The arid northwestern China is the most vulnerable region to climate change, where the variability of seasonally extreme temperature events has profound implications for both its hydrological, ecological, and human systems. In this study, we applied 15 indicators of extreme temperature to analyze the spatial and temporal variation of its occurrence in arid northwestern China for a recent 40-year period (1979 to 2018). These extreme temperature event dynamics were then combined with atmospheric and oceanic circulation to explore their response mechanisms. Our results revealed the following: (1) Over the 40-year period, the annual average temperature in this arid zone increased at a rate of 0.4 degrees C/decade (p = 0.09), exceeding the national average rate (0.28 degrees C/decade). Apart from a few indicators, extreme temperature events (TXx, TNx, TXn and TNn) generally increased at least twice as fast as average temperature during the four seasons, especially in spring, when TNn (0.98 degrees C/decade) rose five times faster than did the average temperature (0.2 degrees C/decade). (2) Spatially, except for the Kunlun Mountains and Tarim Basin, seasonal warming occurred in most parts of the studied arid zone, being most prominent in the summer. In this season, the average number of warm nights increased (3.23 days/decade), while the average number of cold nights decreased (2.69 days/decade). (3) After the 1990s, extreme temperature events accelerated significantly. The Cold Spell Duration Indicator decreased 42% in spring and the Warm Spell Duration Indicator increased 300% in summer, from 1979 -1998 to 1999-2018, which may hasten the formation of snow and glacier melt flooding events in the spring and summer. Spatiotemporal variability in seasonally extreme temperature events was closely related to atmospheric and oceanic circulation, particularly for the AMO (r = 0.8). Altogether, these findings enhance our understanding of how to better assess shifts in extreme temperature events in response to a changing climate in arid zones.

Aerosol particles of Black carbon in the snow cause a significant decrease in the albedo spectrum of the snow, which results in climatic radiation changes seriously, and will delay or advance the snow melting time, badly affecting the characteristics of surface runoff and processes of water cycle in the arid region. This problem is receiving increasing attention in ecological hydrology issues in the arid region. The data of field measurement were obtained by ASD spectrometer, Snow Folk and HR-1024 external field spectrum radiometer. SNICAR model was used to simulate the snow spectrum spectral characteristics under different parameters. Discussed the sensitivity of BC and snow particle size in different spectral ranges. The results showed that : In the snow spectral curve, the zenith angle changes from 0 degrees to 80 degrees, the albedo at 600 nm in the visible spectrum increases by 0. 045, and the albedo at 1 000, 1 200 and 1 300 nm in the near-infrared band increases by 0. 16, 0. 225 and 0. 249, respectively. The zenith angle is at 60 degrees, when snow particle size increases from 100 to 800 mu m, the albedo reduction can reach 0. 15, and snow particle size in the range of 100 similar to 300 mu m is significantly higher than the albedo in the range of 400 similar to 800 mu m. And the increase of the snow particle size can enhance the absorption effect of the light spectrum absorbing particles; Different BC concentrations have little effect on the spectral albedo in the near-infrared region, but are mainly concentrated in the visible light band. At 800 and 1 100 nm, the BC concentration of 5 mu g . g(-1) reduces the spectral albedo by 0. 13. The BC of 5 mu g . g(-1) can reduce the spectral albedo at 350 and 550 nm by 0. 25 and 0. 23. Compared with the different snow sizes, the decrease of BC concentration on the broad-band albedo of snow spectrum can be found in BC. In the case of the increase in the particle size of the snow, the light absorption effect of BC is increased, and at the higher concentration, the more the absorption increases; from the spectral index, the BC is sensitive in the visible light range of 350-740 nm, and the correlation coefficient is higher; The snow size is sensitive in the near-infrared band 1 100 similar to 1 500 nm, especially around 1 000 and 1 300 nm. The correlation between BC and snow particle size in the sensitive band of the snow spectral curve is high. Finally, the snow albedo simulated by the model is compared with the measured data. The R-2 is 0. 738, and the simulation effect is good. It can lay a data foundation for the study of the snow albedo in the arid region.