Prell, W. L., Niitsuma, N., et al , 1991Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 117

17. LATE PLEISTOCENE EVOLUTION OF SURFACE AND MID-DEPTH HYDROGRAPHY ATTHE OMAN MARGIN: PLANKTONIC AND BENTHIC ISOTOPE RECORDS AT SITE 7241

Rainer Zahn2 3 and Thomas F. Pedersen2

ABSTRACT

Stable isotope records of coexisting benthic foraminifers Uvigerina spp. and Cibicidoides spp. and planktonic G. ru-ber (white variety) from Site 724 are used to study the late Pleistocene evolution of surface and intermediate water hy-drography (593 m water depth) at the Oman Margin. Glacial-interglacial δ 1 8 θ amplitudes recorded by the benthic fora-minifers are reduced when compared to the estimated mean ocean changes of θ 1 8 θ s e a w a t e r . Epibenthic δ 1 3C remains at itsmodern level or is increased during glacial times. This implies that Red Sea outflow waters which are enriched inδ18θseawater ond δ 1 3 C E C O have been replaced during glacial periods by intermediate waters still positive in δ13Cjxx>2 t>u t

more negative in δ 1 8 θ s e a w a t e r . Glacial-interglacial amplitudes of the planktonic δ 1 8 θ record exceed those of the meanocean δ 1 8 θ s e a w a t e r variation and imply decreased surface water temperatures (SST) during glacial times. Throughoutmost of the records these cooling events correlate with enhanced rates of carbon accumulation. However, both negative(colder) SST and positive C o r g accumulation rate anomalies do not correlate with potential physical upwelling maximaas inferred from the orbital monsoon index. This is in conflict with the established hypothesis that upwelling in thewestern Arabia Sea should be strongest during maxima of the southwest monsoon.

INTRODUCTION

The chemical structure of the world ocean has changed dur-ing the Pleistocene as a function of global climate. Benthic fora-miniferal δ13C and Cd/Ca records from pelagic core sites, forinstance, point to lower ventilation rates and higher nutrientcontents in the deep ocean during glacial episodes (Curry et al.,1985; Duplessy et al., 1988; Boyle, 1988a). In contrast, benthicδ13C and Cd/Ca data from shallow core sites (< 1.5 km waterdepth) in the North Atlantic and North Pacific Oceans implythat mid-depth nutrient levels were lower during glacial times(Zahn et al., 1987; Curry et al., 1988; Duplessy et al., 1988;Boyle 1988b). This apparent contrast between mid-depth anddeep ocean chemical cycling during glacial-interglacial timessuggests that vertical fractionation processes in the ocean mayplay an important role in the global climate system (Boyle,1988b). ODP Site 724 from a water depth of 593 m is well suitedto study the long-term development of mid-depth hydrographicvariations in the northern Indian Ocean. Also, its position un-der the Oman coastal upwelling regime makes it equally suitablefor studying the evolution of monsoonal upwelling in the west-ern Arabian Sea. In this paper we use records of benthic andplanktonic foraminiferal stable isotopes and of organic carbonto decipher the evolution of coastal upwelling and intermediatedepth hydrography at the Oman Margin.

REGIONAL OCEANOGRAPHY

Surface water circulation in the Arabian Sea is modulated bythe seasonal variation of the monsoon wind system. In responseto the southwest monsoon, strong upwelling occurs off the coastsof Oman and Somalia during the summer months (May throughOctober) due to the wind-induced offshore Ekman transport.During this season the upwelling of cool waters leads to an en-hanced heat loss in the Arabian Sea which results in a zonal

Prell, W. L., Niitsuma, N., et al., 1991. Proc. ODP, Sci. Results, 117: Col-lege Station, TX (Ocean Drilling Program).

2 The University of British Columbia, Department of Oceanography, 6270University Boulevard, Vancouver, B.C. V6T 1W5, Canada.

