Remote Sensing Applications-Society and Environment最新文献

{"title":"A Two-stage Registration Method for UAV and HMLS Point Clouds in Subtropical Forest","authors":"Zeyu Yang ,&nbsp;Zhiqiang Guo ,&nbsp;Ziyan Zhang ,&nbsp;Xiaozi Zhou ,&nbsp;Yuanyong Dian","doi":"10.1016/j.rsase.2025.101709","DOIUrl":"10.1016/j.rsase.2025.101709","url":null,"abstract":"<div><div>HMLS (Handheld Mobile Laser Scanning) and UAV (Unmanned Aerial Vehicle) LiDAR are increasingly utilized in forest inventory due to their efficiency and portability. However, challenges such as occlusions, low vertical overlap, and varying point cloud density complicate the fusion of these two datasets. In this study, we propose a novel two-stage method to match HMLS and UAV LiDAR data with different point density at complicate forest with dense canopy cover. The first stage optimizes voxel size selection for varying cloud densities and performs feature extraction. The second stage addresses gross error elimination through the truncated least squares method and performs feature matching using K-D Tree nearest neighbor indexing in combination with Singular Value Decomposition (SVD). The method was tested in 27 forest plots with varying vertical overlaps and stand conditions across Hubei Province, China and compared with four registration methods: Coarse-to-Global Adjustment Strategy (CGAS), Optimized Coarse-to-Fine Algorithms (OCFA), Generalized-ICP (GICP), and Bidirectional-Pearson Improved Method (BPIM). Results show that the proposed approach significantly improves registration accuracy, with error reductions of up to 0.096 m, 0.284 m, and 0.425 m under lower (0.37–0.56), moderate (0.58–0.73), and higher (0.77–0.95) canopy cover, respectively. Stand conditions and tree species influence registration accuracy. The results demonstrate higher accuracy in plots with lower canopy cover, steeper slopes, and fewer shrubs. Coniferous forests, with straighter trunks and fewer branches, provide more distinct feature points, leading to better accuracy than broadleaf forests. Additionally, UAV and HMLS matching accuracy is influenced by flight altitude, with higher altitudes increasing registration errors due to the decreased point density of UAV LiDAR.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"39 ","pages":"Article 101709"},"PeriodicalIF":4.5,"publicationDate":"2025-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145004159","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}

{"title":"Evaluation of MODIS NDVI product in a heterogeneous urban environment using five upscaling methods and Landsat 8 product","authors":"Haijun Luan ,&nbsp;Zhenhong Lin ,&nbsp;Chenshuo Xing ,&nbsp;Lanhui Wang ,&nbsp;Jian Deng ,&nbsp;Shangharsha Thapa ,&nbsp;Jiajin Zhang ,&nbsp;Weibin Wang ,&nbsp;Hongyi Yao ,&nbsp;Zheng Duan","doi":"10.1016/j.rsase.2025.101687","DOIUrl":"10.1016/j.rsase.2025.101687","url":null,"abstract":"<div><div>In order to accurately assess the quality of low-resolution biogeophysical parameter products, accurate scale transformations are essential. However, different scaling models often lead to inconsistent transformation results. More worryingly, many biogeophysical parameters are not scale-invariant, such as the Normalized Difference Vegetation Index (NDVI), which makes the quality assessment of low-resolution products even more challenging. Therefore, we propose an integrated approach that utilizes multiple upscaling methods and high-quality, moderate-resolution surface reflectance products to evaluate the quality of low-resolution MODIS NDVI products, eliminating the need for extensive <em>in-situ</em> observation data. In this study, the full-scale transformation of Landsat 8 OLI NDVI in heterogeneous urban environments is realized using five upscaling methods, including two reflectance-level Taylor series expansion (TSE) models, the simple averaging method, the Chen NDVI model, and the point spread function (PSF) method. Finally, the overall quality of the MOD13Q1 product in the study area was evaluated based on the upscaled NDVI images. Our study provides quantitative insights into the underlying causes of scale effects in NDVI, including the spatial heterogeneity of the surface and the nonlinearity of the NDVI model. Furthermore, the TSE method, which integrates land