Interdecadal Variability Of Relationship Between Winter Temperature In China And Its Impact Factors In Research And Prediction Methods

In recent decades, the winter temperature in China and the relationship between the wintertemperatures and its key factors may be changed. This theis focuses on the interdecadalchanges of the relationship between key influencing factors and winter temperature in China.On the base of this, an appropriate strategy is proposed for building prediction models in orderto improve the skill of climate prediction.Based on the monthly temperature data of160stations in China and the NCEP/NCARreanalysis data from1951to2012, the inter-decadal variations of winter temperature in Chinaand its key factors are analyzed. The results show that the relationship between winter temper-ature and factors has changed. That is, the relationship of different factors and winter temper-atures has changed during the cold period and warm period. Based on such analysis, a forecaststrategy is proposed to use statistical forecasting methods and may improve the skill of wintertemperatures forecast.The main conclusions are summarized as follows:(1) The winter temperatures in China have significant interdecadal variabilities. The warmperiod was defined from1951to1985and then a cold period followed. In the second period,the scope and intensity of warming has significantly weakened from2004to2011. From thecold period to the warm one, the most obvious warming area was in northern regions of Chi-na.The interdecadal variation is obviously different in winter temperature between the northernChina and southern China. Winter temperature over China has multitime scales, such as2-4years,5-8years, and more than20years.(2) Most of inter-decadal variability of impact factors changed ahead of the winter temper-atures in China, such as Arctic Oscilation (AO), the East Asian winter monsoon (EAWM),Western Pacific subtropical high (WPSN), Arctic sea ice in September, and index of Nino3SST. But the inter-decadal variability of Siberian High (SH) and the SST in Indian Ocean hasthe same pace with winter temperature while and SST in Kuroshio is delayed.(3) The Interdecadal variability of relationship between winter temperatures and its impactfactors have changed in temporal and spatial scales. In temporal scale, the relationship be-tween the winter temperature and SST in part of the ocean area，thermal effects of the Tibetan Plateau have significant weakened on interdecadal scale after the cold period. However, theinfluences of EAWM index, SH index and the WPSH area index on winter temperature havestrenthened. The relationships between each two impact factors have taken place obviously ondecadal scale. In spatially scales, the regions which have past the statistically significant levelbetween winter temperature and affecting factors have changed after the cold period. The highcorrelation coefficient regions between EAWM and winter temperature have enlarged signifi-cantly. Meanwhile, when the western Pacific subtropical high is strong than normal, the wintertemperature is remarkable warmer in warm priod than that in cold period in the southwest andsoutheast coastal China. Positive phase of AO has caused a significant cooling in south China.(4)Based on the fact that the interdecadal variablity of relationship between winter temper-ature over China and its impact factors have changed, a new strategy is proposed for build-ing multi-scale linear regression equation for prediction models under different period includ-ing cold period and warm one in order to improve the skill of climate prediction. Such me-thod may guarantee the stability and effectiveness of the prediction.The work show the facts that the relationship between winter temperature and its affect-ing factors has changed at inter-decadal time scale.Moreover, a new stategy on how to use sta-tistical method to eatablish prediction model under different period. All of the research was afoundation for further studies on about the mechanism of inter-decadal variation predictor andpredictant.