Using Simulated Ground Motions to Constrain Near-Source Ground-Motion Prediction Equations in Areas Experiencing Induced Seismicity

Recent seismic activity increases in historically quiescent areas such as Oklahoma, Texas, and Arkansas have spurred the need for investigation into expected ground motions associated with these seismic sources. The neoteric nature of this seismicity increase corresponds to a scarcity of ground-motion recordings within similar to 20 km of earthquakes M-w 3.0 and greater. To aid the effort of constraining near-source ground-motion prediction equations (GMPEs) associated with induced seismicity, we develop a framework for integration of synthetic ground-motion data from simulated earthquakes into the GMPE development process. We demonstrate this framework by developing a GMPE for a target region encompassing north-central Oklahoma and south-central Kansas. We first gather a catalog of recorded ground motions from M-w 3-4 earthquakes that occurred in the target region. Using constraints on the region's material structure, including well log data that provide insight into the characteristics of shallow sedimentary layers, we perform point-source simulations intended to mimic a selection of recorded earthquakes from the target region. Simulated earthquake sources are constrained by available moment tensors and locations. Once we determine that our simulations produce realistic ground motions, we combine recorded and synthetic ground-motion data to produce a composite ground-motion catalog. We use this composite catalog to develop a regionally specific GMPE for our target region. This framework can be exported to other regions where near-source ground-motion data are sparse and can be used to improve constraints on near-source GMPEs, which could directly benefit seismic-hazard estimates.