An Efficient Landslide–Tsunami Model with Two-Phase Rheology: Challenges and Implementation

Subaerial landslides-induced tsunamis are challenging due to their interaction with air and water. This implies the need for modelling granular collision stresses, fluid-grain interaction and impact shock, and also the resulting cratering. Fully three-dimensional models can be too computationally expensive in practical use such as for analysing parameter sensitivity. Depth-averaged models offer an alternative by enabling faster simulations. The challenge then is in retaining essential physical processes in the depth-averaged process. This work focuses on the development of a two-phase depth-averaged model designed to simulate both subaerial and submarine landslides and the waves they generate using a variant of the μ(I) landslide model for the granular rheology. The approach aims to capture the interaction between solid and fluid phases while maintaining computational efficiency. We compare results using this new model (presently under development) with simpler existing models. We also cover some key challenges encountered during model formulation and implementation, including mathematical and numerical issues such as ill-posedness, instability, and the need for well-balanced schemes. We examine the suitability of various numerical schemes and solvers for this application and present landslide parameter sensitivity analysis. Finally, we compare our approach with alternative modelling frameworks to evaluate performance and reliability and briefly discuss gaps in current depth-averaged modelling approaches. By addressing these questions, we attempt steps towards advancing efficient and robust tools for simulating landslide-generated waves and improving coastal hazard assessment through embedding more advanced landslide formulations in depth averaged models.