Developing effective wetland management strategies requires a deep understanding of their natural processes and the changes induced by external pressures. Satellite observations, using both passive and active sensors, offer an excellent opportunity for wetland monitoring from local to global scales, and are often the only source of information in remote and non-instrumented areas.
This session focuses on studies that use multitemporal satellite observations to understand different processes and components (e.g., water dynamics, vegetation changes, disturbances, soil moisture, biodiversity) of wetland ecosystems (e.g., marshes, swamps, fens, bogs, peatlands, lakes, ponds) with different regimes (e.g., permanent, temporary) and support the development of new applications and monitoring strategies.
This session also encourages but is not limited to studies using multi-sensors and machine learning technologies.
Orals: Thu, 7 May, 16:15–18:00 | Room 1.14
Although covering a small fraction of Earth’s surface, wetlands play an important role in the global carbon cycle. They store about 35 percent of terrestrial carbon and have a high capacity for carbon sequestration and long-term retention. However, because the high water table in wetlands often creates anaerobic conditions, they can also emit greenhouse gases (GHG) such as methane and nitrous oxide. When drained, cleared or otherwise disturbed, large amounts of stored organic carbon can be released into the atmosphere as carbon dioxide.
Recent Earth Observation satellite systems such as Sentinels, SWOT or Planet provide new ways to map and capture wetland dynamics at high spatial and temporal resolutions. In combination with earlier missions (e.g., Landsat, PALSAR), they can provide essential information in support of modelling wetland hydrology and biochemistry. At the Global Wetland Center we leverage these multiple satellite systems and sensors to improve global accounting of GHG emissions for wetlands. Our vision is to extend wetland mapping based on categorizations relevant for GHG estimation with continuous priority variables and drivers that can inform about the spatial and temporal dynamics of GHG emissions. Furthermore, making use of differential programming, modern computer vision and knowledge-guided machine learning forced by EO and in-situ observation, we also work towards a better understanding of how management and disturbances (e.g., land conversion, fire, restoration) affect GHG dynamics.
Reducing the uncertainty of the global greenhouse gas budget of wetlands is an ambitious endeavour. The Global Wetland Center started contributing to this grand objective by leveraging methods for large-scale high-resolution mapping of wetland types and of flooded forest extent and inundation frequency. We used machine learning with Sentinel-1, Sentinel-2, and ancillary data to produce a 10-m wetland-type map across Europe, supporting wetland restoration. We also mapped seasonal dynamics of water beneath the forest canopy in the Amazon and Congo basins taking advantage of multi-year SAR data and virtual altimetry stations. Other activities at the GWC focus on using EO to calibrate catchment scale hydrological models in tropical wetlands where in situ data is scarce.
Together, these research activities aim to deliver new observation-driven insights into wetland processes that can directly support improved modelling of greenhouse gas emissions, reducing uncertainties in emission estimates and strengthening the scientific basis for wetland management and climate change mitigation.
More information: https://globalwetlandcenter.ku.dk
Acknowledgement: The Global Wetland Center is funded by the Novo Nordisk Foundation (grant NNF23OC0081089).
How to cite: Horion, S., Senty, P., Kovacs, G. M., Van der Poel, L., Franze, S., Lang, N., Kittel, C. M. M., Tøttrup, C., Fensholt, R., and Schurgers, G.: Space Monitoring of Wetlands for Climate Solutions – the Global Wetland Center initiative, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18201, https://doi.org/10.5194/egusphere-egu26-18201, 2026.
Systematic wetlands monitoring is crucial for informed management, timely conservations actions and consistent temporal reporting. Recent global wetland assessments show that Southeast Asia is losing wetlands at faster than rest of world, with the floodplain wetlands declining by about 1.2% per year, primarily due to agricultural conversion. Thus, making wetlands monitoring and reporting crucial. The national wetland inventory of Thailand, last complied in 2002, requires an update. In this context, the present study introduces a satellite-based Wetland Monitoring Tool (WMT) for assessing five freshwater Ramsar wetlands in Thailand.
The WMT integrates multispectral Sentinel-2 and Synthetic Aperture Radar (SAR) Sentinel-1 observations to classify wetlands during the pre-monsoon period (March to May) from 2019 to 2025 in Google Earth Engine (GEE). Spectral indices (NDWI, MNDWI, NDVI, NDMI, NBR, AWEI, and Tasseled Cap components) are derived to identify pure pixels using Otsu thresholding. Sentinel-1 observations are used to complement optical observation results by improving identification of inundated and flooded vegetation, particularly in scenarios characterized by denser canopy cover or under cloud interference. Final wetland classification is done using a harmonized pixel-based, rule-based framework with an adaptive water detection approach for enhanced class separation across heterogeneous wetland ecosystems. Classification performance is evaluated using Overall Accuracy (OA), Producer’s Accuracy (PA), and the Kappa coefficient.
Results show vegetation increased across all five Ramsar wetlands, especially in Khao Sam Roi Yot (+823 ha) and Lower Songhkram River Wetland (+641 ha). Four of the five wetlands showed decline in open water, with Bung Khong Long Non-Hunting Area losing maximum area ( -163.4 ha; -13%) and Nong Bong Kai Non-hunting Area losing largest proportional area (-67.4 ha; -28%). Only Khao Sam Roi Yot shows increase in open water (+159.4 ha; 6.5%). Emergent and flooded vegetation declined significantly in smaller wetlands, especially in Nong Bong Kai, where they declined 86.2% and 73.4%, respectively. In contrast, land class declined significantly in larger wetlands, particularly in Khao Sam Roi Yot (−1,141.8 ha; -28.4%) and Lower Songkhram (−534.7 ha; −9.8%), indicating rise in wetland vegetation classes within Ramsar boundaries. The integrated optical-SAR approach in WMT enhances wetlands classification, can be scaled at national and regional level and demonstrates potential for standardized, long-term mapping and reporting for improved wetland management and decision making.
