How do the soil properties of a landslide affect the tsunami it generates?

What is it about?

The advanced 2D landslide model BingClaw is coupled to the tsunami model BoussClaw to simulate the giant Storegga slide and transatlantic tsunami that occurred about 8100 y BP, originating off the Norwegian coast. The run-out distance and lobe structure of the slide are well known, and the traces of the tsunami have been identified along the Norwegian coast, on the Shetland and Faroe Islands, in Scotland and Iceland. Combining the landslide and tsunami data, the soil parameters (corresponding to a Herschel–Bulkley fluid with remolding) in the landslide simulation can be constrained to a degree that would never be possible with just one of the two data sets or with a less advanced landslide model. The good agreement between simulations and observations indicates that BingClaw includes the necessary ingredients for reliable simulation of such complex events.

Featured Image

Why is it important?

Tsunamis generated by landslides can have devastating consequences and need to be included in hazard assessment and mitigation where this is possible. Today's tsunami models allow quite precise calculation of tsunami propagation and run-up―if the conditions generating the wave are known. In the case of landslide-generated tsunamis, it is the slide properties (acceleration, velocity, flow depth, density, shape) that set the initial/boundary conditions, but most often grossly simplified block models are used for this. 1D and 2D (depth-averaged) landslide models improve substantially on this, yet they still use simplistic friction models to describe the landslide motion. Typically, the material parameters have to be chosen far from what is measured in situ or in the laboratory because the sliding soil transforms along the path (remolding, incorporation of ambient water, hydroplaning, etc.). This makes prognostic use of the models doubtful or even impossible. We developed BingClaw, a 2D model with a more advanced rheological formulation (Herschel–Bulkley fluid including remolding), hydrodynamic drag and added-mass effects. This model allows us to simulate subaerial and subaqueous landslides of clay-rich soils and even quick-clay slides with parameter values equal to the actual measured soil properties so that the model can be used prognostically. BingClaw is also able to mimic retrogressive release of landslides to some degree. In this way, the tsunami generated by such landslides is simulated much more realistically than in a model in which the entire mass starts moving at once.

Perspectives