Evapotrairation is one of the major components of hydrologic cycle. Accurate estimation of this parameter is essential for the studies such as water balance, irrigation ystem design and management, and water resources management. A lysimeter study was conducted to determine the evapotrairation (ET) rates for grass crop under outdoor and indoor condition from 23 September 2007 to 22 September 2008 in Isfahan University of Technology. In this research three drainage lysimeter indoor and three weighing microlysimeters for indoor and outdoor the glasshouse were used to determine the reference evapotrairation. With the objective of evaluating the performance of pan evaporation in estimating the water consumption in greenhouse, class A pan and reduced pan installed inside the glasshouse and another class A pan was installed outside. The results showed that the mean monthly ET 0 value measured by the microlysimeters installed outside and inside the glasshouse were 139 mm and 103.8 mm respectively, and 108.22 mm for the drainage lysimeter inside the glasshouse. Because of the large area occupied by a drainage lysimeter and similar accuracy with weighing microlysimeter in the greenhouse, microlysimeter has been suggested to determine reference evapotrairation inside the greenhouses. In the present work, incoming radiation for greenhouse was found to be 39 percent of the incoming radiation outdoor. As a result of this ET 0 in the glasshouse was found to be 69 percent of outdoor ET 0 . Thus, greenhouse agriculture provides a way of increasing crop water use efficiency. The daily and mean monthly output from Stanghellini and Solar radiation based method have been tested against reference evapotrairation data computed by the lysimeter to assess the accuracy of each model in estimating grass reference evapotrairation in greenhouse. The accuracy of Stanghellini equation is the best but the accuracy of the solar radiation equation is the worse for estimating daily and mean monthly evapotrairation. The Stanghellini model underestimated the measured ET rates by 7 percent with R 2 value of 0.96. In order to evaluate the individual and combined effects of climate factors on the measured ET, linear multiple regression analyses were done in order to determine empirically based ET models. The results revealed it appeared that the empirical models did at least as well if not better than the Stanghellini model. This study showed that empirical ET models could be used to predict ET rates if there is a difficulty in running the Stanghellini model due to a lack of input parameters.
