Pub Date : 2026-02-04DOI: 10.1016/j.rse.2026.115273
Yogendra Sharma, Kuo En Ching, Ruey-Juin Rau, Teresito C. Bacolcol, John E. Fungo, Alfie Pelicano, Kaj M. Johnson, Yo Fukushima
{"title":"Seismic implications of creeping and coupled segments along the Philippine fault in Leyte from GNSS and InSAR data","authors":"Yogendra Sharma, Kuo En Ching, Ruey-Juin Rau, Teresito C. Bacolcol, John E. Fungo, Alfie Pelicano, Kaj M. Johnson, Yo Fukushima","doi":"10.1016/j.rse.2026.115273","DOIUrl":"https://doi.org/10.1016/j.rse.2026.115273","url":null,"abstract":"","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"311 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2026-02-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146135000","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-03DOI: 10.1016/j.rse.2026.115270
Omid Reisi Gahrouei, Luc Guindon, Pauline Perbet, David L.P. Correia, Jean-François Cøté, Martin Béland
Reliable mapping of windthrow, a common abiotic disturbance in Canadian forests, is crucial for effective forest management and conservation, and remote sensing is becoming a useful tool for this purpose. Due to the spectral similarity of disturbances like windthrow, logging, and insect outbreaks, pixel-based approaches show limitations for accurate windthrow detection at large scales, underscoring the need for a method to distinguish windthrow events across the Canadian boreal forest. In this study, we developed and evaluated a new approach to detect windthrow-affected forest areas, primarily in the boreal forests of Eastern Canada, covering over 90 million hectares from 2019 to 2024. The method integrates deep learning (DL) models with 10 m Sentinel-2 annual composites and incorporates the Continuous Change Detection and Classification (CCDC) algorithm to improve timestamp estimation. We trained and evaluated the performance of machine and DL approaches, namely, LinkNet, CResU-Net, ResU-Net, DeepLabv3+, and Random Forest. Among them, CResU-Net, an enhanced variant of ResU-Net with a convolutional block attention module, demonstrated the best performance, achieving an overall accuracy of 98.79%, a producer’s accuracy of 75.31%, and an Intersection over Union (IoU) of 58.7% using reference data from a major windthrow event in Canada, the May 2022 Canadian derecho. Including CCDC improved our yearly timestamp to monthly timestamp for 63% of the windthrow affected area with a delay of 1–2 months relative to the event. A historical windthrow map from 2019 to 2024 was generated, highlighting the potential of combining DL and Sentinel-2 imagery for accurate and scalable windthrow detection and characterization.
风阻是加拿大森林中一种常见的非生物干扰，可靠地绘制风阻图对有效的森林管理和养护至关重要，遥感正在成为实现这一目的的有用工具。由于风投、伐木和昆虫爆发等干扰的光谱相似性，基于像素的方法在大尺度上显示出准确的风投检测的局限性，强调需要一种方法来区分加拿大北方森林的风投事件。在这项研究中，我们开发并评估了一种新的方法来检测受风吹影响的森林区域，主要是在加拿大东部的北方森林，从2019年到2024年覆盖了9000多万公顷。该方法将深度学习（DL）模型与10 m Sentinel-2年度复合材料相结合，并结合连续变化检测和分类（CCDC）算法来改进时间戳估计。我们训练并评估了机器和深度学习方法的性能，即LinkNet， CResU-Net, ResU-Net， DeepLabv3+和Random Forest。其中，基于卷积块注意力模块的ResU-Net增强版CResU-Net表现最佳，总体准确率为98.79%，制作者准确率为75.31%，联合路口（IoU）准确率为58.7%，使用的参考数据来自加拿大2022年5月的一次重大风浪事件。包括CCDC在内，我们将63%的大风影响区域的年度时间戳改进为每月时间戳，相对于事件延迟了1-2个月。生成了2019年至2024年的历史风投地图，突出了将DL和Sentinel-2图像结合起来进行准确和可扩展的风投检测和表征的潜力。
