Modeling Causal Links Between Subsurface Water Injection and Induced Earthquakes in Oklahoma

Keywords: Natural Disaster, Earthquake, Environmental Justice, Oklahoma, Machine Learning, Causality, Space-time
Session Type: Virtual Paper Abstract
Day: Saturday
Session Start / End Time: 2/26/2022 02:00 PM (Eastern Time (US & Canada)) - 2/26/2022 03:20 PM (Eastern Time (US & Canada))
Room: Virtual 20

Abstract

Modeling Causal Links Between Subsurface Water Injection and Induced Earthquakes in Oklahoma

Induced earthquakes are natural disasters linked to subsurface wastewater injection due to an enhanced hydrocarbon extraction technique. These artificially induced natural disasters threaten local societies as they caused approximately one thousand earthquakes with small magnitudes, on average 3 Mw, in the span of a year in 2016. Despite empirical evidence on the relationship between increased subsurface pressure and induced earthquakes, a knowledge gap exists due to uncertainty pertinent to subsurface fluid flow, lithology, and the built environment's impact on seismic activity. This work proposes a spatio-temporal analysis of propensity associated with water injection and induced earthquakes. Potential causal links between far-field water injection and property damage in Oklahoma are defined for varying space-time neighborhoods. Confounding factors that cause disproportionate exposure to induced earthquakes are also modeled. Results indicate demographics and local lithology serve as significant confounding factors in exposure to induced earthquakes.