from stage6 to stage5e where the maximum rates amount to 9 cm/kyr, although the lithology is equival- ent. Correspondingly, the diatomaceous ooze core sec-.
Geol Rundsch (1996) 85 : 554–566
Q Springer-Verlag 1996
ORIGINAL PAPER
M. Frank 7 R. Gersonde 7 M. Rutgers van der Loeff G. Kuhn 7 A. Mangini
Late Quaternary sediment dating and quantification of lateral sediment redistribution applying 230Thex: a study from the eastern Atlantic sector of the Southern Ocean Received: 1 June 1995 / Accepted: 16 February 1996
Abstract High-resolution records of the natural radionuclide 230Th were measured in sediments from the eastern Atlantic sector of the Antarctic circumpolar current to obtain a detailed reconstruction of the sedimentation history of this key area for global climate change during the late Quaternary. High-resolution dating rests on the assumption that the 230Thex flux to the sediments is constant. Short periods of drastically increased sediment accumulation rates (up to a factor of 8) were determined in the sediments of the Antarctic zone during the climate optima at the beginning of the Holocene and the isotope stage 5e. By comparing expected and measured accumulation rate of 230Thex, lateral sediment redistribution was quantified and vertical particle rain rates originating from the surface water above were calculated. We show that lateral contributions locally were up to 6.5 times higher than the vertical particle rain rates. At other locations only 15% of the expected vertical particle rain rate were deposited. Key words High-resolution dating 7 230Thex constant flux models 7 Sediment focusing 7 Sediment winnowing 7 Particle flux reconstruction
Introduction Paleoceanographic reconstruction from sediment cores depends greatly on an accurate stratigraphy. However, M. Frank (Y) 1 7 A. Mangini Heidelberger Akademie der Wissenschaften, INF 366, D-69120 Heidelberg, Germany Fax: c06221/563405 e-mail: fk6uphys1.uphys.uni-heidelberg.de R. Gersonde 7 M. Rutgers van der Loeff 7 G. Kuhn Alfred-Wegener-Institut für Polar- und Meeresforschung, Columbusstrasse, D-27568 Bremerhaven, Germany Present address: 1 University of Oxford, Department of Earth Sciences, Parks Road, Oxford OX1 3PR, England Tel.: c44 18 65 27 20 55, FAX: c44 18 65 27 20 72, E-Mail: martinf6earth.ox.ac.uk
in high latitudes the application of the d 18O method is difficult because of low concentrations of biogenic carbonate and 14C-dating being restricted to the past 35 kyr or so (1 kyrp1000 years). It is now possible to measure oxygen isotopes in very small carbonate samples. Thus, a reliable time frame can be constructed, by correlation to standard isotope stratigraphies (Imbrie et al. 1984; Martinson et al. 1987) and, thus calibrated, biofluctuation stratigraphies are useful as independent dating tools (Burckle and Cooke 1983; Abelmann and Gersonde 1988). For many of the sediments in high southern latitudes, however, the d 18O method is not applicable at all, because the foraminifera are dissolved in parts or all of the cores. Although d 18O records may be affected by regional influence of meltwater and thus difficult to interpret (cf. Niebler 1995; Vogelsang 1990), they do generally allow calculation of sedimentation rates between the major climatic transitions by interpolation (Martinson et al. 1987). This method, however, tends to miss short-term fluctuations in sedimentation rates within the major climatic stages of the order of several hundred to several thousand years. In the area of concern, such changes are linked to significant variations of the sea-ice distribution and the hydrography of the surface water, and thus to short-term climate change (Gersonde et al., in preparation). The first goal of this study was to date sediments of a transect crossing the eastern Atlantic sector of the Antarctic circumpolar current (ACC) using the 230Thex method to confirm earlier datings based on other methods. Data on short-term changes in sedimentation rates are obtained by generating high-resolution profiles of the natural radionuclide 230Th. Resedimentation phenomena: focusing and winnowing An important problem for paleoceanographic reconstructions based on sediment accumulation rates is the quantification of lateral sediment redistribution. Lat-
555
eral transport affects particles which sink through the water column and also causes resedimentation of already deposited particles especially adjacent to shallow shelf areas. The motor of this process is either bottomcurrent activity or transport in the nepheloid bottom layer (cf. McCave 1983). Fine-grained particles are preferentially affected by resedimentation. If employing accumulation rates (e.g., of Corg, biogenic opal, or biogenic barium) to calculate proxy fluxes, the assumption is that lateral redistribution of sediments is unimportant. Yet, such redistribution may completely mask original vertical particle flux (rain rate) in the Southern Ocean, as already shown by Hays et al. (1976). The lateral contribution may amount to more than 10 times the vertical particle flux as demonstrated for Southern Ocean sediments by means of 230 Thex normalization (Francois et al. 1993; Kumar 1994). A strong lateral contribution to sediment trap material sampled in the Bransfield strait and at the Maud rise, which are close to topographic elevations, was deduced from particle fluxes into sediment traps which were up to a factor of 10 higher in the deep traps than in the shallow ones. A significant part of this laterally transported material had been resedimented from the shallow water sediments, as documented by the high amounts of lithogenic components derived from the shelves (Abelmann and Gersonde 1991). Redistribution of biogenic silica from the Maud rise was also reported by van Bennekom et al. (1988). The development of modern sonar systems on board of most research ships (e.g., Parasound system of RV “Polarstern”) has facilitated finding places with a sufficiently thick Quaternary sediment cover for high-resolution studies (Gersonde and Hempel 1990; Gersonde Fig. 1 Locations of the five sediment cores forming the transect over the eastern Atlantic sector of the Antarctic circumpolar current (ACC). The present-day position of the fronts is taken from Peterson and Stramma (1991), and the data of the sea ice distribution are from the Sea Ice Climatic Atlas (1985)