3 (Current address) Geomar, Wischhofstra/3e 1-3, Gebàude 4, D-230 Kiel-14,Federal Republic of Germany.

temperature gradient of >4°C between the upwelling center andthe open ocean (Wyrtki, 1971; Düing and Leetmaa, 1980). Atthe same time the concentrations of all major nutrients (silicate,phosphate, nitrate) reach a maximum (Wyrtki, 1971) due to aninjection of nutrients from below the thermocline into the sur-face waters. Carbon productivity reaches some 1.5 g m~2 day"1

(Kabanova, 1968; Cushing, 1973; Krey, 1973) which, for exam-ple, is twice the productivity observed at the upwelling centeroff Northwest Africa (0.7 g m~2 day"1; Schemainda et al.,1975). Advection of nutrient-rich upwelled water offshore fromthe Somali-Arabian upwelling center supports high productionin surface waters more than 400 km from the continental mar-gin (Nair et al., 1989). During the winter months (Novemberthrough April) the prevailing northeast monsoon winds result inan onshore Ekman transport of surface waters which leads to acollapse of all upwelling activity along the continental marginof the Western Indian Ocean.

The high rate of input of organic matter and its subsequentoxidation below the thermocline is associated with a very highoxygen demand at intermediate depths. The resulting oxygendeficiency is fostered by the sluggish intermediate water circula-tion which partly results from the cul-de-sac-type geometry ofthe northern Arabian Sea (Wyrtki, 1973). In addition, the strongtropical-subtropical thermocline prevents downward mixing ofoxygenated surface waters thus enhancing the effect of the lowrenewal rate of intermediate waters (Swallow et al., 1983). As aresult of these factors, the oxygen minimum in the NorthernArabian Sea is one of the most severe encountered in the worldocean. Waters between 200 and 1500 m water depth are oxygen-depleted to an extent that they contain on occasion free H2S andare often relatively deficient in NOj (Qasim, 1982).

The intermediate-depth hydrography at the Oman continen-tal margin is influenced by outflow waters from the Red Sea(Siedler, 1968; Düing and Schwill, 1968; Premchand et al.,1986a). Evaporation in the Red Sea by far exceeds fresh watersupply and the outflow waters are considerably warmer andmore saline than the adjacent ocean. Red Sea waters can there-fore be easily recognized in the temperature-salinity (T-S) fieldas a mid-depth T-S maximum (Fig. 1). The volume transport be-tween the Red Sea and the Arabian Sea is closely tied to themonsoonal cycle and may vary by a factor of five between the

291

R. ZAHN, T. F. PEDERSEN

30

oo

— 20CD

CDQ .

ECD

10

Surface water-

>fled Sea water

(600 m)

Deep water34 35 36

Salinity37

Figure 1. Temperature-salinity (T-S) field at Geosecs Station 413 in thewestern Arabian Sea. The core layer of the Red Sea outflow forms awell-defined T-S maximum at 600 m.

summer and winter seasons (Souvermezoglou et al., 1989). Pre-vailing northeasterly winds strengthen the outflow during win-ter because of a positive mass balance which results from an en-hanced wind-driven advection of surface waters from the Gulfof Aden into the Red Sea (Bethoux, 1987). In contrast, enhancedwind-driven outflow of surface waters during the southwest mon-soon in summer decreases the deep outflow and forces a subsur-face inflow to balance the mass deficiency (Bethoux, 1987). RedSea outflow waters also considerably increase the nutrient levelsat intermediate depths in the Arabian Sea through an estimatedcarbon export of 26 × 1012 mol yr"1 (Bethoux, 1988). PersianGulf waters on the other hand spread in the Arabian Sea atdepths shallower than 300 m (Düing and Schwill, 1968; Prem-chand et al., 1986b) and thus do not contribute to the interme-diate depth waters at Site 724.

