Rupture Model of the M5.8 Pawnee, Oklahoma, Earthquake From Regional and Teleseismic Waveforms

Abstract

The 2016 M5.8 Pawnee, Oklahoma, earthquake is the largest earthquake to have been induced by wastewater disposal. We infer the coseismic slip history from analysis of apparent source time functions and inversion of regional and teleseismic P waveforms, using aftershocks as empirical Green's functions. The earthquake nucleated on the shallow part of the fault, initially rupturing toward the surface, followed shortly thereafter by slip deeper on the fault. Deeper slip occurred below the aftershocks and at greater depths than most induced seismicity in the region, suggesting that small‐ to moderate‐sized earthquakes may not occur on deeper parts of faults in Oklahoma because they are further from failure than shallower fault sections. Comparisons with models of pore pressure perturbations further suggest that the earthquake may have initiated within a region of higher pore pressure perturbation but was not confined to this zone. These observations inform source physics and understanding of maximum magnitudes.

Source Complexity of the 2015 Mw 4.0 Guthrie, Oklahoma Earthquake

“…Along with previous observations of complex rupture processes for moderate earthquakes in Oklahoma (López‐Comino & Cesca, ; Moschetti et al, ; Sun & Hartzell, ), our analysis demonstrates that simple source models such as omega‐square models (e.g., Boatwright, ; Brune, ; Kaneko & Shearer, ; Madariaga, ), widely used to estimate stress drops and radiated energy, are insufficient to explain the source physics of induced earthquakes, even as small as M4. Our results suggest that long‐standing debates concerning earthquake scaling (e.g., Abercrombie, ; Mayeda et al, ) and the difference (or similarity) between induced seismicity and natural seismicity (e.g., Boyd et al, ; Huang et al, ; Ji & Archuleta, ; Sumy et al, ; Wu et al, ) result, at least in part, from limitations of the simple source models assumed (e.g., see Shearer et al, ).…”

“…In the case of induced earthquakes in Oklahoma, small‐to‐moderate earthquakes dominate the catalog (Petersen et al, ). Previous studies show complex rupture processes with multiple slip patches or subevents for the 2011 Mw 5.7 Prague, the 2016 Mw 5.1 Fairview, and the 2016 Mw 5.8 Pawnee earthquakes (López‐Comino & Cesca, ; Moschetti et al, ; Sun & Hartzell, ). Although dense array observations (Sweet et al, ) allow detailed source analysis of relatively small events (e.g., M~2; Fan & McGuire, ), analyses using regular monitoring networks for smaller events in Oklahoma has not yet been conducted.…”

Slip Characteristics of Induced Earthquakes: Insights From the 2015 Mw 4.0 Guthrie, Oklahoma Earthquake

“…Moreover, an examination by Moschetti et al. (2019) of the Pawnee earthquake did not find agreement between modeled pore pressure change along the fault and the location of its slip patches. On the other hand, the non‐induced intraplate 2011 M w 5.8 Mineral Virginia earthquake also has multiple slip patches (Hartzell et al., 2013).…”

“…The finite‐fault slip inversions of the Morgan Hill, Parkfield M6, Joshua Tree, Brawley and Coyote Lake earthquakes are downloaded from the SRCMOD Earthquake Finite Fault Database (http://equake-rc.info/srcmod/; Mai & Thingbaijam, 2014). The Pawnee (Moschetti et al., 2019), and Akita‐Daisen (Yoshida et al., 2020) earthquake slip inversions are available in the respective papers supplemental. The Prague (Sun & Hartzell, 2014), Cushing (Meng et al., 2021), Borrego (Ross et al., 2017), Hida (Ide, 2001), Parkfield (Uchide & Ide, 2010), and Gyeongju (Uchide & Song, 2018) data are obtained directly from the authors.…”

“…Previous investigations of coseismic slip for induced earthquakes have observed multiple phases of slip temporally and spatially. The 2011 Prague earthquake contained multiple slip patches (Sun & Hartzell, 2014), and rupture models of the 2016 Pawnee earthquake showed multiple peaks of slip and moment release (Grandin et al., 2017; Moschetti et al., 2019). Previous studies have shown that the nucleation of these events was affected by prior seismicity and injection, so these two factors could play a role in their rupture processes (Chen et al., 2017; Norbeck & Horne, 2016; Pennington & Chen, 2017; Sumy et al., 2014).…”