Estimating fish mortality rates from catch curves: A plea for the abandonment of Ricker (1975)'s linear regression method.

This article explores the estimation of fish mortality rates using catch curves and critiques the linear regression method proposed by Ricker (1975). The authors argue that this method is biased and advocate for the use of more recent and accurate methods with age-frequency datasets from various sources. The article emphasizes the importance of appropriate error structures, not transforming count data, and not omitting zero counts or truncating low counts at old age. The authors conclude that log-linear models should be used instead of linear models for catch-curve analyses to obtain more accurate mortality rate estimates. The text also discusses different methods for estimating mortality rates in fisheries research, including the corrected regression coefficient bias (CRCB), Pella-Massey (PM), and generalized beta (GB) methods. The authors recommend abandoning the unweighted linear regression method in favor of these alternative methods and suggest further research to assess the accuracy of mortality estimates obtained from catch-curve analyses. [Extracted from the article]