Conference Agenda

ESA’s BIOMASS launched in April 2025 is the first spaceborne P-band Synthetic Aperture Radar (SAR) ever, fully polarimetric (PolSAR), with primary objective of globally estimating forest properties and secondary goals among which the estimation of Digital Terrain Model (DTM) under vegetation. This can be done through multi-baseline InSAR processing of 3/7 acquisitions, depending on mission phase, separated by a 3-days lag [1].

Preliminary to forest products or DTM estimation, phase calibration of the multi-baseline interferometric (InSAR) stack is mandatory. The three main phase disturbances to be compensated are baseline errors, ionosphere and troposphere. BIOMASS InSAR calibration addresses the first two with a dedicated processing [2] and subsequently produces additional ground phases with the two-fold purpose of residual phase calibration and ground steering, i.e., setting height reference to terrain topography. This allows generating data stacks ready for tomographic (TomoSAR) processing and estimation of forest products.

We discuss in this presentation the strategies devised for BIOMASS DTM retrieval, starting from ground phases. First, we review different approaches to retrieve ground phases (i.e., purely InSAR versus PolInSAR) and address precise topography locking with high-resolution spectral estimation methods [3].

To finally estimate DTM, topography must be separated from residual low-pass disturbances, corresponding mainly to troposphere (APS, i.e., Atmospheric Phase Screen in InSAR literature). Effective APS compensation is challenging in difficult environments such as dense forests, where volume scattering, water vapor variability and a limited number of acquisitions make difficult to resort to traditional InSAR approaches [4]. We discuss a data-driven InSAR APS correction strategy designed for BIOMASS, first removing stratified troposphere, then reconstructing full turbulent phase from open areas. We also assess the superior performance and independence of this approach with respect to external correction services [5], which is desirable for an operational BIOMASS algorithm.

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 POLINSAR 2026

[3] Salvaterra, Francesco; Ferro-Famil, Laurent; Banda, Francesco; Tebaldini, Stefano, “High-Resolution Techniques for Topography Estimation and Terrain Ground Steering within the ESA BIOMASS Processor”, submitted to POLINSAR 2026

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

[5] 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 ESA BIOMASS mission is a single platform, fully polarimetric P-band SAR launched in April 2025 with the task of monitoring world forest and Above Ground Biomass distribution.

Besides supporting the generation of primary mission products, i.e. above ground biomass, forest height, the interferometric acquisitions can also be used to estimate the ground topography.

The estimation of a DTM over forested areas requires to separate the response of the ground from the one of the overlying volume, and to accurately estimate the elevation from which it originates. TomoSAR is a natural solution to this problem as it allows spatial discrimination in the vertical direction to identify the different scattering sources [1-3].

The small bandwidth of the signals measured by the BIOMASS sensor limits the vertical resolution to coarse values in the tomographic mode and to extremely coarse ones in the dual-baseline case. The resulting lack of accuracy and a limited contrast between the responses of the ground and of the canopy may seriously affect the performance of DTM retrieval using classical tomographic imaging techniques or phase center estimation [5].

High-Resolution (HR) spectral analysis techniques represent an alternative to Fourier-based approaches, characterized by significantly improved resolution values. Whereas Fourier-based techniques discriminate sources from reconstructed intensity profiles, HR techniques separate sources by associating the response to low-rank models whose simplicity guarantees both identifiability and robustness [4,5].

In the BIOMASS ground phase estimation processor, the HR spectral analysis is applied after compensating the image stack for low-pass phase screens, mainly caused by tropospheric disturbances. This compensation can be performed using a Phase linking approach (PL) or SKP decomposition.

In BIOMASS processor, the outputs of the HR spectral analysis and the low-pass phase screen calibration are combined to estimate the ground phase, defined as the interferometric phase corresponding to the ground scattering, necessary for ground steering.

This contribution evaluates the performance of HR methods for the estimation of ground phases in the context of the BIOMASS mission, and addresses the following points:

1) association with the SKP and PL methods.

2) assessment of the robustness of the proposed method with respect to residual phase screens.

Performance evaluation is carried out using data from ESA’s TropiSAR, AfriSAR, and TomoSense campaigns, analyzing the influence of polarization choice and calibration methods. Results derived from the BIOMASS ground processor will also be presented.

[1] Soja, M.J., Quegan, S., d’Alessandro, M.M. et al. (2021) Mapping above-ground biomass in tropical forests with ground-cancelled P-band SAR and limited reference data. Remote Sensing of Environment, 253. 112153. ISSN 0034-4257

[2] Y. Huang, L. Ferro-Famil and A. Reigber, "Under-Foliage Object Imaging Using SAR Tomography and Polarimetric Spectral Estimators," in IEEE Transactions on Geoscience and Remote Sensing, vol. 50, no. 6, pp. 2213-2225, June 2012.

[3] Y. Huang and L. Ferro-Famil, "3-D Characterization of Urban Areas Using High-Resolution Polarimetric SAR Tomographic Techniques and a Minimal Number of Acquisitions," in IEEE Transactions on Geoscience and Remote Sensing, vol. 59, no. 11, pp. 9086-9103, Nov. 2021.

