Conference Agenda

The NASA-ISRO Synthetic Aperture Radar (NISAR) mission successfully launched on July 30, 2025, reaching its planned science orbit and altitude in mid-October. This dual-frequency mission, operating in L- and S-bands, monitors Earth's land and cryosphere with a regular 12-day interferometric repeat cycle. The NASA and ISRO science data systems currently process approximately 40 Terabits of raw radar data daily (~35 Tb/day L-band and ~5 Tb/day S-band) into high-level, analysis-ready science products. This presentation reports on the status of NISAR L-band products (~40 TB/day), operationally produced at NASA’s Jet Propulsion Laboratory (JPL) and distributed through the Alaska Satellite Facility (ASF DAAC). We present calibration highlights with an emphasis on the performance of the SweepSAR technique, calibration accuracy and science product quality.

In the SweepSAR imaging technique, a ~240 km ground swath is illuminated by a wide transmit beam, while 12 narrow beams are swept in fast-time to digitally beamform (DBF) a composite receive swath. Because the wide swath's return time exceeds the Pulse Repetition Interval (PRI), transmit events inevitably overlap with the receive window, creating "blind range" gaps in the recorded echo. In a constant PRF mode, these gaps occur at fixed locations, leading to persistent artifacts in the focused imagery. To mitigate this, NISAR primarily utilizes a Dithered PRF (or staggered PRI) mode, which varies the pulse timing along-track. This shifting of gap locations ensures that no two consecutive pulses are missing data at the same range, allowing for interpolation to produce contiguous, gap-free science products. We will report on the performance of Dithered acquisition mode which is the dominant mode in the current observation plan for NISAR.

Using data from calibration sites, we evaluate the geometric, radiometric, interferometric, and polarimetric accuracy of L-band (LSAR) data. We also discuss challenges posed by high ionospheric Total Electron Content (TEC) variations, particularly at high latitudes, and analyze the performance of mitigation algorithms within the production system. Furthermore, we assess the status of Radio Frequency Interference (RFI) at a global scale and report on the effectiveness of detection and mitigation strategies.

Finally, we report on the quality of Level-1 and Level-2 L-band science data products. These include traditional Range-Doppler Single Look Complex (RSLC) images, analysis-ready Geocoded Single Look Complex (GSLC) images, and radiometrically terrain-corrected Geocoded polarimetric Covariance (GCOV). We also cover Geocoded wrapped and unwrapped interferograms (GUNW) at a near-global scale, as well as specialized products for cryosphere regions, including RIFG, RUNW, ROFF, and GOFF intended to measure ice sheets velocity.

Abstract

The PALSAR-3 L-band SAR instrument on the ALOS-4 satellite, launched on 1-July-2024, is operated in the same 14-day repeat-orbit configuration as ALOS-2 PALSAR-2, supporting PALSAR-2 – PALSAR-3 cross-sensor repeat-pass interferometry. During the joint operation period, this enables acquiring interferometric pairs with short 6- and 8-day intervals. In the long-term, the main role of PALSAR-3 is to replace PALSAR-2 and to enable the continuation of L-band SAR data time series. Of particular interest is the possibility to continue the L-band interferometric time series built up with PALSAR-2 since 2014 into the future. In our work we discuss cross-mode and cross-sensor interferometry, considering PALSAR-2 and PALSAR-3 stripmap and ScanSAR mode data.

1. Cross-sensor and cross-mode differential interferometry

Both PALSAR-2 and PALSAR-3 can be operated in several stripmap and ScanSAR modes. Some of the parameters differ between the modes, which needs to be considered in the interferometric processing. To maximize the coherence, the signal is reduced to the common azimuth and range spectra. Furthermore, the different range and azimuth samplings needs to be considered in the co-registration.

As a first step, we assure that the carrier frequency of the two scenes is identical. In cases with slightly different carrier frequencies, we determine a new carrier frequency such that it is in the center of the common range bandwidth. Changing the carrier frequency is accomplished by the application of a range phase ramp that corresponds to the effect of the frequency difference.

