Muskellunge Management: Fifty Years of Cooperation Among Anglers, Scientists, and Fisheries Biologists
A Comparison of Muskellunge Weight Estimation Equations to a Modified Length-Girth Technique
Jonathan R. Meerbeek and Derek P. Crane
Abstract.—Managers seldom sample sufficient numbers of Muskellunge Esox masquinongy to develop population specific weight–total length (TL) models and often rely on existing models to predict Muskellunge weight. However, substantial variation exists among these models and managers and anglers may be forced to report inaccurate Muskellunge weight estimates. We used TL, maximum girth, pectoral girth, sex, and adult female reproductive status data obtained from 869 anesthetized adult Muskellunge collected from three natural lakes in Iowa between 2012 and 2015 to identify the most important variables for estimating Muskellunge weight. We also compared the predictive performance of our model to five popular models for estimating Muskellunge weight using test data from Iowa Muskellunge and data from an outside population. Stepwise model selection identified a model including terms for TL, maximum girth, and pectoral girth as having the lowest Bayesian information criterion value; however, post hoc analyses of model accuracy found that a two-variable model including terms for TL and maximum girth was negligibly different from the three-variable model. Therefore, we selected the TL–maximum girth model as the best model and for use in additional analyses of model accuracy. Of the six models compared, our model and the Casselman and Crossman (1986) model performed well across the range of weights tested using the Iowa test data set; on average, estimated weights were within 2.7–5.5% of measured weights. All other models either grossly under- or overestimated Muskellunge weight. Model accuracy for all tested models decreased substantially when tested using data from the Georgian Bay Muskellunge population, especially for fish greater than 30 lb [13.6 kg]. Although our model and the Casselman and Crossman (1986) model will provide managers with weight estimates that are suitable for use in most fishery-management applications, they may not be suitable for anglers attempting to determine if angled Muskellunge are of record weight. In populations where extreme morphological differences may occur or robust weight–TL models need to be developed, we recommend that population specific models be developed to estimate Muskellunge weight.