Abstract: Wetland degradation and ecosystem structural degeneration caused by climate change and unsustainable human activity have become more common since the middle of the 20th century. It is difficult to carry out a ground-based investigation because of complex spatial mosaic of grassland interspersed with bodies of water and marsh wetlands in the eastern part of the Qinghai-Tibet Plateau. We therefore used remote sensing technology to monitor real-time dynamic change in wetland distribution, to enhance understanding of the changes in the ecological environment of the Qinghai-Tibet Plateau. Using Landsat TM/ETM+/OLI images, the areas of marsh wetlands in the eastern Qinghai-Tibet Plateau were extracted by visual interpretation from 1991, 2000, 2010, and 2016 records. The area change of marsh wetlands, and direction and rate of movement were analyzed based on a dynamic transfer matrix methodology. Landscape indices at the patch level were used to quantify the spatial and temporal dynamics of the marsh wetlands. We also used the grey correlation method together with meteorological data and statistical information to analyze the factors driving marsh wetland change. It was found that: 1) The marsh wetlands are mainly distributed in the northeast of the study region. For the 1991, 2000, 2010 and 2016 analyses, the total marsh wetland areas were 6739.89, 6231.39, 5849.59 and 5649.35 km², respectively, meaning that a reduction in total area of 1090.54 km² was recorded during the 26 year study period; 2) The annual rate of wetland loss decreased gradually during the 26 years from -7.54% to -3.42%, and the lost marsh wetlands were mainly transformed into alpine grassland; 3) The number of patches increased initially and then decreased, while patch density continued to increase, indicating an increasing degree of fragmentation of the marsh wetlands. The largest patch index decreased initially and then increased slightly, while the landscape shape index increased initially and then decreased slightly, reflecting the decreasing dominance of marsh wetlands and the complexity of landscape shape. The splitting index increased initially and then decreased slightly, while the aggregation index decreased throughout the study period, reflecting a tendency for marsh wetlands to be fragmented and discrete; 4) Human factors are the primary reason for reduction of marsh wetland area in the eastern Qinghai-Tibet Plateau, followed by climatic factors. Specifically, the factors in order of influence were: Gross output value of animal product, gross domestic product, population increase, temperature and precipitation.

Key words: marsh wetlands, Qinghai-Tibetan Plateau, remote sensing, landscape pattern, dynamic change, driving factors