Remote Sensing Applications-Society and Environment最新文献

{"title":"Mapping spatio-temporal of ground-level ozone in Jakarta, a tropical capital city: A machine learning approach with multi-source satellite data","authors":"Balqis Meiliana ,&nbsp;Muhammad Hilal Arrizqon ,&nbsp;Parwati Sofan ,&nbsp;Firman Hadi","doi":"10.1016/j.rsase.2025.101793","DOIUrl":"10.1016/j.rsase.2025.101793","url":null,"abstract":"<div><div>Ground-level ozone (GLO) is a harmful air pollutant with significant impacts on human health and the environment. In Indonesia, GLO monitoring is limited, particularly in the densely populated capital, Jakarta, which has only five ground monitoring stations, underscoring the need for alternative approaches. This study aimed to map the spatial-temporal distribution of GLO concentrations in Jakarta from 2022 to 2024 using satellite data and machine learning. We integrated atmospheric, biophysical, and anthropogenic variables into three models: Linear Regression, Random Forest, and Light Gradient Boosting Machine (LightGBM). LightGBM achieved the highest predictive accuracy (R² = 0.73) when spatial geolocation was included. In this setting, SO₂, the north-south wind component (V10), and Nighttime Light (NTL) emerged as the third most influential predictors. Spatial analysis revealed higher GLO concentrations in industrial and densely built-up areas, especially in North and West Jakarta. Seasonal trends showed peaks during the dry season (74.33 μg/m³) and significant declines in the rainy season (10.16 μg/m³), driven by solar radiation and atmospheric stability. The highest GLO levels were observed in 2023, coinciding with El Niño-related warming. Local Climate Zone (LCZ) analysis further indicated that built-up areas had higher GLO concentrations compared to vegetated zones. This study demonstrates the potential of combining remote sensing and machine learning to estimate GLO in tropical megacities with limited monitoring infrastructure. The findings can support data-driven urban planning and policies aimed at reducing ozone pollution and promoting green urban development.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"40 ","pages":"Article 101793"},"PeriodicalIF":4.5,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145528802","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":"Machine learning for mapping glacier surface facies in Svalbard","authors":"Sagar F. Wankhede ,&nbsp;Shridhar D. Jawak ,&nbsp;Adeeb H. Noorudheen ,&nbsp;Akankshya Nayak ,&nbsp;Abhilash Thakur ,&nbsp;Keshava Balakrishna ,&nbsp;Alvarinho J. Luis","doi":"10.1016/j.rsase.2025.101753","DOIUrl":"10.1016/j.rsase.2025.101753","url":null,"abstract":"<div><div>Glaciers are dynamic and highly sensitive indicators of climate change, necessitating frequent and precise monitoring. As Earth observation technology evolves with advanced sensors and mapping methods, the need for accurate and efficient approaches to monitor glacier changes becomes increasingly important. Glacier Surface Facies (GSF), formed through snow accumulation and ablation, serve as valuable indicators of glacial health. Mapping GSF provides insights into a glacier's annual adaptations. However, satellite-based GSF mapping presents significant challenges in terms of data preprocessing and algorithm selection for accurate feature extraction. This study presents an experiment using very high-resolution (VHR) WorldView-3 satellite data to map GSF on the Midtre Lovénbreen glacier in Svalbard. We applied three machine learning (ML) algorithms—Random Forest (RF), Artificial Neural Network (ANN), and Support Vector Machine (SVM)—to explore the impact of different image preprocessing techniques, including atmospheric corrections, pan sharpening methods, and spectral band combinations. Our results demonstrate that RF outperformed both ANN and SVM, achieving an overall accuracy of 85.02 %. However, nuanced variations were found for specific processing conditions and can be explored for specific applications. This study represents the first clear delineation of ML algorithm performance for GSF mapping under varying preprocessing conditions. The data and findings from this experiment will inform future ML-based studies aimed at understanding glaciological adaptations in a rapidly changing cryosphere, with potential applications in long-term spatiotemporal monitoring of glacier health.