摘要
采用奇异谱分析(SSA)与自回归向量(AR)预报模型相结合的方法,对热带地区大气季节内振荡(MJO)指数向量作自适应滤波意义下的预报试验。结果表明,通过对MJO原始序列进行SSA的分解重建,无论采用对重建的分量序列进行AR(P)建模的方案,还是利用对重建合成序列进行AR(P)建模的方案,均可得到两周以上的MJO指数预报能力,其提前20天指数预报值与实况之间平均相关系数达到0.5,与直接对MJO原始序列进行AR建模相比较,该方法有较高的预报技巧和超前预报能力,预报效果也较稳定,故将SSA-AR方案进一步完善,可望作为MJO指数业务预报的有效模型。
Using both the Singular Spectrum Analysis（SSA） and Auto-regression（AR） model,this study presents a short-term forecast experiment on the vectors of the tropical atmospheric intraseasonal oscillation（MJO） index under the condition of an auto-adaptive filter.The results show that the SSA-AR model was very efficient to forecast the MJO index for a period of 1～20 days ahead.For a 20 day forecast,the correlation coefficient between the predicted and validated observations,when averaged for the two indices,was 0.5,which was better than the result of the AR model.Therefore,the forecast skills of this SSA-AR prediction scheme were steady and its independent sample tests and real forecasts were of high precision.If the SSA-AR scheme is further revised,it is possible that it becomes an efficient model for operational forecast of MJO.
出处
《热带气象学报》
CSCD
北大核心
2010年第3期371-378,共8页
Journal of Tropical Meteorology
基金
国家科技支撑计划2009BAC51B01
江苏省高校自然科学重大基础研究项目07KJA17020
国家自然科学基金资助项目40875058共同资助
