The Salinity, Temperature, and␦18O of the Glacial Deep OceanJess F. Adkins,1* Katherine McIntyre,1Daniel P. Schrag2We use pore fluid measurements of the chloride concentration and the oxygenisotopic composition from Ocean Drilling Program cores to reconstruct salinityand temperature of the deep ocean during the Last Glacial Maximum (LGM). Ourdata show that the temperatures of the deep Pacific, Southern, and Atlanticoceans during the LGM were relatively homogeneous and within error of thefreezing point of seawater at the ocean’s surface. Our chloride data show thatthe glacial stratification was dominated by salinity variations, in contrast withthe modern ocean, for which temperature plays a primary role. During the LGMthe Southern Ocean contained the saltiest water in the deep ocean. This reversalof the modern salinity contrast between the North and South Atlantic impliesthat the freshwater budget at the poles must have been quite different. A strictconversion of mean salinity at the LGM to equivalent sea-level change yieldsa value in excess of 140 meters. However, the storage of fresh water in iceshelves and/or groundwater reserves implies that glacial salinity is a poorpredictor of mean sea level.The general circulation of the modern deepocean is dominated by two distinct watermasses: In the Atlantic, warm, salty NorthAtlantic Deep Water (NADW) is formednorth of Iceland and in the Labrador Sea,whereas in the Southern Ocean, cold, freshAntarctic Bottom Water (AABW) forms oncontinental shelves. Because of its greaterdensity, AABW underlies NADW in the At-lantic at all latitudes south of ⬃40°N. Nodeep waters are formed in the North Indian orNorth Pacific Oceans today. The mean deep-ocean tracer properties are strongly weightedtoward those of the Indo-Pacific abyss, wheredeep waters are roughly a 50/50 mixture ofAABW and NADW (1) and the average over-turning time is about 800 years (2).Measurements of passive tracer fields (3–5) and the radiocarbon age difference be-tween surface and deep waters (6–9) suggestthat during the LGM the Atlantic was floodedby waters of a Southern Ocean origin, theLGM analog to NADW shoaled but still ex-ited the Atlantic, and the mean overturningrate was somewhat slower than today. Thesechanging patterns of deep tracer distributionsare consistent over many glacial cycles (10).Although these tracer distributions for theLGM have provided significant insight intothe mechanisms of climate change, a betterunderstanding of the deep circulation of thepast requires constraints on the seawater den-sity, a function of temperature, salinity, andpressure. Oxygen isotope ratios (␦18O) withinthe carbonate shells of benthic foraminiferaare a function of both temperature and sea-water isotopic composition, but distinguish-ing between these two effects is a long-stand-ing problem in paleoceanography (11–13). Aseries of recent studies using pore fluids (14 –16) and Mg/Ca ratios (17–19) have attemptedto separate the temperature effect, but rela-tively little is known about the glacial salin-ity. Following a preliminary study in theNorth Atlantic (14), we used the pore-fluidapproach (20, 21) to independently constrainthe seawater salinity by using profiles ofchlorinity from four Ocean Drilling Program(ODP) sites (Table 1). By combining thesedata with ␦18O measurements from the samelocations (16 ), we have determined the spa-tial differences in salinity and temperature forthe glacial deep ocean (22).Sediment-pore water samples for these mea-surements were squeezed at sea and sealed inglass ampoules for transfer to our shore-basedlabs (23). Sample resolution varied betweenand within sites but ranged from every 1.5 m atsite 981 to every9matthebottom of site 1123(Figs. 1 and 2). At all sites there is a pore-fluidmaximum in ␦18O and [Cl] corresponding tothe LGM ice-volume peak. If these systemswere exactly analogous to diffusion in a pipe,1MS 100-23, Department of Geological and PlanetarySciences, California Institute of Technology, Pasadena,CA 91125, USA.2Department of Earth and PlanetarySciences, Harvard University, Cambridge, MA 02138,USA.*To whom correspondence should be addressed. E-mail: [email protected] 19.4 19.5 19.619.319.2 19.4 19.519.619.35 19.45 19.55 19.65Depth (mcd)100Depth (mcd)19.319.2 19.4 19.5 19.6 19.7 19.8Site 981=3.3 ± 0.3%Site 1063=2.7 ± 0.1%Site 1093=6.9 ± 0.5%Site 1123=4.2 ± 0.2%[Cl] (g/Kg)[Cl] (g/Kg)020406080100020406080Fig. 1. The top 100 mof pore-fluid [Cl] forfour ODP sites and ourmodel fits to the data.Solid lines are modelresults using the coralsea-level curve as thebottom-water [Cl] his-tory. Dashed graylines are a small alter-ation to this curveduring the Holocene.There is more scatterin the [Cl] than the␦18O because evapo-ration and fresh tap-water addition duringsqueezing have amuch larger effect onthe [Cl] values than onthe ␦18O values. Atsite 1093, several ofthe points in the upper20 m were rejectedbecause they evapo-rated between sam-pling and measure-ment at Caltech. Weare confident in ignor-ing these points be-cause the shipboard[Cl] profile agrees withall of our points ex-cept these. Site 1123has a wide scatter atthe LGM peak andcould be fit by a vari-ety of other curves.The top boundary isconstrained to match the local bottom-water value. This is a much stronger constraint for [Cl] thanfor ␦18O because the modern database of salinity values is so much larger than for oxygen isotopes.Our pore fluids are standardized to the same reference (IAPSO) as modern conductivity temper-ature-depth sensors. mcd, meters composite depth.R EPORTSwww.sciencemag.org SCIENCE VOL 298 29 NOVEMBER 2002 1769then we would expect the pore-fluid peaks in[Cl] and in ␦18O to be about 25 m below the seafloor. However, sediment accumulation overthe last 20 thousand years (ky), compaction-driven advection, and the duration of the lastglacial period all work to alter the actual depth.At our sites, the amplitudes from bottom-waterminima to glacial maxima range from ⬃0.25 to0.5 g/kg in chlorinity and 0.3 to 0.4 per mil (‰)in ␦18O(24). We measure [Cl] rather thansalinity because the ratios of the major ions inseawater are not constant with depth in thesediment column. Specifically, sulfate reduc-tion and cation-clay interactions change thechemical composition of salinity. Chloride, out-side of methane clathrate formation, is conser-vative in pore fluids and faithfully tracks thediffusively altered history of