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"40 ","pages":"Article 101753"},"PeriodicalIF":4.5,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145747673","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":"Detecting seasonal snow transitions in SAR time series with Horizontal Visibility Graphs","authors":"Giuliana Beltramone ,&nbsp;Alejandro C. Frery ,&nbsp;Marcelo C. Scavuzzo ,&nbsp;Matias Bonansea ,&nbsp;Anabella Ferral","doi":"10.1016/j.rsase.2025.101772","DOIUrl":"10.1016/j.rsase.2025.101772","url":null,"abstract":"<div><div>Seasonal snow studies in the Andes face major challenges due to limited ground access and the region’s complex mountain topography. Despite the crucial role of snow in hydrological and climatic processes, research on seasonal snow dynamics in Latin America remains scarce, particularly regarding the use of advanced remote sensing techniques. Synthetic Aperture Radar (SAR) image time series offer a powerful tool to detect the onset of snowmelt processes in such data-scarce environments. However, the potential of SAR to discriminate between bare soil and fresh snow remains largely unexplored, representing an important gap in current remote sensing methodologies. In this study, we analyzed SAR time series via Horizontal Visibility Graphs (HVG) with modularity-based community detection. The proposed method detected a previously unreported bare soil transition (summer-autumn) that conventional threshold-based techniques missed, as shown by SAVI/NMDI validation. These findings demonstrate that community extraction from HVG provides a robust and insightful framework for analyzing temporal snow dynamics. This approach is critical for water resource management in the Patagonian Andes, where ground monitoring is limited, and snowmelt timing directly impacts regional runoff, hydropower generation, and flood risk mitigation.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"40 ","pages":"Article 101772"},"PeriodicalIF":4.5,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145578729","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":"Revealing the association mechanisms between PM2.5 and O3 pollutants and ecosystem services in Beijing-Tianjin-Hebei and surrounding areas","authors":"Wenxing Hou,&nbsp;Zhuowei Hu,&nbsp;Mi Wang,&nbsp;Yongcai Wang,&nbsp;Xiangping Liu,&nbsp;Siyuan Li,&nbsp;Junjie Wang,&nbsp;Li Zhao","doi":"10.1016/j.rsase.2025.101788","DOIUrl":"10.1016/j.rsase.2025.101788","url":null,"abstract":"<div><div>This study pioneers integration of bivariate spatial autocorrelation with interpretable machine learning (XGBoost + SHAP) for the first comprehensive 23-year analysis of pollutant-ecosystem interactions in China's most critical economic region. By combining long-term remote sensing, ground observations, and extended time-series analysis, the study systematically evaluates the associations between PM_2.5, O₃, and three key ecosystem services: net primary productivity, annual water yield, and soil conservation. Results show: (1) PM_2.5 maintains persistent negative correlations with NPP and SC, especially in industrialized clusters, and although recent air-quality policies have weakened these correlations, high-pollution and low-service clusters remain, suggesting delayed ecological recovery relative to emission reductions. (2) O₃ demonstrates strengthening positive correlations with AWY (Moran's I: 0.005–0.284) and its interaction with temperature emerges as an important factor associated with slowed net primary productivity growth in agricultural regions. (3) Regional heterogeneity is evident: mountainous ecological zones with dense vegetation display lower pollutant concentrations and higher service provision, highlighting potential buffering roles of ecosystems. These findings provide data-driven evidence for differentiated PM_2.5 and O₃ management and offer a replicable methodological framework for assessing compound environmental pressures under climate change.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"40 ","pages":"Article 101788"},"PeriodicalIF":4.5,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145528816","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 the estimation of equivalent water thickness in neglected and underutilized taro crops using UAV acquired multispectral thermal image data and index-based image segmentation","authors":"Helen S. Ndlovu ,&nbsp;John Odindi ,&nbsp;Mbulisi Sibanda ,&nbsp;Onisimo Mutanga","doi":"10.1016/j.rsase.2025.101758","DOIUrl":"10.1016/j.rsase.2025.101758","url":null,"abstract":"<div><div>Taro, recognized as a future smart neglected and underutilized crop as a result of its resilience to abiotic stresses, has emerged as valuable for diversifying crop farming systems and sustaining local livelihoods. Nonetheless, a significant research gap exists in spatially explicit information on the water status of taro, contributing to the paradox of its ability to adapt to diverse agro-ecological conditions. Precision agriculture, including the use of unmanned aerial vehicles (UAVs) outfitted with high-resolution multispectral and thermal imagery, has proven effective in farm-scale monitoring and provides near-real-time information on crop water status. Hence, this study sought