Stable stress drop ratio estimation for potentially induced earthquakes in Oklahoma

Stress drop is a key source parameter that influences the earthquake hazard in Oklahoma where potentially injection-induced earthquakes show high activity. Determing if the stress drops of induced earthquake are similar to natural events would provide an important constraint on the hazard. Making stable and accurate stress drop measurements has been a challenging issue because of the strong dependence of stress drop on the corner frequency. Shaw et al. (in this meeting) introduced a new approach for the stable stress drop ratio estimation, using the high- and low-frequency asymptotic level of the spectral ratio that avoids the need to measure corner frequencies. We compared the stress drop ratio of earthquakes in Oklahoma, measured by this new approach, with stress drops measured by the traditional approach based on the spectral fitting.We formed spectral ratios for pairs of co-located events in two clusters to remove path effects; both clusters contain 9 events (distance < 5 km, 2.7 < Mw < 4.2). The spectral ratios between the smaller events and the larger event in the cluster were calculated using the Multi-Window Spectral Ratio method. We analyzed 5.12 seconds after the twice the S wave arrival time (coda wave), and stacked the spectral ratios of all components of 6 USGS stations. Corner frequencies and moment ratio of each event pair were fit with the Boatwright model. Using the derived corner frequencies and the asymptotic level of the spectral ratio, we measured the stress drop ratios of each event in two methods. As a result, the stress drop ratio obtained from asymptotic spectral values ranged from 1 to 2 while the values obtained from the corner frequencies ranged from 1 to 5. This suggests the stress drop ratio based on the spectral asymptotes is an effective way to reduce the error of the estimation, which should greatly improve our ability to characterize true stress drop variability.