1993). Thus, quite commonly sediments are preferentially sampled at locations of thick cover where lateral supply (focusing) contributed sediment. The second goal of our study therefore was to quantify lateral sediment redistribution applying 230Thex and by estimating the vertical rain rates of particles to the sediments which originate directly from the water column above and comparing these with accumulation rates. Presuming that the supply of a particular proxy in the vertical flux is in the same proportion as in the lateral flux, the vertical rain rates may also deliver a reliable basis for the calculation of proxy accumulation rates. Study area The sediments studied are located below the eastern Atlantic sector of the ACC, which presently is the ocean’s largest current system with respect to volume transport (ca. 130 Sv) and which in many places extends to the sea floor. Two main hydrographic fronts, the Subantarctic front (SAF) and the Antarctic polar front (APF; Fig. 1), separate warm and nutrient- and H4SiO4–poor surface waters in the north (Subantarctic zone, SAZ) from cold and nutrient- and H4SiO4–rich surface waters in the south (Antarctic zone, AZ). The area between the APF and the SAF is called polar frontal zone (PFZ; see Peterson and Stramma 1991, for a summary). In the AZ and parts of the PFZ, particle concentrations in the surface waters and also particle fluxes are seasonally very high, and particle composition is dominated by biogenic opal, mainly diatoms, resulting in the
556
opal belt of the Southern Ocean. The southern boundary of this high particle flux area is the seasonally variable northern limit of the sea ice cover. Below the sea ice the production of diatoms is diminished greatly (cf. Hays et al. 1976; Gersonde and Zielinski, in preparation). Geochemistry of
230
Th
230 Th (t1/2p75 200 years) is produced by decay of 234U in the water column. The supply rate of 230Th is assumed constant because of the homogenous U concentration of 3.3 mg/l in the ocean, which suggests that U is unaffected by biological cycling. The production amounts to 2.6 disintegrations per minute (dpm) 7 cm –2 . kyr –1 7 1000 m water column –1. 230Th is highly particle-reactive, i.e., it is quickly adsorbed to particles. The mean residence time of 230Th in the oceans’ water column is between 5 and 40 years (Anderson et al. 1983; see Frank et al. (1994) for a summary on the geochemistry of 230Th). The equilibrium activity of 234U which is present in the sediment particles and the 230Th activity produced from authigenic U must be subtracted from the measured 230Th activity to get the unsupported activity ( 230Thex), which originates from the water column (see Francois et al. 1993). Due to the high particle-reactivity, basin-to-basin transport of 230Th can be considered negligible, so that regionally the deposition of 230Thex should more or less match its production in the water column. However, where particle flux is exceptionally high, 230Thex may be imported from surrounding areas (boundary scavenging). Thus, under conditions of strong upwelling and high mass flux, as for example in the upwelling areas off West Africa or off Peru, isopycnal transport of 230 Th may lead to an increased flux of 230Thex to the sediments (Mangini and Diester-Haass 1983). From sediment trap data Francois et al. (1990) deduced that this lateral import may amount to 50% of the 230Th production (normalized to water depth) in the Panama basin, corresponding to a high particle flux of approximately 10 g 7 cm –2 . kyr –1. At particle fluxes of 5 g 7 cm –2 kyr –1 or below no excess of the production was observed. Such an isopycnal supply of 230Th may also be important south of the APF and north of the sea ice edge in the Atlantic sector of the Southern Ocean, where upwelling of deep and 230Th-enriched water masses may cause increased supply of 230Th to the water column and thus also accumulation rates of 230Th which may regionally exceed local production (Rutgers van der Loeff and Berger 1993). The maximum lateral contribution of 230Th by advectional transport in the water column should also be approximately 50% in excess of production in the water column. This was estimated from sediment traps from the high particle flux area in the Bransfield strait and the King George basin, where a large contribution by laterally transported material was deduced from the 210Pb content of the trap material (Rutgers van der Loeff and Berger 1991).
Keeping in mind the above uncertainties, the flux of Thex to the sea floor is assumed to match its production, both globally and regionally. If that is the case at a certain location, the 230Thex method may be used as a dating tool for sediments back to approximately 300 kyr. If the calculated flux of 230Thex to the sea floor deviates significantly from the expected value, it is ascribed to lateral redistribution of sediment particles, loaded with adsorbed 230Thex. 230
Material and methods 230
Thex constant flux modeling
If a constant flux of 230Thex to the sediments is assumed, variations of the concentrations of initial (corrected for decay) 230Thex in the sediment may be interpreted directly in terms of variations of dilution with accumulating sediment, and thus in terms of variation of sedimentation rate (cf. Francois et al. 1990). Using a 230 Thex constant flux model it is possible to calculate both instantaneous sedimentation rate and age for each sample. Assuming that lateral isopycnal transport of 230 Th is negligible, errors include statistical uncertainties of the 230Th measurement and errors of the independent stratigraphies. The flux or accumulation rate F of 230Thex is calculated as FpC(x)7S7s
(1)
where C(x) is the concentration of (decay-corrected) 230 Thex at a depth x in the sediment, S is the sedimentation rate, and s is the dry bulk density. It follows for the uppermost sample of a sediment core S1 p
F Dx1 p C(x1)7s1 Dt1
(2)
where Dx1 is the depth-in-core interval of the slit sample and Dt1 is the period of time represented by this slit sample. C(x 1) is the decay-uncorrected 230Thex concentration because it is the uppermost sample. Solving for Dt1 gives for the uppermost sample:
Dt1 p
Dx17C(x1)7s1 F
(3)