{"title":"Windthrow mapping in boreal forests using a spatio-temporal deep learning approach and Sentinel-2 imagery","authors":"Omid Reisi Gahrouei, Luc Guindon, Pauline Perbet, David L.P. Correia, Jean-François Cøté, Martin Béland","doi":"10.1016/j.rse.2026.115270","DOIUrl":"https://doi.org/10.1016/j.rse.2026.115270","url":null,"abstract":"Reliable mapping of windthrow, a common abiotic disturbance in Canadian forests, is crucial for effective forest management and conservation, and remote sensing is becoming a useful tool for this purpose. Due to the spectral similarity of disturbances like windthrow, logging, and insect outbreaks, pixel-based approaches show limitations for accurate windthrow detection at large scales, underscoring the need for a method to distinguish windthrow events across the Canadian boreal forest. In this study, we developed and evaluated a new approach to detect windthrow-affected forest areas, primarily in the boreal forests of Eastern Canada, covering over 90 million hectares from 2019 to 2024. The method integrates deep learning (DL) models with 10 m Sentinel-2 annual composites and incorporates the Continuous Change Detection and Classification (CCDC) algorithm to improve timestamp estimation. We trained and evaluated the performance of machine and DL approaches, namely, LinkNet, CResU-Net, ResU-Net, DeepLabv3+, and Random Forest. Among them, CResU-Net, an enhanced variant of ResU-Net with a convolutional block attention module, demonstrated the best performance, achieving an overall accuracy of 98.79%, a producer’s accuracy of 75.31%, and an Intersection over Union (IoU) of 58.7% using reference data from a major windthrow event in Canada, the May 2022 Canadian derecho. Including CCDC improved our yearly timestamp to monthly timestamp for 63% of the windthrow affected area with a delay of 1–2 months relative to the event. A historical windthrow map from 2019 to 2024 was generated, highlighting the potential of combining DL and Sentinel-2 imagery for accurate and scalable windthrow detection and characterization.","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"147 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2026-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146101339","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-03DOI: 10.1016/j.rse.2026.115275
Shiliu Wang, Lianhuan Wei, Meng Ao, Yingjie Chen, Mi Wang, Xiaosong Feng, Yian Wang, Shanjun Liu, Cristiano Tolomei, Christian Bignami
Time-series analysis is a crucial component of Synthetic Aperture Radar (SAR) pixel offset tracking (POT), directly impacting the accuracy of deformation monitoring. However, most existing time-series approaches struggle with precise long-term monitoring, particularly for landslide experiencing long-term large deformation. The widely used Pixel Offset–Small Baseline Subset (PO-SBAS) method supports long-term analysis but often underestimates deformation in high-gradient regions. To overcome this limitation, we introduce a Robust Sequential Pixel Offset Tracking (RS-POT) method. RS-POT initially derives short-term deformations using a single-reference POT approach, then sequentially integrates them into a complete deformation time series through a robust fusion strategy. This ensures both long-term continuity and high accuracy in capturing large deformations. Simulation results show that RS-POT provides accurate deformation estimates in 94.42% of the study area. In a case study of the Qiantaishan landslide, RS-POT outperforms PO-SBAS through reducing the root mean square error (RMSE) with reference to global navigation satellite system (GNSS) measurements by 51.51% in the azimuth direction and 43.61% in the line-of-sight (LOS) direction. Additionally, due to fewer image pairs being required, RS-POT improves computational efficiency by 122% compared to PO-SBAS. Further simulations confirm that RS-POT performs reliably under large-gradient deformation conditions, and it is applicable to various landslide types, including traction, thrust, and homogeneous landslides. These results demonstrate that RS-POT offers a more accurate and efficient solution for long-term landslide deformation monitoring.
