Scholarlink

<p>Forest inventory practitioners require information at finer geographic scales than ever to provide information for assessing forest sustainability, carbon storage, and other policy and management needs. Reliable assessments of estimation uncertainties and timely delivery of information are also increasingly critical. Small area estimation (SAE) techniques have been successfully applied in forest inventory to address these needs, borrowing strength from broad scale inventory data paired with auxiliary information – often remote sensing – to improve the precision of sample-based estimates. This work presents a comparative analysis of SAE results using auxiliary canopy height information from two remote-sensing sources, both canopy height models (CHM), one derived from digital aerial photogrammetry and existing digital elevation models, the other from spaceborne lidar and multitemporal surface-reflectance observations. Area-level Fay-Herriot models were developed using field observations from the US National Forest Inventory (NFI) applied to three states in the southeastern US. Supplemental filtering of CHMs to align with the NFI’s definition of forest was explored. Precision was increased in county-level estimates of both forest wood volume and live tree aboveground biomass using either of the CHM products, with greater precision gains possible by excluding remote-sensing data from non-forested areas. Results demonstrated precision gains possible from incorporating state-wide CHM data from either aerial or spaceborne remote sensing platforms with NFI sample data to estimate timber volume and biomass for counties in three eastern states in the USA. Further work is needed to determine where synthetic model predictions without random effects might serve in place of composite Fay-Herriot estimators.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Comparing Canopy Height Models from Regional-Scale Aerial Photogrammetry with Global Spaceborne Lidar-Derived Data for Estimating Forest Volume and Biomass

  • 期刊论文

  • 发表日期：

    2025年5月27日

查看全文
  • Qianqian Cao,
  • Philip J. Radtke,
  • John W. Coulston,
  • Valerie A. Thomas,
  • Randolph H. Wynne,
  • David M. Walker

摘要

Forest inventory practitioners require information at finer geographic scales than ever to provide information for assessing forest sustainability, carbon storage, and other policy and management needs. Reliable assessments of estimation uncertainties and timely delivery of information are also increasingly critical. Small area estimation (SAE) techniques have been successfully applied in forest inventory to address these needs, borrowing strength from broad scale inventory data paired with auxiliary information – often remote sensing – to improve the precision of sample-based estimates. This work presents a comparative analysis of SAE results using auxiliary canopy height information from two remote-sensing sources, both canopy height models (CHM), one derived from digital aerial photogrammetry and existing digital elevation models, the other from spaceborne lidar and multitemporal surface-reflectance observations. Area-level Fay-Herriot models were developed using field observations from the US National Forest Inventory (NFI) applied to three states in the southeastern US. Supplemental filtering of CHMs to align with the NFI’s definition of forest was explored. Precision was increased in county-level estimates of both forest wood volume and live tree aboveground biomass using either of the CHM products, with greater precision gains possible by excluding remote-sensing data from non-forested areas. Results demonstrated precision gains possible from incorporating state-wide CHM data from either aerial or spaceborne remote sensing platforms with NFI sample data to estimate timber volume and biomass for counties in three eastern states in the USA. Further work is needed to determine where synthetic model predictions without random effects might serve in place of composite Fay-Herriot estimators.