GIS-based habitat modeling using logistic multiple regression: a study of the Mt Graham red squirrel: J. M. C. Pereira & R. M. Itami, Photogrammetric Engineering & Remote Sensing, 57(11), 1991, pp 1475–1486

Abstract

Multivariate statistical techniques were applied to the development of habitat suitability models for the Mt. Graham red squirrel, an endangered species. A digital map database and a Geographic Information System (GIS) were used to support the analysis and provide inputs for two logistic multiple regression models. The models attempted to predict squirrel presence or absence, the dichotomous dependent variable. Independent variables were a set of envi- ronmental factors in the first model, and locational coordinates in the second case, where a logistic trend surface was developed. Bayesian statistics were then used to integrate the models into a combined outcome. Potential habitat losses resulting from development of an astronomical observatory were assessed using the environmental model and were found to represent about 3 percent of the currently available habitat.