Strength and geometry of the glacial Atlantic Meridional Overturning Circulation

Abstract

International audienceThe strength and geometry of the Atlantic meridional overturning circulation is tightly coupled to climate on glacial-interglacial and millennial timescales(1), but has proved difficult to reconstruct, particularly for the Last Glacial Maximum(2). Today, the return flow from the northern North Atlantic to lower latitudes associated with the Atlantic meridional overturning circulation reaches down to approximately 4,000 m. In contrast, during the Last Glacial Maximum this return flow is thought to have occurred primarily at shallower depths. Measurements of sedimentary Pa-231/Th-230 have been used to reconstruct the strength of circulation in the North Atlantic Ocean(3,4), but the effects of biogenic silica on Pa-231/Th-230-based estimates remain controversial(5). Here we use measurements of Pa-231/Th-230 ratios and biogenic silica in Holocene-aged Atlantic sediments and simulations with a two-dimensional scavenging model to demonstrate that the geometry and strength of the Atlantic meridional overturning circulation are the primary controls of Pa-231/Th-230 ratios in modern Atlantic sediments. For the glacial maximum, a simulation of Atlantic overturning with a shallow, but vigorous circulation and bulk water transport at around 2,000 m depth best matched observed glacial Atlantic Pa-231/Th-230 values. We estimate that the transport of intermediate water during the Last Glacial Maximum was at least as strong as deep water transport today

231Pa/230Th evidence for a weakened but persistent Atlantic meridional overturning circulation during Heinrich Stadial 1

“…Data north of 45°S were used for the LGM and HS1 to account for the northward shift of the Southern Ocean opal belt during the LGM, in keeping with Yu et al 1 The average 231 Pa/ 230 Th ratio in LGM sediments (0.065±0.005; n ¼ 79) is similar to the value obtained by Yu et al 1 (0.059±0.007; n ¼ 26; Fig. 2); this study 15,16,[20][21][22][23]25,27,28,30,39,40,[54][55][56] . The average ratio for HS1 (0.074 ± 0.009; n ¼ 26) is somewhat greater than the LGM value, but not significantly different at the 95% confidence level; this study 15,16,21,22,25,27,30,39,40,54 .…”

“…The observation of a basin-wide deficit of 231 Pa in the Holocene is recognized as being consistent with its net transport from the North Atlantic by the AMOC 1,7,20 . Our data set shows that some 231 Pa was also being exported from the Atlantic throughout the LGM and HS1.…”

“…It is unlikely that nepheloid layers, or any other scavenging mechanism, could increase this removal by an additional 30% ARTICLE of the total for longer than a few decades, much less do so without any increase in the scavenging of 231 Pa. In summary, our favoured interpretation of the existing evidence is that variable lateral export of 231 Pa out of the Atlantic by AMOC 1,15,16,20 caused the observed changes in the net sedimentary 231 Pa burial in the North Atlantic. Because the North Atlantic is bounded to the east and west by continents, and to the north by shallow ridges, the most likely candidate for the site of eventual 231 Pa sink remains the Southern Ocean 1,6 .…”

“…2a). The Holocene data from the main basin show a broad, overall decrease in 231 Pa/ 230 Th with increasing water depth throughout the water column 20,21 . The opal-rich areas of the Southern Ocean and North Atlantic have higher 231 Pa/ 230 Th ratios, reflecting the scavenging efficiency of opal 17 .…”

“…Previous studies have documented a strong relationship between sedimentary composition, specifically opal content, and the 231 Pa/ 230 Th ratio in both sediment traps 17 and downcore records [22][23][24] . Some authors have sought to minimize the effects of scavenging by excluding diatom-rich intervals 25 or high-flux sites, where 231 Pa/ 230 Th and opal flux correlate the most strongly 20 . We do not exclude any data from our 25-core depth survey; instead we use opal and particle fluxes to ascertain how much of the observed difference between our time periods may be due to changes in mass flux or opal flux ( Supplementary Figs 2-5 Supplementary Fig.…”