Evaluating confidence intervals for habitat suitability models

摘要: Implicit in the use of wildlife habitat models is assumption that different ratings reflect differences quality. Using a gray squirrel suitability index (HSI) model, we compared 6 habitats within Huron-Manistee National Forest, Michigan. We assessed variation final HSI scores from around model input variables using Monte Carlo simulation (parametric bootstrapping) and nonparametric bootstrapping. In our demonstration, as divergent 0.38 0.81 did not differ statistically. Habitat modeling should account for variability scores, well indicators animal response. Modeling attempts ignore this are unlikely to determine whether faulty relationships product or result variation.