Numerical Investigation of Landslide-tsunami Propagation and Transformation in a Wide Spectrum of Water Body Geometries

Landslide-tsunamis are generated by mass movements such as soil or rock impacting into a water body, which resulted in major catastrophes in the recent past. Thus far, scientists mainly investigate landslide-tsunamis in two idealised geometries: wave flume (2D, confined wave propagation) and wave basin (3D, unconfined wave propagation). The landslide-tsunami height in these geometries can differ by an order of magnitude or even more, and limited knowledge of the wave features in intermediate geometries, between 2D and 3D, is currently available.

This work addresses this shortcoming by focusing on the landslide-tsunami propagation in intermediate geometries in the far field where the waves are reasonably stable. The non-hydrostatic Non-linear Shallow Water Equation model SWASH was used to simulate linear, Stokes, cnoidal and solitary waves, as landslide-tsunamis show similarities to these theoretical wave types. Simulations were performed in 6 different idealised water bodies including intermediate geometries characterised by wave propagation confined by diverging side walls at θ = 7.5, 15, 30 and 45° as well as in 2D (θ = 0°) and 3D (θ = 90°).

The wavefront length was found to be an excellent parameter to correlate the wave decay in all geometries along the slide axis and to confirm the decay rate with Green`s law. Diffraction theory further confirms some of the tsunami features in 3D. In addition, semi-empirical equations were derived to predict the wave decay of the theoretical waves based on available wave parameters from 2D studies. Further, numerical simulations of experimental landslide-tsunami time series were performed in 2D to separate the primary wave decay caused by frequency dispersion, which is essentially excluded by the investigated theoretical wave types. Frequency dispersion was thereby found to be negligible small for waves in shallow water, i.e. for solitary- and cnoidal-like waves, at least for the purpose of initial landslide-tsunami hazard assessment. However, frequency dispersion becomes more important for Stokes-like waves in deeper water.

Future work will focus on expanding the numerical simulations to include also wave generation, as well as comparing the results to real water bodies to develop more complete criteria for landslide-tsunami hazard assessment.