Performance of Penman-Monteith FAO56 in a Semiarid Highland Environment Bogachan Benli 1 ; Adriana Bruggeman 2 ; Theib Oweis 3 ; and Haluk Üstün 4 Abstract: Reliable estimates of evapotranspiration are essential for irrigation and water resources planning and management. Although several methods are available for computing reference evapotranspiration ET o , the provision of complete and accurate climate data is often a problem. Therefore, weighing lysimeter data from a semiarid highland environment were used to evaluate the performance of six commonly used reference evapotranspiration estimation methods with different data requirements Penman-Monteith-FAO56, Priestley- Taylor, Radiation-FAO24, Hargreaves, Blaney-Criddle, Class A pan. The lysimeter experiments were conducted at Ankara Research Institute of Rural Services in Turkey, during the April–October cropping seasons of the years 2000–2002. The average ET o for the three seasons, computed from the lysimeter data, was 964 mm. The Penman-Monteith-FAO56 method was also evaluated for cases where relative humidity, wind speed, solar radiation, or all three parameters would be missing. This resulted in a total of 10 different methods. The RMS errors RMSE and index of agreement for the daily data and the monthly averages as well as the mean absolute error MAE for the seasonal totals were computed to compare these methods. The methods were ranked based on the sum of the ranks for all five evaluation criteria. The Penman-Montheith-FAO56 method with the full data set, with replacement of wind speed, and with replacement of relative humidity took the top three spots, with MAEs for the seasonal totals ranging between 40 and 70 mm. The Hargreaves method came in fourth MAE 54 mm, followed by the Penman-Montheith-FAO56 method with replacement of all three parameters MAE 57 mm. The RMSE for the monthly average ET o was 0.43 and 0.50 mm·days -1 for the Penman-Monteith-FAO56 without and with replacement of all three parameters and 0.48 mm· days -1 for Hargreaves. Thus, if only temperature data would be available, the much easier to use Hargreaves method would be preferred above the Penman-Montheith-FAO56 equation with replacement of humidity, radiation, and wind speed data, for this semiarid highland environment. DOI: 10.1061/ASCEIR.1943-4774.0000249 CE Database subject headings: Lysimeters; Data processing; Arid lands; Evapotranspiration; Irrigation; Water resources. Author keywords: Lysimeter; Hargreaves; Radiation-FAO24; Class A Pan-FAO24; Blaney-Criddle-FAO24; Priestley-Taylor; Missing data. Introduction The estimation of reference evapotranspiration ET o  from weather data has many practical applications in crop and water management. Irrigation planners rely on estimates of ET o , com- puted from meteorological data, when no direct measurements have been made with lysimeters or soil moisture monitoring de- vices. Numerous ET o equations have been developed and used by irrigation researchers and practitioners according to availability of historical and current weather data. One of the problems that have stirred continuous debate is the best choice of ET o estimation methods. The Food and Agricultural Organization of the United Nations FAO Land and Water Development Division has ac- tively pursued this issue during the last three decades. A revision of the procedures for estimating reference evapotranspiration was summarized in the FAO-24 report Doorenbos and Pruitt 1977. The writers recommended four methods depending on data availability—Penman, radiation, temperature, and pan evapora- tion. FAO established an expert consultation in 1990 to revise the methods proposed in FAO-24. One of the main recommendations by the expert consultation was to replace the FAO-24 Penman method with the Penman-Monteith method. Penman-Monteith gives more consistent ET o estimates and has been shown to per- form better than other ET o methods when compared with lysim- eter data in a range of environments Jensen et al. 1990; Chiew et al. 1995. The Penman-Monteith method for estimating reference evapotranspiration was published as FAO-56 by Allen et al. 1998. This method will here be referred to as PM-FAO56. Evaluation of evapotranspiration equations is generally done with lysimeters, with weighing lysimeters providing the most exact means of continuously measuring the evapotranspiration throughout the cropping period Wright 1991. In an analysis of lysimeter data from eight Mediterranean locations, Steduto et al. 1996 found that PM-FAO56 underestimated lysimeter evapo- transpiration at high rates. These writers indicated that the use of day-time wind speed, humidity, and canopy resistance, as op- 1 Project Manager, United Nations Development Programme UNDP, 2. Cad. No: 11, Cankaya 06610, Ankara, Turkey corresponding author. 2 Research Scientist, Energy, Environment, and Water Research Cen- ter, The Cyprus Institute, P.O. Box 27456, 1645 Nicosia, Cyprus. 3 Director, International Center for Agricultural Research in the Dry Areas ICARDA, P.B. 5466, Aleppo, Syria. 4 Director, International Agricultural Training Center, Istanbul, Yolu 3. Km. No: 36, Yenimahalle, Ankara, Turkey. Note. This manuscript was submitted on July 14, 2009; approved on April 9, 2010; published online on April 13, 2010. Discussion period open until April 1, 2011; separate discussions must be submitted for individual papers. This paper is part of the Journal of Irrigation and Drainage Engineering, Vol. 136, No. 11, November 1, 2010. ©ASCE, ISSN 0733-9437/2010/11-757–765/$25.00. JOURNAL OF IRRIGATION AND DRAINAGE ENGINEERING © ASCE / NOVEMBER 2010 / 757 Downloaded 19 Oct 2010 to 130.126.32.13. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org