Conference Agenda

This paper introduces Adaptive sum of Kronecker products for QUAntitative- Soil Moisture Anomalies retrieval (AKQUA-SMA), a novel polarimetric framework for SMA estimation. The core innovation lies firstly in the use of Temporal-SKP (T-SKP) [1], which exploits the temporal-polarimetric domain to isolate two scattering components: the moisture-related contribution, namely the Latent contribution, from the other scattering mechanism, namely Ground, i.e., above-ground scattering components. To do so, the SKP solution is chosen through exhaustive search as the one that minimizes the l1-norm of the error between the decomposed coherence values in the Latent structure matrix and the theoretical expected moisture related complex coherences. In particular, the search grid originates from the model proposed in [2], which has been slightly modified to account for a double-scattering mechanism. Then, the Ground component is chosen as the one that minimizes the phase residues, i.e., the most triangular scattering mechanism, and used for phase calibration similarly to what is done for the BIOMASS tomography. Finally, absolute N (zero-mean) SMA values are regressed by IRLS estimation from the estimated variations in the N(N-1)/2 InSAR pairs.

The AKQUA-SMA algorithm was applied to the Hydrosoil dataset [3], which was collected over a 20m × 58m agricultural field using a C-Band ground-based PolSAR (GB-PolSAR). The campaign aimed to simulate the frequent monitoring capability of the HydroTerra mission [3] for soil moisture and vegetation parameter retrieval. The data comprises two phases: the Barley Crop (March–June 2020), a Dual-Pol dataset (18,055 acquisitions), and the Corn Crop (July–November 2020), a Quad-Pol dataset (12,945 acquisitions), both acquired every 10 minutes. This SAR data is supplemented with essential ancillary information, including probe-based volumetric moisture, plant density, and crop height. Due to the limited field size, the entire area was treated as a single resolution cell. Processing utilized a full overlapping sliding window approach, scanning the dataset in steps of six acquisitions with five temporal samples overlapping between adjacent windows. The first results reveal good estimation accuracy, with a RMSE of less than 3% for the entire Barley campaign dataset (3 months, 18,055 acquisitions). For the Corn dataset, which is a challenging crop type, AKQUA-SMA provides estimates with an accuracy of less than 2% in the period from the bare soil stage up to slightly vegetated field (crop height<=40cm), which is the maximum height reached by barley. Experiments are currently underway on SARSimHT-NG dataset to study the behavior of the developed approach in the case of other crop types.

References

[1] Tebaldini, Stefano. ”Algebraic synthesis of forest scenarios from multibaseline PolInSAR data.” IEEE Transactions on Geoscience and Remote Sensing 47.12 (2009): 4132-4142.

[2] De Zan, Francesco, et al. ”A SAR interferometric model for soil moisture.” IEEE Trans actions on Geoscience and Remote Sensing 52.1 (2013): 418-425.

[3] Aguasca, Albert, et al. ”Hydrosoil, soil moisture and vegetation parameters retrieval with a C-band GB-SAR: Campaign implementation and first results.” 2021 IEEE Inter national Geoscience and Remote Sensing Symposium IGARSS. IEEE, 2021

Characterizing spatial soil moisture patterns is essential for numerous applications in high-latitude permafrost regions, as soil moisture controls thermal and biogeochemical processes, supports flux upscaling, and the assessment of greenhouse gas composition. However, conventional remote sensing approaches often face difficulties in these environments due to the pronounced landscape heterogeneity characteristic of Arctic permafrost regions. Seasonal thawing and freezing of the near-surface soil drive subsidence-heave cycles that typically produce vertical displacements of less than 10 cm.

In this study, we examine the capability of Sentinel-1 interferometric synthetic aperture radar to infer near-surface soil moisture by exploiting the relationship between seasonal surface subsidence and thawing degree days (DDT), a metric of cumulative seasonal heating. Using Sentinel-1 data from Arctic lowland permafrost regions, we derive deformation rates in the DDT domain and evaluate their correspondence with in situ soil moisture measurements. We focus on two study regions with in situ soil property data - central Yamal (northwestern Siberia) and Inuvik (northwestern Canada) - and evaluate the InSAR-derived deformation results against in situ subsidence measurements and in relation to near-surface soil moisture.

