Science of Remote Sensing最新文献_第4页

A spatio-temporal machine learning method for estimating high-resolution XCO2 in China 中国高分辨率XCO2估算的时空机器学习方法

IF 5.2 Q1 ENVIRONMENTAL SCIENCES

Science of Remote Sensing

Pub Date : 2025-12-27 DOI: 10.1016/j.srs.2025.100359

Fengxue Ruan , Fen Qin , Jie Li , Weichen Mu

Column-averaged dry air mole fraction of carbon dioxide (XCO₂) data is of great significance for addressing global climate change, monitoring carbon emissions. Currently, satellite XCO₂ exhibit significant spatial discontinuity, which makes it difficult to meet the needs of research at small spatial scales. Although machine learning methods have been widely used to fill the gaps in satellite XCO₂ data, mainstream methods are mostly data-driven mode, which, to some extent, limits the accuracy and generalization ability of the models. Given the limitations of existing studies in mining the spatiotemporal characteristics of XCO₂, this study innovatively proposes a new spatiotemporal XGBoost model (XGBKT) to generate high-resolution XCO₂ dataset covering the entire territory of China. This model focuses on the three major spatiotemporal characteristics of XCO₂, namely spatial correlation, temporal heterogeneity, and temporal periodicity. Through the spatiotemporal encoding strategy, these characteristics are skillfully transformed into features that the XGBoost model can efficiently utilize, thereby enabling the model to explore the spatiotemporal distribution pattern of XCO₂ and significantly improve its estimation accuracy and reliability. The research results indicate that: The XGBKT model significantly enhances the estimation performance and generalization ability of machine learning models, demonstrating clear advantages compared to mainstream machine learning methods; The XGBKT model validates the effectiveness of spatiotemporal characteristics, thereby further strengthening the interpretability of machine learning models. Overall, XGBKT is an effective method for accurately estimating XCO₂, providing a reliable data foundation for the fine-scale quantification of regional carbon cycling.

柱平均干空气二氧化碳摩尔分数（XCO2）数据对于应对全球气候变化、监测碳排放具有重要意义。目前，卫星XCO2具有明显的空间不连续性，难以满足小空间尺度的研究需求。虽然机器学习方法已被广泛用于填补卫星XCO2数据的空白，但主流方法多为数据驱动模式，这在一定程度上限制了模型的准确性和泛化能力。针对现有研究在挖掘XCO2时空特征方面的局限性，本研究创新性地提出了一种新的时空XGBoost模型（XGBKT），以生成覆盖中国全境的高分辨率XCO2数据集。该模型关注XCO2的三大时空特征，即空间相关性、时间异质性和时间周期性。通过时空编码策略，将这些特征巧妙转化为XGBoost模型能够有效利用的特征，使模型能够探索XCO2的时空分布格局，显著提高模型的估计精度和可靠性。研究结果表明：XGBKT模型显著提高了机器学习模型的估计性能和泛化能力，与主流机器学习方法相比优势明显；XGBKT模型验证了时空特征的有效性，从而进一步增强了机器学习模型的可解释性。总体而言，XGBKT是一种准确估算XCO2的有效方法，为区域碳循环的精细量化提供了可靠的数据基础。

{"title":"A spatio-temporal machine learning method for estimating high-resolution XCO2 in China","authors":"Fengxue Ruan , Fen Qin , Jie Li , Weichen Mu","doi":"10.1016/j.srs.2025.100359","DOIUrl":"10.1016/j.srs.2025.100359","url":null,"abstract":"<div><div>Column-averaged dry air mole fraction of carbon dioxide (XCO<sub>2</sub>) data is of great significance for addressing global climate change, monitoring carbon emissions. Currently, satellite XCO<sub>2</sub> exhibit significant spatial discontinuity, which makes it difficult to meet the needs of research at small spatial scales. Although machine learning methods have been widely used to fill the gaps in satellite XCO<sub>2</sub> data, mainstream methods are mostly data-driven mode, which, to some extent, limits the accuracy and generalization ability of the models. Given the limitations of existing studies in mining the spatiotemporal characteristics of XCO<sub>2</sub>, this study innovatively proposes a new spatiotemporal XGBoost model (XGBKT) to generate high-resolution XCO<sub>2</sub> dataset covering the entire territory of China. This model focuses on the three major spatiotemporal characteristics of XCO<sub>2</sub>, namely spatial correlation, temporal heterogeneity, and temporal periodicity. Through the spatiotemporal encoding strategy, these characteristics are skillfully transformed into features that the XGBoost model can efficiently utilize, thereby enabling the model to explore the spatiotemporal distribution pattern of XCO<sub>2</sub> and significantly improve its estimation accuracy and reliability. The research results indicate that: The XGBKT model significantly enhances the estimation performance and generalization ability of machine learning models, demonstrating clear advantages compared to mainstream machine learning methods; The XGBKT model validates the effectiveness of spatiotemporal characteristics, thereby further strengthening the interpretability of machine learning models. Overall, XGBKT is an effective method for accurately estimating XCO<sub>2</sub>, providing a reliable data foundation for the fine-scale quantification of regional carbon cycling.</div></div>","PeriodicalId":101147,"journal":{"name":"Science of Remote Sensing","volume":"13 ","pages":"Article 100359"},"PeriodicalIF":5.2,"publicationDate":"2025-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145938946","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Deep learning for epistemic uncertainty in SMAP-derived soil moisture estimates over the Kulfo watershed, Ethiopia 埃塞俄比亚Kulfo流域smap衍生土壤湿度估算中认知不确定性的深度学习

IF 5.2 Q1 ENVIRONMENTAL SCIENCES

Science of Remote Sensing

Pub Date : 2025-12-26 DOI: 10.1016/j.srs.2025.100357

Demiso Daba Dugassa , Aschalew Cherie Workneh , Babur Tesfaye Yersaw , Getachew Enssa Sedeta , Mulusew Bezabih Chane , Sintayehu Yadete Tola , Sufiyan Abdulmenan Ousman , Zelalem Anley Birhan

Precise soil moisture estimation is critical for irrigation scheduling and water resource management, especially in semi-arid regions like the Kulfo watershed, Ethiopia, where water availability is highly climate-dependent. However, quantifying the reliability of soil moisture estimates remains a key challenge. Standard satellite products like SMAP provide a measure of aleatoric uncertainty, but this offers limited insight into the confidence of a predictive model. This study develops a deep learning framework to produce a more reliable and better-calibrated measure of epistemic uncertainty, which directly quantifies analytical confidence. Using data from the Soil Moisture Active Passive (SMAP) mission, time-series data were processed in Google Earth Engine (GEE) and structured for deep learning models. Ten different models were tested: one basic long short-term memory (LSTM) model that doesn't consider uncertainty and nine uncertainty-aware model (including five deep ensemble models, Monte Carlo Dropout (MC Dropout), and Quantile Regression models). The Deep Ensemble model achieved the highest accuracy (RMSE = 0.131, R² = 0.993) and a 94.51 % more reliable uncertainty estimate than the baseline. The MC Dropout model also delivered a much-improved confidence estimate, showing a 41.79 % enhancement. Among the quantile regression models, the 5th percentile model produced a 40.58 % better confidence estimate. Among the quantile regression models, the 5th percentile model produced a 40.58 % better confidence estimate. This study demonstrates that coupling an LSTM-based model with the Deep Ensemble method provides a highly reliable and precise measure of model-specific analytical confidence for soil moisture estimates, offering a more trustworthy approach for decision-making in data-limited regions.

