Remote Sensing of Environment最新文献

{"title":"Using an improved radiative transfer model to estimate leaf area index, fractional vegetation cover and leaf inclination angle from Himawari-8 geostationary satellite data","authors":"Yaoyao Chen, Xihan Mu, Tim R. McVicar, Yuanyuan Wang, Yuhan Guo, Kai Yan, Yongkang Lai, Donghui Xie, Guangjian Yan","doi":"10.1016/j.rse.2024.114595","DOIUrl":"https://doi.org/10.1016/j.rse.2024.114595","url":null,"abstract":"Quantitative vegetation structural parameters such as leaf area index (LAI), fractional vegetation cover (FVC), and leaf inclination angle (LIA) are important for controlling biophysical processes, such as carbon dynamics and transpiration. The generation of remote sensing vegetation structural products using geostationary satellite data may allow for near real-time monitoring of vegetation change and associated biophysical processes. However, operational algorithms for retrieving the vegetation structure from geostationary satellite imagery are rare. Herein we developed a bidirectional model of reflectance and difference vegetation index (DVI) which requires LAI and other vegetation parameters as inputs, allowing these parameters to be estimated <em>via</em> an optimization scheme. The developed radiative transfer model specifically considers the high-frequency and multi-angle features of geostationary satellite data to separate the sun-angle related variables from the sun-angle independent variables. This parameterization facilitates the retrieval of vegetation structural products by reducing the number of variables while maintaining the generality of the model. The inversion of this physical radiative transfer model produced daily LAI and FVC with a spatial resolution of 1 km from the bidirectional reflectance factor (BRF) of Himawari-8 high-frequency observations for Australia. In contrast to most other readily available LAI products, this approach to generating Himawari-8 LAI did not rely on MODIS LAI or land cover data. Compared with field-measured data, the RMSE of Himawari-8 LAI was 1.009 and the bias was −0.354, and for FVC the RMSE was 0.132 and the bias was −0.014; these were more accurate than MODIS LAI and GLASS LAI, and GEOV3 FVC, respectively. The intercomparison of these products showed that the Himawari-8 LAI and FVC products performed well having realistic spatio-temporal distributions. For the first time, a mean leaf inclination angle (MLIA) product was generated only using satellite data. Similarity was found between the spatial patterns of MLIA and the land cover map over Australia. Independent validation data showed that the uncertainty of MLIA was generally less than 10°. The high-frequency nature of geostationary satellite imagery coupled with the radiative transfer model developed herein enables the derived vegetation structural products to facilitate improved monitoring of both short-term (<em>i.e.</em>, daily to weekly) and long-term (<em>i.e.</em>, seasonal to annual) vegetation dynamics.","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"100 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2025-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142936311","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Decoding primary forest changes in Haiti and the Dominican Republic using Landsat time series","authors":"Falu Hong, S. Blair Hedges, Zhiqiang Yang, Ji Won Suh, Shi Qiu, Joel Timyan, Zhe Zhu","doi":"10.1016/j.rse.2024.114590","DOIUrl":"https://doi.org/10.1016/j.rse.2024.114590","url":null,"abstract":"Forest loss has greatly reduced habitats and threatened Earth's biodiversity. Primary forest (PF) has an irreplaceable role in supporting biodiversity compared with secondary forest (SF). Therefore, distinguishing PF and SF using remote sensing observations is critical for evaluating the impact of forest loss on biodiversity. However, continuous monitoring of PF loss through remote sensing time series observations remains largely unexplored, particularly in developing tropical regions. In this study, we used the COLD algorithm (COntinuous monitoring of Land Disturbance) and Landsat time series data to quantify PF loss on the island of Hispaniola, comprising Haiti and the Dominican Republic, from 1996 to 2022. We considered the resilience of PF to different disturbance agents and identified the primary drivers of PF loss in Hispaniola through a sample-based approach. Accuracy assessment based on the stratified random sample shows that the overall accuracy of land cover classification is 80.5% (±5.2%) [95% confidence interval]. The user's, producer's, and overall accuracies of PF loss detection are 68.8% (±9.3%), 73.6% (±38%), and 99.4% (±0.5%), respectively. Map-based analysis reveals a more pronounced decline in PF coverage in Haiti (0.75% to 0.44% at 324 ha/year) compared to the Dominican Republic (7.14% to 5.67% at 2,704 ha/year), with substantial PF loss occurring both inside and outside protected areas. Furthermore, Haiti exhibits a higher degree of PF fragmentation, characterized by smaller and fewer PF patches, than the Dominican Republic, posing significant challenges for biodiversity conservation. The remaining PFs are found on steeper slopes in both Haiti and the Dominican Republic, suggesting that flatter, more accessible areas are more vulnerable to PF loss. Fire, tree-cutting, and hurricanes were identified as the primary drivers of PF loss, accounting for 65.7%, 20.9%, and 9.0% of the PF loss area in Hispaniola, respectively. These findings underscore the urgent need for conservation policies to protect remaining PF in Hispaniola, particularly in Haiti.","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"1 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2025-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142936310","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Sentinel-1 SAR-based global 1-km resolution soil moisture data product: Algorithm and preliminary assessment","authors":"Dong Fan, Tianjie Zhao, Xiaoguang Jiang, Almudena García-García, Toni Schmidt, Luis Samaniego, Sabine Attinger, Hua Wu, Yazhen Jiang, Jiancheng Shi, Lei Fan, Bo-Hui Tang, Wolfgang Wagner, Wouter Dorigo, Alexander Gruber, Francesco Mattia, Anna Balenzano, Luca Brocca, Thomas Jagdhuber, Jean-Pierre Wigneron, Jian Peng","doi":"10.1016/j.rse.2024.114579","DOIUrl":"https://doi.org/10.1016/j.rse.2024.114579","url":null,"abstract":"High spatial resolution of satellite-based soil moisture (SM) data are essential for hydrological, meteorological, ecological, and agricultural studies. Especially, for watershed hydrological simulation and crop water stress analysis, 1-km resolution SM data have attracted considerable attention. In this study, a dual-polarization algorithm (DPA) for SM estimation is proposed to produce a global-scale, 1-km resolution SM dataset (S1-DPA) using the Sentinel-1 synthetic aperture radar (SAR) data. Specifically, a forward model was constructed to simulate the backscatter observed by the Sentinel-1 dual-polarization SAR, and SM retrieval was achieved by minimizing the simulation error for different soil and vegetation states. The produced S1-DPA data products cover the global land surface for the period 2016–2022 and include both ascending and descending data with an observation frequency of 3–6 days for Europe and 6–12 days for the other regions. The validation results show that the S1-DPA reproduces the spatio-temporal variation characteristics of the ground-observed SM, with an unbiased root mean squared difference (ubRMSD) of 0.077 m³/m³. The generated 1-km SM product will facilitate the application of high-resolution SM data in the field of hydrology, meteorology and ecology.","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"2 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2025-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142936308","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Abiotic influences on continuous conifer forest structure across a subalpine watershed","authors":"H. Marshall Worsham, Haruko M. Wainwright, Thomas L. Powell, Nicola Falco, Lara M. Kueppers","doi":"10.1016/j.rse.2024.114587","DOIUrl":"https://doi.org/10.1016/j.rse.2024.114587","url":null,"abstract":"Understanding the abiotic drivers of high-elevation forest physiognomy is essential for forecasting how mountain ecosystems will respond to emerging environmental pressures. Most prior studies of these relationships have relied on small samples of the full landscape, resulting in limited power to detect dominant covariates and their interactions. Here we report the first evaluation of abiotic influences on a complement of accurate, wall-to-wall estimates of conifer forest structure and composition at the watershed scale. In a subalpine conifer domain in the Colorado Rocky Mountains (USA), we developed a novel method for deriving stand structure metrics from waveform LiDAR data, which showed high fidelity with field inventory. We quantified the relationships between structural and compositional metrics and climate, topographic, edaphic, and geologic factors. Our results showed that peak snow water equivalent (SWE), snow disappearance rate, and elevation explained most of the variation in forest structure. The highest stand density, basal area, maximum canopy height, and quadratic mean diameter occurred in sites with SWE around one standard deviation below mean, but with long snow residence times. Stand density decreased linearly with elevation, while other metrics peaked between 3000 m.a.s.l. and 3200 m.a.s.l. Substrate properties had weaker influence. Continuous mapping of through-canopy forest structure enabled our novel findings of the dominant