Conference Agenda

The European Space Agency's BIOMASS mission, launched on April 29, 2025, represents a groundbreaking advancement in spaceborne radar remote sensing. Operating at P-band (435 MHz), BIOMASS is specifically designed to monitor global forest biomass and structure through its unique combination of polarimetric and interferometric capabilities. The mission acquires fully polarimetric data (HH, HV, VV) across multiple interferometric baselines through its repeat-pass tomographic mode, generating rich multi-dimensional datasets that encode both the physical scattering properties and vertical structure of forested ecosystems.

However, extracting accurate vegetation parameters from BIOMASS data requires a critical preprocessing step: removing the ground topographic phase contribution that contaminates volume scattering measurements. This ground-volume separation challenge is precisely where the Sum Kronecker Product Decomposition (SKPD) technique becomes essential.

Ground phase calibration problem

BIOMASS interferometric phases contain contributions from both the ground surface topography and the distributed volume scattering within the forest canopy. For accurate tomographic focusing and biomass estimation, the ground topographic phase must be precisely estimated and removed. Traditional phase calibration methods struggle in densely forested areas where ground scattering may be weak or completely obscured by canopy returns. The SKPD technique solves this by simultaneously exploiting BIOMASS's polarimetric and tomographic dimensions.

SKPD Decomposition for BIOMASS Data

The technique processes BIOMASS's multi-baseline, multi-polarization covariance matrices through SKPD, factorizing the observations into separable components:

W ≈ Σ λ_k · (C_k ⊗ R_k)

where C_k captures the polarimetric signatures across HH, HV, and VV channels, while R_k contains the interferometric coherence structure across BIOMASS's tomographic baselines. This factorization leverages the fundamental physical principle that ground and volume mechanisms exhibit distinct behaviors in both domains simultaneously.

Exploiting BIOMASS's dual information channels

Interferometric discrimination exploits the stability difference between scattering mechanisms. Ground surfaces produce highly coherent interferometric returns that maintain strong phase correlation across BIOMASS's multiple baselines. Volume scattering from the forest canopy, distributed vertically over tens of meters, exhibits progressive decorrelation as baseline separation increases—a signature clearly visible in the interferometric coherence matrices.

Polarimetric discrimination leverages electromagnetic scattering physics. Ground surfaces beneath vegetation primarily exhibit Bragg surface scattering with dominant co-polarized returns (HH/VV ratios characteristic of rough surfaces). Forest canopy volume scattering generates significant depolarization through multiple scattering among branches, leaves, and trunks, producing elevated HV cross-polarized energy. These distinct polarimetric covariance structures provide complementary separation power.

Joint Optimization and Phase Linking

The algorithm determines optimal mixing coefficients g and v that linearly combine the SKPD basis matrices to reconstruct ground and volume components:

R_g = g·R_{1} + (1-g)·R_{2}
C_g = ((1-v)·C_{1} - v·C_{2})/(g-v)

The ground interferometric coherence matrix R_g is selected by maximizing coherence magnitude—exploiting ground scattering's temporal stability—while ensuring mathematical validity through positive definiteness constraints. Tomographic focusing with Capon beamforming validates that the ground component produces a sharp surface peak while the volume exhibits distributed vertical structure.

Once R_g is identified, phase linking extracts the topographic phase calibration:

φ_topo = phase_linking(R_g)

This calibration phase is then removed from all observations, enabling accurate tomographic imaging of the forest volume uncontaminated by ground topography.

Impact on BIOMASS science products

This ground-volume separation technique is fundamental to BIOMASS mission objectives. Accurate ground phase calibration enables precise forest height mapping, essential for biomass estimation algorithms. The separated volume component provides clean canopy scattering for structure parameter retrieval. The technique's ability to work in dense tropical forests—where ground echoes are weak—makes it particularly valuable for BIOMASS's primary mission of monitoring Earth's most carbon-dense ecosystems.

By synergistically combining BIOMASS's polarimetric and tomographic acquisitions, this approach transforms raw satellite measurements into calibrated observations ready for quantitative forest monitoring.

