Migration has the potential to violate the assumptions of tag recapture models used in the assessment of toothfish stocks. Evaluating the potential bias introduced into stock assessments and tag-based biomass estimates when the distributions of tagged fish, fishing effort and the underlying stock distribution are spatially heterogeneous is currently one of the highest priorities of the Scientific Committee.
This study used a fisheries simulation model to evaluate the impacts of migration on a biomass estimates and sustainable catch limits from a closed population tag-based CASAL stock assessment. Our simulations demonstrated that emigration of tagged fish out of the assessed area results in initial and current spawning biomass, and stock status being over-estimated.
We then estimated migration of rates of Kerguelen Plateau Patagonian toothfish between the Divisions 58.5.1 and 58.5.2 using a catch-conditioned modification of the method used by Hilborn (1990). We considered longline tag releases from two periods, 2007-2014 and 2009-2014 and used tag recaptures to 2015 in both cases. The annual migration rate from Division 58.5.2 to 58.5.1 was estimated to be 1.1% using longline releases from 2007-2014 and 0.7% using longline releases from 2009-2014. The annual migration rate from Division 58.5.1 to 58.5.2 was estimated to be 0.4% and was insensitive to the time period of longline tag releases used.
For annual migration rates <2%, consistent with those we observed in Patagonian toothfish on the Kerguelen Plateau, bias in spawning biomass estimates from simulations was <5%. We demonstrated that the bias can be corrected by increasing the tag-shedding parameter in the stock assessment by the value of the migration rate, providing a simple yet effective approach to correct for the effects of emigration. The key benefit of making such an ad-hoc increase to the tag-shedding parameter is that it avoids the need to develop new assessment frameworks to account for emigration.
The impact of increasing the tag-shedding parameter in the 2015 assessment of Division 58.5.2 toothfish by 1% would have reduced the unfished spawning stock biomass by around 1.4% and the TAC by 3.8%.
We recommend that for tag-based assessments, the tag-shedding parameter should be increased by the estimated migration rate to adjust for migration related bias.