Development of a regional stock–recruitment model for understanding factors affecting walleye recruitment in northern wisconsin lakes

We used data from 162 lakes in northern Wisconsin during 1990–1999 to develop a stock–recruitment model for walleye Stizostedion vitreum and to identify important factors affecting recruitment. We used the linear version of the Ricker stock–recruit model and regression tree analysis to model age-0 walleye density, which ranged from 0.24 to 505 per km of shoreline (mean = 51.2). Adult walleye density ranged from .040 to 5.9 per surface hectare (mean = 1.5). The density of adult (sexually mature) walleyes, calendar year, and angling harvest rate of yellow perch Perca flavescens were significant descriptors of age-0 walleye density. Adult walleye density explained 10% of the variation in age-0 walleye abundance, which exhibited significant compensatory density dependence. Calendar year and yellow perch density explained nearly 32% of the residual variation in age-0 walleye density. The average age-0 walleye density in high-recruitment years (1991, 1994, and 1995) was 74.2/km, whereas the average age-0 walleye density in low-recruitment years (1990, 1992, 1993, and 1996–1999) was 38.5/km. Yellow perch harvest rates influenced walleye recruitment in low-recruitment years, as lakes with angler harvest rates of yellow perch lower than 0.79/h had lower recruitment of age-0 walleyes; in high-recruitment years, by contrast, yellow perch density did not affect recruitment. Combining data from many walleye stocks within a region allowed us to determine regional factors affecting walleye recruitment. The combined data can be used to develop walleye harvest forecasting models.