汛期降水的奇异谱分析及预报试验  

SINGULAR SPECTRUM ANALYSIS AND ITS FORECASTING EXPERIMENTS ON SUMMER RAINFALL

作　　者：黄嘉佑[1] 黄茂怡[1] 

机构地区：[1]北京大学地球物理系,北京100871

出　　处：《应用气象学报》2000年第A06期58-63,共6页Journal of Applied Meteorological Science

基　　金：国家“九五”重中之重科技攻关项目“我国短期气候预测系统的研究”96-908-04-01-3专题;国家重点基础研究发展规划项目“我国重大气候和天气灾害形成机型和预测理论的研究”C1998040901-1的资助。

摘　　要：用奇异谱分析方法对宜昌地区和我国汛期降水的气象时间序列进行预测 ,比较奇异谱分析与谱分析效果的差异 ,检验其在实际短期气候预测中的应用 .结果表明 ,奇异谱分析方法比自回归模型有较好的预测效果 ,对我国 1 60站汛期降水的预测准确率达 50 % .In order to investigate the skill and characteristics of singular spectrum analysis(SSA) in forecasting meteorological time series and to identify the similarity and difference between SSA and spectrum analysis, and also to examine the applicability of SSA in the operational short range climate forecasting, SSA is applied to the rainfall field of Yichang District and China. The results indicate that the forecasting skill of SSA is higher than conventional auto regression models. The SSA model gives a skill of 50% in the forecast of summer rainfall when applied to 160 stations in China.

关 键 词：奇异谱分析 预报 

分 类 号：P457.6[天文地球—大气科学及气象学]