With this relation Dt1, the age of the period of time covered by the first sample of the sediment core, is calculated and the 230Thex concentration of the second sample is then decay-corrected. Dt2 of the second sample is then
Dt2 p
Dx27C(x2)7s27e l( F
230
TH)7t 1
(4)
where C(x2) is the decay-uncorrected 230Thex concentration of the second sample and t1 is the age of the lower boundary of the first sediment interval (if the first sample interval is 0–10 cm, then t1 corresponds to a depth of 10 cm). All other time intervals may be calculated
557
this way. The age t for a certain depth x is then calculated by adding all Dt’s from the top of the core down to x. If no additional stratigraphic information on a sediment core is available, the average 230Thex flux, calculated from the average sedimentation rate estimated by the 230Thex–dating method is inserted for F. This method simply works by exponentially fitting the downcore decreasing 230Thex concentrations and thus determining an average sedimentations rate and an average flux for the whole sediment core. In case there is more stratigraphic information, the flux F is calculated for certain age intervals between fixed time horizons so that the sum of the Dt’s corresponds to the age previously fixed. The instantaneous sedimentation rate of a sample n is calculated as Sn p
Dxn Dtn
(5)
If F (Fa) of 230Thex deviates significantly from the expected production value (Fpp2.63 dpm 7 cm –2 7 kyr –1 7 1000 m water column –1) within an independently dated sediment section, redistribution processes must have occurred. Nevertheless, assuming constant flux and constant sediment redistribution, sedimentation rates and ages may be calculated in such sections as described above and yield a better evaluation of the “true” pattern of the sedimentation rates than an interpolation between two age-fixed points. Fa/Fp ratios ~1 indicate a lateral export of particles loaded with 230Thex (winnowing) and Fa/Fp ratios `1 indicate an import (focusing; Bacon 1984; Suman and Bacon 1989; Francois et al. 1990, 1993; Scholten et al. 1994; Frank et al. 1995). Thus, average vertical sediment rain rates (Fvertical) for an independently dated sediment section may be calculated by dividing the measured accumulation rate (MAR) by the Fa/Fp ratio: Fvertical p
MAR Fa/Fp
(6)
Further assuming that the flux of 230Thex and the Fa/ Fp ratio were constant within a given section in the sediment (Francois et al. 1993) the vertical rain rate Fivertical of any particulate substance i may be calculated for each sample by Fivertical p
P7z7fi C( 230Th ex)
(7)
with p is production of 230Th in the water column; z is water depth; fi is weight fraction or concentration of i in the sediment; C( 230Th ex) is decay-corrected 230Thex concentration One of the great advantages of the calculation of Fivertical consists of a near independency from stratigraphic errors in assigned ages up to B20 kyr. Whereas an error of 20 kyr, e.g., between oxygen isotope stage boundaries in the case of incomplete records due to carbonate dissolution, will significantly change the interpolated sediment accumulation rates, it will hardly affect
the decay-corrected 230Thex concentration and thus will also not influence Fivertical significantly due to the long half-life of 230Th. This approach is limited by some uncertainties: The dissolution of labile biogenic phases in the sediments may alter the ratio between fi and C( 230Th ex) (Yang et al. 1990), which means that the rain rates presented are referred to as “preserved fluxes” of component i. Another uncertainty may arise from a grain-size fractionation between bigger particles of biogenic origin such as radiolarians and fine-grained particles such as aluminosilicates. The bigger particles may not be affected by lateral transport as strongly as the fine-grained fraction (Scholten et al. 1994). It follows that, in the case of focusing, the rain rates represent a minimum estimate and, in the case of winnowing, they are a maximum value. However, if strong bottom currents account for lateral particle transport (above 10 cm/s; see Pudsey 1992), the effect of 230Thex grain size fractionation should be small (Frank et al. 1995). Core material Five gravity cores and the corresponding multicores, delivering undisturbed surface sediments, forming a transect over the ACC from 567S to 437S, were analyzed in this study (Fig. 1; Table 1). All cores were recovered during “Polarstern” cruise ANT VIII/3 (Gersonde and Hempel 1990), except the northernmost PS2082-1 and PS2082-3, which were recovered during “Polarstern” cruise ANT IX/4 (Bathmann et al. 1992). In Fig. 2 the lithology of the sediments is displayed. The transect starts at 567S in the AZ (core PS1772-8 and multicore PS1772-6), where the cores were recovered at a position roughly corresponding to the present (1973–1982) average northern extent of the sea ice maximum (Sea Ice Climatic Atlas 1985; Fig. 1). The location of the cores is on the southern slope of the SW Indian ridge. Sediments there consist mainly of diatomaceous mud with two layers of lightly colored diatomaceous ooze (150–360 and 600–675 cm). Only the upper 640 cm of the core are included in this study. Approximately 27 north of the present average sea ice maximum, gravity core PS1768-8 and multicore PS1768-1 were recovered north of Bouvet Island in the permanently sea-ice-free AZ (Fig. 1). This core also consists of the typical sedimentary components of the opal belt, which are alternating layers of diatomaceous muds and diatomaceous oozes (oozes at 0–138, 595– 625, 640–695, and 730–830 cm core depth; Gersonde and Hempel 1990). From the PFZ two sediment cores were analyzed. Core PS1756-5 and multicore PS1756-6 were recovered from a location approximately 17 north of the presentday position of the Polar front, and south of Meteor rise. The core shows a monotonous record of diatomaceous mud with only one layer of diatomaceous ooze (673–760 cm). The top of the piston core (0–12 cm) and
558 Fig. 2 Lithological description of the gravity cores taken from the cruise reports (Gersonde and Hempel 1990; Bathmann et al. 1992)
