Fish Species of Greatest Conservation Need in Wadeable Iowa Streams: Relative Importance of Spatial Scales

We used fish and habitat data collected from 84 wadeable stream reaches in the Mississippi River drainage of Iowa to predict the occurrences of seven fish species of greatest conservation need and identify the relative importance of habitat variables measured at a small scale (e.g., depth, velocity) and a large scale (e.g., stream order, elevation).  Multiple logistic regression analyses were used to predict fish species occurrences using all possible combinations of five large-scale variables, thirteen small-scale variables, and all eighteen large- and small-scale variables.  We used Akaike information criterion to weight model parameters and calculate model-averaged predictions.  On average, correct classification rates (CCR = 0.80) and Cohen’s kappa values (κ = 0.59) were greatest for multiple-scale models, intermediate for small-scale models (CCR = 0.75, κ = 0.49), and lowest for large-scale models (CCR = 0.73, κ = 0.44).  However, the best performing model varied by species, and each species was found to be associated with a unique combination of large- and small-scale variables.  Our results support the necessity of understanding the factors that constrain the distribution of fishes across spatial scales to ensure that management actions can be made at the most appropriate scale.