Modeling solar radiation attenuation and reflection within canopy using integrated point clouds between UAV-LiDAR and handheld LiDAR

Accurate leaf area density (LAD) is essential for radiative transfer simulations of forests, yet measurement methods are often limited to terrestrial LiDAR scanning (TLS). Numerous measurements are required to minimize occlusion; however, upper-canopy measurements are insufficient. Recently, new measurement methods such as handheld LiDAR scanning (HLS) and UAV-LiDAR have demonstrated the potential to reduce occlusion by enabling mobile data collection. However, the data acquired by UAV-LiDAR and HLS have not been directly used for radiative transfer calculations because their measurement accuracy is inferior to that of TLS, primarily because of the limitations of the inertial monitoring unit accuracy. The aim of this paper is to propose a radiance simulation model that considers the attenuation and reflection of direct irradiance within a forest canopy, utilizing LiDAR measurements. In the proposed approach, the point cloud is first converted into voxels. Then, the attenuation of the irradiance reaching an arbitrary voxel was calculated using the Beer-Lambert law with the point cloud between the light source and voxel as the input value. Finally, the radiance reflected in the direction of the sensor was calculated. The proposed method was applied to point clouds of a larch forest obtained through UAV-LiDAR and HLS, as well as to a combined dataset. The reliability of the proposed method was evaluated using the correlation coefficient (r) for the Sentinel-2 top of the atmosphere product red band reflectance. The results indicated that the r calculated from the radiance based on the UAV-LiDAR and coupled point cloud exceeded 0.75, indicating the validity of the method.