Author(s): Yifan Li; Chendi Zhang; Shunyu Yao; Guotao Zhang

Linked Author(s):

Keywords: Disaster-inducing factors; Flash floods; Graph neural network; Regionalization; The Hengduan Mountains

Abstract: The geomorphology, meteorology and underlying surface conditions of the Hengduan Mountains region are complicated. And this region is seriously threatened by flash flood disasters. Flash flood regionalization contributes significantly to quantifying the spatial heterogeneity of disaster-inducing factors of flash floods. In this study, we applied three state-of-the-art unsupervised graph neural network (GNN) methods, which considered both the attribute characteristics and spatial topology structure of the input data, to establish the regionalization of flash floods at raster-scale in the Hengduan Mountains. Our research provides a scientific basis for flood control and disaster reduction in southwest mountainous areas.

DOI: https://doi.org/10.3850/iahr-hic2483430201-454

Year: 2024