Evaluation of various heat-pulse methods for estimation of sap flow in orchard trees: comparison with micrometeorological estimates of evaporation

A microcomputer-controlled heat-pulse system for the measurement of sap velocity in trees is described. Several published methods for determining sap velocity from the temperature rise measured either above or below a heater inserted into the stem were compared and evaluated. All methods could be improved by the use of curve-fitting procedures, with a particularly useful approach involving direct estimation of the parameters of the diffusion equation using the non-linear curve-fitting package maximum likelihood programme. An alternative approach that was based on measurement of the value of the maximum temperature was proposed and tested. This was found to be particularly robust and sensitive to changes in flow rate. Although sap-flow velocity varied markedly with depth in the stem (as shown by the rate of dye movement), the maximum temperature at any given flow rate was found to be relatively insensitive to sensor depth. Estimated sap-flow velocities were compared with evaporation rates estimated either by weighing (for potted trees) or by the Penman-Monteith equation. Several independent methods were used for estimating the values of boundary layer resistance and net radiation that were required for application of the Penman-Monteith equation. There was a close relationship between flow and evaporation with some evidence for hysteresis. Although absolute calibration of the sap-flow estimates is difficult, the methods described are all useful for relative studies and all responded rapidly to altered flow rates caused by changing weather conditions.