to evaluate the applicability of multispectral and thermal infrared UAV imagery in understanding taro's water status. Leveraging deep learning techniques to evaluate the use of thermal remote sensing and three index-based segmentation techniques in predicting the canopy equivalent water thickness (EWT) of taro crops, this study sought to determine EWT as a proxy to its water status in smallholder farmlands. The study findings illustrate a significant difference in the prediction accuracies of taro EWT with and without the thermal band (<em>P &lt; 0.05</em>). Additionally, results (R² = 0.92, RMSE = 8.04 g/m², and rRMSE = 15.31 % including the thermal band and 0.91, 8.73 g/m², and 16.64 % excluding the thermal band) reveal the value of the Excess Green minus Excess Red (ExGR) technique in accurately predicting EWT_canopy. This study serves as a foundation for developing an effective and efficient monitoring framework that provides a spatially explicit overview of neglected and underutilized crops such as taro.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"40 ","pages":"Article 101758"},"PeriodicalIF":4.5,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145528803","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":"Snow-resilient mapping reveals three decades of surface water expansion on the Qinghai–Tibet Plateau","authors":"Shaofan Tang ,&nbsp;Yilong Peng ,&nbsp;Yongzhi Chen ,&nbsp;Jiaming Liu ,&nbsp;Pingping Zhang ,&nbsp;Ying Zhang ,&nbsp;Tingting He ,&nbsp;Jianhua Li","doi":"10.1016/j.rsase.2025.101769","DOIUrl":"10.1016/j.rsase.2025.101769","url":null,"abstract":"<div><div>As the “Water Tower of Asia”, the Qinghai–Tibet Plateau (QTP) plays a pivotal role in maintaining regional hydrological balance and regulating the global climate system.Its high elevation and extensive snow cover, however, present substantial challenges for accurate surface water detection via remote sensing. This difficulty is primarily due to spectral confusion between water and snow, particularly during the winter and spring months. In response to these challenges, we have developed a novel Snow-Water Separation Detection Method (SWS) that utilizes a combination of multispectral indices to effectively minimize snow interference and enhance the precision of water body identification. Validation of this method shows that 98.87 % of the 2687 evaluated open water samples conformed to classification standards, while misclassifications were exceptionally low, with only 0.02 % of the 2545 terrestrial samples and 0.52 % of the 2203 open water samples incorrectly identified as water. This demonstrates the method's robust discriminatory capacity. We employed the SWS method on approximately 159,000 Landsat images from 1990 to 2024 across the QTP to construct a high-resolution, long-term dataset of surface water dynamics (QTP-SW). Analysis of this dataset indicates a net increase of 18,053 km² in permanent water bodies and 2296 km² in seasonal water bodies over the past 35 years.River surface area also expanded by 1788 km². These changes are significantly correlated with regional climate trends, underscoring the pronounced impact of climate change on water resources. The SWS method exhibits strong generalizability and is well-suited for application in other cold, high-altitude regions. The publicly available QTP-SW dataset (<span><span>https://10.5281/zenodo.15870117</span><svg><path></path></svg></span>) offers a valuable resource for studies on lake evolution, cryospheric dynamics, and sustainable water resource management.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"40 ","pages":"Article 101769"},"PeriodicalIF":4.5,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145747680","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":"Robust assessment of snow persistence dynamics in Iran (2002–2024) using MODIS satellite imagery and multi-source evaluation","authors":"Neamat Karimi,&nbsp;Amirhossein Sarbazvatan","doi":"10.1016/j.rsase.2025.101806","DOIUrl":"10.1016/j.rsase.2025.101806","url":null,"abstract":"<div><div>Snow persistence (SP) (defined as the proportion of time a surface remains covered by snow) plays a critical role in Iran's hydrological and climatic systems, particularly in regulating water supply for arid and semi-arid regions. This study provides a comprehensive assessment of SP variability across Iran from 2002 to 2024 using the MODIS MOD09A1 8-day surface reflectance product (500 m). Snow-covered pixels were identified using the Normalized Difference Snow Index (NDSI ≥0.4), and SP was calculated as the proportion of snow-covered days for each hydrological year. The analysis was limited to regions with mean annual SP &gt; 2 % to exclude areas with negligible snow occurrence. Long-term trends were quantified using the non-parametric Mann–Kendall