Keywords: Wetlands, Ramsar, Sentinel-1, Sentinel-2, Synthetic Aperture Radar, Google Earth Engine, Thailand
How to cite: Prasad, S., Kasambara, T., Saluja, R., and Piman, T.: Monitoring Freshwater Ramsar Wetlands in Thailand Using Integrated Sentinel-1 SAR and Sentinel-2 Optical Observations, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18588, https://doi.org/10.5194/egusphere-egu26-18588, 2026.
Wetlands are dynamic ecosystems whose health and functionality are continually shaped by seasonal fluctuations and long-term shifts in hydrology, climate, and land use. Effective wetlands mapping and monitoring requires methods capable of capturing temporal dynamics, spectral separability, and spatial patterns. Time series satellite observations are invaluable in this regard, as they reveal variations in vegetation phenology, water extent, and other key characteristics over time. To fully leverage temporal information, we employ the Continuous Change Detection and Classification (CCDC) algorithm, which robustly models temporal dynamics by detecting both abrupt and gradual changes, ensuring consistency across seasonal cycles and long-term trends.
To overcome the limitations of individual sensors, we integrate multi-source satellite data. Sentinel-2 provides detailed spectral information related to vegetation conditions and water properties, while Sentinel-1 C-band SAR enables consistent, cloud-penetrating monitoring of surface water dynamics. PALSAR-2 L-band SAR complements them by capturing sub-canopy inundation and vegetation structure. This synergy of optical and multi-frequency SAR data enables a comprehensive characterization of both surface and sub-surface wetland properties across varying environmental conditions.
Deep learning architectures such as U-Net outperform traditional pixel-based classifiers (e.g., Random Forests) by leveraging spatial context for object-level predictions. However, large‑scale wetland typology mapping remains challenging due to input‑dependent label noise arising from the integration of multi‑source maps at various spatial resolutions. We propose an uncertainty‑aware segmentation framework that fuses multi‑source satellite data and explicitly models heteroscedastic aleatoric uncertainty. Concretely, we combine a spatial overlap loss (Dice) with a heteroscedastic negative log-likelihood (NLL) to improve robustness to noisy labels and yield calibrated, per‑pixel uncertainty maps for quality control.
We evaluate the performance of different feature representations derived from multi-source satellite data—including statistical metrics (minimum, maximum, and standard deviation), satellite embeddings, and CCDC-derived temporal features—using both Random Forests and deep learning models. Preliminary results indicate that CCDC features effectively capture temporal wetland dynamics, while spatial context plays a critical role in distinguishing specific wetland types such as marshes, forested wetlands, rivers, and lakes. The resulting uncertainty maps are spatially coherent and consistent with our expectations, showing higher uncertainty along wetland boundaries and lower uncertainty in homogeneous regions, ultimately contributing to more accurate and reliable wetland typology classification.
How to cite: Zhang, P., Druce, D., Kruger, G., Ghariani, W., Kondylatos, S., and Toettrup, C.: Uncertainty-aware Deep Learning for Wetlands Typology Mapping from Multi-Source Satellite Remote Sensing Data , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19100, https://doi.org/10.5194/egusphere-egu26-19100, 2026.
Wetland restoration in Europe under the EU Nature Restoration Law prioritises the recovery of surface water connectivity. However, our understanding of how water moves and connects in European wetlands remains limited, hindering effective restoration and monitoring. To address this, we first analysed changes in surface water extent across ~50 European Ramsar wetlands from 2015 to 2024 using Sentinel-1 data. Second, we developed a novel set of surface water connectivity metrics describing how water bodies emerge, merge, fragment, and persist through time. Quantification of these metrics revealed that ~80% of wetlands exhibit highly variable and unpredictable connectivity patterns across years. While surface water extent generally followed seasonal cycles, the timing of water-body merging and fragmentation diverged from water extent trends in many wetlands. In northern European wetlands, surface water connectivity was stable and predictable across years and closely linked to temperature and precipitation, reflecting strong seasonality and snowmelt regimes. In contrast, many central European wetlands showed irregular connectivity, with large interannual variability in water extent and connectivity often decoupled from simple seasonal wet–dry cycles. These patterns are likely shaped by the interaction of erratic temperature and precipitation, large upstream catchments, and human-modified floodplains. Our results demonstrate how satellite-based monitoring of surface water connectivity can be used to identify distinct hydrological regimes and evaluate future restoration efforts. Furthermore, given the strong heterogeneity in surface water connectivity across Europe, we suggest that restoration should not be evaluated using static extent-based indicators alone, but rather with multitemporal connectivity metrics that reflect how water actually moves and reorganises within wetlands.
How to cite: Robinson, A. and Jaramillo, F.: Sentinel-1 reveals the dynamics of surface water connectivity across European wetlands, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13982, https://doi.org/10.5194/egusphere-egu26-13982, 2026.
Peatland condition is a key indicator of ecosystem integrity and carbon storage potential, making reliable monitoring of degradation processes essential at regional to national scales. Mire breathing—the cyclic surface motion driven by hydrological dynamics—serves as an important proxy for peatland degradation. Drained or degraded peatlands typically exhibit weak oscillatory behaviour combined with persistent subsidence trends, whereas near-natural peatlands show pronounced seasonal surface dynamics. Quantifying peatland subsidence therefore provides a structural indicator for assessing peatland condition and its implications for carbon storage dynamics and greenhouse gas emissions.
