Modelos de distribuição geográfi ca de Amaioua guianensis Aubl. em Minas Gerais, Brasil

Abstract

The objective of this work is to evaluate the performance of 4 models using different environmental datasets in an attempt to predict the geographic distribution of Amaioua guianensis Aubl. in the State of Minas Gerais. The specie occurrence points was recovered from the TreeAtlan 1.0 database for Minas Gerais. The environmental datasets used for the work involved climatic coverings related to temperature and precipitation, elevation, distance from the ocean, MODIS vegetation indices, soil classes and litology. The Maximum Entropy method (Maxent) was used for modeling specie distribution. Four alternative environmental datasets was input to Maxent: (a) all variables, bioclimatic variables, variables selected by CCA and variables selected using the Jackknife test applied to all variables. The performance of the models was analyzed using the area under ROC curve and extrinsic omission rates. The simulations show that the selection of environmental variable produced the most accurate results.