Pub Date : 2024-12-12DOI: 10.1016/j.rse.2024.114570
John Xun Yang , Yong-Keun Lee , Shuyan Liu , Christopher Grassotti , Quanhua Liu (Mark) , William Blackwell , Robert Vincent Leslie , Tom Greenwald , Ralf Bennartz , Scott Braun
The NASA TROPICS mission encompasses a constellation of CubeSats equipped with microwave radiometers, dedicated to investigating tropical meteorology and storm systems. In a departure from traditional microwave sounders, the TROPICS Microwave Sounder (TMS) employs new frequencies at F-band near 118 GHz and features an additional G-band channel at 205 GHz. We have expanded the capabilities of the Microwave Integrated Retrieval System (MiRS), a state-of-the-art one-dimensional variational (1DVAR) algorithm, for the retrieval of geophysical variables with the TROPICS Pathfinder early-phase data. Here we focus on assessing the retrieved precipitation in terms of rainfall and graupel. TROPICS captures well the spatial distribution and temporal evolution of Hurricane Ida and Super Typhoon Mindulle. TROPICS depicted the eyewall replacement cycle of Mindulle as it weakened and reintensified. The global precipitation distribution and dynamics are well represented by TROPICS. We compare TROPICS with other precipitation datasets, including Global Precipitation Mission (GPM) GPM Microwave Imager (GMI) and Dual-Frequency Precipitation Radar (DPR) products. For example, when compared with GMI, MiRS TROPICS instantaneous precipitation yields a correlation coefficient of 0.5 and an RMSE of 2.0 mm/h. For graupel, MiRS TROPICS retrievals show a correlation of 0.52 and an RMSE of 0.53 kg/m2. The retrieval performance is comparable to other sensors such as the Advanced Technology Microwave Sounder (ATMS), while the higher number of channels of ATMS, including its low-frequency channels serve to better constrain retrievals. TMS observes at higher spectral frequencies than the coincident ATMS channels, showing higher sensitivity to rainfall and graupel. The TMS high-frequency channels and lower orbit allow for greater resolution of precipitation features, while lower-frequency ATMS channels excel at resolving hurricane warm-core structures. The results underscore the value of the TROPICS mission for precipitation measurement and demonstrate the successful integration of TROPICS processing capability within the MiRS retrieval algorithm framework.
{"title":"Evaluating rainfall and graupel retrieval performance of the NASA TROPICS pathfinder through the NOAA MiRS system","authors":"John Xun Yang , Yong-Keun Lee , Shuyan Liu , Christopher Grassotti , Quanhua Liu (Mark) , William Blackwell , Robert Vincent Leslie , Tom Greenwald , Ralf Bennartz , Scott Braun","doi":"10.1016/j.rse.2024.114570","DOIUrl":"10.1016/j.rse.2024.114570","url":null,"abstract":"<div><div>The NASA TROPICS mission encompasses a constellation of CubeSats equipped with microwave radiometers, dedicated to investigating tropical meteorology and storm systems. In a departure from traditional microwave sounders, the TROPICS Microwave Sounder (TMS) employs new frequencies at F-band near 118 GHz and features an additional G-band channel at 205 GHz. We have expanded the capabilities of the Microwave Integrated Retrieval System (MiRS), a state-of-the-art one-dimensional variational (1DVAR) algorithm, for the retrieval of geophysical variables with the TROPICS Pathfinder early-phase data. Here we focus on assessing the retrieved precipitation in terms of rainfall and graupel. TROPICS captures well the spatial distribution and temporal evolution of Hurricane Ida and Super Typhoon Mindulle. TROPICS depicted the eyewall replacement cycle of Mindulle as it weakened and reintensified. The global precipitation distribution and dynamics are well represented by TROPICS. We compare TROPICS with other precipitation datasets, including Global Precipitation Mission (GPM) GPM Microwave Imager (GMI) and Dual-Frequency Precipitation Radar (DPR) products. For example, when compared with GMI, MiRS TROPICS instantaneous precipitation yields a correlation coefficient of 0.5 and an RMSE of 2.0 mm/h. For graupel, MiRS TROPICS retrievals show a correlation of 0.52 and an RMSE of 0.53 kg/m<sup>2</sup>. The retrieval performance is comparable to other sensors such as the Advanced Technology Microwave Sounder (ATMS), while the higher number of channels of ATMS, including its low-frequency channels serve to better constrain retrievals. TMS observes at higher spectral frequencies than the coincident ATMS channels, showing higher sensitivity to rainfall and graupel. The TMS high-frequency channels and lower orbit allow for greater resolution of precipitation features, while lower-frequency ATMS channels excel at resolving hurricane warm-core structures. The results underscore the value of the TROPICS mission for precipitation measurement and demonstrate the successful integration of TROPICS processing capability within the MiRS retrieval algorithm framework.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"318 ","pages":"Article 114570"},"PeriodicalIF":11.1,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142816038","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-12DOI: 10.1016/j.rse.2024.114563
Zanpin Xing , Lin Zhao , Lei Fan , Gabrielle De Lannoy , Xiaojing Bai , Xiangzhuo Liu , Jian Peng , Frédéric Frappart , Kun Yang , Xin Li , Zhilan Zhou , Xiaojun Li , Jiangyuan Zeng , Defu Zou , Erji Du , Chong Wang , Lingxiao Wang , Zhibin Li , Jean-Pierre Wigneron
Most existing soil moisture (SM) products from earth observations and land surface models over the Qinghai-Tibetan Plateau (QTP) have coarse resolutions or are mostly generated with high spatial resolutions based on downscaling methods. The former could hinder the applications in hydrological and ecological analyses at the regional scale and the performance of the latter could be limited by the intricate relationship between SM and downscaling factors in regions with complex topography. To address this issue, this paper aims to retrieve a 1 km SM product from 2017 to 2021 using Sentinel-1 Synthetic Aperture Radar (SAR) observations based on a semi-empirical method specific to the QTP region (SMS-1) as different from the previous downscaled SM products. The main interest in our retrievals is that the semi-empirical modeling approach allows exploring the relationships between microwave backscatters and the soil and vegetation parameters spatially based on well-defined mathematics. The SMS-1 retrievals were evaluated against the observations from five in-situ networks over the QTP and against six other existing downscaled 1 km SM products. The temporal evaluation against in-situ measurements showed that SMS-1 retrievals performed better than most 1 km SM products obtained from Machine Learning methods (median R = 0.57, ubRMSD = 0.064 m3/m3, RMSD = −0.107 m3/m3 and bias = −0.042 m3/m3) except for SMSg. Furthermore, the SMS-1 retrievals presented reasonable spatial patterns that are consistent with the spatial distribution of the grassland-type map. Our Sentinel-1 SAR-based method can therefore potentially serve as a foundation for the advance of active microwave remote sensing SM algorithm to retrieve spatially high-resolution SM.
