A regionalized partially nonergodic ground-motion data driven model for low to moderate seismicity areas: using RESIF-RAP, ESM, RESORCE and NGA-West2 data

The aim of this work is to develop a regionalized partially nonergodic data driven ground-motion model (R-GMM) that provides functional forms (ffs) for each of the world’s 13 regions. The ffs are derived from machine learning (ML) of a given dataset drawn from four databases: namely RESIF-RAP, ESM, RESORCE and NGA-West2. The ML is performed by the neural network approach whose explanatory parameters are the moment magnitude M_W, Joyner–Boore distance R_JB, average shear wave velocity in the first 30 m V_S30, nature of V_S30: (measured or estimated) and the focal depth. The model thus established, estimates the ground motion intensity measures IMs. These IMs are represented by the peak ground acceleration (PGA) and the peak ground velocity (PGV respectively, as well as the 13-period acceleration pseudo-spectra from 0.04 to 4.00 s PSA for a damping of 5%. The 13 regions subject of this study are distinguished by their epistemic uncertainties. The aleatory variability is considered as heteroscedastic depending on the M_W and the R_JB. The consideration both of R-GMM, the heteroscedasticity and the ML approach leads to a significant reduction of the aleatory variability. Furthermore, using strong-motion data enabled us to predict the strong-motion intensity measures in regions that were previously considered to be at lower and moderate seismic risk, as is the case in metropolitan France.