{"title":"Robust sequential pixel offset tracking (RS-POT): A novel monitoring approach for landslides with long-term large deformations","authors":"Shiliu Wang, Lianhuan Wei, Meng Ao, Yingjie Chen, Mi Wang, Xiaosong Feng, Yian Wang, Shanjun Liu, Cristiano Tolomei, Christian Bignami","doi":"10.1016/j.rse.2026.115275","DOIUrl":"https://doi.org/10.1016/j.rse.2026.115275","url":null,"abstract":"Time-series analysis is a crucial component of Synthetic Aperture Radar (SAR) pixel offset tracking (POT), directly impacting the accuracy of deformation monitoring. However, most existing time-series approaches struggle with precise long-term monitoring, particularly for landslide experiencing long-term large deformation. The widely used Pixel Offset–Small Baseline Subset (PO-SBAS) method supports long-term analysis but often underestimates deformation in high-gradient regions. To overcome this limitation, we introduce a Robust Sequential Pixel Offset Tracking (RS-POT) method. RS-POT initially derives short-term deformations using a single-reference POT approach, then sequentially integrates them into a complete deformation time series through a robust fusion strategy. This ensures both long-term continuity and high accuracy in capturing large deformations. Simulation results show that RS-POT provides accurate deformation estimates in 94.42% of the study area. In a case study of the Qiantaishan landslide, RS-POT outperforms PO-SBAS through reducing the root mean square error (RMSE) with reference to global navigation satellite system (GNSS) measurements by 51.51% in the azimuth direction and 43.61% in the line-of-sight (LOS) direction. Additionally, due to fewer image pairs being required, RS-POT improves computational efficiency by 122% compared to PO-SBAS. Further simulations confirm that RS-POT performs reliably under large-gradient deformation conditions, and it is applicable to various landslide types, including traction, thrust, and homogeneous landslides. These results demonstrate that RS-POT offers a more accurate and efficient solution for long-term landslide deformation monitoring.","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"90 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2026-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146101340","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-31DOI: 10.1016/j.rse.2026.115271
Yuan Xiong, Gaoxiang Yang, Lei Zhang, Weiguo Yu, Yapeng Wu, Jun Lu, Chongya Jiang, Xia Yao, Yan Zhu, Weixing Cao, Tao Cheng
Rapid, accurate, and large-scale in-season prediction of winter wheat yield is essential for enhancing food security and guiding agricultural policies. Traditional data-driven methods with satellite imagery face challenges in large-scale prediction of winter wheat yield because of the limited ground sampling data available for model training. Although unmanned aerial vehicle (UAV) images have been integrated with satellite imagery for generating reference data in monitoring vegetation dynamics, the UAV and satellite synergy has not yet been investigated for cross-scale sample augmentation and information fusion in large-scale prediction of winter wheat yield. To address these issues, this study proposed a novel framework integrating ground, UAV, and satellite data with data-driven algorithms to improve regional-scale yield prediction without the need of adding field measured yield samples. The potential contributions of UAV data to yield sample augmentation were examined for compensating the lack of ground samples and improving regional-scale wheat yield prediction. Subsequently, an optimal yield prediction strategy was developed through augmented sample quality and spatial variability analysis with cross-scale information fusion. The proposed framework was evaluated with extensive field-level yield measurements over three consecutive seasons of winter wheat across Jiangsu Province, China.
The results demonstrated that synthesizing UAV and satellite data achieved superior performance across four data-driven algorithms as compared to using satellite data alone, with the ground-UAV-satellite Deep Neural Networks (DNN) model showing the most significant improvement (R2: 0.39 vs 0.85, RMSE: 1.05 vs 0.43 t/ha). Additionally, optimizing UAV-derived upscaled samples with the spatial variability indicator (Entropy for the anthesis-filling stage, Entropy_F) proved more effective for yield prediction than the conventional Winter Wheat Vegetation Fraction (WVF). The optimal strategy combination further enhanced the ground-UAV-satellite model, which resulted in the highest accuracy (R2 = 0.90, RMSE = 0.34 t/ha) across six counties. When the optimal ground-UAV-satellite model was transferred to the province, it exhibited strong transferability across seasons (2021–2022: R2 = 0.52, RMSE = 0.94 t/ha; 2022–2023: R2 = 0.62, RMSE = 0.90 t/ha; 2023–2024: R2= 0.45, RMSE = 0.96 t/ha). These findings suggest that the proposed cross-scale sample augmentation and information fusion approach is highly valuable for enhancing large-scale crop yield prediction accuracy, particularly in smallholder farming systems with limited ground samples.
