Automated satellite-based glacial lake inventory and change detection in High Mountain Asia

高山亚洲地区基于卫星的冰川湖泊自动化清查与变化检测

阅读:4

Abstract

The High Mountain Asia (HMA) region has the highest concentrations of high-altitude lakes in the world and experienced more frequent glacial lake outburst floods (GLOFs) in recent decades. Continuous monitoring and mapping of glacial lakes at high spatiotemporal resolution are crucial for understanding this rapidly evolving and vast landscape as well as associated disasters. While existing automated methods show significant promise for mapping glacial lakes, their completeness and accuracy need further improvement to regularly produce glacial lake databases in highly dynamic regions like HMA. In this study, we present a fully automated method integrating open-source remote sensing datasets, including Landsat-8, Sentinel-1, Sentinel-2, and Copernicus Digital Elevation Model (DEM), to map glacial lakes and generate a comprehensive inventory of glacial lakes across the HMA region. Our 2022 inventory comprising 31,698 glacial lakes across HMA covers an area of approximately 2,240 km[Formula: see text]. Most lakes are situated between 4000-5400 m asl elevations, with the Eastern Himalaya exhibiting the greatest lake area coverage. We achieved robust accuracy exceeding 96% in the glacial lake size bin (20,000-100,000 m[Formula: see text]), demonstrating the effectiveness of our method in mapping and detecting small lakes while successfully delineating all larger glacial lakes (> 100,000 m[Formula: see text]). The method was applied to two observation periods (2016-17 and 2022-24), enabling analysis of changes in lake area and spatial distribution patterns. Across HMA subregions, Qilian Shan region showed the highest expansion rate of 22.5% between the observation periods, while the Pamir region showed the least changes (2.9%) in this period, contributing to a 5.5% net change in overall glacial lake area at the HMA scale. Our automated approach provides substantial improvements over previous methodologies in data integration, accuracy and completeness, which can be utilized for routine updating of glacial lakes in HMA and elsewhere.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。