Box Jenkins Multivariate Modelling and Co-Integration: Two Statistical Methods with Potential Usefulness in Climatic Studies

Climatic data are mainly in the form of time series. In an exploratory analysis of such data we often try to verify or discover causal relationships between two, or among a set of time series. We are also interested to know whether or not the level around which the series fluctuates is constant over time since this may have important physical implications; for the series of surface temperature, for instance, a change in level in the recent decades may indicate the existance of the greenhouse effect.