SECT. 2] LARGK-SCALE INTERACTIONS 283 



somewhat longer interval.^ Thus no demonstrable increase in th(^ jjjeostrophic 

 mass transport of the stream is deducible ; over such a long time, changes in 

 How should surely have become adjusted with the pressure field (V^eronis and 

 Stommel, 1956). However, it is impossible at the present stage of dynamic 

 oceanography to predict, even in order of magnitude, the relation between 

 warming in the northern surface waters and the mass flow of the stream. The 

 heat transport by layers is not deducible from the mean mass flow, which is all 

 that can be computed from the wind-driven current models. Nor are the mixing 

 processes during meanders well enough studied to estimate how much warm 

 water is given off, and where and how much is recycled in the gyre. Whether 

 or not the mean position of the stream has shifted in this interval , or its meander 

 pattern has changed, probably cannot be settled retroactively. In the region of 

 tight surface-temperature gradients, a translation of the mean pattern by as 

 little as 20 nautical miles could bring a 2° warmer isotherm to the surface at a 

 given spot. 



The second part of Bjerknes' study develops an explanation of the cooling 

 south of Iceland and the slight warming in the Sargasso Sea in terms of mass 

 readjustments of the ocean layers to the increased wind circulations. The basis 

 of the argument is clearly summarized in Fig. 90. It runs briefly as follows : an 

 anticyclonic vortex in the ocean which is decreasing in intensity with depth 

 must be of the warm core type ; in other words, the warm surface layer must 

 have maximum thickness near the center. In that way, in fact, the anti- 

 cyclonic winds around the Atlantic high do maintain in a permanent fashion a 

 downward bulge of the lower limit of the warm surface water-mass. Thus 

 Bjerknes suggests that anticyclonic anomaly winds, more or less concentric with 

 the anticyclonic ocean currents, would add more depth to the warm surface 

 water at the center. Inflation of the warm layer would tend to raise very 

 slightly the equilibrium temperature of the ocean surface, because the water 

 there would have become a little less exposed to mixing with the cold deep 

 water. Conversely, a cyclonic current system decreasing with depth will be 

 characterized by minimum thickness of the warm surface layer. Therefore, an 

 increase in circulation around the normal Icelandic low pressure cell in the 

 atmosphere would reinforce the cyclonic vortex of oceanic flow in high latitudes 

 and make the warm upper layer thinner, thus exposing the water at the ocean 

 surface to more mixing with the cold deep water. 



In attempting to pursue these deductions further, we run abruptly into the 

 unsolved problem of water-mass production and destruction, and the small- 

 scale turbulent or convective transfers through the thermocline, of which 

 oceanographers are just becoming able to attempt serious models (Stern, 1960). 

 As Stommel has so cogently pointed out in the conclusion of his book, The Gulf 

 Stream, the lack of physical knowledge on this key point permits opposite 

 conclusions to be drawn relating circulation strength and warming of the 



1 The writer is deeply indebted to her colleague Henry Stomrriel for his generous sharing 

 of his vast information, experience and oceanographic insight in constructing the discussion 

 on this page. 



