Sap-flow techniques are vital tools for quantifying tree transpiration and assessing water use at the stand level. Among various sap-flow measurement methods, the Granier-type thermal-dissipation method (TDM) is the most widely adopted globally due to its simple principle and low cost. However, its original calibration coefficients often underestimate high flow rates.

In a study published in New Phytologist, researchers from Xishuangbanna Tropical Botanical Garden (XTBG) of the Chinese Academy of Sciences and their collaborators demonstrated that the original calibration for the widely adopted TDM often significantly underestimates water flow, particularly in fast-transporting species like lianas and ring-porous trees, leading to inaccurate environmental data.

Researchers comprehensively studied 31 woody plant species including 18 diffuse-porous species, two ring-porous species, six palms, and five lianas, covering a wide spectrum of wood hydraulic properties. They not only systematically recalibrated the key parameters of the TDM model but also investigated multiple factors influencing model coefficients and their accuracy.

Researchers found that the calibration coefficients, i.e., scaling factors and exponents, varied significantly across species and were strongly influenced by wood anatomical properties, especially the vessel-lumen area fraction and hydraulic conductivity.

Statistical models and the maximum applied pressure during parameter calibration significantly influenced the recalibrated parameters. Crucially, plant hydraulic traits, especially the xylem-specific hydraulic conductivity and the proportion of vessel lumen area, were found to be major drivers of intra- and inter-species variation in model parameters.

Moreover, researchers found that using original coefficients for species with high hydraulic conductivity such as ring-porous trees and lianas resulted in a severe underestimation of sap-flow, and that uncertainties in scaling processes including sapwood-area and radial and azimuth effects also contributed significantly to the overall estimates.

By applying the recalibrated coefficients to estimate transpiration in a rubber plantation, researchers confirmed that the TDM model severely underestimated stand-level transpiration, while the new parameters yielded more reasonable estimates.

"Accurately estimating tree water uptake is a multi-step process. Our study demonstrates that for reliable results, we must move beyond the original calibration and account for both species-specific coefficients and the significant uncertainties involved in scaling measurements from a single sensor to an entire tree or forest," said CHEN Yajun from XTBG.