In this study, interferometric time-series analysis based on Sentinel-1 Synthetic Aperture Radar (SAR) data was applied to monitor large-scale peatland subsidence using a Small Baseline Subset (SBAS) approach implemented in MintPy. Additionally, hourly Radolan precipitation data were integrated to relate subsidence dynamics to hydrological forcing, and first in-situ measurements from extensometers were used to capture actual ground motion. The analysis covers peatlands across the federal state of Mecklenburg-Vorpommern (north-eastern Germany) for the period 2017–2024 and demonstrates a scalable monitoring framework applicable to national peatland inventories.
The results reveal pronounced spatiotemporal subsidence patterns across the study region. At three representative sites, mean subsidence rates range from −4.3 to −9.6 cm yr⁻¹ in line of sight (LOS). In addition, distinct site-specific mire breathing signals were identified, with seasonal amplitudes between 5 and 15 cm (LOS). The time series show enhanced subsidence during summer months and partial surface recovery during wetter periods, highlighting the strong control of hydrological conditions on peatland surface dynamics.
Overall, the findings demonstrate the capability of SBAS-based InSAR time-series analysis to capture both long-term subsidence trends and short-term oscillatory responses in peatlands. Comparison with in-situ extensometer measurements confirms the validity of the remote-sensing-derived deformation signals. This supports large-scale peatland mapping and monitoring efforts and provides a remote-sensing-based component relevant for greenhouse gas accounting and monitoring, reporting and verification (MRV) frameworks. Future work will focus on improving methodological robustness and validating the InSAR-derived deformation signals using in-situ subsidence measurements.
This research was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Project-ID 531801029 (TRR 410).
How to cite: Pienkoß, L. and Marzahn, P.: Assessing mire breathing patterns across Mecklenburg Vorpommern, Germany using a Sentinel-1 SBAS approach, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4912, https://doi.org/10.5194/egusphere-egu26-4912, 2026.
Spatially explicit knowledge of soil organic carbon and groundwater variability in peatlands is essential to support climate action, such as planning and implementing restoration projects. Geospatial machine learning is a key tool to obtain such knowledge at high spatial resolution, based on linking maps of explanatory variables with point information to train regression or classification models. Explanatory variables are usually identified through expert knowledge and derived from satellite remote sensing, digital elevation models or maps of soil types or land use. Identifying and processing relevant explanatory variables at large scale is non-trivial and cumbersome.
Geospatial foundation models, such as Google’s Alpha Earth change how satellite data and other geospatial data can be utilized in downstream machine learning tasks. Such models provide analysis-ready unified layers, i.e. embeddings, that are semantically rich representations capturing the underlying input data. In the case of Alpha Earth, input data cover archives of Sentinel-1 and Sentinel-2 as well as other geospatial data sources.
In the present study we introduce Alpha Earth embeddings into the modelling of soil organic carbon and groundwater across Danish peatlands at 10 m resolution. We use existing datasets and models of the two variables to benchmark the potential of foundation models for low-barrier large-scale modelling. The models trained against solely Alpha Earth embeddings are contrasted with models trained against explanatory variables selected through expert knowledge as well as with hybrid models combining basic topographical variables with Alpha Earth embeddings.
The Alpha Earth model of soil organic carbon produces meaningful spatial patterns while having a 6% decrease in performance (RMSE) with respect to the expert model. The true positive rate to predict peaty and peat soils is 0.68 and 0.65 for the expert and Alpha Earth model, respectively. The hybrid model increases the performance slightly with respect to the Alpha Earth model and all models achieve very comparable result of mapping the overall peat extent.
The Alpha Earth model predicting groundwater has a 3% performance decrease with respect to the expert model (RMSE). When introducing synthetic training data for drained and wet conditions to the groundwater model, the Alpha Earth model shows limited performance. However, the hybrid model can utilize the synthetic data in a more meaningful way and achieves satisfactory results with respect to performance and spatial patterns.
In addition, we carry out feature importance analysis to explain the Alpha Earth embeddings, which is clear limitation in the usage of foundation models, where explainability is typically not provided.
How to cite: Koch, J., Denager, T., Stisen, S., Greve, M. H., Møller, A. B., and Beucher, A. M.: Modelling Carbon and Groundwater in Peatlands using Alpha Earth Embeddings, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13089, https://doi.org/10.5194/egusphere-egu26-13089, 2026.
Aquatic vegetation contributes to lake methane emissions, but changes in aquatic vegetation in northern (>40° N) lakes remain unknown, hindering evaluations of its importance in estimating lake emissions. Here we use Landsat imagery to monitor aquatic vegetation (mainly emergent and floating vegetation) in 2.7 million northern lakes from 1984 to 2021. Vegetation was observed in 1.2 million lakes, with a total maximum vegetation area of 12.0 × 104 km2, a mean vegetation occurrence of 1.68 ± 3.8% and a greenness of 0.66 ± 0.05. From the 1980s–1990s to 2010s, significant (P < 0.05) increases in maximum vegetation area (+2.3 × 104 km2) and vegetation occurrence (+73.7%) were observed and 72.5% of lakes experienced higher greenness. Vegetation expansion was affected by the temperature in sparsely populated regions, whereas lake area and fertilizer usage played vital roles in densely populated areas. The methane emission estimate that includes contributions from both aquatic vegetation and open water (1.31 [ 0.73, 1.89] Tg CH4 yr−1) is 13% higher than that calculated for open water (1.16 [0.63, 1.68] Tg CH4 yr−1). The long-term net increase in total methane emissions including aquatic vegetation is 125% higher than that of open water due to vegetation expansion. This highlights the necessity of incorporating aquatic vegetation in estimates of methane emissions from northern lakes.
