Answers are grounded in this paper and cite their sources.
Accurate estimation of forest canopy height is critical for understanding forest dynamics and ecosystem functions in complex mountainous regions. Traditional optical remote sensing methods are widely used for estimating forest canopy height at large scales. However, challenges persist in vast mountainous areas, where rugged terrain and dense, high-biomass forest cover impede both human access and field survey. This study integrates Ice, Cloud, and Land Elevation Satellite-2 (ICESat-2) LiDAR, Sentinel-1 microwave, and Sentinel-2 optical data within a random forest machine learning framework, and a scale generalization methodology is applied to estimate canopy height in the alpine forests of Morihe, China. We conclude the following: (a) The ICESat-2 data were calibrated by unmanned-aerial-vehicle-based observations; the 95th percentile of relative height from ICESat-2 displays the strongest correlation ( r = 0.80) among all measured parameters. (b) Incorporating multisource data, including microwave variables (VV and VH) and topographic variables (altitude and slope), improves model performance relative to that of the optical-only model, with R 2 increasing from 0.77 to 0.83 and the root mean square error and mean absolute error decreasing by 0.21 and 0.10, respectively. (c) Compared with existing products, our product was better suited for canopy height estimation in complex mountainous areas, with R 2 increasing from 0.34 to 0.83 and the root mean square error and mean absolute error decreasing by 0.63 and 0.52, respectively. Overall, multisource remote sensing data were used to produce wall-to-wall canopy height maps at a 10-m resolution, which reliably capture height distributions in complex mountainous areas, demonstrating the advantage of data fusion for enhancing mapping accuracy and ecosystem monitoring.
No takes yet. Share an insight, caveat, or question.
Song et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: