Application of GK17 Ground-Motion Model to Preliminary Processed Turkish Ground-Motion Recordings Dataset and GK Model Adjustment to the Turkish Environment by Developing Partially Nonergodic Model

Abstract

Abstract The global ergodic GK17 (Graizer, 2018) ground-motion model developed using the Next Generation Attenuation-West2 database (Ancheta et al., 2014) for the active crustal regions (ACRs) was applied to the dataset of recordings from the two largest moment magnitude Mw 7.8 and 7.5 earthquakes in Türkiye (Buckreis et al., 2023a). The GK17 model demonstrates acceptable performance while mostly underpredicting spectral accelerations (SAs) at near-fault up to ∼100 km and far-field more than ∼400 km rupture distances for periods T < 1 s. The GK17 model was modified by applying additional Rrup distance and VS30 residuals corrections, creating an updated GK nonergodic model tuned for Türkiye called the GK_T model. After varying anelastic attenuation of SAs QSA, it was concluded that the average for the ACRs value of QSA=120 produces the best results. Comparisons of the model site amplification functions (SAFs) show that the GK_T SAFs are lower compared to the GK17 model. The Türkiye-specific GK_T partially nonergodic model shows better agreement with the recorded data than the ergodic GK17 model, especially at short periods and short rupture distances. As expected, application of the Türkiye-specific model results in a lower sigma. It is also lower than the within-event sigma of the GK17 model, except for the period range of 1.5 < T < 3 s. Improvements to the GK_T model can be expected with an addition of deep sediment (basin depth) information to the current dataset and other earthquakes in the region.