role of snowpack in explaining structural and compositional variation, and of elevation thresholds. Our reproducible approach facilitates assessment of forest-topoclimate relationships in other conifer-dominated landscapes and improves understanding of the baseline patterns controlling forest structure, which is needed for predicting long-term ecological change.","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"24 3 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2025-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142935356","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Generation of robust 10-m Sentinel-2/3 synthetic aquatic reflectance bands over inland waters","authors":"Rejane S. Paulino, Vitor S. Martins, Evlyn M.L.M. Novo, Claudio C.F. Barbosa, Daniel A. Maciel, Raianny L. do N. Wanderley, Carina I. Portela, Cassia B. Caballero, Thainara M.A. Lima","doi":"10.1016/j.rse.2024.114593","DOIUrl":"https://doi.org/10.1016/j.rse.2024.114593","url":null,"abstract":"Inland waters comprise various aquatic systems, including rivers, lakes, lagoons, reservoirs, and others, and satellite data play a crucial role in providing holistic and dynamic observations of these complex ecosystems. However, available medium-spatial resolution satellite sensors, such as Sentinel-2 Multi-Spectral Instrument (MSI), are typically designed for land monitoring and lack suitable spectral bands and radiometric quality for water applications. This study developed a novel synthetic band generation method, called Sentinel-2/3 Synthetic Aquatic Reflectance Bands (S2/3Aqua), for computing eight 10-m synthetic spectral bands from multivariate regression analysis between Sentinel-2 MSI and Sentinel-3 OLCI image pair. Three multivariate regressor models, Multivariate Linear Regressor (MLR), Multivariate Quadratic Regressor (MQR), and Random Forest Regressor (RFR), were applied and assessed to replicate the Sentinel-3 spectral consistency on 10-m Sentinel-2 images. A cyanobacteria modeling was developed based on <em>in-situ</em> observations (<em>n</em> = 54), and we demonstrated, for the first time, the application of 10-m harmful algal bloom mapping over two eutrophic tropical urban reservoirs (Promissão and Billings, Brazil). Additionally, the generalization of S2/3Aqua was assessed by comparing its spectral signatures across different water optical types. Overall, the comparison between S2/3Aqua and Sentinel-3 bands achieved a mean absolute error of 6 % and a mean difference close to zero. We found that MLR exhibited a higher accuracy with <em>in-situ</em> observations (with a 28 % bias) and was more suitable than other tested models. S2/3Aqua also performed satisfactorily across all eight spectral bands, including at 620 and 681 nm, with a mean difference of less than 0.003 reflectance units. The cyanobacteria mapping showed a high level of agreement between S2/3Aqua and Sentinel-3 for low concentrations of Phycocyanin (less than 50 mg m⁻³), and S2/3Aqua effectively captured the spatial variability of narrower and smaller blooms. Finally, S2/3Aqua provides reliable synthetic spectral bands that can effectively be used in several aquatic system studies, including monitoring potentially harmful algal blooms.","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"363 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142917711","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Can real-time NDVI observations better constrain SMAP soil moisture retrievals?","authors":"Sijia Feng, Lun Gao, Jianxiu Qiu, Xiaoping Liu, Wade T. Crow, Tianjie Zhao, Chao Tan, Shaohua Wang, Jean-Pierre Wigneron","doi":"10.1016/j.rse.2024.114569","DOIUrl":"https://doi.org/10.1016/j.rse.2024.114569","url":null,"abstract":"NASA's Soil Moisture Active Passive (SMAP) satellite mission provides an unprecedented opportunity to monitor global surface soil moisture (SM). The retrieval of SMAP official SM product relies on the inversion of a zeroth-order <em>τ-ω</em> radiative transfer model constrained by climatological Normalized Difference Vegetation Index (NDVI) derived vegetation optical depth (VOD) and constant surface roughness. However, NDVI climatology cannot capture vegetation variation in response to climate extremes and agricultural practices, which can cause non-negligible errors in SMAP SM products. To resolve this issue, we develop a new Dynamic Dual-Channel Algorithm (DDCA) by constraining the <em>τ-ω</em> model using VOD and surface roughness derived from the real-time dynamic NDVI observations acquired from MODIS and VIIRS, where surface roughness is estimated through the classic DCA with VOD determined via dynamic NDVI. Considering that NDVI is not a perfect proxy for VOD, its derived surface roughness may contain VOD information to some extent. To reduce uncertainties in surface roughness, four