{"title":"Retrieval of 1 km surface soil moisture from Sentinel-1 over bare soil and grassland on the Qinghai-Tibetan Plateau","authors":"Zanpin Xing , Lin Zhao , Lei Fan , Gabrielle De Lannoy , Xiaojing Bai , Xiangzhuo Liu , Jian Peng , Frédéric Frappart , Kun Yang , Xin Li , Zhilan Zhou , Xiaojun Li , Jiangyuan Zeng , Defu Zou , Erji Du , Chong Wang , Lingxiao Wang , Zhibin Li , Jean-Pierre Wigneron","doi":"10.1016/j.rse.2024.114563","DOIUrl":"10.1016/j.rse.2024.114563","url":null,"abstract":"<div><div>Most existing soil moisture (SM) products from earth observations and land surface models over the Qinghai-Tibetan Plateau (QTP) have coarse resolutions or are mostly generated with high spatial resolutions based on downscaling methods. The former could hinder the applications in hydrological and ecological analyses at the regional scale and the performance of the latter could be limited by the intricate relationship between SM and downscaling factors in regions with complex topography. To address this issue, this paper aims to retrieve a 1 km SM product from 2017 to 2021 using Sentinel-1 Synthetic Aperture Radar (SAR) observations based on a semi-empirical method specific to the QTP region (SM<sub>S-1</sub>) as different from the previous downscaled SM products. The main interest in our retrievals is that the semi-empirical modeling approach allows exploring the relationships between microwave backscatters and the soil and vegetation parameters spatially based on well-defined mathematics. The SM<sub>S-1</sub> retrievals were evaluated against the observations from five <em>in-situ</em> networks over the QTP and against six other existing downscaled 1 km SM products. The temporal evaluation against <em>in-situ</em> measurements showed that SM<sub>S-1</sub> retrievals performed better than most 1 km SM products obtained from Machine Learning methods (median <em>R</em> = 0.57, ubRMSD = 0.064 m<sup>3</sup>/m<sup>3,</sup> RMSD = −0.107 m<sup>3</sup>/m<sup>3</sup> and bias = −0.042 m<sup>3</sup>/m<sup>3</sup>) except for SM<sub>Sg</sub>. Furthermore, the SM<sub>S-1</sub> retrievals presented reasonable spatial patterns that are consistent with the spatial distribution of the grassland-type map. Our Sentinel-1 SAR-based method can therefore potentially serve as a foundation for the advance of active microwave remote sensing SM algorithm to retrieve spatially high-resolution SM.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"318 ","pages":"Article 114563"},"PeriodicalIF":11.1,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142809979","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 : 2024-12-12DOI: 10.1016/j.rse.2024.114549
Yan Wang , Xiaolin Zhu , Tao Wei , Fei Xu , Trecia Kay-Ann Williams , Helin Zhang
Accurate and timely mapping of rural settlements using medium-resolution satellite imagery, such as Landsat data, is crucial for evaluating rural infrastructure, estimating ecological service values, assessing the quality of life for rural populations, and promoting sustainable rural development. Current mapping techniques, including pixel-based and object-based classifications, primarily focus on identifying artificial surfaces, often failing to capture the complete spatial footprint of rural settlements. These settlements consist of diverse land cover elements, such as houses, roads, agricultural buildings, ponds, parks, and woodlands, which together form entities with distinct local characteristics. To address this limitation, we introduce a novel classification strategy: Entity-Based Image Analysis (EBIA). Inspired by cognitive principles of human visual perception, EBIA groups related land cover elements and differentiates settlements from their background. The key innovation of EBIA lies in its ability to incorporate semantic features within rural settlements, transforming pixel-level land cover classification results (Phase 1) into entity-level settlement mapping results (Phase 2). Our results demonstrate that EBIA effectively maps the comprehensive footprint of rural settlement entities, achieving F1 scores ranging from 0.79 to 0.88 across five globally selected experimental areas. Furthermore, EBIA can be utilized to monitor changes in rural settlements using long-term Landsat imagery. As a new classification strategy, EBIA holds potential for mapping other geographic entities.
