The effect of reference-condition-based temperature modification on the trend of reference evapotranspiration in arid and semi-arid regions

• Published in: Agricultural water management 2017-12, Vol.194, p.204-213
• Main Authors: Khanmohammadi, Neda, Rezaie, Hossein, Montaseri, Majid, Behmanesh, Javad
• Format: Article
• Language: English
• Subjects: ET0; FAO56 Penman-Monteith; Iran; Reference condition; Trend analysis
• ISSN: 0378-3774; 1873-2283
• DOI: 10.1016/j.agwat.2017.09.010

•Temperature data recorded in arid and semi-arid regions were modified on the basis of reference condition.•The reference evapotranspiration (ET0) was calculated using observed and modified temperature.•The trends of two calculated ET0 were different.•To calculate correct ET0 trend, the temperature data must be modified based on the reference condition. Reference evapotranspiration (ET0) has a significant role in agricultural water management. Commonly, FAO56 Penman-Monteith as standard method is used to calculate ET0. To apply this method, meteorological data recorded at reference stations are required. The use of non-reference stations data in ET0 calculation can affect trend of this parameter. In the present study, the effect of reference-condition-based temperature modification on ET0 trend was evaluated. For this purpose, the collected data from 27 non-reference meteorological stations during 1960–2014 were applied. The mentioned stations are located in arid or semi-arid regions of Iran. The trends of unchanged and modified values of ET0 in studied stations were calculated using the Mann-Kendall method. Trend results showed that unchanged values of ET0 had a positive trend, while the modified ones had a negative trend in more than 50% of stations at seasonal and annual scales. According to R2 and RMSE results, there was no correlation between the trends of unchanged and modified values of ET0 and temperature modification based on the reference condition can change the trend values of ET0. Also, the results of T-statistic showed that the difference between two values of trend in arid regions was more than that of semi-arid ones. This difference in stations with arid climate was significant at 95% confidence level.