It is shown that Lagged correlations for and cross-correlations betwee
n observed hemispheric-mean temperature data differ markedly from thos
e for unforced (control-run) climate model simulations. The difference
s can be explained adequately by assuming that the observed data conta
in a significant externally forced component involving both natural (s
olar) and anthropogenic influences and that the global climate sensiti
vity is in the commonly accepted range. Solar forcing alone cannot rec
oncile the differences in autocorrelation structure between observatio
ns and model control-run data.