摘要
应用相空间理论对赤道单站日表层水温及青岛月平均气温的９点滑动平均资料进行了分析，结果表明：青岛月平均气温的９点滑动平均的相关维数和嵌入维数分别为５．８３、８，当反厄尔尼诺发生时赤道日ＳＳＴ的相关维和嵌入维数分别为３．９３、１１，当厄尔尼诺发生时其相关维数和嵌入维数分别为４．３７、９．揭示了无论是厄尔尼诺还是反厄尔尼诺发生赤道日ＳＳＴ都具有混沌特征，并且厄尔尼诺发生其内部结构复杂化使得支撑奇怪吸引子的雏数增加，而嵌入维数减少．这种现象是由于表层水温所隐含非线性作用显著增强所导致．根据以上分析结果应用单纯形投影算子（ｓｉｍｐｌｅｘ ｐｒｏｊｅｃｔｉｏｎ ｏｐｅｒａｔｏｒ）预测方法对赤道单站日ＳＳＴ和青岛月平均气温的９点滑动平均资料进行了预测，其前８点最大相对误差分别为０．４２％，７．４％．
Phase space theory is used to analyze the day SST of equatorial local area and nine moving average month air temperature ofQingdao (NMAMAT). The results show that the correlation dimeare and abedding dimension of NMAMAT and SST are 5.83, 8, 3.93(anti-EI Nino) or 4.37 (EI Nino), 11 (anti-EI Nino) or 9 (EI Nino), resnectivaly. The SST of EI Nino and anti-EI Nino can be described as chaotic phenomenon. The correlation dimension increase and embedding dimension decrease of EI Nino SST show that the SST in ner-structure during E1 Nino period is complex and implies nonlinear interaction increase. The above parameters and the forecast method ofsimplex projection operator are used to predict the time series data of SST and NMAMAT, and the biggest predication relative errors of SSTand NMAMAT are 0.42%, 7.4% within the first eight points, respectively.
出处
《海洋学报》
CAS
CSCD
北大核心
1999年第4期21-26,共6页
基金
国家自然科学基金!49476254
关键词
单纯形投影算子
相关维数
表层水温
气温
青岛
Simplex projection operator
correlation dimension
embedding dimension
