Abstract: The land use in Lhasa is divided into six types inclusive of farmland, woodland, grassland, water area, building land and unused land according to the remote sensing data in 2000 and 2010 on Erdas platform. Then the distribution of land use and land use transition matrix in the two periods are obtained using ArcGIS spatial analysis techniques, and the land use changes of the years 2020, 2030 and 2040 are predicted by using Markov model. On this basis, the carbon emissions per year of these years are calculated and the emissions every 10 years in the region are predicted. The final results show that the change of ecological land structure and the increase in ecological land (forest) area are the main causes of carbon sequestration. Overgrazing, increasing consumer demand for animal husbandry and development of tourism are the main causes of carbon emissions during land use change. Finally, some suggestions to control carbon emission and limit the decrease of ecological land are put forward: protecting ecological environment, restricting overgrazing, and promoting low-carbon tourism.

Key words: ArcGIS, carbon emissions, land use change, ecological land, Markov model