Remote Sensing of Environment最新文献

{"title":"Structural diversity, tree species richness and biomass production in new forests on agricultural land","authors":"Yamina M. Rosas,&nbsp;Inger Kappel Schmidt,&nbsp;Torben Riis-Nielsen,&nbsp;Lars Vesterdal,&nbsp;Per Gundersen,&nbsp;David Bille Byriel,&nbsp;Mathias Just Justesen,&nbsp;Davide Barsotti,&nbsp;Sebastian Kepfer-Rojas","doi":"10.1016/j.rse.2025.114978","DOIUrl":"10.1016/j.rse.2025.114978","url":null,"abstract":"<div><div>Changes in forest structure (FS) significantly influence forest conditions and dynamics, ecosystem services, including biomass accumulation and biodiversity. Several LiDAR metrics have been described to analyse FS at a high spatial resolution, including vertical and horizontal attributes. This study aimed to investigate FS development in spruce (<em>Picea abies</em>), oak (<em>Quercus robur</em>) and beech (<em>Fagus sylvatica</em>) through the first 50 years after planting and their relationship to biomass stocks and richness of arrival tree species. LiDAR and ground truth data were extracted from 60 plots. To characterize the FS, LiDAR metrics were calculated using a point cloud of 5.4 pulses/m2 under leaf-off conditions from 2019. Moreover, trees were measured in 2022 following the Danish National Forest Inventory protocols to calculate biomass and estimate tree species richness. FS changes and their relationship were analyzed by employing linear mixed models and principal component analysis (PCA). Results show that spruce stands exhibited rapid growth in height and the highest canopy cover values over time. Contrarily, oak stands developed a multi-layer canopy, increasing light availability and tree size variability with age. At later stages, beech stands reached similar height as spruce and showed the highest standard deviation of height (StDH) with age. The PCA revealed that tree maximum height and StDH were primarily age-dependent, while the internal vertical and horizontal variability tree species dependent. Spruce showed a constant homogeneous FS, oak was characterized by a high internal and external FS heterogeneity, while beech gradually decreased the FS variability over time. Old, tall and structurally homogeneous beech and spruce stands supported the highest biomass accumulation. Conversely, more structurally heterogeneous oak stands tended to harbor greater tree species richness. The study underscores the importance of incorporating both vertical and horizontal dimensions to understand FS changes and the trade-offs between biomass production and biodiversity in monoculture afforestation.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"330 ","pages":"Article 114978"},"PeriodicalIF":11.4,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144890613","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":"Multi-E2E: An end-to-end urban land-use mapping framework integrating high-resolution remote sensing images and multi-source geographical data","authors":"Ruiyi Yang ,&nbsp;Yanfei Zhong ,&nbsp;Yu Su","doi":"10.1016/j.rse.2025.114966","DOIUrl":"10.1016/j.rse.2025.114966","url":null,"abstract":"<div><div>The urban land-use map reflects the distribution of the different functional lands in the city, serving as a valuable reference for urban planning and management. High-resolution remote sensing (HRS) images provide detailed spatial information about parcels but lack socio-economic information, which is crucial for identifying land-use types. To enhance the mapping performance and obtain more comprehensive land-use information, the integration of HRS images and points of interest (POIs) with socio-economic information is crucial. Nonetheless, the existing land-use mapping methods based on HRS images and POIs are generally confronted with the following challenges: 1) due to the reliance on prior knowledge, the existing methods cannot automatically capture the complex relationships between multi-source data and land-use categories; 2) there are substantial semantic disparities between HRS images and POIs, so that the simple fusion approaches cannot fully utilize the complementary information; and 3) the existing methods are generally based on the assumption of complete modalities, resulting in them failing to work on POI-deficient parcels. In this paper, to address these issues, an end-to-end urban land-use mapping framework integrating