Biological response to millennial variability of dust and nutrient supply in. the Subantarctic South Atlantic Ocean

Transcription

1 Biological response to millennial variability of dust and nutrient supply in the Subantarctic South Atlantic Ocean Robert F. Anderson 1,2 *, Stephen Barker 3, Martin Fleisher 2, Rainer Gersonde 4, Steven L. Goldstein 1,2, Gerhard Kuhn 4, P. Graham Mortyn 5, Katharina Pahnke 6, Julian P. Sachs 7 *corresponding author <boba@ldeo.columbia.edu> 1 Lamont-Doherty Earth Observatory of Columbia University, PO Box 1000, Palisades, New York USA Department of Earth and Environmental Sciences, Columbia University 3 School of Earth and Ocean Sciences, Cardiff University, Cardiff CF10 3AT, UK 4 Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research, Am Alten Hafen 26, Bremerhaven, Germany 5 Institute of Environmental Science and Technology (ICTA) and Department of Geography, Universitat Autonoma de Barcelona (UAB), Edifici Cn Campus UAB Bellaterra Spain 6 Max Planck Research Group - Institute for Chemistry and Biology of the Marine Environment (ICBM), University of Oldenburg, Carl-von-Ossietzky-Str. 9-11, Oldenburg, Germany 7 University of Washington, Ocean Sciences Building, Room 521, Box , Seattle, WA USA Key Words: Southern Ocean, dust, iron, biological productivity Short title for page headings: South Atlantic dust, nutrients and productivity 1

2 Abstract Fluxes of lithogenic material and fluxes of three paleo productivity proxies (organic carbon, biogenic opal and alkenones) over the past 100,000 years were determined using the 230 Th-normalization method in three sediment cores from the Subantarctic South Atlantic Ocean. Features in the lithogenic flux record of each core correspond to similar features in the record of dust deposition in the EPICA Dome C ice core. Biogenic fluxes correlate with lithogenic fluxes in each sediment core. Our preferred interpretation is that South American dust, most likely from Patagonia, constitutes a major source of lithogenic material in Subantarctic South Atlantic sediments, and that past biological productivity in this region responded to variability in the supply of dust, probably due to biologically available iron carried by the dust. Greater nutrient supply as well as greater nutrient utilization (stimulated by dust) contributed to Subantarctic productivity during cold periods, in contrast to the region south of the Antarctic Polar Front (APF), where reduced nutrient supply during cold periods was the principal factor limiting productivity. The anti-phased patterns of productivity on opposite sides of the APF point to shifts in the physical supply of nutrients and to dust as co-factors regulating productivity in the Southern Ocean. 2

3 1. Introduction Establishing the contribution by the ocean s biological pump to past changes in the global carbon cycle, especially the climate-related changes in atmospheric CO 2 concentration [1], is a long-sought goal of paleoclimate research [e.g., 2, 3, 4]. Biological utilization of nutrients in the Southern Ocean is particularly important in this regard as it regulates the preformed nutrient inventory for most of the deep ocean and, therefore, the global average efficiency of the biological pump [4, 5]. Nutrient utilization is inefficient in the Southern Ocean today, in part because phytoplankton growth is limited by the scarcity of iron [e.g., 6]. Martin [7] suggested that iron limitation in the Southern Ocean may have been relieved during the Pleistocene ice ages by the greater deposition of continental mineral aerosols (dust) that was discovered in Antarctic ice cores. Dust would have supplied iron to marine phytoplankton. However, early tests of this hypothesis using geochemical indicators of biological productivity extracted from marine sediment cores concluded that the region south of the Antarctic Polar Front (APF) was characterized by lower productivity during the ice ages [8-10], contrary to the expectations from Martin s hypothesis. The principal exception to this view was derived by examining the species assemblages of diatoms preserved in glacial-age sediments, which suggested greater rather than lower ice-age productivity [11]. Despite this one dissenting finding, the common view is that the increase in dust supply did not enhance biological productivity south of the APF [for a recent synthesis see 12]. In contrast to the Antarctic zone, south of the APF, sediment records from Subantarctic sites (defined here as the zone north of the APF, extending as far as the 3

4 Subtropical Convergence) revealed ice-age enhancement of biological productivity in all sectors of the Southern Ocean [12, 13]. This is particularly true in the South Atlantic, where high levels of productivity are inferred from sediments deposited during the last ice age [14]. Initial studies concluded that ice-age productivity in the Subantarctic South Atlantic was stimulated by dust that originated in Patagonia [8], immediately upwind of the region and known to have been the primary source of dust in Antarctic ice cores [e.g., 15, 16]. However, this view was challenged by a number of studies concluding that most of the lithogenic material in South Atlantic sediments is delivered by ocean currents rather than via the atmosphere [17-27], both during ice ages and during interglacial periods. Martinez-Garcia and coworkers [28, 29] resurrected the notion that biological productivity in the Subantarctic South Atlantic is stimulated by dust in finding a tight coupling between dust and productivity over the past 4 Myr. Specifically, at ODP Site 1090 ( S, E, 3702m, Figure 1) the accumulation rates of iron and of n- alkanes (organic compounds derived from leaf waxes of land plants) were correlated with one another across glacial-interglacial cycles, as well as with geochemical indicators of biological productivity. Each of these parameters in ODP 1090 sediments was further correlated with dust deposition in the EPICA Dome C (EDC) ice core. The tight coupling between fluxes of lithogenic material and of leaf waxes, together with the correlation with EDC dust fluxes, supports the view that most of the lithogenic material in Subantarctic South Atlantic sediments is supplied as dust [8]. Although the findings of Martinez-Garcia and coworkers are compelling, correlations across glacial-interglacial cycles may be misleading. Many variables 4