{"title":"Entity-based image analysis: A new strategy to map rural settlements from Landsat images","authors":"Yan Wang , Xiaolin Zhu , Tao Wei , Fei Xu , Trecia Kay-Ann Williams , Helin Zhang","doi":"10.1016/j.rse.2024.114549","DOIUrl":"10.1016/j.rse.2024.114549","url":null,"abstract":"<div><div>Accurate and timely mapping of rural settlements using medium-resolution satellite imagery, such as Landsat data, is crucial for evaluating rural infrastructure, estimating ecological service values, assessing the quality of life for rural populations, and promoting sustainable rural development. Current mapping techniques, including pixel-based and object-based classifications, primarily focus on identifying artificial surfaces, often failing to capture the complete spatial footprint of rural settlements. These settlements consist of diverse land cover elements, such as houses, roads, agricultural buildings, ponds, parks, and woodlands, which together form entities with distinct local characteristics. To address this limitation, we introduce a novel classification strategy: Entity-Based Image Analysis (EBIA). Inspired by cognitive principles of human visual perception, EBIA groups related land cover elements and differentiates settlements from their background. The key innovation of EBIA lies in its ability to incorporate semantic features within rural settlements, transforming pixel-level land cover classification results (Phase 1) into entity-level settlement mapping results (Phase 2). Our results demonstrate that EBIA effectively maps the comprehensive footprint of rural settlement entities, achieving F1 scores ranging from 0.79 to 0.88 across five globally selected experimental areas. Furthermore, EBIA can be utilized to monitor changes in rural settlements using long-term Landsat imagery. As a new classification strategy, EBIA holds potential for mapping other geographic entities.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"318 ","pages":"Article 114549"},"PeriodicalIF":11.1,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142809720","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-12DOI: 10.1016/j.rse.2024.114547
Francisco Mena , Deepak Pathak , Hiba Najjar , Cristhian Sanchez , Patrick Helber , Benjamin Bischke , Peter Habelitz , Miro Miranda , Jayanth Siddamsetty , Marlon Nuske , Marcela Charfuelan , Diego Arenas , Michaela Vollmer , Andreas Dengel
Accurate crop yield prediction is of utmost importance for informed decision-making in agriculture, aiding farmers, industry stakeholders, and policymakers in optimizing agricultural practices. However, this task is complex and depends on multiple factors, such as environmental conditions, soil properties, and management practices. Leveraging Remote Sensing (RS) technologies, multi-modal data from diverse global data sources can be collected to enhance predictive model accuracy. However, combining heterogeneous RS data poses a fusion challenge, like identifying the specific contribution of each modality in the predictive task. In this paper, we present a novel multi-modal learning approach to predict crop yield for different crops (soybean, wheat, rapeseed) and regions (Argentina, Uruguay, and Germany). Our multi-modal input data includes multi-spectral optical images from Sentinel-2 satellites and weather data as dynamic features during the crop growing season, complemented by static features like soil properties and topographic information. To effectively fuse the multi-modal data, we introduce a Multi-modal Gated Fusion (MMGF) model, comprising dedicated modality-encoders and a Gated Unit (GU) module. The modality-encoders handle the heterogeneity of data sources with varying temporal resolutions by learning a modality-specific representation. These representations are adaptively fused via a weighted sum. The fusion weights are computed for each sample by the GU using a concatenation of the multi-modal representations. The MMGF model is trained at sub-field level with 10 m resolution pixels. Our evaluations show that the MMGF outperforms conventional models on the same task, achieving the best results by incorporating all the data sources, unlike the usual fusion results in the literature. For Argentina, the MMGF model achieves an value of 0.68 at sub-field yield prediction, while at the field level evaluation (comparing field averages), it reaches around 0.80 across different countries. The GU module learned different weights based on the country and crop-type, aligning with the variable significance of each data source to the prediction task. This novel method has proven its effectiveness in enhancing the accuracy of the challenging sub-field crop yield prediction. Our investigation indicates that the gated fusion approach promises a significant advancement in the field of agriculture and precision farming.
{"title":"Adaptive fusion of multi-modal remote sensing data for optimal sub-field crop yield prediction","authors":"Francisco Mena , Deepak Pathak , Hiba Najjar , Cristhian Sanchez , Patrick Helber , Benjamin Bischke , Peter Habelitz , Miro Miranda , Jayanth Siddamsetty , Marlon Nuske , Marcela Charfuelan , Diego Arenas , Michaela Vollmer , Andreas Dengel","doi":"10.1016/j.rse.2024.114547","DOIUrl":"10.1016/j.rse.2024.114547","url":null,"abstract":"<div><div>Accurate crop yield prediction is of utmost importance for informed decision-making in agriculture, aiding farmers, industry stakeholders, and policymakers in optimizing agricultural practices. However, this task is complex and depends on multiple factors, such as environmental conditions, soil properties, and management practices. Leveraging Remote Sensing (RS) technologies, multi-modal data from diverse global data sources can be collected to enhance predictive model accuracy. However, combining heterogeneous RS data poses a fusion challenge, like identifying the specific contribution of each modality in the predictive task. In this paper, we present a novel multi-modal learning approach to predict crop yield for different crops (soybean, wheat, rapeseed) and regions (Argentina, Uruguay, and Germany). Our multi-modal input data includes multi-spectral optical images from Sentinel-2 satellites and weather data as dynamic features during the crop growing season, complemented by static features like soil properties and topographic information. To effectively fuse the multi-modal data, we introduce a Multi-modal Gated Fusion (MMGF) model, comprising dedicated modality-encoders and a Gated Unit (GU) module. The modality-encoders handle the heterogeneity of data sources with varying temporal resolutions by learning a modality-specific representation. These representations are adaptively fused via a weighted sum. The <em>fusion</em> weights are computed for each sample by the GU using a concatenation of the multi-modal representations. The MMGF model is trained at sub-field level with 10 m resolution pixels. Our evaluations show that the MMGF outperforms conventional models on the same task, achieving the best results by incorporating all the data sources, unlike the usual fusion results in the literature. For Argentina, the MMGF model achieves an <span><math><msup><mrow><mi>R</mi></mrow><mrow><mn>2</mn></mrow></msup></math></span> value of 0.68 at sub-field yield prediction, while at the field level evaluation (comparing field averages), it reaches around 0.80 across different countries. The GU module learned different weights based on the country and crop-type, aligning with the variable significance of each data source to the prediction task. This novel method has proven its effectiveness in enhancing the accuracy of the challenging sub-field crop yield prediction. Our investigation indicates that the gated fusion approach promises a significant advancement in the field of agriculture and precision farming.