Study of the NDVI with multi-scale and time-series analysis using SPOT imagery during the period 1998-2012 in Uruguay

Vegetation indices are a relevant source of information for spatial monitoring of vegetation at multiple scales. Among them, the Normalized Difference Vegetation Index (NDVI) is one of the most commonly used. This study aims to describe and analyze the spatial patterns of the NDVI in terrestrial systems in Uruguay at the onset of the 21st Century. A multiscalar approach (country, basin and sites) was applied using time series analysis of NDVI values obtained from SPOT 4 and 5 images through the program Instrument Vegetation (VGT). The analyzed time series showed a significant fit of the Autocorrelated Integrated Moving Averages (ARIMA) model, with an autocorrelation of order 2 and a level of integration of order 1, ARIMA (211). A significant decline of the NDVI over all spatial units was found, with agricultural units (site scale) showing the most negative slope. This study provides baseline data on changes in vegetation productivity for Uruguay, and develops an accurate and robust methodology for spatio-temporal analysis of NDVI series. Remote sensing techniques are shown to be relevant to improve the management of environmental systems.