Answers are grounded in this paper and cite their sources.
Accurate mapping of forest aboveground biomass (AGB) is essential for regional forest carbon assessment and ecosystem monitoring. Spaceborne LiDAR missions, including the Global Ecosystem Dynamics Investigation (GEDI) and the Ice, Cloud, and land Elevation Satellite-2 (ICESat-2), provide valuable observations of forest vertical structure and have become important data sources for large-scale AGB estimation. However, obtaining reliable AGB estimates over large areas remains challenging due to the sparse distribution of spaceborne LiDAR observations. To address this limitation, an adaptive fusion method of GEDI and ICESat-2 data was proposed to integrate sample-driven LiDAR interpolation with an optical AGB map for spatially continuous AGB mapping. Firstly, footprint-level AGB was estimated from GEDI and ICESat-2 observations using LiDAR metrics, Sentinel-2 predictors, and terrain variables. Secondly, a sample-driven interpolation scheme was applied to generate continuous AGB estimates from sparse LiDAR observations by jointly incorporating sensor reliability, spatial proximity, and feature-space similarity. Finally, an optical AGB map (OAM) was generated from Sentinel-2 and terrain variables, and then the interpolated AGB map (IAM) and the OAM were adaptively fused at the pixel level using local discrepancy and boundary strength to produce the final AGB map. Experimental results showed that adaptive fusion of OAM and IAM improved forest AGB mapping accuracy. Compared with OAM alone (R² = 0.53, RMSE = 24.17 Mg/ha), the fused AGB map achieved an R² of 0.68 and an RMSE of 18.39 Mg/ha. The results demonstrate the potential of the proposed framework for improving forest AGB mapping under sparse LiDAR sampling conditions.
No takes yet. Share an insight, caveat, or question.
Qi et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: