sandeel stocks but reported on those prior to this meeting. The working group conducted update assessments for five of the herring stocks. An analytical assessment was performed for North Sea sprat, and data limited assessments (ICES category 3 and 5) were conducted for English Channel sprat and Division 3.a sprat. The assessment of sprat in the Celtic Sea (spr.27.67a–cf–k) was not updated this year.The North Sea autumn spawning herring (her.27.3a47d) SSB in 2017 was estimated at 1.9 mill tonnes while F2–6 in 2017 was estimated at 0.21, which is below the management plan target F2–6 and below FMSY. Fishing mortality on juveniles, mean F0–1 is 0.032, below the agreed ceiling. Recruitment in 2017 is estimated to be very low. The estimate of 0-wr fish in 2017 (2016 year class) is estimated to be at approximately 18 billion, which is very low even compared to other recent low recruitments. Year classes since 2002 are estimated to be consistently weak with year classes 2015 to 2017 some of the weakest on record. ICES considers that the stock is still in a low productivity phase.The Western Baltic spring spawning herring (her.27.20-24) assessment was updated. The SSB in 2017 was relatively stable compared to recent years and is estimated to be around 104 000 tonnes. Fishing mortality has been estimated at 0.33 which is is above the estimate of FMSY (0.31). Recruitment has been low since 2006 and 2017 is the lowest observed in the time-series. Under a historical perspective the estimate of SSB of 104 170 tonnes in 2017 is considered low, below both Bpa and Blim, The stock has decreased consistently during the second half of the 2000s and given the continued low recruitments the stock seems not to be able to recover to these higher biomass levels.The Celtic Sea autumn and winter spawning stock (her.27.irls) is estimated to be at a low level, declining from a recent high biomass that peaked in 2011. SSB is currently estimated at 36 000 tonnes in 2017, coming down from 136 000 tonnes in 2011. Mean F(2–5 rings) was estimated at 0.4 in 2017, having increased from 0.06 in 2009. Recruitment has been good in recent years with several strong cohorts (2004, 2008, 2009, 2010 and 2013) entering the fishery but has come down substantially in the most recent years with the poorest year class in 2015.The 2017 SSB estimate of 6.a/7.b, c herring (Her.27.6a7bc) the combined stock of 6.aN and 6.aS/7.b, c was 113 000 tonnes, well below Bpa. Continued low recruitment has caused a decline of the stock which is expected to continue in the near future although fishing mortality is low at 0.05–0.1 in recent years.Irish Sea autumn spawning herring (her.27.nirs) assessment shows a stable SSB in 2017 compared to previous years at around 27 000 tonnes. The stock has experienced large incoming year classes in most recent years. Fishing mortality is estimated at the lowest level in the time series at 0.15, below FMSY. Catches have been relatively stable since the 1980s, and close to TAC levels in recent years.North Sea sprat (spr.27.4) is an ICES category 1 stock with a fully accepted analytical assessment. The stock is estimated to be at the highest level since the 1970s in 2018, driven by relatively high recruitment in 2017. The stock appears to be well above Bpa (142 000 tonnes) in 2018 at 408 808 tonnes. Fishing mortality in the last years has fluctuated between 0.4–2.3. A recent management strategy evaluation (WKMSYREF2) suggested that the Bescapement management strategy is not precautionary.Sprat in Division 3.a (spr.27.3a) was benchmarked in 2013 (WKSPRAT), however, it is an ICES category 3 stock that uses three survey series, from which an abundance index is derived, and landings data. ). The surveys show variability over time without a clear trend. The most recent annual change is positive compared to the 4 years before.Catch advice for sprat in the English Channel (7.d, e) (spr.27.7de) was based on criteria for an ICES category 3-based method. Data available are landings and a short time series of acoustic biomass (2013–2017). The acoustic biomass indicates an overall decline in the stock size. Quantitative advice was provided for Sprat in the Celtic Sea (spr.27.67a–cf–k) in 2017 using an ICES category 5-based method where only data on landings are available; this assessment was not updated this year as the advice is still valid and the perception of the stock has not changed. A sprat benchmark is taking place in the summer/autumn of 2018.The HAWG reviewed the assessments performed on seven sandeel stocks and the related advice of these stocks. Section 11 of this report contains the assessments of sandeel in Division 3.a and Subarea 4. Standard issues such as the quality and availability of data, estimating the amounts of discarded fish, availability of data through industry surveys and scientific advances relevant for small pelagic fish were discussed. All data and scripts used to perform the assessment and perform the forecast calculations are available on GitHub and accessible to anyone.

