| Estimating the range and extent of a species is a critical element of developing management and conservation plans. With the advent of large-scale species occurrence data bases and remote sensing technologies, the ability of researchers to predict the distribution of focal taxa has been greatly enhanced in recent years. Accounting and controlling for spatial autocorrelation is a commonly applied technique in determining the effect of large scale environmental factors in predicting the distribution of terrestrial organisms. For many fish species, however, their distribution is confined to a network of streams and rivers that cannot easily be modeled by traditional methods.I found that while non-spatial logistic regression models have environmental variables that are significantpredictors of Yellowstone cutthroat trout occurrence, suchmodels are the least accurateof the classes that I examined. Mystudy illustrates the importance of considering spatial autocorrelation in modeling species distributions patterns over large geographic areas. Keywords:Yellowstone cutthroat trout, stream modeling, geographic information systems,remote sensing, spatial autocorrelation |