Kinematics of M6.4 and M7.1, 2019 Ridgecrest, California, earthquakes inferred by the fuzzy inversion method

Abstract:

The two events with magnitudes M6.4 and M7.1 were the largest events in the seismic sequence that hit California in July 2019. The two events were separated by only 34 hours. The aftershock pattern revealed a complex geometry of the causative faults, showing two perpendicular planes, dipping almost vertically, which ruptured during the seismic events. The main shock, with M7.1, is a shallow event at a depth of ~6km (NEIC). It caused a large surface rupture of about 50 km.

In this work, we study the Ridgecrest earthquakes using the novel fuzzy finite-fault inversion method. Accordingly, the complexity of slip parameterization is reduced by employing a machine-learning based fault discretization, in which we select only a few number of balanced-with-data parameters to describe slip at various frequencies. In this study, we combined two datasets: strong-motion accelerograms (31 stations by SCSN) and high-rate GNSS displacements (22 stations by USGS), in order to invert the spatial slip distribution, independently and consistently at different frequencies. Our method determines the most adequate number of fault parameters based on a maximum likelihood analysis of model uncertainty, a criterium that maximizes the power of data reproduction. In the frequency bands where both data types are viable (0.1<f<0.5Hz), we divide the entire data set into two subsets, one for training and the other for validation. Thus, we evaluate the robustness of the trained parameters using a dataset that was not used in the inversion. Form the results, we discuss the prompting effect of the M6.4 event on the hypocenter of M7.1 event.