of the multicore consist of foraminiferal ooze and siliceous foraminiferal ooze. At the second location from the PFZ, core PS1754-1 and multicore PS1754-2 were recovered at a position approximately 17 south of the Subantarctic front at the bottom of a morphological depression at the foot of a submarine mountain on Meteor rise, close to Ocean Drilling Project (ODP) site 704. The mountain tops at a water depth of approximately 1000 m. The sediment consists of siliceous foraminiferal ooze, with the upper 20 cm being foraminiferal ooze. The transect ends at 437S in the Agulhas basin with core PS2082-1 and multicore PS2082-3 from a position north of the Subantarctic front. The core site is located at the bottom of a small morphological depression. Several layers of calcareous nannoplankton mud appear in the diatomaceous mud. The top of the core (upper
50 cm) is enriched in nannofossils; the layer between 6 and 19 cm consists of foraminiferal ooze. Only the upper 1091 cm of the entire recovery of 1391 cm were analyzed because of the expected age limitation of the 230 Thex measurements. The gravity cores were sampled continuously as 25 cm slit samples and then split in the lab with a resolution between 2 and 25 cm per sample depending on macroscopical changes in lithology and known stratigraphic transitions. The measurement of each sample thus represents the average value for the corresponding depth interval. The 230Th activities were determined by a-spectroscopy following a chemical preparation as described by Frank et al. (1994). Physical properties (wet bulk density in grams per cubic centimeter and porosity in percent) for all five cores were analyzed for the calculation
559 Table 1 Locations of the gravity cores (GC) and multicores (MUC) of the transect over the ACC
Name
Instrument
Location
PS1772-8 PS1772-6 PS1768-8 PS1768-1 PS1756-5 PS1756-6 PS1754-1 PS1754-2 PS2082-1 PS2082-3
GC MUC GC MUC GC MUC GC MUC GC MUC
55 727.5 bS, 55 727.5 bS, 52 735.6 bS, 52 735.5 bS, 48 743.9 bS, 48 753.7 bS, 46 746.2 bS, 46 746.0 bS, 43 713.2 bS, 43 713.1 bS,
1 709.8 bE 1 710.0 bE 4 728.5 bE 4 727.6 bE 6 742.8 bE 6 743.7 bE 7 736.7 bE 7 736.1 bE 11 744.3 bE 11 745.5 bE
Water depth (m)
Total recovery (cm)
Sampled core length (cm)
4135 4136 3270 3298 3787 3803 2471 2476 4610 4661
1329 24.5 896 34 862 27 356 25 1391 26
640 24.5 896 34 862 27 200 25 1091 26
Fig. 3 Results of the 230Thex dating of the five cores: the uncorrected 230Thex concentrations (in dpm/g) are plotted vs depth in centimeters. The dashed curves are the exponential fits to the downcore decreasing 230Thex concentrations. The error bars represent the statistical errors of one standard deviation from the mean. The drastic decreases in the 230Thex concentrations from 360 to 150 cm depth in core PS1772-8 and from 20– 160 and 780–830 cm depth in core PS1768-8 are omitted for the exponential fitting because the sedimentation rates in these sections were significantly different from the mean
of the dry bulk density with an uncertainty smaller than 5%: porosity dry bulk densitypwet bulk density– 1.0267 100
1
2
Results In Fig. 3 the decay-uncorrected 230Thex concentrations and the exponential fits of the 230Thex dating are
shown. From these fits average sedimentation rates for the whole cores may be calculated. The results of the 230 Thex dating, which only represent coarse estimates due to the deviations of the measured 230Thex concentration records from their exponential fits, are in good accordance with previous stratigraphic work for 4 of the 5 cores (Table 2), with the exception of PS1768-8 (see Discussion). In PS1756-5 no significant downcore decrease in the 230 Thex concentrations is observed. Two causes are pos-
560 Table 2 Comparison of average sedimentation rates deduced from previous studies and from the 230Thex-dating method. Assuming a mean error of the 230Thex-dating method of about 10%, the average sedimentation rates are in good accordance with each other, except core PS1768-8 (see text) Core
PS1772-8 PS1768-8 PS1756-5 PS1754-1 PS2082-1
Average sedimentation rate previous work (cm/kyr)
230 Thex dating method (cm/kyr)
1.9 6.4 5.9 1.2 4.3
2.2 4.4 P 1.4 4.8
at three core depths (Table 3). The raw 14C ages were reservoir-corrected according to Bard (1988) and recalibrated vs U/Th ages after Bard et al. (1990). Refined age information from modeling
230
Thex constant flux
PS1772-8
sible: Either the sedimentation rates in this core were so high that no significant decay of 230Th has occurred, or the sedimentation rates were constantly increasing through time, or both. In between the time marks given in Table 3, refined age information was obtained from 230 Thex constant flux modeling, which resulted in the final dating of the sediments in Figs 4 and 5. Baseline stratigraphic age interpretation was focused on the determination of isotopic stage and substage boundaries using integrated lithostratigraphic, isotope stratigraphic, and siliceous microfossil biofluctuation stratigraphies (Mackensen et al. 1994; Zielinski 1993; Niebler 1995; Fischer et al., in preparation; and unpublished results of A. Abelmann, G. Bohrmann, R. Gersonde, H.-W. Hubberten, A. Mackensen, and G. Kuhn). The absolute age assignment of isotope stage boundaries is after Martinson et al. (1987). The uppermost section of core PS1768-8 was dated by AMS- 14C
In the upper part of core PS1772-8 the stage boundaries at 12 and 130 kyr (Martinson et al. 1987) were fixed by ascribing the diatomaceous ooze sections (Fig. 2) to interglacials. Diatom stratigraphy corroborates and completes this interpretation by correctly fixing the 2–3 and 4–5 stage boundaries (R. Gersonde, unpublished results). Initial 230Thex values decrease from an average of 13 dpm/g to approximately 2 dpm/g in the upper ooze section (150–360 cm) and to approximately 5 dpm/ g in the lower ooze section (Fig. 4). These 230Thex variations, which are interpreted as changes in the sedimentation rates by 230Thex constant flux modeling, amount to between 1 and 4 cm/kyr in the diatomaceous mud sections and increase to values of up to 33 cm/kyr in the stage 5 ooze section and up to 6 cm/ky in the stage 9 ooze section. From the 230Thex constant flux modeling it was calculated that the diatomaceous ooze section between 150 and 360 cm core depth was deposited in only approximately 11 kyr (from ca. 130–119 kyr B.P.) representing the beginning of interglacial substage 5e. During the isotope stages 2–4 no significant changes in the sedimentation rates occurred. Isotope stage 7 is not represented by a diatomaceous ooze, and the lower ooze section is attributed to isotope stage 9.