METHODS

Stable IsotopesIsotope analyses were run on a VG Isogas Prism mass spec-

trometer using an on-line VG Isogas Autocarb automated car-bonate preparation device. Release and preparation of the car-bon dioxide gas for isotope analysis follows the standard tech-niques described in Shackleton and Opdyke (1973) with theexception that reactions occur at 90°C in a common acid bath.That is, up to 40 sample reactions occur in the same reactionvessel containing 5 cm3 of 100% orthophosphoric acid. Poten-tial memory effects are below 1 % of the isotope offset betweensamples as revealed by sequential analysis of carbonates with anisotope offset of about 20‰ in δ 1 8θ. Calibration of the massspectrometer to the international PeeDee Belemnite (PDB) car-bonate standard was done primarily using National Bureau ofStandards (NBS) carbonate standards 19 and 20. External preci-sion of the system for δ 1 8θ and δ13C was better than 0.06‰ andO.04‰, respectively (1 a standard deviation, n = 75; deter-mined from a replicate laboratory standard run with the sam-ples over a 3-month period). As a further indication of preci-sion, 6 replicate analyses each of 20 specimens of Neoglobo-quadrina pachyderma from the same sample taken from a highsedimentation rate core (8 cm k.y. ~ l) from the Northeast Pacificyielded a 1 σ standard deviation of 0.06‰ in both δ 1 8θ andδ13C.

Stable isotope measurements were run for the upper 32 m ofCore 724C, which was sampled continuously at 20 cm intervals.Records were obtained from the planktonic foraminifer Globi-gerinoides ruber (white variety) and from the two benthic gen-era Cibicidoides spp. and Uvigerina spp. Identification of thebenthic foraminifers on the species level follows the nomencla-ture given in Resig (1981) and Lutze (1£86). Accordingly, mea-surements were run on Uvigerina auberiana d'Orbigny, Uviger-ina senticosa Cushman, Cibicidoides wuellerstorfi (Schwager),and Cibicides mundulus (Brady, Parker and Jones). The isotopesamples contained on average 25 planktonic and 5-7 benthicspecimens which were picked from the size fractions 300-355µm and 250-355 µm, respectively. All samples were mechani-cally crushed, ultrasonically rinsed in methanol, and roasted invacuo for 30 min at 430°C. This cleaning procedure was appliedso as to release occluded fine-grained carbonate particles and toremove impurities and organic coatings from the tests whichcould alter the isotope values. All isotope data are listed in Ap-pendix A.

Organic Carbon

Organic carbon (Corg) was determined as the difference be-tween total carbon measurements made by combustion and ther-mal conductivity detection of CO2 using a Carlo-Erba NA 1500CNS analyzer (1 a relative standard deviation is better than 2%),and carbonate carbon measurements made by acidification withhot 20% HC1 and coulometric quantification of the evolvedCO2 using a Coulometrics analyzer. All carbonate carbon mea-surements were made in duplicate and the mean used to increasethe precision of the Co r g (by difference) data. The 1 σ relativestandard deviation of the carbonate carbon analyses was 1%.The Corg data presented in this paper are considered precise towithin ± 3% (1 a). All Co r g data are listed in Appendix B.

RESULTS

Oxygen Isotope Stratigraphy

The isotope record obtained from Cibicidoides spp. is in-complete and shows relatively large scatter (Fig. 2) reflecting thelow abundance of this genus in the sediment samples. Whereasthe low number of specimens (1-3) available for some measure-ments may seriously lower the reliability of the Cibicidoidesspp. isotope record, this problem is of minor importance for theUvigerina spp. record (Fig. 2) which was measured on samplesgenerally containing 5-10 specimens. Since we will use the inter-specific δ13C difference between Uvigerina spp. and Cibicidoi-des spp. later as a proxy for the potential gradient of pore-waterδ13CECo2> ^ ls especially important to check for the reliability ofthe Cibicidoides spp. isotope data. We have done this by usingthe statistically better constrained Uvigerina spp. record as areference. All Cibicidoides spp. data which deviate by more than0.2‰ from the generally accepted δ 1 8θ difference of 0.64‰ be-tween both genera (Shackleton and Opdyke, 1973) have beendiscarded from the record (Fig. 3). Accordingly, 63 out of a totalof 113 Cibicidoides spp. data have been rejected. The glacialcore sections were most severely affected by this rigorous cut-offprocess. However, in view of the importance of the data for ourlater discussion, this "clean-out" appears thoroughly appropri-ate.

