The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume XL-7/W3
https://doi.org/10.5194/isprsarchives-XL-7-W3-329-2015
https://doi.org/10.5194/isprsarchives-XL-7-W3-329-2015
29 Apr 2015
 | 29 Apr 2015

Land-use and land-cover dynamics monitored by NDVI multitemporal analysis in a selected southern Amazonian area (Brazil) for the last three decades

D. Borini Alves, F. Pérez-Cabello, and M. Rodrigues Mimbrero

Keywords: NDVI, multitemporal dynamics, Landsat series, fire, deforestation, Brazilian Amazon

Abstract. This study aims to analyse the dynamics of land-use and land-cover (LULC) in a selected southern Amazonian area (Brazil), monitoring and distinguishing trajectories in NDVI (Normalized Difference Vegetation Index) variations for the last three decades. The area, with a total of 17336 km², has been subject to significant LULC changes associated with deforestation progress and use of fire. Considering available Landsat time series, it was selected an image per year from 1984 to 2013 (path/row -231/66), at a particular period of year, atmospherically corrected using LEDAPS tools. NDVIs values were generated for each selected image. Furthermore, the images of 1984 and 2010 still underwent a classification of LULC differentiate five categories: water, forest, secondary/degraded forest, savannah/pasture and crop/bare soil. The trajectories in NDVI variation values were analysed by R software, considering intersections of classified categories. The pixels identified as forests on the images of 1984 and 2010 displayed stable trajectories of NDVI values, with average value 0.824 and coefficient of variation 3.9%. While the pixels of savannah/pasture, which was periodically affected by fire, had an average NDVI value 0.585 and coefficient of variation 15,1%. The main regressive trajectory was the transition “forest to crop/bare soil", identifying 1999 as the starting point in the drop in NDVI values, associated with an increase of the deforested areas. Therefore, the results show distinct trajectories associated with NDVIs and LULC changes that assist in better understanding the dynamics of ecological processes and the human impacts operating in the area.