Erratum to: Stochastic simulation of ground motions based on NGA-West2 strong motion records

The original version of this article contained faults in Conclusions section. The corrected version appears in this erratum.

6.   Conclusions

A stochastic method for generating synthetic ground motions based on a given set of earthquake and site characteristics is introduced and utilized to predict ground shaking level of scenario earthquakes in the WUS region. The method is developed based on a modulated, filtered white-noise process and characterizes the temporal and spectral nonstationarities in terms of six parameters that desicribe the evolving intensity, predominant frequency, and bandwidth of the ground acceleration process. Six parameters of the stochastic record-based approach that model the time- and frequency-domain nonstationarity of ground motion are determined based on a NGA-WEST2 database with 2,571 pairs of horizontal ground motion records with magnitude range of 4.63 to 7.51 and distance range of 0.21 km to 299.35 km. The six parameters are then assigned probability density functions from fitting the frequency histogram. Using the assigned PDF, model parameters are transformed to the standard normal space where a regression analysis method considering randomeffects is used to represent the intra- and inter-event variability, and the correlations among the transformed model parameters are calculated. Based on the prediction equations and correlations obtained, a sets of random realizations of the model parameters are generated. Using each set of model parameters, one realization of groundmotion can be generated. Stochastic ground motions for scenario earthquakes for M5, M6, and M7 at 16 distances are generated, and the PGA, PSA at 0.2, 1.0, and 2.0 s are calculated for all realizations. The median of the synthetic GMI is then compared with the existing NGA-West2 GMMs. Overall, the range of all realizations can cover the estimates of the five NGA-West2 GMMs. At long periods (T≥1 s) for all scenario magnitudes, the median over all realizations is close to the 2014 NGA-West2 GMMs, which indicating the validity of the prediction model of this study.

Figure  7.  Median (black solid lines), median ±σ (black dashed lines), and range of 50 realizations (green shaded area) of PGA and corresponding estimates from the five NGA-West2 GMMs for selected M, rupture distances, and v_S30=670 m/s

DownLoad: Full-Size Img PowerPoint

Figure  8.  Median (black solid lines), median ±σ (black dashed lines), and range of 50 realizations (green shaded area) of PSA at 0.2 s and corresponding estimates from the five NGA-West2 GMMs for selected M, rupture distances, and v_S30=670 m/s.

DownLoad: Full-Size Img PowerPoint

Figure  9.  Median (black solid lines), median ±σ (black dashed lines), and range of 50 realizations (green shaded area) of PSA at 1 s and corresponding estimates from the five NGA-West2 GMMs for selected M, rupture distances, and v_S30=670 m/s

DownLoad: Full-Size Img PowerPoint

Figure  10.  Median (black solid lines), median ±σ (black dashed lines), and range of 50 realizations (green shaded area) of PSA at 2 s and corresponding estimates from the five NGA-West2 GMMs for selected M, rupture distances, and v_S30=670 m/s

DownLoad: Full-Size Img PowerPoint