Monitoring desertification in a Savannah region in Sudan using Landsat images and spectral mixture analysis

Two Landsat images, acquired in 1987 and 2008, were analyzed to evaluate desertification processes in central North Kurdufan State (Sudan). Spectral Mixture Analysis (SMA) and multitemporal comparison techniques (change vector analysis) were applied to estimate the long-term desertification/re-growing of vegetation cover over time and in space. Site-specific interactions between natural processes and human activity played a pivotal role in desertification. Over the last 21 years, desertification significantly prevailed over vegetation re-growth, particularly in areas around rural villages. Changes in land use and mismanagement of natural resources were the main driving factors affecting degradation. More than 120,000 km(2) were estimated as being subjected to a medium-high desertification rate. Conversely, the reforestation measures, adopted by the Government in the last decade and sustained by higher rainfall, resulted in low-medium re-growth conditions over an area of about 20,000 km(2). Site-specific strategies which take into account I he interactions of the driving factors at local scale are thus necessary to combat desertification, avoiding any implementation of untargeted measures. In order to identify the soundest strategies, high-resolution tools must be applied. In this study the application of spectral mixture analysis to Landsat data appeared to be a consistent, accurate and low-cost technique to identify risk areas. (C) 2011 Elsevier Ltd. All rights reserved.