Identifying potentially induced seismicity and assessing statistical significance in Oklahoma and California

We develop a statistical method for identifying induced seismicity from large data sets and apply the method to decades of wastewater disposal and seismicity data in California and Oklahoma. The study regions are divided into grid blocks. We use a longitudinal study design, seeking associations between seismicity and wastewater injection volume along time series within each grid block. In each grid block, we find the maximum likelihood estimate for a model parameter that relates induced seismicity hazard to total volume of wastewater injected each year. To assess significance, we compute likelihood ratio test statistics in each grid block and each state, California and Oklahoma. Resampling with permutation and random temporal offset of injection data is used to estimate p values from the likelihood ratio statistics. We focus on assessing whether observed associations between injection and seismicity occur more often than would be expected by chance; we do not attempt to quantify the overall incidence of induced seismicity. The study is designed so that, under reasonable assumptions, the associations can be formally interpreted as demonstrating causality. Wastewater disposal is associated with other activities that can induce seismicity, such as reservoir depletion. Therefore, our results should be interpreted as finding seismicity induced by wastewater disposal and all other associated activities. In Oklahoma, the analysis finds with extremely high confidence that seismicity associated with wastewater disposal has occurred. In California, the analysis finds moderate evidence that seismicity associated with wastewater disposal has occurred, but the result is not strong enough to be conclusive.