A comparison of four evapotranspiration models in a green house environment

Advances in evapotranspiration modeling have led to a variety of formulas and terms used in irrigation control research. Four models that represent the progression from outdoor, "big leaf" estimates to indoor, greenhouse-specific formulations are compared in this article. Each of the models was algebraically reduced to common and consistent terms. The data from a lysimeter study were used to compare evapotranspiration combination models using evapotranspiration (ET) rates of a 'Red Sunset' Red Maple tree (Acer rubrum 'Red Sunset) grown in a controlled-environment greenhouse. The measured ET was compared with two empirical climatic factors (solar irradiance and vapor pressure deficit) and with calculated ET based on four evapotranspiration models: (1) Penman, (2) Penman-Monteith, (3) Stanghellini, and (4) Fynn. A measure of the model performance is given by the Nash-Sutcliffe R-2. or model efficiency. The relationship between measured and calculated ET for the Stanghellini model was calculated to have a model efficiency of R-2 = 0.872, while the other models yielded correlations of R-2 = 0.214, R-2 = 0.481, R-2 = -0.848 for Penman, Penman-Monteith, and Fynn, respectively. The differences between the models are discussed, revealing the importance of the leaf area index factor and a sub-model for irradiance in the plant canopy. Based on the coefficient of determination, vapor pressure deficit alone yielded a good linear correlation to measured ET (r(2) = 0.884), while solar irradiance alone resulted in r(2) = 0.652.