Petersen A.T., Ball, L.G., and Cohoon K.P., 2002. Predicitng distributions of Mexican birds using ecological niche modeling methods. Ibis. 144:27-32.

            An analysis of the ability of the Genetic Algorithm for rule-set Production (GARP) to model ecological niches of 25 Mexican bird species is presented by the authors. To produce the models, they obtained bird data (points of occurrence) from 32 scientific collections, and geographical data from 7 km resolution raster images. For the model testing, Petersen et. al., withheld the data of two States of Mexico (Oaxaca and Jalisco), from the model building process; then, in a GIS (ArcView 3.1) they overlaid the withheld points on the predicted areas. They found that 78-90% of the distributional models performed better than random models in the predicting states. Even though this represents a good performance, the authors expect that using higher resolution data, the accuracy of GARP prediction will be even improved.