Estimating multifleet catchability coefficients and natural mortality from fishery catch and effort data: comparison of Bayesian state–space and observation error models

Abstract

Catchability and natural mortality are key quantities in fisheries stock assessment. However, it is difficult to estimate these two parameters simultaneously using only fishery catch and effort data. A Bayesian state–space modified delay–difference model is outlined that can estimate time series of catchability by fleet as well as natural mortality. This model, and three variants thereof, is fitted to data for grooved tiger prawns ( Penaeus semisulcatus ) in Australia’s Northern Prawn Fishery during the period of the year when there is little recruitment. A model that allows for both observation and process error and estimates natural mortality is best, in terms of model selection criteria as well as fit diagnostics. The posterior median estimate for catchability for the primary target fleet ranges from 6.15 × 10−4to 1.09 × 10−4during 1980–2007, while the posterior median estimate for catchability for a fleet with P. semisulcatus as its byproduct is about 20% of that for the primary fleet. Fishing efficiency increased at approximately 2% annually during 1980–2007, while the weekly natural mortality is estimated to be 0.053 week–1.