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"318 ","pages":"Article 114547"},"PeriodicalIF":11.1,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142816039","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-12DOI: 10.1016/j.rse.2024.114518
Dan J. Dixon, Yunzhe Zhu, Yufang Jin
<div><div>Canopy height mapping is critical for assessing forest structure, forest resilience, carbon stocks, habitat, and biodiversity, all of which are threatened by changing climate and weather extremes. While current tools utilizing lidar (e.g., GEDI) and multispectral imagery (e.g., Landsat, Sentinel-2, airborne imagery) produce canopy height products, significant challenges remain, particularly in capturing fine-scale spatial details across large areas with high frequency. PlanetScope CubeSat imagery, with its 3 m spatial resolution and near-daily frequency, offers a unique opportunity to estimate woody plant structure by capturing fine-scale texture and temporal patterns that shift throughout the year. In this study, we adapted a 3D Spatio-Temporal Convolutional Neural Network (ST-CNN) to estimate canopy height at 3 m resolution, utilizing sequential PlanetScope time series over five months, summer Sentinel-1 radar imagery, and solar illumination layers as inputs. We generated a large and diverse reference database covering 2,296 sample scenes (each scene = 768 × 768 m, totaling <span><math><mo>∼</mo></math></span>135,000 ha) using a semi-automatic labeling process that leverages 23 aerial lidar surveys conducted in California between 2016 and 2021. Trained on a random selection of 2,046 scenes, the accuracy assessment on the remaining 250 scenes demonstrates strong performance across various ecoregions, capturing 80.8% of the observed variance in live canopy height with a mean absolute error (MAE) of 3.6 m and a bias of -0.53 m compared with aerial lidar. Analysis of all 681 GEDI footprints over the same testing scenes estimates the MAE of 6.5 m, bias of -1.82 m, and R<span><math><msup><mrow></mrow><mrow><mn>2</mn></mrow></msup></math></span> of 0.58 for the GEDI L2A Vector Canopy Top Height RH98 product. The ST-CNN model accurately identifies heterogeneous canopy structures, and shows sensitivity to canopies reaching 50 to 60 m in height. We found a major contribution from the PlanetScope time series, compared to a single PlanetScope image, and marginal benefits of including Sentinel-1 and terrain-based solar irradiance layers to improve performance on dense canopies or diverse topography. Example applications demonstrate the ability to generalize to different years, maintaining consistent predictions between years and capturing changes in canopy height over a seven year period (2017–2023) within 400 plots representing regrowth, minimal change, selective logging, and clear cut areas. We also demonstrate improved canopy height estimation compared to existing products from Landsat (MAE = 8.41 m) and Sentinel-2 (MAE = 7.19 m). A visualization tool displays our data alongside existing products for the Sierra Nevada in 2022. The Planet ST-CNN model, using a 15-day PlanetScope satellite time series, offers a scalable approach for annual canopy height estimation in California, achieving a high level of detail, often down to the resolution of in
{"title":"Canopy height estimation from PlanetScope time series with spatio-temporal deep learning","authors":"Dan J. Dixon, Yunzhe Zhu, Yufang Jin","doi":"10.1016/j.rse.2024.114518","DOIUrl":"10.1016/j.rse.2024.114518","url":null,"abstract":"<div><div>Canopy height mapping is critical for assessing forest structure, forest resilience, carbon stocks, habitat, and biodiversity, all of which are threatened by changing climate and weather extremes. While current tools utilizing lidar (e.g., GEDI) and multispectral imagery (e.g., Landsat, Sentinel-2, airborne imagery) produce canopy height products, significant challenges remain, particularly in capturing fine-scale spatial details across large areas with high frequency. PlanetScope CubeSat imagery, with its 3 m spatial resolution and near-daily frequency, offers a unique opportunity to estimate woody plant structure by capturing fine-scale texture and temporal patterns that shift throughout the year. In this study, we adapted a 3D Spatio-Temporal Convolutional Neural Network (ST-CNN) to estimate canopy height at 3 m resolution, utilizing sequential PlanetScope time series over five months, summer Sentinel-1 radar imagery, and solar illumination layers as inputs. We generated a large and diverse reference database covering 2,296 sample scenes (each scene = 768 × 768 m, totaling <span><math><mo>∼</mo></math></span>135,000 ha) using a semi-automatic labeling process that leverages 23 aerial lidar surveys conducted in California between 2016 and 2021. Trained on a random selection of 2,046 scenes, the accuracy assessment on the remaining 250 scenes demonstrates strong performance across various ecoregions, capturing 80.8% of the observed variance in live canopy height with a mean absolute error (MAE) of 3.6 m and a bias of -0.53 m compared with aerial lidar. Analysis of all 681 GEDI footprints over the same testing scenes estimates the MAE of 6.5 m, bias of -1.82 m, and R<span><math><msup><mrow></mrow><mrow><mn>2</mn></mrow></msup></math></span> of 0.58 for the GEDI L2A Vector Canopy Top Height RH98 product. The ST-CNN model accurately identifies heterogeneous canopy structures, and shows sensitivity to canopies reaching 50 to 60 m in height. We found a major contribution from the PlanetScope time series, compared to a single PlanetScope image, and marginal benefits of including Sentinel-1 and terrain-based solar irradiance layers to improve performance on dense canopies or diverse topography. Example applications demonstrate the ability to generalize to different years, maintaining consistent predictions between years and capturing changes in canopy height over a seven year period (2017–2023) within 400 plots representing regrowth, minimal change, selective logging, and clear cut areas. We also demonstrate improved canopy height estimation compared to existing products from Landsat (MAE = 8.41 m) and Sentinel-2 (MAE = 7.19 m). A visualization tool displays our data alongside existing products for the Sierra Nevada in 2022. The Planet ST-CNN model, using a 15-day PlanetScope satellite time series, offers a scalable approach for annual canopy height estimation in California, achieving a high level of detail, often down to the resolution of in","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"318 ","pages":"Article 114518"},"PeriodicalIF":11.1,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142809719","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-12DOI: 10.1016/j.rse.2024.114543
