PSI - Issue 78

Fadel Ramadan et al. / Procedia Structural Integrity 78 (2026) 859–866

860

1. Introduction In old-generation design codes, the vertical ground motion is often neglected, despite the recent observations demonstrates its relevance at short periods and in near-source conditions, where it can often exceed the horizontal component (Ramadan et al., 2021). Moreover, the seismological community is gradually shifting toward the use of Fourier spectra, which better represent the energy content of seismic waveforms (Van Houtte et al. 2015; Bora et al., 2015; Sgobba et al., 2023). In addition, the development of FAS model play a key role in ground motion simulation and structural demand assessment, enabling the generation of synthetic accelerograms and providing a direct frequency-based evaluation of seismic input. This study provides an overview of the characteristics of vertical component of ground motion in terms of both spectral acceleration and Fourier amplitude spectra, focusing on Italian shallow crustal events. The study on spectral acceleration and peak parameters was already discussed in Ramadan et al. (2021) paper and is based on two datasets ( one regional of medium to large events and one global of strong events ) and includes the development of a vertical-to-horizontal (VH) model for Italy, adjusted to account for near-source effects. In addition, a Fourier-based ground motion model for the vertical component is developed in this paper for the first time and the predictions are compared with the corresponding horizontal ITA18 GMM available in the literature (Lanzano et al., 2022). Finally, an application example using fully consistent horizontal and vertical target spectra is presented to demonstrate the practical applicability of the VH model for the spectrum-compatible selection of recorded ground motions. 2. Datasets In the GMMs calibration, two datasets were employed: 1. The ITA18 dataset (Lanzano et al. 2022b; https://shake.mi.ingv.it/ita18-flatfile/) was originally compiled to support the development of the horizontal ground motion model proposed by Lanzano et al. (2019) for shallow active crustal regions in Italy. It includes 5,778 strong-motion recordings from 156 seismic events, approximately 90% of which are mainly associated with Italian seismicity. The moment magnitudes (Mw) range from 3.5 to 8.0, and recordings were obtained from 1,684 stations, with Joyner–Boore distances (RJB) extending up to 200 km (Figure 2.1). Events are classified by focal mechanism as normal faulting (NF, 47%), thrust faulting (TF, 28%), and strike-slip (SS, 25%). For further details on the dataset and its processing, the reader is referred to Lanzano et al. (2019) and Ramadan et al. (2021); 2. The NESS1.0 dataset (Pacor et al. 2021), by contrast, is a global archive of strong-motion data that includes only high-quality recordings acquired under near-source conditions (see http://ness.mi.ingv.it/). It comprises approximately 800 three-component waveforms recorded by about 700 accelerometric stations, generated by 74 crustal earthquakes with Mw ≥ 5.5 and RJB distances up to 140 km (Fig. 2.1). 3 . VH ITA18 – NESS For PGA, PGV and SA, it was preferred to calibrate a model not directly for the vertical component, but for the spectral ratio between the vertical and horizontal components (VH), to scale the horizontal predictions available from the ITA18 model of Lanzano et al. (2019). The use of spectral ratio VH instead of direct calibration of the vertical component is preferred for seismic design purposes because it allows to perform the PSHA (probabilistic seismic hazard assessment) for both horizontal and vertical components, avoiding the treatment of their correlation. Since the model is already published, we report here only the strategy followed for model calibration, while details on functional form, coefficients and standard deviations are available in the work of Ramadan et al. (2021). The derivation of the final model is done in two steps: 1. Calibration of V/H model for shallow crustal earthquakes in Italy as a function of magnitude, distance (Joyner-boore and Rupture distances), VS30 and focal mechanism, based on the ITA18 dataset, briefly called VH-ITA18;

2