National-scale calibrated GEDI AGBD models for effective assessment of growth conditions across forest strata

Abstract

Forests play a critical role in global carbon storage, biodiversity conservation, and climate change mitigation. Japan, with 67 % of its land area covered by forests, is one of the world's most forested nations. NASA’s Global Ecosystem Dynamics Investigation (GEDI) mission has widely been demonstrated to accurately estimate aboveground biomass density (AGBD). GEDI relative height metrics (RHs), derived from vertical energy distributions, serve as key predictors for AGBD estimation while capturing the variability in the vertical structure of the observed area. This study collected widely distributed co-located airborne laser scanning and field measurement data across Japan. Using the GEDI simulator, noise-free simulated GEDI waveform data were generated to train the AGBD models for different forest strata, including canopy-level AGBD and total AGBD, as well as for various forest types, including needleleaf, broadleaf, and their combinations. All GEDI L2A data collected between 2019 and 2023 were processed and aggregated to generate a nationwide map of AGBD for Japanese forests at a spatial resolution of 0.01° (approximately 1.1 km in latitude and 0.90 km in longitude at Tokyo). The results showed strong agreement with Japan's 4th National Forest Inventory (NFI), with national-scale AGBD estimates averaging 195.81 Mg/ha, compared to 197.66 Mg/ha from the NFI. This consistency was also evident across different forest types: for needleleaf forests, our estimate of 213.02 Mg/ha closely matched the NFI value of 216.21 Mg/ha, while for broadleaf forests, our estimate of 185.43 Mg/ha was in close agreement with the NFI estimate of 180.31 Mg/ha. The mean canopy-level AGBD and under-canopy AGBD were estimated at 150.11 Mg/ha and 45.70 Mg/ha, respectively, with under-canopy AGBD accounting for 25.33 % of the total AGBD. Our results demonstrate that GEDI can accurately capture AGBD across different forest strata, emphasizing the potential of GEDI RHs for estimating vertical forest structures. This study underscores the importance of quantifying forest carbon stocks, which is critical for biodiversity conservation and sustainable forest management.