Fanghui Deng , Mark Zumberge
Space geodetic techniques have achieved centimeter to even millimeter precision in measuring earth surface deformation. However, a large data gap remains in the offshore area. Offshore man-made structures (e.g., oil/gas platforms) anchored to the ocean bottom provide an opportunity to study seafloor motion in some areas. Although satellite InSAR (Interferometric Synthetic Aperture Radar) has been widely used to study earth surface deformation, its application to offshore regions is extremely limited. Continuous GNSS (Global Navigation Satellite System) observations at several tens of offshore platforms in the Adriatic Sea have recently been released. Measuring the same platforms with InSAR provides a great opportunity to assess the feasibility of applying this technique to study seafloor motion on a regional scale using offshore structures. We processed a six-year-long time series of SAR images from the Sentinel-1A satellite using the Permanent Scatterer InSAR (PS-InSAR) method. We assessed the feasibility of phase unwrapping using synthetic data with different velocity fields and noise levels. Correct phase unwrapping could be achieved in the Adriatic Sea and two other large offshore oil/gas fields: the Gulf of Mexico and the North Sea. Different calibration strategies were applied, and we suggest that the InSAR results could be calibrated with limited and even no GNSS stations. Our InSAR results show good agreement with the GNSS measurements and the InSAR observations from the European Ground Motion Service. In addition, our InSAR results provide deformation measurements at about twenty offshore structures where GNSS stations are not present. Most of the offshore structures have a subsidence rate of no more than 5 mm/year, while a few of them reach about 10 mm/year. Our work demonstrates that it is feasible to apply the InSAR technique to measure displacement of discrete offshore man-made structures (fixed to the ocean bottom) on a regional scale but still on a case-by-case basis. Pre-acquired information including geological settings, existing geodetic observations, and human activity records (e.g., hydrocarbon production) are useful information to assess the feasibility and to validate the InSAR results.
{"title":"Seafloor motion from offshore man-made structures using satellite radar images – A case study in the Adriatic Sea","authors":"Fanghui Deng , Mark Zumberge","doi":"10.1016/j.rse.2024.114543","DOIUrl":"10.1016/j.rse.2024.114543","url":null,"abstract":"<div><div>Space geodetic techniques have achieved centimeter to even millimeter precision in measuring earth surface deformation. However, a large data gap remains in the offshore area. Offshore man-made structures (e.g., oil/gas platforms) anchored to the ocean bottom provide an opportunity to study seafloor motion in some areas. Although satellite InSAR (Interferometric Synthetic Aperture Radar) has been widely used to study earth surface deformation, its application to offshore regions is extremely limited. Continuous GNSS (Global Navigation Satellite System) observations at several tens of offshore platforms in the Adriatic Sea have recently been released. Measuring the same platforms with InSAR provides a great opportunity to assess the feasibility of applying this technique to study seafloor motion on a regional scale using offshore structures. We processed a six-year-long time series of SAR images from the Sentinel-1A satellite using the Permanent Scatterer InSAR (PS-InSAR) method. We assessed the feasibility of phase unwrapping using synthetic data with different velocity fields and noise levels. Correct phase unwrapping could be achieved in the Adriatic Sea and two other large offshore oil/gas fields: the Gulf of Mexico and the North Sea. Different calibration strategies were applied, and we suggest that the InSAR results could be calibrated with limited and even no GNSS stations. Our InSAR results show good agreement with the GNSS measurements and the InSAR observations from the European Ground Motion Service. In addition, our InSAR results provide deformation measurements at about twenty offshore structures where GNSS stations are not present. Most of the offshore structures have a subsidence rate of no more than 5 mm/year, while a few of them reach about 10 mm/year. Our work demonstrates that it is feasible to apply the InSAR technique to measure displacement of discrete offshore man-made structures (fixed to the ocean bottom) on a regional scale but still on a case-by-case basis. Pre-acquired information including geological settings, existing geodetic observations, and human activity records (e.g., hydrocarbon production) are useful information to assess the feasibility and to validate the InSAR results.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"318 ","pages":"Article 114543"},"PeriodicalIF":11.1,"publicationDate":"2024-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142809718","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 : 2024-12-11DOI: 10.1016/j.rse.2024.114557
Neha Hunka , Paul May , Chad Babcock , José Armando Alanís de la Rosa , Maria de los Ángeles Soriano-Luna , Rafael Mayorga Saucedo , John Armston , Maurizio Santoro , Daniela Requena Suarez , Martin Herold , Natalia Málaga , Sean P. Healey , Robert E. Kennedy , Andrew T. Hudak , Laura Duncanson