精确的土壤湿度估算对于灌溉计划和水资源管理至关重要，特别是在像埃塞俄比亚库尔福流域这样的半干旱地区，那里的水资源供应高度依赖气候。然而，量化土壤湿度估计的可靠性仍然是一个关键的挑战。像SMAP这样的标准卫星产品提供了一种任意不确定性的测量，但这对预测模型的置信度提供了有限的见解。本研究开发了一个深度学习框架，以产生更可靠和更好校准的认知不确定性测量，直接量化分析置信度。利用土壤湿度主动式被动探测（SMAP）任务数据，在谷歌Earth Engine （GEE）中对时间序列数据进行处理，并构建深度学习模型。测试了10种不同的模型：1种不考虑不确定性的基本长短期记忆（LSTM）模型和9种不确定性感知模型（包括5种深度集成模型、蒙特卡罗Dropout （MC Dropout）和分位回归模型）。Deep Ensemble模型的精度最高（RMSE = 0.131, R2 = 0.993），不确定性估计比基线高94.51%。MC Dropout模型也提供了一个大大提高的置信度估计，显示出41.79%的提高。在分位数回归模型中，第5百分位模型的置信度估计提高了40.58%。在分位数回归模型中，第5百分位模型的置信度估计提高了40.58%。该研究表明，基于lstm的模型与深度集合方法的耦合为土壤湿度估算提供了高度可靠和精确的模型特定分析置信度度量，为数据有限地区的决策提供了更可靠的方法。

{"title":"Deep learning for epistemic uncertainty in SMAP-derived soil moisture estimates over the Kulfo watershed, Ethiopia","authors":"Demiso Daba Dugassa , Aschalew Cherie Workneh , Babur Tesfaye Yersaw , Getachew Enssa Sedeta , Mulusew Bezabih Chane , Sintayehu Yadete Tola , Sufiyan Abdulmenan Ousman , Zelalem Anley Birhan","doi":"10.1016/j.srs.2025.100357","DOIUrl":"10.1016/j.srs.2025.100357","url":null,"abstract":"<div><div>Precise soil moisture estimation is critical for irrigation scheduling and water resource management, especially in semi-arid regions like the Kulfo watershed, Ethiopia, where water availability is highly climate-dependent. However, quantifying the reliability of soil moisture estimates remains a key challenge. Standard satellite products like SMAP provide a measure of aleatoric uncertainty, but this offers limited insight into the confidence of a predictive model. This study develops a deep learning framework to produce a more reliable and better-calibrated measure of epistemic uncertainty, which directly quantifies analytical confidence. Using data from the Soil Moisture Active Passive (SMAP) mission, time-series data were processed in Google Earth Engine (GEE) and structured for deep learning models. Ten different models were tested: one basic long short-term memory (LSTM) model that doesn't consider uncertainty and nine uncertainty-aware model (including five deep ensemble models, Monte Carlo Dropout (MC Dropout), and Quantile Regression models). The Deep Ensemble model achieved the highest accuracy (RMSE = 0.131, R<sup>2</sup> = 0.993) and a 94.51 % more reliable uncertainty estimate than the baseline. The MC Dropout model also delivered a much-improved confidence estimate, showing a 41.79 % enhancement. Among the quantile regression models, the 5th percentile model produced a 40.58 % better confidence estimate. Among the quantile regression models, the 5th percentile model produced a 40.58 % better confidence estimate. This study demonstrates that coupling an LSTM-based model with the Deep Ensemble method provides a highly reliable and precise measure of model-specific analytical confidence for soil moisture estimates, offering a more trustworthy approach for decision-making in data-limited regions.</div></div>","PeriodicalId":101147,"journal":{"name":"Science of Remote Sensing","volume":"13 ","pages":"Article 100357"},"PeriodicalIF":5.2,"publicationDate":"2025-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145939363","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Divergent GPP dynamics in alpine and temperate grasslands: Hierarchical climatic controls across the Qinghai-Tibetan and Mongolian Plateaus 高寒和温带草原GPP动态差异：青藏高原和蒙古高原的分层气候控制

IF 5.2 Q1 ENVIRONMENTAL SCIENCES

Science of Remote Sensing

Pub Date : 2025-12-25 DOI: 10.1016/j.srs.2025.100360

Yusi Zhang , Gang Bao , Zhonghua He , Yuhai Bao , Zhihui Yuan , Siqin Tong

Alpine grasslands on the Qinghai-Tibet Plateau and temperate grasslands on the Mongolian Plateau are key components of the global carbon cycle but differ markedly in their responses to climate change. To investigate the spatiotemporal variations in Gross Primary Productivity (GPP) and their response to climate change in these two types of grasslands, we developed a novel Random Forest Regression-Light Use Efficiency-Solar Induced Fluorescence (RFR-LUE-SIF) model that integrates machine-learning regression with physiological efficiency principles and satellite-derived SIF observations. This framework bridges tower-based GPP observations with large-scale remote-sensing estimates, improving model interpretability and accuracy. The model reproduced observed GPP with high fidelity (R² = 0.91), identifying EVI, NIRv, and GOSIF as the most influential predictors. Spatially, alpine grassland GPP decreases from southeast to northwest, while temperate grassland GPP declines from northeast to southwest. From 2001 to 2023, both grassland types exhibited increasing GPP trends, with temperate grasslands showing a faster rise, indicating stronger climatic sensitivity. Further, partial correlation analysis and Structural Equation Modeling (SEM) reveal that alpine grassland productivity is generally more sensitive to temperature, particularly under adequate moisture conditions, whereas temperate grasslands exhibited stronger dependence on precipitation and vapor pressure deficit (VPD). The proposed RFR-LUE-SIF model provides a scalable, data-driven, and physiologically consistent approach for assessing grassland carbon dynamics and their hierarchical climatic responses across contrasting ecosystems.