cover types, and the PSF method were first practically applied to the study of upscaling NDVI. The integration of land cover types in the TSE method and the consideration of specific weights for “small pixels” in the PSF method offer nuanced insights. Our findings affirm the overall high quality of the MOD13Q1 product and the overall bias between the MOD13Q1 images and the corresponding upscaled NDVI images for the entire study area, Xiamen city, which ranged from 0.0176 to 0.0225 in absolute value (mean difference) and from 0 to 0.0071 in absolute value (standard deviation difference). This study advances our understanding of NDVI scale effects and demonstrates that the proposed method serves as an efficient and effective way to evaluate the overall quality of low-resolution constructed biogeophysical parameters that lack scale-invariant characteristics in expansive areas with insufficient <em>in-situ</em> observation data.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"39 ","pages":"Article 101687"},"PeriodicalIF":4.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144865725","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}

{"title":"Biodiversity monitoring in urban community gardens using proximal sensing and drone remote sensing","authors":"Yasamin Afrasiabian ,&nbsp;Felix Contiz ,&nbsp;Elisa Van Cleemput ,&nbsp;Monika Egerer ,&nbsp;Kang Yu","doi":"10.1016/j.rsase.2025.101685","DOIUrl":"10.1016/j.rsase.2025.101685","url":null,"abstract":"<div><div>In urban community gardens, artificially managed ground cover types, including vegetative and non-vegetative ground components, are both critical to ecological functioning. Yet, how these non-vegetative components influence spectral diversity in ways that are different from natural systems has not been addressed. This study investigated the potential of combining spectral and structural diversity variables, corresponding to the Spectral Variation and Height Variation Hypotheses, respectively, to monitor plant and ground cover diversity. These variables were derived from in situ hyperspectral measurements, drone-based multispectral imagery, and three-dimensional canopy height models. We examined four biodiversity variables, including plant species richness, total plant abundances, ground cover entropy, and ground cover richness, across five urban community gardens over two years. Spectral diversity was calculated based on the Coefficient of Variation (CV), Spectral Angle Mapper (SAM), and Shannon's Entropy (Entropy) indices at multiple spectral ranges. Structural diversity variables, including canopy height variation and image texture features. Our results showed that Red-Edge and Near-infrared (NIR) bands effectively captured compositional variation in ground cover, while visible wavelengths better reflected subtle differences in vegetative components. Texture features and height-based structural variables provided valuable insights into canopy complexity, particularly improving predictions of plant abundance and ground cover entropy. Finally, we found that integrating spectral and structural diversity variables further enhanced predictive performance due to considering canopy biochemical and structural differences. This multi-metric approach outperformed single-source analyses, underscoring the value of combining complementary remote sensing data for better interpreting urban garden biodiversity. Our findings highlight the importance of characterizing canopy structural heterogeneity in advancing biodiversity monitoring within these complex urban ecosystems.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"39 ","pages":"Article 101685"},"PeriodicalIF":4.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144852932","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}

