RESUMEN
Accurate determination of sap flow over a wide measurement range is important for assessing tree transpiration. However, this is difficult to achieve by using a single heat pulse method. Recent attempts have been made to combine multiple heat pulse methods and have successfully increased the sap flow measurement range. However, relative performance of different dual methods has not yet been addressed, and selection of the numerical threshold used to switch between methods has not been verified among different dual methods. This paper evaluates three different dual methods with respect to measurement range, precision and sources of uncertainty: (method 1) the heat ratio (HR) and compensation heat pulse method; (method 2) the HR and T-max method; and (method 3) the HR and double ratio method. Field experiments showed that methods 1, 2 with three needles and 3 compare well with the benchmark Sapflow+ method, having root mean square deviations of 4.7 cm h-1, 3.0 cm h-1 and 2.4 cm h-1, respectively. The three dual methods are equivalent in accuracy (P > 0.05). Moreover, all dual methods can satisfactorily measure reverse, low and medium heat pulse velocities. However, for high velocities (>100 cm h-1), the HR + T-max (method 2) performed better than the other methods. Another advantage is that this method has a three- instead of four-needle probe configuration, making it less error prone to probe misalignment and plant wounding. All dual methods in this study use the HR method for calculating low to medium flow and a different method for calculating high flow. The optimal threshold for switching from HR to another method is HR's maximum flow, which can be accurately determined from the Péclet number. This study therefore provides guidance for an optimal selection of methods for quantification of sap flow over a wide measurement range.