GNGTS 2022 - Atti del 40° Convegno Nazionale

GNGTS 2022 Sessione 1.3 127 PERFORMANCE OF EMPIRICAL AND MODEL DERIVED ATMOSPHERE CORRECTIONS OF INSAR TIME SERIES M. Bartola 1 , C. Braitenberg 1 , Y. Fialko 2 1 Department of Mathematics and Geoscience, University of Trieste, Trieste, Italy 2 Institute of Geophysics and Planetary Physics, UC San Diego, San Diego, California The reliability of differential interferometry SAR time series depends on the robustness in the calculated Line of sight (LOS) ground velocity. Orbital errors and atmospheric noise are more limiting for small-amplitude and low-rate signals, compared to large-amplitude rapid displacements caused by e.g. earthquakes or volcanic unrest. Orbital errors are reduced given precise orbital information while atmospheric noise is still a considerable challenge. Methods used to mitigate the atmospheric noise can be divided into those using SAR data, and external datasets. Each method has its own strengths and weaknesses (Li et al. , 2022), but none of them can remove entirely the atmospheric contribution that can consist of two parts: the stratified atmosphere and the turbulent mixing. In this studyweuse twodifferentmethods, theCommon Scene Stacking (CSS) (Tymofyeyeva and Fialko, 2015) and the Global Atmospheric Correction Online Service (GACOS) (Yu et al. , 2017; Yu et al. , 2018a; Yu et al. , 2018b). The CSS is an iterative method for the estimation of the atmospheric contribution based on the fact that interferograms sharing a common scene contain the same delay contributions. GACOS is model based and integrates the HRES- ECMWF and the GPS to generate high resolution maps of phase delays due to atmospheric moisture. We used these methods on a rugged area straddling the Friuli Venezia Giulia and Veneto regions. In the area of interest, different weather conditions over the year are present, karstic channels characterize the underground in which, in combination with strong flood events, the water is over- pressurized and triggers surface deformation, and furthermore there are active seismic faults. Fig. 1 - Comparison between the original displacement time series, the atmospheric corrections and GPS. The GPS line is the time series of the GPS CANV with a moving average filter of 5 days. The GPS data refer to the Nevada Geodetic Laboratory. A): Original time series, CSS correction residuals and GPS; B): GACOS correction residual, CSS correction residuals after the application of GACOS.