5 correlate with one another over 100-kyr Milankovitch cycles simply because the enormous variability of climate boundary conditions influences these variables, without any inherent causal connection among them. Consequently, here we expand on earlier work by examining South Atlantic sediment records at much greater temporal resolution, thus affording examination of dust-productivity linkages in the absence of major changes in climate boundary condition (e.g. glacial-interglacial cycles), instead focusing on more subtle changes in forcing at higher frequency. We find a tight coupling between geochemical indicators of biological productivity and the accumulation of lithogenic material in three cores from the Subantarctic South Atlantic Ocean, and we further demonstrate that these features exhibit a close correspondence to dust deposition in the EDC ice core on millennial time scales. We conclude that phytoplankton in the region responded to increased dust supply with an increase in nutrient utilization and increased export production. 2. Materials and Methods Accumulation rates of sedimentary constituents were evaluated for three South Atlantic cores (Figure 1) using the 230 Th-normalization approach [30]. This approach makes a necessary correction for the redistribution of sediments by deep-sea currents (sediment focusing), which can cause the local accumulation of sediment to exceed the regional average particle rain rate by as much as a factor of 20 in the cores studied here [31]. Concentrations of lithogenic material were estimated for each core using the measured 232 Th concentration. Concentrations of Th isotopes ( 230 Th and 232 Th) were measured by isotope dilution using methods described elsewhere [32, 33]. 5

6 Thorium-232 abundance has a fairly narrow range in upper continental crust [34, 35], and among dust sources worldwide [36]. Here we use an average content of 10 ppm to estimate the lithogenic component of each sample, acknowledging that there is a potential uncertainty of as much as 20% in this estimate depending on the source of the lithogenic material. Assessing the temporal variability of lithogenic supply is more important here than the absolute flux, so we accept the uncertainty in the absolute flux inherent in this approach in favor of the precision that it offers for detecting small changes in the lithogenic flux over time. We make no initial assumption about the origin of the lithogenic material. Undoubtedly, for each site some fraction is delivered by ocean currents as well as by dust. We will use the amplitude of the temporal variability that correlates with dust deposition in the EDC ice core [37] to provide semi-quantitative constraints for the fraction of the lithogenic component of sediments that consists of dust. Mackensen et al. [38] presented an initial estimate of paleoproductivity for core PS (44.15 S, W, 3783m). In addition to thorium (see above) we add to the study of PS new measurements of organic carbon (C-org, determined at AWI), using an elemental analyzer. Biogenic opal was also measured, both at AWI using the method of Müller and Schneider [39] and at LDEO using the method of Mortlock and Froelich [40]. Opal concentrations measured at LDEO were systematically greater than those measured at AWI, most likely reflecting a difference in the efficiency of Si extraction [41]. Despite the systematic offset, the downcore pattern of opal content derived using the two methods was internally consistent. As for the lithogenic fluxes, the precision with which we can assess temporal variability is more important for our 6

7 objectives than the accuracy of the opal flux, so the opal results from LDEO are used without implying any judgment about the relative accuracy of the two data sets. Concentrations of C37 alkenones, a suite of organic compounds produced by coccolithophorids, and thorium have been described previously [31] for TN057-06PC4 (42.91 S, 8.9 E, 3751m) and TN PC2 (41.13 S, 7.81 E, 4981m). Here we report the summed concentration of the C 37 methyl ketones with two (C 37:2Me ) and three (C 37:3Me ) double bonds, and refer to that quantity simply as alkenones. Although more than 10 different alkenones and the related alkenoates have been reported in marine sediments, C 37:2Me and C 37:3Me are the two compounds used for paleotemperature reconstructions and are by far the most abundant alkenones in Subantarctic sediments. Those results are supplemented here with additional measurements of thorium for both cores. 3. Age models Lithogenic fluxes in PS2498-1, on the age model of Mackensen et al. [38] (Figure 2B), share a number of features with the EDC dust record (Figure 2A); for example, elevated fluxes during the intervals ~18-30 ka and ~60-70 ka, as well as three shorter intervals of elevated flux between 30 and 60 ka. Given the similarity of the records, we selected a number of tie points to align lithogenic fluxes in PS with EDC dust fluxes (solid black lines in Figure 2). In addition, the age model was modified using radiocarbon ages reported by Gersonde et al. [42] for the core top down to a depth of 151 cm (age of 22 ka). The original age-depth relationship for PS is compared with the revised age model in Figure 3. Oxygen isotope data for PS [38] placed on the 7

8 adjusted age model fit the stacked record of Lisiecki et al. [43], the benchmark for Pleistocene chronology of deep-sea sediments, at least as well as when using the original age model (not shown, but available from the lead author). For TN057-06PC4, we began with the age model of Hodell et al. [44] as modified by C. Charles (personal communication, 2000). Further adjustments to the age model were made to align the lithogenic fluxes of TN PC4 with dust deposition in the EDC record (Figure 2, dashed grey lines). As for PS2498-1, the oxygen isotope record for the adjusted age model fit the Lisiecki stack at least as well as with the original. The revised age-depth relationships for PS and TN exhibit less variability over time than with the original age models (Figure 3). While a more uniform accumulation rate does not necessarily mean that an age model has been improved, the internal consistency among all results following the adjustments to these age models (see also below) supports the use of the new age models. Several age models have been published for TN PC2. Building upon previous work [45, 46], we first tied features in the TN record to features in the oxygen isotope record of the North GRIP ice core. The absolute age of both records beyond 60 ka was then adjusted by aligning the North GRIP oxygen isotope record with the northern Alps (NALPS) speleothem record (Barker et al., in prep), which has absolute age control from U-Th dating [47]. Adjustments to the age model of the EDC ice core were then made to place it on the NALPS chronology as well, maintaining the relationship between EDC and North GRIP developed by previous studies. While further adjustments to the absolute age of the records may be necessary, more important is that 8

9 all of the marine and ice core records used here have been tied to a common age model to facilitate intercomparison of the various results. 4. Results Lithogenic fluxes in the three Subantarctic cores are compared with the EDC dust flux record in Figure 4. Fluxes of opal and of C-org in PS2498-1, as well as the lithogenic flux, are shown with the EDC dust record in Figure 5. Alkenone fluxes from TN and from TN are presented in Figure 6, together with the lithogenic fluxes from each core, to compare with the EDC dust record. Alkenone concentrations are also shown in Figure 6D to show that variability of the 230 Th-normalized alkenone flux is determined mainly by variability in the concentration of alkenones rather than by variability of 230 Th-normalized mass flux. 5. Discussion (a) Sources of lithogenic material Cores PS and TN were recovered from sites on the eastern flank of the Mid Atlantic Ridge [38] and on the Agulhas Ridge [44], respectively. Their location on elevated topography and their distance from continental margins isolates these sites from turbidites and from contour currents delivering sediment from continental sources. Therefore, one can have confidence that lithogenic sediment delivered to these sites was transported either via the atmosphere or by major ocean currents, such as the Antarctic Circumpolar Current, allowing for minor contributions from material transported by icebergs. 9

