Applying improved spectral analysis to an induced earthquake sequence in Oklahoma and implications on earthquake triggering

Xiaowei Chen, & Rachel E. Abercrombie

Published December 10, 2018, SCEC Contribution #9042

Earthquake stress drop is an important parameter in ground motion prediction and earthquake source physics. However, it is notoriously known for its uncertainty. We develop an improved spectral analysis based on stacking. This method solves for an empirical correction spectrum (ECS) with stacked spectra across multiple magnitude bins. It is similar to the stacking method in Shearer et al. (2006) and Trugman and Shearer (2016), but it does not make any assumptions about the intrinsic scaling relationship of stress drop. Carefully designed synthetic test suggests that this method significantly improves the stability and reduces uncertainty of stress drop estimates with a small dataset. We apply this method to a well-recorded earthquake sequence in Oklahoma near Guthrie with over 700 earthquakes from M2 to M4, and carefully examine the influence of wave-type (P wave, S wave, and S-wave with coda), signal-to-noise ratio, number of stations recording each earthquake, and frequency band used for spectral fitting. P-wave produces the most significant scaling relationship, while S-wave and S-wave (coda) only result with minor scaling factor. Spectral fitting with wider frequency range significantly reduces the standard deviation of stress drops. We then compare with spectral ratio analysis based on individual event pairs. The results suggest that data quality control during spectral ratio analysis has strong impact on the resulted stress drop estimates. We further evaluate the role of spectral complexity (i.e., deviation from Brune-type source model) on the consistency in stress drop measurement. The improved results suggest that this induced earthquake sequence initiates with lower stress drop near structural bend, and large earthquakes tend to occur within high stress drop patches. The spatial pattern implies that fault strength variations influence the spatiotemporal evolution of earthquake sequence.

Chen, X., & Abercrombie, R. E. (2018, 12). Applying improved spectral analysis to an induced earthquake sequence in Oklahoma and implications on earthquake triggering. Oral Presentation at American Geophysical Union.