Effects of tree-to-tree and radial variations on sap flow estimates of transpiration in Japanese cedar

To scale sap flux density (F(d)) measurements of individual trees to stand level transpiration (E), the mean stand Fd (J(S)) is among the most important factors, because E is frequently estimated as a product of J(S) and stand sapwood area. Major uncertainties in J(S) estimations are caused by tree-to-tree and radial variations in F(d). Thus, by using the Monte Carlo sampling technique with original tree-to-tree Fd data measured at several depths, we examined how many trees and radial depths need to be sampled for calculation of appropriate J(S). In this study, Fd was measured in 14 mature Japanese cedar (Cryptomeria japonica) trees at several radial depths, 0-2, 2-4 and 4-6 cm, using Granier-type sensors. The results revealed that the potential error in J(S) with the largest available sample size was not stable even though radial profiles in F(d) for all sample trees were measured. They also showed that to consider radial variation in F(d) it might be necessary to sample at least six trees, but not necessarily more than this. The results imply that, for E estimations, there is more tree-to-tree variation in F(d) than in radial variation in F(d). (c) 2005 Elsevier B.V. All rights reserved.