Evaluation and Comparison of a Habitat Suitability Model for Postdrift Larval Lake Sturgeon in the St. Clair and Detroit Rivers

We evaluated composition and spatial distribution of riverine nursery habitat for larval Lake Sturgeon Acipenser fulvescens in the Middle Channel of the St. Clair River, Michigan, and Fighting Island Channel of the Detroit River, Ontario, using a habitat suitability model (HSM) and fish collections. Although model outputs indicated similar portions of high-quality habitat in the Middle Channel (16.9%) and Fighting Island Channel (15.7%), larval abundance and dispersal patterns varied between these systems. Analysis with Akaike's information criterion indicated that a regression model using sand-silt substrate performed best at predicting the observed water-volume-standardized CPUE (number of larvaeh(-1)m(-3)) in the Middle Channel. Of 93 larvae that were collected in the Middle Channel, most were found to cluster at three distinct areas of high- and moderate-quality habitat, which was composed predominately of sand-silt substrate. Lengths of larvae varied by as much as 9mm, and the degree of yolk sac absorption also varied, indicating that larvae in the Middle Channel remained within the channel after a short drift downstream. Of the 25 larvae that were collected in Fighting Island Channel, distribution was sporadic, and occurrence did not significantly correlate with measured habitat variables. Larvae were relatively homogeneous in size and yolk sac stage, indicating that newly emerged larvae did not utilize available habitat in Fighting Island Channel but instead drifted into the main channel of the Detroit River. Dispersal patterns indicate variability in young Lake Sturgeon ecology, which is dependent on local habitat conditionsmost notably, substrate composition. Furthermore, modeled larval-habitat associations found in this study were compared to a similar study on larval Lake Sturgeon from the North Channel of the St. Clair River. Model outputs from all three systems accurately accounted for observed larval dispersal patterns among both rivers. This supports the transferability of an HSM parameterized for Lake Sturgeon from individual river reaches within two large river systems.