HRS images and multi-source geographic data (Multi-E2E) is proposed. The Multi-E2E framework automatically establishes the mapping from multi-source data to land-use categories through a data-driven approach, and generates informative fused representations with an adaptive fusion module (AFM). In Multi-E2E, the labeled HRS image-POI pairs are constructed using the areas of interest (AOIs), and the interactions between modalities are facilitated by the end-to-end architecture. To identify POI-deficient parcels and ensure that the modality-specific encoders are adequately supervised, a unimodal supervision module (USM) is introduced in the Multi-E2E framework. Experiments conducted with multi-source samples from 34 Chinese cities and the urban regions of Beijing, Wuhan, Hong Kong, Macao, and Helsinki validate the effectiveness and generalizability of the proposed framework for urban land-use mapping applications. The code will be publicly available at <span><span>https://github.com/Rayoll/Multi_E2E</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"330 ","pages":"Article 114966"},"PeriodicalIF":11.4,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144890615","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":"Harmoni-Planet: A holistic harmonization method for PlanetScope constellation imagery leveraging a graph-based greedy optimization strategy","authors":"Ruilin Chen ,&nbsp;Wei Yang ,&nbsp;Xuehong Chen ,&nbsp;Zhuoning Gu ,&nbsp;Benfeng Yin ,&nbsp;Yuanming Zhang ,&nbsp;Jin Chen","doi":"10.1016/j.rse.2025.114986","DOIUrl":"10.1016/j.rse.2025.114986","url":null,"abstract":"<div><div>The PlanetScope CubeSat constellation provides unprecedentedly high spatiotemporal resolution for Earth observations but is limited by radiometric inconsistencies resulting from sensor degradation and spectral configuration differences. Existing harmonization methods often rely on internal or external references, limiting harmonization to only a subset of spectral bands or restricting applicability to localized spatiotemporal scales. To address these limitations, we propose Harmoni-Planet, a novel harmonization method that leverages graph-based greedy optimization to achieve holistic radiometric consistency across all PlanetScope bands without requiring reference data. The method consists of two components: (1) graph construction, which integrates unharmonized images into a graph with nodes representing images and edges connecting intersecting images, and (2) graph optimization, which iteratively minimizes radiometric inconsistencies between each image and its intersecting images to optimize consistency. Harmoni-Planet was validated across four geographically diverse regions (the Nile, Beijing, Indonesia, and Greenland), achieving substantial mean absolute error (MAE) reductions of 53 %, 53 %, 57 %, and 25 %, respectively, and outperforming the CubeSat-enabled spatiotemporal enhancement method (CESTEM; 40 %, 41 %, 47 %, and − 69 %, respectively) and the official harmonization algorithm (25 %, 33 %, −9 %, and 9 %, respectively). Harmoni-Planet significantly improves the spatial and temporal comparability of imagery across all PlanetScope bands, regardless of whether images are acquired by the same or different generations of satellites. In addition, it supports flexible scene-based and strip-based implementations, effectively resolving both intra-strip and cross-strip inconsistencies. It also demonstrates robust potential for near-real-time harmonization of newly acquired imagery to accommodate the rapidly expanding data volume of the PlanetScope constellation. Furthermore, Harmoni-Planet supports seamless integration with standard third-party reflectance products such as Landsat-8 and Sentinel-2. Harmoni-Planet provides a practical solution to address cross-sensor radiometric inconsistencies, substantially improving the quality and reliability of PlanetScope data for diverse Earth observation applications.