青藏高原高寒草原和蒙古高原温带草原是全球碳循环的重要组成部分，但对气候变化的响应存在显著差异。为了研究这两种草地的总初级生产力（GPP）的时空变化及其对气候变化的响应，我们建立了一种新的随机森林回归-光利用效率-太阳诱导荧光（RFR-LUE-SIF）模型，该模型将机器学习回归与生理效率原理和卫星衍生的SIF观测相结合。该框架将基于塔的GPP观测与大尺度遥感估算相结合，提高了模型的可解释性和准确性。该模型高保真地再现了观测到的GPP (R2 = 0.91)，确定EVI、NIRv和GOSIF是最具影响力的预测因子。从空间上看，高寒草地GPP从东南向西北递减，温带草地GPP从东北向西南递减。2001 - 2023年，两种草地类型的GPP均呈上升趋势，其中温带草地上升较快，气候敏感性更强。此外，偏相关分析和结构方程模型（SEM）表明，高寒草地生产力总体上对温度更敏感，特别是在水分充足的条件下，而温带草地生产力对降水和水汽压亏缺（VPD）的依赖性更强。提出的RFR-LUE-SIF模型提供了一种可扩展的、数据驱动的、生理上一致的方法来评估草地碳动态及其在不同生态系统中的分层气候响应。

{"title":"Divergent GPP dynamics in alpine and temperate grasslands: Hierarchical climatic controls across the Qinghai-Tibetan and Mongolian Plateaus","authors":"Yusi Zhang , Gang Bao , Zhonghua He , Yuhai Bao , Zhihui Yuan , Siqin Tong","doi":"10.1016/j.srs.2025.100360","DOIUrl":"10.1016/j.srs.2025.100360","url":null,"abstract":"<div><div>Alpine grasslands on the Qinghai-Tibet Plateau and temperate grasslands on the Mongolian Plateau are key components of the global carbon cycle but differ markedly in their responses to climate change. To investigate the spatiotemporal variations in Gross Primary Productivity (GPP) and their response to climate change in these two types of grasslands, we developed a novel Random Forest Regression-Light Use Efficiency-Solar Induced Fluorescence (RFR-LUE-SIF) model that integrates machine-learning regression with physiological efficiency principles and satellite-derived SIF observations. This framework bridges tower-based GPP observations with large-scale remote-sensing estimates, improving model interpretability and accuracy. The model reproduced observed GPP with high fidelity (R<sup>2</sup> = 0.91), identifying EVI, NIRv, and GOSIF as the most influential predictors. Spatially, alpine grassland GPP decreases from southeast to northwest, while temperate grassland GPP declines from northeast to southwest. From 2001 to 2023, both grassland types exhibited increasing GPP trends, with temperate grasslands showing a faster rise, indicating stronger climatic sensitivity. Further, partial correlation analysis and Structural Equation Modeling (SEM) reveal that alpine grassland productivity is generally more sensitive to temperature, particularly under adequate moisture conditions, whereas temperate grasslands exhibited stronger dependence on precipitation and vapor pressure deficit (VPD). The proposed RFR-LUE-SIF model provides a scalable, data-driven, and physiologically consistent approach for assessing grassland carbon dynamics and their hierarchical climatic responses across contrasting ecosystems.</div></div>","PeriodicalId":101147,"journal":{"name":"Science of Remote Sensing","volume":"13 ","pages":"Article 100360"},"PeriodicalIF":5.2,"publicationDate":"2025-12-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145939362","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Assessment of ground deformation in Mandalay, Myanmar, using InSAR with Sentinel-1 data after the March 2025 earthquake 基于InSAR和Sentinel-1数据的缅甸曼德勒2025年3月地震后地面变形评估

IF 5.2 Q1 ENVIRONMENTAL SCIENCES

Science of Remote Sensing

Pub Date : 2025-12-24 DOI: 10.1016/j.srs.2025.100358

Behzad Taghi-Lou, Michael Schultz, Andreas Braun, Volker Hochschild

This study quantifies coseismic ground deformation associated with the March 28, 2025 earthquake sequence (Mw 6.7–7.7) in the Mandalay Region of Myanmar. Differential interferometric synthetic-aperture radar (DInSAR) was applied to Sentinel-1A single-look complex (SLC) data, using pre-event images acquired six and four days before the earthquakes and post-event images acquired five and nineteen days after. The interferograms were processed in ESA SNAP, phase-unwrapped with SNAPHU, and geocoded using the 30 m SRTM digital elevation model (DEM). A coherence threshold of 0.35 was applied to ensure reliable phase retrieval. Peak line-of-sight (LOS) displacements reach +0.24 m (uplift) and −0.62 m (subsidence) for the ascending pair, and +0.22 m (uplift) and −0.64 m (subsidence) for the descending pair. Decomposition of the paired LOS maps yields vertical motion between +0.74 m and −1.57 m and east–west offsets between −1.99 m (westward) and +1.29 m (eastward), corresponding to a horizontal-to-vertical displacement ratio of 1.27. The deformation field closely follows the surface trace of the dextral Sagaing Fault, confirming it as the primary seismogenic structure. Opposite-sense deformation is observed across the fault, with subsidence and eastward motion on the western block and uplift and westward motion on the eastern block. These results provide a high-resolution deformation field for this event and highlight the value of open-access Sentinel-1 data for rapid seismic hazard assessment in data-scarce regions.

本研究量化了与2025年3月28日缅甸曼德勒地区地震序列（Mw 6.7-7.7）相关的同震地面变形。差分干涉合成孔径雷达（DInSAR）应用于Sentinel-1A单视复合体（SLC）数据，使用地震前6天和4天获取的事件前图像以及地震后5天和19天获取的事件后图像。干涉图在ESA SNAP中进行处理，使用SNAPHU进行相位解包裹，并使用30 m SRTM数字高程模型（DEM）进行地理编码。相干阈值为0.35，保证了相位恢复的可靠性。上升对的峰值视线位移达到+0.24 m（隆升）和- 0.62 m（沉降），下降对的峰值视线位移达到+0.22 m（隆升）和- 0.64 m（沉降）。对成对的LOS地图进行分解，得到垂直运动在+0.74 m和- 1.57 m之间，东西偏移量在- 1.99 m（向西）和+1.29 m（向东）之间，对应的水平与垂直位移比为1.27。变形场与右旋实皆断层的地表轨迹密切相关，证实了实皆断层是主要的发震构造。整个断层呈现相反意义的变形，西部地块为沉降东移，东部地块为隆升西移。这些结果为该事件提供了高分辨率变形场，并突出了开放获取的Sentinel-1数据对数据稀缺地区快速地震危害评估的价值。

{"title":"Assessment of ground deformation in Mandalay, Myanmar, using InSAR with Sentinel-1 data after the March 2025 earthquake","authors":"Behzad Taghi-Lou, Michael Schultz, Andreas Braun, Volker Hochschild","doi":"10.1016/j.srs.2025.100358","DOIUrl":"10.1016/j.srs.2025.100358","url":null,"abstract":"<div><div>This study quantifies coseismic ground deformation associated with the March 28, 2025 earthquake sequence (Mw 6.7–7.7) in the Mandalay Region of Myanmar. Differential interferometric synthetic-aperture radar (DInSAR) was applied to Sentinel-1A single-look complex (SLC) data, using pre-event images acquired six and four days before the earthquakes and post-event images acquired five and nineteen days after. The interferograms were processed in ESA SNAP, phase-unwrapped with SNAPHU, and geocoded using the 30 m SRTM digital elevation model (DEM). A coherence threshold of 0.35 was applied to ensure reliable phase retrieval. Peak line-of-sight (LOS) displacements reach +0.24 m (uplift) and −0.62 m (subsidence) for the ascending pair, and +0.22 m (uplift) and −0.64 m (subsidence) for the descending pair. Decomposition of the paired LOS maps yields vertical motion between +0.74 m and −1.57 m and east–west offsets between −1.99 m (westward) and +1.29 m (eastward), corresponding to a horizontal-to-vertical displacement ratio of 1.27. The deformation field closely follows the surface trace of the dextral Sagaing Fault, confirming it as the primary seismogenic structure. Opposite-sense deformation is observed across the fault, with subsidence and eastward motion on the western block and uplift and westward motion on the eastern block. These results provide a high-resolution deformation field for this event and highlight the value of open-access Sentinel-1 data for rapid seismic hazard assessment in data-scarce regions.</div></div>","PeriodicalId":101147,"journal":{"name":"Science of Remote Sensing","volume":"13 ","pages":"Article 100358"},"PeriodicalIF":5.2,"publicationDate":"2025-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145938943","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Beyond spectral signals: Geographic features drive bathymetric accuracy in the turbid Sancha Lake using machine learning 超越光谱信号：利用机器学习，地理特征驱动浑浊的三岔湖测深精度

