This project develops statistical approaches for integrated analysis of natural and induces seismicity. Underground fluid injection has induced an increasing number of seismic events, raising concerns among stakeholders.
The quantitative assessment of seismic hazard follows the paradigm of Bayesian modeling, in which the prior uncertainty can be updated upon acquisition of new measures, followed by posterior confidence intervals for seismic hazards.
We demonstrate the effectiveness of the developed approaches by analyzing the Oklahoma earthquake catalog.
Spatial map of Oklahoma, with epicenters and magnitude ranges of the seismic events.
Empirical and inferred seismic event rates in Oklahoma, from 2000 to 2014.
Source of funding
National Energy Technology Laboratory’s Regional University Alliance (NETL-RUA), RES contract DE-FE0004000, Task 200 (PI: Mitchell Small).
Jan-Dec 2016: CMU Scott Institute Seed Grants: “Hazard and risk modeling of induced seismicity” with M. Small, ($72,720).
Jan-Dec 2019: CMU Scott Institute Seed Grants: “Forecasting induced seismicity by integrating physics-based models with machine learning” (PI), with K. Dayal and A. Singh (Co-PIs), [$75,000].
Wang, P., Small, M.J., Harbert, W., Pozzi, M. "A Bayesian approach for assessing seismic transitions associated with wastewater injections," Bulletin of the Seismological Society of America (2016: pre-issued online).
Wang, P., Pozzi, M., Small, M. J., Harbert, W. "Statistical method for real-time detection of changes in seismic risk at deep-well injection sites," Bulletin of the Seismological Society of America, 105:2852-2862, doi:10.1785/0120150038 (2015).