A Quick Detection of Lake Area Changes and Hazard Assessment in the Qinghai–Tibet Plateau Based on GEE: A Case Study of Tuosu Lake

Abstract

Under the background of global warming, lake changes in the Qinghai–Tibet Plateau and their impacts have been widely concerned. To avoid downloading and preprocessing numerous remote sensing images, we used JavaScript to preprocess the Landsat OLI and Sentinel-2 MSI images based on the Google Earth Engine (GEE) platform and extract the water area of Tuosu Lake by combining NDWI, MNDWI, and SVM methods. The area changes of Tuosu Lake in 2000–2021, and their impact on the surrounding infrastructure were further analyzed. The results show that 1) in different methods of surface water extraction, the segmentation based on NDWI is the most efficient for delineating the Tuosu Lake, and the optimal threshold is 0.15 ± 0.03. 2) During 2000–2021, the area of Tuosu Lake increased by about 45.79 km2, which can be roughly divided into a slow growth period (2000–2013, 1.20 km2/a), a stable period (2013–2016, 0.49 km2/a), slow expansion period (2016–2017, 1.97 km2/a), and rapid expansion period (2017–2021, 7.43 km2/a). In particular, in 2019–2020, the lake area increased sharply by 9.22 km2. 3) Affected by the expansion of Tuosu Lake, the highway along the northern lake had been seriously threatened, and about 19.17 km (83.28%) of it was submerged by 2021. In addition, the distance between Tuosu Lake and the Qinghai–Tibet Railway has been shortened year by year, with the shortest distance of 0.85 km in 2021.