IF 5.2 Q1 ENVIRONMENTAL SCIENCES

Science of Remote Sensing

Pub Date : 2025-12-23 DOI: 10.1016/j.srs.2025.100355

Xiaojuan Li , Wei Zhang , Hongrui Zheng , Zhongqiang Wu , Hongliang Lu

Accurate bathymetric mapping in inland water bodies presents significant challenges for conventional optical remote sensing due to complex water quality conditions and variable bottom types. This study introduces a novel Spectral-Geospatial XGBoost Regression (SG-XGBoost) model that revolutionizes depth estimation by integrating comprehensive spectral transformations with explicit geographic coordinates through gradient boosting methodology. Applied to Sancha Lake, a morphologically complex reservoir in China's Upper Yangtze watershed, the model achieved exceptional performance with R² = 0.91 and RMSE = 1.66m, representing 70 % improvement over traditional empirical methods (Stumpf, Log-Linear) and 21 % advancement beyond Random Forest. The iterative error correction and sophisticated regularization of the gradient boosting methodology not only enable the effective exploitation of spatial-spectral interactions but also ensure better accuracy is maintained across all depth ranges (2–31m). The feature importance analysis revealed an unexpected finding, the geographic coordinates dominated predictive power (85 % contribution), while spectral features contributed minimally, challenging fundamental assumptions about optical bathymetry. The iterative error correction and sophisticated regularization of the gradient boosting methodology not only enable the effective exploitation of spatial-spectral interactions but also ensure better accuracy is maintained across all depth ranges (2–31m). Bathymetric maps generated by SG-XGBoost successfully captured fine-scale morphological features invisible to conventional approaches, including channels <30m wide and subtle depth variations of 1–2m. Despite limitations in extreme turbidity and site-specificity requiring readjustment for new water bodies, this research establishes gradient boosting with spatial-spectral integration as a transformative approach for inland water bathymetry, with broader implications for aquatic remote sensing applications including water quality monitoring and habitat mapping.

由于复杂的水质条件和多变的海底类型，内陆水体的精确测深测绘对传统光学遥感提出了重大挑战。本研究引入了一种新的光谱-地理空间XGBoost回归（SG-XGBoost）模型，该模型通过梯度增强方法将综合光谱变换与显式地理坐标相结合，从而彻底改变了深度估计。应用于三岔湖这一形态复杂的长江上游水库，该模型取得了优异的效果，R2 = 0.91, RMSE = 1.66m，比传统的经验方法（Stumpf、Log-Linear）提高了70%，比随机森林方法提高了21%。梯度增强方法的迭代误差校正和复杂正则化不仅可以有效地利用空间-光谱相互作用，而且可以确保在所有深度范围（2 - 31米）保持更好的精度。特征重要性分析揭示了一个意想不到的发现，地理坐标主导了预测能力（85%的贡献），而光谱特征贡献最小，挑战了光学测深的基本假设。梯度增强方法的迭代误差校正和复杂正则化不仅可以有效地利用空间-光谱相互作用，而且可以确保在所有深度范围（2 - 31米）保持更好的精度。SG-XGBoost生成的水深图成功捕获了传统方法看不到的精细尺度形态特征，包括30米宽的通道和1 - 2米的细微深度变化。尽管在极端浊度和需要重新调整新水体的地点特异性方面存在局限性，但本研究建立了梯度增强与空间光谱集成作为内陆水域测深的变革方法，对包括水质监测和栖息地测绘在内的水生遥感应用具有更广泛的意义。

{"title":"Beyond spectral signals: Geographic features drive bathymetric accuracy in the turbid Sancha Lake using machine learning","authors":"Xiaojuan Li , Wei Zhang , Hongrui Zheng , Zhongqiang Wu , Hongliang Lu","doi":"10.1016/j.srs.2025.100355","DOIUrl":"10.1016/j.srs.2025.100355","url":null,"abstract":"<div><div>Accurate bathymetric mapping in inland water bodies presents significant challenges for conventional optical remote sensing due to complex water quality conditions and variable bottom types. This study introduces a novel Spectral-Geospatial XGBoost Regression (SG-XGBoost) model that revolutionizes depth estimation by integrating comprehensive spectral transformations with explicit geographic coordinates through gradient boosting methodology. Applied to Sancha Lake, a morphologically complex reservoir in China's Upper Yangtze watershed, the model achieved exceptional performance with R<sup>2</sup> = 0.91 and RMSE = 1.66m, representing 70 % improvement over traditional empirical methods (Stumpf, Log-Linear) and 21 % advancement beyond Random Forest. The iterative error correction and sophisticated regularization of the gradient boosting methodology not only enable the effective exploitation of spatial-spectral interactions but also ensure better accuracy is maintained across all depth ranges (2–31m). The feature importance analysis revealed an unexpected finding, the geographic coordinates dominated predictive power (85 % contribution), while spectral features contributed minimally, challenging fundamental assumptions about optical bathymetry. The iterative error correction and sophisticated regularization of the gradient boosting methodology not only enable the effective exploitation of spatial-spectral interactions but also ensure better accuracy is maintained across all depth ranges (2–31m). Bathymetric maps generated by SG-XGBoost successfully captured fine-scale morphological features invisible to conventional approaches, including channels <30m wide and subtle depth variations of 1–2m. Despite limitations in extreme turbidity and site-specificity requiring readjustment for new water bodies, this research establishes gradient boosting with spatial-spectral integration as a transformative approach for inland water bathymetry, with broader implications for aquatic remote sensing applications including water quality monitoring and habitat mapping.</div></div>","PeriodicalId":101147,"journal":{"name":"Science of Remote Sensing","volume":"13 ","pages":"Article 100355"},"PeriodicalIF":5.2,"publicationDate":"2025-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145977577","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Satellite-derived seasonal fluctuations in surface displacement and soil moisture: Implications for landslide activity 地表位移和土壤湿度的卫星季节性波动：对滑坡活动的影响

