Predictive modeling distribution of pioneer species in the state of Minas Gerais, Brazil

Abstract

The objective of this work was to determine the potential distribution of 23 pioneer species in the state of Minas Gerais, Brazil, as well as to identify the environmental variables that influence their distributions. The Maxent algorithm was chosen to associate the occurrence of species with the following bioclimatic variables: diurnal temperature variation, isothermality, temperature seasonality, driest month precipitation, precipitation seasonality (coefficient of variation), and actual evapotranspiration. The normalized difference vegetation index (NDVI), flora conservation status, and the spatial heterogeneity of vegetation types were also evaluated, besides erodibility (susceptibility of soil to erosion), groundwater availability, soil texture, organic matter content, mineral occurrence (existing mineral species by lithological unit), pedological simplified map, slope and altitude. The species Anadenanthera colubrina was the most suitable for the Caatinga biome, followed by Casearia sylvestris and Plathymenia reticulate, indicated for the Atlantic Forest and the Cerrado biomes, respectively. The use of Maxent is recommended as a tool to guide conservation plans that require the indication of species, aiming to recover degraded or deforested vegetation areas.