Desertification,which is referred to land degradation in drylands,is often triggered by human activities and natural factors.To understand how land use changes and their drivers influence long-term land cover dynamics we used remote sensing data and multilevel statistical model to analyze spatio-temporal patterns of desertification in the Uxin Banner of Inner Mongolia,China.We found that desertification-induced land degradation expanded during the 1977-1987,but this process reversed during 1987-2007.Multilevel statistical modeling results suggest rural population density,the density of livestock,and accessibility to markets all have significant effects on desertification reversion at the local level,while changes in precipitation and land use policies are important at the region level.Particularly,were found land use policies were principal drivers of vegetation recovery.We therefore conclude that multilevel statistical model is effective and practical tool for understanding mechanisms of land degradation in Uxin Banner and other similar regions.
multilevel statistical modelscaledesertificationland use policyUxin Banner
NETL
NSTL
维普
万方数据
