Results of evaluating the performance of empirical estimators of natural mortality rate

Amy Y. Then, Virginia Institute of Marine Science
John M. Hoenig, Virginia Institute of Marine Science

Abstract

This document has been issued as VIMS Data Report 62 and provides additional results and analyses for Then et al. 2015. Evaluating the predictive performance of empirical estimators of natural mortality rate using information on over 200 fish species. ICES Journal of Marine Science 72(1): 82-92. Natural mortality rate, M, of fish is a highly influential stock assessment parameter. The M parameter is also difficult to estimate directly and reliably. Various empirical estimators have been developed to estimate M indirectly, based on relationships established between M and predictor variables such as growth parameters, lifespan and water temperature (e.g., Beverton and Holt, 1959; Alverson and Carney, 1975; Pauly, 1980; Hoenig, 1983). Despite the importance of these estimators, there is no consensus in the literature on how well they work in terms of prediction error or how their performance may be ranked. Then et al. (in press) evaluated estimators based on various combinations of maximum age (tmax), von Bertalanffy growth parameters (K) and asymptotic length (L∞), and water temperature (T), by seeing how well they reproduce independent, direct estimates of M for more than 200 unique fish species. They also considered the possibility of combining different estimators using a weighting scheme to improve estimation of M. This report documents additional analyses and results to supplement the results in the journal article. The estimators, evaluation criteria, and other important details are given in the journal article.