How to cite: Jinying, L., Huabing, H., and Xuejiao, H.: Expansion of aquatic vegetation in northern lakes amplified methane emissions, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5347, https://doi.org/10.5194/egusphere-egu26-5347, 2026.
This study presents a comprehensive framework for assessing current and future ecosystem services in coastal wetlands, by integrating satellite image classification with simplified predictive modelling. The research was conducted in Ria Formosa, one of Portugal's most important coastal lagoons. Hard classification and soft regression Random Forest algorithms were employed to estimate the fractional coverage of marsh zones in 2025, with the former applied to very high-resolution satellite data (Worldview-3, with metric pixel size) and the latter to high and medium-resolution satellite imagery (PlanetScope and Sentinel-2, with metric to decametric pixel sizes). The results provide valuable insights into the challenges associated with variable satellite sources for automated mapping of ecological succession. Future projections up to 2100, informed by land cover change simulations from the SLAMM model, investigated potential ecological trajectories under sea-level rise (SLR) scenarios. Ecosystem adjustments to SLR were further translated into estimates of future blue ecosystem services (i.e., organic carbon stocks), based on reference values reported in the literature. Under the IPCC SSP5-8.5 pathway, significant transitions were projected, including relevant portions of present-day high marshes converting to low marsh, low marshes transforming into tidal flats, and tidal flats devolving into subtidal bare sediment. The modelling framework suggests that coastal squeeze will lead to a meaningful decline in salt marsh extent. The projected ecogeomorphologic adjustments to SLR allowed for pinpointing vegetated areas of gains and losses in carbon stocks by 2100. Foreseen changes will have key implications for the ecological balance of these wetlands, as the significant loss of high marsh habitat may compromise the ecological succession functioning and potentially lead to a decline in biodiversity within these zones. The integrative approach employed, which combines remote sensing and simplified modelling, offers crucial insights into the dynamics and resilience of wetland ecosystemsunder SLR conditions, supporting informed management and conservation efforts in the face of environmental changes.
Acknowledgements: This study contributes to the projects C-Land (CEXC/4647/2024), SYREN (ALGARVE-FEDER-00853600-SYREN-17135), and DEVISE (2022.06615.PTDC), funded by the Fundação para a Ciência e a Tecnologia, as well as to the CLARKS project (CLARKS - 2024-1-ES01-KA220-SCH-000246633) under ERASMUS+ 2024, and to RestLands (ID 705677) funded by Planet Labs.
How to cite: Carrasco, A. R., Kombiadou, K., Carneiro, I., Duarte, J., and Matias, A.: Framework for assessing wetland blue carbon provision through remote sensing and modelling , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11461, https://doi.org/10.5194/egusphere-egu26-11461, 2026.
How to cite: Liu, Y. and Huang, H.: Decade Dynamic Monitoring of Exotic Mangroves in Guangdong Province, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5027, https://doi.org/10.5194/egusphere-egu26-5027, 2026.
The invasion and subsequent removal of Spartina alterniflora in recent years have induced pronounced changes in landscape patterns and ecological processes of China’s coastal wetlands. Long-term and continuous remote sensing monitoring of its dynamics is therefore of considerable scientific and management importance. However, long-term wetland monitoring based on multi-source remote sensing data commonly faces several challenges, including inconsistencies in temporal information caused by asynchronous image acquisition, limited and unstable training samples, and pronounced spatial noise resulting from highly fragmented wetland landscapes. These issues constrain the stability and reliability of long-term classification results. To address these challenges, a spatiotemporal-context-enhanced multi-source remote sensing time-series (SCE-TS) is proposed. First, phenological dynamics from different sensors are preserved through feature-level joint modeling, avoiding phenological distortions introduced by forced temporal alignment. Subsequently, representative and stable temporal prototypes are extracted through repeated feature selection and clustering of local temporal features, and combined with feature enhancement strategies to improve the representation of class-specific temporal characteristics under limited sample conditions. Furthermore, spatial neighborhood convolution is incorporated during feature construction to integrate temporal information from the central pixel and its surrounding neighbors, thereby mitigating the effects of mixed pixels and pixel-level temporal instability. Finally, an improved cascade forest model is employed for classification and mapping. The Yellow River Delta (YRD) and Yancheng wetlands (YC), characterized by distinct geographic settings and landscape structures, were selected as study areas. Using 602 Sentinel-1 and Sentinel-2 images, wetland classification map was generated for the period from 2016 to 2025. Experimental results show that the proposed method achieves overall accuracies of 95.02% in the YRD and 94.48% in the YC, while maintaining stable classification performance across multiple years. Long-term monitoring results further indicate that the removal of Spartina alterniflora has promoted the recovery of wetland vegetation structure; however, post-removal wetland ecosystems still exhibit complex dynamics in terms of wetland area and carbon storage. Overall, the proposed method provides a robust framework for evaluating invasive species management and supporting sustainable wetland ecosystem monitoring.
How to cite: Wang, Z., Liu, Y., Wang, L., Zhao, J., and Lu, Z.: Long-Term Coastal Wetland Mapping Using SCE-TS: A Spatiotemporal-Context-Enhanced Multi-Source Remote Sensing Time-Series Approach, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15636, https://doi.org/10.5194/egusphere-egu26-15636, 2026.
