Improved reference evapotranspiration methods for regional irrigation water demand estimation

Abstract

Temperature- and radiation-based methods are widely used for reference evapotranspiration (ETo) estimates and in regional irrigation water demand studies, especially in areas with limited weather data. A major limitation is that without local calibration, the constant parameters used in these methods usually cannot ensure the same reliability in regional analysis. On the other hand, the controlling variables of ETo under different climatic conditions and how they are related to the structures of these methods have not been adequately addressed yet. Herein, we evaluated the performance of three commonly used temperature-based methods, i.e., Thornthwaite (Th_T), Blaney-Criddle (BC_T), and Hargreaves and Samani (HS_T), and five radiation-based methods, i.e., Makkink (Ma_R), Priestley-Taylor (PT_R), Jensen and Haise (JH_R), Turc (Tu_R), and Abtew (Ab_R), for regional ETo estimates and calibrated them using the Penman-Monteith method. Monthly meteorological data (1961–2010) from 15 weather stations in Texas, United States, covering humid, subhumid, semiarid, and arid climates, were used. The HS_T method is recommended for regional analysis if reliable wind speed and relative humidity data are not available. The radiative component was the driving factor of ETo, accounting for 72− 84% of its variation in the warm season in all climates. Therefore, temperature- and radiation-based methods performed well and showed little variation in the warm season. However, we demonstrated that the vapor pressure deficit impact on ETo variations could be as high as 43–49% in the cool season in subhumid, semiarid, and arid climates, and wind speed impact on ETo variations was 4–12% and 10–16% in humid, semiarid, and arid climates, respectively. To accurately reflect these impacts, we adjusted the HS_T and PT_R methods by multiplying a calibration-free coefficient. By doing so, the accuracy of ETo under cool, dry, and windy conditions is considerably improved compared with the locally calibrated methods. The adjusted HS_T and PT_R methods can significantly improve the efficiency and accuracy of applying the less data-intensive ETo methods to predict regional or global irrigation water demand.