Abstract
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.
Highlights
In recent decades, the southern Brazilian Amazon experienced increases in the dynamics of land use and land cover (LULC) associated with the expansion of crop and pasture areas, road and communication network construction, and population density growth (Espindola et al, 2012; Morton et al, 2006; Nepstad et al, 2001)
To monitor and better understand the dynamics of ecological processes and human impacts related to these changes in LULC, remote sensing data represents an essential source of analysis, allowing to generate systematic information in most different spatial and temporal scales (Nagendra et al, 2013)
We identify as stable trajectories of NDVI those pixels where no LULC was detected: 'Forest to Forest' (FF); 'Sec./degraded forest to Sec./degraded forest' (Fs-Fs); 'Savannah/pasture to Savannah/pasture' (SP-SP); and 'Crop/bare soil to Crop/bare soil' (CB-CB)
Summary
The southern Brazilian Amazon experienced increases in the dynamics of land use and land cover (LULC) associated with the expansion of crop and pasture areas, road and communication network construction, and population density growth (Espindola et al, 2012; Morton et al, 2006; Nepstad et al, 2001). This dynamics results in pressure on forest resources in the area due to the advance of deforestation, degradation and intensive use of fire (Aragão et al, 2008; Lima et al, 2012; Vasconcelos et al, 2013). NDVI relates spectral information of the red and near infrared generating a variable able to estimate quantity, quality and development of vegetation (Pettorelli et al, 2005)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
