Operational determination of daily actual evapotranspiration of irrigated tomato crops under Mediterranean conditions by one-step and two-step models: Multiannual and local evaluations

Abstract

The performance of two operational models for calculating daily actual evapotranspiration (ET) was presented in this study. Both models used only climatic variables collected by standard weather stations. The analysis was performed during 2000, 2001 and 2006 on three irrigated varieties of tomato crop cultivated in three different plots located in the Province of Foggia (Southern Italy). The first model (AL model) was a two-step model proposed by Allen et al. (1998). The second model (RK model) was a one-step model proposed by Rana and Katerji (2009).The comparison was performed for three growth seasons (223 days) between daily values of evapotranspiration calculated with these two models and daily values of evapotranspiration measured by the eddy covariance technique, and it underlined different performances of the AL and RK models to estimate the evapotranspiration. In the AL model, the slope of the linear regression between measured and simulated values (0.80) was significantly different from one, and the intercept (1.06) was significantly different from zero. The coefficient of determination (r2) was equal to 0.52, and the standard error was equal to 1.1mm. In the RK model, the slope of the linear regression (1.05) was slightly greater than one, and the intercept was not significantly different from zero. The coefficient of determination (r2) was equal to 0.83, and the standard error was equal to 0.63mm.The variability year by year of the statistical parameters (slope, intercept, r2 and standard error) of the linear regression between measured and calculated daily evapotranspiration values calculated with the AL model underlined an important interannual variability. In the RK model, the slope and intercept of the linear regression showed stability from year to year. Only the determination coefficients and standard errors showed certain interannual variability, although the range of the found variations was reasonable when compared to the AL model.The reasons for the weak performance of the AL model in the Mediterranean region, which has also been confirmed by other authors, were analysed and discussed in detail in this study.Finally, the conclusion of this study highlighted the reliability of the one-step approach to calculate the water requirement of tomato crops and other irrigated species.