Examination of the SBT operating model to inform conditioning and projection specifications

Abstract

Stock assessments and constant catch projections have been conducted using the Operating Model (OM) specified at the CCSBT technical meeting held in Seattle in July 2009. The current analysis shows that: (1) for the base case, higher steepness and lower M10 (natural mortality at age 10) are preferred in the grid sampling based on the likelihood (in contrast to the prior-based weight for steepness), and this leads to more optimistic future projections despite lower current spawning biomass (3.7% of the unfished biomass for the likelihood-based and 4.9% for the prior-based approaches), (2) when incomplete mixing of fish tagged is taken into consideration, the model fit, particularly to tag recaptures, is improved, and projection results are somewhat more optimistic, (3) the low recruitment estimate in 2006 seems to be primarily a consequence of LL1 catch-at-size data from 2008, (4) when CPUE is not adjusted for overcatch (i.e., S = 0), lower steepness and higher M are preferred, and higher S scenarios generally lead to more optimistic projection results despite worse fit to observed CPUE series, and (5) several sensitivity trials which accord less reliability to the Japanese longline CPUE favour higher M values and lead to more pessimistic results.