Predicting the Occurrence of Stunted Bluegill Populations from Wisconsin Lake Features

Abstract

Bluegill stunting (poor growth and small size) is a fishery management problem in Wisconsin. Wisconsin has over 15,000 lakes and surveying each lake is not feasible due to lack of resources (personnel, money and time). Classifying lakes based on ecological and limnological similarities may provide a way to account for differences among lakes without having to survey all lakes. My objective was to classify stunted and non-stunted bluegill populations using features of Wisconsin lakes. Before I addressed my main objective, two subordinate objectives were addressed to establish data needs and define a stunted bluegill population for Wisconsin lakes. First, I determined if size selectivity of bluegills differed between electrofishing and Fyke netting in Wisconsin lakes. Second, I determined if size structure was related to body condition and growth of bluegill populations in Wisconsin lakes. Proportional stock density (PSD) estimated from the two primary gear types did not significantly differ, allowing me to combine data for future analyses. Mean length at age-4 was positively related to PSD, however the relationship was noisy. Bluegill relative weight (Wr) was not significantly related to PSD, suggesting, one index by itself may not provide adequate understanding of the dynamics of a bluegill population. Stunted bluegill populations were defined as having a PSD < 20 and a mean length at age-4 < 5 inches. I used linear discriminant analysis (LDA) to classify stunted and non-stunted bluegill populations based on lake features. Overall, the linear discriminant function (LDF) was 82% accurate in model creation and 85% accurate with validation. Stunted bluegill populations were predicted with 77% accuracy in model creation and 79% accuracy with validation. Non-stunted bluegill populations were predicted with 85% accuracy in model creation and 90% accuracy with model validation. My model using easy to measure lake features can be used to classify stunted and non-stunted bluegill populations in Wisconsin lakes and allow managers to set broad-scale regulations to optimize angling opportunities for bluegills.