A Novel Holistic Framework for Evaluation of Landslide Susceptibility in the Hengduan Mountains, the Qinghai-Tibet Plateau

Abstract

Amplifying landslide hazards in the backdrop of warming climate and intensifying human activities call for an integrated framework for accurately evaluating landslide susceptibility at fine spatiotemporal resolutions, which is critical for mitigation of increasingly high landslide disaster risks. Yet dynamic landslide susceptibility mapping is still lacking. Using high-quality 14,435 landslide and non-landslide data, here we developed a novel holistic framework for evaluating landslide susceptibility, considering landslide-relevant internal and external factors based on cloud computing platform and algorithmic models, that enables dynamic updating of landslide susceptibility map at the regional scale, particularly in regions with highly complicated topographical features such as Hengduan Mountains considered in this study. We compared Classification and Regression Trees (CART), Support Vector Machines (SVM), and Random Forest (RF) classifiers to screen out the best portfolio model for landslide susceptibility mapping on the Google Earth Engine (GEE) platform. We found that Random Forest (RF) classifier integrated with synergy mode had the best modelling performance with 90.48% and 89.24% accuracy and precision respectively. We also found that forests and grasslands had the controlling effect on the occurrence of landslides, while human activities had a notable inducing effect on the occurrence of landslides within the Hengduan Mountains. The non-landslide areas were 4.41 times and 10.38 times higher than the landslide areas in the forest-dominated and grassland land-dominated regions, while the landslide areas were 3.66 times and 4.97 times higher than the non-landslide areas in the cropland-dominated and impervious surface-dominated regions, respectively. This study highlights the performance of the holistic landslide susceptibility evaluation framework proposed in this study and provides a viable technique for landslide susceptibility evaluation in other regions of the globe.

Keywords: Landslide susceptibility assessment, Multisource data, Random forest, Google Earth Engine, Hengduan Mountains