Parameterizing canopy resistance using mechanistic and semi-empirical estimates of hourly evapotranspiration: critical evaluation for irrigated crops in the Mediterranean

In this paper two models are presented for calculating the hourly evapotranspiration lambda E (W m(-2)) using the Penman-Monteith equation. These models were tested on four irrigated crops (grass, soya bean, sweet sorghum and vineyard), with heights between 0.1 and 2.2 m at the adult growth stage. In the first model (Katerji N, Perrier A. 1983. Modelisation de l'evapotranspiration reelle ETR d'une parcelle de luzerne : role d'un coefficient cultural. Agronomie 3(6): 513-521, KP model), the canopy resistance r(c) is parameterized by a semi-empirical approach. In the second model (Todorovic M. 1999. Single-layer evapotranspiration model with variable canopy resistance. Journal of Irrigation and Drainage Engineering-ASCE 125: 235-245, TD model), the resistance r(c) is parameterized by a mechanistic model. These two approaches are critically analysed with respect to the underlying hypotheses and the limitations of their practical application. In the case of the KP model, the mean slope between measured and calculated values of lambda E was 1.01 +/- 0.6 and the relative correlation coefficients r(2) ranged between 0.8 and 0.93. The observed differences in slopes, between 0.96 and 1.07, were not associated with the crop height. This model seemed to be applicable to all the crops examined. In the case of the TD model, the observed slope between measured and calculated values of lambda E for the grass canopy was 0.79. For the other crops, it varied between 1.24 and 1.34. In all the situations examined, the values of r(2) ranged between 0.73 and 0.92. The TD model underestimated lambda E in the case of grass and overestimated it in the cases of the other three crops. The under- or overestimation of lambda E in the TD model were due: (i) to some inaccuracies in the theory of this model, (ii) to not taking into account the effect of aerodynamic resistance r(a) in the canopy resistance modelling. Therefore, the values of r(c) were under-or overestimated in consequence of mismatching the crop height. The high value of air vapour pressure deficit also contributed to the overestimation of lambda E, mainly for the tallest crop. The results clarify aspects of the scientific controversy in the literature about the mechanistic and semi-empirical approaches for estimating lambda E. From the practical point of view the results also present ways for identifying the most appropriate approach for the experimental situations encountered. Copyright (C) 2010 John Wiley & Sons, Ltd.