Posters on site: Fri, 8 May, 08:30–10:15 | Hall X1
Since decades, even centuries, blanket and raised bogs in the central-northern United Kingdom have been exposed to climatic and anthropogenic stressors like drainage, land management, evapotranspiration, peat extraction and/or atmospheric pollution. As a result, those landscapes undergo substantial mass changes and soil compaction, but the quantification of these processes remains a challenge: In-situ observations are cost-extensive and/or often biased due to an unstable local measurement reference.
To overcome this, we analyze remotely-sensed surface uplift rates from the European Ground Motion Service (EGMS) that were extracted from four years (2019-2023) of radar-interferometric (InSAR) time-series. We also consulted point-wise uplift rates from two decades of positioning (GNSS) measurements to reference ongoing bedrock uplift. We validate these observations with analytical models that mimic a the load response while also accounting for a remaining glacial isostatic adjustment.
The surface rate maps show ~50 km-wide uplift bulges that correlate with land classified as heathlands and bogs. Maximum uplift surrounding the heather/bogs reaches 2-7 mm/yr. The bogs/heathlands themselves, however, exhibit distinct subsidence due to mass loss (carbon, water) of up to 10 mm/yr, which is already corrected for the simultaneous bedrock uplift due to unloading. Based on these observations, we can reproduce the spatial bedrock uplift pattern with our load model. The explanation of the signal amplitudes requires further fine-tuning of the model parameters and a better understanding of the in-situ bio-chemical processes. This approach will enable us to quantify the amount of water-vs.-carbon loss in this particular landscape.
How to cite: Metzger, S., Dill, R., Klemann, V., Sachs, T., and Zahra, A.: Quantifying mass loss in raised bogs and heathlands in the central-northern United Kingdom from satellite-geodetic uplift rates, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14452, https://doi.org/10.5194/egusphere-egu26-14452, 2026.
Monitoring of peatland groundwater levels is crucial for effective bog rewetting. Traditional in-situ measurement methods are often costly and impractical in countries with expansive peatland areas. Synthetic Aperture Radar (SAR) may be an approach for remotely estimating groundwater levels over large areas due to its sensitivity to soil properties, such as soil moisture. In this study, we investigated the relationships between terrain-corrected Sentinel-1 C-band SAR backscatter intensity (γ0) and groundwater level (GWL) at two temperate raised bogs: (1) a near-intact site (Cors Fochno bog, Wales, United Kingdom; 11 dip wells; Area = ~ 6.3 km2); and (2) an industrially-extracted, ‘bare-peat’ site (Castlegar Bog, Co. Galway, Ireland; 34 dip wells; Area = ~ 3.2 km2). Both sites have recently undergone rewetting measures, primarily bunding and drain blockage. For the industrially extracted Castlegar Bog, initial linear regression analysis between γ0 and GWL yielded average correlation coefficients (r) of 0.33 and 0.47 for VV and VH polarization, respectively. However, average correlation values increased when the dataset was separated into pre- and post-rewetting periods. Values of 0.55 and 0.64 for VV and VH, respectively, were found before restoration, and 0.43 and 0.54 for VV and VH, respectively, were found after restoration. For the near-intact Cors Fochno bog, SAR intensity exhibited very weak correlation with GWL, with average r values of 0.34 and 0.16 for VV and VH polarizations, respectively. Average correlation values changed to 0.41 and 0.14 for VV and VH after accounting for and filtering out rainfall events preceding each acquisition. Consequently, our results indicate a limited capability of SAR backscatter intensity to serve as a reliable proxy for GWL in near-intact temperate raised peatlands. We hypothesize that the limited correlation is attributable to two main factors. Firstly, GWL in near-intact sites typically remains approximately 10 cm below the surface with minimal fluctuation, thereby maintaining a consistently saturated peat layer and limiting variance in dielectric properties beyond background noise levels. Secondly, vegetation acts as a buffer, temporarily retaining rainfall in above-ground and near-surface layers, which increases local volumetric water content, leading to surface saturation that affects the SAR backscattering mechanism. On the other hand, these findings indicate that SAR-based monitoring of GWL using C-band data is effective in highly degraded or extracted temperate peatlands, where water table fluctuations are pronounced and where vegetation impacts on the SAR signal are reduced due to extensive bare peat exposure.
How to cite: Khoshlahjeh Azar, M., Hrysiewicz, A., Donohue, S., Regan, S., Renou-Wilson, F., Reddin, E., Williamson, J., and P. Holohan, E.: Assessing the Potential of Sentinel-1 SAR Backscatter intensity for Monitoring Groundwater Levels in Temperate Raised Bogs, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20531, https://doi.org/10.5194/egusphere-egu26-20531, 2026.
Raised bogs are ombrotrophic peatland ecosystems whose long-term functioning depends on the stability of a vertically growing peat surface sustained by tightly coupled ecohydrological feedbacks. In intact systems, the peat surface undergoes reversible seasonal shrink–swell behaviour (“bog breathing”), allowing bogs to self-regulate in response to water-table fluctuations. Prolonged hydroclimatic stress or artificial drainage for land use activities can disrupt this oscillatory regime, leading to peat consolidation, loss of water-storage capacity, and progressive surface subsidence. Despite extensive national and EU-level protection, the stability of near-intact raised bogs has rarely been assessed at national scales.
Here, we use vertical ground-motion time series from the European Ground Motion Service (EGMS) data to quantify seasonal and interannual peat surface dynamics across raised bog Special Areas of Conservation (SACs) in Ireland between 2019 and 2023. Surface elevation changes were analysed for 52 paired raised bog sites representing both active raised bogs (habitat type 7110) and degraded raised bogs capable of natural regeneration (7120), using over 39,000 EGMS measurement points located within bog SAC boundaries. Long-term elevation trends were quantified using linear regression, alongside analysis of seasonal surface oscillations associated with bog breathing.
