| The change of snow depth has a far-reaching impact on hydrology,ecosystem and agriculture by changing the energy and water balance between the surface and the atmosphere.In this study,five global climate models with complete SSPs scenarios in CMIP6 were selected,Based on the long time series data set of snow depth in Xinjiang from 1979 to 2014,the temporal variation and spatial distribution characteristics of snow depth in Xinjiang from 1979to 2014 were analyzed.We also evaluated the ability of the climate model we selected to simulate snow depth in the study area,and used machine learning methods to analyze the effects of different meteorological elements such as temperature,precipitation,wind speed,humidity,and sunshine on snow depth.influence.Then,the interannual variation characteristics of snow cover depth to 2100 and the temporal and spatial variation characteristics of snow cover depth in three different periods in the 21st century:2021-2040(recent),2041-2060(middle),2081-2100(late)relative to the base period(1995-2014)are estimated.The results analyzed by machine learning show that temperature and precipitation play a leading role in the variation of snow depth in Xinjiang.Therefore,we also analyzed the temporal trends of temperature and precipitation in Xinjiang from 2100,as well as the temperature,temperature,and precipitation under different SSPs scenarios.The quantitative relationship between precipitation and snow depth,the main conclusions are as follows:1.The average snow depth in Xinjiang region from 1979-2014 shows a trend of slow fluctuation,with a decreasing rate of 0.96cm/10 years.The error between the simulated average snow depth of the five climate models from 1979 to 2014 and that of the snow depth dataset is about 4%,and there is not much difference between the results of the multi-model average and that of the snow depth dataset.Compared with the snow depth data set,the snow depth simulated by the climate model can reflect the basic characteristics of the variation of snow depth in a year to some extent.The maximum distribution of snow depth in a year is from January to March,and the error between the simulated value of the climate model and the observed data from January to March is relatively small.Therefore,in this study,we focused on the spatial-temporal variation characteristics of snow depth from January to March,and compared the spatial distribution of snow depth simulated by different climate models from January to March of each year from 1979 to 2014 with the spatial distribution of snow depth shown by snow depth data sets.In most areas of Xinjiang,the snow depth simulated by climate model and the snow depth data set have similar spatial distribution characteristics.The climate model selected in this study can be used to predict the spatial-temporal changes of snow depth in Xinjiang in the future.Based on the machine learning method,the changes of different meteorological factors and snow depth are analyzed,and it is concluded that the relative importance of temperature and precipitation to snow depth in Xinjiang is far greater than other meteorological factors.2.Climate patterns after deviation correction for the temperature and precipitation in Xinjiang area two meteorological elements and the correlation between the observed values were greater than 0.8,after a deviation correction of climate models to the temperature and precipitation in Xinjiang region have better simulation ability,and weather patterns in the Xinjiang region of the ability of simulating the temperature than precipitation.The climate of Xinjiang will be warm and humid in the future.Under different SSPs scenarios,the changes of air temperature and precipitation in the middle and long term compared with the base period in the 21st century are greater than those in the near term compared with the base period.3.The snow depth in Xinjiang shows a trend of slow fluctuation,with an average increase of0.41cm·(10a)-1from 2100.In the three different periods of the 21st century,the increase of snow depth under high emission scenarios is relatively large,and the increase of snow depth under high emission scenarios is about 20%in the late stage.The spatial distribution characteristics of Xinjiang are different in three different periods in the 21st century compared with the baseline period.To be specific,the increase of snow depth in the early 21st century compared with the baseline period is smaller in the three periods,and the areas with reduced snow depth are mainly distributed in the western part of the Tarim Basin.In the middle of the 21st century,the increase of snow depth mainly occurs in the eastern Part of the Junggar Basin,while the snow depth will decrease in most areas of the Tarim Basin.In the late 21st century,the snow depth increased greatly mainly in the eastern part of the Junggar Basin.4.Under different SSPs scenarios,the evolution characteristics of snow depth and the quantitative relationship between temperature,precipitation and other meteorological factors in Xinjiang from the end of the 21st century are different to some extent.With the increase of emission scenarios,the correlation coefficient between snow depth and temperature,precipitation is also increasing.Under SSP1-1.9 and SSP2-4.5 scenarios,air temperature has a greater impact on snow depth,but under SSP5-8.5 scenarios,precipitation has a greater impact on snow depth. |
PDF Full Text Request