Sap flux density (SFD) measurements methods have been widely used in woody plants to accurately quantify tree water use (transpiration). Two different methods viz., heat ratio method (HR) and the compensation heat pulse (CHP) were tested against a pre-calibrated mini-weighing lysimeter (gravimetric method) that served as the control. The experiments were conducted on large potted Citrus sinensis (‘Midknight’ Valencia trees) in a glasshouse at the University of Pretoria. The main aim was to compare these methods and establish the most reliable and accurate method to be used to estimate transpiration in citrus trees. A strong positive linear relationship between the HR method and gravimetric method (R2 = 0.98) was obtained when a wound width of 2.0 mm (width of the widest probe) was used, with tree transpiration underestimated by an average of 1.08 % day-1. The CHP method was satisfactorily accurate as indicated by an R2 value of 0.91 and an overestimation of tree transpiration by an average of 1.23 % day-1. However, a large portion of missing data was one of the challenging factors to adopt the CHP method. The measurements also showed evidence of a time lag between sap flow measured with the CHP and HR method and transpiration measured with the gravimetric method, indicating some degree of capacitance in these potted trees. Although validation of the methods in potted citrus against a weighing lysimeter showed that both the HR and CHP methods accurately estimated transpiration, it was important to assess the methods in commercial orchards where long term measurements were to be made. The HR and CHP methods were tested in a commercial Citrus sinensis (‘Washington’ Navels) orchard for their ability to quantify transpiration under field conditions. Field trials were conducted on Patrysberg farm near Citrusdal from 10 - 14 March 2015. Micrometeorological measurements for the estimation of evapotranspiration (ET) were performed using an EC (eddy covariance) system and soil evaporation (Es) was determined using the microlysimeters. Transpiration (Tres) was calculated as a residual of ET and Es and compared with transpiration (Tsap) determined by the CHP and HR method. In comparison to Tres the HR method (R2 = 0.97) resulted in a better agreement than the CHP method (R2 = 0.80) when a wound width of 2.5 mm corresponding to the width of the widest probe was used. However, both the HR and CHP methods underestimated orchard transpiration by 42 % and 36 % respectively, when compared to Tres. The underestimation was mainly attributed to an underestimation of the wound width and possible additive errors, which includes the scaling of whole tree water use to orchard water use and errors in the determination of soil evaporation. Even though the SFD methods underestimated transpiration, the good agreement between the SFD methods and Tres enabled the calibration of SFD methods. As wound width is one of the most difficult parameters to determine for the upscalling of heat pulse velocities to sap flow in the field, wound width was used to calibrate the CHP and the HR methods. For the CHP method a virtual wound width of 3.6 mm resulted in the best agreement between Tres and Tsap (1.4% overestimation of transpiration), whilst for the HR method a virtual wound width of 4.4 mm resulted in the best agreement between the two measurement methods (0.4% underestimation). At the end of the measurements the actual wound width was determined and was on average 4.7 mm for HR method probe sets. In situ stem staining was also performed at this time to determine the sapwood depth and heartwood radius. Using these measurements the calculated Tsap underestimated Tres by 5 % on average per day. Provided accurate estimates of wound width and sapwood conducting area are obtained at the end measurement period the HR and CHP methods can be used with confidence in Citrus sinensis. However, it is advised that during the measurement period a calibration is performed against an independent measure of transpiration, as wound width and sapwood conducting area do vary between trees and are difficult to determine with a great degree of accuracy.
Dissertation (MSc (Agric))--University of Pretoria, 2018.