The δ1 8θ records obtained from G. ruber and from Uviger-ina spp. (Fig. 2) display distinct δ 1 8θ maxima at sub-bottomdepths of 2 m, 7 m, 12 m, between 23 m and 25 m, and at 30 m.We interpret these maxima as documenting glacial oxygen iso-tope stages 2, 4, 6, 8, and 10. In addition, we used the δ 1 8θ min-imum of G. ruber at 9.4 m as the stratigraphic fix point for thelast interglacial substage 5e. No control points were used for the

292

LATE PLEISTOCENE PLANKTONIC AND BENTHIC ISOTOPE RECORDS: SITE 724

COQQ .

0

áo

Core sections X18,-, _ s16 r

5 10 15 20 25 30 35

1 -

2 -

3 -

II

i !\ i

• I

IO|

1

rJ

1

o° 1

9 ® '

i l \ C61 o1 °11

—1 , 1 1 •

Cibicidoides spp.1 Jf

o 1° - ° Io 1 % •

1o

, , 1 . , •

10 15 20 25 30 35

3.0

Q. 2.5CO

CO

"O 2.0

:s 1.5

o1.0

0.5

0 6 4 ‰

cut-off lines (± 0.2‰)

δ O (‰PDB)

o " good " data

+ " bad" data

1.5 2.0 2.5 3.0

Uvigerina spp.

Figure 3. Uvigerina spp. vs. Cibicidoides spp. comparison of δ 1 8 θ. AllCibicidoides spp. data which deviate by more than ± 0.2‰ from the ac-cepted δ 1 8 θ difference of 0.64‰ between both genera were discardedfrom the isotope record.

2 -

1 1 Uvigerina spp.

0 5 10 15 20 25

Depth (mbsf)

30 35

Figure 2. Oxygen isotope records of planktonic foraminifer G. ruber(white variety) and of benthic foraminifers Uvigerina spp. and Cibici-doides spp. Open symbols are rejected data (see Fig. 3). The data gapsin the Cibicidoides spp. record are caused by the discontinuous occur-rence of this genus (see also Fig. 5). Full dots in the G. ruber isotope re-cord mark the position of the stratigraphic fix points which were usedfor developing the time model for Site 724. The age control points arelisted in Table 1. All isotope data are listed in Appendix A.

broad δ 1 8θ minimum between 16 and 23 m. We believe that thisminimum correlates with interglacial oxygen isotope stage 7,but the structure of the benthic and planktonic δ1 8θ profiles inthis core section is inconclusive and does not allow for more de-tailed stratigraphic correlation to the Specmap δ 1 8θ stack (Im-brie et al., 1984) which was used as a reference. Magnetic sus-ceptibility data at Site 724 show a good correlation among holesA, B, and C implying that there are no major gaps or overlapsbetween sub-cores 1 through 4 (Peter deMenocal, written comm.,1988), and that the sediment sequence, and thus also the isotoperecord, of Hole 724C is complete.

The stratigraphic fix points which were used to develop theage model for Hole 724C have been picked by employing thecorresponding age control points from the Specmap global stackof δ1 8θ (Imbrie et al., 1984) (see Table 1). Accordingly, the up-per 32 m of Hole 724C spans the past 350,000 yr (Fig. 4) whichimplies mean sedimentation rates of about 9 cm k.y.~1. Linearinterpolation between the depth position of the oxygen isotopestage boundaries reveals downcore variations of sedimentationrates between maximum values of 15 cm k.y.~1 (oxygen isotope

Table 1. Age control points usedto develop the time model forHole 724C.

