Spatial–temporal evolution of glacial lakes and multi-phase geological hazards susceptibility evaluation of Midui Gully, Xizang

Abstract

The rate of glacier retreat and glacial lake formation on the Qinghai-Tibet Plateau is closely related to geological hazards susceptibility. Inadequate field data on glaciers, glacial lakes, geological hazards are limited, and geological hazard events are still the subject of detailed research. In this study, the multi-year data of Midui gully (the watershed area is 169.74 km²) in Linzhi, Xizang were used to analyze the evolution of glaciers, glacial lakes, and geological hazards susceptibility. This study used a unit-variable linear regression model to evaluate the spatial–temporal evolution of glaciers and glacial lakes from 2001 to 2017. The results showed that glaciers retreated at a rate of 0.0063 km²/a, while glacial lakes increased linearly at 0.0169 km²/a. The results of evaluating the spatial–temporal evolution of geological hazards susceptibility using the certainty factor and artificial neural network (CF-ANN) model indicate that there is a slow-growing tendency of geological hazards susceptibility, extremely high susceptibility (EHS) increased by 3.76%, high susceptibility (HS) decreased by 1.16%, moderate susceptibility (MS) decreased by 1.69%, low susceptibility (LS) increased by 2.18%, and extremely low susceptibility (ELS) decreased by 3.09%. The 7-phase of the area under the curve (AUC), recall, accuracy, and F1 score were greater than 96%, 0.94, 0.90, and 0.92, respectively. This study shows that glaciers are retreating, glacial lakes are expanding, and geological hazards susceptibility is increasing in Xizang. The geological hazards susceptibility obtained from the CF-ANN model has a good predictive value for high-altitude areas and provides effective guidance for further land use and tourism resource planning.