Table 3 Age models of the cores. For core depths without additional marks, ages were assigned by d 18O-stratigraphy (Niebler 1995; Mackensen et al. 1994) following Martinson et al. (1987). Depths in parentheses mark diatom stratigraphic results (Gersonde, unpubl.), which for PS1756-5 were combined with a d 13C record of the organic matter (Fischer et al., in prep.). Depths
marked with Th indicate that these values were fixed by 230Thex constant flux modelling. The ages and depths of core PS1768-8 marked with * were determined by 14C-AMS measurements of organic matter (Gersonde et al., in prep.). Between these age fixpoints, the age of each sample was calculated by 230Thex constant flux modelling
d 18O stage boundary or event
Age (kyr)
P P 1/2 P 2/3 3.13 3/4 4/5 5.4 5/6 6.4 6/7 7.4 7/8 8/9
10.06* 11.26* 12 14.16* 24 44 59 74 111 130 153 190 225 244 300
Depth (cm) PS1772-8
PS-1768-8
PS1756-5
PS1754-1
PS2082-1
P P (17) P (55) P (80) (100) 136 Th (360) P (460) P (530) (590)
54* 78* 95 Th 142* 244 P 469 585 730 830 P P P P P
P P (12) P (240) P (440) (530) (720) (770) P P P P P
P P 22 P 54 P 78 88 140 Th 160 P P P P P
P P 33 P 245 300 370 440 523 570 730 900 980 1040 P
561 Fig. 4 Initial decay-corrected concentrations of 230Thex (in dpm/g)vs age (in kyr). The lightly shaded sections mark the interglacial isotope stages and the darkly shaded section marks the stage 5e following Martinson et al. (1987). The error bars represent the statistical errors of one standard deviation from the mean. Note the different scales of the profiles
PS1768-8 For this core a d 18O record of the planktonic foraminferal species Neogloboquadrina pachyderma sin. was established (Niebler 1995). Three AMS 14C datings in the upper 142 cm of the core showed that the diatom ooze deposition already started at approximately 14.7 kyr B.P. (Gersonde et al., in preparation). The diatomaceous ooze sections in the lower part of the core are attributed to the three warm substages of interglacial stage 5. The initial 230Thex concentrations show significant decreases from 5–10 dpm/g to approximately 2 dpm/g during the end of the last deglacial period and the beginning of the Holocene (14.7–10 kyr B.P.) and during the isotope substage 5e. Applying the 230Thex constant flux model, a huge increase in the sedimentation rates from approximately 5–9 cm/kyr to up to 25 cm/kyr during this period (14.7–10 kyr B.P.) is determined, which is corroborated by the 14C datings. Such an increase in the sedimentation rate did not occur at the transition from stage 6 to stage 5e where the maximum rates amount to 9 cm/kyr, although the lithology is equivalent. Correspondingly, the diatomaceous ooze core section from 755–830 cm was deposited during approxi-
mately 12.5 kyr (130–117.5 kyr B.P.). A slight decrease in sedimentation rates from 5–9 cm/kyr to 3–5 cm/kyr is observed in the ooze sections representing the warm interglacial substages 5a and 5c. PS1756-5 Due to the lack of lithological changes and of biogenic carbonate, it was difficult to establish a stratigraphy in core PS1756-5. Further age information from the decrease in the 230Thex concentrations was not possible due to the insignificant decrease below 10 cm core depth (Fig. 3). The stratigraphy is based mainly on a d 13C profile, measured at the organic matter of the sediment (Fischer et al., in preparation) and diatom stratigraphic results (R. Gersonde, unpublished results). The initial 230Thex profile yields highest values of approximately 10 dpm/g during the Holocene and the sections representing stages 5 and 6 (Fig. 4). Decreased concentrations around 3 dpm/g are observed during stages 2, 3 and 5e. The 230Thex constant flux model showed that the section representing isotope stage 2 is characterized by relatively high sedimentation rates between 17 and 25 cm/kyr compared with 1–10 cm/kyr in
562
the stage 1 and 3–5 sections. There was a rapid decrease in the sedimentation rates at the beginning of the Holocene from approximately 13 cm/kyr to approximately 1 cm/kyr within a period of approximately 3.5 kyr (ca. 12–8.5 kyr B.P.). The section corresponding to stage 5e was fixed between 720 and 770 cm.
PS1754-1 The stratigraphy is deduced from d 18O records of the foraminiferal species Neogloboquadrina pachyderma sin., Globigerina bulloides, and Globigerina inflata (Niebler 1995). Due to severe disturbances in the lower part of the core, only the section above 200 cm core depth was included in the interpretation. The initial 230Thex concentrations vary between 3 and 6 dpm/g and show peaks during the Holocene, at the beginning of stage 5e and at the end of stage 5 (Fig. 4). The 230Thex constant flux model yields generally low sedimentation rates which show only slight variations with small increases to approximately 2–3.5 cm/ kyr in the stage 2 and 6 sections, compared with 0.5– 2 cm/kyr in stages 1 and 3–5. The decrease in the sedimentation rates at the end of glacial stage 2 preceded the isotope stage boundary at 12 kyr by approximately 4 kyr.
PS2082-1
d 18O stratigraphies of the benthic foraminiferal species Cibcidoides spp. and Fontbotia wuellerstorfi, and the planktonic species Neogloboquadrina pachyderma sin. and Globigerina bulloides, were established for core PS2082-1 (Mackensen et al. 1994). The profile of the initial 230Thex shows relatively high values between 10 and 40 dpm/g with maxima in the interglacial core sections (Fig. 4). Correspondingly, the sedimentation rates are up to 26 cm/kyr in the glacial stage 2 and up to 8 cm/kyr in the glacial stage 6, whereas the interglacial values range from 2 to 5 cm/ kyr. A continuous increase in the sedimentation rates within the stage 6 from approximately 3 cm/kyr at the beginning to approximately 8 cm/kyr at the transition to isotope stage 5 results from the 230Thex constant flux modeling.
