NORTH: I wonder whether EOFs are the best way of describing the data. True, they are objective, but can you tell from analyzing them whether the Gulf Stream's temperature is changing or whether it is just shifting location?
DESER: If you look at the anomalies during the cold 30 years, or one of those decades, the warming first appears near Cape Hatteras. Then you see anomalies along the Gulf Stream east of Cape Hatteras, then east of Newfoundland, then in higher latitudes. A northward shift of about 50 km would be sufficient to produce the observed warming, and seems to me to be the simplest explanation.
MCWILLIAMS: The Gulf Stream warming pattern in your last figure is much broader than 50 km. While a mechanistic interpretation is tempting, an advective supply of heat from the Gulf Stream would be related to wind driving, whereas instead the forcing of the anti-cyclonic gyre is actually weakening.
GHIL: I just wanted to comment on the causes of the decadal variability. It seems to me that certain phase differences between fluctuation or oscillation are clear indications of coupling. A coupled atmosphere-ocean phenomenon seems the most plausible to me.
CAYAN: Have you looked at SST minus air temperature together with your Gulf Stream pattern? I wonder whether part of the reason for the change could be a change in the air mass distribution over the Gulf Stream. In extratropical regions the sensible flux is driven a lot by the temperature lapse rate, and to some extent by the wind.
DICKSON: I should like to show a slide, my Figure 14, that suggests two things. The first is that temperature and salinity show similar patterns, and I think were both involved, meaning that both the circulation and the ecosystem might have had large changes. The second is that the century-long trend covered a much larger area of the North Atlantic than we've been talking about—right up to the ice in the north, in fact.