Performance of 12 reference evapotranspiration estimation methods compared with the Penman–Monteith method and the potential influences in northeast China

Abstract

Reference evapotranspiration (ET0) is an important factor for agricultural activity and water management. Its estimation methods have significant regional differences. The Penman–Monteith (P‐M) method, which is recommended by the Food and Agriculture Organization (FAO), has been widely accepted, but it requires a large amount of data. This study analysed the performance of 12 ET0 estimation methods in different sub‐regions and periods by comparing them with the FAO P‐M method. Northeastern China was divided into eight sub‐regions according to their terrains and climates. Meteorological data obtained from 126 stations in these sub‐regions between the 1950s and 2014 were used to calculate the ET0 values. The performances were identified via four indices: root mean squared error, mean absolute error, average ratio and Spearman's correlation co‐efficient. The optimal method for agricultural areas was then modified by using linear regression analysis. The results are as follows: the Valiantzas2, Romanenko2 and H‐Makkink methods are recommended as alternative methods during the crop‐growing period, whereas the Turc and Hargreaves–Samani methods can generate significant biases. The Valiantzas2 and H‐Makkink methods are the optimal ones for estimating ET0 values in agricultural areas. Based on the outputs from the FAO P‐M method, ET0 values are most sensitive to temperature, accounting for 59.5%. Other methods that provide similar results can be considered as alternatives. Finally, the modified equation can provide the most accurate results, but drought events also affect the accuracy of the modified method.