Sub-bottomdepth (m)

0.221.102.906.427.629.029.42

11.6214.2215.8222.9224.3227.3228.3229.52

Age(k.y. BP)

6.012.024.059.071.0

110.0122.0135.0171.0183.0249.0287.0320.0331.0341.0

Specmapeventa

1.12.03.04.05.05.45.56.26.56.68.28.59.29.3

10.2

a From Imbrie et al. (1984).

stages 1 and 2) and minimum values of 3.5 cm k.y.~1 (oxygenisotope stage 5) (Fig. 5). Apparently, there is no systematic cor-relation between variation of sedimentation rates and the stateof glacial-interglacial climates, which hints that additional pro-cesses such as sediment focusing and winnowing may be super-imposed on the climatically-controlled variations of eolian sedi-ment input and changes of biological productivity. Time resolu-tion of the isotope records at a sample interval of 20 cm variesbetween 3 k.y. and 11 k.y. as a function of sedimentation rate(calculated as Nyquist period: 2× [sample interval/sedimenta-tion rate]; Balsam and Posmenier, 1981).

Comparison with standard isotope profiles such as the ben-thic δ1 8θ record of equatorial Pacific core V19-30 (Shackletonand Pisias, 1985) shows that the glacial-interglacial δ 1 8θ ampli-tudes at Site 724 are often reduced (Fig. 4). This effect is mostprominent in the benthic δ 1 8θ record. In addition, interglacialstage 7 is not well resolved in our records. These irregularitieshave been confirmed by more than 40 re-runs of foraminiferalsamples from the respective core sections and we can thereforeexclude analytical errors as a cause for the deviations of our iso-

293

R. ZAHN, T. F. PEDERSEN

Oxygen isotope stages ( l Sl = glacials )

oCO

>

400

C\JN

CD

CO

Benthic foraminifers

G. ruber

400

Benthic foraminifers

100 200 300

Age [1000yr B.P.]

400

Figure 4. The benthic and planktonic δ 1 8 θ records in the time domain. The benthic record includes Uvi-gβrina spp. and Cibicidoides spp. data. The Cibicidoides spp. δ 1 8 θ data were converted to the Uvigerinascale of δ 1 8 θ by adding 0.64‰. Comparison with the pelagic benthic δ 1 8 θ record of equatorial Pacificcore V19-30 (3°S, 83°W, 3091 m; Shackleton and Pisias, 1985) shows that the glacial-interglacial δ 1 8 θamplitudes at Site 724 are considerably lower than those observed in the deep ocean. This effect is mostprominent in the benthic δ 1 8 θ record and points to systematic δ 1 8 θ s e a w a t e r changes of the intermediatedepth waters at Site 724.

tope records from standard deep-sea isotope profiles. As notedabove, magnetic susceptibility data do not indicate major gapsor overlaps between the core sections, which excludes this sourceof potential stratigraphical disturbance. The magnitude of f ra-miniferal δ 1 8θ variation through time is determined by varia-tions of global ice volumes and by local variations of ö^O^^g,.and water temperature. In view of the high sedimentation rates(see above), an influence of vertical mixing by burrowing orga-nisms of foraminiferal shells across stratigraphic boundaries,which has the potential to considerably reduce the δ 1 8θ varianceof the downcore records, may be of minor importance at Site724. The quasi-coherent variation of the benthic and planktonicδ 1 8θ curves in Hole 724C (Fig. 4) implies that both records con-tain the global ice volume effect. That is, the irregular structuresof the δ 1 8θ profiles are likely the result of variations of the am-bient water hydrography at Site 724 or of downslope migrationof foraminiferal shells winnowed by local undercurrents andtidal flow regimes from shallower upper slope or shelf sediments.

