Variabilidade espaço-temporal do NDVI gerado a partir do satélite terra/modis sobre o Nordeste brasileiro

Abstract

Land degradation, droughts and desertification on Brazilian Northeast Region (NEB) has been demonstrating consequences on the political, social and social spheres. Vegetation-based information derived from remote sensing has assisted the monitoring and history trend identification of degraded areas worldwide. In addition to tools for the acquisition and preprocessing of MODIS data, the present work perform an analysis on the NEB vegetation behaviour by using the NDVI and python programming language in the years 2000 to2017. Also, influencing factors concerning the NDVI reliability and its spatio-temporal patterns has been listed. Through time-series analysis, statistical measures were extracted aiming the identification of patterns and trends. The results shows vegetative grow during the years 2000-2011 with an consecutive abrupt negative trend on 2012 influenced by north Atlantic Ocean Sea Surface Temperature (SST). NDVI Spatial patterns shows similarity with mean precipitation and spatial delimitations of the biomes Caatinga and Mata Atlântica, as well as high altitude enclaves over the NEB semiarid portion. 16.3% of the NDVI data were found to be affected by the presence of clouds, spatially distributed on north of Maranhão state and NEB east coast. Specially on these regions it is recommended the utilisation of pixel reliability and quality bands. Future work might include investigation of defective data filtering techniques and spatial correlation analysis on NDVI and Precipitation data.