Smallmouth Bass Nest Site Selection and Success in Four Northern Wisconsin Lakes: Linking Habitat Selection with Habitat Quality

Abstract

Habitat models based on nest site selection and spawning success are important in understanding the effects of riparian area and littoral zone alterations on smallmouth bass (Micropterus dolomieu) survival and recruitment. A common approach to developing habitat models is the use of resource selection functions that determine the relative probability of use of particular subsets of available habitat. Application of these models presupposes that habitat selection infers habitat quality and study of both are critical to testing this assumption. The objectives of this study were to assess smallmouth bass nest site selection, to evaluate nest site quality, and to assess the relations between habitat selection and habitat quality. Smallmouth bass nest characteristics and egg survival were quantified on four lakes in the summers of 1997 to 2000. Habitat variables were quantified at each nest site and along random transects that were placed perpendicular to the shoreline to a depth of three meters, which corresponds to the maximum depth that nests are found in the study lakes. Eggs and pre swim-up fry were estimated at each nest site to quantify individual nest success. Resource selection functions (i.e., logistic regression) were used to elucidate habitat features used disproportionately as nest sites. Nest site quality was quantitatively evaluated using egg survival and fry production as dependent variables and nest site characteristics as independent variables. Linear regression was used to assess the relations between habitat selection (i.e., probability of selection) and habitat quality (i.e., survival and fitness). Habitat variables that predicted nest site selection as well as those that predicted nest site quality were variable across lakes and across years. The percent gravel in a nest and its proximity to cover (e.g., wood or rock) were the most common variables present in both habitat selection models and habitat quality models. Our results also show that habitat selection only inferred habitat quality when adult population densities were high (> 10 adults per hectare) and thus high quality habitat was saturated. These results suggest that models based on habitat selection may only reflect habitat quality under specific conditions where habitat is limited.