摘要
对太平洋海温场与我国夏季降水之间的典型相关分析表明,海温场第一典型分布型的季节变化为:秋季东南高西北低、冬季东西低中间高、春夏季西北高东南低。典型主分量的时间变化规律清楚地反映出ElNino的演变过程。用CCA短期气候预测模型对我国夏季降水的预报试验结果显示,我国的降水对因子场的响应存在明显的地区性差别,且预报效果同各区降水指数与典型主分量相关性的高低关系密切。
The relation between previous Pacific SST and summer rainfall in China is investigated through the agency of canonical correlation analysis (CCA). Evidence suggests that the first canonical pattern of the SST predictor has its seasonal distribution as follows: the SE part is higher in values than the NW in autumn, the east and west are lower than the central part in winter, the NW higher than the SE in spring and summer, and temporal variation of the canonical principal component pattern clearly reflects the El Nino episode. Also, the CCA predicted summer precipitation experiments show that the rainfall response to the SST varies considerably from place to place and the forecast skill depends strongly upon the amount of correlativity between the rainfall index and canonical principal component.
出处
《南京气象学院学报》
CSCD
1998年第1期130-137,共8页
Journal of Nanjing Institute of Meteorology
基金
国家气象中心"九五"课题
关键词
典型分布型
预报效果
降水
海温场
相关分析
canonical pattern, canonical principal component, forecast skill, independent sample test, correclation coefficient for prediction