The authors acknowledge the support in part from the Centre National d'Etudes Spatiales/Terre, Ocean, Surfaces Continentales, Atmosphere (CNES/TOSCA - projects GEDITOMO3D, Aquabio, BayesTomo, BIOMALT and TS4Biomass).

References:

D. Ho Tong Minh et al., "Capabilities of BIOMASS Tomography for Investigating Tropical Forests," in IEEE Transactions on Geoscience and Remote Sensing, vol. 53, no. 2, pp. 965-975, Feb. 2015.

D. Ho Tong Minh et al., Potential value of combining ALOS PALSAR and Landsat-derived tree cover data for forest biomass retrieval in Madagascar, Remote Sensing of Environment, Volume 213, 2018.

D. Ho Tong Minh and S. Tebaldini, "Interferometric Phase Linking: Algorithm, application, and perspective," in IEEE Geoscience and Remote Sensing Magazine, vol. 11, no. 3, pp. 46-62, Sept. 2023

I. El Moussawi et al., L-Band UAVSAR Tomographic Imaging in Dense Forests: Gabon Forests. Remote Sens. 2019, 11, 475.

Y. Bai et al., "An Empirical Study on the Impact of Changing Weather Conditions on Repeat-Pass SAR Tomography," in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 11, no. 10, pp. 3505-3511, Oct. 2018,

Y. -N. Ngo et al., "Tropical Forest Vertical Structure Characterization: From GEDI to P-Band SAR Tomography," in IEEE Geoscience and Remote Sensing Letters, vol. 19, pp. 1-5, 2022.

Launched in April 2025 by the European Space Agency, the BIOMASS mission introduces the first fully polarimetric P-band Synthetic Aperture Radar (SAR) system in space [1]. While its primary goal is the global characterization of forest structure and biomass, an equally important secondary objective is the retrieval of Digital Terrain Model (DTM) beneath vegetation canopies. Achieving this requires multi-baseline interferometric processing of stacks composed of 3 to 7 acquisitions, depending on the operational phase, acquired with a 3-day revisit interval.

A central challenge in this framework is the reliable calibration of the interferometric phase prior to any forest parameter or DTM retrieval, especially considering P-band SAR is highly sensitive to ionosphere. In the BIOMASS processing chain, ionospheric effects are mitigated through innovative dedicated correction procedures [2]. Building upon this calibrated stack, ground phase estimates are derived to serve two essential purposes: refining residual phase errors and enforcing terrain-referenced height alignment (ground steering). This step produces interferometric stacks suitable for subsequent tomographic (TomoSAR) analysis and forest product generation.

The retrieval of the DTM from ground phases, however, remains non-trivial. First, the ground contribution must be robustly localized. We investigate alternative strategies for ground phase estimation, comparing conventional InSAR-based approaches with polarimetric (PolInSAR) techniques. Accurate topographic locking is then achieved through high-resolution spectral estimation methods, enabling precise identification of the terrain response within the vertical reflectivity profile.

After ground phase estimation, residual nuisance large-scale phase components must be removed. These low-frequency disturbances, primarily associated with tropospheric propagation effects—commonly described as Atmospheric Phase Screen (APS) [3]—can significantly bias DTM estimation. Correcting APS in dense forest environments is particularly challenging due to strong volume scattering, spatial and temporal variability of water vapor, and the limited number of available acquisitions.

To address these constraints, we propose a fully data-driven APS mitigation strategy tailored to BIOMASS. The method first estimates and removes the stratified tropospheric component. Subsequently, the turbulent phase contribution is reconstructed using information extracted from open or sparsely vegetated areas. This internally driven correction scheme reduces reliance on external atmospheric services [4], a key requirement for an operational BIOMASS DTM processing algorithm.

Overall, the presented framework integrates polarimetry, interferometry, spectral estimation and atmospheric compensation into a coherent processing chain designed to enable reliable DTM retrieval beneath forest canopies at a global scale.

