PICO
Remote sensing of seasonal snow is a key tool in this regard, as it provides a wide range of techniques and data across various spatial and temporal scales. This technology is essential for monitoring snow properties and their hydrological impacts, enabling a better understanding of the interaction between snow and its environment at a small scale, rapid snow changes, rain-on-snow events, and snow-vegetation interactions.
This session focuses on studies linking remote sensing of seasonal snow to hydrological applications to: (i) quantify snow characteristics (e.g., SWE, snow grain size, albedo, pollution load, snow cover area, snow depth and snow density), (ii) understand and model snow-related processes and dynamics (snowfall, melting, evaporation, wind redistribution and sublimation), (iii) assess the snow hydrological impacts and snow environmental effects.
We welcome contributions that integrate methods and data from diverse technologies, including time-lapse imagery, laser scanning, radar, optical photography, thermal and hyperspectral sensing, as well as emerging applications, across a range of spatial scales (from plot-level to global) and temporal scales (from instantaneous observations to multi-year time series).
Snow cover variability in high-altitude Himalayan river basins plays a critical role in regulating regional hydrology, water availability, and climate-cryosphere interactions. Understanding its temporal behaviour is particularly important for snow-fed basins such as the Budhi Gandaki Basin, Nepal, where seasonal meltwater significantly influences downstream flow regimes. Previous studies often overlooked detailed monthly-scale time series analysis and robust trend diagnostics of snow cover variability. In this study, we investigate the temporal dynamics of monthly snow cover area in the Budhi Gandaki Basin (area: 3857.85 km²) using continuous observations from 2020 to 2024 using the Moderate Resolution Imaging Spectroradiometer (MODIS) 8-day snow cover products (MOD10A2) with 500 m resolution. A comprehensive time-series framework was applied, incorporating descriptive statistics, linear trend analysis, seasonal climatology, anomaly assessment, and non-parametric trend detection. The MOD10A2 snow cover composites were spatially averaged over the basin and temporally aggregated to monthly resolution using a proportional day-overlap weighting scheme to estimate basin-averaged monthly snow covered area (SCA). Monthly snow cover data were then transformed into a continuous time series to evaluate overall variability and long-term behaviour. Seasonal characteristics were quantified through monthly climatology and interannual variability using mean and standard deviation metrics. Trend significance was examined using the Mann-Kendall test, while the magnitude of change was estimated using Sen’s slope. Results reveal a mean snow cover index of 0.56 with substantial variability showing a standard deviation of 0.16 snow cover fraction (SCF), indicating pronounced seasonal and interannual fluctuations. Monthly climatology shows maximum snow cover during winter months, peaking in February (mean: 0.762 SCF) and January (mean: 0.756 SCF), while minimum values occur during the summer monsoon period, particularly in July (mean: 0.424 SCF) and June (mean: 0.426 SCF). Linear trend analysis indicates a gradual declining tendency in snow cover at a rate of -0.0024 SCF per month. However, the Mann-Kendall test yields a Z statistic of -1.77 with a p-value of 0.076, suggesting that the observed decreasing trend is not statistically significant at the 95% confidence level. Anomaly analysis further highlights episodic deviations from the climatological mean, with maximum positive and negative anomalies of +0.204 SCF and -0.162 SCF, respectively, reflecting short-term climate-driven variability. Overall, the findings indicate a weak but persistent declining tendency in snow cover, modulated strongly by seasonal and interannual variability rather than a statistically significant long-term trend. This study provides an improved understanding of snow cover dynamics in the Budhi Gandaki Basin and offers valuable insights for hydrological modeling, climate impact assessments, and sustainable water resource management in snow-fed Himalayan river systems.
Keywords: Snow cover variability; Time series analysis; Himalayan river basin; Seasonal climatology; Trend detection
How to cite: Tiwari, H., Thakur, A. K., and Garg, R. D.: Temporal Variability and Trend Analysis of Monthly Snow Cover in a Himalayan river basin, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2219, https://doi.org/10.5194/egusphere-egu26-2219, 2026.