10 Lithogenic fluxes in PS and in TN (Figure 4 B, C) exhibit a close correspondence to the pattern of dust deposition in the EDC ice core (Figure 4A), including two major intervals of elevated dust flux (approximately 18 to 35 ka and 60 to 70 ka), two shorter but intense episodes of dust deposition centered at ~41 ka and at ~50 ka, as well as smaller features (e.g., ka and ~89 ka). The observed correspondence is consistent with an aeolian source for a large fraction of the lithogenic material in PS and TN Furthermore, during intervals of maximum lithogenic flux, fluxes are nearly twice as large on the Mid-Atlantic Ridge (PS2498-1) as on the Agulhas Ridge (TN057-06), in agreement with previous observations of an eastward decrease of lithogenic flux across the Subantarctic South Atlantic during the Last Glacial Maximum (LGM) [8], and consistent with expectations as dust is washed out of the atmosphere while transported downwind from a South American source. In contrast to the other cores, TN was recovered from a drift deposit in the deep Cape Basin where contour currents deposit sediment entrained along the margin of Africa [48]. Mineralogical [22] and isotopic [25] composition of lithogenic material at this location are consistent with an African source for a portion of the sediment, with an additional time-varying source of material originating at higher latitudes [22]. This combination of two major sources of lithogenic material at the site of TN creates a record of lithogenic accumulation (Figure 4D) that has a smaller amplitude of variability compared to the other cores (Figure 4B, C). Building on the prior work of Kuhn and Diekmann [22], we interpret the record of lithogenic accumulation in TN to reflect a time-varying aeolian component, similar to the record in TN057-06, superimposed on a relatively constant source of sediment from Africa. Despite its reduced amplitude of 10

11 variability, the record of lithogenic flux in TN contains most of the features that correspond to the EDC dust record, including some of the smaller peaks (e.g., ~30 ka). Our results do not permit us to quantify precisely the fraction of lithogenic material in the marine cores that was delivered as dust. Future studies involving geochemical provenance tracers will help in this regard. Nevertheless, based on our finding that the principal features in the lithogenic flux record correspond to features in the EDC dust record, we conclude that the majority of the lithogenic material in PS and in TN057-06, as well as a large fraction of lithogenic material in TN057-21, was delivered by atmospheric transport, most likely from South America. (b) Biological response to dust Biogenic constituents of the sediments are interpreted as indicators of biological production. Accumulation of biogenic material in marine sediments is influenced by preservation as well as by production, and preservation is difficult to quantify. Therefore, rather than attempt to quantify preservation, we employed two biogenic tracers (opal and organic carbon) for which preservation is not expected to co-vary. Preservation of opal is sensitive to temperature, but in the cold stable environment of the abyssal ocean the preservation of opal depends mainly on sediment accumulation rate [49]. Preservation of organic carbon, on the other hand, is most sensitive to the concentration of oxygen in bottom water [50], which is not expected to co-vary with sediment accumulation rate. Therefore, co-variability of opal and of C-org is interpreted to reflect changes in production rather than variable preservation. Accumulation rates of C-org (Figure 5C) and of opal (Figure 5D) in PS are correlated with one another, as well as with the accumulation rate of lithogenic material 11

12 (Figure 5B) and with EDC dust deposition (Figure 5A). For the eastern sites (TN057 cores) we use alkenones, long-chained ketones produced exclusively by certain haptophyte algae, predominantly the coccolithophorids E. huxleyi and G. oceanica in the open ocean [51, 52], as a paleo productivity indicator (Figure 6). Alkenones were also employed as a paleo productivity indicator by Martinez-Garcia and coworkers [28, 29] in their study of ODP1090 (Figure 1). Features in TN are less well defined because of the roughly 5-fold lower accumulation rate in this core compared to TN057-21, which allows for greater smoothing of the record by bioturbation. Nevertheless, despite the differing depositional regimes (pelagic deposition on a topographical elevation vs. deep drift deposit), the principal features of the TN alkenone record (Figure 6E) are consistent with those in TN [Figure 6D, F and reference 31], supporting the validity of these tracers as indicators of time-varying changes in biological productivity. Paleoproductivity proxy records in each Subantarctic core exhibit features that covary with the lithogenic flux in the same sediment core, even over millennial time scales (Figures 5 and 6). Tight coupling over short time intervals lessens the potential influence of changes in global climate boundary conditions, thereby strengthening the view that the productivity record reflects a biological response to varying supply of dust, and to the iron that it carries [8, 28]. However, factors other than dust may also regulate biological productivity in the Southern Ocean, and these are discussed in the next section. Before proceeding, we want to reiterate that whereas the age models for PS and TN were adjusted by tuning to the EDC dust record, the age model for TN was tuned only by alignment with the oxygen isotope record of the North GRIP ice core (Barker et al., in preparation). Therefore, the correspondence of the 12

13 paleoproductivity record of TN (Figure 6D, F) with the records from the other cores (Figures 5 and 6), as well as with the EDC dust record (Figure 6A), supports the reliability of tuning the age models of PS and TN to the EDC dust record. (c) Supply and utilization of nutrients The Subantarctic Zone is a region of persistently high nitrate concentrations (Figure 1A; phosphate, not shown, has a similar distribution). Therefore, the greater fluxes of organic material during periods of elevated dust flux, which correspond to the colder intervals in Antarctica [37], could reflect a more efficient utilization of nitrate due to relief from iron limitation, or a greater supply of nutrients, or both. Nitrogen isotopes in foraminifera-bound organic matter from ODP1090/TN have been interpreted recently to indicate greater nitrate utilization efficiency during periods of elevated dust supply [53], providing compelling evidence for iron fertilization. Although the results presented here are consistent with a strong sensitivity of biological productivity to dust supply in the Subantarctic South Atlantic Ocean, dust supply cannot be the sole factor regulating productivity. For example, all of the core sites are located in regions where surface waters today are impoverished in dissolved silicon (Si, Figure 1B). At the site of PS the flux of opal during the late glacial period (20-35 ka; g cm -2 kyr) was about six-fold greater than during the last 10 kyr (~0.05 g cm -2 kyr; Figure 5D). Nearly identical results were reported previously for a site further to the east [PS2082-2, S, E, reference 9]. Substantially greater opal fluxes during the LGM are recorded at sites further south, but still well to the north of the APF (~50 S in this region), ranging between 0.6 g cm -2 kyr [PS1754-1/-2, S, E, reference 9] and 1.6 g cm -2 kyr [RC15-93, S, W, 13