test and Theil–Sen slope estimator. Results indicate a significant nationwide decline in SP, averaging −6.2 %, equivalent to an average reduction of ≈23 snow-covered days over 22 years. The most pronounced decreases occurred in the Zagros and Alborz Mountains, key snow accumulation zones for Iran's major river basins. Along the elevation gradient, SP exhibited a U-shaped pattern, with the strongest losses (≈−7 %) between 1900 and 2600 m a.s.l., while high elevations (&gt;2600 m) showed relatively stable SP. A significant upward shift in the mean elevation of snow-covered areas was observed (29 m yr⁻¹; cumulative +638 m), reflecting elevation-dependent warming (0.07–0.09 °C yr⁻¹). Using a Random Forest model (R² = 0.98, RMSE = 0.27), temperature was found to account for 58 % of the total influence on SP decline compared to 42 % for precipitation, confirming that warming trends are the dominant climatic driver of long-term snow reduction across Iran. Phenological analysis revealed that while the first snowy day (FSD) changed minimally, the last snowy day (LSD) advanced by approximately 23 days, shortening the snow season duration. MODIS-derived SP results were cross-evaluated with high-resolution composites from Landsat-8/9 and Sentinel-2 imagery, showing strong agreement (R² = 0.88). Additional validation using ERA5 and FLDAS reanalysis datasets confirmed consistent spatial patterns and magnitudes of SP decline in Iran. These integrated results highlight an accelerating reduction in Iran's snow persistence, emphasizing the urgent need for adaptive water-resource management and climate-resilience strategies in snow-dependent regions.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"40 ","pages":"Article 101806"},"PeriodicalIF":4.5,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145623636","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":"Remote sensing-based rice mapping in Brazil: Identifying the best approach for segmenting different spectral compositions using deep learning","authors":"MD Samiul Islam ,&nbsp;Andre Dalla Bernardina Garcia ,&nbsp;Ieda Del’Arco Sanches ,&nbsp;Victor Rohden Prudente ,&nbsp;Irene Cheng","doi":"10.1016/j.rsase.2025.101770","DOIUrl":"10.1016/j.rsase.2025.101770","url":null,"abstract":"<div><div>This study explored the mapping of irrigated rice fields using remote sensing data and deep learning techniques, focus-ing on the evaluation of various spectral band combinations and polarizations from Sentinel-1 and Sentinel-2 satellites. Three deep learning models, called UNET, FAPNET, and PLANET, were implemented to perform segmentation of rice fields within a region located in southern Brazil. Specifically, the FAPNET model outperformed others when using vegetation-related spectral bands (NIR, RED), while the PLANET model demonstrated greater efficacy with water-related bands (SWIR, VH, VV). However, PLANET struggled with multi-band configurations, and its slower convergence indicated the need for refined training strategies. The integration of optical and SAR data did not lead to significant performance improvements for these models, suggesting that their architectures are limited in processing more than three input channels. In contrast, the UNET model exhibited greater robustness when handling diverse data combinations, achieving balanced performance even with the integration of optical and SAR data. This suggests that while FAPNET and PLANET specialize in extracting features from specific spectral bands, UNET is more adaptable to multi-source inputs. These findings highlight the role of thoughtful model and data choice, illustrating that special-ized structures perform well with particular data setups, whereas more generalized models are superior at synthesizing various data sources. Future research should focus on enhancing PLANET's multi-band capabilities and improving FAPNET's sensitivity to SWIR, advancing segmentation precision across a broader range of spectral profiles. This study contributes to the field of crop mapping through remote sensing by providing evidence that indiscriminate data fusion is not always the optimal approach, advocating for model and spectral band choices tailored to the specific application requirements.