The third Workshop on Data Limited Stocks of Short-Lived Species (WKDLSSLS3) had as main aim to further develop assessment, catch advice and management methods for short-lived stocks in ICES categories 3–4, focusing on the provision of advice rules that are within the ICES MSY framework.

Regarding assessment methods (TOR 1) for data-limited short-lived species, a review of implementing precautionary harvest control rules based on SPiCT assessments was presented to the group (based on Mildenberger et al. 2021). Two approaches were tested: using biomass thresholds and limits, and using “uncertainty buffers”. General management strategy evaluation (MSE) identified the range 0.15-0.45 of uncertainty buffer fractiles as better performing. Shorter-lived species require higher thresholds than longer-lived species.

Exploratory approaches for the assessment of two category-5 sprat stocks were presented to the group. Both stocks have issues of stock identity and lack survey coverage over their spatial distribution and would benefit from acoustic surveys. A preliminary operating model for sprat in the Celtic Seas was parameterised incorporating life history information from neighbouring stocks and literature, though some outstanding issues still needed to be addressed (e.g. virgin biomass, spatial structure or steepness). For the Scottish Mallaig Sprat, a collaborative project between Marine Scotland Science and fishing industry investigates spatial limits of the stock while improving the landing sampling. VMS data confirmed that the fishery is restricted to the inshore areas of the Minch, Mull and Skye, and the biological sampling revealed catches were dominated by age 1 fish, in contrast to the groundfish survey in subarea 6.a, suggesting that the Mallaig sprat could be a subset of a wider population. A preliminary SPiCT assessment was

presented using a commercial CPUE and unsuitable demersal survey indices.

In relation to TOR 2, on the evaluation of the appropriateness of the management procedures based on direct use of abundance indices (for category 3 stocks), first a summary of the work and conclusions of WKLIFE VII-X (ICES, 2017, 2018, 2019, 2020) on empirical (i.e. model-free) management procedures was presented, highlighting limitations for the faster-growing species and suggesting alternative management procedures (e.g. harvest rate-based rules or escapement strategies). Recommending specific harvest rates requires caution as optimum harvest rate levels can be narrow and depend on simulation assumptions.

The effect of seasonal advice schedule (July-June) was investigated for English Channel sprat. During the stock’s interbenchmark, an annual MSE was not able to investigate within-year processes. A novel intra-annual MSE (Mildenberger et al., 2021) was parameterised for the stock, accounting for seasonal growth and exploitation. The timing and lag between events within the year (e.g. survey observation, implementation of advice, recruitment) affect the performance of Harvest Control Rules (HCR). The interbenchmark decision of 8.57% Constant Harvest Rate (CHR) seems to be appropriate. Annual simulation studies make coarse approximations of what are assumed to be smooth biological processes and likely underestimate risk of higher CHRs and overestimate catch of lower CHRs.

A MSE was presented investigating the performance of alternative dynamic harvest rate rules (HR) under the in-year calendar. The HRs were adapted from Carruthers et al. (2016) transforming them from TAC modifiers to HR modifiers. Additionally, a novel Perturbation rule (Pert) was included in the MSE. It was shown that some of these rules were able to reduce risks to values at or below 25% in most of the cases, with relative yields ranging from 50% to 150% MSY, depending mostly on the initial exploitation status. Compared with the default 1-over-2 rule with 80% uncertainty cap, the retained harvest rate rules implied in the long term less reduction of catch for higher risks, though below 0.2 in the long term.

Finally, issues of the initial implementation of the 1-over-2 rule to moderately exploited stocks were presented, with sardine in subarea 7 as case study. By default, mean catches from the latest two years are used to initialise the rule, which implied very low catch options as result of abnormally low harvest rates and landings in these years. It was suggested that expert knowledge of the stock and the fishery should inform decisions about initial application of the rule.

Overall, the work of WKDLSSLS is considered unfinished. Further research on the definition of optimal harvest control rules for data-limited short-lived stocks is ongoing. Therefore, the suggested either tuned constant harvest rate or the trend rule (1-over-2 with symmetrical 80% Ucap and biomass safeguard) should be taken as an interim (provisional) proposal while guidelines are refined in 2022