Our results show that locations with higher near-surface soil moisture exhibit greater subsidence rates per DDT than drier sites, confirming a link between soil wetness and thaw-related deformation. Building on this relationship, we propose an interpretation scheme for classifying soil moisture categories and assess its performance against in situ observations and commonly used remote-sensing-based soil moisture indicators. Compared with coarse-resolution satellite products, which frequently underestimate in situ soil moisture, the InSAR-derived metric shows lower errors and resolves smaller-scale soil moisture patterns.

While the method provides only static information and does not capture short-term or seasonal variability, InSAR-derived subsidence rates represent a valuable proxy for overall soil moisture states in heterogeneous permafrost landscapes. The applicability of the approach is limited by the spatial resolution of Sentinel-1, which constrains the detection of fine-scale permafrost features such as high- and low-centered polygons, which are associated with characteristic wet and dry patterns. Nevertheless, future research will focus on applying our methodology to additional Arctic regions to further explore its potential and test its transferability. Future work may also benefit from comprehensive longer-wavelength observations that will become available from missions such as NISAR, which may improve coherence and enhance soil-moisture-related deformation retrievals.

In nature, biological events follow a cyclical rhythm that enables ecosystem functioning. The seasonal development of plants shapes terrestrial ecosystem dynamics and strongly influences carbon exchange and land-atmosphere interactions. Temperate deciduous forests, one of the dominant forest types across Europe, exhibit pronounced seasonal transitions. In spring, canopy leaf-out marks the onset of the growing season, as new foliage emerges and the forest structure rapidly reorganizes. This shift from winter dormancy to active growth is accompanied by the beginning of photosynthetic activity and an accelerated increase in carbon uptake. Accurately detecting the timing of this shift, commonly expressed as the start of season (SOS), is therefore of high ecological and climatic importance.

Traditionally, phenological monitoring has relied on ground-based observations; however, such measurements are labor-intensive and spatially limited, making large-scale or continuous regional assessment nearly impossible. Consequently, satellite remote sensing has become a key tool for monitoring phenological dynamics, while ground observations remain essential for validating satellite-derived phenological metrics.

Optical satellite data, particularly vegetation indices derived from multispectral sensors such as Sentinel-2, have been widely used to estimate SOS. Nevertheless, optical observations are highly sensitive to cloud cover, which is frequent during early spring when phenological changes occur most rapidly. Moreover, optical indices primarily reflect canopy greenness and photosynthetic activity, potentially overlooking structural changes occurring before full leaf expansion. In vertically complex deciduous forests, spring green-up progresses sequentially across layers, from forest floor vegetation to shrubs and juvenile trees, and finally to the overstory canopy formed by mature tree crowns. As a result, optical signals integrate multiple structural components and may not accurately represent canopy onset alone.

Radar remote sensing offers complementary information, as Synthetic Aperture Radar (SAR) operates independently of illumination and cloud cover and is sensitive to vegetation structure and its dielectric properties. This study evaluates the potential of Sentinel-1 interferometric coherence to characterize early spring forest dynamics. The analysis was conducted in Betula pendula-dominated stands located in central Poland, featuring a uniform canopy layer and heterogeneous understory and forest floor vegetation. Single Look Complex (SLC) data acquired in Interferometric Wide mode (VV and VH polarizations) were processed to generate backscatter time series and short temporal baseline interferograms (6-12 days, depending on the availability of S1A and S1C data). Mean coherence was calculated over forest polygons and normalized by temporal baseline to account for varying revisit intervals. Rapid vegetation development was expected to cause structural reorganization within the forest volume, increasing temporal decorrelation. The period of most intense vegetation transition was defined as the interval exhibiting the strongest negative coherence change per day.

Results indicate that the strongest negative coherence change aligned closely with ground-observed canopy structural reorganization and leaf-out, demonstrating that rapid temporal decorrelation provides a reliable indicator of SOS. Distinct polarization-dependent responses emerged. VV polarization exhibited an earlier coherence decrease preceding canopy emergence, likely reflecting structural and dielectric changes associated with forest floor and understory development. The absence of significant precipitation during this period suggests that the observed decorrelation was primarily driven by vegetation dynamics rather than soil moisture variations caused by rainfall. In contrast, the strongest coherence decrease in VH polarization occurred during rapid canopy leaf expansion, coinciding with full canopy development observed in the field.

