INFORMATION EXTRACTION AND CHANGE ANALYSIS OF MAJOR LAKES IN TIBETAN PLATEAU BASED ON LANDSAT REMOTE SENSING IMAGES

Abstract. The water resources of Tibetan plateau, particularly the lakes, has been influenced by global climate change and also reacted to global change. It is important to study the lake changes in the Tibetan plateau. This paper aimed to analyse the changes detected from remote sensing images for the typical lakes in Tibetan Plateau, including Qinghai Lake, Nam Co and Selin Co, using different information extraction methods, such as the single-band threshold method, NDWI, MNDWI, MSCT, MLC, SVM and object-oriented classification method. Results show that object-oriented classification method is better at removing ice, snow, cloud and mountain shadow, while the NDWI and MNDWI generated the worst results for lake extraction. After that, the lake's long temporal changes for Qinghai Lake, Nam Co and Selin Co derived from 12 images in different periods have been analysed using GIS technology. Results show that between 1985 and 2015, Qinghai Lake has expanded after a steady expansion; Nam Co was in the state about stable to expansion to stable. The area increased by 70.03 km², a growth rate of 2.33 km² a^-1 and had a small change in area; Selin Co expanded after a stable state, and the area is always growing, and Selin Co's growth rate is larger than Qinghai Lake and Nam Co at the same time.