14 reference 8]. Opal fluxes this large require a source of Si much greater than exists today to sustain the diatom productivity implicated by the opal flux. A greater supply of nutrients to the Subantarctic zone during glacial periods has also been inferred from the carbon isotopic composition (δ 13 C) of planktonic foraminifera in TN [54, 55]. Many of the features in the alkenone record of TN are correlated with the δ 13 C of G. bulloides in this core, in the sense that increasing productivity corresponds to increasing concentrations of nutrients (decreasing δ 13 C; Figure 7). However, δ 13 C of planktonic foraminifera is sensitive to other factors in addition to nutrient concentration [e.g., 56, 57], so a nutrient-productivity correlation cannot be inferred unambiguously. Furthermore, over two intervals (~65 to 60 ka and ~18 to 10 ka), the paleo productivity indicator drops dramatically while the δ 13 C remains low (consistent with, but not necessarily requiring, high nutrients). Over these intervals the paleoproductivity indicator correlates much better with dust supply (Figures 5, 6). On the other hand, there are occasions when this correlation breaks down. In TN057-21, for example, alkenone concentrations are relatively high during times of low or declining lithogenic flux at approximately 39, 48 and 57 ka (Figure 6). In PS2498-1, elevated fluxes of opal and of Corg also persist until ~39 ka, despite declining dust flux (Figure 5). Each of these intervals, corresponding to Heinrich Stadials 4, 5 and 5A, coincides with low δ 13 C in G bulloides (Figure 7), which we interpret to indicate that nutrient supply may play at least as great a role as dust flux in regulating productivity at these times. Enhanced upwelling south of the APF during these intervals [58] may have injected nutrients into the Subantarctic zone [59] as well as in the Antarctic Zone. We conclude, therefore, that both increased supply of nutrients and increased nutrient 14

15 utilization efficiency, stimulated by dust, contributed to the rise in Subantarctic productivity during cold periods, including relatively cold intervals of millennial duration during Marine Isotope Stage (MIS) 3 (~30-60 ka) and late MIS 5 (prior to ~80 ka) as well as throughout MIS 2 and 4. (d) Contrasting conditions across the APF Subantarctic productivity can now be considered in the context of a substantial body of work describing patterns of productivity throughout the entire Southern Ocean. In contrast to modern (Holocene) conditions, when export production south of the APF exceeds that of the Subantarctic Zone, the situation was reversed during the LGM, when export production of the Subantarctic Zone was much greater than to the south of the APF [e.g., 8, 9, 10]. Kohfeld et al. [12] provide a comprehensive synthesis of the evidence for the last glacial period, while Jaccard et al. [60] show that the seesaw pattern of productivity across the APF was a regular response to climate variability throughout the late Pleistocene. Two general scenarios have been proposed to explain lower export production of the Antarctic Zone during glacial periods. First, expansion of sea ice during glacial periods [61] may have limited the growth of phytoplankton in the Antarctic Zone [e.g., 10, 62]. Implicit in this scenario is the corollary that a greater fraction of the nutrients upwelled south of the APF would remain unused and thus be subject to Ekman transport northward into the Subantarctic Zone. That is, the greater supply of nutrients to the Subantarctic Zone during glacial periods could simply reflect lower utilization in the Antarctic Zone superimposed on conditions of upwelling and surface water transport similar to those that exist today. The principal alternative scenario is that nutrient supply 15

16 by upwelling south of the APF was reduced during glacial periods [e.g., 63], most likely tied to increased stratification of the polar ocean [for reviews see 64, 65]. Nitrogen isotopes archived in Southern Ocean sediments have been interpreted to favor the second hypothesis (reduced nutrient supply). Throughout the Antarctic Zone, the nitrogen isotopic composition of organic compounds preserved within diatom frustules indicate greater nitrate utilization during glacial periods, coincident with lower export production, compared to interglacials [66-68]. Relatively efficient utilization of nitrate during glacial periods is inconsistent with inhibition of phytoplankton growth by sea ice, but it allows for iron fertilization in response to the greater supply of dust. That is, the lower rate of nutrient supply during glacial periods trumps the effect of iron fertilization, reducing export production within the Antarctic Zone compared to ironlimited interglacial periods. The combination of lower supply and greater utilization of nitrate south of the APF during glacial periods eliminates the northward transport of unused nutrients as the primary source to fuel Subantarctic productivity. Instead, during glacial periods nutrients must have been supplied to the Subantarctic Zone by mixing from below Today, the zone of maximum upwelling and nutrient supply is located to the south of the APF [69]. Two scenarios can be envisioned to account for the northward displacement of nutrient supply during glacial periods. First, the APF may have migrated northward by several degrees of latitude [reviewed by 61], carrying the band of upwelling with it into latitudes more favorable for phytoplankton growth. Alternatively, the hydrographic features associated with the APF may have remained locked roughly in their present position through interaction with bottom topography [70]. Under this 16

17 scenario, a northward shift of the southern westerlies [71] may have decoupled the zone of upwelling from the APF. Upwelling, driven primarily by wind stress curl in the Subantarctic Zone under these conditions, may have ventilated much shallower layers of the deep ocean than occurs today. While this scenario is speculative, ventilation of shallower water masses in the Southern Ocean during glacial periods is consistent with a growing body of evidence for greater mid-depth stratification during the LGM [72, 73]. Of course, the two scenarios are not mutually exclusive, as the westerlies and the frontal systems may have migrated concurrently. Lastly, we note that the seesaw pattern of nutrient supply across the APF that is evident over glacial-interglacial cycles is also apparent over much shorter time scales, with some exceptions. Anderson et al. [58] interpreted opal fluxes to indicate increased upwelling south of the APF during the most intense northern hemisphere stadials of the last glacial cycle, which correspond to relatively warm Antarctic Isotope Maxima (AIM). Dust fluxes to Antarctica are at relative minima during AIM events, and peak during the coldest intervals [37]. Therefore, if the lithogenic fluxes to Subantarctic sites correspond to dust fluxes in Antarctica, then the covariance between dust, export production and nutrient supply (Figures 5-7) indicates maximum supply of nutrients to the Subantarctic Zone during Antarctic cold intervals, opposite to the pattern observed south of the APF. The principal exceptions, noted in the previous section, occur during Heinrich Stadials 4, 5 and 5A, when nutrient supply to the Subantarctic Zone remained high during the transition into Antarctic warm intervals. Records with greater temporal resolution will be needed to resolve the precise phasing between temperature, dust, nutrients and productivity at these times. 17

