Gillnet catchability of Walleye (Sander vitreus): comparison of North American and provincial standards

Abstract Multi-mesh index gillnet surveys are widely employed for the assessment of fish populations, including Walleye (Sander vitreus), an important commercial and recreational species in North America. In this paper, we estimate gillnet catchability of Walleye for two standard sampling protocols: the North American (NA) and the Fall Walleye Index Netting (FWIN), allowing the conversion of Catch Per Unit Effort (CPUE) into absolute estimates of fish density or abundance. We combined mark-recapture studies from 14 lakes in Ontario and Quebec, 11 of which were concurrently sampled using NA and FWIN methods, totalling 53 recapture surveys from 1996 to 2012. To account for lake-specific influences on catchability, we analyzed the data with Generalized Linear Mixed Models (GLMMs) using lake as a random factor and a model-averaging approach to calculate mean estimates and variances. Catchability for Walleye ≥ 350mm total length was 1.04 ha/gang for the FWIN, meaning each fish caught in a FWIN net gang (1.8 x 61 m) represents roughly one fish per hectare. Catchability for one gang of NA net (1.8 x 24.8 m) was 0.52 ha/gang, which is 20% higher than FWIN catchability after accounting for net area. Two important sources of variation in catchability were: (i) lake-specific differences and (ii) seasonal changes within the NA protocol. The mechanisms underlying such variation are still not clear and will require additional studies covering a broader geographic range to reveal potential environment-catchability relationships. We provide recommendations for estimating Walleye density and confidence intervals, and for how to combine our catchability estimates with size-selectivity curves to predict density at any specific size interval.