Constraining stochastic slip rupture models with geodetic coupling models: An example application to PTHA in the Cascadia

Abstract:

The notion that heterogeneities exist in fault characteristics, rupture propagation, slip partitioning, among others are accepted and common assumptions by the scientific community; yet influence of heterogeneities like fault slip deficit associated with slab locking are overlooked in rupture modeling techniques. Coastal GPS measurements have shown variations in plate motions along a given subduction zone. These variations have been used to calculate and create locking models for given areas. In this work we model stochastic slip ruptures using the von Karman autocorrelation function (ACF) with application of the Karhunen-Loeve (KL) expansion for random slip patch generation. Unlike previous techniques, here the locking models are implemented within the KL expansion. While difficult to observe variations in rupture properties of single ruptures from including the locking models, when looked as a whole, clear patterns are presented. In order to understand the implication of the locking models in the rupture modeling workflow, we apply these models to probabilistic tsunami hazard analysis (PTHA) for the Cascadia subduction zone. The three classes of rupture models - two with inclusion of different locking models, and one without - ranging in magnitude between M7.8 and M9.1 are then implemented into the open source tsunami modeling software GeoClaw in order to calculate estimated tsunami arrival heights along the Pacific northwest. By analyzing hazard curves for the coastal points there is a clear effect in the modeled hazards produced by assuming different locking scenarios. Influence in heterogeneity produced by locking models is clear, however, our ability to constrain locking models for subduction zones is dependent on the scientific community’s push for advances in seafloor geodesy.