18 6. Summary and Outlook Accumulation rates of lithogenic material in three sediment cores from the Subantarctic South Atlantic Ocean exhibit features over the past 100,000 years that correspond to the pattern of dust deposition in the EPICA Dome C ice core, indicating that most of the lithogenic material in the sediments was derived from South American dust sources. Accumulation rates of biogenic opal and of organic indicators of biological productivity also correlate with the lithogenic fluxes in these cores. Combining these results with evidence for nutrient utilization [53] and for nutrient supply, we conclude that export production in the Subantarctic Zone was stimulated during cold periods by the synergistic effects of greater nutrient supply together with increased nutrient utilization, supported by elevated dust fluxes. Nutrient supply and export production south of the APF varied in a pattern that was anti-phased with their variability in the Subantarctic Zone, whereas nutrient utilization was high in both regions during the Last Glacial Maximum (LGM) and, presumably, during earlier cold intervals as well. Conditions south of the APF indicate that export production during the LGM was limited more by the supply of macronutrients than by iron. The anti-phased pattern of nutrient supply on opposite sides of the APF, combined with the dust fluxes presented here, further indicates that the physical forcing that brings nutrients to the surface varied almost concurrently with dust supply. Climate related shifts in the latitude of the southern hemisphere westerlies have been suggested to influence nutrient supply in the Subantarctic South Atlantic [45, 74] as well as in the Antarctic Zone [58]. Meridional shifts in the southern westerlies may also 18

19 have modulated the growth and retreat of mountain glaciers in the mid latitudes of the Southern Hemisphere [75-78], which, in turn, may have regulated the source of glaciogenic sediments entrained by the winds as they passed over Patagonia [79]. Thus, a north-south displacement of the southern westerlies provides a unifying hypothesis that is consistent with the paleoclimate records described above. Future studies should consider why these features do not appear to be prominent in model simulations [for recent discussion see 80, 81]. Acknowledgements New results reported here were generated with support from the US NSF via award OCE Helpful comments from Wally Broecker, Chris Charles, Joerg Schaefer, Gisela Winckler, Aaron Putnam and Ben Bostick are much appreciated. Insightful reviews from two anonymous referees substantially improved the paper. Opal data from PS were generated by Patricia Malone, who passed away while this manuscript was under review. Her many years of support for the paleoceanography community at Lamont are much appreciated. She will be dearly missed by many friends and colleagues. Data accessibility Results presented here that have not been published previously will be submitted to the PANGAEA < and NCDC < data archives. DOI will be provided upon acceptance of the manuscript. 19

20 References 1. Lüthi, D., et al High-resolution carbon dioxide concentration record 650, ,000 years before present. Nature 453, Broecker, W.S Glacial to interglacial changes in ocean chemistry. Progress in Oceanography 11, Archer, D., A. Winguth, D. Lea, and N. Mahowald 2000 What caused the glacial/interglacial atmospheric pco 2 cycles? Reviews of Geophysics 38, Sigman, D.M. and E.A. Boyle 2000 Glacial/interglacial variations in atmospheric carbon dioxide. Nature 407, Marinov, I., A. Gnanadesikan, J.R. Toggweiler, and J.L. Sarmiento 2006 The Southern Ocean biogeochemical divide. Nature 441, de Baar, H.J.W., J.T.M. Dejong, D.C.E. Bakker, B.M. Loscher, C. Veth, U. Bathmann, and V. Smetacek 1995 Importance of iron for plankton blooms and carbon-dioxide drawdown in the Southern-Ocean. Nature 373, Martin, J.H Glacial-Interglacial CO 2 change: The iron hypothesis. Paleoceanography 5, Kumar, N., R.F. Anderson, R.A. Mortlock, P.N. Froelich, P. Kubik, B. Dittrich- Hannen, and M. Suter 1995 Increased biological productivity and export production in the glacial Southern Ocean. Nature 378, Frank, M., R. Gersonde, M.M. Rutgers van der Loeff, G. Bohrmann, C.C. Nürnberg, P.W. Kubik, M. Suter, and A. Mangini 2000 Similar glacial and interglacial export bioproductivity in the Atlantic sector of the Southern Ocean: Multiproxy evidence and implications for glacial atmospheric CO 2. Paleoceanography 15, Chase, Z., R.F. Anderson, M.Q. Fleisher, and P. Kubik 2003 Accumulation of biogenic and lithogenic material in the Pacific sector of the Southern Ocean during the past 40,000 years. Deep-Sea Research II 50, Abelmann, A., R. Gersonde, G. Cortese, G. Kuhn, and V. Smetacek 2006 Extensive phytoplankton blooms in the Atlantic sector of the glacial Southern Ocean. Paleoceanography 21, PA1013 doi: /2005pa Kohfeld, K.E., R.M. Graham, A.M. de Boer, L.C. Sime, E.W. Wolff, C. Le Quéré, and L. Bopp 2013 Southern Hemisphere westerly wind changes during the Last Glacial Maximum: paleo-data synthesis. Quaternary Science Reviews 68, Lamy, F., R. Gersonde, G. Winckler, O. Esper, A. Jaeschke, G. Kuhn, J. Ullermann, A. Martinez-Garcia, F. Lambert, and R. Kilian 2014 Increased Dust Deposition in the Pacific Southern Ocean During Glacial Periods. Science 343, Anderson, R.F., Z. Chase, M.Q. Fleisher, and J. Sachs 2002 The Southern Ocean's biological pump during the Last Glacial Maximum. Deep-Sea Research, Part-II 49, Basile, I., F.E. Grousset, M. Revel, J.R. Petit, P.E. Biscaye, and N.I. Barkov 1997 Patagonian origin of glacial dust deposited in East Antarctica (Vostok and Dome 20

