The best alternative for estimating reference crop evapotranspiration in different sub-regions of mainland China

Reference crop evapotranspiration (ET (o)) is a critically important parameter for climatological, hydrological and agricultural management. The FAO56 Penman-Monteith (PM) equation has been recommended as the standardized ET (o) (ET (o,s)) equation, but it has a high requirements of climatic data. There is a practical need for finding a best alternative method to estimate ET (o) in the regions where full climatic data are lacking. A comprehensive comparison for the spatiotemporal variations, relative errors, standard deviations and Nash-Sutcliffe efficacy coefficients of monthly or annual ET (o,s) and ET (o,i) (i = 1, 2, …, 10) values estimated by 10 selected methods (i.e., Irmak et al., Makkink, Priestley-Taylor, Hargreaves-Samani, Droogers-Allen, Berti et al., Doorenbos-Pruitt, Wright and Valiantzas, respectively) using data at 552 sites over 1961–2013 in mainland China. The method proposed by Berti et al. (2014) was selected as the best alternative of FAO56-PM because it was simple in computation process, only utilized temperature data, had generally good accuracy in describing spatiotemporal characteristics of ET (o,s) in different sub-regions and mainland China, and correlated linearly to the FAO56-PM method very well. The parameters of the linear correlations between ET (o) of the two methods are calibrated for each site with the smallest determination of coefficient being 0.87.