2026· IEEE Transactions on Geoscience and Remote Sensing· Vol 64, pp. 4302724-4302724· 0 citations· 73 references
Abstract
Quantifying the spatiotemporal distribution of mountain snowpack is critical for hydrology, as it represents a vital freshwater resource for over a billion people. While C-band synthetic aperture radar (SAR) from Sentinel-1 provides high-resolution estimates over mountains, its use for daily snow monitoring is hindered by infrequent revisits and signal attenuation in wet snow. To address these limitations, this work proposes the adaptive snow-phenology shape model fitting (ASP-SMF), a physically constrained fusion algorithm that generates all-weather continuous daily 500-m snow-depth maps for Northern Hemisphere mountains. Designed for data-scarce regions, ASP-SMF integrates Sentinel-1, ERA5-Land, and Interactive Multisensor Snow and Ice Mapping System (IMS) data independently of in situ calibration. The algorithm first adaptively segments the highly dynamic snow time series into three distinct phenological phases using physical rules of snowfall and snowmelt. Subsequently, it utilizes Sentinel-1 retrievals and IMS snow-free constraints to correct the temporal, amplitude, and offset biases of the reanalysis data via a robust reduced major axis (RMA) regression and uncertainty-aware sequential least-squares programming (SLSQP) optimization. Comprehensive validation against 565 sites and 55 airborne LiDAR/photogrammetry surveys (2017–2021) demonstrates that ASP-SMF: 1) reduces errors relative to Sentinel-1 under dry snow conditions, reducing root-mean-square error (RMSE) by 26% (from 0.27 to 0.20 m) and increasing R2 from 0.49 to 0.72; 2) exhibits pronounced improvement under wet snow conditions, reducing RMSE by 23% (from 0.61 to 0.47 m) and boosting R2 by 145% (from 0.22 to 0.54); and 3) shows comparable performance to the SNOw Data Assimilation System (SNODAS) product without in situ data assimilation, with a higher Kling–Gupta efficiency (KGE) (0.57 versus 0.37) against Airborne Snow Observatory (ASO) LiDAR in the Rocky Mountains. Critically, this independence from in situ inputs enables ASP-SMF to deliver spatiotemporally consistent snow monitoring in ungauged regions, providing a useful framework for high-resolution assessment of mountain snow water resources.
A deep learning–based emulator is developed to reproduce SWE simulated by Noah-MP using a Convolutional Long Short-Term Memory (ConvLSTM) network trained on meteorological forcings from the Weather Research and Forecasting model and augmented with static physiographic variables and remotely sensed snow cover and leaf area index data.
Stanley Akor, A. Flores, I. Alabi et al.· Journal of Hydrometeorology· 0 citations
Reliable monitoring of lake turbidity is often constrained by the trade-off between the spatial coverage of passive optical imagery and the vertical profiling capability of active LiDAR observations. This study proposes a multi-source retrieval framework integrating ICESat-2 ATL03 photon data with Sentinel-2 multispectral imagery for large-scale turbidity mapping in Lake Erie. An adaptive quadtree pruning strategy combined with Otsu thresholding was applied to isolate high-confidence surface water photons. Vertical distribution descriptors, including penetration depth and attenuation-related photon metrics, were quantified along 500-m segments. A Random Forest inversion model was established using these LiDAR-derived features, achieving an RMSE of 2.67 NTU. To overcome the spatial discontinuity of ICESat-2 tracks, the LiDAR-derived turbidity estimates were incorporated as virtual buoy constraints to calibrate temporally matched Sentinel-2 reflectance products. A Bayesian-optimized fusion framework was subsequently developed to generate spatially continuous turbidity fields. Validation results indicate that the synergistic model achieved an RMSE of 2.95 NTU, representing a 39% improvement over conventional optical-only retrieval methods. The proposed framework demonstrates the potential of cross-modal remote sensing synergy for large-scale inland water quality monitoring.
Rong He, Heng Chen, Guang-Hui Zhu· GEOINFORMATICS· 0 citations
Snow on sea ice modulates the Earth's radiation budget and sea ice mass balance, yet satellite retrievals of Arctic snow depth remain subject to substantial algorithmic uncertainty. We develop a multi‐model machine learning (ML) framework for retrieving spring snow depth over Arctic sea ice from Advanced Microwave Scanning Radiometer 2 (AMSR2). Three structurally distinct ML models are trained with high‐resolution airborne measurements from the Alfred Wegener Institute IceBird campaigns in April 2017 and 2019. Evaluation using a composite airborne data set from NASA's Operation IceBridge for spring 2013–2015 yields correlation coefficients of 0.66–0.70 and mean errors of 0.1–0.3 cm across the three models. A performance‐weighted fusion scheme is then introduced to combine the three ML models into a merged framework (AMSR2‐Merged) and generate a daily snow depth record for March–April from 2013 to 2023. Independent validation against ground‐based observations during the MOSAiC expedition shows a mean error of 0.16 cm for AMSR2‐Merged, with error reductions of 50%–98% relative to established empirical algorithms and 16%–88% relative to individual ML models. Over the study period, regionally averaged spring snow depths range from ∼15 to 23 cm in the Barents Sea to ∼23–25 cm in the central Arctic Ocean, with the largest interannual variability reaching ∼5 cm. Notably, the Arctic‐wide spring snow depth is ∼10 ± 5 cm lower than that in the
Warren
climatology (1954–1991). The uncertainty of AMSR2‐Merged ranges from ∼2.5 to 6 cm and peaks in the marginal ice zones, where AMSR2 data and inter‐model spread are the dominant sources.