Changes in the absolute abundance of foraminiferal speciesused for isotope measurements may significantly enhance arti-

facts introduced by bioturbation and lead to vertical transla-tions of the isotope signals (Berger and Heath, 1968; see alsodiscussion by Climap Project Members, 1984). To evaluate thepotential influence that such effects may have had on the iso-tope records at Site 724 we have checked for the abundance dis-tribution of Uvigerina spp. and Cibicidoides spp. during thelast two climatic cycles, i.e., back to oxygen isotope stage 8.Specimen abundances of both genera were determined for thesame size fraction from which the benthic isotope samples werepicked (i.e., 250-355 µm). As Figure 6 shows, Uvigerina spp.are by far more abundant in the sediment samples than Cibici-doides spp. Massive incursions of Uvigerina spp. specimens areobserved during mid-stage 7, late stage 6, and during early stage3 (Fig. 6). Cibicidoides spp. maxima occur during stage 7 andearly stage 6, in late stage 4, and mid-stage 3 (Fig. 6). The mostdominant abundance maximum occurs synchronously for bothgenera during mid-stage 3. However, abundances of both generado not significantly change across the stratigraphic boundarybetween full glacial isotope stage 6 and full interglacial isotopesubstage 5e (Fig. 6). The Uvigerina spp. maximum at 145 k.y.

294

LATE PLEISTOCENE PLANKTONIC AND BENTHIC ISOTOPE RECORDS: SITE 724

Sedimentation rates

Oxygen isotope stages ([ | = glacials)

100 150 200 250 300 350 400

Age [1000yr B.P. ]

Figure 5. Sedimentation rates at Site 724 calculated by linear interpola-tion between the depth positions of oxygen isotope stage boundaries asimplied by the time model. Time resolution of the isotope records wascalculated as a function of sampling interval (20 cm) and sedimentationrate, and varies between 3 k.y. (isotope stages 1 and 2) and 11 k.y. (iso-tope stage 5).

Oxygen isotope stages ( R i l l = glacials )

2500

300

50 100 150 200 250 300

Age [1OOOyr B.P. ]

Figure 6. Abundance of Cibicidoides spp. and Uvigerina spp. at Site724 during the past 270,000 yr. The abundance record is used to evaluatethe potential effect on the glacial-interglacial benthic δ 1 8 θ amplitudesof bioturbational admixture of foraminiferal shells across the strati-graphic boundaries. Apparently no major change in abundance occursat the oxygen isotope stage boundary 5/6. The concurrently reducedδ 1 8 θ amplitude may therefore document a change of the oxygen isotopecomposition of intermediate waters at the Oman Margin.

B.P. (> 4600 specimens per gram of size fraction 250-355 µm) isabout 3 m below the core section where depleted δ 1 8 θ values ofG. ruber indicate the last interglacial 5e. Therefore, reduction inamplitude of the isotope records is not likely to be the result ofbioturbation. Based on the evidence to date we conclude thatthe depletion of the glacial-interglacial δ 1 8 θ amplitudes at Site724 points to systematic variations in temperature and the

δ18<λ of surface and intermediate waters in the WesternArabian Sea. The implications of this are discussed below.

Carbon Isotope Records

Comparison of the benthic and planktonic δ1 3C records withmean ocean variations of δ1 3C is again obtained by using theisotope record from core V19-30 (Shackleton and Pisias, 1985)(Fig. 7). Both the G. ruber and Uvigerina spp. records at Site724 show depleted δ13C values during glacial times, and the car-bon isotope profiles are in this respect coherent with those ofequatorial Pacific core V19-30. Mismatches between the carbonisotope profiles at Site 724 and core VI9-30 occur during latestage 7. Our δ 1 8 θ records from this section correlate rather poorlywith the global δ 1 8 θ stratigraphy (see above), and the intercoreδ1 3C differences shown in Figure 7 could be due to inaccuraciesin our chronostratigraphy. δ1 3C variations of Cibicidoides spp.at Site 724 appear to be decoupled from the variations observedin the δ13C record of coexisting Uvigerina spp. specimens (Fig.7). Because of its presumed epibenthic microhabitat (Corliss,1985; Lutze and Thiel, 1987), Cibicidoides spp. are consideredto record faithfully δ1 3C changes of ambient bottom water Σ C O 2