{"title":"Monitoring mangrove responses to seasonal changes, hurricane-induced disturbance, and recovery in the South Florida Everglades: A spatio-temporal analysis of decade-long (2013–2023) Landsat-8 observations","authors":"Selena Chavez ,&nbsp;Shimon Wdowinski ,&nbsp;David Lagomasino ,&nbsp;Edward Castañeda-Moya","doi":"10.1016/j.rsase.2025.101688","DOIUrl":"10.1016/j.rsase.2025.101688","url":null,"abstract":"<div><div>Mangrove forests play a critical role in coastal ecosystems by buffering shorelines against the destructive forces of storms and storm surges, but in doing so, they often endure significant structural damage, including defoliation, tree snapping, and branch loss. Using decade-long remote sensing Landsat 8 data, we calculated the Normalized Difference Vegetation Index (NDVI) and Normalized Difference Moisture Index (NDMI) to assess patterns and trends within the decade-long time series for each index in mangrove forests of southwestern Florida Everglades. Before calculating NDVI and NDMI, we cloud-filtered and calculated the monthly spectral means of the study region from March 2013 to March 2023. Using both NDVI and NDMI, we found seasonal variations in the value of both indices, in which the value increased during the wet season and decreased during the dry season of the Everglades. We also detected the impact of Hurricane Irma on mangroves in 2017 due to a sudden drop in the indices’ values right after the storm. The time series showed a slow recovery of indices values compared to pre-storm values. Using an exponential recovery model, we calculated that most mangrove areas recovered within two to four years. However, some small mangrove areas show no recovery, which we attribute to saltwater ponding and areas without freshwater flow and hydrological connectivity.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"39 ","pages":"Article 101688"},"PeriodicalIF":4.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144858557","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}

{"title":"Estimating crop evapotranspiration using drone imagery, ground canopy temperature, and machine learning techniques","authors":"Fitsum T. Teshome ,&nbsp;Haimanote K. Bayabil ,&nbsp;Bruce Schaffer ,&nbsp;Yiannis Ampatzidis","doi":"10.1016/j.rsase.2025.101661","DOIUrl":"10.1016/j.rsase.2025.101661","url":null,"abstract":"<div><div>Efficient irrigation management relies on accurately estimating crop evapotranspiration (ETc), yet conventional methods often face limitations, such as cost, spatial coverage, data requirements, and the need for local calibration. This study had two main objectives: 1) to quantify daily ETc of sweet corn (SC) and green beans (GB) using crop water stress index (CWSI) calculated from canopy temperatures (Tc) and crop coefficient (Kc) estimated vegetation indices and 2) to evaluate the potential of machine learning (ML) models in estimating daily ETc and Tc. Irrigation experiments were conducted during the winter seasons of 2020–2021 and 2021–2022 at the University of Florida's Tropical Research and Education Center (TREC), Florida, USA. Networks of above-canopy infrared thermocouples (IRTs) and soil moisture (SM) sensors were used to collect Tc and SM. Multispectral images were also collected using an unmanned aerial vehicle (UAV)-based RedEdge-MX sensor. Sub-hourly changes in SM during dry periods were aggregated to estimate the daily measured ETc of SC and GB. Time series of CWSI were generated from Tc, while Kc was estimated using eleven vegetation indices (VIs) generated from drone imagery. Moreover, four ML models, i.e., CatBoost (CB), Random Forest (RF), k-Nearest Neighbors (kNN), and Extreme Gradient Boosting (XGB), were evaluated for simulating ETc. Six models, i.e., CB, kNN, RF, XGB, Light Gradient Boosting Machine (LGB), and Deep Learning (DL), were also evaluated for simulating the Tc of SC and GB. The results showed that the CWSI approach was acceptable in estimating ETc with an average MAE of 0.90 mm day^-1 for SC and 0.62 mm day^-1 for GB. Four out of eleven vegetation indices (VIs) demonstrated superior performance in estimating daily ETc, including the Soil Adjusted Vegetation Index (SAVI), Normalized Green-Red Difference Index (NGRDI), Red Edge Normalized Difference Vegetation Index (RENDVI), and NIR-RE normalized difference vegetation index (NIRRENDVI). The ML models captured ETc and Tc with better accuracy. Averaged root mean square error (RMSE) for ETc across the four models was 0.86 mm day^-1 for SC and 0.89 mm day^-1 for GB. The average RMSE of the ML models for simulating Tc was ±1.1 °C for SC and ±1.5 ^°C for GB. Overall, CWSI, spectral reflectance-based Kc, and ML models proved to be useful tools for estimating ETc at finer spatial and temporal scales with reasonable accuracy.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"39 ","pages":"Article 101661"},"PeriodicalIF":4.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144841964","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}

