Examining Spatial Variations in California Earthquake Dynamics Using a High- to Low-Frequency Spectral Ratio

Earthquakes radiate varying amounts of high-frequency energy, reflecting differences in source dynamics, propagation and attenuation, and near-station site effects. Building on results from the Ridgecrest Stress Drop Validation Exercise, which revealed large uncertainties in stress drop estimates, we apply a new data-driven approach to measure differences in seismic radiation directly from the observations. We compute a ratio of high-frequency to low-frequency amplitude from individual spectra. By empirically correcting path, station, and magnitude effects, we measure the relative amount of high-frequency radiation in earthquakes compared to nearby calibration events. Our results show that this normalized spectra ratio is correlated with earthquake stress-drop estimates from previous studies of the 2019 Ridgecrest aftershock sequence, suggesting that it can be used to infer differences in source dynamics. We expand this approach to examine over 30 years of California seismicity, including hundreds of thousands of earthquakes, and identify spatial differences in high-frequency radiation and inferred source mechanics. Spatial variations in high-frequency radiation occur over both local and regional scales, and are visible within high-seismicity regions, including Parkfield, Geysers, and Mammoth Lakes. Our large-scale study of spatial differences in earthquake radiation offers an observation-based alternative to traditional stress-drop estimation methods, and we will discuss its implications for fault properties and strong ground motion prediction in California.