different parameterization schemes are considered, including daily-scale, monthly average, yearly average, and constant surface roughness. Validation results against in-situ measurements demonstrate that DDCA is typically superior to the SMAP baseline algorithm – Regularized Dual-Channel Algorithm (RDCA) – across different continents, land covers, and climates, especially when parameterized with surface roughness at relatively coarse time scales (i.e., monthly or annually), indicating that averaging daily surface roughness at monthly and yearly scales can effectively reduce its uncertainties. One exception is that daily-scale roughness works well for grassland, likely because NDVI can accurately approximate VOD in grassland and its derived surface roughness is of high quality. Further analysis demonstrates that the improvement of DDCA SM over the SMAP official SM (SMAP_L3_SMPE) is particularly remarkable in cases of drought and agricultural practices. Overall, these results highlight the necessity to account for accurate vegetation dynamics during SMAP SM retrieval.","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"28 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2025-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142917354","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An interpretable attention-based deep learning method for landslide prediction based on multi-temporal InSAR time series: A case study of Xinpu landslide in the TGRA","authors":"Chao Zhou, Mingyuan Ye, Zhuge Xia, Wandi Wang, Chunbo Luo, Jan-Peter Muller","doi":"10.1016/j.rse.2024.114580","DOIUrl":"https://doi.org/10.1016/j.rse.2024.114580","url":null,"abstract":"The prediction of landslide deformation is crucial for early warning systems. While conventional geotechnical in-situ monitoring is restricted due to its high cost and spatial limitations over large regions, deep learning-based methodologies with remote sensing data have become increasingly prevalent in contemporary predictive research, yet this frequently engenders the enigmatic “black box” issue. To address this, we improve the landslide displacement prediction framework by combining interpretable deep learning based on an attention mechanism and Multi-Temporal Interferometric Synthetic Aperture Radar (MT-InSAR) techniques. MT-InSAR is first used to extract a landslide displacement time series from Copernicus Sentinel-1 SAR images. Then Variational Mode Decomposition (VMD) is employed to separate the nonlinear displacement time series into trend, seasonal, and noise components. The Auto-Regressive Integrated Moving Average (ARIMA) model and Bidirectional Gated Recurrent Unit (BiGRU) are applied to predict trend and seasonal displacements, respectively. The inputs for these predictions are determined by analyzing landslide influencing factors. This study uses the Xinpu landslide in the Three Gorges Reservoir Area of China to evaluate the proposed method and compare its performance with existing models. The CNN-Attention-BiGRU algorithm effectively captures the nonlinear relationship between landslide deformation and its triggering factors, outperforming conventional deep learning models such as BiLSTM, BiGRU, and CNN-BiGRU, achieving improvements in Root Mean Square Errors (RMSEs) by 21%—55% and Mean Absolute Errors (MAEs) by 23%—56%. By applying deep learning with an attention mechanism, our proposed method considers the underlying principles of landslide deformation, and factors with higher relative importance for prediction modeling are interpreted to be concentrated annually between April and August, enabling a more effective and more accurate prediction of large-scale landslide kinematics for the studied reservoir region.","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"39 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2025-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142912034","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A novel self-similarity cluster grouping approach for individual tree crown segmentation using multi-features from UAV-based LiDAR and multi-angle photogrammetry data","authors":"Lingting Lei, Guoqi Chai, Zongqi Yao, Yingbo Li, Xiang Jia, Xiaoli Zhang","doi":"10.1016/j.rse.2024.114588","DOIUrl":"https://doi.org/10.1016/j.rse.2024.114588","url":null,"abstract":"Automatic collection of tree-level crown information is essential for sustainable forest management and fine carbon stock estimation. UAV-based light detection and ranging (LiDAR) and UAV-based multi-angle photogrammetry (UMP) data depict the 3D structure of forests at a fine-grained level by generating detailed point clouds, making them potential alternatives to labor-intensive forest inventories. However, the accuracy of the individual tree crown segmentation algorithms that have been developed is unstable in forest stands