Abstract: Based on GIMMS (Global Inventory Modeling and Mapping Studies) data from 1982 to 2006, we calculated the vegetation phenological data in arid zone of Central Asia and Xinjiang, China using threshold method with the Timesat software. The Mann-Kendall trend analysis and Theil Sen slope tools were used to assess the spatial-temporal change. Combined with land cover data and digital elevation model, varieties of phenological metrics were also analyzed for different vegetation cover types and different elevation zones. According to our results, no significant phenological change was detected for the whole study area, but obvious change was found in some local areas. The area where the start of growth season (SOS) had advanced was also found to have a significant extended length of growth season (LOS). Different vegetation cover types showed remarkably different patterns in phonological changes. The end of growth season (EOS) of Deciduous Broadleaf Forest delayed significantly. Except for the Open Shrub, Closed Shrub and Bare Ground, most vegetation cover types showed an extended LOS. The phenology in different elevation varied obviously. Analyses further showed that the water and heat conditions in the 2000-3000 m elevation zone were improved, which may result in the most advanced SOS, delayed EOS and prolonged LOS in this zone.

Key words: remote sensing, vegetation phenology, Central Asia, Xinjiang, arid zone