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"330 ","pages":"Article 114986"},"PeriodicalIF":11.4,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144890616","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":"Retrieving yearly forest growth from satellite data: A deep learning based approach","authors":"Martin Schwartz ,&nbsp;Philippe Ciais ,&nbsp;Ewan Sean ,&nbsp;Aurélien de Truchis ,&nbsp;Cédric Vega ,&nbsp;Nikola Besic ,&nbsp;Ibrahim Fayad ,&nbsp;Jean-Pierre Wigneron ,&nbsp;Sarah Brood ,&nbsp;Agnès Pelissier-Tanon ,&nbsp;Jan Pauls ,&nbsp;Gabriel Belouze ,&nbsp;Yidi Xu","doi":"10.1016/j.rse.2025.114959","DOIUrl":"10.1016/j.rse.2025.114959","url":null,"abstract":"<div><div>High-resolution mapping of forest attributes is crucial for ecosystem monitoring and carbon budget assessments. Recent advancements have leveraged satellite imagery and deep learning algorithms to generate high-resolution forest height maps. While these maps provide valuable snapshots of forest conditions, they lack the temporal resolution to estimate forest-related carbon fluxes or track annual changes. Few studies have produced annual forest height, volume, or biomass change maps validated at the forest stand level. To address this limitation, we developed a deep learning framework, coupling data from Sentinel-1 (S1), Sentinel-2 (S2) and from the Global Ecosystem Dynamics Investigation (GEDI) mission, to generate a time series of forest height, growing stock volume, and aboveground biomass at 10 to 30-m spatial resolution that we refer to as FORMS-T (FORest Multiple Satellite Time series). Unlike previous studies, we train our model on individual S2 scenes, rather than on growing season composites, to account for acquisition variability and improve generalization across years. We produced these maps for France over seven years (2018–2024) for height at 10 m resolution and further converted them to 30 m maps of growing stock volume and aboveground biomass using leaf type-specific allometric equations. Evaluation against the French National Forest Inventory (NFI) showed an average mean absolute error of 3.07 m for height (r² <!-->=<!--> <!-->0.68) across all years, 86 m³ ha^-1 for volume and 65.1 Mg ha^-1 for biomass. We further evaluated FORMS-T capacity to capture growth on a site where two successive airborne laser scanning (ALS) campaigns were available, showing a good agreement with ALS data when aggregating at coarser spatial resolution (r² <!-->=<!--> <!-->0.60, MAE<!--> <!-->=<!--> <!-->0.27 m for the 2020–2022 growth of trees between 10 and 15 m in 5 km pixels). Additionally, we compared our results to the NFI-based wood volume production at regional level and obtained a good agreement with a MAE of 1.45 m³ ha^-1 yr^-1 and r² of 0.59. We then leveraged our height change maps to derive species-specific growth curves and compared them to ground-based measurements, highlighting distinct growth dynamics and regional variations in forest management practices. Further development of such maps could contribute to the assessment of forest-related carbon stocks and fluxes, contributing to the formulation of a comprehensive carbon budget at the country scale, and supporting global efforts to mitigate climate change.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"330 ","pages":"Article 114959"},"PeriodicalIF":11.4,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144892272","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":"Deep learning of seasonal peak snow water content of global boreal forest and arctic using spaceborne L-band radiometry","authors":"Divya Kumawat ,&nbsp;Ardeshir Ebtehaj ,&nbsp;Sujay Kumar ,&nbsp;Andreas Colliander","doi":"10.1016/j.rse.2025.114963","DOIUrl":"10.1016/j.rse.2025.114963","url":null,"abstract":"<div><div>Estimating peak snow water equivalent (SWE) across the Northern Hemisphere is critical for assessing seasonal water availability for both ecosystems and human needs. This study is the first to demonstrate a direct link between peak SWE and the temporal variability of L-band surface emission under a moderately dense vegetation canopy. We introduce SWEFormer, a novel deep transformer neural network that retrieves peak SWE primarily using time series of L-band brightness temperatures from NASA’s Soil Moisture Active and Passive (SMAP) satellite. The model is trained using an incremental learning approach that transfers low-level information from reanalysis data for spatially coherent high-level learning from sparse <em>in situ</em> observations. SWEFormer outperforms leading global products, including ERA5, GlobSnow, and AMSR-based estimates, particularly in complex boreal watersheds, where previous global SWE estimates suffer from significant uncertainties, as vegetation canopy often markedly attenuates high-frequency microwave signatures of snowpack.