[4] Y. Huang, Q. Zhang, and L. Ferro-Famil, “Forest Height Estimation Using a Single-Pass Airborne L-Band Polarimetric and Interferometric SAR System and Tomographic Techniques,” Remote Sensing, vol. 13, no. 3, p. 487, Jan. 2021

[5] P. -A. Bou, L. Ferro-Famil, F. Brigui and Y. Huang, "Tropical forest characterisation using parametric SAR tomography at P band and low-dimensional models," in IEEE Geoscience and Remote Sensing Letters

Digital elevation modeling in desert environments presents unique challenges for radar-based remote sensing. The varying penetration capabilities of radar bands result in terrain representations that differ significantly in depth and detail. X-band signals penetrate only a few centimeters, C-band up to approximately 50 cm, and L-band between 2–3 meters. In contrast, the P-band, used by the Biomass satellite currently in orbit, can penetrate more than 5 meters—potentially reaching bedrock in areas with shallow sand cover. In areas where radar penetration is deep, advanced image processing techniques become crucial for distinguishing between the multiple subsurface layers contributing to the signal.

This study investigates the potential of P-band imagery to enhance terrain modeling in desert regions by comparing it with DEMs derived from X, C, and L-band data. The objective is to provide a preliminary assessment of P-band’s ability to represent subsurface morphology and to anticipate the performance of advanced processing techniques—such as tomography and Polarimetric Interferometric SAR (PolInSAR)—that will be applied to Biomass data for topographic extraction beneath desert surfaces.

The analysis utilizes existing data such as Copernicus (X-band) and SRTM (C-band) DEMs, alongside an InSAR-derived DEM based on ALOS PALSAR imagery over an area in Egypt. Validation is conducted using two complementary approaches:

• External validation compares DEMs against ground truth data, primarily Ground Penetrating Radar (GPR) measurements, to evaluate elevation and slope accuracy compared with bedrock elevation reference.

• Internal validation assesses geomorphological consistency based on physical and statistical principles.

This experiment will be completed by 2D analysis of Biomass polarimetric features to assess to sensitivity of P-band to deeper targets.

The outcomes of this study contribute to the development of a validation framework for future bedrock DEMs generated from Biomass P-band data. This framework aims to clarify the capabilities and limitations of P-band radar for subsurface terrain modeling in arid regions, with implications for geoscientific research and environmental monitoring.

The European Space Agency’s (ESA) BIOMASS mission represents the first spaceborne Synthetic Aperture Radar (SAR) operating at P-band, providing an unprecedented perspective for Earth observation. It is also the first mission to systematically utilize SAR tomography for three-dimensional mapping of terrestrial structures. While its primary objective is to investigate the global biosphere, the mission also offers significant potential for imaging subsurface geological structures in arid regions.

This paper presents a preliminary interferometric analysis of ESA BIOMASS data over desert regions. After its launch, the satellite entered a six-month commissioning phase, during which it collected repeat-pass interferometric pairs with a three-day revisit cycle and a range of spatial baselines, from near-zero to variable separations. Using a straightforward regional InSAR processing approach, the resulting interferograms yields the following initial findings: (i) time-series BIOMASS interferometric data with nearly zero spatial baselines and three-day temporal separation enable the detection of spatiotemporal sand dune movements through interferometric phase measurements; (ii) BIOMASS interferometric data with appropriate spatial baselines facilitate elevation estimation of potential subsurface structures at a study site in the eastern Sahara. These preliminary results indicate that the interferometric capabilities of BIOMASS mission may unlock new opportunities for desert environment research. A thorough analysis will follow upon completion of precise calibration and rigorous validation.

This study presents initial findings from the analysis of Biomass P-band polarimetric synthetic aperture radar (PolSAR) data acquired over the Tibesti Mountains, an arid region characterized by predominantly sedimentary and volcanic rocks, as well as ancient calderas. Only one P-band scene was available for this area, but it provided a valuable opportunity to assess the potential of P-band data for geological applications. We compiled comprehensive geological information, including the integration of hyperspectral datasets from Prisma satellite, and performed classification using various methods. Multiple polarimetric decompositions were evaluated for their suitability in geological mapping and compared with results from historical L-band PolSAR data (ALOS, SAOCOM).

The analysis began with the L1a SCS product, which was ingested using a Jupyter notebook reader. The data were subsequently converted to the DIM format and processed within the Python environment using SNAPISTA and SNAP software. Terrain correction was applied to ensure compatibility with GIS layers and hyperspectral data. Preliminary results allow to analyse overall data quality and reveal intriguing geological features that have not been previously documented, highlighting the unique value of P-band SAR in resolving surface roughness and subsurface structures due to its enhanced penetration capabilities. While the geology of the examined area is relatively straightforward, this work demonstrates the added perspective that P-band data can offer. Future work will focus on analyzing data from the northern slopes of the Tibesti Mountains, where the geology is more complex and includes significant metamorphic rock formations.