Estimating individual tree heights of sugi (Cryptomeria japonica D. Don) plantations in mountainous areas using small-footprint airborne LiDAR

Abstract

Recently, it was shown that individual tree heights could be accurately estimated using small-footprint airborne light detection and ranging (LiDAR) remote sensing. Because most of the areas studied previously were limited to flat terrain, we investigated the accuracy of LiDAR-derived individual tree height estimates for different types of topographical features in mountainous forests with a steeper and more complex topography. Several middle-aged (40–50 years old) sugi (Cryptomeria japonica D. Don) plantations are found in the mountainous regions in Japan; hence, we chose 48-year-old sugi plantations to investigate the accuracy of these estimates. The surveyed area was divided into three types of topographical features; steep slope (mean slope ± SD; 37.6° ± 5.8°), gentle slope (15.6° ± 3.7°), and gentle yet rough terrain (16.8° ± 7.8°). Before estimating tree heights, the number of detected trees within each topographical feature was researched. In each of these terrains, the percentage of trees detected correctly was 74%, 86%, and 92%; the average error between LiDAR-derived and field-measured tree heights was 0.227 m, −0.473 m, and −0.183 m; and the accuracy of the LiDAR-derived tree height estimates, given as root mean square error (RMSE), was 0.901 m, 0.846 m, and 0.576 m, respectively. Consequently, the procedure presented in this study could detect most canopy trees and estimate individual tree heights with an accuracy better than 1 m, even in a forest with a mean slope angle of approximately 38°; thus, indicating that small-footprint airborne LiDAR will be a useful tool for accurately estimating the heights of individual canopy trees in sugi plantations in mountainous areas.