Seasonal snow cover in high mountain regions plays an important role in river flow hydrographs, water availability and land atmosphere interactions. In the Western Himalaya, mapping snow cover accurately is difficult due to frequent cloud cover, steep terrain, strong shadow effects and confusion between snow, clouds and bright rocky surfaces in satellite images. The study shows that these issues are clearly observed in the Western Himalayas, where the lack of sufficient ground reference data further limits the use of fully supervised classification methods. To address these challenges, the objective of the present study is to develop a snow cover mapping framework that leverages self-supervised learning for robust feature representation from satellite remote sensing data in the Western Himalaya. The method focuses on learning useful snow related features directly from large volumes of unlabelled satellite images, reducing the need for extensive manually labelled training data. Multi temporal optical satellite images are used so that the model can learn stable snow patterns across different seasons, illumination conditions and surface states. A convolutional neural network is trained using a contrastive self-supervised learning strategy, where different augmented versions of the same image patch are treated as similar samples, while patches from different locations are treated as dissimilar. The self-supervised encoder is coupled with a lightweight decoder in an encoder-decoder segmentation architecture, enabling pixel wise snow mapping while preserving spatial detail under limited supervision. Simple data augmentations, such as brightness changes, contrast adjustments and random cropping are applied to improve the model’s ability to recognize snow under varying conditions while preserving its key spectral and spatial characteristics. After self-supervised pretraining, the learned feature representations are fine tuned for snow and non-snow classification using a limited set of labelled samples derived from reference snow products and manual interpretation. This greatly reduces the dependence on large labelled datasets compared to conventional supervised learning methods. Snow cover maps are generated for different seasons and elevation zones to examine spatial and temporal variability of snow distribution across the basin. The results are compared with traditional index based methods, such as Normalized Difference Snow Index (NDSI) thresholding, especially in areas affected by clouds, shadows and mixed land cover. The study shows that the self-supervised learning provides a practical and reliable framework for snow cover mapping in data scarce and high altitude regions. The methodological framework developed in this study can be utilized for other basins also to have improved understanding of snow cover dynamics.
Keywords: Snow cover; self-supervised learning; remote sensing; Himalaya
How to cite: Thakur, V., Keshari, A. K., and Tak, S.: Monitoring of Spatio-Temporal Snow Cover using AI Based Self-Supervised Learning in Data Scarce Himalayan River Catchment, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3445, https://doi.org/10.5194/egusphere-egu26-3445, 2026.
Reliable information on snow cover dynamics is essential for water resource management, climate impact assessments, and ecological studies. In mountainous regions, where spatial variability is high and long-term observations are limited, satellite-based snow products often present the primary source of information. However, their performance is constrained in complex terrain by cloud cover, coarse spatial resolution, and data gaps. Therefore, the validation of satellite-derived snow cover data is crucial for reducing uncertainty. In this study, we employed a low-cost, ground-based time-lapse camera to monitor snow cover (SC) and to support the validation, gap-filling, and improved reliability of satellite-derived snow cover products.
Time-lapse photography was obtained using a camera trap installed at the Skalnaté Pleso Observatory in the High Tatra Mountains (Slovakia). The camera captured daily images of a south-eastern slope during four snow seasons (2021/22–2024/25). An automated image-processing workflow was applied to derive snow cover percentage from the photographs, including horizon-based image alignment, masking of non-relevant areas, and automatic snow classification based on blue-band intensity thresholds. The resulting camera-derived SC was compared with satellite-based fractional snow cover (FSC) from Sentinel-2 products (Fractional Snow Cover and Gap-filled Fractional Snow Cover marked as S_FSC and S_GFSC) and MODIS products (MOD10 and MYD10 marked as M_TERRA_FSC and M_AQUA_FSC) within the camera’s field of view.
The analysis revealed substantial differences in data availability between ground-based and satellite observations, with time-lapse photography providing more continuous records during periods of frequent cloud cover. Camera-derived SC captured short-term snow accumulation and melt dynamics that were often missed or temporally smoothed in satellite products. Relative to camera observations, Sentinel products overestimated SC by 11.3 % (S_GFSC) and 9.1 % (S_FSC), whereas MODIS products underestimated SC by -9.5 % (M_AQUA_FSC) and -7.7 % (M_TERRA_FSC).
Data gaps in satellite products were addressed using a Random Forest machine-learning approach trained on SC derived from terrestrial time-lapse photography. To avoid sensor-mixing biases, separate models were trained for each Sentinel-2 and MODIS product. By integrating local meteorological variables such as daily air temperature, precipitation, snow depth, and global radiation, the models were able to capture the non-linear nature of snow dynamics. Our study demonstrates that combining time-lapse photography with satellite products and in situ meteorological measurements enables more accurate reconstruction of snow cover dynamics, particularly in periods of rapid snow accumulation and melt in alpine environments.
Acknowledgement: This study was funded by the project VEGA 2/0048/25.
How to cite: Krempaský, J., Lukasová, V., Mrekaj, I., Onderka, M., and Varšová, S.: Machine Learning–Based Gap-Filling of Satellite Snow Products Using Time-Lapse Photography and Meteorological Data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10848, https://doi.org/10.5194/egusphere-egu26-10848, 2026.
The presence or absence of snow in the landscape strongly modulates land surface energy exchanges and governs key ecosystem processes in high-mountain catchments. In mid-latitude mountain regions, such as the Pyrenees, which are dominated by intermittent and ephemeral seasonal snowpacks, pronounced intra-annual spatial variability in snow cover complicates the accurate characterisation of snow temporal dynamics.