Discussion and conclusions Lateral sediment redistribution and vertical particle flux In Fig. 5 the sediment accumulation rates (calculated as 230 Thex constant flux derived sedimentation rate multiplied by dry bulk density) and the vertical particle rain rates (Eq. 7) are plotted for each sample vs age for the
five cores. As described in the introduction, a ratio between the sedimentation accumulation rates and the vertical particle rain rates `1 indicates focusing and ratios ~1 are an indicator for winnowing. The sediments of core PS1772-8 were not significantly affected by sediment redistribution. The vertical rain rates were only slightly higher than the sediment accumulation rates, suggesting a weak winnowing, with the exception of isotope stage 2 where a weak focusing is observed. The contrast between the winnowing of a factor of 4 during stage 3 and the focusing of approximately 100% during stage 2 may represent an artifact caused by a small mistake in the position of the transition from stage 3 to stage 2, which is possible due to the relatively low sedimentation rates in these sections. However, such a mistake does not affect the calculation of the vertical rain rate. The rain rates have been relatively constant throughout the core at 1 g/cm 2 7 kyr and increase to values of up to 8 g/cm 2 7 kyr during stage 5e, which represents the maximum value throughout the transect. Further north, in core PS1768-8, the Holocene and last glacial sediments received approximately 4.5 to 6 times more particles by focusing than by vertical supply, which was also observed in the AZ by Francois et al. (1993) and Kumar (1994). Possible advective contributions to the flux of 230Thex may increase the errors of the calculations for this core by approximately 50% (Francois et al. 1990), but do not influence the significance of focusing, because the measured flux of 230Thex exceeds its production in the water column by up to 600%. Focusing intensity decreases downcore and during stage 5e it amounted to a factor of approximately 1.6, which is the reason for the relatively high discrepancy between the average sedimentation rates deduced from previous results and the 230Thex dating (Table 2). The rain-rate record yields values between 1 and 2 g/cm 2 7 kyr. Maxima reach equally high rates for the core sections representing stage 5e and the Glacial/Holocene transition (14.7–10 kyr B.P.) of about 4 g/cm 2 7 kyr. In the stage 3 and 5e sections of core PS1756-5 only approximately 40% of the vertical rain was deposited in the sediments, which means that an export of particles, probably due to increased bottom-current velocities, occurred. Rain rates reach maximum values of 3.7 g/ cm 2 7 kyr in stage 2 and of 4.5 g/cm 2 7 kyr at the end of stage 3. During glacial stage 4 and interglacials 1 and 5 the rates range between 1 and 2 g/cm 2 7 kyr, whereas during stage 5e again 3.5 g/cm 2 7 kyr were reached. The location of core PS1754 was a region where winnowing occurred throughout the core sections representing stages 3–6. During stage 3 only 15%, and during stage 4 only 50%, of the vertical particle rain was deposited, whereas during stages 1 and 2 no particle redeposition occurred within the errors. Variations in the rain rates were small (between 1 g/cm 2 7 kyr during the interglacials 1 and 5e and 2.5 g/cm 2 7 kyr during stages 2 and 3).
563 Fig. 5 Sediment accumulation rates (dashed line) and vertical particle rain rates (solid line; in g/cm 2 7 kyr) vs age (in kyr). The lightly shaded sections mark the interglacial isotope stages and the darkly shaded section marks the stage 5e climate optimum following Martinson et al. (1987). The error bars represent the statistical errors of one standard deviation from the mean of the 230 Thex measurements. Note the different scales of the profiles
The sediments of PS2082-1 were affected by focusing throughout the core, with strongest lateral supply of approximately the 6.5-fold vertical rain rate during glacial stage 2 and up to the threefold vertical rain rate during glacial stage 6. Vertical rain rates were low compared with the sediments in the south with minima around 0.5 g/cm 2 7 kyr during stages 1 and 5 and maxima of around 1 g/cm 2 7 kyr during stage 2 and 6. Reconstruction of particle sedimentation conditions In Fig. 6 a comparison of the sediment accumulation and vertical rain in six time slices along the transect of this study is presented to illustrate the importance of sediment redistribution for reconstructions of particle fluxes in the past. Without correction for sediment redistribution, the pattern of the sediment accumulation suggests that the area of high particle fluxes remained at a position south of the APF during the past 140 kyr (Fig. 6a). During stages 2 and 6 sediment accumulation rates were comparably high north of the SAF and south of the APF. Peak accumulation rates of up to 17 g/cm 2 7 kyr imply that during the transition from the last glacial to the Holocene, the northern AZ was the area of highest par-
ticle fluxes, whereas the PFZ accumulation rates did not reach the values north of the SAF and south of the APF during the past 140 kyr. After correction for sediment redistribution, the particle flux reconstruction yields a significantly different pattern (Fig. 6b). The high particle flux area, which was located south of the APF during the Holocene and stage 5e, shifted to the north by between 2 and 57 during the glacial stages 2–4. This is in accordance with results of Charles et al. (1991). The main controlling factor for the mainly biogenic particle flux was the extent of the sea ice, which expanded further north during the glacials. The amount of vertical particle rain was comparable or somewhat lower during glacial stage 2 in the PFZ than during the last glacial/Holocene transition in the AZ. However, vertical particle fluxes north of the SAF by far did not reach values as high as in the AZ or in the PFZ. During the interglacial Substage 5e the high vertical particle flux area was enlarged into the PFZ and reached its maximum rain rate values of the transect at about 6 g/cm 2 7 kyr in the southern AZ, which probably was caused by sea ice-free conditions near the ice edge (Gersonde et al., in preparation). With the knowledge of the focusing intensity, the paleoproductivity reconstructions of Mackensen at al. (1994) for core PS2082-1 have to be reinterpreted. The
564
basic pattern of higher glacial productivity at this location would remain the same after the focusing correction. The amount of productivity, however, is lower than the values given by Mackensen et al. by approximately a factor of 5 during glacial stage 2 and by approximately a factor of 3 during glacial stage 6 after focusing correction (see also Nürnberg et al., in preparation). Both the sediment accumulation rates and the vertical particle rain rates have to be taken into account for the calculation of paleoenvironmental proxy accumulation rates such as biogenic barium, opal, or Corg, and corresponding paleoproductivity reconstructions. The vertical rain rates yield a reconstruction of the supply from the surface waters and the sediment accumulation rates have to be considered with respect to preservation rates of biogenic Ba or opal. Detailed paleoceanographic and paleoenvironmental interpretation of proxy fluxes calculated on the basis of these results will be published elsewhere (Gersonde et al., in preparation; Nürnberg et al., in preparation).
