Dataset of Glacial Lakes Inventory on the Tibetan Plateau in 2023

Glacial lakes, as distinctive hydrological and geomorphological features formed during glacial ablation, have exhibited rapid expansion in alpine and polar regions worldwide in recent years. Systematic cataloging of these lakes through documenting their spatial distribution and geometric parameters provides critical scientific value for understanding glacier evolution, assessing climate change impacts, evaluating freshwater resources, and enabling early warnings for glacial lake outburst floods (GLOFs). To address persistent challenges including the controversial delineation of glacial lake distribution areas and the heavy reliance on manual visual interpretation in remote sensing extraction, this study developed an innovative approach that integrates mountain units with glacier boundaries to define glacial lake extent. By employing Sentinel-2 MSI imagery and Attention DeepLab V3+ deep learning model, we achieved comprehensive automated extraction and systematic inventorying of glacial lakes across the entire Tibetan Plateau. The results reveal that in 2023, the Plateau contained 29,089 glacial lakes (≥0.005 km²) with a combined area of 2,020.65 km², predominantly concentrated in the southeastern regions including the Nyainqêntanglha Mountains (6,306 lakes covering 483.65 km²), Eastern Himalayas (3,619 lakes covering 265.25 km²), and Hengduan Mountains (5,131 lakes covering 256.37 km²). The inventory demonstrates that while small lakes (0.01-0.05 km²) numerically dominate (62.66% of total count), they contribute only 20.39% of the total area, whereas the relatively scarce large lakes (>1 km², comprising merely 0.68% of the count) account for a substantial 24.22% of the total area. In terms of typology, non-glacial-contact lakes predominate (73.88% by count, 73.50% by area), followed by moraine-dammed lakes (24.27%) which are primarily distributed across the central-eastern Himalayas and Nyainqêntanglha Mountains, with supraglacial lakes being the least common (1.85%). This comprehensive dataset not only accurately reflects the current status of glacial lake resources on the Tibetan Plateau but also provides essential data support for critical applications including GLOF risk assessment and regional hydropower development planning.