Modelagem da distribuição das florestas atlânticas nebulares na Serra da Mantiqueira

Abstract

We aimed estimate the potential distribution of Atlantic cloud forests in the Serra da Mantiqueira, southeastern Brazil, by modelling the spatial distribution of cloud forests getting this closer as possible to the actual area. The predictive spatial distribution of cloud forests in the Serra da Mantiqueira was estimated using the Maxent algorithm. The distribution modelling was performed with a set of different environmental variables: standard bioclimatic variables, topographic variables and a cloud variable. After we estimated the potential cloud forest distribution area we used remote sensing techniques aiming approximate the actual cloud forest area and remove mistakenly parts that were modeled as cloud forest. We mapped four land use classes: cloud forest, planted forest, high altitude complex and others. The final mapped cloud forest areas were compared with a map of existing protected areas aiming analysis the protection cover of those areas. The results indicated an excellent prediction of cloud forest distribution showing close agreement with the bioclimatic data. Although the environments are suitable for the cloud forest presence using the predictive modelling method, the actual occurrence differs from this according to the mapped land use, being cloud forest extent half of the predicted modeled area. Most areas belong to the Protected Area with Sustainable Use of Natural Resources category. Due the high rate forest loss present in the Atlantic forest the cloud forests may disappear in a few years by a combination of human disturbance and climate change. The presently obtained results reinforce the need to assign a high conservation priority to the Serra da Mantiqueira creating protected areas with full protection status.