Report of the Benchmark Workshop on Redfish Stocks (WKREDFISH)
The ICES Benchmark Workshop on Redfish in Northeast Arctic waters (WKREDFISH) convened at two meetings, one data compilation workshop 21–23 November 2017 and the final benchmark meeting, 29 January – 2 February 2018. Both meetings took place at the ICES Headquarters in Copenhagen.
As part of the benchmark workshop a special request from Norway and Russia on HCR evaluation for redfish and by-catch limits of redfish in the Barents Sea shrimp fishery was addressed. WKREDFISH_2018 did not fully address the request but proposes a road map for setting bycatch limits for redfish. The HCR part of the request will be dealt with by a separate workshop before autumn 2018.
In WKREDFISH_2018 two stocks were benchmarked: Beaked redfish (Sebastes mentella) in subareas 1 and 2 (reb.27.1-2) and golden redfish (Sebastes norvegicus) in subareas 1 and 2 (reg.27.1-2). The most important conclusions for each stock were:
Beaked redfish in 1 and 2
No new information on stock id or sub-stock structure was presented during the benchmark. The assessment of beaked redfish has changed considerably since the last benchmark in 2012. The main points are that the model is now run in Template Model Builder (TMB) whereas in 2012 the implementation of the model was in ADMB. Autoregressive models are now implemented for recruitment and the annual component of fishing mortalities and selectivity in the demersal and pelagic fleets. Additionally, a right trapezoid population matrix is used, and older ages being aggregated into age blocks. Finally, data from the pelagic surveys in the Norwegian Sea were included in the assessment. These changes are considered an improvement and accepted by WKREDFISH_2018.
WKREDFISH_2018 proposes the following biomass reference points:
- Blim as 324 000 tonnes. The stock-recruit scatterplot shows no clear evidence of relation between recruitment and spawning stock biomass but rather trends in recruitment and SSB over time. This can be classified as a type 1 stock and Blim thus be approximated by the lowest observed SSB: 324 000 t. Result from the Schaefer biomass model for S. mentella in subareas 1 and 2, updated at the AFWG (2016) from previous benchmark assessment for the period 1952–2015 show no lower SSB prior to the assessment time-series.
- Bpa at 450 000 tonnes. Bpa can then be estimated from Blim taking into account the uncertainty in estimating SSB in the most recent year: 𝐵𝑝𝑎=𝐵𝑙𝑖𝑚*𝑒(1.645×𝜎), with 𝜎=0.2 (Note that the model output for SSB in 2016 gives 𝜎=0.09 but this value is considered an under-estimate which doesn’t account for sources of uncertainties external to the model, such as uncertainties in M or q).
As for fishing mortality reference points WKREDFISH_2018 proposes F0.1 = 0.080 defined for ages 19 and older. As there is a request from Norway and Russia to evaluate harvest control rules for beaked redfish where FMGT and other reference points will be evaluated, WKRED did not propose candidate Flim and Fpa points.
Golden redfish in 1 and 2
No new information on stock id or sub-stock structure was presented during the benchmark. The assessment of golden redfish has changed considerably since the last benchmark in 2012.
The assessment model proposed by WKREDFISH_2018 is a modification of the existing Gadget model used for this stock.
The model is a single-stock, single-area model with an annual timestep, length range 1–80 cm+ in 1 cm size categories, age range 3–30+, which runs from 1986 to 2017. Two surveys are used for tuning: the winter survey (BS-NoRu-Q1) and the coastal survey (NOcoast-Aco-Q4). Neither of these gives good coverage of the larger fish, and thus the larger fish are only covered by the commercial catches. The modelled fleets are gilfleet, trawl (including the very minor handline catches) and longline. However, for practical reasons, the trawl and longline fleets are combined into a single fleet (trawl and longline) prior to 2009. Annual catches are considered exact, and each fleet has length distributions (in 1 cm categories from 1986) and age length data (from 2005, 5 cm length categories). Natural mortality is fixed at M=0.05. All fleets are modelled with asymmetric dome-shaped selectivity (in length), except for the longline fleet which has logistic selectivity. The main improvement in the model since WKRED-2012 is the use of annual, rather than quarterly, time-steps in the model and the splitting of trawl and longline catches. The model was considered suitable as basis for advice and was ac-cepted by WKREDFISH_2018.
WKREDFISH_2018 proposed the following biomass reference points:
No stock recruitment relationship is presented for this stock. Within the model, re-cruitment is modelled as an annual recruitment value with no relationship to the SSB.
- Blim: Blim is based on the Lowest Observed Stock Size at which reasonable re-cruitment was observed. This is assumed to be the 2003 year class, at which time the SSB is estimated to be 44 000 tonnes.
- Bpa: Using the ICES default multiplier of 1.4 for Bpa gives a Bpa value of 61 600 tonnes.
The stock is currently well below the biomass limit reference point, and thus FMSY is not recommended as the current fishing level. However, it was considered useful to try and estimate a candidate FMSY reference point, which can be used to compare against management performance. Using yield per recruit analysis WKREDFISH_2018 pro-poses F0.1(15+), estimated to be 0.0525, as a candidate FMSY.
Future research and data requirements were identified, also by the external reviewers.
Published under the auspices of the following ICES Steering Group or Committee