Stress drop and ground-motion source comparison of the July 2019 Ridgecrest earthquake sequence: A community validation study

Abstract:

We estimate stress drops for the July 2019 Ridgecrest, California, earthquake sequence using several approaches, including spectral (based on a Brune source representation) and ground-motion based methods, and compare those estimates to each other as well as to stress drops measured in other studies in order to understand sources of uncertainty and variability. We use a large instrumental dataset from the earthquake sequence, including the M7.1 and M6.4 events and ~150 aftershocks of M3.5+. The methods that we use to estimate stress drop are as follows: First, during the processing of these records, Brune stress drops are fit to each of the three components of each record. We take these to be an initial estimate of stress drop for these events, representing the largest plausible uncertainty, without correction for path or site effects. Secondly, we perform an empirical Green’s function deconvolution to robustly estimate the earthquake corner frequency. This method negates source and path effects through use of co-located records. Thirdly, we calculate the Arias intensity stress drop for the earthquakes, using 95% Arias intensity ground motion and duration tabulated during the processing. We also perform a mixed-effects analysis to partition total high-frequency ground motion residuals relative to a ground motion prediction equation into source, site and path components. We also solicit stress drop estimates from the community for these same events. We find a high degree of correlation between the stress drops estimates and our high-frequency ground-motion event terms, indicating that the source residuals reflect physical phenomena, and that the stress drops reflect the genesis of high-frequency ground motion. Cross-validation between these approaches allows us to understand the relative contributions of the aleatory and epistemic components of uncertainty and how these estimated stress drops reflect true source processes. We specifically consider what uncertainty is introduced when path and site effects are not properly considered. Finally, we encourage the community to contribute stress drop or source characterization results from the Ridgecrest sequence to our validation study.