SAR backscatter (σ⁰) time series have been widely used to monitor seasonal vegetation dynamics. However, the present analysis shows that in contrast to coherence, backscatter exhibited rapid short-term fluctuations associated with rainfall events throughout the observation period. These precipitation-driven variations masked changes related with vegetation, limiting the reliability of backscatter for detection of phenological events. Interferometric coherence exhibited less pronounced short-term variability associated with rainfall compared to backscatter.

To enable comparison with optical indicators, the Normalized Difference Vegetation Index (NDVI) derived from Sentinel-2 imagery was also analyzed. During the study period (January–June 2025), 45 Sentinel-2 acquisitions were available, however, only 14 scenes met the cloud-cover threshold (<5%) over the study area. In April, when the most rapid vegetation transitions occurred, only two cloud-free acquisitions were cloud-free, resulting in limited temporal sampling during this critical onset phase. NDVI captured the main phase of canopy development, characterized by a rapid increase during full leaf expansion, but showed limited sensitivity to earlier structural changes preceding canopy closure.

These findings demonstrate that Sentinel-1 interferometric coherence captures early spring structural reorganization in deciduous forests and offers a physically interpretable, weather-independent complement to optical phenological indicators. The study highlights the potential of InSAR observables for ecosystem monitoring and supports the development of radar-based phenological metrics beyond conventional backscatter analysis.

Persistent Scatterer Interferometry (PSI) is a remote sensing technique used to document and monitor ground surface deformation by exploiting multiple interferometric Synthetic Aperture Radar (SAR) images. PSI relies on the identification of radar targets, so-called Persistent Scatterers (PS), that exhibit stable backscattering behavior over long periods. Especially with the launch of the Sentinel‑1 (S1) satellites, data availability became far less restrictive. The S1 mission provides users worldwide with a continuous stream of SAR images free of charge. At the same time, the large number of available images raises questions about their suitability for PSI processing. Not all images contribute equally to the formation of stable interferometric networks, and environmental conditions at the time of acquisition can strongly influence the stability of the backscattered signal. For example, the S1‑based ground motion service InSAR Norway primarily processes images acquired during the summer months, as snow cover disrupts the stability of the backscattered radar signal and reduces PS density. This example illustrates that even with abundant data, careful selection of acquisitions may be necessary to ensure high-quality PSI results.

In an earlier study, we evaluated different strategies for processing a continuous stream of SAR images, including the use of consecutive subsets. During this work, we observed a clear seasonal pattern in PS density, suggesting that environmental conditions may systematically influence the number of detectable PS. This finding prompted the question of whether, similar to the approach used by InSAR Norway, images acquired during certain times of the year may be less suitable for PSI depending on the specific use case and geographic setting. While snow cover is a major limiting factor in northern and alpine environments, other regions may be affected by vegetation cycles, soil moisture variations, or seasonal precipitation patterns.

In the work presented here, we assess variations in PS density across 300 SAR images recorded between November 2016 and December 2021 for the coastal city of Patras and its surrounding areas. The SAR images were processed in consecutive, non‑overlapping subsets of 30 images, each covering either the period from early winter (November or December) to early summer (May or June), or the reverse. The resulting PS densities are examined for seasonal variations associated with different land‑cover and for the influence of environmental factors such as soil moisture, ambient temperature, and precipitation. This analysis provides insights into the temporal suitability of SAR acquisitions for PSI and highlights the importance of environmental context in designing effective processing strategies.

We present a new approach, based on the range autocorrelation function and the observed range pixel to pixel phase shift, that basically unwraps the phase of a single SLC image and yields the path lengthening () within a single focused SLC in the range direction (i.e. in the Line of Sight). One of the possible applications of this approach is the use of conventional SAR systems for ocean surface topography mapping at high spatial resolution. We call this method SAR Original Phase retrieval (SOf). Here, based on Sentinel-1 Strip-Map images, we show results on the Indian Ocean and compare them with conventional satellite altimetry, on two dates. Where the signal-to-noise is strongest, we observe statistical fluctuations less than 1 cm in comparison with SSH (Sea Surface Height) based on conventional Altimetry.