IF 5.2 Q1 ENVIRONMENTAL SCIENCES

Science of Remote Sensing

Pub Date : 2025-12-17 DOI: 10.1016/j.srs.2025.100354

Chiao-Yin Lu , Yu-Chang Chan , Chung-Ray Chu , Che-Hsin Liu , Shih-Chiang Lee , Yu-Chung Hsieh , Jyr-Ching Hu , Chih-Hsin Chang

Seasonal surface fluctuations are commonly influenced by environmental variations, with both water presence and geological conditions playing key roles in surface deformation. These short-term signals often complicate the interpretation of time series data, particularly in complex and mountainous regions where in situ hydrological data such as groundwater levels, pore water pressure, and soil moisture are scarce or difficult to obtain. This study investigates two representative slow-moving landslides exhibiting seasonal variations by applying the small baseline subset (SBAS) method, one of the core approaches within the multitemporal InSAR (MTInSAR) framework. The SBAS-derived time series are further analyzed in combination with satellite-derived soil moisture data. Given the limited availability of high-resolution hydrological observations, satellite-derived soil moisture was adopted as a validated proxy to represent the average hydrological conditions over each landslide area. The results reveal notable patterns of seasonal surface fluctuations driven by hydrological variations, and demonstrate that their expression is further modulated by lithological conditions. Based on available data, we infer that in sedimentary rock areas, a high water storage coefficient causes hydrological loading to dominate seasonal surface displacement, resulting in a negative correlation. In contrast, pore water pressure plays a dominant role in metamorphic rock areas, leading to a positive correlation. This study demonstrates the potential of satellite-derived hydrological data to complement InSAR time series in regions with scarce in situ monitoring. These findings offer valuable and useful insights for a further understanding slow-moving landslide behaviors, particularly in distinguishing and identifying accelerated movement signals from seasonal fluctuations, and for improving slope failure early warning systems.

季节性地表波动通常受到环境变化的影响，水的存在和地质条件在地表变形中起着关键作用。这些短期信号往往使时间序列数据的解释复杂化，特别是在地下水水位、孔隙水压力和土壤湿度等原位水文数据稀缺或难以获得的复杂山区。本研究通过应用multitemporal InSAR （MTInSAR）框架内的核心方法之一——小基线子集（SBAS）方法，调查了两个具有代表性的表现出季节性变化的缓慢移动滑坡。sbas导出的时间序列与卫星导出的土壤湿度数据结合进一步分析。鉴于高分辨率水文观测的可用性有限，采用卫星获得的土壤湿度作为有效的代理来代表每个滑坡区的平均水文条件。结果揭示了由水文变化驱动的季节地表波动的显著模式，并表明其表达进一步受到岩性条件的调节。根据现有数据，我们推断在沉积岩区，高储水系数导致水文负荷主导季节性地表位移，形成负相关。而在变质岩区孔隙水压力起主导作用，两者呈正相关关系。这项研究表明，卫星衍生水文数据在缺乏现场监测的地区补充InSAR时间序列的潜力。这些发现为进一步理解缓慢移动的滑坡行为提供了有价值和有用的见解，特别是在区分和识别季节性波动中的加速运动信号，以及改进边坡破坏早期预警系统方面。

{"title":"Satellite-derived seasonal fluctuations in surface displacement and soil moisture: Implications for landslide activity","authors":"Chiao-Yin Lu , Yu-Chang Chan , Chung-Ray Chu , Che-Hsin Liu , Shih-Chiang Lee , Yu-Chung Hsieh , Jyr-Ching Hu , Chih-Hsin Chang","doi":"10.1016/j.srs.2025.100354","DOIUrl":"10.1016/j.srs.2025.100354","url":null,"abstract":"<div><div>Seasonal surface fluctuations are commonly influenced by environmental variations, with both water presence and geological conditions playing key roles in surface deformation. These short-term signals often complicate the interpretation of time series data, particularly in complex and mountainous regions where in situ hydrological data such as groundwater levels, pore water pressure, and soil moisture are scarce or difficult to obtain. This study investigates two representative slow-moving landslides exhibiting seasonal variations by applying the small baseline subset (SBAS) method, one of the core approaches within the multitemporal InSAR (MTInSAR) framework. The SBAS-derived time series are further analyzed in combination with satellite-derived soil moisture data. Given the limited availability of high-resolution hydrological observations, satellite-derived soil moisture was adopted as a validated proxy to represent the average hydrological conditions over each landslide area. The results reveal notable patterns of seasonal surface fluctuations driven by hydrological variations, and demonstrate that their expression is further modulated by lithological conditions. Based on available data, we infer that in sedimentary rock areas, a high water storage coefficient causes hydrological loading to dominate seasonal surface displacement, resulting in a negative correlation. In contrast, pore water pressure plays a dominant role in metamorphic rock areas, leading to a positive correlation. This study demonstrates the potential of satellite-derived hydrological data to complement InSAR time series in regions with scarce in situ monitoring. These findings offer valuable and useful insights for a further understanding slow-moving landslide behaviors, particularly in distinguishing and identifying accelerated movement signals from seasonal fluctuations, and for improving slope failure early warning systems.</div></div>","PeriodicalId":101147,"journal":{"name":"Science of Remote Sensing","volume":"13 ","pages":"Article 100354"},"PeriodicalIF":5.2,"publicationDate":"2025-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145939361","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Harmonized Tasseled Cap Transformation coefficients for Landsat 8 and 9 OLI sensors using surface reflectance from near-coincident underfly observations Landsat 8和9 OLI传感器的协调流苏帽转换系数，利用接近一致的下飞观测的表面反射率

IF 5.2 Q1 ENVIRONMENTAL SCIENCES

Science of Remote Sensing

Pub Date : 2025-12-14 DOI: 10.1016/j.srs.2025.100353

Shuai Wang , Lijuan Yang , Tingting Shi , Jin Chen

Tasseled Cap Transformation (TCT) is a widely applied remote sensing technique for dimensionality reduction and physical feature enhancement, valued for its interpretability and efficiency. While TCT coefficients have been developed for numerous sensors, no dedicated coefficient set has been proposed to date for Landsat 9 Operational Land Imager-2 (OLI-2) sensor. This study addresses this gap by deriving the first TCT coefficients for Landsat 9 OLI-2 based on surface reflectance data. To ensure cross-sensor consistency, we simultaneously recalculated Landsat 8 Operational Land Imager (OLI) coefficients using strictly matched underfly image pairs and identical sample selection protocols. This harmonized derivation strategy minimizes methodological and sampling-induced discrepancies, enhancing compatibility between the two Landsat sensors. More importantly, the proposed coefficients offer improved performance in capturing wetness-related information across diverse ecological settings. This makes them especially suitable for applications involving soil moisture monitoring, vegetation stress detection, and hydrological modeling in spatially and temporally heterogeneous landscapes. The validation results demonstrate that the newly proposed coefficients effectively enhance spectral differences among surface features with varying brightness, greenness, and moisture content. Moreover, the TCT components from Landsat 8 OLI and Landsat 9 OLI-2 exhibit strong agreement across all three components (R² > 0.96, approaching 1), underscoring the high consistency of the derived coefficients. Sensitivity analyses further reveal that the wetness and greenness components remain highly stable under varying sample selection conditions, while the brightness component, though slightly more sensitive, still maintains angular differences within 5°. The results highlight the improved physical consistency and cross-sensor compatibility of the proposed coefficients, facilitating more robust long-term environmental monitoring and multi-source data integration in Earth observation studies.

