A Ground Motion Model from GNSS Peak Ground Displacement

Abstract:

Earthquake ground motion models (GMM) are used to inform a range of earth science and engineering applications, including seismic hazard evaluations, loss estimates, and seismic design standards. A typical GMM is characterized by simple metrics describing the earthquake source (e.g. magnitude and mechanism), observation distance, and various site terms (e.g. Vs30). GMMs are used to inform United States Geological Survey earthquake response products such as ShakeMap, a ground shaking model that allows rapid assessment of the impact of an earthquake. Most often, GMMs are derived from broadband seismometer and strong-motion accelerometer observations, yet these traditional seismic instruments saturate at high magnitude, leading to inaccurate recordings of the low-frequency ground motions. The integration of geodetic data sources, particularly for characterizing the unsaturated ground motion of large-magnitude events, has proved valuable as a complement to traditional seismic approaches and led to the development of a GMM based on peak ground displacement (PGD) estimated from high-rate Global Navigation Satellite Systems (GNSS) data. The original point-source formulation (Crowell et al., 2013) has been updated with newly acquired high-rate GNSS observations (Melgar et al., 2015; Ruhl et al., 2018) and used to accurately estimate the magnitudes of recent significant events (Hodgkinson et al., 2020). Yet, residuals between observed PGDs and those predicted from the current GMM reveal spatial patterns that indicate the importance of fault dimensions, directivity, and radiation pattern effects on resulting ground motions. We present an updated GMM for moderate to large magnitude earthquakes (Mw6+) based on GNSS-estimated PGD that more accurately accounts for azimuthal and distance dependence. We evaluate how the original GMM can be biased by a suboptimal distribution of stations and demonstrate the improved accuracy from our new ground motion model.