Understory Terrain Estimation by Synergizing Ice, Cloud, and Land Elevation Satellite-2 and Multi-Source Remote Sensing Data

Abstract

Forest ecosystems are incredibly valuable, and understory terrain is crucial for estimating various forest structure parameters. As the demand for monitoring forest ecosystems increases, quickly and accurately understanding the spatial distribution patterns of understory terrain has become a new challenge. This study used ICESat-2 data as a reference and validation basis, integrating multi-source remote sensing data (including Landsat 8, ICESat-2, and SRTM) and applying machine learning methods to accurately estimate the sub-canopy topography of the study area. The results from the random forest model show a significant improvement in accuracy compared to traditional SRTM products, with an R2 of 0.99, ME of 0.22 m, RMSE of 3.59 m, and STD of 3.59 m. In addition, we assessed the accuracy of understory topography estimates for different landforms, canopy heights, forest cover types, and forest coverage. The results demonstrate that the estimation results are minimally impacted by ground elevation, forest cover type, and forest coverage, indicating good stability. This approach holds promise for accurately estimating understory terrain at regional and global scales, providing crucial support for monitoring and protecting forest ecosystems.