Tool for a multirisk management of landslide displacement effects through geophysical-remote sensing approaches

Multi-risk management of complex landslides events requires a solid understanding of the potential displacement effects through adoption of appropriate investigating methodologies (Mangifesta et al., 2024). In such contexts, the integrated geophysical-remote sensing techniques emerged as a powerful tool (Hussain et al. 2022). In this study, we present an integrated approach that combines passive seismic and remote sensing techniques to monitor the displacement effect of landslide at San Vito Romano (RM). The general objective is to develop a quantitative tool capable of correlating the surficial/ground displacements detected by satellite interferometry with some geophysical proxies of soil mechanics of displaced volume itself (for some details see further in the text). The project also involves the development of an artificial neural network designed to use, as input, landslide mobility indices along with preparatory and/or triggering factors, and to provide, as output, the landslide displacement. Once trained, this network will enable the simulation of potential landslide movements in response to different forcing conditions, generating predictive scenarios useful for risk assessment. The selected case study concerns the landslide of San Vito Romano, which affects the entire urban center, an area long subject to widespread slope instabilities and frequent landslide phenomena (Seitone et al., 2025). The landslide under investigation is classified as active, with a retrogressive behavior, slow displacement, and involving the siliciclastic deposits of the Frosinone Formation (Upper Tortonian). The alternation of marl and sandstone layers represents an important predisposing factor for the slope instability, while soil moisture and rainfall are identified as the primary preparatory factors. The main triggering factors, on the other hand, we considered cumulative rainfall and seismic events. To monitor seismic noise, four velocimeters (SARA Electronic Instruments) were installed: two within the landslide area and two outside of it. The instruments, operating since February 2024, have a frequency of 2.5 Hz and continuously record ambient seismic noise. The changing empirical Green’s function of an active landslide medium can be taken as proxy for monitoring landslide displacement effects obtained with wave scatters as in coda wave interferometry, change in HVSR peak and peak polarization considering the diffuse nature of ambient noise wavefield. However, geophysical results require the separation of unwanted climatological, thermal and other such effects through development of statistical tools. It promotes a deeper understanding of the role of preparatory and triggering factors, such as rainfall and earthquakes, in the evolution of the landslide activity. From the continuous recordings, landslide mobility indices were calculated: the variation in natural period (dT/T), in peak polarization (dP/P), and velocity (dV/V). The first two parameters were obtained using the Python library HVSRpy (Vantassel. 2020), which allows the calculation of the Horizontal-to-Vertical Spectral Ratio (HVSR) over time. The processing was carried out with a time window of 200 s, Konno-Ohmachi smoothing, and the geometric mean for the horizontal components. To calculate the velocity variation, the data were processed with MSNoise, an open-source software for seismic noise analysis (Lecocq et al., 2014). Ground displacement was estimated using satellite interferometric techniques, specifically the DInSAR method with the SBAS (Small Baseline Subset) approach. This technique enables the generation of maps showing the average deformation velocity and the monitoring of temporal displacement evolution at specific points within the study area (CNR-IREA). Specifically, images acquired by the Sentinel-1 satellite were used. From the initial processing of seismic noise, the variation in the natural period over time was obtained by analyzing 11 months of recordings. A variation in the natural period was observed exclusively within the landslide body, alternating between phases with higher peaks and phases with lower peaks. Analysis of the variation in S-wave velocity, revealed a peak change of 2% in August. This result confirms that seismic noise measurements can provide useful information about the ground conditions. In parallel, satellite data from 2022 to 2024 were processed, showing a nearly constant displacement over time. References • Hussain, Y., Schlögel, R., Innocenti, A., Hamza, O., Iannucci, R., Martino, S., & Havenith, H. B. (2022). Review on the geophysical and UAV-based methods applied to landslides. Remote Sensing, 14(18), 4564. • Institute for Environmental Research and Elevation (IREA). SBAS. IREA CNR. http://www.irea.cnr.it/index.php?option=com_k2&view=item&id=97:sbas&Itemid=127. • Lecocq, T., Caudron, C., & Brenguier, F. (2014). MSNoise, a python package for monitoring seismic velocity changes using ambient seismic noise. Seismological Research Letters, 85(3), 715-726. • Mangifesta, M., Aringoli, D., Pambianchi, G., Giannini, L. M., Scalella, G., & Sciarra, N. (2024). A Methodologic Approach to Study Large and Complex Landslides: An Application in Central Apennines. Geosciences, 14(10), 272. • Seitone, F., et al. (2025). Comparative analysis of geological and seismic microtremors models of a large translational landslide: The case of San Vito Romano (Central Italy). Engineering Geology, 107934. • Vantassel, J. (2020). jpvantassel/hvsrpy: latest (Concept). Zenodo. http://doi.org/10.5281/zenodo.3666956.