{"title":"Improved prediction of winter wheat yield at regional scale with limited ground samples by unmanned aerial vehicle and satellite synergy","authors":"Yuan Xiong, Gaoxiang Yang, Lei Zhang, Weiguo Yu, Yapeng Wu, Jun Lu, Chongya Jiang, Xia Yao, Yan Zhu, Weixing Cao, Tao Cheng","doi":"10.1016/j.rse.2026.115271","DOIUrl":"10.1016/j.rse.2026.115271","url":null,"abstract":"<div><div>Rapid, accurate, and large-scale in-season prediction of winter wheat yield is essential for enhancing food security and guiding agricultural policies. Traditional data-driven methods with satellite imagery face challenges in large-scale prediction of winter wheat yield because of the limited ground sampling data available for model training. Although unmanned aerial vehicle (UAV) images have been integrated with satellite imagery for generating reference data in monitoring vegetation dynamics, the UAV and satellite synergy has not yet been investigated for cross-scale sample augmentation and information fusion in large-scale prediction of winter wheat yield. To address these issues, this study proposed a novel framework integrating ground, UAV, and satellite data with data-driven algorithms to improve regional-scale yield prediction without the need of adding field measured yield samples. The potential contributions of UAV data to yield sample augmentation were examined for compensating the lack of ground samples and improving regional-scale wheat yield prediction. Subsequently, an optimal yield prediction strategy was developed through augmented sample quality and spatial variability analysis with cross-scale information fusion. The proposed framework was evaluated with extensive field-level yield measurements over three consecutive seasons of winter wheat across Jiangsu Province, China.</div><div>The results demonstrated that synthesizing UAV and satellite data achieved superior performance across four data-driven algorithms as compared to using satellite data alone, with the ground-UAV-satellite Deep Neural Networks (DNN) model showing the most significant improvement (<em>R</em><sup><em>2</em></sup>: 0.39 vs 0.85, <em>RMSE</em>: 1.05 vs 0.43 t/ha). Additionally, optimizing UAV-derived upscaled samples with the spatial variability indicator (Entropy for the anthesis-filling stage, <em>Entropy_F</em>) proved more effective for yield prediction than the conventional Winter Wheat Vegetation Fraction (WVF). The optimal strategy combination further enhanced the ground-UAV-satellite model, which resulted in the highest accuracy (<em>R</em><sup><em>2</em></sup> = 0.90, <em>RMSE</em> = 0.34 t/ha) across six counties. When the optimal ground-UAV-satellite model was transferred to the province, it exhibited strong transferability across seasons (2021–2022: <em>R</em><sup><em>2</em></sup> = 0.52, <em>RMSE =</em> 0.94 t/ha; 2022–2023: <em>R</em><sup><em>2</em></sup> = 0.62, <em>RMSE =</em> 0.90 t/ha; 2023–2024: <em>R</em><sup><em>2</em></sup> <em>=</em> 0.45, <em>RMSE =</em> 0.96 t/ha). These findings suggest that the proposed cross-scale sample augmentation and information fusion approach is highly valuable for enhancing large-scale crop yield prediction accuracy, particularly in smallholder farming systems with limited ground samples.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"335 ","pages":"Article 115271"},"PeriodicalIF":11.4,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146075017","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-31DOI: 10.1016/j.rse.2026.115245
Shu Xu , Jinshan Cao , Peng Huang , Yining Yuan , Yi Fang , Huijun Chen , Mengchao Wu , Tengfei Long
Microvibrations degrade the geometric quality of optical Earth observation satellite imagery by introducing intra-scene spatial distortions, necessitating robust detection and compensation strategies. Like many optical Earth observation satellites, the Sustainable Development Goals Science Satellite 1 (SDGSAT-1) faces similar challenges. To address this, this paper presented a novel microvibration detection and compensation framework based on line-by-line bundle adjustment for SDGSAT-1. By minimizing directionally weighted residuals of a modified rigorous imaging model—constructed from tie points (TPs) and virtual ground control points—the framework enabled high-temporal-resolution estimation of microvibrations, parameterized through the discrete attitude microvibration model, via least-squares optimization. A new sensitivity indicator was also proposed to evaluate the adequacy of the weighting scheme for microvibration detection during optimization and to guide the dynamic adjustment of TP weights. Applied to SDGSAT-1 data, the method successfully characterized microvibrations at 1.0 Hz in the along-track direction and 0.4 Hz in the cross-track direction for the first time. Experimental results demonstrated that the proposed framework effectively suppressed microvibration-induced geometric distortions, consistently outperforming both raw imagery and classical approaches: it achieved a 33.16% reduction in RMSE compared to uncorrected data and improved cross-track precision by 27.82% over the conventional method. The impact of charge-coupled devices operating at heterogeneous imaging speeds was evaluated, with results showing no significant degradation in detection performance. These results validated the framework's effectiveness in enhancing geometric accuracy through robust microvibration modeling and compensation.
