What you see (above) is a plot of the smoothed (w/85 month filter) sea surface temperature annomalies of the Northern Pacific (0 to 65 Deg N) between January 1854 and May 2008 (top graph) plotted above the Earth's LOD (Length-of-Day). Note that the LOD has been shifted by roughly 8 years forward in time compared to the sea surface temperature anomaly data. Note the remarkable correlation between these two paramemters.
Subtle varaitions in the Earth's LOD appear to be followed eight years latter by similar variations in the sea surface temperatures. Also note that I am not claiming that correlation means causation.
In fact, I will try to show that both of these parameters are in fact driven by long term variations in the lunar tides.