Across the study sites, almost all raised bogs exhibit clear seasonal surface oscillations alongside a persistent decline in mean surface elevation over the five-year observation period. Across the protected sites at the national scale, this corresponds to a median surface lowering of approximately 3 mm per year, with similar magnitudes observed in both active and degraded raised bogs. Mean subsidence rates are slightly more negative but remain within a narrow range, and variability across sites is moderate. These preliminary results indicate that long-term surface lowering represents a shift away from stable peat surface equilibrium in raised bogs designated for protection in Ireland, affecting not only degraded sites but also bogs classified as “active”.
Our findings indicate that water-level monitoring alone may not be sufficient to assess raised bog condition. Declining surface elevation reduces peat specific yield, meaning that apparently stable or high water levels can mask a loss of hydrological storage capacity and self-regulation. Consequently, raised bogs may appear hydrologically “healthy” while undergoing structural degradation and progressive subsidence. The surface elevation decline observed across almost all protected raised bogs highlights the need to integrate surface motion metrics into peatland monitoring, conservation assessment, and restoration planning to avoid irreversible ecohydrological degradation.
How to cite: Ahmad, S., Habib, W., and Liu, H.: Satellite-derived evidence of recent peat surface elevation decline across the protected raised bogs of Ireland , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21222, https://doi.org/10.5194/egusphere-egu26-21222, 2026.
Peatland degradation is a critical global climate issue, releasing millions of tonnes of CO₂ annually due to drainage and changes in land use. As countries strive to meet net-zero targets, restoring degraded peatlands has become a priority for carbon sequestration and biodiversity conservation. However, monitoring peatland recovery remains a challenge, especially for large-scale restoration projects. This research is driven by the need for low-cost validation of peatland re-wetting schemes, enabling robust monitoring of peat physical condition and hydrological recovery, with implications for carbon accounting and agriculture’s contribution to net-zero targets. This study addresses this gap by applying remote sensing data from Interferometric Synthetic Aperture Radar (InSAR) to track peat surface motion in Cambridgeshire's Great Fen, one of the UK’s largest lowland peatland restoration initiatives. Sentinel-1 InSAR data (2015–2025) were used to quantify ground motion and derive deformation-based proxies for peat carbon flux. Our analysis revealed distinct subsidence patterns for undrained, early-restored, and later-restored farms, enabling first-order, deformation-based carbon flux estimation under common parameter assumptions. Early-restored farms experienced subsidence rates of up to 1.17 cm/year and deformation-associated carbon flux proxies of 14.50 tons CO₂/ha/year, compared to 1.40 cm/year and 17.37 tons CO₂/ha/year in later-restored sites. National Nature Reserves (Holme Fen and Woodwalton), which remained undrained, recorded the lowest subsidence (~0.48 cm/year) and lowest deformation-associated carbon loss proxy (5.98 tons CO₂/ha/year), linked to restoration timelines and peat moisture regimes. These estimates, interpreted as relative indicators rather than direct measurements of net ecosystem carbon balance, demonstrate InSAR’s utility for tracking peatland condition and relative peat carbon vulnerability across restoration timelines. Seasonal fluctuations aligned with soil moisture and precipitation anomalies, indicating a strong hydrological control on peat surface motion. Together, these findings show that InSAR provides a high-resolution, cost-effective tool for continuous monitoring of peatland physical dynamics, supporting comparative assessment of restoration outcomes and climate-relevant land management decisions.
How to cite: Khouakhi, A., Patil, A., Girkin, N., and Holman, I.: Assessing Peat Surface Motion using Interferometric Synthetic Aperture Radar (InSAR) in The Great Fen Area of Cambridgeshire, UK , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10244, https://doi.org/10.5194/egusphere-egu26-10244, 2026.
Tropical wetlands are among the Earth’s most critical ecosystems, playing a vital role in regulating global water and carbon cycles, buffering extreme weather events, and supporting biodiversity that sustains millions of people. Despite their importance, these ecosystems are highly vulnerable to climate change, and our understanding of their seasonal extent, their role in climate mitigation, and their response to changing climatic conditions remains limited. This lack of knowledge hinders the development of effective climate adaptation strategies and constrains projections of future carbon emissions. To address these gaps, this study and related ongoing work aim to develop an integrated framework that combines multiple methodologies, data sources, and analytical tools to improve the monitoring of surface water inundation in major floodplain systems.
A key component of this framework is understanding the availability, characteristics, and interpretative value of different remote sensing datasets. As a case study, this work focuses on two major wetland systems: the Sudd in South Sudan and the Pantanal in South America. Satellite observation datasets are compiled and categorized into static and dynamic products. The static datasets include GlobCover, GLWDv2 and SWAMP, while the dynamic datasets comprise WAD2M, JRC Global Surface Water (GSW), CYGNSS water mask, and GRACE Total Water Storage. Static datasets are used to assess long-term changes in wetland extent and classification, whereas dynamic datasets capture seasonal variability in wetland extent. Together, these datasets enable comparison between seasonal dynamics and long-term trends, providing improved insight into future projections of wetland extent. The dynamic datasets are projected onto a common grid to assess their consistency and agreement with one another. These datasets are also used to develop a Long Short Term Memory (LSTM) network, capable of capturing both seasonal variability and long-term trends, which is then applied to project future wetland dynamics.
