Análise cronosequencial de áreas degradadas em restauração utilizando diferentes índices de vegetação em imagens de satélite

Abstract

The devastation of natural areas for the removal of products that supply the human need makes the recovery of degraded areas necessary. As one of the steps to achieve a successful restoration, the monitoring of the development of these areas is a necessary practice, and, through Remote Sensing, it turns out to be more practical and feasible. In order to evaluate the use of this technology, the present work used three different vegetation indices (NDVI, SAVI and EVI) in satellite images to perform a chronosequential analysis of twelve parcels in recovery located at the UTFPR, Dois Vizinhos campus, which are divided into three different treatments (Passive Restoration, Nucleation and Reforestation with Native Species). For this, images from the Landsat 7 and 8 satellites from October 2010 to July 2020 were used (two images per year, one in summer and another in winter). These images were obtained with the atmospheric correction already performed and from that the vegetation indices were calculated. For the analysis of the results, the indices were compared in relation to the years, seasonality, treatments and also the relationship with the vegetation data previously obtained in the field, in order to be able to represent the evolution of restoration over the years using indices obtained by satellite imaging. Taking into account all the analyses made between the indices, only the NDVI showed a difference in relation to the EVI and SAVI, which consequently were very similar. The use of the indices to analyze the area showed to be competent to visualize the development of the vegetation and its relationship with the hydric factor in the area, but this did not show a good correlation with the density data obtained in the field. It is suggested to reproduce this work in larger areas or even use sensors with better resolutions.