Physics-Informed Risk Analysis Of Anthropogenic Seismicity

Since 2001, after decades of a steady rate of magnitude ≥3 earthquakes in theUnited States, the annual number of earthquakes has increased exponentially fromapproximately 20 events per year in 2001 to up to 188 events per year in 2011.This increase is suspected to be human-induced. Modern physics tools, such asBayesian Data Analysis, can elucidate processes that trigger seismicity, both an-thropogenic and otherwise. This study introduces analytical methods and examinesanthropogenic processes, such as wastewater and CO2 injections, which can triggerseismicity. Statistical modeling of fluid injection and extraction has been enabledusing Bayesian inference. The San Andreas Fault (SAF) and nearby lake fillings(e.g., Salton Sea) are analyzed as a complex system of interest. Results confirm thatBayesian inference improves fault parameter estimation, aiding short-term seismicforecasting. Implementing physics-informed data assimilation technologies discussedherein is recommended as a policy strategy.