Comparison of Catch-Curve Methods for Estimation of Mortality

Several methods are available to estimate mortality from catch curves. These methods include least squares, a maximum-likelihood method developed by Chapman and Robson, and a “simple” estimator proposed by Ssentongo and Larkin. All three of these methods are based on similar models for mortality, but data requirements are different, and they can give considerably different estimates of mortality. These three methods are compared by applying them to catch curves for natural populations, to catch curves generated by a deterministic model, and to catch curves generated by a stochastic model. In general, the least-squares estimator is the most precise under conditions encountered in assessment of fisheries, but if variation is unusually high, the maximum-likelihood estimator developed by Chapman and Robson is more precise. The “simple” estimator developed by Ssentongo and Larkin can result in serious error.