流苏帽变换（TCT）是一种应用广泛的遥感降维和物理特征增强技术，以其可解释性和高效性而受到重视。虽然已经为许多传感器开发了TCT系数，但迄今为止还没有为Landsat 9操作陆地成像仪-2 （OLI-2）传感器提出专用系数集。本研究通过基于地表反射率数据推导Landsat 9 OLI-2的第一个TCT系数来解决这一问题。为了确保跨传感器的一致性，我们使用严格匹配的下飞图像对和相同的样本选择协议，同时重新计算了Landsat 8操作陆地成像仪（OLI）系数。这种协调的推导策略最大限度地减少了方法和采样引起的差异，增强了两个陆地卫星传感器之间的兼容性。更重要的是，所提出的系数在捕获不同生态环境中与湿度相关的信息方面提供了更好的性能。这使得它们特别适用于涉及土壤湿度监测、植被应力检测和空间和时间异质性景观的水文建模的应用。验证结果表明，新提出的系数有效地增强了不同亮度、绿色和含水量地物之间的光谱差异。此外，Landsat 8 OLI和Landsat 9 OLI-2的TCT分量在所有三个分量之间表现出很强的一致性（R2 > 0.96，接近1），强调了推导系数的高度一致性。灵敏度分析进一步表明，湿度和绿色成分在不同的样品选择条件下保持高度稳定，而亮度成分虽然稍微敏感，但仍然保持在5°内的角差。结果表明，所提出的系数提高了物理一致性和跨传感器兼容性，有助于在地球观测研究中进行更稳健的长期环境监测和多源数据集成。

{"title":"Harmonized Tasseled Cap Transformation coefficients for Landsat 8 and 9 OLI sensors using surface reflectance from near-coincident underfly observations","authors":"Shuai Wang , Lijuan Yang , Tingting Shi , Jin Chen","doi":"10.1016/j.srs.2025.100353","DOIUrl":"10.1016/j.srs.2025.100353","url":null,"abstract":"<div><div>Tasseled Cap Transformation (TCT) is a widely applied remote sensing technique for dimensionality reduction and physical feature enhancement, valued for its interpretability and efficiency. While TCT coefficients have been developed for numerous sensors, no dedicated coefficient set has been proposed to date for Landsat 9 Operational Land Imager-2 (OLI-2) sensor. This study addresses this gap by deriving the first TCT coefficients for Landsat 9 OLI-2 based on surface reflectance data. To ensure cross-sensor consistency, we simultaneously recalculated Landsat 8 Operational Land Imager (OLI) coefficients using strictly matched underfly image pairs and identical sample selection protocols. This harmonized derivation strategy minimizes methodological and sampling-induced discrepancies, enhancing compatibility between the two Landsat sensors. More importantly, the proposed coefficients offer improved performance in capturing wetness-related information across diverse ecological settings. This makes them especially suitable for applications involving soil moisture monitoring, vegetation stress detection, and hydrological modeling in spatially and temporally heterogeneous landscapes. The validation results demonstrate that the newly proposed coefficients effectively enhance spectral differences among surface features with varying brightness, greenness, and moisture content. Moreover, the TCT components from Landsat 8 OLI and Landsat 9 OLI-2 exhibit strong agreement across all three components (R<sup>2</sup> > 0.96, approaching 1), underscoring the high consistency of the derived coefficients. Sensitivity analyses further reveal that the wetness and greenness components remain highly stable under varying sample selection conditions, while the brightness component, though slightly more sensitive, still maintains angular differences within 5°. The results highlight the improved physical consistency and cross-sensor compatibility of the proposed coefficients, facilitating more robust long-term environmental monitoring and multi-source data integration in Earth observation studies.</div></div>","PeriodicalId":101147,"journal":{"name":"Science of Remote Sensing","volume":"13 ","pages":"Article 100353"},"PeriodicalIF":5.2,"publicationDate":"2025-12-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145799954","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

High-resolution winter cover crop mapping with PlanetScope imagery: Comparative analysis of Random Forest, Convolutional Neural Network, and unsupervised classification PlanetScope图像的高分辨率冬季覆盖作物制图：随机森林、卷积神经网络和无监督分类的比较分析

IF 5.2 Q1 ENVIRONMENTAL SCIENCES

Science of Remote Sensing

Pub Date : 2025-12-12 DOI: 10.1016/j.srs.2025.100351

Kanru Chen

Effective monitoring of agricultural conservation practices is essential for evaluating environmental outcomes and guiding land management strategies. This study assessed three satellite-based classification methods to estimate winter cover crop adoption across Benton County, Indiana, from 2021 to 2023, and compared their results with traditional field-based transect surveys. Using 3-m PlanetScope imagery and a consistent preprocessing and validation pipeline, we implemented (1) unsupervised Iso-cluster classification with five vegetation indices, (2) supervised Random Forest (RF) models, and (3) deep learning-based Convolutional Neural Networks (CNNs), tailored separately for December and April imagery. Supervised methods outperformed the unsupervised approach. RF models achieved F1 scores of 0.98 (December) and 0.96 (April), while CNNs reached 0.97 and 0.92, respectively. Unsupervised classification yielded lower accuracy (F1 ≤ 0.77), particularly under heterogeneous spring conditions. While transect surveys reported 24–128 % higher cover crop acreage than satellite-based estimates, the spatial and temporal patterns captured by both methods were similar, highlighting trends such as higher adoption after corn than soybean and substantial seasonal variation. Multi-year analysis revealed that less than 1 % of fields maintained continuous cover cropping across three consecutive winters, indicating predominantly intermittent adoption. These findings underscore the value of satellite imagery for full-coverage, repeatable assessments of conservation practice adoption. Scalable, remotely sensed classification enables timely evaluation of program effectiveness and supports adaptive land management to improve soil health and water quality at county scales.

