Spatio-temporal Dynamics of Land-use and Land-cover in the Mu Us Sandy Land, China, using the Change Vector Analysis
Karnieli, Arnon¹; Qin, Zhihao²; Wu, Bo³; Yan, Feng^3
1Ben Gurion University of the Negev, ISRAEL; ²Institute of Agro-Resources and Regional Planning, CHINA; ³Institute of Desertification Studies, Chinese Academy of Forestry, CHINA

This presentation is aimed at exploring the contradictory previously-published findings regarding the spatial extent of desertified vs. vegetation recovery areas in the Mu Us Sandy Land, China, and their respective causes, either natural or anthropogenic. In order to fulfill this goal, the Change Vector Analysis (CVA) technique was implemented on four sets of Landsat images from 1978 to 2007. Two variables were used - the Normalized Difference Vegetation Index (NDVI) and albedo. While the former variable represents rehabilitation, the latter depicts sandification. Using these variables, four categories of land-use and land-cover changes (LULCC) were computed - vegetation, exposed sands, water bodies, and wetlands.

The CVA enables us to separate the magnitude of the change in a continuous manner, rather than in broad qualitative classes. In the current work, magnitude and direction are exhibited in the same map. Each of the four LULCC categories was validated in selected hot spots, employing very high spatial resolution images (e.g., WorldView 2).

It is shown that, over the entire study period, the majority of the area remains unchanged. Despite the gradual increase in livestock and the consequent growing grazing pressure, 6% of the area, mainly in the east, is under human-induced rehabilitation processes, in terms of increasing vegetation cover. In other areas, it is proved that bare sands (5.1%) have expanded to the central-north and the southwest. These findings are in agreement with the long-term trends of a combined NOAA-AVHRR and MODIS dataset of the same variables, i.e., NDVI and albedo.