Adaptation evaluation for reference evapotranspiration methods in different regions of Sichuan

Reference crop evapotranspiration (ET0) is an important parameter for water cycle and water balance. Accurate estimation of ET0 becomes vital in water resource evaluation and agricultural water resource utilization. Numerous methods have been proposed for estimating ET0, among which the Penman-Monteith (P-M) model recommended by Food and Agriculture Organization of the United Nations (FAO) in 1998 is the best one to compute ET0 at present due to its foundation in physics and comprehensively considering the relationships. FAO has accepted the P-M model as the standard equation for the estimation of ET0 since it provides the highest accurate results across the world no matter in an arid or humid environment. But the main problems on computing ET0 by the P-M model are its complicated nonlinear process and requirements of many climatic variables. Thus, there is an urgent need to develop a much simpler and more appropriate model in the areas with limited data. This paper compares the accuracy of the Hargreaves-Samani (HS) method, Makkink method, Priestley-Taylor (PT) method, Irmark-Allen (IA) method, Penman-Van Bavel (PVB) method and 48PM method. To obtain the best calculation methods and assess its adaptability, the Sichuan Province is divided into 4 regions (including the eastern basin (I), the moist mountainous area (II), the southwest valley area (III) and the western plateau zone (IV)), and the daily ET0 in 46 stations from 1954 to 2013 is calculated. Then, this paper takes the ET0 calculated by the P-M model as a standard, and makes a comparison among the 6 methods. The results show that all methods are obviously variable in different regions. The HS method, the PT method and the PVB method are more accurate while the 48 PM method, the IA method and the Makkink method have a larger error. Unfortunately, there is no method that performs good in all of 4 regions except the PT and PVB methods due to the different terrain or climate in different area. The HS method is relatively accurate in Area I and II which have a lower altitude; the 48PM method has a relative error more than 50% in the shallow hilly area and plateau area, while its relative error is lower in basin areas, such as -25%-11.1% in Batang and Dege, and 11.2%-37.5% in southeast area. Finally, we find that the best methods in the area of I and II are the HS method (root mean square error (RMSE) 0.58 mm/d, mean absolute error (MAE) 0.45 mm/d, mean relative error (MRE) 0.02) and the PVB method (RMSE 0.61 mm/d, MAE 0.39 mm/d, MRE-0.15), and the best method in the area of III and IV is the PT method (RMSE 0.55-0.6 mm/d, MAE 0.44-0.45 mm/d, MRE 0.02-0.06). Therefore, the HS method can be attained as the best one for calculating ET0 in the eastern basin, the PVB method can be the best one for the moist mountainous area, and the PT method can be similarly chosen in the southwest valley area and the western plateau zone in Sichuan. © 2016, Editorial Department of the Transactions of the Chinese Society of Agricultural Engineering. All right reserved.