Yi Zhou, Chentong Zhang, Alek Petty et al.· Journal of Geophysical Resea...· 0 citations
Lake ice and snow cover on the Tibetan Plateau play a critical role in regulating lake-surface radiative, thermodynamic, and hydrological processes, yet accurate monitoring remains challenging at moderate spatial resolutions. Over frozen lakes, widely used snow products (e.g., MOD10A1) frequently misclassify lake ice as snow and cannot simultaneously retrieve subpixel fractional ice cover (FIC) and fractional snow cover (FSC), leading to systematic biases in snow estimation and surface albedo. To address this limitation, we develop a retrieval framework that explicitly resolves lake ice–snow ambiguity through simultaneous subpixel estimation of lake ice and snow fractions by integrating multisource remote sensing observations. High-resolution Sentinel-2 data are used to construct subpixel reference fractions, which are fused with daily moderate-resolution imaging spectroradiometer (MODIS) observations within a physics-informed extreme gradient boosting (XGBoost) regression framework incorporating spectral, thermal, and spatiotemporal information. Model transferability is rigorously evaluated using spatially independent lakes across the Tibetan Plateau. Results demonstrate robust generalization across independent lakes, achieving a root-mean-square error (RMSE) of 0.122 for both FSC and FIC. Compared with MOD10A1, the proposed approach reduces snow-cover RMSE by approximately 40% and substantially improves consistency with Sentinel-2 reference data. Physically interpretable predictors mitigate ice–snow spectral confusion and reduce albedo overestimation in conventional snow products. By explicitly differentiating lake ice and snow at the subpixel scale, this framework provides a physically consistent approach for characterizing lake–atmosphere interactions, with strong potential for advancing hydrological modeling and generating long-term fractional ice and snow datasets over the Tibetan Plateau.
Yongzhi Gao, Xiang Zhao, Xiaozheng Du et al.· IEEE Transactions on Geoscie...· 0 citations
Accurately estimating Snow Water Equivalent (SWE) in glacierized regions is critical for glacier modeling and mass balance analysis yet remains challenging due to sparse observations and uncertainties in precipitation products. The objectives are twofold: (a) assess how well three widely used precipitation datasets—Stage IV (a radar–gauge composite product), Integrated Multi‐satellite Retrievals for GPM version 07 (IMERG V07, a satellite‐based global product), and ERA5 (a global atmospheric reanalysis)—represent end‐of‐season SWE over seven Alaskan glaciers across two winter seasons, using airborne radar SWE retrievals (ARBN) from NASA's Operation IceBridge campaign as the reference, and (b) determine whether combining precipitation with auxiliary variables via machine learning (here XGBoost) modeling can reliably predict SWE beyond airborne coverage. Results reveal systematic biases: Stage IV generally overestimates SWE, IMERG V07 underestimates it, and ERA5 aligns most closely with ARBN. Model performance depends strongly on training data representativeness, with significant discrepancies arising when training and testing SWE distributions diverge. The most effective framework integrates IMERG V07, ERA5 and key auxiliary variables–snowfall fraction, total precipitable water and 2‐m air temperature–with IMERG V07 and ERA5 found to be the dominant predictors. Nevertheless, accuracy is constrained by sparse in situ measurements and the paucity of high‐resolution data in complex terrain. These findings underscore the need for high‐quality targeted observations and integrated strategies that leverage diverse precipitation data and representative training data sets to advance SWE estimation and assessment in cryospheric environments.
Yang Song, A. Behrangi· Earth and Space Science· 0 citations
Warm-season precipitation over Sichuan, China, is jointly modulated by complex terrain, monsoon water vapor transport, and local convective activities, leading to significant spatiotemporal heterogeneity. However, radar observations over mountainous areas are frequently impaired by terrain blockage, beam shielding, and insufficient network coverage, which cause missing data and spatial discontinuity, thereby restricting the accurate monitoring of precipitation systems. To alleviate these problems, this study develops an Efficient Multi-Scale Attention (EMA) U-Net model integrated with Digital Elevation Model (DEM) information, termed EMA-U-Net-DEM, to retrieve radar composite reflectivity by utilizing multi-channel observations from the Fengyun-4A (FY-4A) Advanced Geostationary Radiation Imager (AGRI). In the experiments, FY-4A AGRI multi-spectral measurements were used as model inputs, while radar composite reflectivity products from the Severe Weather Automatic Nowcasting (SWAN) system were applied as reference labels. The modeling and validation were carried out using warm-season (June–August) datasets over Sichuan Province. The results indicate that the proposed EMA-U-Net-DEM exhibits better performance than the traditional U-Net and several typical attention-based benchmark models. Quantitatively, the model achieves a root mean square error (RMSE) of 6.728 dBZ, a mean absolute error (MAE) of 4.788 dBZ, a coefficient of determination R2 of 0.656, a peak signal-to-noise ratio (PSNR) of 25.243 dB, and a structural similarity index measure (SSIM) of 0.793. Categorical verification further reveals that the model yields the highest critical success indices (CSI) of 0.850, 0.560, and 0.364 in the reflectivity ranges of 0–25 dBZ, 25–45 dBZ, and 45–70 dBZ, respectively, demonstrating its superior ability in characterizing weak precipitation backgrounds, moderate precipitation structures, and intense convective cores. The performance enhancements are mainly attributed to the strengthened multi-scale feature extraction by the EMA module and the effective topographic constraints introduced by DEM data. This study confirms that the fusion of FY-4A multi-spectral observations and topographic information can effectively improve radar composite reflectivity retrieval over complex terrain, providing a feasible solution for precipitation monitoring, quantitative precipitation estimation, and severe weather nowcasting in mountainous regions with limited radar coverage.
