Predicting Walleye (Stizostedion vitreum) Presence, Reproduction, and Abundance in Northern Wisconsin Lakes

Abstract

Walleye (Stizostedion vitreum) spearing and angling harvest levels are set annually for lakes located in the ceded territory of northern Wisconsin, based on adult population estimates. For most lakes, population estimates are obtained from linear regression models that describe the relationship between past population estimates and lake size. Lakes are classified according to the recruitment source (i.e. stocked or natural) of the population, and regression models differ accordingly. This thesis evaluates the current recruitment classification system, identifies features of lakes that predict both the presence of walleye and the presence of self-sustaining walleye populations, and improves precision and accuracy of current regression models. The relationship between lake size and walleye abundance was quantified for two life stages, age-0 and adult, in 172 northern Wisconsin lakes. Age-0 and adult walleye abundance was higher in lakes sustained by natural reproduction than in lakes sustained by stocking. This suggests that patterns in walleye abundance differ between lakes with self-sustained and stocked populations in northern Wisconsin lakes and that these patterns may be established during the first year of life, at least partially due to recruitment source. Limnological features of lakes described walleye presence and absence in 2037 lakes, and presence of self-sustaining and stocked walleye populations in 392 lakes in northern Wisconsin. Lake area, maximum depth, and water source were the best predictors of walleye presence in lakes. The presence of self-sustaining populations of walleye was best predicted using pH and watershed area. I conclude that physical, chemical, and biological features of lakes may be used to discriminate lakes where walleye presently exist from lakes that presently do not contain walleye, and lakes with self-sustaining from stocked walleye populations. The relationship between lake size and adult walleye abundance, estimated by spring mark-recapture surveys, was quantified for 166 lakes in northern Wisconsin, 1990-1997. Four models described differences in adult walleye abundance as a function of lake size for lakes with self-sustaining versus stocked populations and for drainage versus landlocked lakes. Models were improved by adding environmental variables, to further explain variation among lakes. I conclude that accuracy of estimates of adult walleye abundance derived from regression models based on recruitment source may be improved by incorporating limnological variables.