In the current Earth Observation scenario the Differential Synthetic Aperture Radar Interferometry (DInSAR) technique has reached a key role thanks to its ability to investigate surface displacements affecting large areas of the Earth, with centimeter- to millimeter-level accuracy and rather limited costs, in both natural and anthropogenic hazard scenarios [1]. Originally developed to analyze single deformation episodes such as an earthquake [2] or a volcanic unrest event [3], the DInSAR methods are also capable to investigate the temporal evolution of the surface deformations. Indeed, the so-called advanced DInSAR techniques properly combine the information available from a set of multi-temporal interferograms relevant to an area of interest, in order to compute the corresponding deformation time series [4-5]. Among several advanced DInSAR algorithms, a widely used approach is the one referred to as Small BAseline Subset (SBAS) technique [5] and to its computationally efficient algorithmic solution referred to as Parallel Small BAseline Subset (P-SBAS) technique [6].
In this work, we show the results achieved within the project referred to as DInSAR-3M, funded by the Italian Space Agency (ASI), which is aimed to improve the generation, through advanced DInSAR methodologies, of multi-frequency surface deformation time series and mean velocity maps, spatially and temporally dense, for the multi-scale analysis of natural and anthropogenic phenomena.
In particular, we present several improvements of the available P-SBAS processing chain which were necessary to effectively generate advanced DInSAR products from SLC stripmap SAR image temporal sequences (Level-1A products) acquired by the twin L-band sensors of the Argentinian SAOCOM-1 constellation.
Specifically, we focus in the following on the two steps to which most of the activities have been devoted. The first one allows us to generate the SLC products specifically relevant to the zone to be investigated, referred hereafter to as area of interest (AoI), and the second one, which allow us to improve the quality of the orbital information.
For what concerns the implementation of the AoI SLCs generation, we remark that the SAOCOM-1 L1 images are made available through “slices”, having a typical azimuth extension of about 80/100 km. Accordingly, particularly for large scale DInSAR analysis, they have to be properly merged into a single SLC image relevant to the AoI. This slice-merging operation, which is an ordinary procedure in DInSAR scenarios, is unfortunately not straightforward for the SAOCOM SLC data. Indeed, two sub-steps have been implemented, which we refer as:
- Slice resampling on a common temporal grid;
- Phase shift estimation and compensation.
About the slices resampling on a common temporal grid procedure, it is important to highlight that different slices of the same SAOCOM-1 acquisition are characterized by the same Pulse Repetition Frequency (PRF) but they typically show slightly shifted temporal references. Accordingly, a resampling step is needed to properly align the timing of successive slices to be subsequently fused in a single slice. Moreover, in order to finalize the slice images merging procedure, it is also necessary to carry out a phase shift estimation and compensation step. Indeed, following the temporal resampling of adjacent SLC slices, phase inconsistencies may appear when generating DInSAR interferograms, due to unexpected phase offsets between adjacent slices belonging to the same SAOCOM acquisition (see Fig. 1 of the attached file). To better clarify this issue, in Fig. 1-(c) we show an example of a 300 km azimuth extended differential interferogram over the Piemonte region in Italy. As evident in Fig. 1-(c) and even more in Fig. 1-(d,e,f), the result of the merging procedure is affected by phase jumps, which may have a negative impact on the phase unwrapping procedure and, therefore, on the displacements retrieval operation. Fortunately, the presence of a significant overlap between adjacent slices (see Fig. 1-(a,b)) allows us to easily estimate the existing phase shift, which we can identify in correspondence of the peak of the SLC’s phase difference histogram. In Fig. 1-(g,h) we report the differential interferometric phase and the corresponding interferometric coherence after applying the above discussed phase compensation procedure, which properly accounts for the phase difference between adjacent slices.
Finally, for a high quality interferograms generation, the implementation of a second step was needed. Indeed, the orbital information of the SAOCOM-1 SAR images are often characterized by a low accuracy. Accordingly, if no orbital correction is applied this unavoidably leads to an incorrect estimation of the topographic phase component within the DInSAR interferogram generation process and, therefore, it introduces artefacts in the interferometric phase (that, at the first order, can be represented by a sort of phase ramp) which may significantly degrade the quality of the DInSAR products if no appropriate correction is introduced. Accordingly, in order to improve the quality of the generated DInSAR interferograms, we have implemented an additional step within the P-SBAS processing chain; this follows the rationale of the algorithm described in [8], by properly exploiting the redundancy of the generated interferograms and retrieving an orbit correction for each single SAR acquisition of the exploited dataset.
At the conference time we will present the P-SBAS results achieved by processing multi-temporal SAOCOM-1 image datasets relevant to different hazard scenarios. In particular, we will show the results retrieved for areas affected by slow-moving hydrogeological phenomena (Tuscany region, central Italy), and over volcanic zones (Campi Flegrei Caldera, Mt. Etna and Stromboli volcano, southern Italy), thus highlighting the effectiveness of the implemented new developments of the P-SBAS processing chain.
[1] A. K. Gabriel, R. M. Goldstein, and H. A. Zebker, “Mapping small elevation changes over large areas: Differential interferometry,” J. Geophys. Res., vol. 94, no. B7, pp. 9183–9191, 1989.
[2] G. Peltzer and P. A. Rosen, "Surface displacement of the 17 May 1993 Eureka Valley earthquake observed by SAR interferometry", Sci., vol. 268, no. 5215, pp. 1333-1336, Jun. 1995.
[3] Borgia, A., Lanari, R., Sansosti, E., Tesauro, M., Berardino, P., Fornaro, G., ... & Murray, J. B. (2000). Actively growing anticlines beneath Catania from the distal motion of Mount Etna's decollement measured by SAR interferometry and GPS. Geophysical Research Letters, 27(20), 3409-3412.
[4] A. Ferretti, C. Prati and F. Rocca, "Permanent scatterers in SAR interferometry", IEEE Trans. Geosci. Remote Sens., vol. 39, no. 1, pp. 8-20, Jan. 2001.
[5] P. Berardino, G. Fornaro, R. Lanari and E. Sansosti, "A new algorithm for surface deformation monitoring based on small baseline differential SAR interferograms", IEEE Trans. Geosci. Remote Sens., vol. 40, no. 11, pp. 2375-2383, Nov. 2002.
[6] F. Casu.; S. Elefante; P. Imperatore; I. Zinno; M. Manunta; C. De Luca; R. Lanari, “SBAS-DInSAR Parallel Processing for Deformation Time-Series Computation”. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., vol. 7, 3285–3296, 2014.
[7] Y. Roa, P. Rosell, A. Solarte, L. Euillades, F. Carballo, S. García, P. Euillades,”First assessment of the interferometric capabilities of SAOCOM-1A: New results over the Domuyo Volcano, Neuquén Argentina”, Journal of South American Earth Sciences, Vol. 106, 102882, 2021.
[8] A. Pepe, P. Berardino, M. Bonano, L. D. Euillades, R. Lanari and E. Sansosti, "SBAS-Based Satellite Orbit Correction for the Generation of DInSAR Time-Series: Application to RADARSAT-1 Data," IEEE Trans. Geosci. Remote Sens., vol. 49, no. 12, pp. 5150-5165, Dec. 2011,