Although a wide range of snow products and remote sensing platforms are currently available, many of them have significant limitations when applied to complex, mountainous environments such as the Pyrenees. These limitations include data gaps caused by cloud cover and confusion between snow and clouds; reduced accuracy in areas affected by topographic shadows; insufficient illumination due to low solar elevation at the time of satellite overpass; and the trade-off between spatial and temporal resolution. Furthermore, as most existing products are designed for large-scale applications, they can introduce significant errors when high spatial detail is required. This is particularly pertinent in catchment- and sub-catchment-scale hydrological, biogeochemical, and ecological studies.
In this context, we propose a methodological approach that combines the daily temporal resolution of snow gap-filled MODIS products with Sentinel-2-derived snow cover as the ground truth, using k-nearest neighbour (k-NN) classification. We generated daily binary snow presence/absence maps at a spatial resolution of 20 m over the study area using a logistic regression model incorporating general explanatory variables such as elevation, slope, aspect, monthly solar radiation and the spatial and temporal information of snow cover, such as distance-to-snow maps derived from MODIS.
Preliminary results show that the logistic regression framework generates daily snow cover maps that are spatially and temporally consistent, substantially reducing data gaps and improving the representation of intermittent and ephemeral snow zones, which are expected to become increasingly prevalent under future climate change. Model outputs were evaluated against independent ground-based observations, including snow pole measurements, telenivometer data, showing good agreement across elevation gradients and seasons. Together, these results demonstrate the potential of the proposed approach to capture fine-scale spatio-temporal variability in snow cover, providing a robust basis for catchment-scale analyses of snow–hydrology and snow–biogeochemistry interactions in high-mountain regions.
How to cite: Navarro Planes, M., Pons, X., and Gómez Gener, L.: Estimating daily high-resolution snow cover in the Central Pyrenees using logistic regression with Sentinel-2 and MODIS data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19704, https://doi.org/10.5194/egusphere-egu26-19704, 2026.
Modelling streamflow in mountainous areas is challenging. The presence of snow is an additional factor to consider, as it is the primary driver of streamflow dynamics in mountain catchments. In addition, this complexity increases in the Mediterranean mountains, where snow dynamics are more variable, with specific characteristics, including shallow snowpack, high density, and evaposublimation rates that cannot be neglected when assessing water resource availability during the dry season. Many approaches, with varying levels of complexity, have been used to model streamflow in mountainous areas. In general, these models are calibrated and evaluated against streamflow without considering their performance with respect to snow dynamics. That is, river discharges are well represented, but not due to the correct reasons.
This study assesses the implications of selecting snow parameterizations for streamflow modelling in Mediterranean mountain catchments, considering not only streamflow but also snow performance. Five different hydrological models, with different conceptualizations – lumped, semi-distributed, and fully distributed – and with different levels of complexity regarding snow parameterization – degree-day, radiation-day, and mass and energy balance approach– have been used. These models are: (1) GR4J associated with CemaNeige (lumped with degree-day snow model), (2) SWAT (semidistributed with degree-day snow model), (3) HYPE (semidistributed with radiation-day snow model), (4) GEOFRAME (semidistributed with temperature-radiation-day snow model), and (5) WiMMed (distributed with mass and energy balance snow model). Models were calibrated against streamflow observations and evaluated for snow performance using remote-sensing-derived snow-cover area. A spectral mixture analysis carried out using Landsat imagery, considering the three main land cover types over the region: snow, shallow vegetation, and rocks, was performed to define the fraction of snow in each cell. The values of these pixels were aggregated at the catchment scale for comparison with the simulations. The Guadalfeo River basin in southern Spain has been selected as representative of a Mediterranean mountain-coastal catchment for this analysis.
Preliminary results indicate that the complexity of snow dynamics is better captured by the more complex approach, namely, the fully distributed mass and energy balance snow model. However, the assessment indicates that simpler approaches can be valid when analyzing changes and seasonality rather than actual values. This observation underscores the potential to use this model in an ensemble to compute hydrological uncertainty, as is common in hydrological seasonal prediction and climate studies.
Acknowledgments: This work is part of the project PCI2024-153496, funded by MCIU/AEI/10.13039/501100011033 and EU
How to cite: Gómez-Beas, R., Egüen, M., Formetta, G., Polo, M. J., and Pimentel, R.: Are snow patterns well modelled when simulating river discharge in Mediterranean mountain catchments? A multimodel approach assessment using remote sensing data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17516, https://doi.org/10.5194/egusphere-egu26-17516, 2026.
The COOL project, funded by the Italian Space Agency (ASI), focuses on the development of an advanced, modular Level-2 (L2) processor for the PRISMA Second-Generation (PRISMA-SG) hyperspectral mission. The processor is designed to generate high-quality L2 products, including surface reflectance, water vapor content, and aerosol optical thickness, while addressing the unique challenges introduced by the off-nadir acquisition geometry of the PRISMA-SG sensor.
