Detection and Characterization of Ground Surface Deformations from Wavelet Analysis of GNSS Time Series
Résumé
Ground surface is subjected to deformations from various origins, acting on very different time scales: fast break during an earthquake, diurnal/semi-diurnal/monthly and other periodic deformations related to tides, seasonal loads related to hydrological events etc… The amplitude of these deformations is also very variable, from few millimeters up to few meters. GNSS is one of the geodetic tools that can be used to quantify these deformations. The goal is then to identify and to differentiate the various deformation mechanisms involved in a particular region from GNSS time series. Fourier transform analysis allows the identification of periodic phenomena in GNSS time series if the amplitude of the studied deformation is greater than the GNSS measurement precision. The characterization of non-stationary signals is more difficult and requires specific analysis. Different methods of singularity detection exist, which have recently been adapted to GNSS time series processing. This study proposes a method based on the wavelet transform. This time-frequency analysis is particularly adapted to detect and locate in time the information on the various frequencies contained in a non-stationary signals. The wavelet transform is also robust in the presence of noise. Theoretical aspects are well known, but they require adaptation to the case of GNSS time series, in particular for the choice of the best suited analyzing wavelet for the detection of deformation processes. This method has been applied to the French Armorican massif where many GNSS data are available. In this region, several phenomena are at the origin of ground surface deformations: 1- tides, with tidal amplitudes among the largest in the world. They generate deformations related to both ocean tide loading and earth tides, with amplitudes up to ten centimeters . 2- diffuse seismicity mainly related to regional tectonics. In this region, regular seismic events of magnitude up to 4 can occur that are likely to be recorded by GNSS receivers, even if the deformation associated with these events is small. 3- more locally, changes in hydrological loads, seasonal or anthropogenic, within confined aquifers located in the fractured basement produce ground surface deformations up to few centimeters.