{"title":"Enhancing mountain grassland mapping: A comparative study with PRISMA hyperspectral, multispectral, and SAR data","authors":"E. Patriarca,&nbsp;L. Stendardi,&nbsp;E. Dorigatti,&nbsp;R. Sonnenschein,&nbsp;B. Ventura,&nbsp;M. Claus,&nbsp;M. Castelli,&nbsp;B. Tufail,&nbsp;C. Notarnicola","doi":"10.1016/j.rsase.2025.101666","DOIUrl":"10.1016/j.rsase.2025.101666","url":null,"abstract":"<div><div>Mountain grasslands are increasingly threatened by climate change, land abandonment, and overexploitation. Remote sensing is a valuable tool for monitoring these changes through vegetation mapping. However, challenges such as frequent cloud cover, short growing seasons, and limited field data can reduce the accuracy of results. In this study, we evaluated the effectiveness of different remote sensing data for classifying mountain grasslands in the Sciliar-Catinaccio Nature Park, Italy. We compared classification results using a hyperspectral PRISMA image (Sept 29, 2023), multispectral data from a single-date Sentinel-2 image (Sept 25, 2023), and Spectral-Temporal Metrics (STM) derived from a Sentinel-2 time series from 2021 to 2023. Additionally, we assessed the impact on accuracy of combining optical datasets with Synthetic Aperture Radar (SAR) data, including a time series of 2023 Sentinel-1 backscatter and coherence metrics. Using the Recursive Feature Elimination algorithm (<em>RFE</em>), we selected the most relevant features for classification and applied both Random Forest (RF) and Support Vector Machines (SVM) classifiers. SVM outperformed RF, performing better with the limited training data available. SAR data did not significantly enhance classification and was therefore excluded by the RFE algorithm. PRISMA-based classification achieved up to 74 % accuracy, while single-date Sentinel-2 imagery reached 52 %. The use of STM improved classification performance, yielding an overall accuracy of 77 %. The highest accuracy (87 %) was achieved by combining PRISMA and STM features. These findings suggest that while individual optical datasets may not provide optimal classification accuracy, integrating data from multiple optical sensors significantly enhances the mapping of mountain grasslands.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"39 ","pages":"Article 101666"},"PeriodicalIF":4.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144893402","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}

{"title":"Integrating SMAP and CYGNSS data for daily soil moisture and agricultural drought monitoring in Nghe An province, Vietnam","authors":"Tich Phuc Hoang ,&nbsp;Minh Cuong Ha ,&nbsp;Phuong Lan Vu ,&nbsp;José Darrozes ,&nbsp;Phuong Bac Nguyen","doi":"10.1016/j.rsase.2025.101664","DOIUrl":"10.1016/j.rsase.2025.101664","url":null,"abstract":"<div><div>In the context of climate change, droughts are increasingly frequent and severe, affecting broader regions. Consequently, effective drought monitoring is crucial for risk management and understanding climate change impacts. Soil moisture estimation using satellite data is a pivotal metric for developing time-series agricultural drought monitoring maps. This study proposes a methodology for constructing soil moisture and agricultural drought maps for Nghe An Province, Vietnam, using the SMAP dataset along with soil moisture estimations from CYGNSS data and additional ancillary data. The Self-Attention-based Imputation for Time Series (SAITS) model, employing self-attention mechanisms to impute missing values in multivariate time series, is used to construct the soil moisture dataset from SMAP, resulting in complete datasets with a training loss RMSE_SAITS <span><math><mo>=</mo></math></span> 0.073 <span><math><mrow><msup><mrow><mi>cm</mi></mrow><mrow><mn>3</mn></mrow></msup><mo>/</mo><msup><mrow><mi>cm</mi></mrow><mrow><mn>3</mn></mrow></msup></mrow></math></span>. Additionally, leveraging a Random Forest Regression model, CYGNSS data combined with meteorological, topographic, and soil texture information enable the estimation of daily soil moisture values, exhibiting a strong