Testing a Ground-Based Canopy Model Using the Wind River Canopy Crane

Abstract

A ground-based canopy model that estimates the volume of occupied space in forest canopies was tested using the Wind River Canopy Crane. A total of 126 trees in a 0.25 ha area were measured from the ground and directly from a gondola suspended from the crane. The trees were located in a low elevation, old-growth forest in the southern Washington Cascades. The ground-based model was based on six measurements and assumptions about the individual crown shape (e.g., conic, parabolic), the crane-based measurements required up to 377 measurements per tree. The two models were then compared, both by species and by crown position, to see where major discrepancies occurred. At the stand scale, ground-based and crane-based models of canopy structure were similar. At the scale of individual trees, however, groundbased estimates of crown volume differed significantly from the more detailed models of crown shape afforded by direct canopy access with the crane. Douglas-fir crowns were overestimated by 10.6%, Pacific yews were overestimated by 0.8% and western hemlocks were underestimated by 1.9%. While errors for yew and hemlock were smaller than for Douglas-fir, their standard deviations are much higher: 0.09 for Douglas-fir and 0.13 and 0.12 for Pacific yew and western hemlock, respectively. Most of the error resulted from model estimates of the lower crown, as epicormic branching and uneven shading caused highly irregular lower crowns in Douglas-fir. Over 85% of the differences between the two models among all the Douglas-fir trees were in the lower halves of the crowns. Similarly, 74% of the hemlock error and 58% of the Pacific yew error resulted from differences in their lower crowns. At the stand-level, the ground-based model of crown volume and the vertical distribution of foliage provided estimates consistent with more the detailed measurements made using the canopy crane.