Then, we co-register the SLC data to a common geometry. A reference geometry with a high enough sampling is used to avoid aliasing. Typically, this is done by selecting the data set with the highest spatial resolution as the reference. If necessary, this can be assured through an oversampling of the selected reference SLC, e.g. in the case where one scene has the highest range resolution and the other scene the highest azimuth resolution. After the co-registration, the common band filtering is applied. In the case of stripmap – stripmap pairs we typically use the same band-pass filter for the entire scenes. In the case of stripmap – ScanSAR pairs, the common band filtering applied is spatially adaptive to consider the along-track variation of the ScanSAR SLC spectrum. The co-registered and common-band filtered SLCs are then used to generate differential interferograms.

Within their common 14-day repeat-orbit, PALSAR-2 follows upon PALSAR-3 with a 6-day delay. Accordingly, the shortest interferometric intervals possible are 6 days and 8 days. Short intervals are of particular interest for the mapping of fast displacements. We used e.g. such short-interval pairs to map alpine rock-glacier velocities.

Also considering stripmap – ScanSAR pairs increases the number of pairs available over a certain area. Examples investigated confirm the feasibility of the processing and demonstrate the usefulness of the results.

The characteristics of the ScanSAR mode of PALSAR-3 differs significantly from the ScanSAR mode of PALSAR-2. Therefore, ScanSAR – ScanSAR cross-sensor interferometry is not supported. The azimuth spectral patterns are not synchronized. Nevertheless, parts of the spectrum do overlap. Tests showed that interferometry is still possible for a significant fraction of the common acquisition area covered, but at a reduced spatial resolution.

2. Persistent scatterer Interferometry

Persistent Scatterer Interferometry with PALSAR-2 time-series of stripmap and/or ScanSAR acquisitions are widely used to map ground-motion [1,2,3]. PALSAR-3 now offers the opportunity to continue such time-series into the future beyond the lifetime of PALSAR-2. Detecting mm/year scale displacement rates requires using interferometric time series of several years. Therefore, being able to continue PALSAR-2 time series is clearly preferred over starting a new PALSAR-3 only time series, as reliable results can immediately be obtained without delay until a long enough data stack has been acquired. Both stripmap and ScanSAR acquisitions can be considered in PSI time series. To interpret the phase of point-like scatterers, no common band filtering needs to be applied in the processing.

3. Acknowledgements

PALSAR-2 and PALSAR-3 data used in this work are copyright JAXA. The PALSAR-2 and PALSAR-3 data were made available to us through ALOS PI Projects EORA4N012 (PI Wegmüller) and ER4A2N045 (PI Strozzi).

4. References

[1] Wegmüller, U.; Magnard, C.; Strozzi, T.; Caduff, R.; Jones, N.; Landslide velocity mapping using ALOS-2 PALSAR-2 ScanSAR data, Procedia Computer Science 2024, Vol. 239, pp. 2278-2285, ISSN 1877-0509, https://doi.org/10.1016/j.procs.2024.06.419.

[2] Strozzi T., R. Caduff, N. Jones, A. Manconi, and U. Wegmüller, “L-Band StripMap-ScanSAR Persistent Scatterer Interferometry in Alpine Environments with ALOS-2 PALSAR-2,” in Proc. IEEE Int. Geosci. Remote Sens. Symp., 2022, pp. 1644–1647. https://doi.org/10.1109/IGARSS46834.2022.9884743.

[3] Strozzi, T., Jones, N., Agliardi, F., De Pedrini, A., Frey, O., Bernhard, P., Caduff, R., Ambrosi, C., and Manconi, A.: Monitoring the displacement of large alpine rock slope instabilities with L-band SAR interferometric techniques, https://doi.org/10.5194/egusphere-2025-5347, 2025.

In recent years, China’s Earth observation infrastructure has undergone a marked transition from isolated SAR missions to a more diversified and increasingly coordinated SAR system. This rapid development is creating new opportunities for InSAR-based Earth observation, particularly for nationwide deformation monitoring, hazard assessment, and operational geoscience applications. Beginning with the GaoFen-3 mission, several notable SAR systems have entered service in recent years, including the L-band bistatic mission LuTan-1, the geosynchronous SAR mission LuTan-4, and new commercial small-satellite SAR constellations represented by Tianyi. In parallel, additional mission concepts are being advanced, including low-inclination SAR systems, P-band SAR, and SRTM-like interferometric missions. Together, these developments suggest that China is moving toward a more complete and multi-layered InSAR observation architecture.