with high terrain undulation and high canopy density, mainly due to the various crown sizes and interlocking crowns resulting in varying degrees of over- or under-segmentation. Here, we propose self-similarity cluster grouping (SCG) algorithm for individual tree crown segmentation that integrates multivariable calculus of crown surfaces and spectral-texture-color spatial information of crown. Firstly, according to the property that DSM and its multi-order gradient information can characterize the crown surface variation and concavity-convexity features, first- and second-order edge detection operators were used to preliminarily determine the crown patch edges in order to reduce under-segmentation. Then, we developed a self-similarity weight function controlled by the spectral, texture and color spatial information of the tree crown patches to increase the similarity difference between adjacent crown patches of the same tree and those of neighboring trees, and designed the strategy for cluster grouping crown patches to complete individual tree crown segmentation. The performance of the proposed SCG algorithm was verified in Mytilaria, Red oatchestnu, Chinese fir and Eucalyptus plots in subtropical forests of China using LiDAR and UMP data. The overall accuracy of F-score (&lt;em&gt;f&lt;/em&gt;) was above 0.85 for crown segmentation, and the rRMSE for crown width, crown area and crown circumference extractions reached 0.13, 0.22 and 0.14, respectively. On this basis, we evaluated the effect of spatial resolution of DSM on the segmentation accuracy of SCG algorithm, and found that the crown segmentation accuracy was proportional to the spatial resolution. Compared to the normalized cut algorithm, marker-controlled watershed algorithm and threshold-based cloud point segmentation algorithm, the SCG algorithm improved the overall accuracy &lt;em&gt;f&lt;/em&gt; of individual tree crown segmentation by 0.06, 0.13 and 0.05 for LiDAR and 0.06, 0.21 and 0.10 for UMP, respectively. Furthermore, the effectiveness and generalizability of the SCG algorithm was verified in other Mytilaria, Red oatchestnut, Chinese fir and Eucalyptus plots in subtropical forests and Larch and Chinese pine plots in temperate forests using UMP data. The crown segmentation accuracy was better than 0.82, and the crown width extraction accuracy was up to 89 %. Overall, our proposed SCG algorithm reduces the over- and under-segmentation in complex forest structures and provides technica","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"114 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2024-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142908437","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A long-term global Mollisols SOC content prediction framework: Integrating prior knowledge, geographical partitioning, and deep learning models with spatio-temporal validation","authors":"Xiangtian Meng, Yilin Bao, Xinle Zhang, Chong Luo, Huanjun Liu","doi":"10.1016/j.rse.2024.114592","DOIUrl":"https://doi.org/10.1016/j.rse.2024.114592","url":null,"abstract":"Recently, Soil Organic Carbon (SOC) content has declined across global Mollisols region due to erosion, intensive agriculture, and other factors, weakening the soil's capacity to buffer climate change and necessitating urgent monitoring of SOC dynamics. Large-scale SOC content monitoring using remote sensing technology faces challenges in extracting advanced features from remote sensing data and mitigating the negative impact of high spatial heterogeneity in SOC content on prediction accuracy. To address these challenges, we collected 8984 samples, 956,423 Landsat TM/OLI images, shuttle radar topography mission-digital elevation model data, and meteorological data. We developed a Geographic Knowledge Dataset (GEKD) incorporating prior knowledge of soil formation and erosion processes. We then input the GEKD into a Probability Hybrid Model (PHM). In the PHM, we applied a fuzzy Gaussian mixture model to cluster the global Mollisols region and calculate corresponding probabilities. We then built a high-accuracy SOC content prediction model by integrating the Attention mechanism, Convolutional Neural Networks, and Convolutional Long Short-Term Memory Networks (A-CNN-ConvLSTM). Finally, we generated spatial maps of SOC content at a 30 m resolution for 8 periods since 1984 and verified the accuracy of its spatial distribution and temporal variation patterns. The results showed that (1) the highest SOC content prediction accuracy (<em>RMSE</em> = 7.17 g/kg, <em>R</em>^<em>2</em> = 0.72, and <em>RPIQ</em> = 1.92) was achieved when GEKD was input into PHM using the A-CNN-ConvLSTM algorithm. (2) PHM effectively reduces the negative impact of high SOC spatial heterogeneity on prediction accuracy, resulting in smoother spatial distribution at cluster