Comparison of Artificial Neural Network and Physically Based Models for Estimating of Reference Evapotranspiration in Greenhouse

Evapotranspiration (ET) is one of the major components of hydrologic cycle. Accurate estimation of this parameter is essential for studies such as water balance, irrigation system design and management, and water resources management. Generally we used climate data for calculating evapotranspiration from indirect methods. This study investigates the utility of artificial neural networks (ANNs) for estimation of daily grass reference crop evapotranspiration (ET0) and compares the performance of ANNs with the conventional methods (Penman, Penman-Monteith, Stanghellini and Fynn) used to estimate ETo in Greenhouse. In the present study, the meteorological variables including air temperature, solar radiation, wind speed and relative humidity were considered daily. The daily outputs from on four physical ET and artificial neural networks have been tested against reference evapotranspiration data computed by the lysimeter to assess the accuracy of each method in estimating grass reference evapotranspiration in greenhouse. The accuracy of ANNs is the best but the accuracy of the Penman equation is the worse for estimating daily evapotranspiration compared with the other equations. In ranking the equations, Stanghellini equation, Penman–Monteith and Fynn, equation ranked in a second, third and forth places, respectively. The results showed the ANNs, Penman and P-M models overestimated ET, while the Fynn and Stanghellini models underestimated ET. The efficiency values of Penman, Fynn, P-M, Stanghellini and ANNs were 0.68, 0.72, 0.86, 0.907 and 0.93 respectively.