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"40 ","pages":"Article 101770"},"PeriodicalIF":4.5,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145417691","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":"Spatiotemporal responses in sea surface temperature and chlorophyll concentration to typhoons observed by Himawari satellite and multi-satellite reanalysis datasets","authors":"Dimas Pradana Putra ,&nbsp;Po-Chun Hsu","doi":"10.1016/j.rsase.2025.101778","DOIUrl":"10.1016/j.rsase.2025.101778","url":null,"abstract":"<div><div>This study utilized multi-satellite datasets and Himawari geostationary satellite observations to analyze the spatiotemporal responses of sea surface temperature (SST) and chlorophyll concentration (Chl) induced by 1219 tropical cyclone (TC) positions in the western North Pacific from July 2015 to December 2023. By comparing daily TC observations with baseline values from 5 to 7 days before TC passage, SST drop (δSST) and Chl enhancement (δChl) were calculated. The results revealed the formation of a cold wake along the TC track, with significant SST drops concentrated within approximately 1.25° to the right of the TC center. The multi-satellite dataset indicated that δSST reached its minimum about two days after TC passage, whereas the Himawari satellite observed a similar minimum around the fifth day after passage. Despite differences in timing, the SST cooling rankings by TC intensity were consistent between the datasets. Category 4 TCs exhibited the largest cooling effect, with SST drops averaging 2–6 % within a 2.5° diameter and diminishing to 1–3 % at a 5° diameter. For Chl, both datasets displayed a similar spatial enhancement pattern on the right side of the TC track. However, the timing of maximum δChl differed: the multi-satellite dataset observed the δChl peak 3–4 days after TC passage, while the Himawari satellite typically detected the peak between the day of passage and the following day. As TC intensity increased to Category 3 and above, Chl was at least double the pre-TC levels (δChl &gt;2), indicating stronger mixing and nutrient upwelling near the TC core (2.5° diameter). Analysis of the ratio between TC wind speed (U) and translation speed (Vp) (U²Vp⁻¹) showed that higher values were associated with greater SST cooling and Chl enhancement on average. However, the large standard deviation suggested that other factors, such as oceanic background conditions and TC path locations, significantly influenced the results. Overall, we recommend using the multi-satellite SST dataset, which incorporates microwave observations to provide broader and more continuous SST coverage, thereby likely offering a closer representation of the actual ocean surface during TCs. For Chl observations, the Himawari satellite, with its complete hourly geostationary observation periods, may better reflect actual Chl conditions, while the multi-satellite Chl dataset provides high spatial completeness but relies on interpolation-based reconstructions.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"40 ","pages":"Article 101778"},"PeriodicalIF":4.5,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145417693","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":"Aboveground dry biomass modeling with Remotely Piloted Aircraft in Brazilian Savanna: a case study in an experimental area under reduced-impact logging (2005–2021)","authors":"Paola Aires Lócio de Alencar ,&nbsp;Alba Valéria Rezende ,&nbsp;Marcus Vinicio Neves d’Oliveira ,&nbsp;Eder Pereira Miguel ,&nbsp;Hallefy Junio de Souza ,&nbsp;Roberta Franco Pereira de Queiroz","doi":"10.1016/j.rsase.2025.101774","DOIUrl":"10.1016/j.rsase.2025.101774","url":null,"abstract":"<div><div>Savanna ecosystems are critical for carbon storage and biodiversity conservation, yet their structural complexity challenges biomass estimation. This study evaluates the applicability of using RGB imagery acquired by a Remotely Piloted Aircraft (RPA), through the Structure for Motion (SfM) approach, to build a statistical model for estimating aboveground dry biomass (AGB) in a cerrado <em>sensu stricto</em> area of the Brazilian Cerrado biome. The research was conducted from 2005 to 2021 in a 2.1 ha experimental area located at Fazenda Água Limpa (FAL), a research and conservation area of the University of Brasília, Federal District, Brazil. A Canopy Height Model (CHM) was generated by subtracting the Digital Surface Model (DSM) from the Digital Terrain Model (DTM). Height metrics were derived from the CHM, and an Ordinary Least Squares regression model was fitted using data from 21 field plots (20 m × 50 m). The resulting model (AGB = −2.73 - 0.54·Elev_MAD_MODE + 2.56·Elev_P99 + ε) achieved R² = 0.65 and RMSE = 0.41 Mg 0.1 ha^-1 (RMSE% = 18), and was used to map AGB across the sampled plots (2.1 ha) and the entire imaged area (10.4 ha). The estimated mean AGB was 25 ± 6.4 Mg ha⁻¹, consistent with forest inventory data. The total estimated AGB for the full imaged area was 24.6 Mg ha⁻¹. The model demonstrated potential for extrapolation to areas with similar structural characteristics.</div></div>","PeriodicalId":53227,"journal":{"name":"Remote Sensing Applications-Society and Environment","volume":"40 ","pages":"Article 101774"},"PeriodicalIF":4.5,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145528798","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}