Evaluation and modification of potential evapotranspiration methods in Beijing, China

Abstract: In this study, seven widely used potential evapotranspiration (ETo) methods were evaluated by comparing with the FAO-56 Penman-Monteith method (PM method) to provide useful information for selecting appropriate ETo equations under data-limited condition in Beijing, China. Statistical methods and parameters, namely linear regression, root mean squared error (RMSE) and mean bias error (MBE), were used to evaluate the seven ETo methods. Results showed that ETo estimated using Kimberly–Penman method have fairly close agreement with the PM method (referring to standard ETo), considering the coefficient of determination (R2) of 0.96, RMSE of 0.42 mm/day, and a coefficient of efficiency (E) of 0.96. Locally calibrated Penman and Doorenbos–Pruitt methods also have better agreement with the PM method, correspondingly with R2 of 0.99 and 0.95, RMSEs of 0.24 mm/day and 0.21 mm/day, and coefficients of efficiency of 1.02 and 0.99, respectively. The ETo is the most sensitive to vapor pressure deficit (VPD) and net radiation in the Beijing area. Hence, the VPD-based and VPD-radiation combined ETo methods were developed and calibrated. Results showed that the two developed methods performed well in ETo estimation. By fully considering the data-limit situation, the calibrated Turc method, VPD-based method and VPD-radiation-combined method may be attractive alternatives to the more complex Penman−Monteith method in Beijing.