Evidence for Decreased NADW Flux During Glacial Stages 2 and 4 from Nd Isotopes in Southern Ocean Sediments

Abstract

It has been suggested that North Atlantic Deep Water (NADW) production was reduced during the Last Glacial Maximum thereby altering the distribution of heat around the globe, and the partitioning of CO2 between the ocean and atmosphere (Broecker and Denton, 1989). Although the thermohaline circulation concept has great appeal in explaining much of the evidence for climate changes that have taken place in the past, currently employed proxies have produced conflicting results regarding the changes in NADW production in the late Quaternary. In the Southern Oceart ~5C measurements on benthic foraminifera have been interpreted to indicate a switching off of NADW during glacial times. However, this interpretation is in conflict with interpretations based on Cd/Ca measurements of benthic foraminifera (Boyle, 1992). In support of the Cd/Ca results, 23 ipaj230Th data from core top and late glacial marine sediments have been interpreted to indicate that export of deep water formed in the North Atlantic was similar to modem conditions (Yu et al., 1996). The conflicting interpretations based on commonly employed proxies highlights the need for new tracers. We report here on work in progress to validate and apply the Nd isotope composition of marine authigenic precipitates as a proxy for variation in deep water mass composition. The Nd isotope composition of deep seawater varies systematically with water mass properties. The unradiogenic character of Nd in the North Atlantic reflects the ancient shield rocks surrounding the area of NADW formation. The very young rocks surrounding the Pacific Ocean impart a more radiogenic signature to Pacific waters. In the Atlantic Ocean the Nd isotope composition of today's deep water appears to be the product of mixing of Southern Component Waters and Northern Component waters. The observation that the Nd isotope composition of present-day water conforms to water mass properties is an important line of evidence in evaluating the potential for Nd isotopes as a tracer of water mass mixing in the past. However, it is also necessary that there be some extractable phase that sequesters seawater Nd in marine sediment in order to make use of this observation. Published work has demonstrated that ferromanganese precipitates, specifically crusts and nodules, generally record the approximate eNd of the bottom water mass that bathes them, with only a few exceptions (e.g. Albarbde and Goldstein, 1992). Although ferromanganese crusts have been demonstrated to provide a reliable measure of bottom water Nd composition, and recent studies have used these substrates to examine changes in seawater composition over time, crusts are limited in their availability and growth rate, generally integrating over l0 thousand to 1 million years per subsample. Accordingly, we are in the process of testing the application of leaching methods to extract the ferromanganese phase from marine sediments in order to examine secular variation in Nd isotope composition of seawater at glacial to interglacial timescales. A preliminary test of leaches from a suite of core tops has yielded positive results that are entirely overlapping with seawater measurements. Core RCll-83 (41.36~ 9.48~ 4718 m) is located in the Cape Basin of the southeastern Atlantic Ocean. Extensive stable isotope data have been reported for this core (Charles et al., 1996). Charles et al. (1996) have interpreted the large variations in the ~5C to reflect a reduced contribution of NADW to this region during Stage 2 and Stage 4. RC11-83 also records a significant glacial to interglacial contrast in the Nd isotopic composition of the ferromanganese component (Fig. 1). A core top sample and a sample from isotope Stage 3 provide Nd isotope ratios similar to nearby water measurements. In contrast, values much closer to those of the modem Pacific are found in two samples measured from Stage 2. Two measurements come from samples that were taken from transitional intervals based on stable isotope measurements, and these samples also provide intermediate Nd isotope values.