No abstract

No abstract

No abstract

Practical landslide predictions are instrumental to effective landslide risk management. Recently, the use of machine learning (ML) has become a promising alternative means for landslide predictions. This paper discusses the recent progress of a pilot study of ML-powered rainfall-based natural terrain landslide susceptibility analysis in Hong Kong. This study is different to other similar studies in that: (1) data sampling commonly used to deal with an imbalanced dataset is not adopted, and (2) the incorporation of domain knowledge on landslide characteristics for the development of physically meaningful ML models. The results are found to be promising, with the achieved ROC AUC up to 91.5% based on the testing data. The resolution of the susceptibility map is enhanced by approximately three orders of magnitude further than the introduction of additional features critically selected with feature engineering and based on domain knowledge and past experiences.

Steep natural terrain in Hong Kong, combined with deep weathering profile and high seasonal rainfall, is highly susceptible to rain-induced, shallow landslides. With over 35 years of practice in landslide risk management, Hong Kong has progressively built up a series of important databases that facilitate conducting state-of-the-art research and development works on landslides. Amongst others, there is a dense network of raingauges that provides state-of-the-art rainfall records, a high-resolution inventory of historical landslides and a LIDAR-based digital terrain model for natural terrain. This paper will firstly discuss the previous landslide susceptibility map for natural terrain in Hong Kong, and then present a new territory-wide rainfall-based landslide susceptibility analysis for natural terrain that takes cognizance of effect of slope angle. The year-based susceptibility model correlates landslide density with normalized maximum rolling 24-hour rainfall for eight different slope angle classes. The potential applications of the outcomes of the landslide susceptibility analysis will also be discussed.