Recently, the research team led by Professor Ji Jian of Hohai University, in collaboration with researchers from the Polytechnic University of Catalonia, Spain, has made a breakthrough in addressing the challenge of predicting rainfall-induced landslide susceptibility at the regional scale. They tackled the significant challenge of uncertainty in rainfall infiltration models caused by the lack of geotechnical and hydrogeological data at the regional scale. The research results were published under the title "Fast physically-based probabilistic modelling of rainfall-induced shallow landslide susceptibility at the regional scale considering geotechnical uncertainties and different hydrological conditions" in the international top journal in the field of geotechnical numerical computation, *Computers and Geotechnics*.

The research team combined the two movement modes of rainfall surface processes—"lateral flow" and "vertical flow"—and analyzed the mechanisms of groundwater recharge formed by medium- and long-term rainfall and the temporary rise of groundwater level induced by event rainfall. On the basis of further comprehensively considering the multi-source uncertainties of geotechnical parameters and hydrogeological parameters at the regional scale, they adopted the efficient geotechnical reliability algorithm (FORM-HLRF-x) to achieve regional rainfall landslide susceptibility analysis, proposed a new probabilistic physical model FSLAM-FORM (Figure 1), and developed the "Py.GIS-FSLAM-FORM" landslide susceptibility zoning software applicable to the Microsoft operating system, enabling more accurate and efficient regional-scale landslide susceptibility assessment and prediction/early warning. Using the proposed method, the research team conducted a case study in the Val d'Aran region of northern Spain and found a synergistic effect between effective recharge during rainfall and groundwater rise caused by event rainfall. Clustering results further indicated that within the slope range of 30 to 40 degrees and aspect range of 100 to 250 degrees, the occurrence of event rainfall can intensify the probability of landslides (Figure 2).

Untitled image

Figure 1. The proposed FSLAM model: (a) lateral flow caused by antecedent rainfall; (b) vertical flow induced by event rainfall

Untitled image

Figure 2: Very high susceptibility results obtained from FSLAM-FORM probabilistic analysis: (a) regional distribution map; clustering effects considering slope and aspect (b) locations of actual landslide points; (c) different rainfall conditions

The study further explored the influence of regional geotechnical stratigraphic parameter variability on rainfall-induced shallow landslide susceptibility. The results showed that the spatial distribution uncertainty of geotechnical parameters (i.e., coefficient of variation, COV) under different hydrological scenarios can lead to a broader and more dispersed spatial distribution of medium- to high-susceptibility areas, but does not directly expand the extent of very high-susceptibility areas; the combined effect of medium- and long-term effective rainfall and triggering event rainfall intensifies the transition of areas originally in very low or very high susceptibility states toward medium or high susceptibility states (Figure 3).

The model proposed in the paper provides new insights for improving rainfall landslide risk analysis at the regional scale and has certain reference value for rapid landslide susceptibility zoning and landslide disaster prediction and early warning worldwide.

Untitled image

Figure 3. Influence of COV variation on very high-susceptibility areas under different hydrological conditions

The first author of the paper is Cui Hongzhi, a doctoral student from the School of Civil and Transportation Engineering of our university. The corresponding authors are Professor Ji Jian of our university and Professor Marcel Hürlimann of the Polytechnic University of Catalonia, Spain. Co-authors also include Professor Vicente Medina of the Polytechnic University of Catalonia, Spain. The research work was funded by the National Natural Science Foundation of China Joint Key Project "Research on Strong Earthquake Response and Landslide Disaster Protection Mechanisms of Highway Slopes in Deep-Cut Canyons in High-Altitude Regions (U22A20594)".

Related paper information:

Cui, Hongzhi, Medina Vicente, Hürlimann Marcel, and Ji Jian. Fast physically-based probabilistic modelling of rainfall-induced shallow landslide susceptibility at the regional scale considering geotechnical uncertainties and different hydrological conditions. Computers and Geotechnics 172 (2024): 106400. doi: https://doi.org/10.1016/j.compgeo.2024.106400