A quantitative assessment of the SAOCOM-1 L-band DInSAR time-series retrieved through the P-SBAS approach in natural and anthropogenic hazard scenarios of the Italian territory

Spaceborne Differential Synthetic Aperture Radar Interferometry (DInSAR) represents a well-established technique to accurately retrieving ground surface displacements over large areas of the Earth, in both natural and anthropogenic hazard scenarios, with limited costs and with a centimeter to millimeter accuracy. However, the DInSAR technique retrieval capability may be affected by the so-called “temporal decorrelation phenomena” due to possible temporal changes of the imaged scene electromagnetic response. In this regard, the low-frequency SAR sensors, as those operating at the L-band, characterized by a significantly larger wavelength (~23 cm) with respect to the X-band (~3 cm wavelength) and C-band (~5.6 cm wavelength) ones, are particularly suited to mitigate the above-mentioned decorrelation effects, thanks to their capacity of maintaining the interferometric coherence for a long period. Moreover, these L-band SAR systems also imply considerable robustness with respect to the possible occurrence of phase unwrapping errors. These peculiarities have pushed the worldwide space agencies to invest in the development of L-band spaceborne SAR sensors as, for instance, the NISAR mission, jointly developed by NASA and ISRO, the PALSAR-3 mission of JAXA and the ROSE-L mission developed by ESA, as well as the already operative SAOCOM-1 sensors of CONAE. In this work, we focus on the Argentinean SAOCOM-1 constellation which is composed of two twins, full-polarimetric L-band SAR sensors. This system guarantees, over a large part of Europe (with a priority given to the Italian territory coverage), a systematic, DInSAR-oriented acquisition plan of SAR images in the StripMap mode, with a revisit time varying among 16, 24 and 48 days, in order to avoid coverage gaps. Moreover, we largely exploit the Parallel Small BAseline Subset (P-SBAS) approach, which is an advanced DInSAR method that allows us to effectively and efficiently generate displacement time-series with sub-centimeter accuracy. The capability of the P-SBAS algorithm to retrieve C- and X-band DInSAR time-series, relevant to both natural and anthropogenic hazard scenarios has already been widely demonstrated, as well as its capacity to perform analyses at different spatial resolution scales. Accordingly, we present here the results of the L-band SAOCOM-1 P-SBAS analysis carried out at medium spatial resolution (about 30 m) in different ground deformation scenarios affecting the Italian territory. In particular, the presented results are relevant to a portion of the Tuscany region (central Italy), which is affected by significant landslide phenomena. Moreover, we also consider the volcanic contexts of the Campi Flegrei caldera, Mount Etna and Stromboli island, all located in southern Italy. In this case, we fully benefit from the availability of GNSS measurements to provide a quantitative assessment of the retrieved L-band deformation time-series.

Finally, some SAOCOM-1 results, achieved by applying the full resolution P-SBAS approach over the urban areas of Rome and Naples municipalities, are also presented. Such a full spatial resolution (about 5 m of pixel size) analysis allows us to investigate the potentialities of the L-band data to overcome some of the limitations of the current high resolution X-band SAR systems in urbanized scenarios.