Assessing alternative allometric algorithms for estimating leaf area of Douglas-fir trees and stands

Allometric approaches based only on tree diameter at breast height (DBH) ha ve been used to estimate leaf biomass and leaf area index (LAI) in conifero us forests. This approach has been suggested to sometimes produce unrealist ically high LAI estimates, notably in old-growth stands in the Pacific Nort hwest (PNW). Leaf area to sapwood area relationships offer an alternative b asis for estimating LAI and a number of empirical and modeling approaches h ave been used for generating tree-level sapwood area and leaf area estimate s. To evaluate effects of the alternative assumptions among these approache s, we compared five algorithms for estimating leaf area per tree and stand- level LAI in stands dominated by Douglas-fir (Pseudotsuga menziesii var men ziesii). Initially, 383 trees representing the major species and the comple te range of tree sizes in the study area were examined for species, DBH, cr own ratio, and sapwood area. There were significant species-specific relati onships between sapwood and DBH for these trees. Neither crown ratio nor th e ratio of observed tree height to expected tree height were effective in p redicting the sign and magnitude of the residuals in the sapwood area/DBH r elationships. Nine stands (three young, three mature, and three old-growth) in and near the H.J. Andrews Experimental Forest in the western central Or egon Cascades were then studied with species, DBH, and crown ratio recorded for all trees in a set of subplots. At the stand level, the LAT algorithms based on sapwood area at breast height produced significantly lower LAI es timates than the DBH-alone algorithm in mature and old-growth stands. The a lgorithm relying on estimated sapwood area at the base of the crown further reduced LAI estimates relative to those based on sapwood area at breast he ight. Although the differences between the DBH-based and sapwood area-based approaches tended to be greatest in the older stands, estimated LAI was hi gher in the older stands than in the younger stands for all algorithms. Com parative studies using allometric, litterfall, and optical approaches to LA I estimation are needed to resolve issues such as trends in LAI with succes sion and maximum possible LAI in Douglas-fir forests. (C) 2000 Elsevier Sci ence B.V. All rights reserved.