Seasonality of Vegetation Indices in different land uses in the São Francisco Valley

Abstract

Remote Sensing has been used in researches with higher frequency to allow the analysis of the surface without the contact with the targets, allowing to identify biophysical conditions of the vegetation and understanding its photosynthetic dynamics. The objective of this study was to analyze the seasonal behavior of NDVI and EVI2 vegetation indices with and without transforming wavelets in areas with different soil uses in the São Francisco Valley and to verify their relationship with the occurrence of rainfall. The study was carried out in areas of preserved and degraded caatinga, and degraded pasture in the municipality of Petrolina-PE. NDVI was determined on 20 OLI sensor images with dates representing rainy and dry periods between 2013 and 2016. EVI2 data with and without filter were obtained for MODIS pixels. Precipitation data were used to understand the relationship between vegetation indexes and rainfall occurrence. The results showed that the studied areas presented similar behavior among vegetation indexes, but with different intensities, presenting higher values in January 2014, April 2015 and March 2016, coinciding with the occurrence of higher precipitation records in the region. The relationship between vegetation indexes and precipitation was linear. Therefore, it was possible to determine the seasonality of the studied areas, making it possible to understand that through the OLI and MODIS data, it was possible to analyze the dynamics of the vegetation and its relation to the rainfall.