{"title":"Microvibration detection and compensation for SDGSAT-1 based on line-by-line bundle adjustment","authors":"Shu Xu , Jinshan Cao , Peng Huang , Yining Yuan , Yi Fang , Huijun Chen , Mengchao Wu , Tengfei Long","doi":"10.1016/j.rse.2026.115245","DOIUrl":"10.1016/j.rse.2026.115245","url":null,"abstract":"<div><div>Microvibrations degrade the geometric quality of optical Earth observation satellite imagery by introducing intra-scene spatial distortions, necessitating robust detection and compensation strategies. Like many optical Earth observation satellites, the Sustainable Development Goals Science Satellite 1 (SDGSAT-1) faces similar challenges. To address this, this paper presented a novel microvibration detection and compensation framework based on line-by-line bundle adjustment for SDGSAT-1. By minimizing directionally weighted residuals of a modified rigorous imaging model—constructed from tie points (TPs) and virtual ground control points—the framework enabled high-temporal-resolution estimation of microvibrations, parameterized through the discrete attitude microvibration model, via least-squares optimization. A new sensitivity indicator was also proposed to evaluate the adequacy of the weighting scheme for microvibration detection during optimization and to guide the dynamic adjustment of TP weights. Applied to SDGSAT-1 data, the method successfully characterized microvibrations at 1.0 Hz in the along-track direction and 0.4 Hz in the cross-track direction for the first time. Experimental results demonstrated that the proposed framework effectively suppressed microvibration-induced geometric distortions, consistently outperforming both raw imagery and classical approaches: it achieved a 33.16% reduction in RMSE compared to uncorrected data and improved cross-track precision by 27.82% over the conventional method. The impact of charge-coupled devices operating at heterogeneous imaging speeds was evaluated, with results showing no significant degradation in detection performance. These results validated the framework's effectiveness in enhancing geometric accuracy through robust microvibration modeling and compensation.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"335 ","pages":"Article 115245"},"PeriodicalIF":11.4,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146075018","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-31DOI: 10.1016/j.rse.2026.115274
Huihui Feng , Jianhong Zhou , Zhiyong Wu , Jianzhi Dong , Long Zhao , Luca Brocca , Hai He
Remote sensing (RS) soil moisture (SM) retrievals are frequently assimilated into land surface models (LSMs) to enhance their overall performance. However, uncertainty in LSM parameterization limits the capacity of current models to accurately capture the coupling strengths between SM and hydrological fluxes. This limitation reduces the effectiveness of SM data assimilation (DA) in improving estimates of key fluxes such as evapotranspiration (ET) and streamflow. Here, we introduce an improved SM DA framework with the optimization of LSM coupling strengths between SM and fluxes. Specifically, the SM DA framework is developed based on the Variable Infiltration Capacity (VIC) model. The model first calibrates the SM-ET and SM-runoff coupling strengths using RS data to enhance its physical consistency and representation of land surface processes. Subsequently, RS SM retrievals are assimilated into the calibrated VIC model using the Ensemble Kalman Filter to improve ET and streamflow simulations. Results indicate that the developed SM DA framework enhances DA efficiency, with SM correlation increasing from 0.45 to 0.49. It also enhances hydrological flux simulations, increasing ET correlation from 0.77 to 0.80 and improving the Nash-Sutcliffe efficiency for streamflow from 0.21 to 0.71, relative to the default VIC scheme. These improvements are especially evident in (sub-)humid regions, where the VIC model's runoff generation mechanism – based on saturation-excess processes – is well suited to representing local hydrological processes. Overall, the calibration of coupling strengths within LSMs offers a promising pathway to enhance hydrological fluxes simulation through land DA.