boundaries. Compared to the global model, PHM reduced <em>RMSE</em> by 1.66 g/kg and improved <em>R</em>^<em>2</em> and <em>RPIQ</em> by 0.06 and 0.15, respectively. (3) Compared to the commonly used random forest algorithm, A-CNN-ConvLSTM reduced <em>RMSE</em> by 1.50 g/kg and improved <em>R</em>^<em>2</em> and <em>RPIQ</em> by 0.13 and 0.47, respectively. The spatial context features extracted by the CNN structure in the A-CNN-ConvLSTM algorithm are the most effective in improving SOC content prediction accuracy. (4) Currently, the SOC content across continents in the global Mollisols region is ranked as follows: Siberia (27.21 g/kg) &gt; Europe (26.78 g/kg) &gt; Asia (20.48 g/kg) &gt; North America (20.43 g/kg) &gt; South America (16.49 g/kg). Since 1984, SOC content has shown a decreasing trend, with the global Mollisols region losing 1.91 g/kg overall. The Asian Mollisols region experienced the largest decline (2.93 g/kg), while Siberia saw the smallest decrease (1.45 g/kg).","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"42 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2024-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142908406","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mapping the spatial distribution of species using airborne and spaceborne imaging spectroscopy: A case study of invasive plants","authors":"M. Ny Aina Rakotoarivony, Hamed Gholizadeh, Kianoosh Hassani, Lu Zhai, Christian Rossi","doi":"10.1016/j.rse.2024.114583","DOIUrl":"https://doi.org/10.1016/j.rse.2024.114583","url":null,"abstract":"Predicting the spatial distribution of invasive plants remains challenging because of the complex relationships between plant invasion, abiotic, and biotic factors. While conventional species distribution models (SDMs) are often developed using abiotic factors, recent studies have suggested that including biotic factors, particularly plant functional traits, can improve our capability to model the distribution of invasive plants. Remote sensing is capable of estimating plant functional traits across large spatial extents. These remotely-estimated plant functional traits can then be used as predictors in mapping the spatial distribution of species. However, exploring the application of remotely-estimated plant functional traits in mapping the spatial distribution of invasive plants is relatively understudied. In this study, we aimed to (1) develop trait-based approaches for mapping the spatial distribution of an invasive plant, (2) assess the scale-dependency of these trait-based approaches, and (3) determine the capability of spaceborne hyperspectral imagery in mapping the spatial distribution of invasive plants through fusing their data with fine spatial resolution multispectral data. We focused on &lt;em&gt;Lespedeza cuneata&lt;/em&gt; (hereafter, &lt;em&gt;L. cuneata&lt;/em&gt;)&lt;em&gt;,&lt;/em&gt; commonly known as sericea lespedeza, an invasive legume threatening grassland ecosystems of the U.S. Southern Great Plains. To achieve our objectives, we collected &lt;em&gt;in situ&lt;/em&gt; data, including plant functional traits, such as foliar nitrogen, phosphorus, and potassium, and measured average canopy height, and percent cover of &lt;em&gt;L. cuneata&lt;/em&gt; from 900 sampling quadrats. We also collected remote sensing data, including airborne hyperspectral data (400–2500 nm, 1 m spatial resolution), spaceborne hyperspectral data from DLR's DESIS (401.9–999.5 nm, 30 m spatial resolution), and PlanetScope multispectral data (8 bands, 3 m spatial resolution). We also fused DESIS and PlanetScope imagery to produce fine spatial and fine spectral imagery (401.9–999.5 nm, 3 m spatial resolution). We used partial least squares regression (PLSR) to estimate plant functional traits from remotely sensed data and developed approaches for mapping the spatial distribution of invasive plants using remotely-estimated plant functional traits. We developed approaches for mapping the spatial distribution of invasive plants across spatial scales, at 1 m, 3 m, and 30 m spatial resolutions, using (1) abiotic factors only, (2) remotely-estimated plant functional traits only, and (3) remotely-estimated plant functional traits along with abiotic factors. Our findings showed that trait-based approaches for mapping the spatial distribution of invasive plants had higher accuracy than abiotic-based approaches, mapping the spatial distribution of &lt;em&gt;L. cuneata&lt;/em&gt; at fine spatial resolution performed better than at coarse spatial resolution, and fusion of coarse spatial resolution hyperspectral imagery with fine spatial","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"167 1","pages":""},"PeriodicalIF":13.5,"publicationDate":"2024-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142904888","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}