Abstract Traditional field-based site index determination (expected tree height at, e.g. 25, 50, or 100 years) is labor-intensive, time-consuming, and costly. Recognizing the immense potential of airborne light detection and ranging (LiDAR) in forest inventory and monitoring, we evaluated LiDAR as a more efficient approach to characterize the site index. We evaluated two LiDAR-based site index approaches, including a LiDAR-only approach (no field plots) using repeated LiDAR acquisitions and a modeled approach using a field-based age and single-date LiDAR height. The multi-date LiDAR approach without field plots uses the change in canopy height (95th percentile height) between two LiDAR collections to identify an appropriate site index curve. In the single-date LiDAR approach, field-measured plot ages are paired with corresponding LiDAR heights to identify an appropriate site curve. LiDAR heights were bias-adjusted against field-measured canopy heights before site index estimation. Our study was conducted for two Douglas-fir (Pseudotsuga menziesii)–dominated sites, one in southwest Oregon and another in northwestern California. Multi-date and single-date LiDAR site index approaches performed similarly on the Oregon site (12% vs. 13% RMSE, respectively). However, the multi-date LiDAR site index approach fared 4 percentage points worse than the single-date LiDAR site index approach in California (13% vs. 9%, respectively). Despite the slight difference in performances for California, these findings suggest that both LiDAR approaches to site index are operationally viable for use in forest inventory and monitoring in the Western USA. The LiDAR-based site index has the advantage that it can be mapped wall-to-wall at fine spatial scales (e.g. 5–20 m) over large areas (e.g. 10s–1000s of km2).
Aryal et al. (Thu,) studied this question.