The benchmark workshop on sprat (WKSPRAT) was held in ICES headquarters in Copenhagen from 5–9 November 2018. The benchmark process started in June 2017 with a 3 day data compilation meeting, also held at ICES. Three stocks were benchmarked: North Sea sprat, 3.a sprat and the Channel sprat. During the benchmark process, several arguments were put forward on the connection between the North Sea stock and the 3.a stock in Kattegat-Skagerrak. It was therefore agreed to merge the two stocks and assess them as one stock assessment unit. Sprat in area 4 and 3a During the 2018 benchmark, enough evidence were presented to merge the sprat stocks in the Skagerrak-Kattegak and in the North Sea. The catch data and the indices of abundance from 3.a were included in the data from area 4. Three surveys are carried out in the area (IBTS in Q1 and Q3 and HERAS survey) and were all used as tuning indices in the model. The IBTS indices were standardized using a delta-GAM approach: the inclusion of 3.a data increased the internal consistency between all age classes for all indices. The SMS model, previously used to assess the North Sea component, was used to assess the new combined stock. The model year is resolved into 4 seasons, where season 1 corresponds to quarter 3 of the calender year. The term ‘season’ is hence used here to identify the model year, while the term ‘quarter’ refers to the calendar year. The final model formulation includes a power function for the age 0 catchability of IBTS Q1, a constant maturity ogive and the inclusion of the very few catches reported for season 4 (Q2) into season 1 (Q3) of the following year. The new stock assessment shows a considerable improvement in the retrospective pattern, as well a better fitting to some ages of the IBTS surveys and catches (as indicated by survey and catch CVs). The stock reference points were revised following ICES standard guidelines. Blim was estimated as the breakpoint of the segmented stock recruitment relationship and Bpa was derived from Blim. Short-term forecast settings were updated. The Management Strategy Evaluation (MSE) settings for the estimation of the Fcap were presented and discussed. Several scenarios were proposed to be tested and results will be evaluated during an ad-hoc MSE workshop that will take place on 11–12 December 2018. Sprat in area 7d,e Not enough evidences were available to change the boundaries of the English Channel stock. The stock is a category 3, and the data available are still scarce. An acoustic survey (PELTIC) is carried out in the English part of area 7.e since 2013. In 2017, the survey has been extended to include also the French part of 7.e and in 2018 the Eastern Channel was surveyed as well. Also, an IBTS index in Q1 is available for the Eastern Channel from 2007 onwards. The time series from the PELTIC survey has been used since 2017 to provide advice. At the benchmark, the updated time series was presented: the method used to produce the acoustic index was revised compared to previous years and it was changed to StoX (StoX, 2015). Concerns were raised because of the uncertainty in the stock distribution, and the fact that the acoustic survey covers only the Western part of the English Channel. No analytical assessment seems to work so far: both a seasonal and non-seasonal SPiCt were attempted during the benchmark workshop, but the short time series in the acoustic index and the lack of any contrast in the data still does not allow the model to converge. The group agreed to provide a seasonal advice based on an empirical method, i.e. to use both indices for trends, but only the acoustic survey for provision of advice. Two advice rules were proposed: the 1 over 2 rule, and a fixed F rule. Both options will be tested in an MSE framework and presented during the ad-hoc MSE workshop that will take place on 11–12 December 2018.

