Empirical predictive models for fling step and displacement response spectra based on the NESS database

The characterization of the fling-step represents a challenging task due to the shortage of near-source records with permanent tectonic displacement and the limitation in retrieving the fling-amplitude from accelerometric waveforms. In recent years, innovative ground-motion processing techniques have been developed for a more accurate estimation of both fling-displacements and spectral displacements in contrast to traditional bandpass filtering, although their application is still unusual. In this paper, we exploit the newly released dataset of the Near-Source Strong-motion records (NESS2) uniformly processed with the extended BASeline COrrection technique (eBASCO), against which we propose: (1) a new empirically-based ground motion model (GMM) for the prediction of the fling-step, and (2) an adjustment factor of the spectral displacements predicted by a reference GMM to account for the contribution of the fling-step at long periods. Such models are in agreement with observations and existing GMMs, and thus could be advantageously employed in seismic hazard analyses.