Radiocarbon Variability in the Western North Atlantic During the Last Deglaciation Laura F. Robinson,1* Jess F. Adkins,1 Lloyd D. Keigwin,2 John Southon,3 Diego P. Fernandez,1 S-L Wang,1 Daniel S. Scheirer41California Institute of Technology, MS 100-23, 1200 East California Boulevard, Pasadena, CA 91125, USA. 2McLean Lab, Department of Geology and Geophysics, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA. 3Earth System Science Department, 3200 Croul Hall, University of California, Irvine, CA 92697–3100, USA. 4U.S. Geological Survey, 345 Middlefield Road, MS989, Menlo Park, CA 94025, USA. *To whom correspondence should be addressed. E-mail: [email protected] We present a detailed history of glacial to Holocene radiocarbon in the deep Western North Atlantic from deep-sea corals and paired benthic-planktonic foraminifera. The deglaciation is marked by switches between radiocarbon enriched and depleted waters, leading to large radiocarbon gradients in the water column. These changes played an important role in modulating atmospheric radiocarbon. The deep-ocean record supports the notion of a bi-polar seesaw with increased Northern-source deep water formation linked to Northern Hemisphere warming and the reverse. By contrast, the more frequent radiocarbon variations in the intermediate/deep ocean are associated with roughly synchronous changes at the poles. The last deglaciation was punctuated by numerous distinct millennial-scale climate events (1, 2) and understanding the mechanisms behind these changes is a major goal of paleoceanography. The deep ocean stores and transports heat and carbon, so changes in its circulation are likely to influence global climate. Indeed, alternating the main site of deep water formation between the Northern and Southern hemispheres has been linked to switches in the amount of cross-equatorial heat transport (3). This bi-polar seesaw predicts sizable changes in mass transport in the deep North Atlantic and may be the cause of anti-phase warm and cool periods observed in Greenland and Antarctic ice-cores during the last deglaciation (1, 2) (Fig. 1). Well-dated, high-resolution records are needed to make a mechanistic connection between deep-ocean circulation and climate. Passive geochemical tracers from marine sediments show us that during the last glacial maximum (LGM) Northern-source water (NSW) overlay Southern-source water (SSW) with the boundary at ~2,000m in the western North Atlantic (4, 5). The transition from the LGM to the modern state, where North Atlantic Deep Water (NADW) dominates the Western basin, was marked by a series of changes in the deep-ocean circulation pattern (6, 7). To help characterize those changes more completely, we have made 14C/12C measurements of well-dated samples of the deep-sea coral Desmophyllum dianthus (8). A radiocarbon-age can be deduced for a given water mass if its radiocarbon content (14C 5,730 year half-life) is known both when it forms, and when it reaches the deep ocean. By making depth profiles of ∆14C in the past we can investigate variability in deep-ocean ∆14C values, and begin to put constraints on changes in ocean circulation. Deep-sea radiocarbon records can also be used to investigate the role of the ocean in modulating the atmospheric carbon reservoir. The ocean contains ~60 times more carbon than the atmosphere, so small changes in uptake or release of radiocarbon from the ocean may cause significant changes in atmospheric ∆14C. Today, radiocarbon-enriched NADW formation draws down atmospheric 14C more efficiently than radiocarbon-depleted Antarctic Bottom Water (AABW) formation, so varying the proportion of NSW to SSW, or changing the flux of NSW are both likely to change atmospheric ∆14C. Our record of ocean ∆14C lets us constrain the influence of the deep ocean on atmospheric radiocarbon. Using radiocarbon as a circulation tracer has been successful in the modern ocean. NADW and AABW have end-member values of –65 ‰ and –165 ‰ respectively (9–11). Radioactive decay causes deviations below the mixing line of these two end-members, allowing us to calculate the radiocarbon age of the water in the North Atlantic (12). In the modern Western North Atlantic (GEOSECS Station 120, 33°16’N) the water column has a small vertical ∆14C gradient (~10 ‰ /1,000m) consistent with a single, Northern-source water mass (13). By contrast, further South in the Atlantic, NADW is underlain by Southern-sourced AABW. AABW has a characteristic low ∆14C because the “old” Pacific intermediate water from which it forms is not at the surface long enough to re-equilibrate with the atmosphere. In the past, this approach is complicated by variability in the two end-member ∆14C-values at the sites of deep water formation / www.sciencexpress.org / 3 November 2005 / Page 1/ 10.1126/science.1114832(14). For example, increasing the extent of sea-ice cover would allow less air-sea gas exchange, and, therefore, less radiocarbon in AABW. Constraints on the past deep-ocean ∆14C have been acquired using the radiocarbon ages of planktonic and benthic foraminifera (BF-PF) and the aragonitic skeletons of deep-sea corals. In the foraminifera, the planktonic age can be converted to a calendar age, and the benthic 14C/12C ratio can then be used to calculate deep-ocean ∆14C. Early ∆14C reconstructions suffered from problems of species-dependent age variability in planktonic 14C/12C measurements (15–18), but this problem is alleviated by targeting depths with high foraminiferal abundances or high sedimentation rates (19). Deep-sea corals, typically found at water depths of ~500 to 2,500 m are datable by U-Th techniques and are good archives of palaeo-∆14C (20–24). Individual corals with different calendar ages can be compared to one another to give a resolution similar to that of ocean sediment cores. The solitary coral D. dianthus is thought to have a life span of ~100 years (25) so each individual skeleton can be sub-sampled for 14C/12C to construct decadal-resolution records of radiocarbon variability, comparable to the temporal resolution of ice core climate records (20). We collected more than 3,700 D. dianthus corals from the New England Seamounts in May 2003 (26). U-Th isotopic measurements were made by isotope dilution and 27 samples were selected for 14C/12C analysis (27) (table S1). Nine of these corals were sub-sampled to produce high-resolution transects of