As China’s first high-resolution civilian SAR mission, GaoFen-3 has now evolved into a three-satellite constellation in operation. Existing studies have shown that GaoFen-3 can deliver reliable InSAR and MT-InSAR results and, in favourable urban environments, can achieve deformation-monitoring performance comparable to Sentinel-1, with reported velocity accuracy reaching the millimetre-per-year level. GaoFen-3 therefore provides an important domestic foundation for routine C-band InSAR applications and serves as an early demonstration of China’s capability for sustained spaceborne deformation monitoring using indigenous SAR systems.

Building upon this foundation, LuTan-1 represents a major advance as China’s first civilian L-band full-polarimetric SAR constellation for topographic mapping and deformation monitoring. The mission comprises two identical satellites and was designed for interferometric applications from the outset. It combines the coherence advantages of L-band with relatively high spatial resolution, for example providing dual-polarization stripmap imaging at 3 × 3 m resolution over a 50 km swath. The mission operates in two main phases: bistatic DEM generation followed by deformation monitoring. From an InSAR perspective, LuTan-1 is particularly important because it extends China’s monitoring capability into the L-band domain, which is better suited to vegetated, mountainous, and geologically active areas where C-band coherence is often limited. Early operational practice further shows that LuTan-1 is already being used for wide-area geological hazard monitoring and, in Shaanxi Province, for the routine production of monthly provincial DInSAR deformation maps.

In the commercial sector, the Tianyi series illustrates the rapid progress of China’s small SAR satellite capability. From Hisea-1, China’s first commercial small SAR satellite, to Chaohu-1, which completed in-orbit InSAR testing, and then to Fucheng-1, the technical evolution has been notably fast. Fucheng-1, launched in 2023, is particularly important because it demonstrates that a miniaturized commercial SAR platform can support not only repeat-pass InSAR, but also multi-temporal InSAR services when sufficient orbit control and radiometric quality are achieved. Operating at relatively low altitude, Fucheng-1 achieves favourable radiometric performance, while its electric propulsion system enables precise orbit maintenance within a 150 m-radius control tube, thereby improving interferometric consistency. In a time-series analysis over Mianyang, China, Fucheng-1 achieved a reliable monitoring-point density exceeding 18,000 points/km² with coherence greater than 0.7, producing results comparable to TerraSAR-X and Sentinel-1.

Another major milestone is LuTan-4, which marks China’s entry into geosynchronous SAR observation. LuTan-4-01, launched in August 2023, is the world’s first GEO-SAR satellite to enter engineering operation. Unlike conventional low-Earth-orbit SAR systems, LuTan-4 offers a fundamentally different observational geometry, combining all-weather microwave sensing with the potential for high-temporal-sampling observations over broad regions. For InSAR, its significance lies not merely in adding another mission, but in opening a new technical route for observing large-scale and rapidly evolving deformation processes over key areas. At the same time, the practical use of GEO-SAR for operational deformation monitoring still depends on further progress in long-aperture imaging, calibration, phase-stability analysis, and error-control strategies, all of which remain more challenging than in mature LEO InSAR systems.

Taken together, these missions indicate that China’s SAR capability for InSAR has moved beyond the exploratory stage and entered a phase of sustained system-level development. More importantly, the emerging picture is not one of isolated satellites, but of a diversified observation architecture spanning C-band, L-band, commercial small-satellite platforms, and geosynchronous SAR. Such an architecture has clear potential for future multi-mission synergy, including complementary use across frequency bands, orbit types, spatial scales, and revisit characteristics. At the same time, it also raises important scientific and technical challenges, including cross-mission interoperability, geometric and radiometric consistency, processing standardization, and product validation.

Finally, we introduce TERESA (Terrain Registration and Sampling Software, https://github.com/aprilab-dev/teresa), an open-source tool developed on the basis of DORIS to better support InSAR processing of Chinese SAR data. By lowering the technical barrier for handling these emerging datasets, TERESA aims to facilitate broader community access and promote further scientific and operational use of Chinese SAR missions within the international InSAR community.