Earth Observation (EO) data can provide added value to nations’ assessments of vegetation aboveground biomass density (AGBD) with minimal additional costs. Yet, neither open access to global-scale EO datasets of vegetation heights or biomass, nor the availability of computational power, has proven sufficient for their wide uptake in climate policy-related assessments. Using Mexico as an example, one of the primary obstacles to enhancing their National Forest Inventory (NFI) with such global EO datasets is the lack of statistically defensible methodologies that do so, while addressing the nation’s existing reporting needs and gaps. In collaboration with the Comisión Nacional Forestal (CONAFOR), this study develops a geostatistical model that integrates vegetation height and AGBD estimates from NASA’s Global Ecosystem Dynamics Investigation (GEDI) and ESA’s Climate Change Initiative (CCI) with Mexico’s NFI to attain sub-national and geographically-explicit biomass predictions. The posited model includes spatially varying parameters, allowing flexibility to capture non-stationary relations between the EO-based covariates and NFI-estimated AGBD. Inference is conducted with Bayesian methods, allowing the computation of summary statistics, such as the standard deviations for single-location and area-wide predictions of AGBD. This enables the transparent disclosure and traceability of sources of uncertainty throughout the prediction approach. Results indicate strong model performance; the EO-based covariates explain 79% of the variance in NFI-estimated AGBD in a randomly withheld sample of 10% of observations and a heuristic root mean squared error (RMSE) of 21.55 Mg/ha. Approximately 96% of the observations falling within the 95% credible intervals of our predictions, with some systematic under-prediction observed at AGBD ranges of Mg/ha. To ease the operational uptake of the model for policy purposes, source code based in the ‘R’ language with the optional use of urban and (non)forest masks for AGBD predictions is released. It includes demonstrations for predicting AGBD in Mexico’s Natural Protected Areas, terrestrial ecological strata, and community forest management or payment for environmental services projects, which are commonly used delineations in its climate policy reports. For other nations considering the presented approach for policy purposes, the study discusses challenges concerning the use of EO-based covariates and the limitations of the model. It concludes with a broader call toward ensuring consistency in EO data streams, and prioritizing the co-development of EO-NFI integration approaches with nations in the future, thereby directly addressing their long-term climate policy needs.
{"title":"A geostatistical approach to enhancing national forest biomass assessments with Earth Observation to aid climate policy needs","authors":"Neha Hunka , Paul May , Chad Babcock , José Armando Alanís de la Rosa , Maria de los Ángeles Soriano-Luna , Rafael Mayorga Saucedo , John Armston , Maurizio Santoro , Daniela Requena Suarez , Martin Herold , Natalia Málaga , Sean P. Healey , Robert E. Kennedy , Andrew T. Hudak , Laura Duncanson","doi":"10.1016/j.rse.2024.114557","DOIUrl":"10.1016/j.rse.2024.114557","url":null,"abstract":"<div><div>Earth Observation (EO) data can provide added value to nations’ assessments of vegetation aboveground biomass density (AGBD) with minimal additional costs. Yet, neither open access to global-scale EO datasets of vegetation heights or biomass, nor the availability of computational power, has proven sufficient for their wide uptake in climate policy-related assessments. Using Mexico as an example, one of the primary obstacles to enhancing their National Forest Inventory (NFI) with such global EO datasets is the lack of statistically defensible methodologies that do so, while addressing the nation’s existing reporting needs and gaps. In collaboration with the Comisión Nacional Forestal (CONAFOR), this study develops a geostatistical model that integrates vegetation height and AGBD estimates from NASA’s Global Ecosystem Dynamics Investigation (GEDI) and ESA’s Climate Change Initiative (CCI) with Mexico’s NFI to attain sub-national and geographically-explicit biomass predictions. The posited model includes spatially varying parameters, allowing flexibility to capture non-stationary relations between the EO-based covariates and NFI-estimated AGBD. Inference is conducted with Bayesian methods, allowing the computation of summary statistics, such as the standard deviations for single-location and area-wide predictions of AGBD. This enables the transparent disclosure and traceability of sources of uncertainty throughout the prediction approach. Results indicate strong model performance; the EO-based covariates explain 79% of the variance in NFI-estimated AGBD in a randomly withheld sample of 10% of observations and a heuristic root mean squared error (RMSE) of 21.55 Mg/ha. Approximately 96% of the observations falling within the 95% credible intervals of our predictions, with some systematic under-prediction observed at AGBD ranges of <span><math><mrow><mo>></mo><mn>100</mn></mrow></math></span> Mg/ha. To ease the operational uptake of the model for policy purposes, source code based in the ‘R’ language with the optional use of urban and (non)forest masks for AGBD predictions is released. It includes demonstrations for predicting AGBD in Mexico’s Natural Protected Areas, terrestrial ecological strata, and community forest management or payment for environmental services projects, which are commonly used delineations in its climate policy reports. For other nations considering the presented approach for policy purposes, the study discusses challenges concerning the use of EO-based covariates and the limitations of the model. It concludes with a broader call toward ensuring consistency in EO data streams, and prioritizing the co-development of EO-NFI integration approaches with nations in the future, thereby directly addressing their long-term climate policy needs.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"318 ","pages":"Article 114557"},"PeriodicalIF":11.1,"publicationDate":"2024-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142804922","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-11DOI: 10.1016/j.rse.2024.114568
Young Jun Kim , Hyun-cheol Kim , Daehyeon Han , Julienne Stroeve , Jungho Im
Over the last five decades, Arctic sea ice has been shrinking in area and thickness. As a result, increased marine traffic has created a need for improved sea ice forecasting on seasonal to annual time-scales. In this study, we introduce a novel UNET-based deep learning model to forecast sea ice concentration up to 12 months. Based on yearly hindcast validation, the UNET 3-, 6-, 9-, and 12-month predictions provided more accurate and stable predictions than did the four baseline models: the Copernicus Climate Change Service (C3S), the damped anomaly persistence (DP) forecast, and two deep learning approach, the Convolutional Neural Network (CNN) models and Convolutional Long Short-Term Memory (ConvLSTM). During years with large departures from the long-term trend, the proposed UNET model exhibited promising SIC prediction results with root-mean-square errors (RMSEs), which were reduced from 17.35 to 7.07 % compared to the four baseline models. Our findings also confirmed the relative importance of each predictor variable (temperature, incoming solar radiation, wind speed and direction) in long-term prediction. Past SIC conditions, together with surface temperature emerged as the most important factors for SIC prediction, especially in the marginal ice zone. Incoming solar radiation and wind speed and direction showed greater sensitivity in predicting SICs in areas with thin ice. This model offers the potential to shape Arctic development and management plans and strategies, ensuring extended forecasting periods and enhanced prediction accuracy.