有效监测农业保护措施对于评估环境结果和指导土地管理战略至关重要。本研究评估了三种基于卫星的分类方法，以估计2021年至2023年印第安纳州本顿县冬季覆盖作物的采用情况，并将其结果与传统的基于实地的样带调查进行了比较。利用3米PlanetScope图像和一致的预处理和验证管道，我们实现了(1)基于5个植被指数的无监督等聚类分类，(2)监督随机森林（RF）模型，(3)基于深度学习的卷积神经网络（cnn），分别针对12月和4月的图像进行了定制。有监督方法优于无监督方法。RF模型的F1得分分别为0.98（12月）和0.96（4月），cnn分别为0.97和0.92。非监督分类的准确率较低（F1≤0.77），特别是在异质弹簧条件下。虽然样带调查报告的覆盖作物面积比基于卫星的估计高24 - 128%，但两种方法捕获的时空模式相似，突出了玉米种植后的采用率高于大豆，以及明显的季节变化等趋势。多年分析显示，只有不到1%的农田在连续三个冬季保持覆盖种植，这表明主要是间歇性采用。这些发现强调了卫星图像对保护实践采用的全覆盖、可重复评估的价值。可扩展的遥感分类能够及时评估项目有效性，并支持适应性土地管理，以改善县尺度上的土壤健康和水质。

{"title":"High-resolution winter cover crop mapping with PlanetScope imagery: Comparative analysis of Random Forest, Convolutional Neural Network, and unsupervised classification","authors":"Kanru Chen","doi":"10.1016/j.srs.2025.100351","DOIUrl":"10.1016/j.srs.2025.100351","url":null,"abstract":"<div><div>Effective monitoring of agricultural conservation practices is essential for evaluating environmental outcomes and guiding land management strategies. This study assessed three satellite-based classification methods to estimate winter cover crop adoption across Benton County, Indiana, from 2021 to 2023, and compared their results with traditional field-based transect surveys. Using 3-m PlanetScope imagery and a consistent preprocessing and validation pipeline, we implemented (1) unsupervised Iso-cluster classification with five vegetation indices, (2) supervised Random Forest (RF) models, and (3) deep learning-based Convolutional Neural Networks (CNNs), tailored separately for December and April imagery. Supervised methods outperformed the unsupervised approach. RF models achieved F1 scores of 0.98 (December) and 0.96 (April), while CNNs reached 0.97 and 0.92, respectively. Unsupervised classification yielded lower accuracy (F1 ≤ 0.77), particularly under heterogeneous spring conditions. While transect surveys reported 24–128 % higher cover crop acreage than satellite-based estimates, the spatial and temporal patterns captured by both methods were similar, highlighting trends such as higher adoption after corn than soybean and substantial seasonal variation. Multi-year analysis revealed that less than 1 % of fields maintained continuous cover cropping across three consecutive winters, indicating predominantly intermittent adoption. These findings underscore the value of satellite imagery for full-coverage, repeatable assessments of conservation practice adoption. Scalable, remotely sensed classification enables timely evaluation of program effectiveness and supports adaptive land management to improve soil health and water quality at county scales.</div></div>","PeriodicalId":101147,"journal":{"name":"Science of Remote Sensing","volume":"13 ","pages":"Article 100351"},"PeriodicalIF":5.2,"publicationDate":"2025-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145938945","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Development of a 10 m daily seamless surface reflectance data cube based on Sentinel-2 constellation for generating the reference true-value products at Wanglang mountain area, China 基于Sentinel-2星座的10 m日无缝地表反射率数据立方生成参考真值产品

IF 5.2 Q1 ENVIRONMENTAL SCIENCES

Science of Remote Sensing

Pub Date : 2025-12-09 DOI: 10.1016/j.srs.2025.100350

Jinhu Bian , Siyuan Li , Zhengjian Zhang , Yi Deng , Guangbin Lei , Xi Nan , Amin Naboureh , Ainong Li

Mountain ecosystems, characterized by high heterogeneity and rapid dynamics, require high spatiotemporal resolution remote sensing products for accurate monitoring. However, rugged terrain induces significant radiometric distortion in satellite imagery, and persistent cloud cover leads to data gaps, severely limiting the application of optical satellites in these regions. This study developed an integrated method to generate a daily, seamless, 10-m resolution, Sentinel-2-like surface reflectance data cube for the topographically complex Wanglang mountain area in China. Our method synergistically combined a physically-based topographic correction model that incorporates atmospheric and bidirectional reflectance distribution function (BRDF) effects, a harmonic model for temporal reconstruction and gap-filling, and a validation and calibration framework using a tower-based multi-angle in-situ surface reflectance observation. The results demonstrate that the physically-based topographic correction model outperformed the operational L2A product of Sentinel-2, reducing the RMSE in the near-infrared (NIR) band from 0.025–0.035 to 0.0106–0.0206 and effectively eliminating the dependence on solar incidence angle (reducing R² to as low as 0.007 during peak season). The harmonic model accurately reconstructed seamless daily data, achieving well prediction accuracy (R² of 0.83–0.87 for NIR band and 0.81–0.89 for blue band) across different phenological stages. Following linear calibration, the final data cube achieved exceptional radiometric agreement with independent in-situ measurements, with R² > 0.60 for all visible and NIR bands and RMSE as low as 0.0081–0.0171. This study provides not only a high-fidelity and seamless surface reflectance product in Wanglang complex terrains for the research community but also a replicable framework for generating reference true-value products in the challenging mountain area. The resulting surface reflectance data cube serves as a critical foundation for biophysical parameters estimation, thereby enhancing our ability to validate and develop mountain remote sensing algorithms and products, and further understanding vulnerable mountain ecosystems as well.

山地生态系统具有高度异质性和快速动态的特点，需要高时空分辨率的遥感产品进行精确监测。然而，崎岖的地形导致卫星图像出现明显的辐射失真，持续的云层覆盖导致数据空白，严重限制了光学卫星在这些地区的应用。本研究开发了一种集成方法，为中国地形复杂的王朗山区生成每日、无缝、10米分辨率、类似sentinel -2的地表反射率数据立方体。该方法将包含大气和双向反射率分布函数（BRDF）效应的基于物理的地形校正模型、用于时间重建和间隙填充的调和模型以及基于塔的多角度原位地表反射率观测的验证和校准框架协同结合起来。结果表明，基于物理的地形校正模型优于Sentinel-2的实际L2A产品，将近红外（NIR）波段的RMSE从0.025-0.035降低到0.0106-0.0206，有效消除了对太阳入射角的依赖（旺季时R2低至0.007）。调和模型准确地重建了无缝的日数据，在不同物候阶段取得了较好的预测精度（近红外波段R2为0.83-0.87，蓝波段R2为0.81-0.89）。在线性校准之后，最终的数据立方体与独立的原位测量结果达到了非常好的辐射一致性，所有可见光和近红外波段的R2 >； 0.60， RMSE低至0.0081-0.0171。本研究不仅为研究界提供了王朗复杂地形下的高保真无缝地表反射率产品，也为在具有挑战性的山区生成参考真值产品提供了可复制的框架。由此产生的地表反射率数据立方体可以作为生物物理参数估计的重要基础，从而提高我们验证和开发山地遥感算法和产品的能力，并进一步了解脆弱的山地生态系统。

