For the 2014 assessment of Patagonian toothfish (Dissostichus eleginoides) around the Heard Island and McDonald Islands (HIMI) in Division 58.5.2, a prior for catchability q of the random stratified trawl survey (RSTS) was estimated by comparing abundance estimates of the survey with abundance estimates calculated from the tag-recaptures data on the main trawl ground. Here, we use simulations to address potential bias in the estimation of q and conclude that the potential bias that may arise from fishing selectivity cannot be corrected for given the available data.
In the absence of an unbiased estimate of survey catchability, we recommend that a non-informative prior with uniform distribution be used for abundance estimates from trawl surveys in a toothfish stock assessment. The bounds of such as uniform prior should be wide, e.g. 0.1-1.5, since many processes act on catchability at the same time which collectively result in an estimate of catchability that can be well below or above 1.