This work represents an important first step towards reducing uncertainty in global wetland mapping. Building on this foundation, the study aims to use the Joint UK Land Environment Simulator (JULES) to simulate wetland extent and hydrological dynamics across selected tropical wetland regions (Sudd, Pantanal). Model simulations are driven by newly developed ancillary inputs, including land cover parameters, soil properties, and topographic information, to assess their influence on simulated wetland extent and seasonal flooding patterns. The resulting JULES outputs are systematically evaluated against EO-based wetland datasets such as GLWDv2, GlobCover, and WAD2M to identify areas of agreement, model sensitivities, and potential sources of bias. Through this comparative analysis, the study benchmarks the capability of the JULES land surface model to represent tropical wetland dynamics and provides insights into optimal data configurations for large-scale wetland modelling.
As the project develops, machine learning approaches are further applied to forecast wetland dynamics and to inform improvements in the representation of wetlands within climate models. Ultimately, this integrated modelling and data-driven framework aims to contribute to more reliable climate predictions and to provide decision-makers with clearer, evidence-based information for climate adaptation and mitigation planning.
How to cite: Pantula, C., J Parker, R., Balzter, H., Ruiz Villena, C., R Marthews, T., and Fatima, K.: Understanding changes in tropical wetlands with remote sensing, machine learning and land surface modleing, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3075, https://doi.org/10.5194/egusphere-egu26-3075, 2026.
Wetlands contribute to human well-being in multiple ways. These contributions take various forms, including goods such as food and other raw materials, but also through services like carbon sequestration, flood and climate regulation. Despite their importance, wetlands are being lost at annual rates exceeding 0.5% of their global extent over the past sixty years, mainly due to conversion to other land uses. Several multilateral agendas emphasize the importance of wetland monitoring and protection; however, progress toward established targets remains limited by the lack of consistent national datasets and operational monitoring tools. Satellite observations provide globally consistent data that enable systematic wetland monitoring. In particular, synthetic aperture radar (SAR) has been successfully used for mapping wetland flooding dynamics due to their all-weather acquisition capability and the ability to identify below canopy inundation processes. The increasing availability from current and planned SAR missions poses an opportunity to advance space-based wetland monitoring, but their operational implementation requires the development of flexible and scalable frameworks.
This study aims to develop WetSAT-ML, a satellite-based machine learning (ML) framework for mapping wetland flooding dynamics and trends using Sentinel-1 SAR data. The approach combines radar features with supervised and unsupervised ML classification algorithms, to distinguish different inundation categories, including open water, vegetated water, and land. Random Forest and K-means models were trained using two training and validation areas in the South Florida Everglades (USA) and the lower Atrato River Basin (Colombia). These sites were selected considering their wetland variability and the availability of reference data. The first version of the models was trained using all available Sentinel-1 observations from 2024 over the selected regions, capturing the full hydrological seasonality. Reference training and validation datasets included gauge measurements and manually annotated data for both regions. The trained WetSAT-ML models are being evaluated through a proof of concept across five wetland systems in South Asia and South America, including the Meghna River wetlands in Bangladesh; the Atrato River, Ayapel, and Barbacoa wetlands in Colombia; and the Pantanal wetlands spanning the Brazil–Bolivia border. The test sites represent a wide range of hydroclimatic conditions, geomorphological settings, and vegetation cover.
Preliminary results indicate that WetSAT-ML consistently captures spatial patterns and intra-annual inundation dynamics that are coherent with the known hydrological regimes of the test regions. Cross-site comparisons reveal clear differences in key hydroperiod parameters related to inundation persistence, seasonal amplitude, and ecosystem connectivity. Overall, the results provide a foundation for operational wetland monitoring applications. WetSAT-ML is open access, and the first public version is available in a GitHub repository: https://github.com/sei-latam/WETSAT_v2. The next steps of the research will focus on cross-validation using independent datasets and expanding the training database across additional wetland areas.
How to cite: Palomino-Ángel, S., Méndez, C., Zamora, D., Santos, T., Prasad, S., and Piman, T.: WetSAT-ML: A Machine Learning Framework for Mapping Wetland Flooding Dynamics using Sentinel-1 Observations, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-16380, https://doi.org/10.5194/egusphere-egu26-16380, 2026.
Wetlands are among the most hydrologically dynamic ecosystems, particularly across northern boreal and subarctic regions of Canada, where seasonal freeze-thaw cycles and precipitation variability drive pronounced fluctuations in surface water and soil moisture. These landscapes play a critical role in carbon storage, ecosystem functioning, and regional hydrology, yet their remoteness severely limits the availability of long-term in situ soil moisture observations. Consequently, satellite-based microwave remote sensing has become an essential tool for monitoring wetland hydrological dynamics at large spatial scales.
NASA’s Soil Moisture Active Passive (SMAP) mission provides global soil moisture estimates at a coarse spatial resolution (~36 km). However, retrieval performance declines significantly in wetland-dominated regions due to the mixed influence of open water, saturated soils, and vegetation within a single footprint. This mixture complicates the interpretation of passive microwave brightness temperatures and increases uncertainties in soil moisture products. Improving SMAP performance in these environments requires a better understanding of how water dynamics influence the satellite signal.
In northern Canadian wetlands, small surface water bodies such as shallow ponds, ephemeral pools, and saturated depressions exhibit substantial seasonal variability, especially during snowmelt and early summer. Although individually below SMAP’s resolution, their aggregated extent may substantially affect observed brightness temperatures and mimic soil moisture variability. This study investigates whether temporal changes in small surface water extent can serve as a proxy for soil moisture variations within SMAP footprints.
