Estimation of gillnet efficiency and selectivity across multiple sampling units: A hierarchical Bayesian analysis using mark-recapture data

Fisheries research often involves a repetitive sampling protocol for multiple ecological units (our example is gillnetting of lakes). Estimates of abundance based on catch per unit effort can be erroneous due to lake effects on sampling efficiency. Conversely dividing data into individual lakes may lead to poor inference due to sparse data. Hierarchical Bayesian analysis compromises between these two extreme methods by estimating parameters for an individual lake, but borrowing information from other lakes. We estimated size-selective gillnet efficiency with mark-recapture data across a series of lakes subject to a constant netting effort. Hierarchical Bayesian analysis was able to prevent unrealistic selectivity functions that arose from individual lake analysis. Furthermore, the hierarchical approach was able to derive accurate parameter estimates with very few mark-recaptures in sub-sampled data trials. This paper demonstrates the hierarchical methodology for the estimation of fishery selectivity parameters. The results could be used to derive informative priors for future research which uses the proposed gillnet protocol.