Evaluating how precision in estimates of abundance

No abstracts are to be cited without prior reference to the author.

The rules for setting quotas for North-East Arctic cod (NEAc) are based on how estimates of stock parameters relate to defined biological reference points obtained from the Extended Survivors Analysis (XSA) model, calibrated using bottom trawl survey data from the Barents Sea. In this paper we use simulations to investigate how the precision in estimates of relevant stock parameters for NEAc relates to different levels of sampling effort in the trawl survey. We also evaluate the importance of estimates of abundance-indices by age as compared to estimates of catch-at-age for assessments and management advice. It is generally assumed that catch-at-age is known exactly and that uncertainty in estimates of abundance is chiefly caused by errors in the survey indices. However, catch-at-age is estimated, and subject to sampling errors that depend on the design and sampling effort in fisheries-dependent surveys. This must be taken into account when evaluating the performance of fisheries-independent surveys. The yearly winter survey used for tuning is expensive, has large area coverage, and samples from 176-394 trawl stations. It is therefore important to establish the required survey effort to achieve adequate precision in estimates of stock parameters. We explore whether the effective sample size for estimating simple statistics, such as the proportion of ages 7+, or mean age, can serve as a proxy. We adopt a statistical catch-at-age model in AD Model Builder which also allows errors in catch at age when evaluating effects on sampling strategies in the trawl survey on assessment and management advice.