Conference Agenda

Near geostationary Synthetic Aperture Radar (SAR) systems represent a promising solution for near-real-time Earth observation, thanks to its capability of ubiquitous and continuous monitoring, addressing stability and health of large infrastructures, urban mapping and planning and risk management in various areas subject to deformations, besides rapid disaster response.

Such systems compensate the huge attenuation due to the distance by long integration time, of several minutes. However, performance remains strongly affected by the random and unknown fast fluctuation of the Atmospheric Phase Screen (APS), which introduces phase distortions mainly driven by tropospheric turbulence.

This reduces the feasibility to C and X band systems, since lower frequencies are highly penalized by the variation of the background ionosphere, whereas higher frequencies are strongly impacted by the spatio-temporal variation of the ionosphere. Furthermore, the ITU regulations in the orbit occupation by geostationary satellite limit the achievable resolution.

This study proposes a different concept, made by two satellites operating in Ku-band and placed in a close along-track formation. In this case it would be possible to achieve a reasonable fine resolution, with compact payload in terms of antenna size, and – at the same time - retrieve the atmosphere, thanks to short temporal baseline interferometry.

The MIMO-2 architecture provides at one time three temporal baselines to enhance the retrieval of the APS, and an increase of a factor four in Signal-to-Noise Ratio (SNR). These along-track interferograms will benefit of a limited decorrelation due to the short time interval, but also the very small baselines that can be achieved in such formation.

In addition, the AT interferometric configuration supports further applications, like ship detection or monitoring of fast changes. On the top of this, one-day revisit interferometric applications will be made possible over the stable targets.

In this work, the feasibility is addressed by a quantitative and thoughtful analysis of the impact and compensation of ionosphere and troposphere.

The performance of such system will be compared with a system with single geostationary satellite operating in X band.

This study was carried out within the Space It Up! project and received funding from the ASI and the MUR –Contract n. 2024-5-E.0 -CUP n. I53D24000060005. This manuscript reflects the views and opinions of the authors only; the funding bodies cannot be held responsible for them.

A geostationary SAR mission such as Hydroterra+ offers unique capabilities for persistent observation of a wide area, enabling continuous monitoring of the same region with daily or even sub-daily interferometric revisit times.

Its equatorial orbital position makes the system particularly well suited for observing sub-tropical to mid-latitude regions, for example from sub-Saharan Africa to the entire Mediterranean basin within our hemisphere. This would provide products like: Water-Vapor, Soil Moisture, Snow Water Equivalent, Deformation and change maps, frequently sampled in the day, to understand underlying physics and to be assimilated into forecasts models. Furthermore, it allows for a flexible acquisition planning where resolution can be traded with coverage, and revisit time, and pointing can be switched over a super-continental access area within minutes.

In terms of revisit time and South–North line of sight, such systems are complementary to polar-orbiting LEO satellites such as Sentinel-1, which achieve frequent revisits at high latitudes and operate with an East–West line of sight. Clearly, the imaging performance obtained from an altitude of approximately 37,000 km and with today technologies in terms of power and antenna width, cannot match that of a LEO SAR operating at typical ranges of 800–1000 km. Nevertheless, a geostationary system benefits from reduced temporal decorrelation thanks to its short revisit intervals. This characteristic can significantly improve interferometric quality over distributed targets and enables interferometric observations of rapidly evolving phenomena that would otherwise be difficult or impossible to monitor, such as snow accumulation and melt cycles or lava flows.

For both GEO and LEO SAR systems, such as Hydroterra+ and Sentinel-1, interferometric performance results from a trade-off among several factors: Noise Equivalent Sigma Nought (NESZ), that accounts for Radio Frequency Interferences, temporal decorrelation, and additional decorrelation sources such as atmospheric turbulence during the integration time. Interferometric quality also depends on the number of looks used for phase estimation, and therefore on the achievable spatial resolution.

In this work, we present a comprehensive model for assessing the interferometric performance of a C-band geostationary SAR. The analysis is based on large-scale datasets, including the Global Seasonal Sentinel-1 Interferometric Coherence and Backscatter Dataset provided by ASF and the WRF-OL meteorological data from CIMA. We also provide a comparative assessment of the expected performance relative to Sentinel-1.

