Assessment of desertification in Eritrea: land degradation based on Landsat images

Abstract

Remote sensing is an effective way in monitoring desertification dynamics in arid and semi-arid regions. In this study, we used a decision tree method based on NDVI (normalized difference vegetation index), SAVI (soil adjusted vegetation index), and vegetation cover proportion to quantify and analyze the desertification in Eritrea using Landsat data of the 1970s, 1980s and 2014. The results demonstrate that the NDVI value and the annual mean precipitation declined while the temperature increased over the past 40 a. Strongly desertified land increased from 4.82×104 km2 (38.5%) in the 1970s to 8.38×104 km2 (66.9%) in 2014: approximately 85% of the land of the country was under serious desertification, which significantly occurred in arid and semi-arid lowlands of the country (eastern, northern, and western lowlands) with relatively scarce precipitation and high temperature. The non-desertified area, mostly located in the sub-humid eastern escarpment, also declined from approximately 2.1% to 0.5%. The study concludes that the desertification is a cause of serious land degradation in Eritrea and may link to climate changes, such as low and unpredictable precipitation, and prolonged drought.