The processor builds upon state-of-the-art radiative transfer modeling and integrates physics-based atmospheric and topographic correction algorithms based on the MODTRAN6 model. The processing chain is derived and extended from previous work (Santini & Palombo, 2019; Palombo & Santini, 2020; Santini & Palombo, 2022), and incorporates second-order effects, such as adjacency corrections and topographic illumination variations. These algorithms are carefully adapted to the spectral, spatial, and viewing geometry characteristics of PRISMA-SG, aiming to achieve or exceed a Scientific Readiness Level (SRL) of 6.
Validation of the processor relies on both simulated datasets and in-situ measurements over dedicated calibration and validation (CAL/VAL) sites established within the COOL project. Top-of-atmosphere (TOA) radiance signals are simulated over these sites and compared with field measurements to quantify residual errors and assess the sensitivity of the inversion algorithms to off-nadir acquisition effects. These activities ensure the robustness and scientific usability of the derived L2 products in both nadir and off-nadir observation modes.
As a demonstration, the L2 processor was applied to PRISMA-SG images acquired over snow-covered areas in the Italian Alps. The results were compared with the standard L2 products provided by the image supplier. The comparison shows close general agreement in reflectance spectra while correcting artifacts present in the standard products, including topographic effects, adjacency effects, and off-nadir-induced reflectance overestimation. Notably, the corrected reflectance values remain physically consistent and do not exceed unity, a problem often observed in the standard products.
This work consolidates the L2 processing capabilities for PRISMA-SG, providing validated, reliable, and application-ready hyperspectral products. The approach demonstrates the importance of accounting for off-nadir geometry and second-order atmospheric and topographic effects, enabling robust use of PRISMA-SG data for environmental monitoring, snow cover studies, and other Earth observation applications.
How to cite: Santini, F., Palombo, A., Mirzaei, S., Pignatti, S., and Pascucci, S.: Development of an Advanced L2 Processor for PRISMA Second-Generation within the COOL Project: Application to Snow-Covered Terrain, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12971, https://doi.org/10.5194/egusphere-egu26-12971, 2026.
Water management represents a critical challenge for the mining industry, as large volumes of surface water must be controlled and treated during operations to ensure the stability of geotechnical structures and protect the environment. This extends beyond the operational phase, as it is essential to ensure the physical stability of tailings storage facilities and to limit the potential transport of contaminants on reclaimed mine sites. The water balance of tailings storage facilities is regulated by surface water management and often treatment infrastructures. In addition, the water balance is also closely linked to the performance of several engineered cover systems used to reclaim tailings storage facilities. Because tailings storage facilities generally occupy large areas, snow accumulation in winter and rapid melting in spring generate substantial volumes of meltwater within a short period. Thus, the spring freshet is a critical phase for water inventory control, from operation to post-reclamation. In this context, developing tailored monitoring tools is essential to ensure effective spring water management on tailings storage facilities.
This work aims to develop a high spatial resolution drone-based sensing approach for semi-real-time monitoring of the snow water balance in tailings storage facilities during snowmelt. This study is based on the results of several drone-based Structure-from-Motion photogrammetry and LiDAR surveys conducted during snowmelt on a reclaimed tailings storage facility. The site presents two major challenges for these sensing techniques: a flat, featureless area prone to oversaturated whites when covered with snow, and sections of dense low vegetation that reduce LiDAR signal penetration and hinder the generation of accurate digital elevation models. The accuracy and precision of the two remote sensing technologies to evaluate the snow depth were assessed based on manual measurements and conventional GNSS surveys. The impact of the reconstruction software/algorithms and parameters, as well as the number of ground control points (between 3 and 21) used in the reconstructions, on accuracy was also assessed. Finally, a preliminary snow-water equivalent model was developed and integrated within the data processing scheme to provide the changes of snow-water equivalent during snowmelt. Results show that LiDAR is the most accurate and reliable approach to monitor the snow depth. Photogrammetry-derived digital elevation models resulted in an error up to 66 cm. The quality and accuracy of photogrammetric surveys depend on the number of ground control points, the reconstruction algorithm used, and the absence of aerotriangulation tie points in certain areas. A snow-water equivalent model was integrated with LiDAR-derived snow depth data to characterize the temporal evolution of the tailings storage facility water balance during snowmelt. This presents an incremental improvement towards effective spring-water management on tailings storage facilities.
How to cite: Boulanger-Martel, V. and Blatter, N.: Potential of LiDAR and Structure-from-Motion Photogrammetry for High-Resolution Monitoring of Snowmelt Water Balance in Tailings Storage Facilities, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-8352, https://doi.org/10.5194/egusphere-egu26-8352, 2026.