Differential Interferometric Synthetic Aperture Radar (DInSAR) is nowadays a key technique for monitoring surface deformation, achieving centimeter/millimeter-level accuracy. Since the 1990s, large-scale SAR data archives have fostered advances in multi-temporal DInSAR methods, enabling the tracking of natural and anthropogenic deformation phenomena by generating displacement time series. Expanding the SAR sensor availability across different bands (e.g., X, C, and L) further enhances spatial and temporal resolution. However, traditional large satellite missions are expensive to develop and launch. Consequently, a shift is occurring towards modular, flexible solutions based on smaller SAR satellites, driven by ease of design, serial production capabilities, and the ability to deploy entire constellations with a single launch, drastically reducing mission costs. Companies like ICEYE, Capella Space, and Synspective, along with the forthcoming Italian constellation IRIDE (managed by ESA with ASI support), exemplify this trend. In particular, the IRIDE constellation, expected to be completed in 2027, will integrate X-band SAR platforms and optical sensors to deliver advanced monitoring services for the national territory and bolster Civil Protection activities.

Simultaneously, the DInSAR user community increasingly demands more sophisticated and technically demanding interferometric products. Meeting these needs requires shorter revisit times, improved spatial resolution, and diverse acquisition geometries to capture three-dimensional surface deformation.

Therefore, exploring alternative orbital configurations beyond the conventional sun-synchronous orbits (SSO) is crucial. SSOs provide global coverage and simplify energy management through consistent illumination, but they can limit DInSAR coverage performance in mid-latitude regions and poorly capture north-south deformation. Recent advancements in satellite miniaturization and launch logistics now effectively enable consideration of mid-inclination orbits (MIOs), particularly when the Area of Interest is located at medium latitudes and global monitoring is not required. Indeed MIOs offer higher revisit frequencies over target areas and significantly improved sensitivity to north-south ground displacement, enabling more accurate retrieval of 3D deformation.

The present study investigates the orbital configuration and interferometric strategy for the NIMBUS mission of the Italian IRIDE program, consisting of two batches of six X-band small SAR satellites each. These satellites can operate with different SAR acquisition modes, and we select the Stripmap mode as the baseline due to its balance between spatial resolution and coverage (the nominal swath width of each NIMBUS satellite is of about 27.5 km).

The constellation design is guided by two main objectives: (i) ensuring systematic coverage of the Italian territory, and (ii) minimizing the interferometric revisit time. The methodology adopted in this study is based on repeat ground track (RGT) orbits, which are fundamental to interferometric applications. Since each NIMBUS batch consists of six satellites, we select a six-day RGT cycle. Each satellite is assigned a dedicated beam with an average swath width of 27.5 km, designed to fill the spatial gaps between consecutive ground tracks (GTs).

The analysis considers both SSO and MIO options as potential orbital configurations for NIMBUS. We use the same RGT cycle (N=6) and beam assignment strategy for the cases herein reported. In particular, for the two batches, we select two MIO configurations based on the following reasons:

• a 49° right-looking MIO, because it is the minimum orbital inclination angle that guarantees full coverage of the Italian territory in a right-looking acquisition mode exploiting off-nadir angles higher than 20°;
• a 44° left-looking MIO, because it further enhances the angular diversity and DInSAR capabilities of the overall NIMBUS constellation.

This work examines the validity of this design choice and highlights the advantages of MIOs for interferometric applications. In particular, we investigate the coverage capability of the DInSAR performance constellation over the Italian territory. Prior to the NIMBUS launch (first batch by 2026), we can anticipate its performance by analysing real data from commercial satellites operating in 45° MIO orbits. This study utilizes Capella Space satellite data acquired over the Campi Flegrei caldera, an active volcanic area in Southern Italy, processed using the Parallel Small Baseline Subset (P-SBAS) chain to assess MIO ability to recover north-south deformation component. The achieved results provide a valuable precursor study for NIMBUS capabilities and confirm the growing importance of small SAR constellations in advanced DInSAR monitoring.

