Summary: | The purpose of this paper is discussing the estimation of a sample cross correlation function among the time series, global and regional annual average sea surface temperature (SST) deviations of the Indian Ocean, the North Pacific, the South Pacific, the North Atlantic and the South Atlantic. In order to show if a cross correlation estimate is significantly different from zero, the prewhitening one of the time series is very important. We use SARIMA model fitting for the prewhitening and estimate the cross correlation functions. Furthermore, we consider the lagged linear regression model with ARIMA errors by using the series of SST as a leading indicator of the global air temperature. departmental bulletin paper
|