Estimation of tag reporting rates in agestructured multicomponent fisheries where one component has observers

Abstract

For many tagging experiments, it is vital that fishers find and report all tags to scientists. If not, the tag reporting rate needs to be estimated so that fishing and natural mortality rates can be estimated. One way to estimate this rate is to have one fishery component (e.g., with observers) report every tag found from all fish that it catches. If the numbers of fish caught by all fishery components are also known and the tagged fish are mixed with the population (or subpopulation) being harvested, then one can estimate the reporting rate of underreporting fishery components. This procedure can fail if data are pooled over ages. We obtain maximum likelihood estimators for the reporting rate for each age and (or) each fishery component. We show how to estimate reporting rates if mixing of tagged and untagged fish occurs with some delay. We also obtain overall age-dependent reporting rates, which combine reporting rates from all components of the fishery. Our likelihood is part of an integrated likelihood that allows estimation of age-dependent fishing and natural mortalities in addition to the reporting rates. Our procedures are illustrated with some southern bluefin tuna (Thunnus maccoyii) tagging data.