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"330 ","pages":"Article 114963"},"PeriodicalIF":11.4,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144886431","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":"Satellite mapping of emperor penguin habitat dispersal under climate extremes","authors":"Hong Lin ,&nbsp;Xiao Cheng ,&nbsp;Jinyang Du ,&nbsp;John S. Kimball ,&nbsp;Ziyu Yan ,&nbsp;Teng Li ,&nbsp;Tongwen Li ,&nbsp;Yibo Li ,&nbsp;Zilong Chen","doi":"10.1016/j.rse.2025.114984","DOIUrl":"10.1016/j.rse.2025.114984","url":null,"abstract":"<div><div>Emperor penguins serve as early-warning sentinels for the Antarctic ecosystem and climate change. Understanding how climate change influences their habitat use offers insights into the fragile polar ecosystem for supporting the climate actions under the United Nations Sustainable Development Goals (SDGs). However, it remains unclear how the gradual climate change and extreme climatic events affect the dispersal of emperor penguin breeding habitats due to the lack of a systematic and long-term dataset documenting their habitat use. Here, we first develop guano indices and present an automated approach to map emperor penguin breeding habitats at 30-m spatial resolution using Earth observation satellite imagery, achieving a user accuracy of 94.8 %. We further reveal that habitat dispersal is sensitive to four extreme events—heat, blizzard, storm, and low sea ice. Specifically, colonies exposed to intense climate extremes generally exhibit more fragmented distributions, with habitat reuse periods mostly under 3 years and interannual habitat dispersal exceeding 4 km. These four extreme events together explained 21 %–72 % of the variability in annual habitat dispersal. Under a high-emission scenario driven by fossil fuels, the warming-induced annual fragmentation of habitats is projected to be 255 m greater than that under a low-emission scenario using clean energy, leading to higher vulnerability in emperor penguins by disrupting their ability to survive and reproduce. The proposed method enables routine mapping and updating of emperor penguin breeding habitats, and the associated findings demonstrate that extreme climatic events significantly impact habitat use and dispersal patterns, highlighting the urgent need for global climate policies aligned with sustainable development to protect the Antarctic ecosystem.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"330 ","pages":"Article 114984"},"PeriodicalIF":11.4,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144886473","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":"Global radio frequency interference in L-band SAR data from ALOS-1 and JERS-1 satellites","authors":"Yuhang Wang,&nbsp;Cunren Liang,&nbsp;Fan Yang,&nbsp;Ligong Yang,&nbsp;Shangzong Lu,&nbsp;Xue Li,&nbsp;Qiming Zeng","doi":"10.1016/j.rse.2025.114955","DOIUrl":"10.1016/j.rse.2025.114955","url":null,"abstract":"<div><div>Synthetic Aperture Radar (SAR) data acquisition is increasingly challenged by the overcrowded radio frequency bands, leading to the problem of Radio Frequency Interference (RFI), particularly at L-band. Knowledges about global RFI are important for the protection of the microwave bands allocated for active remote sensing, SAR system design and operation, as well as the systematic mitigation of RFI for SAR data end users. It is becoming more important as we are entering the unprecedented golden age of L-band SAR satellites. In this work, we develop algorithms, workflow and system for operationally measuring global RFI experienced by L-band SAR data. We define a data format for recording the RFI measurements, which facilitates the computation of a number of RFI metrics. We process all available L-band SAR data acquired by the JAXA ALOS-1 (2006–2011) and JERS-1 (1992–1998) satellites. For comparison with RFI in C-band data, we also process part of the ESA Envisat (2002−2012) data archive acquired in the same time frame (2005–2007). The final results provide a global view of RFI at L-band, showing its spatial and temporal variations. We find heavy RFI pollutions in East Asia, North America, and the region consisting of Europe, North Africa and the Middle East. Comparison of results from ALOS-1 and JERS-1 indicates the increasing trend of RFI pollution. Comparison with C-band Envisat result shows that RFI at L-band is significantly more severe than at C-band. Comparison of results from different polarizations shows that cross polarization data are more susceptible to RFI due to their lower radar backscatter. The global RFI distribution from L-band SAR data is found similar to that from L-band passive radiometers including SMOS and SMAP.