A Referenced Empirical Ground‐Motion Prediction Model for Peak and Permanent Displacement Based on the Near‐Fault Database

ABSTRACT The development of a suitable ground‐motion prediction model (GMPM) for peak ground displacement (PGD) and permanent displacement (PD) at near‐fault sites is crucial for conducting probabilistic fault displacement hazard analysis (PFDHA) and selecting near‐fault ground motion records. The referenced empirical approach is a highly effective method for developing GMPMs in regions with limited availability of strong ground‐motion data. This study utilizes the referenced empirical approach and employs a regression analysis with a random effect model. It makes use of the global NEar‐Source Strong‐motion (NESS2.0) database to develop horizontal and vertical GMPMs for PGD and PD at near‐fault sites, taking into consideration source characteristics, path effects, site conditions, fault effects, and hanging‐wall influences. The credibility and validity of the developed model were verified through residual analysis and comparison with other developed models. Furthermore, the model's predictive performance was assessed using independent earthquake events to substantiate its reliability. The findings suggest that the developed model more accurately captures the attenuation characteristics of both horizontal and vertical PGD and PD in near‐fault regions, demonstrating strong generalization capability and robustness. The findings offer valuable reference for PFDHA and the selection of near‐field horizontal and vertical ground motion records.