21 C) during glacial stages 2, 4 and 6. Earth and Planetary Science Letters 146, Delmonte, B., P.S. Andersson, H. Schoberg, M. Hansson, J.R. Petit, R. Delmas, D.M. Gaiero, V. Maggi, and M. Frezzotti 2010 Geographic provenance of aeolian dust in East Antarctica during Pleistocene glaciations: preliminary results from Talos Dome and comparison with East Antarctic and new Andean ice core data. Quaternary Science Reviews 29, Petschick, R., G. Kuhn, and F. Gingele 1996 Clay mineral distribution in surface sediments of the South Atlantic: sources, transport, and relation to oceanography. Marine Geology 130, Diekmann, B Sedimentary patterns in the late Quaternary Southern Ocean. Deep-Sea Research Part Ii-Topical Studies in Oceanography 54, Diekmann, B., D.K. Fütterer, H. Grobe, C.D. Hillenbrand, G. Kuhn, K. Michels, R. Petschick, and M. Pirrung 2003, Terrigenous sediment supply in the polar to temperate South Atlantic: Land-ocean links of environmental changes during the Late Quaternary, in The South Atlantic in the Late Quaternary: Reconstruction of Material Budgets and Current Systems, G. Wefer, S. Mulitza, and V. Ratmeyer, Editors. Springer-Verlag: Berlin. p Diekmann, B. and G. Kuhn 1999 Provenance and dispersal of glacial-marine surface sediments in the Weddell Sea and adjoining areas, Antarctica: ice-rafting versus current transport. Marine Geology 158, Diekmann, B., G. Kuhn, V. Rachold, A. Abelmann, U. Brathauer, D.K. Futterer, R. Gersonde, and H. Grobe 2000 Terrigenous sediment supply in the Scotia Sea (Southern Ocean): response to Late Quaternary ice dynamics in Patagonia and on the Antarctic Peninsula. Palaeogeography Palaeoclimatology Palaeoecology 162, Kuhn, G. and B. Diekmann 2002 Late Quaternary variability of ocean circulation in the southeastern South Atlantic inferred from the terrigenous sediment record of a drift deposit in the southern Cape Basin (ODP Site 1089). Palaeogeography Palaeoclimatology Palaeoecology 182, Latimer, J.C. and G.M. Filippelli 2001 Terrigenous input and paleoproductivity in the Southern Ocean. Paleoceanography 16, Latimer, J.C. and G.M. Filippelli 2007 Sedimentary iron records from the Cape Basin. Deep-Sea Research Part II-Topical Studies in Oceanography 54, Latimer, J.C., G.M. Filippelli, I.L. Hendy, J.D. Gleason, and J.D. Blum 2006 Glacial-interglacial terrigenous provenance in the southeastern Atlantic Ocean: The importance of deep-water sources and surface currents. Geology 34, Krueger, S., D.C. Leuschner, W. Ehrmann, G. Schmiedl, A. Mackensen, and B. Diekmann 2008 Ocean circulation patterns and dust supply into the South Atlantic during the last glacial cycle revealed by statistical analysis of kaolinite/chlorite ratios. Marine Geology 253, Noble, T.L., A.M. Piotrowski, L.F. Robinson, J.F. McManus, C.-D. Hillenbrand, and A.J.M. Bory 2012 Greater supply of Patagonian-sourced detritus and 21

22 transport by the ACC to the Atlantic sector of the Southern Ocean during the last glacial period. Earth and Planetary Science Letters 317, Martinez-Garcia, A., A. Rosell-Mele, W. Geibert, R. Gersonde, P. Masque, V. Gaspari, and C. Barbante 2009 Links between iron supply, marine productivity, sea surface temperature, and CO 2 over the last 1.1 Ma. Paleoceanography 24, DOI: /2008PA Martinez-Garcia, A., A. Rosell-Mele, S.L. Jaccard, W. Geibert, D.M. Sigman, and G.H. Haug 2011 Southern Ocean dust-climate coupling over the past four million years. Nature 476, Francois, R., M. Frank, M.M. Rutgers van der Loeff, and M.P. Bacon 2004 Th- 230 normalization: An essential tool for interpreting sedimentary fluxes during the late Quaternary. Paleoceanography 19, PA1018 doi: /2003pa Sachs, J.P. and R.F. Anderson 2003 Fidelity of alkenone paleotemperatures in southern Cape Basin sediment drifts. Paleoceanography 18, 1082, doi: /2002pa Fleisher, M.Q. and R.F. Anderson 2003 Assessing the collection efficiency of Ross Sea sediment traps using 230 Th and 231 Pa. Deep-Sea Research Part II 50, Anderson, R.F., M.Q. Fleisher, L.F. Robinson, R.L. Edwards, J.A. Hoff, S.B. Moran, M.R.v.d. Loeff, A.L. Thomas, M. Roy-Barman, and R. Francois 2012 GEOTRACES intercalibration of 230 Th, 232 Th, 231 Pa, and prospects for 10 Be. Limnology and Oceanography: Methods 10, Taylor, S.R. and S.M. McLennan, The continental crust: its composition and evolution1985, Oxford: Blackwell Scientific Publications Rudnick, R.L. and R. Gao 2003, Composition of the Continental Crust, in The Crust, vol 3 Treatise on Geochemistry, R. Rudnick, Editor Elsevier-Pergamon: Oxford. p McGee, D., F. Marcantonio, and J. Lynch-Stieglitz 2007 Deglacial changes in dust flux in the eastern equatorial Pacific. Earth and Planetary Science Letters 257, Lambert, F., M. Bigler, J.P. Steffensen, M. Hutterli, and H. Fischer 2012 Centennial mineral dust variability in high-resolution ice core data from Dome C, Antarctica. Climate of the Past 8, Mackensen, A., M. Rudolph, and G. Kuhn 2001 Late Pleistocene deep-water circulation in the subantarctic eastern Atlantic. Global and Planetary Change, Muller, P.J. and R. Schneider 1993 An automated leaching method for the determination of opal in sediments and particulate matter. Deep-Sea Research Part I-Oceanographic Research Papers 40, Mortlock, R.A. and P.N. Froelich 1989 A simple method for the rapiddetermination of biogenic opal in pelagic marine-sediments. Deep-Sea Research Part a-oceanographic Research Papers 36, Conley, D.J An interlaboratory comparison for the measurement of biogenic silica in sediments. Marine Chemistry 63,

