Subaerial Landslide-Generated Waves: Numerical and Laboratory Simulations

Abstract

Subaerial landslide-generated waves (SALGWs) are among destructive hazards which have been not often studied in comparison with earthquake-generated tsunamis and submarine landslide-generated waves. This paper represents a brief review of the physical and numerical studies on SALGWs. Samples of the laboratory experiments are provided and it is highlighted that all the available data should be combined and studied collectively to overcome the discrepancies and improve our understandings of SALGWs. Commonly applied numerical approaches to simulate SALGWs are discussed. A Boussinesq-type model (LS3D) considering landslide as a rigid body, and a two-layer shallow-water type model (2LCMFlow) considering landslide as a layer of a Coulomb mixture are utilized to investigate the effects of landslide deformations on the characteristics of the landslide-generated waves (LGWs) based on a set of available experimental data. With a rigid landslide assumption, the maximum height of LGW is about 16% overstimated. Dense material deformes into a thick front—thin tail profile and induce a LGW consists of a larger wave crest than the wave trough while loose material shows a dam-break type behaviour with a LGW having a larger wave trough. A real case of SALGW is simulated by both models. The maximum LGW height predicted by the 2LCMFlow model which is closer to the physics is about 14% less than the equivalent value predicted by the LS3D model. On the other hand, the LS3D model, with the 4th order of accuracy of wave dispersion, simulates the LGW propagation stage more efficiently and with around 30% less runtime. Assessing the effects of the landslide initial submergence on the LGW characteristics shows that a semi-submerged, a submarine, and a subaerial landslides induce the largest wave crest, wave trough, and landslide runout distance, respectively. Combining different conceptual and mathematical models at the various stages of SALGWs initiation, propagation, transformations and runup can advance the current numerical practice, in this field, both from accuracy and computational efficiency point of views.