Estimation of Solar Radiation in Forests Using Drone LiDAR Data in Japanese Artificial Forests

Many Japanese cypress plantation forests have been degraded due to inadequate forest management, resulting in low solar radiation on the forest floor. In recent years to increase forest floor radiation, however, assessment methods for the impact of thinning have not been developed.　The light environment in forests has been estimated by Hemispheric photography with a fisheye lens camera and using image analysis software Hemisfer to determine the amount of solar radiation in the forest and canopy openness situation. To precisely recreate the actual canopy structure, it is challenging due to the limitations of fisheye lenses, which cause distortions as the distance from the center of the captured area grows and the projected area decreases. In addition, taking Hemispherical photos in the forest is labor and time intensive work.

In this study, to explore the forest light environment in cypress plantations, we not only performed image analysis of hemispheric photography, but also estimated forest solar radiation using drone LiDAR data. The study site was a cypress plantation forest located in Mt.Karasawa, Sano City, Tochigi Prefecture. The site is a south-facing slope with a slope angle of approximately 30 degrees. 25 pyranometers were set up in the forest in a grid pattern with 1-meter intervals to measure the spatial distribution of solar radiation in the forest. Total solar radiation was measured by a radiometer installed outside the forest. For Hemispherical image analysis, the software Hemisfer was used to calculate direct and diffused solar radiation in the forest. The drone generated high-density point cloud data with a point cloud density of 2000 pts/m² was converted to 1cm³ voxel data first, then canopy openness was calculated by clipping the area directly above each pyranometer into a cylindrical buffer and calculating the percentage of the total number to the canopy points number. Considering the significant effect of direct sunlight in forest solar radiation, we varied the solar height in 5° intervals to simulate the actual angle of sunlight penetrating the tree canopy and recalculated the openness.

While the Hemispheric photography did not capture the detailed solar radiation changes in the measured data, the UAV LIDAR data succeeded in reproducing solar radiation changes closer to the measured data by considering the canopy openness. Therefore, there is a possibility of more accurate estimation by using LiDAR data together.