Conference Agenda

We investigated P-band SAR interferometry using early interferometric acquisitions of the BIOMASS mission obtained through our Open Cal/Val project. In an initial assessment of the data, we checked the orbit data, noise equivalent sigma zero, spatial resolution, range and azimuth spectra, the presence of radio frequency interference (RFI) and the co-registration accuracy of the STA product. Special care was then given to the identification and mitigation of ionospheric effects. Both split-beam interferograms and azimuth offset refinement fields determined in the co-registration of interferometric image pairs clearly indicated the presence of ionospheric effects. Gradients in the ionospheric path delay cause azimuth positional offsets. Not considering these in the co-registration results in a significant reduction of the interferometric coherence. In our processing we considered the orbit and topography based geometry as well as an offset field refinement determined using cross-correlation based matching techniques. For the correction of the ionospheric path delay phase, we investigated using a split-spectrum technique [1]. Alternatively, the low frequency part of the ionospheric, tropospheric and residual orbital phase was estimated using spatial filtering techniques.

Both the backscatter and the coherence very clearly separate forest and open water areas in the Amazonas area. For short spatial (< 50m) and temporal (3 days) intervals the coherence levels over the tropical forest were generally very high. Over open water very low coherence levels were observed. This is expected, but it also serves as a confirmation of a “clean signal” without significant azimuth ambiguities. In some of the forest areas we observed reduced coherence levels, possibly related to precipitation. This reduction was the smallest at HH polarization, larger at VV polarization and even larger at cross polarization. Our interpretation is that the precipitation slightly changed the direct backscattering as well as the propagation parameters in the vegetation cover. Localized backscatter and phase variations observed might relate to changes in the water level below the canopy.

Hopefully, we will soon be able to also investigate pairs with longer spatial baselines to investigate the dependence of the coherence on the baseline and forest parameters. Furthermore, we look forward to study the P-band coherence in other geographic and thematic areas.

We will present the processing methods applied and discuss the results obtained.

[1] U. Wegmüller, C. Werner, O. Frey, and C. Magnard, “Estimation and Compensation of the Ionospheric Path Delay Phase in PALSAR-3 and NISAR-L Interferograms,” Atmosphere, vol. 15, no. 6, p. 632, May 2024, doi: 10.3390/atmos15060632.

Phase triplets are an interferometric observable derived by summing the interferometric phases around closed loops formed by three SAR acquisitions [1]. In the absence of perturbing effects, the closure sum should be zero. Deviations from this condition, referred to as phase non-closure, are primarily attributed to multi-looking effects, which occur when two or more scattering populations with distinct phase behaviours coexist within a single resolution cell [2]. Previous studies have demonstrated the potential of phase triplets to monitor temporal surface and vegetation changes, including soil moisture variations [1][2], vegetation growth [3], and fluctuations in vegetation water content [2]. However, the physical mechanisms underlying phase non-closure, and in particular its dependence on wave polarization, remain insufficiently understood.

This study aims to advance our understanding of surface and vegetation mechanisms along time by investigating the polarimetric dependence of phase non-closure. To this end, early data from the ESA BIOMASS P-band SAR mission, acquired during the commissioning phase, are employed. With its fully polarimetric capabilities and long wavelength (70 cm), BIOMASS provides sensitivity to scattering processes throughout the entire canopy, from top to bottom, offering a unique opportunity to study the temporal behaviour of forest backscatter under dense vegetation conditions.

Such capabilities are explored over the Gabonese rainforest, a natural environment characterized by high above-ground biomass, complex canopy structure, and diverse tree species composition. Triplets of temporally proximate P-band acquisitions (3 days intervals) are analysed to assess how polarization influences both the magnitude and spatial distribution of phase non-closure. The analysis focuses on small-baseline interferometric pairs to minimize geometric decorrelation, thereby isolating the contributions of volumetric and polarimetric effects. In addition, complementary airborne LiDAR data are used to discriminate forested from non-forested areas and to relate phase non-closure patterns to canopy height and structural heterogeneity.

Results reveal a clear polarization-dependent behaviour: over open areas, HH and VV channels exhibit low and spatially stable closure deviations, whereas in forested regions, HV and cross-polarized combinations display larger and more variable phase non-closure values. These findings suggest that phase non-closure carries valuable information about canopy structure and dielectric heterogeneity. Understanding its polarimetric sensitivity can improve the interpretation of multi-temporal P-band interferometric data for biomass estimation, vegetation dynamics monitoring, and physical model validation in tropical forests.

[1] F. De Zan, A. Parizzi, P. Prats-Iraola and P. López-Dekker, "A SAR Interferometric Model for Soil Moisture," in IEEE Transactions on Geoscience and Remote Sensing, vol. 52, no. 1, pp. 418-425, Jan. 2014.

[2] F. De Zan, M. Zonno and P. López-Dekker, "Phase Inconsistencies and Multiple Scattering in SAR Interferometry," in IEEE Transactions on Geoscience and Remote Sensing, vol. 53, no. 12, pp. 6608-6616, Dec. 2015.

[3] Y. Yuan, M. Kleinherenbrink and P. López-Dekker, "On Crop Growth and InSAR Closure Phases," in IEEE Transactions on Geoscience and Remote Sensing, vol. 62, pp. 1-12, 2024.

The closure phase, constructed by a circular summation of three interferometric phases, each obtained from multilooking a SAR interferogram, consists of a geophysical component and phase noise, often exhibits non-zero values. These non-conservative closure phases challenge the validity of the implicit phase consistency assumption in SAR interferometry. This assumption relies on a geometric interpretation of the interferometric phase, where the expected values of the three interferometric phases are redundant, given that their sum, the closure phase, is equal to zero. Quantifying the statistical significance of non-zero closure phases across different land cover types and examining their correlation with geophysical variations are essential for assessing their effect on interferometric phases and, consequently, improving the accuracy of InSAR applications, such as deformation analysis.