Summary
Fig. 6. a Time slices of the sediment accumulation rates and b rain rates from 567S to 437S. The solid-line time slices represent average values for the Holocene climate optimum (12–9 kyr B.P.), the stages 3 and 5e, whereas the dashed-line time slices represent the LGM (18 kyr B.P.), glacial stages 4 and 6. At core PS1768-8 the Holocene climate optimum value represents the average for the period between 14 and 10 kyr B.P. The question marks at the stage 6 accumulation rate time slice mark the uncertainty of these values, which is caused by the incomplete stage 6 core sections. The vertical lines at 507S and 45.57S mark the present-day positions of the Antarctic polar front (APF) and the Subantarctic front (SAF), respectively. Note the different scales of the plots representing the Holocene and LGM sediment accumulation rates and rain rates
Applying the 230Thex-dating method, average sedimentation rates for the five cores of the transect were estimated. Other stratigraphic investigations could be corroborated, which implies that this dating method may be applied to gain estimates on sedimentation rates and ages of sediment cores where no other stratigraphic information exists. After the core depths of the main climatic transitions in the cores were fixed applying various stratigraphic methods, 230Thex constant flux models were run to achieve a high-resolution record of the sedimentation rates within the isotope stages. It was shown that in the southern core PS1772-8 the diatomaceous ooze section was deposited during only approximately 11 kyr at the beginning of isotope stage 5e. In the northern AZ a drastic increase in the sediment accumulation rates, as well as the vertical particle fluxes, is observed after the retreat of the sea ice in the Holocene climate optimum. In the PFZ and the Subantarctic zone rain rates were increased mainly during the glacials and a continuous increase in the rain rates during isotope stage 6 is determined in the core from the Subantarctic zone. The comparison of the measured fluxes of 230Thex with the values expected from production shows that lateral sediment redistribution locally affected the sedimentation severely. Sediment focusing contributed up to six times more sediment than accumulated vertically in the section representing the last glacial/Holocene transition in core PS1768-8 and in the glacial stage 2 section of core PS2082-1. Winnowing caused by bottom currents during stage 3 only left approximately 15% of the vertical sediment rain in the PFZ. The vertical rain rates calculated via 230Thex represent reliable reconstructions of sediment supply rates,
565
on the basis of which fluxes of proxy tracers may be calculated, which then may be compared between different locations. The observed sediment accumulation rates have to be applied when estimating effects of burial on the degree of preservation of proxies such as biogenic Ba, opal, or Corg. The reconstruction of the vertical sediment rain rates of the transect suggests a glacial northward shift in the high vertical particle flux area by 2–57and no overall increase, but a slight decrease in the overall glacial particle flux. Acknowledgements This study was supported by “Deutsche Forschungsgemeinschaft” grant no. Ma821/9–1. We thank Wolfgang Berger and an anonymous reviewer for their constructive comments, which significantly improved the manuscript. The data are available on request or through the SEPAN database of the Alfred Wegener Institute for Polar- and Marine Research. This is AWI contribution no. 1031 and SFB 261 contribution no. 121.
References Abelmann A, Gersonde R (1988) Cycladophora davisiana stratigraphy in Plio/Pleistocene sediment cores from the Antarctic Ocean (Atlantic sector). Micropaleontology 34 : 268–276 Abelmann A, Gersonde R (1991) Biosiliceous particle flux in the Southern Ocean. Mar Chem 35 : 503–536 Anderson RF, Bacon MP, Brewer PG (1983) Removal of 230Th and 231Pa from the open ocean. Earth Planet Sci Lett 62 : 7– 23 Bacon MP (1984) Glacial to interglacial changes in carbonate and clay sedimentation in the Atlantic estimated from thorium-230 measurements. Isot Geosci 2 : 97–111 Bard E (1988) Correction of accelerator mass spectrometry 14C ages measured in planktonic foraminifera: paleoceanographic implications. Paleoceanography 3 : 635–645 Bard E, Hamlin B, Fairbanks RG, Zindler A (1990) Calibration of the 14C timescale over the past 30 000 years using mass spectrometric U–Th ages from Barbados corals. Nature 345 : 405–410 Bathmann U, Schulz-Baldes M, Fahrbach E, Smetacek V, Hubberten H-W (1992) Die Expeditionen ANTARKTIS-IX/1–4 des Forschungsschiffes “Polarstern” 1990/91. Ber Polarforsch 100 : 1–403 Burckle LH, Cooke DW (1983) Late Pleistocene Eucampia antarctica abundance stratigraphy in the Atlantic sector of the Southern Ocean. Micropaleontology 29 : 6–10 Charles CD, Froelich PN, Zibello A, Mortlock RA, Morley JJ (1991) Biogenic opal in Southern Ocean sediments over the last 450 000 years: implications for surface water chemistry and circulation. Paleoceanography 6 : 697–728 Fischer G, Bohrmann G, Gersonde R (in preparation) The Late Quaternary d 13C record in the Antarctic circumpolar current (eastern Atlantic sector): paleoceanographic implications Francois R, Bacon MP, Suman D (1990) Thorium 230 profiling in deep-sea sediments: high-resolution records of flux and dissolution of carbonate in the equatorial Atlantic during the last 24 000 years. Paleoceanography 5 : 761–787 Francois R, Bacon MP, Altabet MA, Labeyrie LD (1993) Glacial/ interglacial changes in sediment rain rate in the SW Indian sector of Subantarctic waters as recorded by 230Th, 231Pa, U, and d 15N. Paleoceanography 8 : 611–629 Frank M, Eckhardt J-D, Eisenhauer A, Kubik PW, Dittrich-Hannen B, Mangini A (1994) Beryllium 10, thorium 230 and protactinium 231 in Galapagos microplate sediments: implications for hydrothermal activity and paleoproductivity changes during the last 100 000 years. Paleoceanography 9 : 559–578