Public SAR missions such as Sentinel-1 and NISAR have established the foundation for systematic, wide-area ground deformation monitoring. However, applications involving rapid or transient deformation, including fast-moving landslides, short-lived volcanic episodes, or infrastructure failures, demand finer spatial resolution, shorter revisit intervals, and, in some cases, sensitivity to the north-south component of motion that Sun-synchronous polar orbits cannot provide. Commercial New Space SAR constellations are now positioned to fill these observational gaps.
We present approximately two years of repeat-pass InSAR results from Capella's X-band constellation. Starting in mid-2024, the Capella-14 satellite (and later the Capella-13 satellite), were placed in 45-degree (53-degree, resp) mid-inclination orbit with a 2.95-day repeat ground track cycle. Automated orbit maintenance has regularly kept the satellites within the presceibed orbital tube, yielding perpendicular baselines consistently below 500 meters. We characterize the interferometric performance of this system, including coherence as a function of spatial and temporal baseline, and phase stability for both persistent scatterer (PS) and distributed scatterer (DS) targets.
We demonstrate the displacement monitoring capability through several case studies. A multi-year time series (over 120 acquisitions) over the Rosamond, California calibration site provides a controlled assessment of long-term phase stability and decorrelation behavior. We examine the ongoing eruptive activity at Kīlauea volcano (Hawai'i, USA), where episodes of rapid inflation and deflation, often occurring on sub-daily timescales separated by multi-day pauses, are captured by the 3-day revisit and sub-meter resolution. Changes to coherence map clearly show locations of new lava flows. We also demonstrate high-quality results for the Portuguese Bend landslide complex in Rancho Palos Verdes (California, USA), which accelerated to rates exceeding 30 cm/week in 2023. Here we are able to compare to the a longer time series using JPL’s OPERA Surface Displacement from Sentinel-1 (DISP-S1) 12-day revisit time series. We find that the coarser resolution combined with the slower revisit cycle was often insufficeint to resolve rapid movements; Capella's higher temporal sampling provides substantially improved displacement tracking over this site for the studied time period.
Our InSAR processing uses the open-source ISCE3 framework for SLC coregistration and the Dolphin library for PS/DS phase estimation, the same algorithms underpinning NASA's OPERA Surface Displacement product suite. This methodological consistency enables direct, quantitative comparison between medium-resolution C-band (Sentinel-1) and high-resolution X-band (Capella) displacement results over the same regions, providing insight into the complementary roles of public and commercial InSAR systems.
The mid-inclination orbit geometry additionally offers ascending-descending viewing configurations with substantially different look directions than polar-orbiting missions, improving sensitivity to the north-south deformation component and enabling more complete 3D displacement retrieval when combined with Sentinel-1 or NISAR observations. Our results demonstrate that New Space SAR constellations have reached the operational maturity required for sustained, science-quality InSAR time series, and we discuss the implications for multi-mission deformation monitoring strategies.

Interferometric synthetic aperture radar (InSAR) has been widely used to measure Earth’s surface deformation with a millimetre accuracy, providing routine monitoring of infrastructure, mining operations, natural disasters or environmental hazards. Most of currently available SAR satellites operate in the near-polar sun-synchronous orbits (SSOs) with side-looking geometry. SSO-only InSAR monitoring makes satellite’s line-of-sight (LOS) measurements highly sensitive to vertical and east-west displacement with almost no sensitivity to the north-south component of the actual three-dimensional (3D) displacement field. In many applications, either of the east-west or north-south displacements have been assumed negligible so that the LOS InSAR measurements can be projected to either one-dimensional (1D) component in vertical plane or two-dimensional (2D) components in the vertical and east-west planes. Both 1D and 2D analyses with this simplified assumption can however introduce substantial bias when the east-west and north-south components are not negligible. An unbiased solution for 3D time series displacement components can be obtained by at least three stacks of non-coplanar LOS measurements with different viewing geometry. MDA Space will soon be launching CHORUS, a next-generation dual-frequency C- and X-band SAR satellite constellation featuring a unique mid-inclination orbit. CHORUS supports both left- and right-looking imaging capability between incidence angles of 24.5° and 63.9° with a repeat cycle of 9.85 days. The resulting 700 km accessible swath and coverage between ±62.5° latitude (89% global access) can offer substantial wide-area monitoring capabilities. CHORUS’s mid-inclination orbit can provide improved sensitivity to the north-south component of surface displacement field over the wide area. MDA CHORUS combined with existing SSO sensors such as RADARSAT-2, will enable the precise 3D deformation monitoring with its diverse non-polar imaging geometry. Time series of 3D deformation components can be retrieved by mathematical inversion of InSAR LOS measurement stacks collected from different orbits and viewing geometries at different times. However, 3D deformation decomposition inherently becomes an ill-posed inversion problem due to a lack of angular diversity in the SAR datasets. Furthermore, extending 2D InSAR decomposition to 3D InSAR decomposition leads to a more ill-posed inversion problem due to an additional unknown parameter representing north-south displacement component, where any small errors in the measurements can cause a more significant deviation between the true surface displacement and the decomposed solution. Here we introduce the advanced 3D InSAR algorithm to retrieve more reliable time series of 3D displacement components using the combination of different SAR sensors. For this study, temporal stacks of CHORUS-C and RADARSAT-2 interferograms are simulated with seasonal atmospheric and decorrelation noise components assuming hypothetic surface deformation based on Mogi model. Regularization method with various constraints is applied to stabilize the inversion process, enabling recovery of the reliable surface displacement components from the noisy ill-posed linear system of simulated InSAR LOS measurement stacks. Weighted Network Inversion (WNI) can be optionally applied to the stacks of InSAR LOS measurements to mitigate the impact of noise and unwrapping errors, enabling more precise and robust analysis of 3D deformation time series. We demonstrate how the proposed 3D InSAR algorithm effectively improves the recovery of 3D deformation components from the combination of simulated CHORUS and RADARSAT-2 datasets.