Mineral dust is produced as a by-product when mining for metals such as iron and rare-earth elements. Around mines in the Arctic and Subarctic, this dust is transported by wind and deposited on the snow surface, contaminating the seasonal snowpack. The presence of mine dust darkens the snow surface, resulting in a lower snowpack albedo. Due to their lowered albedo, seasonal snowpacks that contain mine dust experience accelerated melting. Nordic countries, including Sweden, are showing an increasing interest in the expansion of mining activities due to increasing demand for metals to use in technology and a desire to produce raw materials within Europe. The Kirunavaara mine in the Swedish Arctic is Europe’s largest iron mine, and is an accordingly large source of mineral dust, which spreads around the adjacent town of Kiruna and the surrounding areas.
One possible approach to quantify mine dust contamination of the seasonal snowpack is using optical remote sensing. The change in spectral reflectance of the contaminated snow surface is used to infer optical properties and the concentration of mine dust in the surface snow. Spectral indices and radiative transfer modelling are applied to the spectral reflectance data to retrieve dust concentrations. We have measured dust concentrations in snow around Kiruna during spring 2025, and measured reflectance of the affected snow surfaces. Snow darkening in Kiruna occurs predominantly in the area located downwind from the mine where dust concentrations in snow are highest. Dust loadings in surface snow around Kiruna reach over 2000ppm, with associated snow broadband albedo values as low as 0.3 in the most heavily contaminated areas. There is a clear relationship between broadband albedo and mine dust concentrations in the surface snow. However, the spectral signatures of the contaminated snow surface show that iron mine dust darkens the snow relatively evenly across all wavelengths of visible light. Combined with the high dust loading, this even darkening effect means that previously established spectral indices for minerals dust in snow are not applicable in the case of iron dust contamination, and an approach tailored specifically to this type of dust is required.
How to cite: Taveirne, M., Zdanowicz, C., Langlois, A., Di Mauro, B., Traversa, G., and Hagermann, A.: Measuring mine dust contamination of snow in northern Sweden using optical remote sensing, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12080, https://doi.org/10.5194/egusphere-egu26-12080, 2026.
Snow albedo is a key control on surface energy balance in snow-covered environments, yet its estimation from multispectral satellite observations remains uncertain due to limited spectral resolution and spatial heterogeneity in snow reflectance. Thus, accurate surface albedo estimates over snow-covered landscapes are critical for the development of reliable satellite-based albedo products. However, validation data in snowy environments remains scarce, especially at high spatial resolution. This is problematic because within a single satellite scene, snow surfaces often exhibit substantial variability that challenges assumptions of spectral homogeneity of snowpack underlying many reflectance-to-albedo parameterizations.
We present a comparative framework that integrates hyperspectral Unmanned Aerial Vehicle (UAV) observations with multispectral satellite data to evaluate the limitations of derived snow albedo within the spectral configurations of Landsat 8, Landsat 9, and Sentinel 2. Our assessment extended across three distinct snowscapes: alpine, prairie, and taiga in Montana (USA), Montana, and Northern Finland; respectively. Our field-based approach employed two commercial hyperspectral sensors (Resonon Pika L and IR-L), to measure surface reflectance across the VIS–NIR–SWIR range (400-1700 nm; Landsat Bands 1-6; Sentinel 2 Bands 1-11) at high spectral (>250 bands) and spatial (0.3 m) resolution.
We isolated snow-only satellite scenes using a Convolutional Neural Network, enabling the identification of heterogeneous snow surfaces within each snowscape. Hyperspectral reflectance measurements were transformed into Landsat- and Sentinel-equivalent band reflectance using weighted sensor response functions, enabling direct band-wise comparison between hyperspectral and multispectral observations.
Our results highlight systematic discrepancies in Landsat reflectance: notably, strong overestimations in Bands 1, 2, and 5, and a consistent underestimation in Band 6 (SWIR1), with surface reflectance biases reaching up to 17%. The CNN-based classification highlighted the high spatial variability in snow reflectance, underscoring the limitations of assuming homogeneous conditions. These findings demonstrate the need to enhance validation strategies for snow-covered regions and provide a scalable protocol that integrates UAV-based acquisitions, high-resolution spectral measurements, and supervised scene analysis. This work contributes to improved characterization of snow albedo uncertainty and supports refinement of satellite-derived snow albedo products for cryospheric applications.
How to cite: Sproles, E. A., Fonseca-Gallardo, D., Hamp, S., Shaw, J. A., Wood, J., Hanula, H.-R., Pirazzini, R., and Logan, R. D.: Evaluating uncertainties in modeled snow reflectance using UAV-based hyperspectral imaging and multispectral remote sensing, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20414, https://doi.org/10.5194/egusphere-egu26-20414, 2026.