References

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

[2] S. Tebaldini, F. Salvaterra, F. Banda, and M. Pinheiro, “Multi-layer ionosphere correction in BIOMASS interferometry,” Submitted to IEEE TGRS 2026

[3] A. Ferretti, C. Prati, and F. Rocca, “Permanent scatterers in SAR interferometry,” IEEE Transactions on geoscience and remote sensing, 2002

[4] C. Yu, Z. Li, N. T. Penna, and P. Crippa, “Generic atmospheric correction model for interferometric synthetic aperture radar observations,” Journal of Geophysical Research: Solid Earth, 2018

The BIOMASS mission, ESA’s seventh Earth Explorer, entered orbit on 29 April 2025, marking the deployment of the first P-band Synthetic Aperture Radar (SAR) instrument in space [1]. Operating at long wavelength and providing full polarimetric measurements, the system is specifically designed to support advanced interferometric (InSAR) and tomographic (TomoSAR) techniques. These characteristics enable unique capabilities for three-dimensional forest observation and subsurface terrain retrieval. The mission is primarily dedicated to quantifying forest above-ground biomass and canopy height, as well as tracking their temporal dynamics. In addition, it supports scientific investigations of ionospheric effects, arid regions, and cryospheric environments, and enables the estimation of Digital Terrain Models beneath vegetated areas.

Following launch, efforts have concentrated on system verification during the In-Orbit Commissioning phase, together with an initial appraisal of mapping performance. In parallel, algorithmic development has progressed within related projects, including the implementation of a TomoSAR processing chain under BIOTOMEX. Tomographic datasets were acquired during the COM4 and COM5 phases, using a seven-pass configuration with baselines nominally spaced at 15% of the critical baseline. This geometry corresponds to an expected vertical resolution of roughly 23 m at equatorial latitudes, providing the first opportunity to evaluate interferometric and tomographic performance in orbit. Additional experimental configurations were tested during COM2 through drifting orbit acquisitions originally intended for antenna characterization, which may also support tomographic reconstruction at higher latitudes with comparable performance levels. After completion of commissioning, the mission transitioned to the dedicated tomographic observation phase, ensuring systematic multi-swath coverage.

This talk reviews the current status of BIOMASS TomoSAR activities and describes the planned operational approach.

References

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

Antarctic ice flow mapping from satellite SAR is a well-established technique, with several products available for users [1][2]. These products rely heavily on Sentinel-1 data, but the C-band wavelength means that phase-based InSAR techniques fail on fast flowing glaciers due to loss of coherence from excessive fringe rates and also often in slow flowing areas due to high sensitivity to surface conditions. The fallback technique of amplitude-based offset-tracking results in noisier velocity maps with lower resolution. Compared to existing sensors, the combination of short temporal baselines and increased penetration of the BIOMASS 70 cm wavelength reduces sensitivity to changes in surface conditions, resulting in reduced temporal decorrelation. Also, the long wavelength reduces fringe rates and simplifies phase unwrapping, but these advantages come at the cost of a strong sensitivity to ionospheric scintillations. For BIOMASS, initial results indicate that, as expected, InSAR techniques work well even on fast flowing outlet glaciers where Sentinel-1 InSAR fails, but the influence of ionospheric scintillations is significant and must be addressed, especially when applying the InSAR method in slow-flowing regions.

The acquisition scenario and radar parameters of BIOMASS represent opportunities but also potential challenges when using BIOMASS for ice flow mapping. During the tomographic and Interferometric phases of the mission, a given ground track is acquired in sets of consecutive images with 3-day temporal separation (7 images in each set in the tomographic phase, 3 in the Interferometric phase) but is then not revisited for a long period (273 days in the Interferometric phase), resulting in maps with nonuniform temporal coverage compared to the dense temporal sampling of Sentinel-1. We examine how this can be used to generate a BIOMASS ice velocity product, and how this could complement Sentinel-1 ice velocity measurements.

