People often pick up on the risk of future cooling over Northern Europe as a "paradoxical" consequence of global warming: most recently in the Times: Britain faces big chill as ocean current slows. I think that people get too enthusiastic about this, because of the exciting "paradoxical" features. One way to tell whether such reports are dodgy or not is by seeing if they reference "The Day After Tomorrow" without pointing out that the film was... just std.hollywoodjunk.
A little disclaimer here: all of this is outside my immeadiate professional expertise, so for these purposes I'm just an interested amateur.
My own view on this (and it was number 7 before) is that its all oversold; that a THC collapse is unlikely (based on the GCMs, of course: ...the probability of a THC shutdown is not high: in fact, it doesn't happen "by itself" in coupled models runs, you have to force it to happen; what you get is instead a slight slowdown and although there is a cooling tendency from the slowdown, the overall effect is warming, even over northern europe. The TAR, section 220.127.116.11 Thermohaline circulation changes is good, as always... was what I said before); and even that the role of the gulf stream is over emphasised.
To take that last point on, try Is the Gulf Stream responsible for Europe’s mild winters? by Seager et al. in QJRMS (google "Seager gulf" if you want the pdf, and other interesting links). They say:
Is the transport of heat northward by the Gulf Stream and North Atlantic Drift, and its subsequent release into the midlatitude westerlies, the reason why Europe’s winters are so much milder than those of eastern North America and other places at the same latitude? Here, it is shown that the principal cause of this temperature difference is advection by the mean winds. South-westerlies bring warm maritime air into Europe and northwesterlies bring frigid continental air into north-eastern North America. Further, analysis of the ocean surface heat budget shows that the majority of the heat released during winter from the ocean to the atmosphere is accounted for by the seasonal release of heat previously absorbed and not by ocean heat-flux convergence. Therefore, the existence of the winter temperature contrast between western Europe and eastern North America does not require a dynamical ocean. Two experiments with an atmospheric general-circulation model coupled to an ocean mixed layer confirm this conclusion. The difference in winter temperatures across the North Atlantic, and the difference between western Europe and western North America, is essentially the same in these models whether or not the movement of heat by the ocean is accounted for. In an additional experiment with no mountains, the flow across the ocean is more zonal, western Europe is cooled, the trough east of the Rockies is weakened and the cold of north-eastern North America is ameliorated. In all experiments the west coast of Europe is warmer than the west coast of North America at the same latitude whether or not ocean heat transport is accounted for. In summary the deviations from zonal symmetry of winter temperatures in the northern hemisphere are fundamentally caused by the atmospheric circulation interacting with the oceanic mixed layer.
Now, this isn't the end of the story, because "minor" effects like a 5 oC change can still come from changes in the ocean circ, and Robert Kunzig tells us:
Other scientists are surprisingly willing to concede much of Seager's challenge to the Gulf Stream myth. Michael Vellinga of the British Meteorology Office's Hadley Centre agrees that the currents do not warm Europe preferentially. But he and others emphasize that they do warm both sides of the Atlantic in winter, by roughly 5 degrees. Although Seager calls that a "modest" effect, it's more than the difference between today's climate and that of the Little Ice Age, from the 16th century to the 19th century, when both Europe and the United States endured many a rude winter. "If you switch [the currents] off, you get massive cooling," says Stefan Rahmstorf of the Potsdam Institute for Climate Impact Research in Germany.
But shutting down the THC wouldn't shutdown the Gulf stream anyway, since its a Western Boundary Current driven by the winds in one of those odd mysteries of ocean dynamics.