Abstract: Based on the temperature data of 105 meteorological stations in Xinjiang from 1961 to 2018, using climate tendency rate, Mann-Kendall test, wavelet power spectrum analysis and inverse distance weighted（IDW） interpolation technology under ArcGIS10.7 environment, the change trend, mutation characteristics, change cycle and refined spatial-temporal distribution of annual average temperature in Xinjiang were explored. The results showed that the spatial distribution of annual average temperature in Xinjiang was uneven, and the overall spatial distribution was “high in southern Xinjiang, low in northern Xinjiang; high in plain and basin （Valley）, low in mountainous area”. In recent 58 years, except Kuqa and Aktao, the temperature in all parts of Xinjiang showed a fluctuating upward trend, but there were regional differences, and the linear warming rates were different. The lowest warming rate was 0.025 7 ℃/a in southern Xinjiang, the warming rate was 0.031 4 ℃/a in Tianshan Mountains, and the largest warming rate was 0.035 4 ℃/a in northern Xinjiang. The key growth areas were located in Hami, Tacheng, Ili and Altay. The annual average temperature increasing rate in Xinjiang decreased from north to south. The temperature increasing rate in the north was greater than that in the northeast, the northeast was greater than that in the middle, the middle was greater than that in the south, and the south was greater than that in the east. The temperature increasing rate in cold regions was faster than that in warm regions. The main periodic fluctuation of temperature in Xinjiang was about 2.6 years, and all of them pass the 95% red noise test. The strongest periodic signal of high temperature fluctuation was in northern Xinjiang, followed by Tianshan Mountains, and the smallest in southern Xinjiang. Before the mid-1990s, except for a few years, the temperature anomaly was negative, and after the mid-1990s, it was positive. The 2~8a scale average spectrum showed that the interannual variation of temperature was weak before the mid-1990s, and gradually increased after the mid-1990s. The corresponding power spectrum temperature was higher, and the periodic signal of fluctuation was also strong, which indicated that the climate of Xinjiang was in the middle and late 1990s. There was a general and significant turning point of temperature increase and mutation.

Key words: Xinjiang, climate tendency rate, mutation, wavelet power spectrum, inverse distance weighted, mean squared error