Sensibilidade de índices de vegetação para classificação do estado de conservação da cobertura florestal

This study aimed to map the conservation state conditions of forest fragments through remote sensing data. For this purpose, NDVI (Normalized Difference Vegetation Index); EVI 2 (Enhanced Vegetation Index 2) and MVI (Moisture Vegetation Index) were tested in this study. Initially, we developed a method to infer, at the site, the state of forest disturbance through ocular reconnaiscence. 66 (sixty six) fragments were classified according to a representative type of their disturbance state conditions. These fragments were used as samples to extract the parameters related to the vegetation indices tested. The vegetation indices were sensitive to differences among vegetation types, but it was not possible to obtain a parameter that could accurately separate the samples as per the field classification. The best result was obtained using the median NDVI parameter which enabled the mapping of disturbance intensity for the forest remnants in the study area, with two classes: disturbed secondary forests, and lowly disturbed secondary forests. Results indicate that remote sensing is one of the possible ways to pass from a local scale to the landscape scale.Pages: 1652-165