Long-term coordinated changes in the convective activity of the North Atlantic

The North Atlantic is a peculiarly convective ocean. The convective renewal of intermediate and deep waters in the Labrador Sea and Greenland/Iceland Sea both contribute significantly to the production and export of North Atlantic Deep Water, thus helping to drive the global thermohaline circulation, while the formation and spreading of 18-Degree Water at shallow-to-intermediate depths off the US eastern seaboard is a major element in the circulation and hydrographic character of the west Atlantic. For as long as time-series of adequate precision have been available to us, it has been apparent that the intensity of convection at each of these sites, and the hydrographic character of their products have been subject to major interannual change, as shown by AAGAARD (1968), CLARKE, SWIFT, REID and KOLTERMANN (1990), and MEINCKE, JONSSON and SWIFT (1992) for the Greenland Sea, in the OWS BRAVO record from the Labrador Sea, (egLAZIER, 1980 et seq.), and at the Panulirus / Hydrostation "S" site in the Northern Sargasso off Bermuda (eg JENKINS, 1982, TALLEY and RAYMER, 1982). This paper reviews the recent history of these changes showing that the major convective centres of the Greenland and Labrador Seas are currently at opposite convective extrema in our postwar record, with vertical exchange at the former site limited to 1000 m or so, but with Labrador Sea convection reaching deeper than previously observed, to over 2300 m. As a result, the deep water of the Greenland Sea has become progressively warmer and more saline since the early '70s as a result of increased horizontal exchange with the Arctic Ocean through Fram Strait, while the Labrador Sea Water has become progressively colder and fresher over the same period through increased vertical exchange; most recently, convection has become deep enough there to reach into the more saline NADW which underlies it, so that cooler, but now saltier and denser LSW has resulted. The horizontal spreading of these changing watermasses in the northern gyre is described from the hydrographic record. The theory is advanced that the scales of atmospheric forcing have imposed a degree of synchrony on convective behaviour at all three sites over the present century, with ventilation at the Sargasso and Greenland Sea sites undergoing a parallel multi-decadal evolution to reach a long term maximum in the 1960s, driven by the twin cells of the North Atlantic Oscillation (NAO). During the NAO minimum of the 1960s, with an extreme Greenland ridge feeding record amounts of fresh water into the northern gyre in the form of the Great Salinity Anomaly, and its partner cell over the Southeast USA causing a southwestward retraction of storm activity (DICKSON and NAMIAS, 1976), the surface freshening and postwar minimum in storm activity in the intervening area of the Labrador Sea also brought a progressive reduction, and ultimately a cessation, of wintertime convection there during the 1960s. In other words, the evolution of winter convective activity during the century was in phase but of different sign at the three sites. In these events, we see strong evidence of a direct impact of the shifting atmospheric circulation on the ocean; while this certainly does not rule out either feedbacks from anomalous ice and SST conditions on the atmosphere, or autonomous oscillations of the ocean's overturning circulation, it does tend to minimise them. Crown copyright (C) 1997 Published by Elsevier Science Ltd