23 42. Gersonde, R., et al Last glacial sea surface temperatures and sea-ice extent in the Southern Ocean (Atlantic-Indian sector): A multiproxy approach. Paleoceanography 18, 1061, doi: /2002PA Lisiecki, L.E. and M.E. Raymo 2005 A Pliocene-Pleistocene stack of 57 globally distributed benthic δ 18 O records. Paleoceanography 20, PA1003 doi: 10:1029/2004PA Hodell, D.A., C.D. Charles, and F.J. Sierro 2001 Late Pleistocene evolution of the ocean's carbonate system. Earth and Planetary Science Letters 192, Barker, S., P. Diz, M.J. Vautravers, J. Pike, G. Knorr, I.R. Hall, and W.S. Broecker 2009 Interhemispheric Atlantic seesaw response during the last deglaciation. Nature 457, Barker, S., G. Knorr, M.J. Vautravers, P. Diz, and L.C. Skinner 2010 Extreme deepening of the Atlantic overturning circulation during deglaciation. Nature Geoscience 3, Boch, R., H. Cheng, C. Spoetl, R.L. Edwards, X. Wang, and P. Haeuselmann 2011 NALPS: a precisely dated European climate record ka. Climate of the Past 7, Tucholke, B.E. and R.W. Embley 1984, Cenozoic regional erosion of the abyssal sea floor off South Africa, in Interregional Unconformities and Hydrocarbon Accumulation, J.S. Schlee, Editor AAPG Memoirs: Woods Hole. p Sayles, F.L., W.R. Martin, Z. Chase, and R.F. Anderson 2001 Benthic remineralization and burial of biogenic SiO 2, CaCO 3, organic carbon and detrital material in the Southern Ocean along a transect at 170 west. Deep-Sea Research II 48, Hartnett, H.E., R.G. Keil, J.I. Hedges, and A.H. Devol 1998 Influence of oxygen exposure time on organic carbon preservation in continental margin sediments. Nature 391, Sachs, J.P., R. Schneider, T.I. Eglinton, K. Freeman, G. Ganssen, J. McManus, and D. Oppo 2000 Alkenones as paleoceanographic proxies. Geochemistry, Geophysics, Geosystems 1, 2000GC DOI: /2000GC Herbert, T.D Review of alkenone calibrations (culture, water column, and sediments). Geochemistry Geophysics Geosystems 2, 2000GC DOI: /2000GC Martinez-Garcia, A., D.M. Sigman, H. Ren, R.F. Anderson, M. Straub, D.A. Hodell, S.L. Jaccard, T.I. Eglinton, and G.H. Haug in press Iron fertilization of the Subantarctic Ocean during the last ice age. Science. 54. Ninnemann, U.S. and C.D. Charles 1997 Regional differences in Quaternary Subantarctic nutrient cycling: Link to intermediate and deep water ventilation. Paleoceanography 12, Mortyn, P., C.D. Charles, and D.A. Hodell 2002 Southern Ocean upper water column structure over the last 140 kyr with emphasis on the glacial terminations. Global and Planetary Change 34, Spero, H.J., J. Bijma, D.W. Lea, and B.E. Bemis 1997 Effect of seawater carbonate concentration on foraminiferal carbon and oxygen isotopes. Nature 390,

24 57. Kohfeld, K.E., R.F. Anderson, and J. Lynch-Stieglitz 2000 Carbon isotopic disequilibrium in polar planktonic foraminifera and its impact on modern and Last Glacial Maximum reconstructions. Paleoceanography 15, Anderson, R.F., S. Ali, L.I. Bradtmiller, S.H.H. Nielsen, M.Q. Fleisher, B.E. Anderson, and L.H. Burckle 2009 Wind-driven upwelling in the Southern Ocean and the deglacial rise in atmospheric CO 2. Science 323, Sachs, J.P. and R.F. Anderson 2005 Increased productivity in the subantarctic ocean during Heinrich events. Nature 434, Jaccard, S.L., C.T. Hayes, A. Martinez-Garcia, D.A. Hodell, R.F. Anderson, D.M. Sigman, and G.H. Haug 2013 Two modes of change in Southern Ocean productivity over the past million years. Science 339, Gersonde, R., X. Crosta, A. Abelmann, and L. Armand 2005 Sea-surface temperature and sea ice distribution of the Southern Ocean at the EPILOG Last Glacial Maximum - A circum-antarctic view based on siliceous microfossil records. Quaternary Science Reviews 24, Charles, C.D., P.N. Froelich, M.A. Zibello, R.A. Mortlock, and J.J. Morley 1991 Biogenic opal in Southern Ocean sediments over the last 450,000 years: Implications for surface water chemistry and circulation. Paleoceanography 6, Francois, R., M.A. Altabet, E.F. Yu, D.M. Sigman, M.P. Bacon, M. Frank, G. Bohrmann, G. Bareille, and L.D. Labeyrie 1997 Contribution of Southern Ocean surface-water stratification to low atmospheric CO 2 concentrations during the last glacial period. Nature 389, Sigman, D.M., A.M. de Boer, and G.H. Haug 2007, Antarctic stratification, atmospheric water vapor, and Heinrich Events: A hypothesis for Late Pleistocene deglaciations, in Ocean Circulation: Mechanisms and Impacts, A. Schmittner, J.C.H. Chiang, and S.R. Hemming, Editors. American Geophysical Union: Washington. p Sigman, D.M., M.P. Hain, and G.H. Haug 2010 The polar ocean and glacial cycles in atmospheric CO 2 concentration. Nature 466, Sigman, D.M., M.A. Altabet, R. Francois, D.C. McCorkle, and J.F. Gaillard 1999 The isotopic composition of diatom-bound nitrogen in Southern Ocean sediments. Paleoceanography 14, Robinson, R.S. and D.M. Sigman 2008 Nitrogen isotopic evidence for a poleward decrease in surface nitrate within the ice age Antarctic. Quaternary Science Reviews 27, Horn, M.G., C.P. Beucher, R.S. Robinson, and M.A. Brzezinski 2011 Southern ocean nitrogen and silicon dynamics during the last deglaciation. Earth and Planetary Science Letters 310, Speer, K., S.R. Rintoul, and B. Sloyan 2000 The diabatic Deacon cell. Journal of Physical Oceanography 30, Matsumoto, K., J. Lynch-Stieglitz, and R.F. Anderson 2001 Similar glacial and Holocene Southern Ocean hydrography. Paleoceanography 16, Toggweiler, J.R., J.L. Russell, and S.R. Carson 2006 Midlatitude westerlies, atmospheric CO 2, and climate change during the ice ages. Paleoceanography 21, doi /2005pa

