Indirect Estimates of Gill Net Selectivity for 18 North American Freshwater Fish Species

Indirect estimates of gill net selectivity were calculated for eighteen fish species sampled throughout South Dakota with North American standard (Standard) gill nets.   Monofilament Standard gill nets (25 m long x 1.83 m deep) include eight randomly ordered panels of 19, 25, 32, 38, 44, 51, 57, and 64 mm bar-mesh.  Five potential models were fit to empirical catch data using the SELECT method including normal, skew-normal, log-normal, bi-normal, and bi-lognormal models.  Models that included bi-modality produced the best fit for 14 species including Channel Catfish Ictalurus punctatus, Northern Pike Esox lucius, Walleye Sander vitreus, and Yellow Perch Perca flavescens.  Uni-modal models best described selectivity for Gizzard Shad Dorosoma cepedianum, Shorthead Redhorse Moxostoma macrolepidotum, and Shovelnose Sturgeon Scaphirhynchus platorynchus.  Selectivity models indicated that most species were captured by a combination of gilling or wedging and tangling, these capture modes explain the prevalence of bi-modal models.  Inspection of model deviances and deviance residuals suggest that models of best-fit provide useful estimations of gill net selectivity for these species.  Expanded use of North American standard gill nets and sampling methods allows these findings to be broadly applicable for understanding selectivity of North American standard gill nets for these commonly sampled freshwater fishes.