Joint Norwegian-Russian Fisheries Commission request to establish a management plan for capelin in the Barents Sea (subareas 1 and 2 excluding Division 2.a west of 5°W)

Harvesting rules were assessed for capelin (Mallotus villosus) in the Barents Sea using a management strategy evaluation (MSE), a modeling framework that simulates population and fishery responses to management actions. The form of the current escapement rule is retained and is defined by B_escapement, the biomass that must escape to spawn with 95% probability after fishing is accounted. The MSE specifically tested four different B_escapement values (100,000, 150,000, 200,000 and 400,000 tonnes) with and without three alternative fixed minimum quotas (25,000, 50,000 or 75,000 tonnes). All four B_escapement values without fixed minimum quotas maintained a low risk (<5%) of spawning biomass falling below B_lim, the SSB below which recruitment is reduced, assuming historical capelin productivity, accurate survey estimates and survey precision is correctly estimated. However, a B_escapement equal to 100,000 tonnes showed notably higher risk (of SSB<B_lim) if the survey is either biased high in reality or survey precision is underestimated by management. High probabilities of fishery closures resulted from B_escapement=400,000 tonnes. All of the alternative rules using fixed minimum quotas showed very high risk of SSB falling below B_lim and the model framework projected reduced future recruitment to the extent of stock collapse over the long term. In general, average catch decreased and the years with a closed fishery increased with higher B_escapement. When selecting a rule, managers should also consider the trade-offs with other consequences and potential impacts on the ecosystem given the critical role played by capelin as the key forage fish for various predators in the Barents Sea.

During a recent benchmark for capelin (Mallotus villosus) in the Barents Sea (ICES 2023), the biological reference point B_lim was changed from 200,000 to 68,000 tonnes to follow the standards for defining B_lim used by ICES (i.e. the SSB below which recruitment is reduced). Due to the redefinition of B_lim, the term B_escapement was introduced as the reference point actually used in the present harvesting rule. B_escapement corresponds to the minimum biomass that must be left to spawn each spring with at least 95% probability (which is in accordance with the precautionary principle) and is set by management. Due to unaccounted uncertainties in the assessment, B_escapement is not necessarily equal to B_lim, but cannot be lower than B_lim. Thus, 200,000 tonnes was set as B_escapement during the capelin benchmark, but alternative values for B_escapement should be explored with an evaluation of the harvesting rule.

The Joint Norwegian-Russian Fisheries Commission (JNRFC) requested in 2023 the evaluation of alternative harvesting rules for Barents Sea capelin and that the results be used to inform managers in 2024. The Institute of Marine Research (IMR) developed a Management Strategy Evaluation framework (MSE) to evaluate the consequences of a set of harvesting rules, both of which were specified by the JNRFC.

Based on discussions with industry and managers in Norway, a list of alternative harvest control rules to be tested was proposed. This list was adopted by Norwegian and Russian scientists and used in the evaluation.

The harvesting rules are as follows and specify different target escapements and the potential introduction of fixed minimum quotas, which would be taken in years when the fishery would otherwise be closed according to the escapement rule:

• Harvest rule A: Existing harvesting rule which states that with a 95% probability one must allow at least 200,000 tonnes (B_escapement) to have the opportunity to spawn. The capelin estimate from the ecosystem survey in autumn is used as a basis.

• Harvest rule B: Like rule A, but with alternative values for B_escapement (400,000, 150,000, 100,000 tonnes).

• Harvest rule C: Like rules A and B, but where a minimum quota of 25,000, 50,000 or 75,000 tonnes is given each year anyway.

These three rules imply a greater number of unique harvesting rules equal to 16, as the four possible values of B_escapement are individually tested, plus each B_escapement paired with each of three possible fixed minimum quotas (4 + 4*3 = 16). A higher relative risk (in terms of falling below B_lim and complete stock collapse) can be inferred for rules with lower B_escapement and fixed minimum quotas compared to the highest B_escapement with fixed minimum quotas. Thus, after testing all four B_escapement values without a fixed minimum quota, only the highest B_escapement is tested with each of the fixed minimum quotas simulations. This approach avoided unnecessary simulation testing of lower B_escapement with fixed minimum quotas that inherently carry greater risk.

