Characterization of ground deformation related with hydrological processes in the northwestern Po Plain by InSAR data

Water mass redistributions in the hydrogeological cycle can produce up to centimeter-scale deformation of the Earth's surface, detectable by satellite geodetic techniques such as Global Navigation Satellite System (GNSS) or Interferometric Synthetic Aperture Radar (InSAR)(Burgmann et al., 2023). These hydrologically-related signals are typically associated with two main processes: the elastic loading response to Terrestrial Water Storage (TWS) variations and the poroelastic response associated with Groundwater Storage (GWS) variations in aquifers (Larochelle et al., 2022). In this thesis, I analyzed InSAR deformation data from the European Ground Motion Service (EGMS) in the interval 2018–2022, to characterize climate-related deformation processes in northwestern Italy. This region hosts the Po Valley, a large alluvial plain rich in water resources, extensively exploited for human use. During this period, the region experienced variable precipitation, with severe drought in 2022 (Montanari et al., 2023). To extract dominant deformation signals, I applied decomposition and clustering techniques to the EGMS dataset and compared the results with precipitation records from the ARPA Lombardia network. I found a set of areas, near the mountain foothills, where deformation is strongly correlated with precipitation: uplift of up to 5 mm in wet periods and subsidence of up to 10 mm during drought. This is consistent with a poroelastic response, likely due to localized concentrations of coarse-grained material like gravel and sand. In one of these areas (around the city of Brescia), I modeled and removed the elastic loading contribution using Equivalent Water Height (EWH) variations from GNSS (Pintori et al., 2023). I then compared the residual deformation to piezometric data to estimate the aquifer’s Young’s modulus. I found results consistent with literature values for gravel deposits, though simplifications in the model may introduce errors in the estimate.