| Popis: |
One of the main sources of noise affecting the Differential Synthetic Aperture Radar Interferometry (DInSAR) products is represented by the Atmospheric Phase Screen (APS) signals which are caused by the temporal and spatial variability of atmospheric conditions between the interferometric radar image pairs. Accordingly, it can be challenging to discriminate atmospheric phase delay signals from the deformation ones, and this is particularly difficult in volcanic areas which are often characterized by the presence of both significant topography and displacements.In this work we present an extensive analysis based on the exploitation of the ECMWF ERA-5 data [1] to filter out the APS contribution from DInSAR products. In particular, we focus on the impact of the ERA-5 corrections on DInSAR deformation time series, as well as on single interferograms. The generation of the exploited DInSAR products is performed through the P-SBAS advanced DInSAR approach [2]. To filter out the APS signal component from the retrieved deformations we have developed an automating processing chain that exploits:the PyAPS Python software implementing the approach described in [3], for the APS evaluation; an ad-hoc developed IDL code, for correcting the generated interferograms and deformation time series. The presented experimental analysis has been carried out by taking into account large Sentinel-1 (S-1) datasets acquired both from ascending and descending orbits over several volcanic areas, which are of particular interest for the presence of both significant atmospheric phenomena and remarkable deformations:Etna (Sicily, Italy); La Palma island (Canary, Spain); Stromboli (Sicily, Italy); Mauna Loa (Hawaii, United States). In these sites, the lateral variation of pressure, temperature and humidity, jointly with a topography-correlated component due to the variation of the atmospheric parameters with height, makes the APS interferometric component difficult to be distinguished from the deformation one. This makes the exploitation of auxiliary data crucial.Moreover, an additional analysis has been carried out by considering the first DInSAR results obtained by exploiting the L-band SAR images acquired by the SAOCOM-1 satellites, over the Ischia island (Campania, Italy).Our results show that the ERA-5 based APS correction is capable to effectively filter out, for the considered sites, the atmospheric phase contributions relevant to the typical seasonal oscillations as well as those correlated with topography. As expected, because of the coarse spatial resolution of the input ERA-5 data (30 km on the horizontal grid), it is indeed less effective for removing the small spatial scales (turbulent) component of the atmospheric phase signals, which requires different filtering approaches to be corrected. [1] https://www.ecmwf.int/en/forecasts/datasets/reanalysis-datasets/era5[2] Manunta, M. et al., The Parallel SBAS Approach for Sentinel-1 Interferometric Wide Swath Deformation Time-Series Generation: Algorithm Description and Products Quality Assessment, IEEE Trans. Geosci. Remote Sens., 2019.[3] R. Jolivet et al, "Systematic InSAR tropospheric phase delay corrections from global meteorological reanalysis data," Geophysical Research Letters, vol. 38, no. 17, 2011. |
