Partially Nonergodic Empirical Ground‐Motion Models for Predicting Horizontal and Vertical PGV, PGA, and 5% Damped Linear Acceleration Response Spectra Using Data from the Iranian Plateau

Abstract

We present new ground‐motion prediction equations (GMPEs) to estimate horizontal and vertical strong ground motion intensity measures (GMIMs) generated by shallow active crustal earthquakes occurring within the Iranian plateau. To this end, a dataset containing 688 records from 152 earthquakes with moment magnitudes ranging from 4.7 to 7.4 and Joyner–Boore distances up to 250 km has been used. The effects of the local site condition are taken into account using the time‐averaged shear‐wave velocities in the upper 30 m ( V S 30). We decided not to include the style‐of‐faulting term in the final functional form because the total standard deviation is reduced 10% by removing this term from the functional form. We used a nonlinear mixed‐effects regression to determine the coefficients of the functional form and to separate out the between‐event and between‐station standard deviations from the total standard deviation. Significant standard deviation of site‐to‐site variability demonstrates that the ergodic assumption is not able to account for the spatial variability of ground motions. We introduced random‐effects coefficients to capture regional variations between different tectonic regions of the Iranian plateau, such as Alborz and Zagros, in the regression analysis to investigate the effects of regionalization on GMPEs. The results showed that, although the effects of regional variations for considered regions are negligible at close distances, they are significant at longer distances. The complexity and performance of the final functional form is justified by comparing Akaike and Bayesian information criteria values over many trial functional forms. Moreover, the distribution of between‐event, site‐to‐site, and event‐station corrected residuals demonstrates that no trends are evident, implying satisfactory performance of the proposed GMPEs. Therefore, the derived GMPEs can be employed to predict GMIMs and to do seismic‐hazard assessments within the Iranian plateau. [Electronic Supplement:][1] Lists of the events and stations considered in the final dataset and derived coefficients of the functional form. [1]: http://www.bssaonline.org/lookup/suppl/doi:10.1785/0120160205/-/DC1