Exploring the relationship between seismic noise signals and modeled river flow data: A case study from Sicily, Italy

In this work, we analyze and integrate the seismic data recorded by one seismic station installed close to the Cimia River (Gela, Sicily), the rainfall data recorded by the meteorological stations installed within the Gela hydrological basin, and the flow data estimated using a rainfall-runoff model. We take into account two meteorological events characterized by different features in terms of duration, total rainfall, rainfall intensity, and flow rate: an impulsive short-term thunderstorm that occurred on 13th October 2022, and a long-duration perturbation that took place in February 2023 during the quasi-Medicane Helios. In particular, we explore the relationships between seismic noise (in terms of spectral content, root mean square amplitude time series, and polarization features) and meteorological and hydrological parameters (specifically, rain rate, total rainfall, and net rainfall within the basin, as well as the flow rate at a relevant cross-section along the river). The results show that only the long-duration event (i.e., Helios) leaves a mark on the seismic signals and that the time lag obtained between the root mean square amplitude and rain rate time series matches with the catchment's concentration time. These findings show a clear correlation between seismic noise and hydrological parameters, allowing us to retrieve information about flood events using seismic data when traditional river discharge data are unavailable.