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"330 ","pages":"Article 114955"},"PeriodicalIF":11.4,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144890612","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":"Development of an enhanced hybrid piecewise logistic model for retrieving land surface phenology in drylands","authors":"Yongchang Ye ,&nbsp;Xiaoyang Zhang ,&nbsp;Jianmin Wang ,&nbsp;Khuong H. Tran ,&nbsp;Yuxia Liu ,&nbsp;Yu Shen ,&nbsp;Shuai Gao ,&nbsp;Shuai An","doi":"10.1016/j.rse.2025.114982","DOIUrl":"10.1016/j.rse.2025.114982","url":null,"abstract":"<div><div>The accurate retrieval of land surface phenology (LSP) for drylands is extremely challenging. Drylands exhibit vegetation characteristics such as sparse and patchy vegetation cover, low seasonal greenness variability, and high spatial heterogeneity. The irregular rainy and dry episodes often complicate vegetation growth, leading to an irregular temporal trajectory with multiple growth stages during the greenup and senescence phases. Moreover, the heterogeneous phenological cycles among the vegetation species in a satellite pixel and other factors may lead to a long period with only a very slight increase or decrease in greenness before or after a vegetation growing cycle. Current phenological retrieval methods, however, commonly assume that vegetation greenness gradually increases in a greenup phase and decreases in a senescence phase, following a single sigmoidal growth trajectory, which is inadequate to describe the irregular growth in drylands. In this study, we developed a novel algorithm to improve on the hybrid piecewise logistic model (HPLM) for improving LSP retrievals, especially in drylands. Our enhanced HPLM (E-HPLM) algorithm addresses two characteristics of irregular growth trajectories: (1) the multiple plateau stages within a greenup or senescence phase, and (2) the long linear tail before the start or after the end of a growing season. Specifically, we identified multiple growth stages within a greenup or senescence phase in order to fit each stage separately with a logistic model, and added a linear parameter to the logistic model to eliminate long linear tails by adjusting the background values. We implemented this new algorithm to retrieve LSPs in global drylands using the 500 m Visible Infrared Imaging Radiometer Suite (VIIRS) dataset from 2013 to 2022. The results were then compared with those of HPLM-retrieved LSPs. We also evaluated the E-HPLM results using phenometrics derived from the PhenoCam observations at site levels and the fused Harmonized Landsat and Sentinel-2 (HLS)-PhenoCam dataset at regional levels. The E-HPLM was able to reduce the uncertainty by ∼10 days in the pixels with plateau stages from 2013 to 2022 in global drylands in comparison with the HPLM algorithm, where the plateau stage appeared in over 74 % of drylands. Compared with the HPLM, the E-HPLM improved overall phenology accuracy by two days for the PhenoCam sites and one to four days in HLS-PhenoCam areas, although the improvements varied with land cover types and aridity levels. The E-HPLM algorithm has the potential to replace the current HPLM algorithm, with improved ability to retrieve LSP in drylands and to generate global LSP products.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"330 ","pages":"Article 114982"},"PeriodicalIF":11.4,"publicationDate":"2025-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144886430","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":"Regression coefficient estimation from remote sensing maps","authors":"Kerri Lu ,&nbsp;Dan M. Kluger ,&nbsp;Stephen Bates ,&nbsp;Sherrie Wang","doi":"10.1016/j.rse.2025.114949","DOIUrl":"10.1016/j.rse.2025.114949","url":null,"abstract":"<div><div>Regressions are commonly used in environmental science and economics to identify causal or associative relationships between variables. In these settings, remote sensing-derived map products increasingly serve as sources of variables, enabling estimation of effects