This paper presents further progress towards an integrated stock assessment model for Antarctic toothfish (Dissostichus mawsoni) in the Amundsen Sea region, defined here as SSRUs 88.2C–H. The region is split into two areas: the North (SSRU 88.2H) comprising large mature fish, and the South (SSRUs 88.2C–G) comprising a mix of large mature fish and small immature fish.
Two-area stock assessment models were first developed for the region in 2014 and refined in 2015 and 2016. Results showed the need to collect mark-recapture data in the South to inform the estimation of biomass in the South. Simulation work undertaken in 2017 showed that if tag recaptures continued in the south, and were spread among research blocks, a model may be developed for management advice.
We present an update on these two-area stock models including data from four years of recapture data from the South. The assessment models were fit to the proportions-at-age in the catch, and the mark-recapture data from the two areas. The results suggest that data from the research plan are starting to inform the model, especially with respect to the size of the population in the south and migration rates between areas. Biomass estimates are in agreement with those calculated using local Chapman and CPUE by seabed area analogy methods. At this stage the model should only be used as indicative due to issues including poor fit to the age data in the south, the lack of year-specific age frequency data to inform these fits, and the limited spatial extent of the tag recaptures in the South (almost all recaptures in research block 882_2).
We recommend that the analyses be developed to incorporate the spatial distribution of the tagged fish and subsequent fishing effort, as well as further investigations of the distribution of fish in relation to their length. We also recommend that all Members contribute validated age data from otoliths collected in Subarea 88.2 to further inform the understanding of the stock, and that the research data from vessels that have lower effective tag survival and detection be significantly improved to ensure the best use of the available information.