[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82924-en":3,"doc-seo-82924-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},82924,8796095461610,"Oliver","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Shifting from Discrete to Continuous Reference Data: QSM-Derived Horizontal Tree Biomass Distribution for Deep Learning Biomass Estimation","Methodologies for LiDAR-based above-ground biomass (AGB) estimation often use discrete plot-level inventory aggregates, which can introduce boundary-effect uncertainties and degrade model performance in small field plots. The work evaluates a Horizontal Biomass Distribution (HBD) reference continuously mapped from Quantitative Structure Models (QSMs). A sparse 3D U-Net was trained on simulated broadleaved forests with three AGB reference types, showing QSM-based models outperform FI at small plot sizes. For 100 m² plots, HBD reference lowers RRMSE by 16.84 ± 4.37% and improves accuracy by 0.22 ± 0.05 versus the FI baseline.","Shifting from Discrete to Continuous Reference Data: QSM-Derived Horizontal Tree Biomass Distribution for Deep Learning Biomass Estimation  \nNils Griese1 *, Christoph Kleinn1 and Nils Nölke1  \n1Department of Forest Inventory and Remote Sensing, University of Göttingen, Göttingen, Germany  \n*Corresponding author: Tel: +49 551 39-233472; [Email: ](Email: nils.griese@uni-goettingen.de)[nils.griese@uni-goettingen.de](Email: nils.griese@uni-goettingen.de)  \nAbstract  \nConventional modeling approaches for LiDAR-based above-ground biomass (AGB) estimation rely on discrete plotlevel inventory aggregates. This methodology introduces boundary-effect uncertainties that may severely degrade model performance within small field plots. To solve this limitation, we evaluate a Horizontal Biomass Distribution (HBD) reference mapped continuously from Quantitative Structure Models (QSMs) . We trained a sparse 3D U-Net on simulated broadleaved forest structures using three AGB reference types: a standard forest inventory (FI) plot-level aggregate, an edge-effect-free QSM plot-level aggregate, and a continuous HBD mapping. Evaluating training plot sizes scaling from 100 to 2500 m2, QSM-based models systematically outperformed FI approaches at small plot sizes. Specifically, for 100 m2 plots, the HBD reference reduced the relative root mean square error (RRMSE) by 16.84 ± 4.37 % and increased 􀜴2 by 0 .22 ± 0.05 against the FI baseline. By replacing plot level aggregates with HBDs as AGB reference, this methodology corrects for edge-effects and shows that using an HBD-based reference enhances model performance for small plot sizes.  \nKeywords: Above Ground Biomass, Simulation, Horizontal Biomass Distribution, Plot Size  \n1. Introduction  \nForest ecosystems play a critical role in the global carbon cycle, acting as substantial sinks for atmospheric carbon dioxide. Consequently, the accurate quantification and monitoring of above-ground biomass (AGB) are essential for sustainable forest management, climate change mitigation strategies, and carbon reporting1,2 . Since living woody biomass of accounts for most of terrestrial AGB3, capturing it is one of the core objectives of contemporary national forest inventories4–7. Traditional forest inventories (FIs) usually provide precise estimates for large areas, owing to their large sample sizes. However, they are labor-intensive and costly, which restricts their temporal resolution. Therefore, remote sensing technologies have become crucial for scaling woody AGB estimates. Among sensor types used for woody AGB modeling, active sensors like Radio Detection and Ranging (RADAR) and Light Detection and Ranging (LiDAR) have shown their capability for structural characterization of forests due to its ability to penetrate the canopy and capture three-dimensional vertical structure unmatched by passive sensor types8– 11. Within the LiDAR domain, platform selection involves a trade-off between spatial coverage and structural detail. While Aerial Laser Scanning (ALS) enables landscape-scale assessments, Terrestrial Laser Scanning (TLS) and Mobile Laser Scanning (MLS) provide highly  \ndetailed single-tree measurements but are limited in spatial extent. Bridging this gap, Unmanned Aerial Laser Scanning (ULS) has emerged as a cost-effective approach for capturing three-dimensional forest structure at the individual-tree level across medium-sized areas.  \nThe most widely adopted method for LiDAR-based biomass modeling is the plot-level approach. In this approach, statistical models are developed to relate plot-level AGB to plot-level metrics derived from the LiDAR point cloud10,12,13 . The used LiDAR metrics typically include manually defined statistical summaries of height distributions and canopy density, which are selected for their strong correlation with standing biomass13–16. Despite the success of the area-based approach, the reliance on manually engineered features can limit model generalizability and fail to capt","cbCaiaoRIyLwVsUx","https://ap.wps.com/l/cbCaiaoRIyLwVsUx","pdf",1019971,1,11,"English","en",105,"# Abstract\n# Introduction","[{\"question\":\"What limitation do discrete plot-level inventory aggregates introduce in small plots?\",\"answer\":\"They create boundary-effect uncertainties because biomass spatial variability within a plot is ignored and edge trees extend across plot boundaries, adding noise to training data and degrading model performance.\"},{\"question\":\"How is the Horizontal Biomass Distribution (HBD) reference constructed?\",\"answer\":\"HBD is continuously mapped from Quantitative Structure Models (QSMs), providing a spatially explicit reference instead of assuming uniform distribution over the plot.\"},{\"question\":\"How does the QSM/HBD reference affect deep learning performance as plot size changes?\",\"answer\":\"QSM-based models systematically outperform forest-inventory (FI) approaches for smaller plot sizes; for 100 m² plots, using the HBD reference reduces RRMSE by 16.84 ± 4.37% compared with the FI baseline.\"}]",1784183998,28,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"shifting-from-discrete-to-continuous-reference-data-qsm-derived-horizontal-tree-biomass-distribution-for-deep-learning-biomass-estimation","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/shifting-from-discrete-to-continuous-reference-data-qsm-derived-horizontal-tree-biomass-distribution-for-deep-learning-biomass-estimation/82924/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-24","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What limitation do discrete plot-level inventory aggregates introduce in small plots?","Question",{"text":75,"@type":76},"They create boundary-effect uncertainties because biomass spatial variability within a plot is ignored and edge trees extend across plot boundaries, adding noise to training data and degrading model performance.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How is the Horizontal Biomass Distribution (HBD) reference constructed?",{"text":80,"@type":76},"HBD is continuously mapped from Quantitative Structure Models (QSMs), providing a spatially explicit reference instead of assuming uniform distribution over the plot.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the QSM/HBD reference affect deep learning performance as plot size changes?",{"text":84,"@type":76},"QSM-based models systematically outperform forest-inventory (FI) approaches for smaller plot sizes; for 100 m² plots, using the HBD reference reduces RRMSE by 16.84 ± 4.37% compared with the FI baseline.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]