Tratamento de imagens de satélite para geração de indicadores da desertificação

Abstract

For the study of desertification, several factors have to be considered, both in relation to physical aspects, as well as to social, economic, infrastructure, etc., all of them through a multitemporal analysis. Thus, this is a phenomenon to which the use of geoprocessing and remote sensing techniques is applied with great property, in the sense of identifying the areas that are being affected, and which factors are exerting the greatest contribution to this. The study area is located in the State of Paraíba in the Borborema mesoregion, located in the sub-basin of Alto Paraíba in the Paraíba region. This area covers the fol-lowing municipalities: Amparo, Barra de São Miguel, Boqueirão, Camalaú, Caraúbas, Congo, Coxixola, Monteiro, Ouro Velho, Prata, São Domingos do Cariri, São João do Cariri, São João do Tigre, São Sebastião de Umbuzeiro, Serra Branca,Sumé and Zabe-lê. Theresearch is characterized as a field study and presents some factors related main-ly to the surveys, to develop in the place where the phenomena occur, their results are usually more reliable (GIL, 2010). The objective of this work wasto create a databaseof the study areas that can be used to monitor the desertification process; as well as us-ing geoprocessing and remote sensing techniques to evaluate changes in vegetation cov-er and land use, and generate indicators related to these factors; evaluate the contribu-tion of the remote sensing products used; compare the results obtained with socioeco-nomic indicators generated in another work plan related to the same study. The method is characterized as field research, based on the use of digital orbital image processing techniques(TM-Landsat-5) in SPRING software, with the objective of generating vege-tation index images (Normalized DifferenceVegetation Index -IVDN and in English NDVI), which represents a good indicator of the situation ofvegetation cover, reflecting the ways inwhich the population is intervening in the environment through economic activities. Through these techniques it was possible to perform a multitemporal analysis of the vegetation cover situation, considering the use of orbital images from the years 1989, 2005 and 2010. The results showed that, in general, there was a loss of plant cov-er, a fact that may indicate thatthe desertification process is occurring, and that is wor-rying. In the three dates studied (1989, 2005 and 2010) there was a significant decrease of the closed shrub-tree vegetation. It was considered that the beginning of an extended period of rainfall did not affect the generation of the results, because the phenomenon was in its initial phase.