Influences of removing linear and nonlinear trends from climatic variables on temporal variations of annual reference crop evapotranspiration in Xinjiang, China.

Reference crop evapotranspiration (ET) is a key parameter in field irrigation scheduling, drought assessment and climate change research. ET uses key prescribed (or fixed or reference) land surface parameters for crops. The linear and nonlinear trends in different climatic variables (CVs) affect ET change. This research aims to reveal how ET responds after the related CVs were linearly and nonlinearly detrended over 1961-2013 in Xinjiang, China. The ET-related CVs included minimum (T), average (T), and maximum air temperatures (T), wind speed at 2m (U), relative humidity (RH) and sunshine hour (n). ET was calculated using the Penman-Monteith equation. A total of 29 ET scenarios, including the original scenario, 14 scenarios in Group I (ET was recalculated after removing linear trends from single or more CVs) and 14 scenarios in Group II (ET was recalculated after removing nonlinear trends from the CVs), were generated. The influence of U was stronger than influences of the other CVs on ET for both Groups I and II either in northern, southern or the entirety of Xinjiang. The weak influences of increased T, T and T on increasing ET were masked by the strong effects of decreased U &n and increased RH on decreasing ET. The effects of the trends in CVs, especially U, on changing ET were clearly shown. Without the general decreases of U, ET would have increased in the past 53years. Due to the non-monotone variations of the CVs and ET, the results of nonlinearly detrending CVs on changing ET in Group II should be more plausible than the results of linearly detrending CVs in Group I. The decreasing ET led to a general relief in drought, which was indicated by the recalculated aridity index. Therefore, there would be a slightly lower risk of water utilization in Xinjiang, China.