The NASA–ISRO Synthetic Aperture Radar (NISAR) mission [1] was successfully launched on July 30, 2025. The mission concluded its Commissioning Phase at the end of October 2025 and formally entered Science Phase operations in early January 2026. Spacecraft and payload performance are currently nominal, with data acquisitions proceeding according to the Science Reference Observation Plan. At the time of writing, dense time series covering ten complete global cycles are already available, with more than 15 cycles expected by the start of the ESA Fringe workshop. A first subset of data (over 100K products spanning level-1 and level-2) has just been released, with the full forward archive expected to open later in June 2026. These releases will provide multiple interferometric pairs for the science community to assess data quality and unlock the scientific potential of NISAR.

The objective of this contribution is to examine polarimetric–interferometric time-series and their sensitivity to biophysical land parameters. For NISAR, the dominant source of interferometric decorrelation stems from temporal changes, given the narrow orbital tube (< 100 m RMS) and generally low noise-equivalent-sigma-zero (NESZ < −23 dB). Complex-valued temporal coherence has been shown to be sensitive to key bio/geophysical parameters, including tree height, vegetation density, soil moisture, and, more generally, the ground-to-volume scattering ratio. Existing models linking temporal coherence and backscatter time series to biophysical parameters demonstrate promising sensitivity, and physics-based algorithms for retrieving canopy structure and other biophysical parameters are under active development [2, 3].

Here, we first provide an overview of NISAR capabilities for land cover applications, including acquisition modes, expected polarimetric and interferometric performance, and the observation strategy. We then analyze dual-polarization and quad-polarization interferometric (complex-valued) coherence time-series over selected sites, primarily at established NISAR calibration and validation sites over land. Our two-layer scattering models [2, 3] are used to guide interpretation and to compare observations with theoretical predictions. Independent lidar and ground measurements are used for validation.

Sequences of temporal coherence are derived from interferograms generated primarily from geocoded single-look-complex (GSLC) products. These products are the standard NISAR products obtained by geocoding the conventional range-Doppler SLC (RSLC) on a geographic UTM grid after removing the topographic fringes and correcting for the bulk of the ionospheric distortions (estimated from models and GNSS data). GSLC-based coherence estimates are assessed against traditional RSLC-based coherence estimated over selected sites to ensure their fidelity and consistency across the time-series. At the workshop, we plan to discuss our experience with modeling and using L-band temporal decorrelation derived from NISAR GSLC products for land applications.

REFERENCES

[1] P. A. Rosen, G. W. Bawden, P. Barela, B. Chapman, H. Fattahi, C. E. Jones, I. R. Joughin, M. Lavalle, R. B. Lohman, M. Simons, P. Siqueira, A. Das, N. M. Desai, R. Kumar, D. Putrevu, R. Sharma, and C. Shrikant, “The NASA–ISRO SAR Mission: A summary,” IEEE Geoscience and Remote Sensing Magazine, vol. 13, no. 2, pp. 8–34, 2025.

[2] M. Lavalle, C. Telli, N. Pierdicca, U. Khati, O. Cartus, and J. Kellndorfer, “Model-based retrieval of forest parameters from Sentinel-1 coherence and backscatter time series,” IEEE Geoscience and Remote Sensing Letters, vol. 20, pp. 1–5, 2023.

[3] C. Telli, M. Lavalle, and N. Pierdicca, “Vegetation height from L-band SAR backscatter and interferometric temporal coherence measurements,” Remote Sensing of Environment, vol. 328, p. 114879, 2025.

After first being recommended by the 2007 National Academy of Science “Decadal Survey” report “Earth Science and Applications from Space: National Imperatives for the Next Decade and Beyond” [1] to support measurements that determine Earth change in three disciplines – ecosystems, solid earth, and cryospheric sciences – the NASA-ISRO SAR (NISAR) mission is at last in orbit. NISAR launched on July 30, 2025, completed commissioning at the end of 2025, and began science operations in its calibration/validation phase beginning in 2026. Over the years of development, several papers have described various aspects of the NISAR mission: its science, observation plan, subsystem technologies, and products suite, summarized in a comprehensive mission-focused paper [2] and science-focused article [3], both completed pre-launch. This paper compares NISAR’s planned and realized operational performance with respect to interferometric SAR, including orbit and pointing control, signal-to-noise and multiplicative noise performance, and general coverage statistics.