correlation with R <span><math><mo>=</mo></math></span> 0.889. Subsequently, integration of the two soil moisture products from SMAP and CYGNSS yields a dataset with a spatial resolution of 1km and a temporal resolution of 1 day. The soil moisture results were compared with moisture data from ERA5 (R <span><math><mo>=</mo></math></span> 0.75, ubRMSE <span><math><mo>=</mo></math></span> 0.055 <span><math><mrow><msup><mrow><mi>cm</mi></mrow><mrow><mn>3</mn></mrow></msup><mo>/</mo><msup><mrow><mi>cm</mi></mrow><mrow><mn>3</mn></mrow></msup></mrow></math></span>) and in-situ data in Nghe An province (R <span><math><mo>=</mo></math></span> 0.709, ubRMSE <span><math><mo>=</mo></math></span> 0.017 <span><math><mrow><msup><mrow><mi>cm</mi></mrow><mrow><mn>3</mn></mrow></msup><mo>/</mo><msup><mrow><mi>cm</mi></mrow><mrow><mn>3</mn></mrow></msup></mrow></math></span>). Finally, the Standardized Soil Moisture Index is calculated to transform the time-series soil moisture data into a standardized normal distribution, generating agricultural drought maps with 9 different levels. This study represents a significant advancement in agricultural drought monitoring, highlighting the immense potential of machine learning techniques when combined with satellite-based soil moisture data. Our approach effectively monitors drought in Nghe An Province, Vietnam, with broader applicability to other regions worldwide.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"39 ","pages":"Article 101664"},"PeriodicalIF":4.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144757515","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}

{"title":"Simulating vegetation potential and quantifying uncertainty for precision forestation in arid regions","authors":"Jia Qu ,&nbsp;Zirui Gai ,&nbsp;Qi Liu ,&nbsp;Dongwei Gui ,&nbsp;Xinlong Feng ,&nbsp;Jianping Zhao ,&nbsp;Tao Lin ,&nbsp;Yunfei Liu ,&nbsp;Qian Jin ,&nbsp;Zeeshan Ahmed","doi":"10.1016/j.rsase.2025.101670","DOIUrl":"10.1016/j.rsase.2025.101670","url":null,"abstract":"<div><div>Large-scale forestation in arid regions with excessive planting density often aggravates water scarcity and disrupts local ecosystems. The Potential Normalized Difference Vegetation Index (PNDVI) reflects the optimal density of natural vegetation in the absence of human intervention, and can guide the planting site, area and density in arid areas. However, its accurate simulation with uncertainty quantification remains understudied. We propose a method to quantify uncertainty in PNDVI prediction by integrating deep learning, variational inference, and multiple environmental variables to fill this gap. The model was applied to the lower Tarim River Basin (LTRB) in northwest China and achieved the best performance with an average accuracy of 88.58 %, which is 10.09 % higher than conventional machine learning models. The overall uncertainty is characterized by a mean value of 0.298, with a standard deviation of 0.142. In the LTRB, regions near the river channel in the central and southeastern areas with low uncertainties are ideal for high-density forestation. This approach can offer scientific decision-support for arid-region forestation planning and has great socio-economic benefits by reducing water consumption, increasing land productivity and reducing ecological restoration costs.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"39 ","pages":"Article 101670"},"PeriodicalIF":4.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144750406","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}

{"title":"Assessing grapevine water status and rootstock effects using vegetation indices from UAV and proximal sensors","authors":"Anderson de Jesus Pereira ,&nbsp;Larissa Godarelli Farinassi ,&nbsp;Bruno Ricardo Silva Costa ,&nbsp;Israel de Oliveira Junior ,&nbsp;Robson Argolo dos Santos ,&nbsp;Lucio André de Castro Jorge ,&nbsp;Luís Henrique Bassoi","doi":"10.1016/j.rsase.2025.101708","DOIUrl":"10.1016/j.rsase.2025.101708","url":null,"abstract":"<div><div>Unmanned aerial vehicle (UAV)-based sensing and proximal sensing are effective techniques for monitoring grapevine canopies through vegetation indices (VI), enabling the assessment of plant growth-related factors. This study aimed to evaluate whether VIs derived from both sensing platforms correlate