Raw data of landslide travel distance

We developed a hybrid physics-based and data-driven approach to predict landslide travel distance. An inventory of shallow landslides triggered by rainstorm was compiled using multiple data sources, which was used as training data. Based on the principle of energy conservation, a travel distance prediction model that only requires two geometric parameters of landslide was then developed. The coefficients of the model are determined using the data of the compiled landslide inventory. The proposed model performs better for cases in our study area compared to commonly used empirical travel distance prediction models. Moreover, our model shows good generalization for other regions no matter that historical landslide inventories are available or not.