Glacier Retreating Analysis on the Southeastern Tibetan Plateau via Multisource Remote Sensing Data

Abstract

Accurate multitemporal glacier change investigations and analyses are lacking on the southeastern Tibetan Plateau (SETP). A combination of photogrammetry, optical remote sensing, and synthetic aperture radar datasets can accurately identify large-scale glaciers; In this article, glaciers in three periods on the SETP (1970s, 2000, 2020) were identified from multisource remote sensing data based on a deep learning method and manual visual interpretation, and multitemporal glacial inventory data from relatively high-frequency source imagery. Totals of 11648, 12993, and 11875 glaciers were identified in the 1970s, 2000, and 2020, with total areas of 13372.08 km2, 11692.31 km2, and 10612.94 km2, respectively. The general distribution of SETP glaciers was identified to be typical of alpine glaciers dominated by small-sized glaciers. The average elevation of glaciers was approximately 5000 m; the slopes were mostly lower than 40°, and the main aspect was southeast, followed by south and southwest. The glaciers retreated from the 1970s to 2020, and a total glacier area of approximately 2759.14 km2 was degraded during this time, with an average annual melting rate of 0.45% yr−1. Rising summer temperatures may be the driving force behind the continuous decline in the glacier area. Overall, the results obtained in this article showed relatively low uncertainty involved in the identification of glaciers compared to some previous studies. The results can provide accurate glacier information for glacier monitoring and modeling studies on the SETP.

Keywords

• Deep learning
• glacier retreat
• remote sensing
• southeastern Tibetan Plateau