Further development of pairwise tag detection performance index and its application to the stock assessment of toothfish in the Ross Sea fishery

Tag release and recapture data are used in integrated age-structured stock assessments of Antarctic toothfish (Dissostichus mawsoni) in Subareas 88.1 and 88.2 to determine abundance and sustainable yields. The assessment model assumes that all vessels have equal tag detection rates; as a consequence including observations from vessels with low detection rates (and/or low tagging survival rates) could lead to an over-estimate of the stock biomass.



In this paper we develop further an index of vessel-specific tag detection performance for the Ross Sea fishery using a case-control methodology which controls for the inter-annual spatial and temporal variability of commercial fishing operations from which tags are released and recaptured. We then develop selection criteria, which can be used to determine the subset of vessels for which  there is confidence in their tag-recapture data. Finally, we apply these selection criteria to the tag data available at the time of the 2011 stock assessment to determine the effect of this selection method on the tag data available and also illustrate its effect on the results of that assessment.



For each vessel in the fishery, every fishing haul from the nominated ‘case’ vessel was matched to one or more ‘control’ hauls from other vessels fishing in the same time and location (i.e. in the same fishing season and within a specified distance).  The index was calculated as the scaled number of tags recaptured by the case hauls relative to the number of tags recaptured by the matched control hauls. By iterating over all events for all vessels, we generated a relative index of tag detection rate for each vessel, with a value of one representing the average performance across all vessels.



The tag data from vessels for which the confidence interval was greater than zero and extended at or above one were selected for inclusion in the stock assessment. When applying this decision rule, 75% of all tags released and 83% of tags recaptured were selected for inclusion in the analysis. In comparison, when the 2011 stock assessment was carried out using a different data selection algorithm, more than 90% of all tags released and recaptured were selected for inclusion.



We recommend this approach be used as a data selection method for the stock assessment of toothfish in the Ross Sea region. The application of such indices could be extended to other regions where tag data is used in stock assessment, and developed for other data types such as catch per unit effort when used to inform stock assessment.