We analyzed the relationship between in situ soil moisture, SMAP brightness temperatures, and surface water extent across two wetland regions: the Attawapiskat River (CA-ARB and C-ARF) and the Kinosheo Lakes (CA-KLP). Soil moisture data from eddy covariance flux towers (2017-2021) were used for snow- and ice-free periods (June-October). Small surface water bodies were mapped using Sentinel-1 SAR imagery and the Canadian Digital Elevation Model (CDEM) data through a random forest classification approach, then aggregated to the SMAP footprint scale for analysis.
Results show strong correlations between in situ soil moisture and surface water extent (r > 0.58), as well as between surface water extent and SMAP brightness temperatures (r > 0.77). These findings indicate that surface water dynamics capture spatially representative hydrological variability within SMAP pixels and help address scale-mismatch issues between point in situ measurements and coarse satellite products. The study demonstrates the potential of surface water extent as a proxy variable to support calibration, validation, and future improvement of SMAP soil moisture retrievals in wetland-dominated regions.
How to cite: Amini, Y., Millard, K., Berg, A., Humphreys, E., and Richardson, M.: Soil Moisture Correlation with Small Water Body Extents in Canadian Wetlands: Application to SMAP Soil Moisture Improvement, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-16084, https://doi.org/10.5194/egusphere-egu26-16084, 2026.
Emergent wetlands occupy the transitional zone between permanent wetlands and upland environments and provide a range of ecosystem services. These services strongly depend on inundation patterns and vegetation dynamics, which are often difficult to monitor using in-situ approaches due to limited accessibility, dense vegetation, and the need to minimise disturbance, particularly in protected areas. Remote sensing offers an effective means to overcome these constraints by providing consistent information over large areas in a timely and cost-efficient manner. Synthetic aperture radar (SAR), in particular, is well suited due to its sensitivity to water occurrence beneath vegetation canopies. In emergent wetlands, double-bounce scattering of the radar waves between standing water and vertical vegetation components typically results in elevated backscatter and coherence compared to other land-surface types. However, this response is influenced by several factors, including radar wavelength, vegetation structure, and water level. We analyse time series of Sentinel-1 backscatter intensity and coherence with the goal of characterising inter and intra-annual variations in surface water extent between 2015 and 2025. We interpreted the SAR-derived metrics using a comprehensive reference dataset including water level, high-resolution imagery from unmanned aerial vehicles, meteorological data and vegetation indices. The study was conducted at a long-term ecosystem research site at the shallow, subsaline Lake Neusiedl, located in the Pannonian lowlands of Eastern Austria. The lake is a Ramsar site of international importance due to its ecological significance, particularly for breeding and migratory birds. More than half its surface is covered by one of the largest continuous reed belts in Europe, dominated by Phragmites australis. During spring, the reed belt shows a clear double-bounce signature whereas in summer, high vegetation and declining water levels lead to a decrease in backscatter and coherence. During a prolonged drought period, which lasted from 2019 to 2022, water extent at Lake Neusiedl decreased significantly followed by a marked recovery starting in 2023. Our results showcase both potential and limitations of water extent retrieval in emergent wetlands based on C-band SAR data and hold important lessons for future wetland monitoring using data acquired at longer wavelengths by SAR missions, such as NISAR and ROSE-L. In addition, the observed coherence patterns offer initial indications of the potential for retrieving wetland water level changes using SAR interferometry.
How to cite: Schlaffer, S., Dorninger, P., and Qiu, S.: Sentinel-1 SAR backscatter and coherence patterns during dry and wet periods at a large emergent wetland in the Pannonian lowlands, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17935, https://doi.org/10.5194/egusphere-egu26-17935, 2026.
Inland freshwater aquaculture which includes a new crop-aquaculture system accounts for 77 % of aquaculture production worldwide and contributed significantly to the global demand for fish products. Previous aquaculture monitoring efforts, however, mainly focused on coastal aquaculture ponds and hardly covered inland freshwater aquaculture areas. Here, based on the time-series Sentinel-1 and 2 data and the Google Earth Engine (GEE) platform, we developed a hierarchical framework for mapping inland freshwater aquaculture with different aquaculture types (both aquaculture ponds and rice-crayfish fields) in the Jianghan Plain (JHP), one of the most important inland freshwater aquaculture regions in China. First, we constructed two phenological temporary windows (T1 and T2) and used an automatic threshold extraction method (OTSU approach) to generate potential aquaculture layers in both two temporary windows. Second, based on the potential aquaculture layer in T1, the actual aquaculture was further distinguished from other water bodies (e.g., lakes, rivers, and ditches) by combining spectral and texture features and utilizing a random forest classifier from 2016 to 2022. Finally, by the differences in variations in aquaculture ponds and rice-crayfish fields in T1 and T2, we generated annual 10m maps of fine aquaculture area in the Jianghan Plain from 2017 to 2022, with overall accuracies (OA) of 87.5 %–98.7 % and Kappa coefficients of 0.81–0.98. We found a significant increase in the total aquaculture area in the JHP, from 2266 ± 66 km2 in 2016 to 4766 ± 111 km2 in 2022, with rice-crayfish fields contributing the most, mainly related to a series of stimulus policies. This study proposed a novel framework for monitoring complex inland freshwater aquacultures over large areas and revealed the rapid expansion of inland aquaculture in South China.
How to cite: Wang, C., Dong, J., Zhou, Y., Cui, Y., Chen, X., Di, Y., Xiao, X., and Zhang, G.: Satellite observed rapid inland aquaculture expansion in Jianghan Plain, China from 2016 to 2022, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11704, https://doi.org/10.5194/egusphere-egu26-11704, 2026.