SSH is one of the critical parameters in ocean science allowing to better constrain the ocean spatio-temporal dynamic as it is highly related to sea currents distribution, sea temperature and sea-atmosphere complexes interactions. Besides, it is an important proxy for the geoid estimation. The phase of a side looking SAR imager such as Sentinel 1, on the sea surface, is already used to provide information about the sea surface currents through the Doppler centroid analysis, along the azimuth axis of the image. However, in the range direction, the SAR phase is commonly considered spatially uncorrelated and, therefore, its autocorrelation function is commonly considered to be random and to have no practical use. Following experiments we conducted on focused SLC SAR signals in the open ocean, with Sentinel-1, it emerges that this theorical statement is not entirely satisfactory. Here, we show results on the Indian Ocean and compare them with conventional satellite altimetry, on two dates. This method might have interesting implications not only for improving SSH retrieval using conventional SAR sensors, but also might present interesting perspectives for Earth surface displacement mapping with conventional InSAR. This study suggests that the infra-pulse phase delay in the focused SAR signal can be seen as an across track deflection from the theoretical SAR incident angle -i.e. the offset angle to the local vertical. In open ocean, this quantity is the definition of the geoid. In conclusion, this result has twofold importance. On the one hand, it shows that the phase of a single SLC does not carry white noise only, expanding the conventional assumption of SAR theory. On the other hand, it shows that the phase of a single SLC can be measured and can be unwrapped, transforming conventional SAR into a relative altimeter, as a complement existing methodologies based on space altimeters (Sandwell 1984; Cazenave et al., 1996; Yu et al., 2024). Furthermore, it opens the door to a whole range of applications beyond space oceanography—such as measuring atmospheric phase delay in a single SAR image.

Sandwell, D., 1984, Along-track deflection of the vertical from Seasat : GEBCO overlays. NOAA Tech Memo NOS NGS-40, Natl Ocean Serv., Rockville, MD. https://repository.library.noaa.gov/view/noaa/2798.

Cazenave, A., P. Schaeffer, M. Berge, C. Brossier, K. Dominh, M. C. Gennero, High-resolution mean sea surface computed with altimeter data of Ers-1 (geodetic mission) and topex-poseidon, Geophysical Journal International, Volume 125, Issue 3, June 1996, Pages 696–704, https://doi.org/10.1111/j.1365-246X.1996.tb06017.x.

Yu Y., Sandwell D.T., Dibarboure G., Abyssal marine tectonics from the SWOT mission. Science 386, 1251-1256 (2024). DOI:10.1126/science.ads4472.

We present a new approach, based on the range autocorrelation function and the observed range pixel to pixel phase shift, that basically unwrap the phase of a single SLC image and gives the path lengthening () within a single focused SLC in the range direction (i.e. in the Line of Sight). One of the possible application of this approach is the use of conventional SAR systems for ocean surface topography mapping at high spatial resolution. We call this method SAR Original Phase retrieval (SOf). Here, based on Sentinel-1 Strip-Map images, we show results on the Indian Ocean and compare them with conventional satellite altimetry, on two dates. Where the signal-to-noise is strongest, we observe statistical fluctuations less than 1 cm in comparison with SSH (Sea Surface Height) based on conventional Altimetry.

SSH is one of the critical parameters in ocean science allowing to better constrain the ocean spatio-temporal dynamic as it is highly related to sea currents distribution, sea temperature and sea-atmosphere complexes interactions. Besides, it is an important proxy for the geoid estimation. The phase of a side looking SAR imager such as Sentinel 1, on the sea surface, is already used to provide information about the sea surface currents through the Doppler centroid analysis, along the azimuth axis of the image. However, in the range direction, the SAR phase is commonly considered spatially uncorrelated and, therefore, its autocorrelation function is commonly considered to be random and to have no practical use. Following experiments we conducted on focused SLC SAR signals in the open ocean, with Sentinel-1, it emerges that this theorical statement is not entirely satisfactory. Here, we show results on the Indian Ocean and compare them with conventional satellite altimetry, on two dates. This method might have interesting implications not only for improving SSH retrieval using conventional SAR sensors, but also might present interesting perspectives for Earth surface displacement mapping with conventional InSAR.