The Magnitude Frequency Distribution of Induced Earthquakes and Its Implications for Crustal Heterogeneity and Hazard

Abstract:

Earthquake activity in the central United States has increased dramatically since 2009, principally driven by injection of wastewater coproduced with oil and gas. The elevation of pore pressure from the collective influence of many disposal wells has created an unintended experiment that probes both the state of stress and architecture of the fluid plumbing and fault systems through the earthquakes it induces. These earthquakes primarily release tectonic stress rather than accommodation stresses from injection. Results to date suggest that the aggregated magnitude-frequency distribution (MFD) of these earthquakes differs from natural tectonic earthquakes in the same region for which the b-value is ~1.0. In Kansas, Oklahoma and Texas alone, more than 1100 earthquakes Mw ≥3 occurred between January 2014 and June 2015 but only 32 were Mw ≥ 4 and none were as large as Mw 5. Why is this so? Either the b-value is high (> 1.5) or the magnitude-frequency distribution (MFD) deviates from log-linear form at large magnitude. Where catalogs from local networks are available, such as in southern Kansas, b–values are normal (~1.0) for small magnitude events (M < 3). The deficit in larger-magnitude events could be an artifact of a short observation period, or could reflect a decreased potential for large earthquakes. According to the prevailing paradigm, injection will induce an earthquake when (1) the pressure change encounters a preexisting fault favorably oriented in the tectonic stress field; and (2) the pore-pressure perturbation at the hypocenter is sufficient to overcome the frictional strength of the fault. Most induced earthquakes occur where the injection pressure has attenuated to a small fraction of the seismic stress drop implying that the nucleation point was highly stressed. The population statistics of faults satisfying (1) could be the cause of this MFD if there are many small faults (dimension < 1 km) and few large ones in a critically stressed crust. Alternatively, the MFD of induced earthquakes may be controlled by small scale stress concentrations in a spatially variable stress field. Resolving the underlying causes of the MFD for induced earthquakes may provide key insights into the hazard posed by induced earthquakes.