Influence of natural temperature gradients on measurements of xylem sap flow with thermal dissipation probes. 2. Advantages and calibration of a noncontinuous heating system

Natural temperature gradients in stems of trees growing in open stands give rise to errors when measuring sap flow by the continuous thermal dissipation method. Previously, we obtained evidence from field measurements that a noncontinuous thermal dissipation system can prevent these errors (Do and Rocheteau 2002). Cyclic heating (i.e., 45 minutes on and 15 minutes off; 45/15) allowed the derivation of an alternate signal, defined as the difference between the temperature signal at the end of the heating period and the temperature signal at the end of the cooling period. By analogy with the continuous system, we calculated an alternate flow index. Analysis in an artificial flow system confirmed that external temperature gradients have an additive effect on the continuous signal, whereas the alternate signal is unaffected by such gradients. The response of the alternate flow index to flow densities was similar for five combinations of heating and cooling times (45/15,40/20,30/30,15/15 and 10/10 min). The relationship was markedly different from Granier's (1985) calibration because measurements in cyclic systems are made under non-steady-state temperature conditions. We recommend the 15115 min cycle, which allows two sap flow measurements per hour. We compared flow density estimates obtained from field measurements with the continuous and cyclic systems over 192 days with 1-day lags between systems. Comparisons based on daily maximum values (between 0.5 and 2.51 dm(-2) h(-1) for the cyclic system) confirmed that differences between the continuous and cyclic systems, which could be greater than 100%, were linked to the effect of temperature gradients on the continuous system. The results demonstrate that, in situations where significant natural temperature gradients (i.e., > 0.2 degreesC) are likely, the cyclic system improves the accuracy of sap flow measurements made with thermal dissipation probes.