{"title":"Transferring soil moisture estimation skills to evapotranspiration and streamflow modeling through remote sensing data assimilation","authors":"Huihui Feng , Jianhong Zhou , Zhiyong Wu , Jianzhi Dong , Long Zhao , Luca Brocca , Hai He","doi":"10.1016/j.rse.2026.115274","DOIUrl":"10.1016/j.rse.2026.115274","url":null,"abstract":"<div><div>Remote sensing (RS) soil moisture (SM) retrievals are frequently assimilated into land surface models (LSMs) to enhance their overall performance. However, uncertainty in LSM parameterization limits the capacity of current models to accurately capture the coupling strengths between SM and hydrological fluxes. This limitation reduces the effectiveness of SM data assimilation (DA) in improving estimates of key fluxes such as evapotranspiration (ET) and streamflow. Here, we introduce an improved SM DA framework with the optimization of LSM coupling strengths between SM and fluxes. Specifically, the SM DA framework is developed based on the Variable Infiltration Capacity (VIC) model. The model first calibrates the SM-ET and SM-runoff coupling strengths using RS data to enhance its physical consistency and representation of land surface processes. Subsequently, RS SM retrievals are assimilated into the calibrated VIC model using the Ensemble Kalman Filter to improve ET and streamflow simulations. Results indicate that the developed SM DA framework enhances DA efficiency, with SM correlation increasing from 0.45 to 0.49. It also enhances hydrological flux simulations, increasing ET correlation from 0.77 to 0.80 and improving the Nash-Sutcliffe efficiency for streamflow from 0.21 to 0.71, relative to the default VIC scheme. These improvements are especially evident in (sub-)humid regions, where the VIC model's runoff generation mechanism – based on saturation-excess processes – is well suited to representing local hydrological processes. Overall, the calibration of coupling strengths within LSMs offers a promising pathway to enhance hydrological fluxes simulation through land DA.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"335 ","pages":"Article 115274"},"PeriodicalIF":11.4,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146075016","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-30DOI: 10.1016/j.rse.2026.115260
Zeyu Wang , Feng Zhang , Jieyi Wang , Long Cao
Recent and upcoming carbon satellites, such as the Orbiting Carbon Observatory-3 (OCO-3) and the Copernicus Anthropogenic Carbon Dioxide Monitoring Mission (CO2M), offer unprecedented opportunities for top-down estimation of urban emissions. Their observations, i.e., Snapshot Area Map (SAM) for OCO-3 and 250 km wide swath for CO2M, enable the detection of urban emissions in a single pass. However, accurately identifying urban plumes remains challenging due to their broad spatial extent, low signal-to-noise ratio, and substantial data gaps in quality-filtered snapshots. To address these challenges, we propose a Transformer-based deep learning (DL) model for interpolation and plume detection. Our approach uses masked pre-training on synthetic CO2M data to learn spatial dependencies and emission-related structures of values before fine-tuning for plume detection tasks. Experimental results on synthetic datasets show that the model reconstructs with mean absolute errors below the instrumental noise and achieves stable plume detection performance across noise levels. It improves gap-filling accuracy especially under regional and swath-missing conditions and significantly outperforms test- and wind-based methods in plume region segmentation accuracy. We further validated the model using 110 SAMs from 39 cities observed by OCO-3, integrating it into a lightweight inversion workflow. The resulting top-down emission estimates show improved consistency with bottom-up inventories compared to baselines (R2 = 0.61, total relative deviation = −0.10), and the city-level aggregation reproduces the bottom-up emission rankings with a Pearson’s r of 0.90. These results confirm the transferability and practical utility of our approach across global cities. This study presents a promising approach for reconstructing and detecting urban emission signals from snapshots, demonstrating clear potential to support the next-generation carbon monitoring satellites.