{"title":"Long-term prediction of Arctic sea ice concentrations using deep learning: Effects of surface temperature, radiation, and wind conditions","authors":"Young Jun Kim , Hyun-cheol Kim , Daehyeon Han , Julienne Stroeve , Jungho Im","doi":"10.1016/j.rse.2024.114568","DOIUrl":"10.1016/j.rse.2024.114568","url":null,"abstract":"<div><div>Over the last five decades, Arctic sea ice has been shrinking in area and thickness. As a result, increased marine traffic has created a need for improved sea ice forecasting on seasonal to annual time-scales. In this study, we introduce a novel UNET-based deep learning model to forecast sea ice concentration up to 12 months. Based on yearly hindcast validation, the UNET 3-, 6-, 9-, and 12-month predictions provided more accurate and stable predictions than did the four baseline models: the Copernicus Climate Change Service (C3S), the damped anomaly persistence (DP) forecast, and two deep learning approach, the Convolutional Neural Network (CNN) models and Convolutional Long Short-Term Memory (ConvLSTM). During years with large departures from the long-term trend, the proposed UNET model exhibited promising SIC prediction results with root-mean-square errors (RMSEs), which were reduced from 17.35 to 7.07 % compared to the four baseline models. Our findings also confirmed the relative importance of each predictor variable (temperature, incoming solar radiation, wind speed and direction) in long-term prediction. Past SIC conditions, together with surface temperature emerged as the most important factors for SIC prediction, especially in the marginal ice zone. Incoming solar radiation and wind speed and direction showed greater sensitivity in predicting SICs in areas with thin ice. This model offers the potential to shape Arctic development and management plans and strategies, ensuring extended forecasting periods and enhanced prediction accuracy.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"318 ","pages":"Article 114568"},"PeriodicalIF":11.1,"publicationDate":"2024-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142804960","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 : 2024-12-09DOI: 10.1016/j.rse.2024.114534
Wenlu Qi , John Armston , Changhyun Choi , Atticus Stovall , Svetlana Saarela , Matteo Pardini , Lola Fatoyinbo , Konstantinos Papathanassiou , Adrian Pascual , Ralph Dubayah
NASA's Global Ecosystem Dynamic Investigation (GEDI) mission provides billions of lidar samples of canopy structure over the Earth's temperate and pantropical forests. Using the GEDI sample data alone, gridded height and biomass products have been created at a spatial resolution of 1 km or coarser. However, this resolution may be too coarse for some applications. In this study, we present a new method of mapping high spatial resolution forest height across large areas using fusion of data acquired by GEDI and TanDEM-X (TDX) Interferometric Synthetic Aperture Radar (InSAR). Our method utilizes GEDI waveforms to provide vertical profiles of scatterers needed to invert a physically-based InSAR model to solve for canopy height. We then use 2-year GEDI canopy height and adaptive wavenumber (kZ)-based calibration models to reduce errors in the inverted canopy height caused by the limited penetration capability of the X-band signal in dense tropical forests and the impact of terrain. We apply this novel method over large areas including Gabon, Mexico, French Guiana and most of the Amazon basin, and generate continuous forest height products at 25 m and 100 m. After validating against airborne lidar data, we find that our canopy height products have a bias of 0.31 m and 0.46 m, and a root mean square error (RMSE) of 8.48 m (30.02 %) and 6.91 m (24.08 %) at 25 m and 100 m respectively, for all sites combined. Compared to existing data products that integrate GEDI with passive optical data using machine learning approaches, our method reduces bias, has a lower RMSE, and does not saturate for tall canopy heights up to 56 m. A key feature of this study is that our canopy height product is complemented with an uncertainty of prediction map which provides information on the predictor's uncertainty around the actual value —an advancement over the standard error maps used in earlier studies, which provide uncertainty around the expectation of the predicted value. This integration approach enables the first-ever accurate and high-resolution mapping of forest canopy heights at unprecedented large areas from GEDI and TDX InSAR data fusion, serving as an essential foundation for pantropical aboveground biomass mapping.
