Estimating Tree Growth Using Crown Metrics Derived from LiDAR Data

In order to determine current tree condition and predict future growth using LiDAR data, tree height, diameter at breast height, diameter 4 m above the ground, tree volume, tree volume growth, diameter at breast height growth and diameter growth 4 m above ground for individual trees were estimated from various crown height metrics and measurements obtained using a small footprint airborne laser scanner flown over a planted forest in Japan. Ground-truth values for tree height, diameter at breast height, diameter 4 m above ground, tree volume, and volume and diameter growth were collected. The actual values were compared with the laser-derived crown height metrics, including: percentiles, maximum, mean, coefficient of variation and crown density, all for the first and last crown height laser pulses. The regressions explained 75-79 % of the variability in ground-truth tree height, diameter at breast height, diameter 4 m above ground and tree volume. Cross-validation of the regressions revealed standard deviations of the differences between predicted and ground-truth values of 1.30 m (6.7 %), 5.2 cm (22.2 %), 3.8 cm (18.7 %) and 0.22 m(3) (43.3 % of ground-truth mean), respectively. The regressions also explained 69-77 % of the variability in ground-truth averages. Cross-validation of the regressions revealed standard deviations of the differences between predicted and ground-truth values of 0.15 cm yr(-1) (43.7 %), 0.1 cm yr(-1) (31.0 %) and 0.008 m(3) yr(-1) (58.5 % of ground-truth mean), respectively. The study confirms that it may be possible to predict individual tree growth based on LiDAR data.