{"title":"Leveraging wide snapshot XCO2 pre-training to estimate urban fossil fuel CO2 emissions from space","authors":"Zeyu Wang , Feng Zhang , Jieyi Wang , Long Cao","doi":"10.1016/j.rse.2026.115260","DOIUrl":"10.1016/j.rse.2026.115260","url":null,"abstract":"<div><div>Recent and upcoming carbon satellites, such as the Orbiting Carbon Observatory-3 (OCO-3) and the Copernicus Anthropogenic Carbon Dioxide Monitoring Mission (CO2M), offer unprecedented opportunities for top-down estimation of urban <span><math><mrow><msub><mi>CO</mi><mn>2</mn></msub></mrow></math></span> emissions. Their observations, i.e., <span><math><mn>80</mn><mo>×</mo><mn>80</mn></math></span> <span><math><mrow><msup><mi>km</mi><mn>2</mn></msup></mrow></math></span> Snapshot Area Map (SAM) for OCO-3 and 250 km wide swath for CO2M, enable the detection of urban emissions in a single pass. However, accurately identifying urban plumes remains challenging due to their broad spatial extent, low signal-to-noise ratio, and substantial data gaps in quality-filtered <span><math><mrow><msub><mi>XCO</mi><mn>2</mn></msub></mrow></math></span> snapshots. To address these challenges, we propose a Transformer-based deep learning (DL) model for <span><math><mrow><msub><mi>XCO</mi><mn>2</mn></msub></mrow></math></span> interpolation and plume detection. Our approach uses masked pre-training on synthetic CO2M data to learn spatial dependencies and emission-related structures of <span><math><mrow><msub><mi>XCO</mi><mn>2</mn></msub></mrow></math></span> values before fine-tuning for plume detection tasks. Experimental results on synthetic datasets show that the model reconstructs <span><math><mrow><msub><mi>XCO</mi><mn>2</mn></msub></mrow></math></span> with mean absolute errors below the instrumental noise and achieves stable plume detection performance across noise levels. It improves <span><math><mrow><msub><mi>XCO</mi><mn>2</mn></msub></mrow></math></span> gap-filling accuracy especially under regional and swath-missing conditions and significantly outperforms test- and wind-based methods in plume region segmentation accuracy. We further validated the model using 110 SAMs from 39 cities observed by OCO-3, integrating it into a lightweight inversion workflow. The resulting top-down emission estimates show improved consistency with bottom-up inventories compared to baselines (R<sup>2</sup> = 0.61, total relative deviation = −0.10), and the city-level aggregation reproduces the bottom-up emission rankings with a Pearson’s r of 0.90. These results confirm the transferability and practical utility of our approach across global cities. This study presents a promising approach for reconstructing and detecting urban emission signals from <span><math><mrow><msub><mi>XCO</mi><mn>2</mn></msub></mrow></math></span> snapshots, demonstrating clear potential to support the next-generation carbon monitoring satellites.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"335 ","pages":"Article 115260"},"PeriodicalIF":11.4,"publicationDate":"2026-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146075415","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.rse.2026.115266
Francesco De Zan , Paolo Filippucci , Luca Brocca
This paper presents a novel algorithm for high-resolution soil moisture retrieval based on Synthetic Aperture Radar (SAR) interferometry and closure phases. The proposed method efficiently processes long SAR time series with minimal computational cost, generating a soil moisture measurement for each acquisition.
Soil moisture data were derived from Sentinel-1 SAR imagery and validated across seven different test sites. Retrieval results were compared with modeled soil moisture data from land surface models, alternative remote-sensing products, and in situ measurements.
The algorithm demonstrates strong correlations with modeled soil moisture, particularly in areas characterized by high interferometric coherence. However, performance was expectedly limited in regions with low interferometric coherence due to factors such as vegetation cover or snow cover.
Looking ahead, this study identifies some relevant directions for future research, including the integration of backscatter information alongside phase data and the adaptation of the algorithm for SAR missions operating at different frequencies (e.g., L-band) or with very dense acquisition schedules (e.g., geosynchronous platforms). These advancements would further enhance the applicability and accuracy of soil moisture retrieval using SAR-based techniques.
