Inter-decadal Variability of the Relationship Between Winter Temperature in China and Its Impact Factors

The inter-decadal variation characteristics of winter temperature in China and its key impact factors are analyzed by using monthly temperature data of 160 stations in China, NCEP/NACR reanalysis data, extended reconstructed sea surface temperatures data, and Arctic sea ice extent data from 1951 to 2012, focusing on the inter-decadal changes of the relationship between key influencing factors and winter temperature in China. Results show that the winter temperature in China before 1985 is in a cold period, and then a warm period follows. A significant inter-decadal variability has occurred for winter temperature anomalies. The scope and intensity of warming tendency has weakened significantly from 2004 to 2011.The main diagnostic analysis conclusions are summarized as follows. The majority of impact factors of winter temperature anomalies in China shows significant inter-decadal shift from 1970s to 1980s, most of which changes ahead of that of winter temperatures in China, such as Arctic Oscillation (AO), the East Asian winter monsoon (EAWM), Western Pacific subtropical high (WPSN), Arctic sea ice cover in September, and Niño3 SST index. But the inter-decadal variability of Siberian High (SH) and the basin-wide SSTA variation in the tropical Indian Ocean (IOBW) has the same pace with that of the winter temperature. The inter-decadal variability of relationship between winter temperature and its impact factors have changed on temporal and spatial scales. On temporal scale, the relationship between the winter temperature and IOBW index has weakened significantly from cold to warm period. But the influences of EAWM index, SH index and the WPSH area index on winter temperature have strengthened. On spatial scales, distributions of correlations between winter temperatures and affecting factors have changed from cold to warm period. The high correlation coefficient regions between EAWM and winter temperature have enlarged significantly. The correlation coefficient between AO index and winter temperature is negative in central-southern China in warm period. The inter-decadal variability of relationship between the impact factors has changed, the relationship between EAWM index and tropical SSTA in winter are significantly weakened from cold to warm period.Based on the basic facts of inter-decadal variations, a multi-factor regression prediction model of winter temperature can be established respectively in cold and warm period. This regression prediction model can reflect inter-decadal characteristics of relationship between the winter temperature and its impact factors. Such strategy may keep the stability and effectiveness of the prediction skill for winter temperature in China.