Using the soil subsidence from satellite Persistent Scatterer InSAR (PSI) to estimate site-effect parameters in high spatial resolution, case of the sedimentary basin of Grenoble (French Alps).

Following previous studies that identified a potential correlation of the subsidence rates of the soils with the sediment thicknesses (H_bed), or with the resonance periods of the soil columns (T₀), we want to verify this correlation in the urbanised valley of Grenoble (French Alps) prone to significant lithological site effect. We take advantage of the high level of geophysical and geotechnical characterisation of this basin to develop a new strategy to estimate site-effect parameters based on subsidence rates measured by satellite Persistent Scatterer InSAR (PSI). Since May 2022, the European Ground Motion Service (EGMS) delivers open-access maps of subsidence rates in high spatial resolution (up to one point every tens of meters) and releases a new dataset every year. Our study aims in particular at testing the feasibility with EGMS data to correlate the soil subsidence rates with the site-effect parameters in the Grenoble basin.

A strong advantage of the subsidence-rate data lies in their high spatial resolution, that, for example, we could exploit in microzonation studies. However, the subsidence-rate data also presents two mains drawbacks. First, the measurement points of subsidence rate do not all have the same quality, due to uncertainties in the PSI method employed by EGMS. Second, when comparing different EGMS updates, the subsidence rates vary from one release to another because of changes in the data referencing applied by EGMS, which causes difficulties to define a site-effect model independent of the EGMS updates. To overcome these two drawbacks, we define a specific protocol to filter and process the subsidence rates in order to obtain a single model of prediction for each site parameter studied, and independent of the chosen EGMS release.

We show that in the Grenoble basin and outside of anthropogenic sources such as water pumping, the subsidence rates are only correlated with site parameters sensitive to properties of the whole soil columns (sediment thickness H_bed and fundamental resonance period T₀) and not with surface parameters (V_S30, geological facies at the near surface, building weights). This suggests that the subsidence rates as measured by satellite PSI are mostly caused by the compaction of the stiff and thick sediments due to their own weight. We finally conclude by quantifying the accuracy and the reliability of the obtained prediction models, in order to assess the input of the PSI satellite data for high-resolution site-effect assessment in urbanized valleys.