Connecting spaceborne lidar with NFI networks: A method for improved estimation of forest structure and biomass

• DOI: 10.1016/j.jag.2024.103797
• Publication Date: 2024-04-05T04:37:55Z
• Authors: Paul B. May; Ralph O. Dubayah; Jamis M. Bruening; George C. Gaines
• Keywords: Environmental sciences; Lidar; Physical geography; Small area estimation; Spatial statistics; Forestry; GE1-350; 01 natural sciences; Regression; GB3-5030; 0105 earth and related environmental sciences

Spaceborne lidar provides a unique opportunity to supplement the field plot measurements of national forest inventories (NFIs) by providing dense vertical canopy structure. For full waveform instruments such as Global Ecosystem Dynamics Investigation (GEDI), take form reflected energy function height within an observed footprint. Many attributes cannot be directly computed from waveforms, and thus statistical models relating target must trained using data. Because discrete footprint samples produced spaceborne have little chance collocation with pre-existing plots, model calibration remains primary challenge. We leveraged density spatial correlation GEDI observations make predictions cumulative over 326,787 NFI plots distributed across contiguous United States. The yield probability distributions, giving not only predicted but quantification prediction uncertainty. product this work is data set comprised paired predictive distributions. waveforms their uncertainties can used train "error-in-variables" techniques that account for filter uncertainty in waveforms. This corresponding allow statistically rigorous training without requirements collocation, yielding downstream attribute are consistent estimates. demonstrate above ground biomass (AGBD) then use produce spatially complete 1 km map AGBD further compare our US estimates at 64,000 hectare scale, showing increase precision proportional collection almost 500,000 additional forested regions CONUS.