An Equivalent Point‐Source Stochastic Simulation of the NGA‐West2 Ground‐Motion Prediction Equations

In this study, we use a genetic algorithm to invert horizontal ground‐motion intensity measures (GMIMs) predicted from the empirical Next Generation Attenuation‐West2 (NGA‐West2) ground‐motion prediction equations (GMPEs) to estimate a consistent and correlated set of seismological parameters to use with an equivalent point‐source stochastic model. The GMIMs are peak ground acceleration and pseudospectral acceleration evaluated over a wide range of magnitudes, distances, and frequencies. The inversion is performed for $M3.5–8.0$⁠, $RRUP=1–300km$⁠, $T=0.01–10s$⁠, and National Earthquake Hazard Reduction Program (NEHRP) B/C site conditions. Seismological parameters are obtained as a function of earthquake magnitude. The near‐source geometric spreading was modeled as both magnitude‐ and frequency dependent to fit the empirical predictions. The agreement between the model and empirical predictions over all magnitudes and distances evaluated in this study is generally within 10%, with some local exceptions. The near‐source geometric spreading is consistent with a distance decay of $R−0.8$ to $R−1.3$ at frequencies of $f≤1Hz$ for $M$ ranging from 3.5 to 8. Near 5 Hz, the distance decay is expressed as $R−1.17$⁠, on average at short distances. At larger frequencies, the near‐source distance decay varies from $R−1.0$ to $R−1.25$⁠. This stochastic model can be used for any application that requires a frequency‐domain representation of the NGA‐West2 GMPEs.