The discretization of system given in model equations is performed by an explicit first-order IFCP Finite Volume Scheme where the discretization of the Coulomb friction term is performed following Fernández-Nieto et al. (2008) (see Fernández-Nieto et al. (2011) for details on the stability, convergence, and efficiency of the numerical scheme). The resulting scheme has been implemented in Graphics Processor Units (GPUs) using CUDA, achieving a speed-up of two orders of magnitude compared to a conventional CPU implementation (see Castro et al. (2011b) for a review and de la Asunción et al. (2012)). This methodology allows us to considerably improve the efficiency of the algorithm as well as the size of the discrete problems that can be solved.
This model reduces to the usual nonlinear shallow-water system when the layer of granular material is not present or when it has zero velocity and this layer reaches equilibrium. Therefore, the model can be used to numerically reproduce the different stages of a landslide tsunami simulation: the landslide tsunami generation, the far field wave propagation and, finally, the coastline inundation and the run-up height reached by the tsunami wave (Macías et al., 2015).
