Hydraulic conductivity-induced systematic parameter variation in a widely used thermal dissipation sap-flow technique

The Granier-type thermal-dissipation method (TDM) is the most widely used sap-flow technique. However, its original calibration coefficients often underestimate high flow rates, limiting their generality. We derived TDM coefficients (scaling factors and exponents) for 31 species, including 18 diffuse-porous, two ring-porous, six palms, and five lianas, representing a broad range of wood properties. Factors influencing the coefficients and their accuracy were also investigated. Furthermore, we compiled 119 published coefficients for 88 additional species covering seven major xylem types. Most recalibrated and published coefficient values were substantially different from the original values, particularly in ring-porous and liana species with high hydraulic conductivity. The coefficient values depend on the statistical models and the applied maximum pressure during the calibration process. Vessel-lumen area fraction and hydraulic conductivity explained the interspecific variation in two coefficients at both segment and species levels. We applied recalibrated and original coefficients to a rubber plantation and found that the original coefficients produced unreasonably low transpiration estimates, while recalibrated coefficients yielded reasonable values. Uncertainties in scaling processes, including sapwood-area and radial and azimuth effects, also contributed significantly to the overall estimates. Our study demonstrates that accurate transpiration estimation must also address scaling-related sources of variation, as they contribute equally to uncertainty as poor calibration.