59588

Dinâmica da regeneração florestal em ambiente de floresta Atlântica e sua modelagem por redes neurais

Favoritar este trabalho

The brazilian Atlantic rainforest is the third largest biome among those that cover the country. Given its characteristics of high anthropic pressure and high endemism of vegetal and animal species, the biome was classified as a mundial hotspot. Thus, actions for conservation and restauration of its forests have been proposed. Between these actions, there is the forest natural regeneration. Given the current socioeconomics aspects of the rural population, many pastures are being abandoned, which allows the natural regeneration’s establishment. The objectives of this work were to analyze the landscape dynamics through orbital images in an area of Atlantic forest and to predict the natural regeneration through neural network modeling. Images from the TM/Landsat-5 and OLI/Landsat-8 sensors were acquired and the visual interpretation allowed the thematic extraction of classes for the analysis of the dynamics of the forest natural regeneration. It was observed that the forest regeneration were mainly found in South facing aspects, because they have an more suitable envinronment for the establishment of the secondary succession. The results for the network modeling weren’t satisfactory, where only 32% of the regeneration were correctly predict.