Forecasting The Average Temperature Of The Northeastern Summer Based On The Composite Circulation Index And Composite Feature Field

Based on the summer temperature data of 110 stations in Northeast China from 1960 to 2010,combined with 74 circulation index data,monthly average reanalysis data of ECMWF and NCEP/NCAR,the summer temperature in Northeast China was predicted by using the methods of singular value decomposition(SVD),EOF analysis and canonical correlation analysis(CCA).To the following conclusions:1.The maximum correlation between the index circulation field in spring and the temperature field in Northeast China in summer was determined.Four indices,namely,the subtropical high ridge in the Northern Hemisphere,the subtropical high ridge in the North Atlantic Ocean,the Asian polar vortex area index and the North American polar vortex area index,were screened.It was found that the subtropical high was stronger in warm summer than in cold summer,and the polar vortex in the Northern Hemisphere was weaker.This is also consistent with the selected circulation index,indicating that the selected index has certain climatological significance.2.Compared with the four selected indices,the correlation between the composite circulation index and the summer mean temperature in Northeast China is higher.Regression equation is constructed by using the composite circulation index at 110 stations in Northeast China.The correlation coefficients of 87 stations have passed the 0.05 significance hypothesis test,and independent sample test shows that CI has certain predictive ability to the summer temperature in Northeast China.3.Using SVD method,we can see that there are certain relations between the geopotential height field,sea level pressure field,sea level temperature field,surface temperature field and snow depth field of 500 h Pa in the previous winter and the summer temperature field in Northeast China.(1)When the current winter 500 h Pa geopotential height is positive(negative)anomaly,the temperature in Northeast China is generally high(low)in summer,and the influence of geopotential height in North China is more obvious.(2)The temperature in Northeast China is generally higher(lower)in summer when the North(positive)anomaly and the South(positive)anomaly of the current winter SLP field are positive(negative)in the north and negative(positive)anomaly in the south,while the temperature in Northeast China is higher(lower)in summer when the current winter Indian air pressure anomaly is positive(negative).At present,when the air pressure in the winter Qinghai-Tibet Plateau is strengthened(weakened),the summer temperature in Northeast China is generally low(high).When the SLP in Northeast China is positive(negative)anomalous in winter,the summer temperature in Northwest China is on the low(high)side.(3)For the positive(negative)SST anomalies in the eastern and Western Pacific Ocean and the negative(positive)SST anomalies in the westerly drift area,the temperature in Northeast China is higher(lower)in summer.(4)When the surface temperature of India and East Asia is positive(negative)anomaly in winter,the temperature of Northeast China is low(high)in summer;when the surface temperature of North China is positive(negative)anomaly in winter and that of South China is negative(positive),the temperature of North China is high(low)in summer and that of South Asia is low(high).(5)When the SD in Siberia is positive(negative)anomaly,the summer temperature in Northeast China is generally low(high);when the SD in eastern Ural and Northeast China is positive(negative)anomaly,the summer temperature in northern Northeast China is high(low).4.Two eigenvector modes are selected for the 500 h Pa height field in the previous winter,four eigenvector modes for the sea level pressure field in the previous winter,five eigenvector modes for the sea level temperature field in the previous winter and two for the surface temperature field in the previous winter.The eigenvector modes and the snow depth field in the previous winter are taken from seven eigenvectors.The time coefficients of the 20 pairs of singular vector left fields are selected as the composite factor fields.The summer temperature field in Northeast China was decomposed twenty-four modes,and the temperature field was established by the time coefficient of the 24 modes.5.Compound factor field constructs temperature prediction model by CCA method,and carries out prediction experiments by cross-validation method.It is found that the temperature prediction ability of composite factor field is better than that of composite circulation index.The independent sample test from 2008 to 2010 shows that the prediction ability of composite factor field is also better.