Abstract: Snow is a critical component of the cryosphere, exhibiting substantial variations in both spatial and temporal dimensions. Accurately capturing the dynamic characteristics of seasonal snow cover is essential for predicting snowmelt runoff, monitoring hydrological cycle, and conducting climate change analysis. Optical satellite remote sensing has proven to be an effective tool for monitoring global and regional snow cover. However, existing fractional snow cover (FSC) data derived from optical imagery often encounters challenges, including large-scale spatial gaps caused by cloud cover and shadows. Meanwhile, passive microwave data, although valuable, typically possess lower spatial resolution, rendering them inadequate for detecting snow cover dynamics under complex surface conditions. In this study, we employed a fractional snow cover fusion estimation method to generate high-resolution (1 km) spatiotemporally continuous FSC estimation datasets for the Tibetan Plateau region from the years 2008 to 2021, regardless of weather conditions. The accuracy of the FSC data was systematically evaluated over the study period, demonstrating excellent consistency with independent datasets, including Landsat-derived FSC (total 20 scenes; RMSE = 0.092–0.193; R = 0.83–0.946) and ground-based snow observations (Approximately 70,000 site records; Overall Accuracy = 0.95; Kappa = 0.95). Furthermore, the FSC datasets produced by this method exhibits superior performance in accurately capturing the complex daily snow cover dynamics compared to other FSC datasets(Overall Accuracy: 0.95 vs. 0.91 vs. 0.85). In conclusion, the daily FSC maps of the Tibetan Plateau generated from 2008 to 2021 using data fusion methods in this study offer high accuracy and complete spatiotemporal coverage. These FSC datasets hold substantial value for climate projections, hydrological studies, and water management at both global and regional scales.
Fig.1 Spatial Distribution of Snow Cover (1 km) for daily FSC data over the Tibetan Plateau from 2008 to 2021. The dates are shown at the bottom of the subplots. The blank areas denote missing values due to various reasons. The range of snow cover variation is from 0 to 1, where 0 indicates no snow cover and 1 indicates full snow cover.
Table.1 Summary of accuracy metrics for the 1km daily fractional snow cover data over the Tibetan Plateau using 10 Landsat scenes FSC data as the reference data.
How to cite: Sun, T. and He, T.: Mapping 1 km Fractional Snow Cover from Passive Microwave Brightness Temperature Data and MODIS Snow Cover Product over The Tibetan Plateau, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10752, https://doi.org/10.5194/egusphere-egu26-10752, 2026.
The seasonal dynamics of snow in Morocco's High Atlas Mountains play a crucial role in the region's water supply, particularly through snowmelt runoff and groundwater recharge in a semi-arid context. Snowmelt provides a significant amount of water to surface and groundwater reservoirs, especially during the summer when precipitation is very rare. However, monitoring snow cover and melt processes in this region remains difficult due to complex topography, high spatial variability, frequent cloud cover in winter, and limited in situ observations. To this end, in this study, we examine the synergistic use of optical and microwave satellite data to improve the monitoring of snow dynamics in the High Atlas.
Optical observations from the Moderate Resolution Imaging Spectroradiometer (MODIS) provide high temporal resolution estimates of snow cover extent, but they are limited by cloud contamination and variable illumination conditions in mountainous terrain. To overcome these limitations, microwave observations from the Sentinel-1 and the Global Microwave Imager (GMI)/Tropical Microwave Imager (TMI) are integrated. The combined optical-microwave framework enables us to improve the temporal continuity and robustness of dynamic snow retrievals, allowing for better characterization of snow accumulation and melt phases across elevation gradients in the High Atlas Mountains, under both clear and cloudy conditions.
The results show that the multi-sensor approach significantly improved the temporal continuity and reliability of snow dynamics monitoring compared to single-sensor approaches. The integration of microwave data allowed for consistent identification of accumulation and melt events, particularly during cloudy periods when MODIS data are not available. In particular, it allowed for better detection of rapid snow events that are often missed by optical data alone and also reduced uncertainty in estimates of snow cover duration, which is essential for assessments of water availability in the High Atlas Mountains. Overall, the approach developed here offers significant potential for improving hydrological modeling and quantifying the contribution of snowmelt to water reservoir storage in semi-arid mountainous regions where data are scarce.
How to cite: Ouatiki, H., Miorqi, C., Lhimer, A., and Chehbouni, A.: Enhancing Snow Dynamics Monitoring in the Moroccan High Atlas Using Combined Optical and Microwave Satellite Observations, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21369, https://doi.org/10.5194/egusphere-egu26-21369, 2026.
