Laserfiche WebLink
ISMP backwater sampling evaluation <br />The fit of individual general linear models for red and sand shiner, and fathead minnow <br />abundance as a function of habitat variables and predator abundance were very low (R' = 0.18 to <br />0.29) and the overall models were not statistically significant (p-values = 0.13 to 0.48). A GLM <br />fitted to the pooled cyprinid abundance data (R2 = 0.17; overall model fit, p = 0. 006), in the <br />absence of the non-significant predator abundance covariate, yielded limited inferences about the <br />relative importance each habitat variable to predict cyprinid abundance. Cyprinid abundance was <br />positively associated with average depth and relatively simple habitat, and negatively, but weakly <br />associated with surface area' and maximum depth. The weak negative relationship of fish <br />density to surface area' suggested that smaller backwaters tended to support higher densities of <br />cyprinids. Similar to the centrarchid GLM models, the negative coefficient for the year] 997 <br />variable in the cyprinid GLM reflected the lower numbers of those species captured in that year. <br />Species richness relationships <br />The 1997 ISMP sampling detected 14 species in the study area while DMR sampling <br />detected 20 (Table 3). In 1998, ISMP sampling detected 14 species in the study area while <br />DMR sampling detected 17. On average, ISMP sampling detected only 66 % of the species that <br />occurred in individual backwaters relative to those detected by DMR sampling (Table 7). As <br />expected, cryptic (e.g., ictalurids) or rare taxa were the ones most frequently overlooked by <br />ISMP sampling. <br />Pooled ISMP and DMR sampling results for individual backwaters suggested that an <br />average of 10 fish species (range 5 to 16) occurred in each backwater in the study area during <br />1997 and 1998 (Table 7). The best fit regression model of species richness as a function of <br />-19-