Abstract:Study of alpine grassland in permafrost zones of Qinghai-Tibetan Plateau has been widely concerned.As the climate warming,researches of alpine grassland types were rarely reported in this region.In this paper,there were 490 vegetation investigation samples and 27 variables derived from 3 remote sensing datasets.Used decision-tree classification model,the distribution of potential vegetation types was predicated by 10 climate system models under 4 climate scenarios in 2050 and 2070.There were alpine swamp meadow,alpine meadow,alpine steppe,alpine desert and barren land in the result of potential distribution.Meanwhile,the current distribution of alpine grassland types was developed for data analysis.The results indicated that in 2050 and 2070 area of bare land and alpine meadow slightly decrease,area of alpine grassland and alpine desert slightly increase and area of alpine marsh meadow did not significantly change in comparison with the current alpine grassland distribution area.The results were basically similar for 4 representative concentration pathways.This study could not only provide data support for the research on alpine grassland climate change,but also provide different methods and tools for exploration carbon cycle in permafrost zones of Qinghai-Tibetan Plateau.