Modeling Grapevine Water Use
Data were collected on Thompson Seedless grapevines grown in a weighing lysimeter to refine a model of vine water use and then compare those results with actual water use. The model was dependent upon the interception of sunlight on each of the three canopy surfaces (top, north and south; row direction east to west) and the relationship between light and stomatal conductance. The canopy also was divided into sunlit and shaded leaf areas. Measures of total vine leaf area, shaded area and leaf area index (LAI) were also measured several times during the growing season. Solar and net radiation, vapor pressure and wind speed were measured above the canopy of the vines within the lysimeter. Once the necessary environmental data had been collected on a specific day and canopy conductance (gc) had been modeled, canopy resistance (l/gc) and aerodynamic resistances were inserted into a resistance-energy balance equation to calculate vine water use. There was a linear relationship between total leaf area per vine and shaded area per vine. As leaf area increased from 2 to 25 m2 per vine, shaded area increased from 0.8 to just less than 5 m2 per vine. The LAI (m2 leaf area m”2 shaded surface area) of the vines in the lysimeter increased up until total vine leaf area was 15 m2 per vine at which time it leveled off (at approximately 5.6 m2 m”2) with further increases in whole vine leaf area. There were linear relationships (r ‘s greater than 0.9) between vine water use and shaded area per vine and %shaded area per unit land area. In addition, the crop coefficient (kc) and total vine leaf area also were linearly related to vine water use (i^’s approximately 0.88). Modeled water use and actual water use on three dates (August 4 and 12 and September 6) late in the growing season were compared with one another. The two earlier dates were chosen as vine leaf area differed from August 4 to 12 due to cane cutting on August 8. Modeled and actual water use on August 4 was 12.3 and 12.28 mm, respectively. The model overestimated water use on 12 August and 6 September by 5.7 and 7.8%, respectively. On all dates tested, the model underestimated water use early in the morning and overestimated water use late in the afternoon. To obtain the above results, it was determined that canopy conductance was responding to VPD and therefore, gc was made a function of VPD (when the VPD was greater than 2.0 kPa.). Due to the lateness in which monies to conduct this study were received, water use was not modeled earlier in the growing season because the appropriate measures of canopy light interception had not been measured. The results collected in 1998 are promising but indicate that further refinement is needed both at the leaf and canopy levels.