Frank M, Eisenhauer A, Bonn WJ, Walter P, Grobe H, Kubik PW, Dittrich-Hannen B, Mangini A (1995) Sediment redistribution versus paleoproductivity change: Weddell Sea Margin sediment stratigraphy for the last 250 000 years deduced from 230 Thex, 10Be and biogenic barium profiles. Earth Planet Sci Lett 136 : 559–573 Gersonde R (1993) Die Expedition ANTARKTIS X/5 mit FS “Polarstern” 1992. Ber Polarforsch 131 : 1–167 Gersonde R, Hempel G (1990) Die Expeditionen ANTARKTIS VIII/3 und VIII/4 mit FS “Polarstern” 1989. Ber Polarforsch 74 : 1–173 Gersonde R, Zielinski U (in preparation) Significance of diatoms as indicators of past Antarctic sea ice extent Gersonde R, Abelmann A, Bohrmann G, Frank M, Heinemeier J, Rutgers van der Loeff MM, Niebler H-S, Mangini A, Rud N, Zielinski U (in preparation) Southern Ocean paleoenvironmental changes during the last 20 000 years (Atlantic sector) Hays JD (1967) Quaternary sediments of the Antarctic Ocean. Prog Oceanogr 4 : 117–131 Hays JD, Imbrie J, Shackleton NJ (1976) Variations in the earth’s orbit: pacemaker of the ice ages. Science 194 : 1121–1132 Imbrie J, Hays JD, Martinson DG, McIntyre A, Mix AC, Morley JJ, Pisias NG, Prell WL, Shackleton NJ (1984) The orbital theory of the Pleistocene climate: support from a revised chronology of the marine d 18O record. In: Berger AL et al. (eds) Milankovitch and climate, part 1. Riedel, Hingham, Massachusetts, pp 269–305 Kumar N (1994) Trace metals and natural radionuclides as tracers of ocean productivity. Ph.D. thesis, Columbia University, New York, pp 1–317 Mackensen A, Grobe H, Hubberten H-W, Kuhn G (1994) Benthic foraminiferal assemblages and the d 13C signal in the Atlantic sector of the Southern Ocean: glacial-to-interglacial contrasts. In: Zahn R et al. (eds) Carbon cycling in the glacial ocean: constraints on the ocean’s role in global change. NATO ASI series, vol I 17, pp 105–144 Mangini A, Diester-Haass L (1983) Excess 230Th in N.W. African sediments traces upwelling in the past. In: Suess E, Thiede J (eds) Coastal upwelling: its sediment record. NATO Conf Ser, Ser IV 10a, pp 455–470 Martinson DG, Pisias NG, Hays JD, Imbrie J, Moore TC Jr, Shackleton NJ (1987) Age dating and the orbital theory of the ice ages: development of a high-resolution 0 to 300 000 year chronostratigraphy. Quaternary Res 27 : 1–29 McCave N (1983) Particulate size spectra, behavior and origin of nepheloid layers over the Nova Scotian continental rise. J Geophys Res 88 : 7647–7666 Niebler HS (1995) Rekonstruktion von Paläo-Umweltparametern anhand von stabilen Isotopen und Faunenvergesellschaftungen planktischer Foraminiferen im Südatlantik. Ber Polarforsch 167 : 198 Nürnberg CC, Bohrmann G, Frank M, Schlüter M, Suess E (in preparation) Barium accumulation in the Atlantic sector of the Southern Ocean: evidence for productivity changes during the last 190 000 years Peterson RG, Stramma L (1991) Upper-level circulation in the South Atlantic Ocean. Prog Oceanogr 26 : 1–73 Pudsey C (1992) Late Quaternary changes in Antarctic bottom water velocity inferred from sediment grain size in the northern Weddell Sea. Mar Geol 107 : 9–33 Rutgers van der Loeff MM, Berger GW (1991) Scavenging and particle flux: seasonal and regional variations in the Southern Ocean (Atlantic sector). Mar Chem 35 : 553–567 Rutgers van der Loeff MM, Berger GW (1993) Scavenging of 230 Th and 231Pa near the Antarctic polar front in the South Atlantic. Deep-Sea Res I 40 : 339–357 Scholten JC, Botz R, Paetsch H, Stoffers P (1994) 230Thex flux into Norwegian–Greenland Sea sediments: evidence for lateral sediment transport during the past 300 000 years. Earth Planet Sci Lett 121 : 111–124
566 Sea Ice Climatic Atlas, vol 1, Antarctica (1985) Naval Oceanography Command Detachment, Ashville, North Carolina, pp 1– 131 Suman DO, Bacon MP (1989) Variations in Holocene sedimentation in the North American basin determined from 230Th measurements. Deep-Sea Res 36 : 869–878 Van Bennekom AJ, Berger GW, Van der Gaast SJ, DeVries RTP (1988) Primary productivity and the silica cycle in the southern ocean (Atlantic sector).Paleogeogr Paleoclimatol Paleoecol 67 : 19–30 Vogelsang E (1990) Paläozeanographie des Europäischen Nordmeeres anhand von stabilen C. und O-Isotopen. Ber Sonderforschungsber 313 (23): 136
Yang YL, Elderfield H, Ivanovich M (1990) Glacial to Holocene changes in the carbonate and clay sedimentation in the equatorial Pacific Ocean estimated from 230Th profiles. Paleoceanography 5 : 789–809 Zielinski U (1993) Quantitative Bestimmung von Paläoumweltparametern des Antarktischen Oberflächenwassers im Spätquartär anhand von Transferfunktionen mit Diatomeen. Ber Polarforsch 126 : 1–148
