Reference Evapotranspiration in Gansu Province Calculated Using Different Methods

【Objective】 Reference evapotranspiration (ET0) measures the potential water loss from soils in terrestrial systems. It varies with a multitude of factors and can be calculated by different methods. In this paper, we compared ET0 calculated by five methods for regions with incomplete meteorological data across Gansu province in northwestern China. 【Method】 We divided the province into four regions: arid area in Hexi (Region I), semi-arid area in northern Longzhong (Region II), semi-humid area in southern Longzhong (Region III), and humid area in Longnan and Gannan (Region IV). The reference evapotranspiration in each region was calculated using the Hargreaves-Samani (Har), the FAO-24 Penman (24PM), the FAO 1948 Penman (48PM), the Makkink (Mak), and the Priestley Taylor (PT) methods, respectively. For each method, we calculated the daily ET0 based on data measured from 1980 to 2019 from 26 meteorological stations across the province. The accuracy and robustness of each method for each region was tested against that estimated using the FAO-56 penman Monteith (P-M) method. 【Result】 ET0 calculated from different methods varied, though the variation depended on region. The results calculated from the Har and Mak methods were correlated with that from the P-M method with the correlation coefficients close to unity, despite that the Har method was less accurate for high altitude regions. The results estimated by the 48PM and 24PM methods were both greater than the ET0 calculated by P-M method. The accuracy of the five simplified methods for estimating ET0 was ranked in the order of Mak > Har >48PM > 24PM > PT for Region I, and Har >Mak> 48PM > 24PM >PT for other three regions. 【Conclusion】 For regions that are lack of meteorological data, the Mak method was the best alternative for estimating ET0 in Region I, while the Har method worked better for other regions in Gansu province.