Geomagnetic activity index AP* was filtered with a 4 month gaussian filter and lagged 1.8 years in time. This lag was determined by aligning key features in the two datasets. Principally the 1957-58 rise and 1963-64 fall, which were sharply defined in both data. The abruptness of these events makes them remarkable in the SST record.

Surface temperature ‘anomaly’ was scaled by a factor of 7 and offset +1.5K to approximately match magnitude of short and long term changes.

The divergence at the end corresponds to about 0.5K in SST.

There is a notable anti-correlation following El Chichon eruption and most significantly during 1998 El Nino. The latter seems to account for the majority of the divergence.

Longer AA* geomagnetic record does not show a good correspondence to the temperature record, prior to the period covered by AP* shown here.