18.104.22.168. Multivariate Fingerprints
Although most univariate detection methods of global climate change due to an enhanced greenhouse effect have their limitations, it is likely that multivariate fingerprint methods, which involve the simultaneous use of several time series, will facilitate attribution. In its most general form one might consider the time evolution of a 3-D spatial field, comparing model results with observations. This may be achieved by comparing changes in mean values and variances, or correlating spatial patterns between simulation and observation (Wigley & Barnett, 1990).