Evaluating single- and multi-species procedures to estimate time-varying catchability functional parameters

Time-varying catchability is prevalent in many management regions, and can be modeled using density dependence and time trend parameters. A catchability model can be estimated using single- or multi-species data, and parameter estimates might be of interest as an input for stock assessment models or in their own right (e.g., catchability trends as an index of technology improvements). An operating model was developed to replicate the catch-at-age, fishery-independent survey, and catch-per-unit-effort data from the Gulf of Mexico. Ordinary least squares was used to estimate catchability trends, density dependence, and annual catchability using 10 different estimation procedures. Procedures included an “imputation” strategy, where data from similar species are used to estimate catchability parameters for a focal species. Estimated trend, density dependence, and annual catchability were compared with index-specific operating model values to determine the precision and accuracy of different estimation procedures in a factorial model design. Multi-species procedures increased precision and accuracy of parameter estimates when compared with single-species procedures, and minimized errors in annual catchability estimates when compared with the assumption of constant catchability. Multi-species procedures also did not introduce large errors in functional parameter or annual catchability estimates, even when density dependence and trend were absent or when between-species variability was high. Procedures that imputed catchability functional parameters from similar species were precise and median unbiased given the quantity and quality of data that are available for the Gulf of Mexico.