We conducted an extensive spatiotemporal statistical analysis using Sentinel-1 acquisitions over the Iberian Peninsula, a region encompasses diverse land cover types and spans multiple climate zones. The results indicate that the non-zero closure phases are statistically significant. Our case study over agricultural fields revealed a clear geophysical signature strongly associated with vegetation phenology[1]. Two primary mechanisms have been proposed to explain the observed signature: variations in dielectric properties and the line-of-sight motion induced by vegetation growth. This C-band study has highlighted the potential of using closure phase observations to detect variations in vegetation water content. However, Sentinel-1 observations are limited in their ability to exploit the physical processes underlying the closure phase signals. The temporal sparsity of orbital passes and the lack of vertical resolution of contributing scatterers restrict the extent to which closure phases can be investigated and exploited.

Analyzing closure phase observations derived from Biomass P-band tomographic data will enhance the understanding of closure phase signatures, as these observations provide different temporal baselines, greater penetration depth, and the ability to resolve the vertical distribution of scatterers. This study aims to statistically quantify P-band closure phase signatures across diverse land cover types, including bare soil, croplands, forests, and glaciers, using multilooked interferograms. We begin by estimating the standard deviation of the closure phases under the null hypothesis that they originate solely from phase noise, followed by an assessment of the statistical significance of geophysical closure phases across sub-regions categorized by different land cover and climate conditions. Subsequently, we evaluate the mean values and percentiles of closure phases to investigate their spatiotemporal variability. When possible, we compare these results with Sentinel-1, depending on data availability and acquisition location.

Our statistical analysis results obtained from multi-sensor observations will enhance the interpretation of interferometric phase signals and provide valuable insights for explicitly accounting for closure phases in SAR stack processing. These findings will also support the development of novel techniques for bio-geophysical parameter retrieval, complementing traditional radar observables such as backscatter and interferometric coherence.

References

[1] Yan Yuan, Marcel Kleinherenbrink, and Paco López-Dekker. On crop growth and insar closure phases. IEEE Transactions on Geoscience and Remote Sensing, 2024.

BIOMASS is collecting for the first time quad-pol tomographic SAR (TomoSAR) data from space for the reconstruction of 3D radar reflectivity at P-band, providing the unique opportunity to characterize 3D forest structure at global scale. The information content in terms of forest structure of BIOMASS 3D reflectivity reconstructions is determined by two critical aspects. The first one is the low range resolution caused by the 6 MHz bandwidth, which in turn limits the achievable TomoSAR resolution. The second one is the 3-day repeat-pass implementation which induces temporal decorrelation e.g. from either scatterer movements inside the resolution cells due to wind or dielectric changes from one satellite pass to the next. Larger values of temporal decorrelation can decrease the vertical separation capability and the radiometric accuracy of the reconstruction.

The objective of this work is to evaluate BIOMASS TomoSAR reconstructions to obtain a first direct understanding of their potentials for forest structure mapping. Preliminary conclusions are expected on the impact of the reduced vertical resolution and the robustness to reflectivity variations induced by dielectric changes from pass to pass. Acquisitions in the commissioning phase and in the first tomographic cycles are used over a variety of forest sites with different structural characteristic and subject to different levels of dielectric changes in time. If possible, besides lidar data, first results from the BIOMASS validation campaign AfriSAR 3 / GABONX 2025 in Gabon will support the analysis. In November 2025, DLR’s airborne platform is acquiring interferometric / TomoSAR measurements in correspondence of a few BIOMASS passes over the La Lope National Park, providing a unique opportunity for comparing BIOMASS TomoSAR reconstructions with high-resolution temporal decorrelation-free ones.

BIOMASS [1] is ESA’s seventh Earth Explorer, successfully launched on April 29, 2025. The first P-band Synthetic Aperture Radar (SAR) in space offers unprecedented scientific opportunities, thanks to the long wavelength, full polarimetric capabilities and orbits specifically designed for repeat pass interferometry (InSAR) and tomography (TomoSAR). Its primary objective is mapping forest biomass and height and their changes over time. Secondary objectives are targeted at studying ionosphere, deserts and ice, as well as retrieving Digital Terrain Model (DTM) under vegetation [1].

Activities from launch to date have been dedicated mainly to In-Orbit Commissioning (IOC) phase. Main focus has been on instrument and system calibration, though several other activities have been carried out in the framework of ongoing projects (e.g., BIOTOMEX). In the latter two phases of IOC (COM4/5) data have been acquired in tomographic configuration, i.e., seven acquisitions with baseline spacing corresponding to 15% of the critical baseline, with a theoretical TomoSAR vertical resolution of about 23 m at the Equator, offering a preliminary opportunity to assess InSAR and TomoSAR performance. An additional opportunity is offered by COM2, with drifting orbits designed to characterize antenna pattern over BIOMASS transponder, which allow TomoSAR at higher latitudes with performance comparable to that of COM 4/5 at the Equator. After IOC the mission transitions to TOM phase, with full tomographic coverage in three different swaths.

In this contribution we present first BIOMASS TomoSAR results obtained so far, with preliminary assessment on IOC acquisitions both at equatorial and higher latitudes.

References

[1] Quegan, Shaun, et al. "The European Space Agency BIOMASS mission: Measuring Forest above-ground biomass from space." Remote Sensing of Environment (2019)