with leaf grapevine and soil water status. The similarity between the two platforms in estimating VI and their effectiveness in detecting variations in plant growth associated with different rootstocks were examined. Grapevines cv. Syrah grafted onto IAC 572 and Paulsen 1103 rootstocks were monitored across a 1.1 ha vineyard in Southeastern Brazil. UAV-based sensing involved capturing images with a multispectral sensor mounted on an UAV, while proximal sensing was conducted by measuring canopy reflectance with an active sensor. The Normalized Difference Red Edge (NDRE) derived from UAV-based sensing showed the strongest correlation with stomatal conductance <em>g</em>_<em>s</em> (r = 0.84, p &lt; 0.001, pseudo-R² = 0.73) and relative leaf water content RLWC (r = 0.73, p &lt; 0.001, R² = 0.60). No significant correlation was observed between the soil water parameter θ at 0–0.2 m and VI estimates from either platform (p &gt; 0.05). Significant correlations (p &lt; 0.001) were found between NDRE (r = 0.70, pseudo-R² = 0,58) and Normalized Difference Vegetation Index (NDVI) (r = 0.72, pseudo-R² = 0,54), derived from both sensing methods. Both indices also detected differences in grapevine vigor related to rootstock influence. UAV-based measures provided stronger correlations with these traits than proximal measurements.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"39 ","pages":"Article 101708"},"PeriodicalIF":4.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145007664","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}

{"title":"Assessment and projection of carbon density in China: An integrated approach combining Boruta-SHAP-machine learning and structural equation modeling","authors":"Ya Wen,&nbsp;Xue Han,&nbsp;Zizhao Ma,&nbsp;Ruirui Zhang","doi":"10.1016/j.rsase.2025.101694","DOIUrl":"10.1016/j.rsase.2025.101694","url":null,"abstract":"<div><div>Carbon density is a key indicator of the carbon sequestration capacity of terrestrial ecosystem. As one of the world's largest carbon emitters, gaining a clear understanding of China's carbon density is essential for achieving the strategic goals of “carbon peak” and “carbon neutrality.” Therefore, this study used machine learning to estimate the current carbon density of vegetation in China. However, machine learning often suffer from the “black box” problem. To address this, the Boruta-SHAP variable selection method was employed to enhance the interpretability of feature importance. Based on the current estimation of China's vegetation carbon density, this study used a structural equation model to explore the driving factors influencing carbon density. The model further predicts China's vegetation carbon density for 2030 and 2060 based on the identified key driving factors. The results of the study indicate the following: (1) Boruta-SHAP effectively visualizes the relative importance of feature variables. Among various vegetation indices, the Leaf Area Index (LAI) is the most important variable for carbon density modeling, with a mean SHAP value of 4.86. (2) In estimating China's carbon density, Boruta-SHAP-Random Forest (R² = 0.636) and Boruta-SHAP-XGBoost (R² = 0.629) outperform Multiple Linear Regression (R² = 0.559). (3) The average vegetation carbon density in China is 48.27 Mg/ha. The Moderately Low Carbon Density Zone accounts for the largest proportion of the total area, primarily located in the northeast region (e.g., Jilin and Heilongjiang), the southern fringe, and the eastern hilly areas. (4) Climate is identified as the main driving factor influencing China's vegetation carbon density, with a path coefficient of 0.72. Among climate variables, maximum temperature exerts the strongest influence, with a path coefficient of 0.85. (5) The average vegetation carbon density in China is projected to be 52.34 Mg C/ha in 2030 and 53.08 Mg C/ha in 2060. The findings of this study are expected to provide theoretical support for optimizing forest management policies and addressing climate change risks, including carbon sink fluctuations caused by extreme events.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"39 ","pages":"Article 101694"},"PeriodicalIF":4.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144865726","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}