The ICES herring assessment working group (HAWG) met for seven days in March 2017 to assess the state of five herring stocks and four sprat stocks. HAWG also provided advice for seven sandeel stocks but reported on those prior to this meeting. The working group conducted update assessments for five of the herring stocks. An analytical assessment was performed for North Sea sprat and data limited assessments were conducted for English Channel sprat, Celtic Sea sprat and 3.a sprat.The North Sea autumn spawning herring SSB in 2016 was estimated at 2.20 m tonnes while F2–6 in 2016 was estimated at 0.26, at the management plan target F2-6 and below Fmsy. Fishing mortality on juveniles, mean F0-1 is 0.05, just below the agreed ceiling. Recruitment in 2017 is estimated to be very low. The estimate of 0-wr fish in 2016 (2015 year class) is estimated to be at approximately 29 billion, being low but in line with recent recruitment. Year classes since 2002 are estimated to be consistently week with year classes 2002 to 2007 to be among the weakest. ICES considers that the stock is still in a low productivity phase. The Western Baltic spring spawning herring assessment was updated. The SSB in 2016 was relatively stable compared to recent years and is estimated to be around 97 000 tonnes. Fishing mortality has been estimated at 0.41 and seems to have increased again after a period of reductions. It is above the estimate of Fmsy (0.32). Recruitment in 2016 is very low and potentially the lowest in the time-series. Under an historical perspective the estimates of SSB are considered still low, and the stock seems not to be able to recover to these higher biomass levels. The Celtic Sea autumn and winter spawning stock is estimated to be at a low level, declining from a recent high biomass that peaked in 2011. SSB is currently estimated at 46 000 tonnes in 2016, coming down from 140 000 tonnes in 2011. Mean F (2–5 rings) was estimated at 0.4 in 2016, having increased from 0.07 in 2009. Recruitment has been good in recent years with several strong cohorts (2003, 2005, 2007, 2011, 2012) entering the fishery but has come down substantially in the most recent years with the poorest year class in 2015. The 2016 SSB estimate of 6.a/7.b, c herring (the combined stock of 6.aN and 6.aS/7.b, c) was 151 000 tonnes, well below Bpa. Low recruitment has caused a decline of the stock while fishing mortality is low at 0.05-0.1 in recent years. Advice has been drafted to setup a monitoring fishery to ensure data relevant for the assessment and genetic studies are secured. Irish Sea autumn spawning herring was benchmarked in 2017 and the assessment shows a stable SSB in 2016 compared to previous years at around 26 000 tonnes, estimated substantially higher than pre-benchmark. The stock increased owing to large incoming year classes in most recent years. Fishing mortality is estimated at the lowest level in the time series at 0.17, below Fmsy. Catches have been relatively stable since the 1980s, and close to TAC levels in recent years. North Sea sprat came down from a time-series high since the early ’80, driven by high recruitment in 2014 and shows another increase owing to the 2016 year class. The stock appears to be well above Bpa (142 000 t) in 2016 at 246 170t. Fishing mortality in the last years has fluctuated between 0.4–1.6. Expected recruitment for 2017 is estimated to be in line with long-term recruitment. Sprat in Division 3.a was benchmarked in 2013 (WKSPRAT) but an analytical assessment is not presented. Short term projections are to be based on a combination of indices providing in year advice for 3.a based on the ICES approach for data limited stocks (Category 3 / 4). (Category 3/4). The surveys show variability over time without a clear trend. The most recent change is negative compared to the 4 years before. Catch advice for sprat in the English Channel (7.d, e) was based on criteria for data limited stocks. Data available are landings and a short time series of acoustic biomass (2013–2016). The acoustic biomass indicates a decline in the stock. Quantitative advice was provided for Sprat in the Celtic Sea (spr-irls) based on criteria for data limited stocks where only data on landings are available. The HAWG reviewed the assessments performed on seven sandeel stocks and the related advice of these stocks. Section 11 of this report contains the assessment of sandeel in Division 3.a and Subarea 4.Standard issues such as the quality and availability of data, estimating the amounts of discarded fish, availability of data through industry surveys and scientific advances relevant for small pelagic fish were discussed.All data and scripts used to perform the assessment and perform the forecast calculations are available on GitHub and accessible to anyone.