Analysis Of Rapid Arctic Sea Ice Declining And Climate Change Based On Statistical Method

Arctic sea ice extent has been on a declining trend over the past decade years,with a particularly pronounced decline in autumn.Autumn sea ice extent reaches its smallest value in 2020 since satellite observations were recorded.Sea ice concentration is also decreasing.As the extent of multiyear ice coverage declines,the Arctic Oceans changes from being dominated by multiyear ice to being dominated by first-year ice.Arctic sea ice plays an important role in the Arctic climate,and in the global climate change by influencing the exchange of energy between the ocean and the atmosphere.This thesis used the sea ice concentration products from NOAA(National Oceanic and Atmospheric Administration),sea ice extent products from NSIDC(National Snow and Ice Data Center)sea surface temperature(SST),surface air temperature(SAT),specific humidity(SH),sea level pressure(SLP)and wind fields from the NCEP/NCAR(National Centers for Environmental Prediction/National Center for Atmospheric Research)reanalysis dataset,all the data covers the period from 1982 to2020.To investigate the spatial and temporal variability of autumn sea ice,and analyze the relationship between sea ice with the variables mentioned above.The overall autumn sea ice concentration declines significantly from the Beaufort Sea to the Barents Sea and the rate of decline is highest along the Beaufort Sea.,The correlation coefficients between sea ice extent(SIE)anomalies and SST,SAT and SH anomalies were calculated.The highest correlation coefficient is in August-October between autumn SIE and SAT,with a value of-0.945.Empirical orthogonal function(EOF)is used to analyze SST,SAT,SH,SLP and wind fields.The first EOF modes of SST,SAT and SH show that the increasing area are mainly concentrated in the Beaufort Sea to the East Siberian Sea.The first EOF mode of SLP is mostly negatively distributed in the central Arctic.Over the Beaufort,Chukchi and East Siberian Seas,the zonal winds weaken while the meridional winds strengthen.The results of the correlation and EOF further validate the influence of ice-temperature,ice-SH and ice-SLP feedback mechanisms in the Arctic.The ARIMA method is used to predict the time series of arctic autumn SIE.After first-order difference,white noise test and significance test,ARIMA(3,1,0)model is selected.The prediction results show that the Arctic autumn SIE will reach 5.68×10~6 km~2 in 2030.M-K test and linear regression fitting were used to synthesize all years before and after the mutation point,and 2002 was selected as the cut-off point.The temporal and spatial variations of autumn sea ice in the Arctic during 1982-2001 and 2002-2020 were investigated.In 1982-2001,the center of sea ice decreasing concentrated in the Chukchi Sea and the Bering Strait.While in 2002-2020,the center shift to the vicinity of the Barents Sea,with the declining rate increasing from-0.44×10~6 km~2/10a to-0.83×10~6km~2/10a.Arctic sea ice is significantly negatively correlated with SST,SAT and SH.The correlation coefficients are higher in 2002-2020 compared with those in 1982-2001,and the period with the highest correlation coefficient is one month earlier.The EOF results show that the rising center of SST,SAT and SH concentrated in the Chukchi Sea and Bering Strait during 1982-2001,while the rising center shifts to the Barents Sea during 2002-2020.The shift in the spatial distribution of SLP and wind fields during the two phases also occurs.The NECP/NCAR daily mean surface temperature data were processed by the moving average method,and the summer length change in the northern hemisphere was calculated by selecting step 5.Based on the daily SAT data of NECP/NCAR,geopotential height(GH)and SH data of ERA5(European Centre for Medium-Range Weather Forecasts Reanalysis Version 5,ECMWF Reanalysis v5)),the variation of summer length in the Northern Hemisphere under the background of melting autumn sea ice were studied.The mean and maximum temperatures in the Northern Hemisphere are gradually increasing,with the maximum value reached in the summer of 2020.The length of summer in the mid-latitudes and polar regions was determined based on the daily SAT from March to November in the 30°-65°N and 65°-90°N regions.The summer length increases significantly in both mid-latitude continental and polar regions,and increases by 22 days from 1982 to 2020.The highest spatial correlation coefficient between summer length and autumn SIE is-0.921 in the Kara Sea.The regression results for air temperature and sea ice indicate that SAT in the polar region increase significantly in spring and autumn,while warming area is concentrated in the 900-500h Pa in summer.In spring,summer and autumn,the SH increases significantly in eastern North America,the Mediterranean region,Central Asia and the North Pacific.The atmospheric vertical structure of the SH rising area are mainly concentrated in 1 000-800 h Pa.GH and wind fields form anticyclones in the Arctic during the spring and further deepen in summer.In autumn,anticyclones form not only in the polar and North Pacific regions,but also in western North America.