The consequences of the harvesting rules are assessed in a long-term perspective from autumn 1987 to spring 2023 (36 years of SSB at spawning time in the spring) where reports are requested on the following parameters:

• Average spawning biomass

• Average quota

• Probability of spawning biomass falling below B_lim (generally referred to as risk)

• Probability of spawning biomass falling below B_escapement

• Number of years in which fishing is not opened

The accuracy of these specific consequences relies on the stock estimate (i.e. the most important input to the harvest rule) being unbiased, the uncertainty of that estimate being correctly characterized, and the estimated B_lim being precise and accurate (ICES 2023); however, none of these three points can be guaranteed and are major uncertainties the requested harvest rules need testing against. Thus, the MSE analysis is conducted in two general parts:

harvest rules are simulation tested with a base scenario characterized by annual stock estimates being unbiased and survey errors being correctly assumed

only harvest rules that have low probability of falling below B_lim is <5% under the base scenario are simulation tested against two additional scenarios; one where surveys overestimate the true stock and one where survey estimates are assumed to be overly precise.

Additionally, a sensitivity analysis is conducted of historical risk to uncertainty in B_lim, and specifically involves projecting the historical 1 April SSB distributions under each alternative B_escapement and then computing the proportions of instances where a random SSB value from this distribution is less than a random B_lim.

4.1 - Methods

The assumptions and formulas used in the MSE are based on the current assessment method for capelin, a within-year forecast of the maturing biomass that is equal to the autumn survey estimate (bifrost). As bifrost accounts for predation mortality due to cod consumption and thus relies on estimates of historical cod abundance, a hindcast MSE was conducted which tests the proposed harvesting rules against the historical observed capelin productivity and conditions. In other words, this MSE addresses the question of how management could have performed in the past using today’s assessment method and harvest rule. Additionally, the results are conditional on the implicit assumption that stock productivity in the future will approximate that in the 1987-2022 time series. In the event that there is evidence for systematic changes in productivity, then the HCR should be re-evaluated.

In short, the MSE simulates in each year the true conditions of the capelin stock (also referred to as the operating model); the possible mismatches between autumn survey observations and the ‘true’ stock; the full stock assessment method based on simulated observations, harvest rule, and quota calculation; and what is actually harvested from the ‘true’ stock based on the quota. Catches are assumed equal to the annual quotas within the MSE, as historical catches have closely adhered to the real advice. This simulation sequence is repeated over a number of years and allows feedback between years (e.g. previous years’ catches directly impact future capelin abundance and reproduction). Once capelin population and fishing dynamics are projected over all years within the pre-specific time period (1987-2022), this sequence is repeated over and over again, each time with different conditions including different recruitment trajectories, different cod abundances and consumption rates, and different survey observations. The consequences of the harvesting rules are calculated over the many possible capelin stock and fishing trajectories.

The MSE framework was custom coded in and the analysis conducted with the R programming software. Code and data files of the MSE results are available on a public Gitlab repo (https://git.imr.no/MSE/barents_capelin_2024).

Technical details of the capelin MSE methodology and results are described in the report this document summarises (CITE REPORT AND ADD TO REFERENCES LIST AT BOTTOM].

4.2 - Results and conclusions

Results based on historical capelin stock dynamics and correct survey assumptions (base scenario) showed all B_escapement values without fixed minimum quotas to have low risk of falling below B_lim (68,000 tonnes) (Figure 1).

The fixed minimum quotas were tested with B_escapement=400,000 tonnes, and the probability of stock collapse exceeded the 5% for all minimum quota values (Figure 2). All similar rules with a lower B_escapement would have a higher chance of stock collapse. We therefore conclude that all of the fixed minimum quota rules have a high risk of causing SSB to fall below B_lim and stock collapse. We note here that the risk does not apply just to the capelin stock, but to the wider ecosystem, given the key role capelin play as a food source in the Barents Sea.

Evaluations based on biased survey indices (i.e. overestimating true stock biomass by 20% or 10%) showed higher risk of falling below B_lim with B_escapement=100,000 tonnes (>5%), while comparatively low risk was seen with the other B_escapement values. Underestimated survey CV input to the stock assessment also resulted in B_escapement=100,000 tonnes showing risk greater than 5% of falling below B_lim.

An asymmetric trade-off is seen between average yield and average SSB, where greater gains in yield occur with less relative reductions in SSB as B_escapement is increased. The average number of years of a closed fishery was similar for the three lowest escapement rules (100,000-200,000 tonnes) where medians were within 1-2 years of each other. Additionally, high probabilities of fishery closures resulted from B_escapement=400,000 tonnes. Overall, values of 150,000 and 200,000 tonnes for B_escapement were showed to be most robust against key uncertainties in survey assumptions, and viable for the fishery (e.g. in terms of fishery openings).

ICES. 2023. Benchmark workshop on capelin (WKCAPELIN). ICES Scientific Reports, 5: 282.X