Fig. 4
From: Physics-based forecasting of man-made earthquake hazards in Oklahoma and Kansas
Forecasted seismicity rate (M ≥ 3) and magnitude exceedance probability in the complete study area. a Observed earthquake rate (solid green line), and forecasted seismicity rates resulting from SI models calibrated through different temporal endpoints (Dec 2011–Dec 2017) (coloured dashed lines). Changing the temporal endpoint of the calibration has only a minor effect on forecasted seismicity rates. A recently published model37, based on rate and state friction, significantly under-predicts the decrease of the earthquake rate caused by reduced injection (dotted grey line). b Forecasted annual exceedance probability of magnitude M in the complete study area. The forecasted probability of potentially damaging earthquakes is significantly enhanced compared to the tectonic background and is the highest in 2015 and 2016. Background probabilities are adopted from Langenbruch and Zoback, 20167. In response to reduced injection rates magnitude exceedance probabilities are decreasing, but still above the tectonic background level through 2020. The overall injection-induced pressure increase from 2009 to 2020 has a potential of 23% to trigger a damaging M ≥ 6 earthquake (6% of M ≥ 6.5). Uncertainties (shaded areas) represent b-value computations (see Supplementary Fig. 7)
