A physics-constrained emulator for time-discretised landslide runout modelling
Organizations: Department of Geological and Hydraulic Engineering, Tongji University, Shanghai, China · Department of Earth Sciences, University of Cambridge, Cambridge, UK · Department of Geography, King’s College London, London, UK · King’s Institute of Artificial Intelligence, King’s College London, London, UK ·
Abstract
Numerical simulations of landslides are commonly used to forecast impact areas. Full physics-based models provide accurate predictions, but their computational expense can preclude their use for large parameter sweeps. To achieve efficient modelling for such cases, we developed a physics-constrained neural network to emulate full landslide runout in a time-discretised manner. This emulator is trained on over 200,000 samples, each consisting of a pair of landslide thicknesses and velocities on a given DEM with a fixed time interval. To constrain the model's physical consistency, we define physics-based loss functions based on mass conservation and the system's kinetic energy loss. The model demonstrates high fidelity across the test dataset, achieving root mean square errors of approximately 0.06 m for thickness and 0.11 m/s for velocity in single-step predictions. In autoregressive rollout, these errors evolved modestly to about 0.21 m and 0.09 m/s after 110 iterations, while greatly improving computational efficiency by more than 50 times.
Figures & tables
| evaluation metrics | data-driven emulator | physics-constrained emulator |
|---|---|---|
| IoU ( h ) | 0.965 ± 0.021 | 0.989 ± 0.009 |
| IoU ( v ) | 0.867 ± 0.091 | 0.904 ± 0.078 |
| RMSE ( h ) | 0.049 ± 0.076 | 0.061 ± 0.062 |
| RMSE ( v ) | 0.012 ± 0.031 | 0.110 ± 0.035 |
| Pearson correlation coefficient ( h ) | 0.999 ± 0.000 | 0.999 ± 0.000 |
| Pearson correlation coefficient ( v ) | 0.994 ± 0.052 | 0.993 ± 0.015 |
Neural Surrogate Modeling
Neural surrogate modeling uses learned models to approximate costly simulations, enabling faster prediction, design, control, and inference.
See all 472 papers