Numerical simulation of tsunami runup in northern Chile based on non-uniform k22 slip distributions

Abstract

A large seismic gap lies along northern Chile and could potentially trigger a M (w) similar to 8.8-9.0 megathrust earthquake as pointed out in several studies. The April 1, 2014, Pisagua earthquake broke the middle segment of the megathrust. Some slip models suggest that it ruptured mainly from a depth of 30 to 55 km along dip and over 180 km in length, reaching a magnitude M (w) 8.1-8.2. The northern and southern segments are still unbroken; thus, there is still a large area that could generate a M (w) > 8.5 earthquake with a strong tsunami. To better understand the effects of source parameters on the impact of a tsunami in the near field, as a case study, we characterize earthquake size for a hypothetical and great seismic event, M (w) 9.0, in northern Chile. On the basis of physical earthquake source models, we generate stochastic k (-2) finite fault slips taking into account the non-planar geometry of the megathrust in northern Chile. We analyze a series of random slip models and compute vertical co-seismic static displacements by adding up the displacement field from all point sources distributed over a regular grid mesh on the fault. Under the assumption of passive generation, the tsunami numerical model computes the runup along the shore. The numerical results show a maximum peak-runup of similar to 35-40 m in the case of some heterogeneous slip models. Instead, the minimum runup along the coast, from the heterogeneous slip models tested, almost coincides with the runup computed from the uniform slip model. This latter assumption underestimates the runup by a factor of similar to 6 at some places along the coast, showing agreement with near-field runups calculated by other authors using similar methodologies, but applied in a different seismotectonic context. The statistical estimate of empirical cumulative distribution functions conducted on two subsets of slips, and their respective runups, shows that slip models with large amount of slip near the trench are more probable to produce higher runups than the other subset. The simple separation criterion was to choose slip models that concentrate at least 60 % of the total seismic moment in the upper middle part of the non-planar rupture fault.