The accuracy of the thermal dissipation technique for estimating sap flow is affected by the radial distribution of conduit diameter and density

Abstract

There are conflicting reports on the accuracy of the thermal dissipation probe (TDP, the Granier method) measurement using the original formula, which is widely used to estimate the transpiration of individual trees and forest stands. In this article, six woody species of three wood types were used to study a possible association between TDP measurement accuracy and wood anatomical characteristics, including the vessel diameter and density, as well as sapwood depth. We found that TDP technique with Granier’s original equation underestimated the sap flux density in six species to various degrees, dependent on conduit size and sap flux. Our calibration using two conifers with small diameters and a high density of tracheids was relatively consistent with Granier’s calibration; however, because there were larger diameters and lower densities of vessels in the two diffuse-porous species, the original calibration significantly underestimated sap flow. Two ring-porous species had the largest diameters and lowest densities of vessels. In particular, Robinia pseudoacacia possessed the shallowest sap wood depth, less than a probe length. Our calibration for the ring-porous species, especially R. pseudoacacia, deviated far from the original calibration, which mostly underestimated the sap flow. The degree of underestimation was well associated with sap wood depth and the radial diameter and density distribution of conduits. Our results demonstrated that a new calibration must be operated for each species together with the sapwood depth determination and more probes may be applied for one stem in the field to obtain the more accurate sap flux. In addition, we investigated the effects of different environmental temperature and perfusing fluid composition on the TDP-based sap flux measurement. We found that an environmental temperature reduction from 25 to 0 °C did not alter the values of the maximum temperature difference (ΔT_m) between a heated probe and a reference probe when there was no sap flow, verifying that ΔT_m measured at night can be used as a reference in daytime.