25 72. Hoffman, J.L. and D.C. Lund 2012 Refining the stable isotope budget for Antarctic Bottom Water: New foraminiferal data from the abyssal southwest Atlantic. Paleoceanography 27, PA1213 doi: /2011pa Lund, D.C., J.F. Adkins, and R. Ferrari 2011 Abyssal Atlantic circulation during the Last Glacial Maximum: Constraining the ratio between transport and vertical mixing. Paleoceanography 26, PA1213 doi: /2010pa Rodríguez-Sanz, L., P. Graham Mortyn, A. Martínez-Garcia, A. Rosell-Melé, and I.R. Hall 2012 Glacial Southern Ocean freshening at the onset of the Middle Pleistocene Climate Transition. Earth and Planetary Science Letters , Denton, G.H., R.F. Anderson, J.R. Toggweiler, R.L. Edwards, J.M. Schaefer, and A.E. Putnam 2010 The last glacial termination. Science 328, Hall, B.L., C.T. Porter, G.H. Denton, T.V. Lowell, and G.R.M. Bromley 2013 Extensive recession of Cordillera Darwin glaciers in southernmost South America during Heinrich Stadial 1. Quaternary Science Reviews 62, Kaplan, M.R., J.M. Schaefer, G.H. Denton, D.J.A. Barrell, T.J.H. Chinn, A.E. Putnam, B.G. Andersen, R.C. Finkel, R. Schwartz, and A.M. Doughty 2010 Glacier retreat in New Zealand during the Younger Dryas stadial. Nature 467, Putnam, A.E., et al Warming and glacier recession in the Rakaia valley, Southern Alps of New Zealand, during Heinrich Stadial 1. Earth and Planetary Science Letters 382, Sugden, D.E., R.D. McCulloch, A.J.M. Bory, and A.S. Hein 2009 Influence of Patagonian glaciers on Antarctic dust deposition during the last glacial period. Nature Geoscience 2, Sime, L.C., K.E. Kohfeld, C. Le Quéré, E.W. Wolff, A.M. de Boer, R.M. Graham, and L. Bopp 2013 Southern Hemisphere westerly wind changes during the Last Glacial Maximum: model-data comparison. Quaternary Science Reviews 64, Völker, C. and P. Köhler 2013 Responses of ocean circulation and carbon cycle to changes in the position of the Southern Hemisphere westerlies at Last Glacial Maximum. Paleoceanography 28, Schlitzer, R Ocean Data View Garcia, H.E., R.A. Locarnini, T.P. Boyer, J.I. Antonov, M.M. Zweng, O.K. Baranova, and D.R. Johnson, World Ocean Atlas 2009, Volume 4: Nutrients (Phosphate, Nitrate, and Silicate)2010, Washington, D. C.: U. S. Government Printint Office. 25

26 Figure Captions Figure 1. Location of marine sediment cores discussed in this paper plotted on maps of mean annual concentration of (A) nitrate and (B) silicic acid in surface waters. Maps were prepared using Ocean Data View [82] using data from the World Ocean Atlas 2009 [83]. Figure 2. (A) Flux of dust in the EPICA Dome C (EDC) ice core [37]. (B) Flux of lithogenic sediment in core PS estimated from 230 Th-normalized fluxes of 232 Th (see text for details) on the original age model of Mackensen et al. [38]. Age control points used to adjust the age model of PS by aligning the lithogenic flux with the EDC dust flux are shown as solid black lines. (C) Flux of lithogenic sediment in core TN estimated from 230 Th-normalized fluxes of 232 Th on the age model of C. Charles (personal communication, 2000). Age control points used to adjust the age model of TN by aligning the lithogenic flux with the EDC dust flux are shown as dashed grey lines. Figure 3. Age depth relationships for PS (A) and TN (B). Original age models from Mackensen et al. [38] for PS and from C. Charles (personal communication) for TN are shown in grey. Age control points for the adjusted age models are shown as triangles for PS and as squares for TN For PS2498-1, the revised age model was fit to the radiocarbon dates of Gersonde et al. [42] to a depth of 151 cm. Age control points for depths greater than 151 cm in PS2498-1, 26

27 and for all depths in TN057-06, were derived by tuning to the EDC dust record (Figure 2). Figure 4. (A) Flux of dust in the EPICA Dome C ice core [37]. Fluxes of lithogenic sediment in Subantarctic South Atlantic cores (B) PS2498-1, (C) TN and (D) TN estimated from 230 Th-normalized fluxes of 232 Th (see text for details). Age models for PS and TN were tuned to the EDC dust record as illustrated in Figures 2 and 3. The age model for TN was derived by tuning to the North GRIP ice core (see text) and, therefore, it is independent of the EDC dust record. Figure 5. (A) Flux of dust in the EPICA Dome C ice core [37]. Fluxes of (B) lithogenic sediment, (C) organic carbon and (D) biogenic opal in Subantarctic South Atlantic core PS Fluxes of sediment constituents were estimated by 230 Th-normalization (see text for details). Figure 6. (A) Flux of dust in the EPICA Dome C ice core [37]. (B) Flux of lithogenic sediment in TN (C) Flux of lithogenic sediment in TN (D) Concentration of alkenones in TN from [31]. (E) Flux of alkenones in TN from [59] supplemented by new 230 Th data. (F) Flux of alkenones in TN Records from TN are in grey to help distinguish between the two TN057 sediment cores. Fluxes of sediment constituents are estimated by 230 Th-normalization (see text for details). 27

28 Figure 7. Concentration of alkenones (grey) and δ 13 C of planktonic foraminifera G. bulloides (black) in TN Alkenone concentrations are from [31] and carbon isotope data are from [55]. 28

29 Figure 1. 29