Are calibrations of sap flow measurements based on thermal dissipation needed for each sample in Japanese cedar and cypress trees?

Key message: As the estimated parameters differed across samples, even from the same site, sample-based calibration is the recommended procedure. A trait-based approach (i.e., the use of structural parameters of the trees) would be an alternative procedure. Abstract: The thermal dissipation method (TDM) is widely used for estimating transpiration by individual trees or stands. Although the importance of TDM calibration experiments is widely recognized, there is still no consensus on whether the calibration should be undertaken in practice for each species, site, or tree sample. The primary reason is that intraspecific variations in the fitting parameters have not been well examined in multiple sites. To address this, we performed TDM calibration experiments using 24 Cryptomeria japonica and Chamaecyparis obtusa samples collected from six regions in Japan and Taiwan. The sap flux density (F_d) based on the original TDM parameters was underestimated for most samples. Using a common set of parameters for 21 samples reduced the systematic underestimation. In addition, root mean square error (RMSE) was reduced by 44%. Site- and sample-based calibration reduced the RMSE by 69% and 75%, respectively. The estimated parameters for the samples varied, even among samples obtained from the same site. The recommended procedure is to obtain sample-specific parameters by performing a calibration experiment after measuring F_d under the target conditions. An alternative procedure is to use the parameters for other trees of the same species at the same site. Further, we discovered that one of the two parameters determined for each sample significantly correlated with the diameter/age ratio and height of the corresponding tree. A trait-based approach for predicting calibration parameters based on associated tree features allows the parameters to be determined without the need for calibration experiments.