We also evaluate different methods to address the ionosphere impact on ice velocity maps. Upcoming versions of the BIOMASS processor are expected to introduce corrections that can reduce the impact of ionospheric scintillations, and we examine how these corrections affect ice velocity measurements. We also investigate methods for screening for residual ionospheric signals, and for combining measurements from different acquisitions to further reduce the impact of ionosphere on ice velocity measurements. This is carried out for different flow regimes, ranging from the slow-flowing ice in the interior to the fast-flowing outlet glaciers. Data from both the commissioning phase and the tomographic phase will be used.

References

[1] Rignot, E., J. Mouginot, and B. Scheuchl. 2017. MEaSUREs InSAR-Based Antarctica Ice Velocity Map, Version 2. Boulder, Colorado USA. NASA National Snow and Ice Data Center Distributed Active Archive Center. https://doi.org/10.5067/D7GK8F5J8M8R.

[2] J. Wuite, M. Hetzenecker, T. Nagler and S. Scheiblauer, ESA Antarctic Ice Sheet Climate Change Initiative (Antarctic_Ice_Sheet_cci): Antarctic Ice Sheet monthly velocity from 2017 to 2020, derived from Sentinel-1, v1, NERC EDS Centre for Environmental Data Analysis, 2021.

The European Space Agency’s BIOMASS mission, the first spaceborne Synthetic Aperture Radar (SAR) operating at P-band, provides an unprecedented opportunity for three-dimensional Earth observation through polarimetry, interferometry, and SAR tomography. Although primarily designed for global forest biomass assessment, the mission’s long wavelength and interferometric capabilities also offer exceptional potential for geological investigations, particularly in arid and hyper-arid regions where limited vegetation cover enhances subsurface signal penetration. This study presents preliminary polarimetric and interferometric analyses of BIOMASS data acquired over desert environments in the Sahara, with particular focus on the northern sector of the Tibesti Massif. The investigated area is characterized by Precambrian lithologies belonging to the Lower Tibestian horizon, including gneisses, hornblende- and quartzitic schists, quartzites, and amphibolitic metavolcanics. These units are arranged in narrow isoclinal folds with NNE–SSW and NE–SW trending axes. Structural complexity is further enhanced by disturbed foliation at contacts between metamorphic rocks and intrusive bodies, as well as along tectonic dislocation zones. These contact zones have historically attracted mineral exploration efforts, including investigations of rare earth element concentrations, yet field-based studies have been severely constrained since the 1980s due to environmental and political limitations. Polarimetric analysis of BIOMASS P-band data reveals previously undocumented geological features, highlighting the sensor’s sensitivity to surface roughness variations, structural fabric orientation, and subsurface bedrock configurations beyond the capabilities of conventional optical imagery. Polarimetric decomposition techniques suggest measurable penetration into dry sandy cover, enabling the detection of underlying structural elements and raising important questions regarding effective penetration depth and morphological characterization of buried features. These findings demonstrate that SAR polarimetry—originally developed for vegetation structure retrieval—can be effectively repurposed for geological structural analysis in complex crystalline terrains. Complementary interferometric processing of repeat-pass BIOMASS acquisitions collected during the commissioning phase (three-day revisit cycle with near-zero to variable spatial baselines) provides additional insights into desert surface dynamics and subsurface morphology. Interferograms derived from near-zero spatial baselines enable the detection of spatiotemporal sand dune displacements through phase variations, illustrating the mission’s sensitivity to subtle surface changes. Interferometric pairs with suitable spatial baselines further allow preliminary elevation retrieval associated with potential subsurface structures in eastern Saharan test sites, indicating the feasibility of exploiting P-band interferometry for geomorphological and shallow subsurface investigations. Together, these results demonstrate that the combined polarimetric and interferometric capabilities of BIOMASS open new avenues for three-dimensional structural mapping in arid environments. The mission’s long wavelength, tomographic potential, and repeat-pass configuration provide a unique framework for integrating surface morphology, subsurface imaging, and dynamic surface processes within a unified remote sensing approach. A comprehensive assessment of performance will follow after completion of precise calibration and validation activities; however, the initial findings already underscore the transformative potential of P-band SAR for geological exploration and desert environment research.