PROFILE HUMAN-INDUCED VEGETATION CHANGE IN ARHORQIN BANNER OF CHINA USING TIME SERIES DATASETS

Discriminating human-induced vegetation change is essential for sustainable managements of arid and semi-arid ecosystems. Residual Trends method (RESTREND), an effective quantitative method, has been widely used to discriminate human-induced vegetation changes in specific arid and semi-arid ecosystems. However, how to define homogeneous spatial neighborhood to determine reference pixel for estimating potential climate-solely-induced vegetation growth is still a challenge. This paper firstly detected vegetation dynamics in ArHorqin Banner of China from 2000 to 2014 by Mann-Kendall method, and then used RESTREND to profile human-induced changes. We optimized strategy in RESTREND by using statistical analysis and trajectory analysis to automatically define flexible homogeneous neighborhood. Results indicated that 18.6% of study area had significantly changes. Both climate change and human activities contributed to the changes. The influence of human activities on vegetation dynamics is more than climate change, and it was the main driver for vegetation decrease in study area.