NISAR uses synthetic aperture radar to map Earth’s surface every 12 days, persistently on ascending and descending portions of the orbit, over all land and ice-covered surfaces. The mission’s primary objectives will be to study Earth land and ice deformation, and ecosystems, in areas of common interest to the US and Indian science communities. This single observatory solution with an L-band (24 cm wavelength) and S-band (10 cm wavelength) radar has a swath of over 240 km at fine resolution, and will operate primarily in a dual-polarimetric mode in an exact repeat orbit. The science teams at NASA and ISRO are working jointly to finalize the joint science plan, calibration and validation plan, and science products, and operational procedures. Both the L- and S-band radars have a fixed swath of 240 km enabled by the phased-array-reflector scan-on-receive design, with selectable polarizations – single, dual, quad, circular – and a range of bandwidths from 5 MHz to 77 MHz. The azimuth resolution is fixed by the 12 m reflector size to be approximately 7-8 m. The combination of polarization state and range resolution is assigned by science target. There are hundreds of possible combinations, but the observation plan favors completeness and consistency, so only a limited subset of modes is used.

NISAR is a science driven mission. For solid Earth, NISAR will use repeat pass interferometry to characterize long-term and local surface deformation on active faults, volcanoes, potential and extant landslides, subsidence, and uplift associated with changes in aquifers and subsurface hydrocarbon reservoirs, and other deforming surfaces. For ecosystems, NISAR will regularly measure changes in radar polarimetric signatures to estimate the amount of woody biomass and its change in the most dynamic ecoregions of the world. The mission will also be able to track changes in the extent of active crops to aid in crop assessments and forecasting, changes in wetlands extent, and characterize freeze/thaw state and permafrost degradation. In the cryosphere, NISAR will provide comprehensive interferometric measurements of Greenland’s and Antarctica’s ice sheets, seasonal dynamics of highly mobile and variable sea ice, and inventory the variability of key mountain glaciers which are retreating in many places at a record pace. In addition, NISAR will be operated to observe potential hazards and disasters on a best-efforts basis to demonstrate rapid assessments in urgent events such as earthquakes, volcanic eruptions, floods, and severe storms. These data will support research into effective rescue and recovery activities, system integrity, lifelines, levee stability, urban infrastructure, and environment quality.

The spacecraft was launched intentionally into an orbit slightly below its 12-day repeat science orbit. The science orbit was achieved mid-October 2025, with a design goal of staying within 350 m of the reference orbit. Since the science orbit was achieved, NISAR has been within 100-200 m of its reference orbit. For low slope surfaces where the critical baseline is on the order of 10-20 km depending on the mode, this offers virtually no baseline decorrelation and allows good fringe visibility in areas with slopes below the incidence angle.

The array-fed reflector antenna system has two feeds, one for L-band and one for S-band, lying side by side on the focal plane. Since neither is at the focus in azimuth, the L-band beam is squinted forward by about 0.9 degrees, while the S-band beam is squinted backwards by about 0.3 degrees. The system was designed to maintain beam pointing such that the mechanical boresight at the focal line in azimuth was always within 1/20^th of a L-band beamwidth of 0 degrees (zero-doppler pointing) for all points along orbit. To date the pointing has been exceptionally stable.

References

[1] Earth Science and Applications from Space: National Imperatives for the Next Decade and Beyond, Ed. by Committee on Earth Science and Applications from Space, National Research Council, National Academies Press, ISBN 978-0-309-10387-9, 2007.

[2] K. Kellogg et al., "NASA-ISRO Synthetic Aperture Radar (NISAR) Mission," 2020 IEEE Aerospace Conference, Big Sky, MT, USA, 2020, pp. 1-21, doi: 10.1109/AERO47225.2020.9172638.

[3] P. A. Rosen et al., "The NASA-ISRO SAR Mission: A summary," in IEEE Geoscience and Remote Sensing Magazine, vol. 13, no. 2, pp. 8-34, June 2025, doi: 10.1109/MGRS.2025.3578258.