9360 - Laserfiche WebLink

Home Browse Search

ISMP backwater sampling evaluation with backwaters that had relatively simple habitat (e.g., little cover). The best fit logistic regression model for largemouth bass suggested that average depth and surface area were positively associated with presence of largemouth bass (Table 5). Inexplicably, maximum depth was negatively, albeit weakly, associated with presence of largemouth bass. The best fit logistic regression model for green sunfish suggested that average depth and surface area were positively associated with presence of green sunfish. Presence of green sunfish was negatively associated with relatively simple habitat. A candidate model with a similar Akaike's Information Criterion (AIC) score retained only average depth and surface area as covariates. A similar analysis was not conducted with the three non-native cyprinid species because they were found in nearly every backwater in the study area. The general linear model (GLM) analysis suggested that largemouth bass density was positively correlated with average backwater depth and surface areal, but negatively associated with backwater maximum depth and relatively simple habitat (Table 6). The negative coefficient for the year1997 variable reflected the lower numbers of largemouth bass captured in that year. The GLM for green sunfish produced essentially the same results; density was positively correlated with average backwater depth and surface area', but negatively associated with backwater maximum depth and relatively simple habitat. The negative coefficient for the year1997 variable reflected the lower numbers of green sunfish captured in that year. Recall that because fish abundance data are divided by backwater surface area to yield fish density, use of surface area as a independent covariate would confound this analysis. Therefore, any positive or negative relationship of backwater size to fish abundance is reflected in the surface area' term. -18-