干旱区生态环境知识资源中心

Desertification is considered as a major worldwide environmental problem mainly caused by the climate changes and human activities during the last decades. Areas affected by desertification processes are gradually losing their level of biological quality and productivity. Among the different indicators of desertification, degradation of vegetation cover and increasing amount of bare soil have been popularly used by researchers. In India, desertification is one of the major sluggish hazards which is found in northwestern part of this country, mainly in the state of Rajasthan. The infringement of the Thar Desert has become a serious problem in the adjoining districts of Bikaner, Churu, and Nagaur. In this study, a linear spectral unmixing (LSU) method has been used for end-member fraction estimation primarily to differentiate the sand percentage and vegetation cover percentage. This linear spectral unmixing model is a widely used technique in remote sensing to estimate the fractions of several individual surface components present in an image pixel and the pure reflectance spectrum of a component which is called end-member. The LSU technique is able to monitor desertification process in terms of fractional changes in bare soil (sand) and vegetation covers. These two land features are the most crucial indicator of desertification and their long-term changes can produce expected result in identification of desertification process in an area. The long-term multispectral satellite data such as Landsat Thematic Mapper (TM) (1990, 1995, and 1999) and Enhanced Thematic Mapper (ETM+) (2003, 2009) have been used in this study. The time series analysis of fractional images of vegetation cover and bare soil has been employed for monitoring desertification processes over a long period. After analyzing the changes, some distinct patches of vegetation depletion coupled with increasing bare soil fraction were identified within the region that clearly indicates the ongoing process of desertification over there.