Mountain snow is a critical component of the hydrological cycle and climate system, making reliable information on snow depth (SD) essential for water resource management and climate studies. Estimating SD in mountainous terrain remains challenging due to complex topography, heterogeneous land cover, and highly variable snow and weather conditions. Synthetic Aperture Radar (SAR) is suitable in such environments, as it provides frequent observations, high spatial resolution, and sensitivity to snow properties independent of cloud cover and illumination. In particular, Sentinel-1 C-band SAR backscatter-based method enables large-scale and continuous SD monitoring, but faces limitations in vegetated, shallow, or wet snow conditions. To overcome these limitations, this study proposes an improved machine learning (ML) framework that incorporates new input variables derived from Sentinel-1 and other optical data, improving upon existing Sentinel-1–based ML approaches for SD estimation. Additionally, the framework is designed for efficient implementation using preprocessed Sentinel-1 data available in Google Earth Engine, thereby minimising the computational burden of handling SAR data and facilitating scalable application across regions and time periods. The methodology is implemented across three climatically and physiographically distinct mountainous regions: the Colorado Rocky Mountains, the European Alps, and the Indian Western Himalayas. Across all three regions, the proposed model substantially performs better than the existing methods, achieving MAE(r) values of 7.9 cm (0.96), 22.3 cm (0.91), and 68.4 cm (0.72), respectively. Since the physical scattering processes governing C-band SAR responses to snow are not yet fully characterized, explainable AI techniques are applied to interpret model predictions and quantify the influence of input variables under varying environmental conditions. The results show region-specific and seasonal dependencies linked to snow type, vegetation cover, and surface conditions, providing new physical insights into the sensitivity of Sentinel-1 C-band backscatter to snow depth.
How to cite: Rajendiran, C. P. and Ramsankaran, R.: Physical Insights into Sentinel-1 SAR-Based Snow Depth Estimation Using Machine Learning and Explainable AI Across Different Mountainous Regions, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19246, https://doi.org/10.5194/egusphere-egu26-19246, 2026.
Seasonal snow in mountain catchments is highly heterogeneous, yet snow water equivalent (SWE) is rarely available at spatial and temporal scales useful for hydrology and ecosystem applications. We present a multi-source retrospective SWE reconstruction framework that produces daily 30-meter resolution SWE over mountains by integrating (i) SAR-based wet-snow information from Sentinel-1 Normalised Radar Backscatter (NRB) product; (ii) optical snow-cover dynamics given using Sentinel-2 and Sentinel-3 data and (iii) in-situ meteorological forcing. A key advantage is that the method does not rely on spatially distributed precipitation fields, which remain a dominant uncertainty in mountain snow modelling.
The approach is built around a pixel “state” concept (accumulation, equilibrium, ablation) that constrains physically plausible SWE evolution through the season. Snow presence is represented by a daily high-resolution snow-cover-area (SCA) time series obtained by gap-filling and downscaling coarse snow-cover fraction with high-resolution optical observations, followed by a state-aware regularization that removes implausible transitions. Snow melt is computed using an enhanced temperature-index (ETI) model driven by air temperature and incoming shortwave radiation. However, ETI formulations do not explicitly resolve cold content and internal energy storage; as a result, they can trigger melt earlier than expected, as they do not account for delays imposed by the snowpack thermal inertia. To constrain the onset of true meltwater conditions, we integrate Sentinel-1 wet-snow maps derived from the new NRB time series, using multi-temporal backscatter changes to detect wet-snow conditions.
The Sentinel-1 NRB product provides radiometrically terrain-corrected backscatter (γ⁰) using the local incidence angle and mapping the data onto a reference coordinate system [1]. This improves consistency over complex topography compared to conventional Level-1 GRD processing. In addition, a novel cloud-native Zarr format enables fast, chunked access to long time series, facilitating regional-scale analyses.
We demonstrate the method in the Maipo region (Andes), where shortwave radiation dominates snowmelt. Preliminary results show that combining daily optical snow-cover dynamics with NRB-informed wet-snow timing enables SWE reconstructions that are temporally consistent across full seasons and, critically, prevents ETI-driven melt before liquid water is detected. Additionally, in the presentation, the NRB products and their assessment for the analysis of timeseries over mountains will be provided.
References
[1] G. H. X. Shiroma, M. Lavalle and S. M. Buckley, "An Area-Based Projection Algorithm for SAR Radiometric Terrain Correction and Geocoding," in IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1-23, 2022.
How to cite: Marin, C., Premier, V., Mang, N., and Castelletti, D.: Evaluating the use of NRB Sentinel-1 product for reconstructing high resolution SWE in mountains, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18431, https://doi.org/10.5194/egusphere-egu26-18431, 2026.
Snow Water Equivalent (SWE) is a critical parameter in the global and regional water cycle and climate system. However, accurately measuring SWE change using satellite remote sensing remains a challenge. While the Interferometric Synthetic Aperture Radar (InSAR) is a promising technique to retrieve SWE from space, its application has been constrained until recently by the lack of spaceborne observations combining optimal low-frequency (e.g., L-band) radar frequencies with short temporal baselines. Furthermore, interferometric coherence is a key factor that affects the accuracy of the unwrapped phase and the subsequent SWE retrieval. However, the repeat-pass InSAR coherence modelling over snow has not been sufficiently investigated.
