用非线性时间序列分析方法对海平面温度距平场进行重构。方法包括3部分:主分量分析,相空间重构,最小二乘拟合。与传统线性拟合方法的区别在于,这里用时间信息代替空间信息。将此方法和线性拟合方法均用CZ模式资料进行检验,两者比较表明,非线性重构的结果很好。
A nonlinear time series analysis method has been used for reconstruction of sea surface temperature (SST) anomalies. The method involves principal components analysis, phase space reconstruction and least squares estimate. It differs from traditional linear fitting methods in the way that temporal information is used instead of spatial information. The method is demonstrated on CZ model data and proves to be a success.