Statistical imaging of the deformation over Belgium using multiple geodetic techniques.

One of the challenges of geodesy is to characterize at the sub-millimeter level the vertical deformation of the ground in response to tectonic, anthropogenic, and climatic forcing. Reaching this level of accuracy is crucial to understand the deformation mechanisms acting in Belgium and it contributes to the mitigation of geo-hazards and the operational management of the territory. To address this challenge, the LASUGEO project, aiming at identifying ground deformation caused by groundwater exploitation, makes use of the observations of three independent geodetic techniques, namely: Global Navigation Satellite System (GNSS), Permanent Scatterers Interferometry Synthetic Aperture Radar (PS-InSAR), and repeated Absolute Gravity measurements (AG). Because GNSS, PS-InSAR, and AG provide independent measurements with different spatial and temporal resolutions, they are highly complementary. However, considering that each technique also comes with its own reference frames, accuracy, and source of biases, the optimal combination of these observations requires an appropriate statistical methodology. To estimate the deformation over Belgium, we performed a joint analysis of the GNSS position time series provided by the Nevada Geodetic Laboratory (Blewitt et al., 2018), the PS-InSAR time series processed at Geological Survey of Belgium (Declercq et al., 2021), and the AG measurement carried out by the Royal Observatory of Belgium (Van Camp et al., 2011). Our statistical analysis is divided in three steps: (1) trajectory modelling of each geodetic time series, that is, the model of the predictable motion (e.g., linear trend, periodic deformation, and instrumental discontinuities), (2) surface reconstruction of the subsidence/uplift rates from each technique, and (3) the comparison of the result of the different techniques. For each step, attention is paid to the realistic estimation of the model uncertainties, by accounting for the influence of the time correlated stochastic variability in the geodetic time series (Williams et al. 2003). We propose to describe the algorithms used and results obtained from the trajectory modelling and surface reconstruction of the subsidence/uplift rates. We show that, by combining a large number of observation, we are able to image vertical deformation at the 1.0 mm/yr level over Belgium (see Figure 1 for the GNSS imaging). We also discuss differences between GNSS, AG and PS-InSAR that could highlight the need to calibrate PS-InSAR relative estimates with GNSS and AG geocentric velocities.

Abstract of an Oral Presentation or a Poster

(15) géologie