Our study presents the first demonstration of SWE change retrieval using spaceborne repeat-pass L-band InSAR observations from the Chinese Lutan-1 mission. The study area focused on the Altay region in Xinjiang, China, during the winter of 2023–2024. Continuous interferometric pairs with 4/8-day temporal baselines are processed for phase changes and then estimate SWE variations. The retrieved SWE change shows a good agreement with in-situ SWE observations during the dry snow period (January 12 to February 9, 2024), with a Root Mean Square Error (RMSE) of 9 mm and a correlation coefficient (R) of 0.48 for the 4-day temporal baselines. However, a heavy snowfall event observed from February 9 to 17, 2024, induced severe decorrelation, leading to phase unwrapping errors that pose a challenge to SWE retrieval. To address the decorrelation mechanism of snow, the InSAR coherence model for snow is established based on the assumption of a bivariate Gaussian distribution for the ground and snow surface. The time-series modeled coherence shows a consistent trend with the observed Lutan-1 coherence, capturing effectively the decorrelation process caused by snowfall events and snow compaction processes. Furthermore, validation of the modeled coherence against Lutan-1 observations shows a strong agreement (R=0.87) over the entire study period from January 12 to March 28, 2024.
Overall, this study demonstrates the capability of spaceborne L-band InSAR with short revisit intervals to effectively retrieve SWE change under appropriate snow conditions. However, the retrieval accuracy is significantly constrained by severe decorrelation during heavy snowfall events. These results highlight both the potential and challenges of operational SWE monitoring from existing and upcoming L-band SAR missions such as Chinese Lutan-1, NASA’s NISAR, JAXA’s ALOS-4, and ESA’s ROSE-L, which are characterized by short repeat cycles, wide swath coverage, and high spatial resolution.
How to cite: Zhou, J., Lei, Y., Pan, J., Li, W., and Shi, J.: Snow Water Equivalent retrieval and InSAR Coherence modeling using L-band Lutan-1 data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20729, https://doi.org/10.5194/egusphere-egu26-20729, 2026.
Snow is the only component of the water cycle that does not have a dedicated earth observation mission. Snow impacts many sectors like the health and well-being of communities, the economy, and sustains ecosystems. Snow also contributes to many costly hazards like floods, droughts, and avalanches. The current lack of information on how much water is stored as snow (snow water equivalent, SWE), and how it varies in space and time, limits the hydrological, climate, and weather services provided by Environment and Climate Change Canada (ECCC). To address this knowledge gap, ECCC, the Canadian Space Agency (CSA) and Natural Resources Canada (NRCan) are working in partnership to advance the scientific and technical readiness for a Ku-band synthetic aperture radar (SAR) mission presently named the ‘Terrestrial Snow Mass Mission’ – TSMM. An observing concept capable of providing dual-polarization (VV/VH), moderate resolution (500 m), wide swath (~250 km), and high duty cycle (~25% SAR-on time) Ku-band radar measurements at two frequencies (13.5; 17.25 GHz) is under development. This Canadian radar mission will provide weekly coverage of the northern hemisphere with Ku-band SAR data, and coupled with modeled data in the Canadian Land Data Assimilation System (CaLDAS), will provide daily snow water equivalent data, to assist hydrological applications and decision-making. It has been proven that Ku-Band backscatter measurements are sensitive to SWE through the volume scattering of the signal by the snow microstructure. Radar measurements are also well known to be able to discriminate between wet and dry snow conditions.
In this presentation, we will review recent progress at ECCC (supported by the mission science team and the international snow community). Key areas of ongoing development include:
(1) The Ku-band radar SWE retrieval algorithm proof of concept, based on the use of physical snow modeling to provide initial estimates of snow microstructure which can effectively parameterize forward model simulations for prediction of snow volume scattering.
(2) Improvements to radiative transfer modelling codes to improve computation efficiency.
(3) Improvements to physical snow modeling in the Canadian land surface model Soil Vegetation Snow version 2 (SVS2).
(4) Development of the capability for direct assimilation of Ku-band backscatter into environmental prediction systems at ECCC.
(5) Segmentation of wet from dry snow based on the time evolution of radar backscatter.
Testbed experiments in which snow physical modeling, SWE retrievals, and data assimilation are analyzed collectively are currently under development. These experiments will be facilitated by the TSMM simulator and will incorporate outputs from SVS2 and are supported by airborne and ground-based Ku-band radar measurements from national and international academic partners.
How to cite: Montpetit, B., Derksen, C., Vionnet, V., Carrera, M., Meloche, J., Leroux, N., and Bergeron, J.: Updates on advancements of the Terrestrial Snow Mass Mission, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-8577, https://doi.org/10.5194/egusphere-egu26-8577, 2026.