such as the impact of conservation zones on deforestation. However, the quality of map products varies, and — because maps are outputs of complex machine learning algorithms that take in a variety of remotely sensed variables as inputs — errors are difficult to characterize. Thus, population-level estimators from such maps may be biased. In this paper, we apply prediction-powered inference (PPI) to estimate regression coefficients relating a response variable and covariates to each other. PPI is a method that estimates parameters of interest by using a small amount of randomly sampled ground truth data to correct for bias in large-scale remote sensing map products. Applying PPI across multiple remote sensing use cases in regression coefficient estimation, we find that it results in estimates that are (1) more reliable than using the map product as if it were 100% accurate and (2) have lower uncertainty than using only the ground truth sample data and ignoring the map product. Empirically, we observe effective sample size increases of up to 17-fold using PPI compared to only using ground truth data. This is the first work to estimate remote sensing regression coefficients without assumptions on the structure of map product errors. Data and code are available at <span><span>https://github.com/Earth-Intelligence-Lab/uncertainty-quantification</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"330 ","pages":"Article 114949"},"PeriodicalIF":11.4,"publicationDate":"2025-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144865703","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":"Learning to reason over multi-granularity knowledge graph for zero-shot urban land-use mapping","authors":"Yansheng Li ,&nbsp;Yu Wang ,&nbsp;Lei Yu ,&nbsp;Bo Dang ,&nbsp;Gang Xu ,&nbsp;Zhenyu Zhong ,&nbsp;Yuning Wu ,&nbsp;Xin Guo ,&nbsp;Kang Wu ,&nbsp;Zheng Li ,&nbsp;Linlin Wang ,&nbsp;Jian Wang ,&nbsp;Jingdong Chen ,&nbsp;Ming Yang ,&nbsp;Yongjun Zhang","doi":"10.1016/j.rse.2025.114961","DOIUrl":"10.1016/j.rse.2025.114961","url":null,"abstract":"<div><div>Accurate urban land-use mapping is an essential undertaking for various urban issues, such as urban planning, disease transmission, and climate change. Recently, learning-based method has emerged as a prevalent approach for urban land-use mapping, although it relies heavily on abundant labeled data. However, since land-use categories are jointly determined by physical and social attributes, obtaining such labels is challenging. This scarcity of labeled data often leads existing learning-based methods to overfit, resulting in models that struggle to recognize diverse land-use categories. To bypass these limitations, this paper for the first time advocates knowledge graph to leverage indirect supervision from related tasks for zero-shot land-use mapping. Toward this goal, this paper introduces a multi-granularity knowledge graph reasoning (mKGR) framework.R Only with indirect supervision from other tasks, mKGR can automatically integrate multimodal geospatial data as varying granularity entities and rich spatial–semantic interaction relationships. Subsequently, mKGR incorporates a fault-tolerant knowledge graph embedding method to establish relationships between geographic units and land-use categories, thereby reasoning land-use mapping outcomes. Extensive experiments demonstrate that mKGR not only outperforms existing zero-shot approaches, but also exceeds those with direct supervision, achieving improvements of 0.16, 0.08, and 0.20 on PA, UA, and OA. Furthermore, this paper reveals the superiority of mKGR in large-scale holistic reasoning, an essential aspect of land-use mapping. Benefiting from mKGR’s zero-shot classification and large-scale holistic reasoning capabilities, a comprehensive urban land-use map of China is generated with low-cost. In addition, a nationwide assessment of 15-minute city walkability over the land-use map provides insights for urban planning and sustainable development. The code and data are available at <span><span>https://github.com/vvangfaye/mKGR</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":417,"journal":{"name":"Remote Sensing of Environment","volume":"330 ","pages":"Article 114961"},"PeriodicalIF":11.4,"publicationDate":"2025-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144865838","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}