Testing Physics and Statics Based Hybrid ETAS Models

Currently, the Epidemic Type Aftershock Sequence (ETAS) model is state-of-the-art for forecasting earthquakes. In most of its applications, aftershocks are isotropically distributed in space, which potentially leads to an underwhelming performance in forecasting especially after large earthquakes. Here we develop a hybrid physics and statics based forecasting model to systematically account for the isotropic deficiency of the ETAS models. The hybrid model uses stress changes, calculated from inverted slip models of large earthquakes, as the basis of the spatial kernel in the ETAS model to get a more reliable spatial distribution of aftershocks. We evaluate five alternative ETAS models which feature different stress-based spatial kernels. We rank the forecasting performance of these models against the base ETAS model, which features an isotropic spatial kernel. In all cases, an expectation-maximization (EM) algorithm is used to estimate the ETAS parameters. The model approach has been tested on synthetic data to check if the known parameters can be inverted successfully. We apply the proposed method to large earthquakes in California like the 1992 Landers earthquake (Mw = 7.3). The probabilistic earthquake forecasts generated by the hybrid model have been tested using established Collaboratory for the Study of Earthquake Predictability (CSEP) test metrics and procedures. We show that the additional stress information, provided to estimate the spatial probability distribution, leads to more reliable spatiotemporal ETAS forecasts of aftershocks as compared to the base ETAS model.