Estimating postspawning movement of walleyes among interconnected lakes of northern Wisconsin

Many species of fish make seasonal movements among geographic areas. Management of such mobile populations is often hindered by lack of knowledge of the patterns and rates of movement among the areas. In northern Wisconsin, spring population estimates (PEs) of walleyes Stizostedion vitreum are used in managing the joint tribal-sport angler fishery. Many lakes in this region are interconnected by streams or waterways that allow movement of walleyes among lakes. Because the movement of fish among lakes complicates the interpretation and use of PEs, we estimated postspawning movement rates of walleyes among interconnected lakes in four different lake chains (of two, two, four, and five lakes). Walleyes were captured and released on three occasions in each chain, with marks to indicate in which lake within the chain they were originally captured. We used Akaike's information criterion to select among four stratified capture-recapture models, with or without mortality and with constant or variable capture rates. We assumed that recaptures followed a Poisson distribution. In all lake chains, a model with mortality and constant capture rates fit approximately as well as one with no mortality and variable capture rates. These two models generated similar estimates of movement rates for surviving fish. Walleyes moved among lakes in all chains as they left spawning areas and apparently moved to feeding areas. In some lakes, model estimates indicated that at least half of all walleyes present at spawning might leave the lake within 1 week. Results for these chains suggested that although spring PEs are appropriate for setting quotas for harvest activities done at or just after spawning, such as spearing, individual lake PEs may be misleading for managing activities that occur throughout the year, such as sport angling.