{"title":"Development of a 10 m daily seamless surface reflectance data cube based on Sentinel-2 constellation for generating the reference true-value products at Wanglang mountain area, China","authors":"Jinhu Bian , Siyuan Li , Zhengjian Zhang , Yi Deng , Guangbin Lei , Xi Nan , Amin Naboureh , Ainong Li","doi":"10.1016/j.srs.2025.100350","DOIUrl":"10.1016/j.srs.2025.100350","url":null,"abstract":"<div><div>Mountain ecosystems, characterized by high heterogeneity and rapid dynamics, require high spatiotemporal resolution remote sensing products for accurate monitoring. However, rugged terrain induces significant radiometric distortion in satellite imagery, and persistent cloud cover leads to data gaps, severely limiting the application of optical satellites in these regions. This study developed an integrated method to generate a daily, seamless, 10-m resolution, Sentinel-2-like surface reflectance data cube for the topographically complex Wanglang mountain area in China. Our method synergistically combined a physically-based topographic correction model that incorporates atmospheric and bidirectional reflectance distribution function (BRDF) effects, a harmonic model for temporal reconstruction and gap-filling, and a validation and calibration framework using a tower-based multi-angle in-situ surface reflectance observation. The results demonstrate that the physically-based topographic correction model outperformed the operational L2A product of Sentinel-2, reducing the RMSE in the near-infrared (NIR) band from 0.025–0.035 to 0.0106–0.0206 and effectively eliminating the dependence on solar incidence angle (reducing R<sup>2</sup> to as low as 0.007 during peak season). The harmonic model accurately reconstructed seamless daily data, achieving well prediction accuracy (R<sup>2</sup> of 0.83–0.87 for NIR band and 0.81–0.89 for blue band) across different phenological stages. Following linear calibration, the final data cube achieved exceptional radiometric agreement with independent in-situ measurements, with R<sup>2</sup> > 0.60 for all visible and NIR bands and RMSE as low as 0.0081–0.0171. This study provides not only a high-fidelity and seamless surface reflectance product in Wanglang complex terrains for the research community but also a replicable framework for generating reference true-value products in the challenging mountain area. The resulting surface reflectance data cube serves as a critical foundation for biophysical parameters estimation, thereby enhancing our ability to validate and develop mountain remote sensing algorithms and products, and further understanding vulnerable mountain ecosystems as well.</div></div>","PeriodicalId":101147,"journal":{"name":"Science of Remote Sensing","volume":"13 ","pages":"Article 100350"},"PeriodicalIF":5.2,"publicationDate":"2025-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145748182","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Weed classification in sugarcane fields in Northeast Thailand from multi-temporal Sentinel-1 and Sentinel-2 data together with random forest algorithm 基于Sentinel-1和Sentinel-2数据的泰国东北部甘蔗田杂草分类研究

IF 5.2 Q1 ENVIRONMENTAL SCIENCES

Science of Remote Sensing

Pub Date : 2025-12-09 DOI: 10.1016/j.srs.2025.100352

Savittri Ratanopad Suwanlee , Muhammad Hanif , Kemin Kasa , Surasak Keawsomsee , Jaturong Som-ard , Vorraveerukorn Veerachitt , Phattamon Heawchaiyaphum , Akkawat Puntura , Mohammad D. Hossain , Sarawut Ninsawat

Timely and accurate weed detection is essential for sustainable crop production and management. The integration of multiple satellite data sources with powerful machine learning has transformed precision agriculture by enhancing the accuracy and automation of object classification, enabling large-scale analysis and real-time predictions. However, challenges remain in effectively managing agricultural practices, particularly in weed control. This study employed Sentinel-1 (S1) and Sentinel-2 (S2) satellite data, combined with vegetation indices and random forest (RF) classification algorithm, to map weed presence in sugarcane fields in Northeastern Thailand. The large number of reference data consisting of 744 points was utilized to train and validate weed identification. The combined S1 and S2 dataset significantly enhanced the detection capabilities of the best RF model, achieving an overall classification result of 96 % accuracy and F1 scores exceeding 93 %. While overall weed levels were low, several high-density zones were clearly detected, underscoring the importance of targeted weed management. The combination of S1 and S2 data improved classification performance, addressing challenges posed by mixed pixels in small fields. Stratifying weed density provided deeper insights into field variability over the large scale. Our work presents a scientifically robust and operationally scalable framework for monitoring weed infestations in sugarcane cultivation. The proposed approach demonstrates strong potential for advancing sustainable precision agriculture by facilitating timely and spatially precise interventions.

及时准确的杂草检测对作物的可持续生产和管理至关重要。将多个卫星数据源与强大的机器学习相结合，通过提高目标分类的准确性和自动化，实现大规模分析和实时预测，改变了精准农业。然而，在有效管理农业实践方面，特别是在杂草控制方面，仍然存在挑战。本研究利用Sentinel-1 （S1）和Sentinel-2 （S2）卫星数据，结合植被指数和随机森林（RF）分类算法，绘制了泰国东北部甘蔗田杂草分布图。利用744个点的大量参考数据对杂草识别进行训练和验证。结合S1和S2数据集显著增强了最佳RF模型的检测能力，总体分类结果准确率达到96%，F1得分超过93%。虽然总体杂草水平较低，但清楚地发现了几个高密度区，强调了有针对性的杂草管理的重要性。S1和S2数据的结合提高了分类性能，解决了小域内混合像素带来的挑战。对杂草密度进行分层可以更深入地了解大规模的田间变异性。我们的工作提出了一个科学可靠和可扩展的框架，用于监测甘蔗种植中的杂草侵害。所提出的方法通过促进及时和空间上精确的干预，显示了推进可持续精准农业的巨大潜力。

{"title":"Weed classification in sugarcane fields in Northeast Thailand from multi-temporal Sentinel-1 and Sentinel-2 data together with random forest algorithm","authors":"Savittri Ratanopad Suwanlee , Muhammad Hanif , Kemin Kasa , Surasak Keawsomsee , Jaturong Som-ard , Vorraveerukorn Veerachitt , Phattamon Heawchaiyaphum , Akkawat Puntura , Mohammad D. Hossain , Sarawut Ninsawat","doi":"10.1016/j.srs.2025.100352","DOIUrl":"10.1016/j.srs.2025.100352","url":null,"abstract":"<div><div>Timely and accurate weed detection is essential for sustainable crop production and management. The integration of multiple satellite data sources with powerful machine learning has transformed precision agriculture by enhancing the accuracy and automation of object classification, enabling large-scale analysis and real-time predictions. However, challenges remain in effectively managing agricultural practices, particularly in weed control. This study employed Sentinel-1 (S1) and Sentinel-2 (S2) satellite data, combined with vegetation indices and random forest (RF) classification algorithm, to map weed presence in sugarcane fields in Northeastern Thailand. The large number of reference data consisting of 744 points was utilized to train and validate weed identification. The combined S1 and S2 dataset significantly enhanced the detection capabilities of the best RF model, achieving an overall classification result of 96 % accuracy and F1 scores exceeding 93 %. While overall weed levels were low, several high-density zones were clearly detected, underscoring the importance of targeted weed management. The combination of S1 and S2 data improved classification performance, addressing challenges posed by mixed pixels in small fields. Stratifying weed density provided deeper insights into field variability over the large scale. Our work presents a scientifically robust and operationally scalable framework for monitoring weed infestations in sugarcane cultivation. The proposed approach demonstrates strong potential for advancing sustainable precision agriculture by facilitating timely and spatially precise interventions.</div></div>","PeriodicalId":101147,"journal":{"name":"Science of Remote Sensing","volume":"13 ","pages":"Article 100352"},"PeriodicalIF":5.2,"publicationDate":"2025-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145748181","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0