{"title":"Mapping large-scale pantropical forest canopy height by integrating GEDI lidar and TanDEM-X InSAR data","authors":"Wenlu Qi , John Armston , Changhyun Choi , Atticus Stovall , Svetlana Saarela , Matteo Pardini , Lola Fatoyinbo , Konstantinos Papathanassiou , Adrian Pascual , Ralph Dubayah","doi":"10.1016/j.rse.2024.114534","DOIUrl":"10.1016/j.rse.2024.114534","url":null,"abstract":"<div><div>NASA's Global Ecosystem Dynamic Investigation (GEDI) mission provides billions of lidar samples of canopy structure over the Earth's temperate and pantropical forests. Using the GEDI sample data alone, gridded height and biomass products have been created at a spatial resolution of 1 km or coarser. However, this resolution may be too coarse for some applications. In this study, we present a new method of mapping high spatial resolution forest height across large areas using fusion of data acquired by GEDI and TanDEM-X (TDX) Interferometric Synthetic Aperture Radar (InSAR). Our method utilizes GEDI waveforms to provide vertical profiles of scatterers needed to invert a physically-based InSAR model to solve for canopy height. We then use 2-year GEDI canopy height and adaptive wavenumber (<em>k</em><sub><em>Z</em></sub>)-based calibration models to reduce errors in the inverted canopy height caused by the limited penetration capability of the X-band signal in dense tropical forests and the impact of terrain. We apply this novel method over large areas including Gabon, Mexico, French Guiana and most of the Amazon basin, and generate continuous forest height products at 25 m and 100 m. After validating against airborne lidar data, we find that our canopy height products have a bias of 0.31 m and 0.46 m, and a root mean square error (RMSE) of 8.48 m (30.02 %) and 6.91 m (24.08 %) at 25 m and 100 m respectively, for all sites combined. Compared to existing data products that integrate GEDI with passive optical data using machine learning approaches, our method reduces bias, has a lower RMSE, and does not saturate for tall canopy heights up to 56 m. A key feature of this study is that our canopy height product is complemented with an uncertainty of prediction map which provides information on the predictor's uncertainty around the actual value —an advancement over the standard error maps used in earlier studies, which provide uncertainty around the expectation of the predicted value. This integration approach enables the first-ever accurate and high-resolution mapping of forest canopy heights at unprecedented large areas from GEDI and TDX InSAR data fusion, serving as an essential foundation for pantropical aboveground biomass mapping.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"318 ","pages":"Article 114534"},"PeriodicalIF":11.1,"publicationDate":"2024-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142793509","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-12-09DOI: 10.1016/j.rse.2024.114546
Yidan Wang , Wei Wu , Zhicheng Zhang , Ziming Li , Fan Zhang , Qinchuan Xin
High-quality and continuous satellite data are essential for land surface studies such as monitoring of land surface phenology, but factors such as cloud contamination and sensor malfunction degrade the quality of remote sensing images and limit their utilization. Filling gaps and recovering missing information in time series of remote sensing images are vital for a wide range of downstream applications, such as land surface phenology extraction. Most existing gap-filling and cloud removal methods focus on individual or multi-temporal image reconstruction, but struggle with continuous and overlapping missing areas in time series data. In this study, we propose a Temporal Attention-Based Multi-Scale Generative Adversarial Network (TAMGAN) to reconstruct time series of Moderate Resolution Imaging Spectroradiometer (MODIS) data. TAMGAN leverages a Generative Adversarial Network (GAN) with a 3-dimensional Convolution Neural Networks (3DCNN) in its generator to reconstruct the missing areas in the annual time series of remote sensing images simultaneously. The temporal attention blocks are designed to capture the changing trends of surface reflectance over time. And multi-scale feature extraction and progressive concatenation are introduced to enhance spectral consistency and provide detailed texture information. Experiments are carried out on MOD09A1 products to evaluate the performance of the proposed network. The results show that TAMGAN outperformed the comparison methods across various evaluation metrics, particularly in handling large and continuous missing areas in the time series. Furthermore, we showcase an example of downstream application by extracting phenological information from the gap-filled products. By effectively filling gaps and removing clouds, our method offers spatial-temporal continuous MODIS surface reflectance data, benefiting downstream applications such as phenology extraction and highlighting the potential of artificial intelligence technique in remote sense data processing.
{"title":"A temporal attention-based multi-scale generative adversarial network to fill gaps in time series of MODIS data for land surface phenology extraction","authors":"Yidan Wang , Wei Wu , Zhicheng Zhang , Ziming Li , Fan Zhang , Qinchuan Xin","doi":"10.1016/j.rse.2024.114546","DOIUrl":"10.1016/j.rse.2024.114546","url":null,"abstract":"<div><div>High-quality and continuous satellite data are essential for land surface studies such as monitoring of land surface phenology, but factors such as cloud contamination and sensor malfunction degrade the quality of remote sensing images and limit their utilization. Filling gaps and recovering missing information in time series of remote sensing images are vital for a wide range of downstream applications, such as land surface phenology extraction. Most existing gap-filling and cloud removal methods focus on individual or multi-temporal image reconstruction, but struggle with continuous and overlapping missing areas in time series data. In this study, we propose a Temporal Attention-Based Multi-Scale Generative Adversarial Network (TAMGAN) to reconstruct time series of Moderate Resolution Imaging Spectroradiometer (MODIS) data. TAMGAN leverages a Generative Adversarial Network (GAN) with a 3-dimensional Convolution Neural Networks (3DCNN) in its generator to reconstruct the missing areas in the annual time series of remote sensing images simultaneously. The temporal attention blocks are designed to capture the changing trends of surface reflectance over time. And multi-scale feature extraction and progressive concatenation are introduced to enhance spectral consistency and provide detailed texture information. Experiments are carried out on MOD09A1 products to evaluate the performance of the proposed network. The results show that TAMGAN outperformed the comparison methods across various evaluation metrics, particularly in handling large and continuous missing areas in the time series. Furthermore, we showcase an example of downstream application by extracting phenological information from the gap-filled products. By effectively filling gaps and removing clouds, our method offers spatial-temporal continuous MODIS surface reflectance data, benefiting downstream applications such as phenology extraction and highlighting the potential of artificial intelligence technique in remote sense data processing.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"318 ","pages":"Article 114546"},"PeriodicalIF":11.1,"publicationDate":"2024-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142793507","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}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号