{"title":"Validation of high-resolution surface soil moisture time series retrieved by means of SAR interferometry","authors":"Francesco De Zan , Paolo Filippucci , Luca Brocca","doi":"10.1016/j.rse.2026.115266","DOIUrl":"10.1016/j.rse.2026.115266","url":null,"abstract":"<div><div>This paper presents a novel algorithm for high-resolution soil moisture retrieval based on Synthetic Aperture Radar (SAR) interferometry and closure phases. The proposed method efficiently processes long SAR time series with minimal computational cost, generating a soil moisture measurement for each acquisition.</div><div>Soil moisture data were derived from Sentinel-1 SAR imagery and validated across seven different test sites. Retrieval results were compared with modeled soil moisture data from land surface models, alternative remote-sensing products, and in situ measurements.</div><div>The algorithm demonstrates strong correlations with modeled soil moisture, particularly in areas characterized by high interferometric coherence. However, performance was expectedly limited in regions with low interferometric coherence due to factors such as vegetation cover or snow cover.</div><div>Looking ahead, this study identifies some relevant directions for future research, including the integration of backscatter information alongside phase data and the adaptation of the algorithm for SAR missions operating at different frequencies (e.g., L-band) or with very dense acquisition schedules (e.g., geosynchronous platforms). These advancements would further enhance the applicability and accuracy of soil moisture retrieval using SAR-based techniques.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"335 ","pages":"Article 115266"},"PeriodicalIF":11.4,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146072690","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.rse.2026.115247
Guangxi Cui , Zhongya Cai , Zhiqiang Liu
The propagation speed of internal waves is a fundamental parameter for understanding their physical mechanisms, dynamic behavior, and environmental impact. However, traditional estimation methods are typically based on numerical simulations or sparse in-situ observations, which limit their accuracy and scalability, and results in a significant scarcity of available phase speed datasets. To overcome these challenges, we propose a physics-informed and data-driven model for estimating internal wave phase speed from satellite imagery. The proposed model incorporates three key innovations: (1) the integration of theoretical equations (KdV, BO, and eKdV equations) as physical constraints to ensure consistency with real-world ocean dynamics; (2) the adoption of an adaptive ensemble learning framework that fuses data-driven and physical-informed features to improve model robustness and prediction accuracy; and (3) the introduction of a transfer learning strategy to mitigate discrepancies between theoretical predictions and observational real-world internal wave results. Experimental results demonstrate that the model achieves superior performance across varying water depths, with an average RMSE of 0.04 m/s, MRE of 2.5%, and R2 of 98.8% on the testing set. Additionally, the model was applied to the South China Sea, revealing a distinct propagation pattern: average phase speed initially increased (from 2.427 m/s to 2.53 m/s), then decreased (to 1.464 m/s), and subsequently increased again (to 1.703 m/s) as internal waves propagated westward across the Dongsha Islands and Hainan Island. The model was further validated on a global scale, achieving an average percentage error of 4.95%, confirming its scalability and generalization capability. This study presents an efficient and automated approach for accurately retrieving internal wave phase speed.
{"title":"A hybrid physics-informed and data-driven model for estimating ocean internal wave phase speeds from remote sensing imagery","authors":"Guangxi Cui , Zhongya Cai , Zhiqiang Liu","doi":"10.1016/j.rse.2026.115247","DOIUrl":"10.1016/j.rse.2026.115247","url":null,"abstract":"<div><div>The propagation speed of internal waves is a fundamental parameter for understanding their physical mechanisms, dynamic behavior, and environmental impact. However, traditional estimation methods are typically based on numerical simulations or sparse <em>in-situ</em> observations, which limit their accuracy and scalability, and results in a significant scarcity of available phase speed datasets. To overcome these challenges, we propose a physics-informed and data-driven model for estimating internal wave phase speed from satellite imagery. The proposed model incorporates three key innovations: (1) the integration of theoretical equations (KdV, BO, and eKdV equations) as physical constraints to ensure consistency with real-world ocean dynamics; (2) the adoption of an adaptive ensemble learning framework that fuses data-driven and physical-informed features to improve model robustness and prediction accuracy; and (3) the introduction of a transfer learning strategy to mitigate discrepancies between theoretical predictions and observational real-world internal wave results. Experimental results demonstrate that the model achieves superior performance across varying water depths, with an average RMSE of 0.04 m/s, MRE of 2.5%, and R<sup>2</sup> of 98.8% on the testing set. Additionally, the model was applied to the South China Sea, revealing a distinct propagation pattern: average phase speed initially increased (from 2.427 m/s to 2.53 m/s), then decreased (to 1.464 m/s), and subsequently increased again (to 1.703 m/s) as internal waves propagated westward across the Dongsha Islands and Hainan Island. The model was further validated on a global scale, achieving an average percentage error of 4.95%, confirming its scalability and generalization capability. This study presents an efficient and automated approach for accurately retrieving internal wave phase speed.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"335 ","pages":"Article 115247"},"PeriodicalIF":11.4,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146072688","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：