Effects of changing climate on reference crop evapotranspiration over 1961–2013 in Xinjiang, China

To know the importance of different climate variables on reference crop evapotranspiration (ETo), a step-by-step sensitivity analysis of ETo to single, two, and multi-climate variables (C) was conducted. ETo in north, south, and entire Xinjiang Province, China, over 1961–2013 was estimated using the Penman-Monteith equation. Trends in the involved six Cs (i.e., minimum temperature—Tmin, average temperature—Tave, maximum temperature—Tmax, wind speed at 2 m—U2, sunshine hour—n, and relative humidity—RH) were detected by the modified Mann-Kendall test. Nineteen scenarios of changed Cs were preset to obtain recalculated ETo values considering the actual trend in each C and the Pearson’s correlation relationship between ETo and Cs. The results showed that ETo was mostly sensitive to Tmax, U2, and n. Sensitivity of ETo to the two overlapped changes of Tmin and Tmax caused larger increases in ETo than Tmax and Tave, Tave and Tmax, Tmax and (−n), Tmax and RH, Tmax and (−U2), and Tmin and Tave, but the overlapped changes (−U2) and (−n) caused larger decreases in ETo than the other two C scenarios. The simultaneously increased Tmax, Tmin, Tave, and RH plus decreased U2 and n had caused the actual decreases in ETo in Xinjiang. In general, the effects of decreased U2 and n on decreasing ETo compensated the effects of increased Tmax on decreasing ETo in Xinjiang.