Comparing actual transpiration fluxes as measured at leaf-scale and calculated by a physically based agro-hydrological model

The primary purpose of this paper is to compare the actual transpiration rates from tomato crops, as measured at leaf scale and estimated by a macroscopic approach in an agro-hydrological model named FLOWS, under variable soil properties and water availability. To this aim, sixteen plots were cultivated with tomatoes in Metaponto, Southern Italy. Soil hydraulic properties (SHP) were obtained using a fast in situ characterisation method. Leaf-area index (LAI) was measured using a leaf-area metre. SHP and LAI were then used in the physically-based FLOWS, which allowed calculating the macroscopic transpiration rates, T-a,T-m. Single-leaf transpiration rates, T-a,T-l, and stomatal conductance, g(s,l), were measured in situ. For comparison with T-a,T-m, g(s,l) was upscaled by the Big-Leaf approach to canopy scale stomatal conductance, g(s,c), which was applied to the Penman-Monteith model to obtain the canopy-scale transpiration, T-a,T-c. Finally, multiple linear regression (MLR) was used to find the statistical relationship between T-a,T-m and T-a,T-c, and the SHP and g(s,c). Results showed that the macroscopic approach smooths the spatial variability of transpiration rates. T-a,T-c increased with the saturated water content, theta(s), and the slope of the water retention curve, n, while T-a,T-m decreased with increasing theta(s) and n. MLR improved significantly by introducing g(s,c) to predict T-a,T-m.