(Curry et al., 1988, and references therein). δ1 3C fluctuations ofUvigerina spp. on the other hand appear to reflect carbon iso-tope variations of pore water ECO2 because of the presumed en-dobenthic habitat of this genus (Zahn et al., 1986; McCorkle etal., 1989). That is, the apparently incoherent variation of δ1 3Cfor coexistent Cibicidoides spp. and Uvigerina spp. at Site 724may be viewed as documenting largely independent variationsof δ1 3CEco2

i n t n e ambient bottom waters and in the pore water,the latter being closely related to variations of the organic car-bon flux into the sediments (McCorkle et al., 1985) and thus toproductivity changes in the surface waters (Zahn et al., 1986).

Organic Carbon Variations

The organic carbon profile at Site 724 shows values between0.5 and 2.0 wt% (Fig. 8A) with the exception of a sharp maxi-mum in mid-stage 3 where the concentration reaches nearly 4%.This value has been confirmed by repeat measurements. Appar-ently the downcore carbon contents do not vary systematicallyas a function of glacial-interglacial climates in that carbon con-tents are high both during interglacial and glacial oxygen iso-tope stages (Fig. 8A).

In order to interpret properly variations of organic carbonthrough time in terms of paleoproductivity changes we have toconvert the profile of organic carbon content (wt%) into a pro-file of carbon accumulation rate (mg c m " 2 1000 yr~1) whichtakes into account changes in sediment porosity and sedimenta-tion rate. This conversion was done by multiplying the organiccarbon data with the mean sedimentation rates shown in Figure5 and the shipboard density data obtained by the Gamma RayAttenuation Porosity Evaluator (GRAPE) (Fig. 8B). The GRAPEwet-bulk-density analog data were converted into dry-bulk-den-sity data using the empirically determined regression equation

Density(dry) = 1.585 × density(GRAPE) - 1.692(r2 = 0.64, n = 13)

which was established by using 13 measured in-situ dry-bulk-density data from the upper 40 m of Hole 724C as a reference.Equation 1 reproduces the in-situ control data to within

295

R. ZAHN, T. F. PEDERSEN

oCO

>

cüQQ.

0

üTo

Oxygen isotope stages ([|§§§] = glacials)

0 -

100 200 300

Benthic foraminifers

400

CM

CD

CDQCL0

è

CDQD.

P

100 200 300

100 200 300

Age [lOOOyr B.P.]

G. ruber

400

1 -

o -

-1 -

2 -

- T — | • — I — | '

*•? "1 To v

l i •U^

1 | il|1 1 1 11 1 1 1

-j ' r

r °o° i

1 ^ 1 1

1 > '1 1

•J r• 1

oo«b 1

} ! ^' i

ii

iull1

1

•

•

Cibicidoides spp.

Uvigerina spp.

400

Figure 7. Benthic and planktonic δ 1 3C variations at Site 724 during the past 350,000 yr. Open symbols arerejected data (see text). The benthic δ 1 3C record of equatorial Pacific core V19-3O is from Shackleton andPisias (1985) and is used here as a reference for deep-ocean δ13C£cc>2 variations. δ 1 3C variations of G. ru-ber and of Uvigerina spp. correlate well with those observed at core V19-30. In contrast, δ 1 3C variationsof epibenthic Cibicidoides spp. are apparently decoupled from those recorded by coexisting Uvigerinaspp., pointing to the importance of δ13Cj;cθ2 gradients between pore waters and ambient intermediate wa-ters at Site 724.

±0.1 g cm 3 ( l σ standard error). The GRAPE density analogdata decrease sharply at the pipe breaks between sections 2 and3, and 3 and 4 (Fig. 8B) probably due to sediment flow-out. Inboth cases the last undisturbed value from the respective uppersections has been used to calculate the carbon accumulationrates (see also Appendix B). Likewise, no GRAPE data wereavailable for Section 728C-4H-4 and here the last density valuefrom Section 728C-4H-3 has been used.

As is the case with the °7o Co r g profile, organic carbon accu-mulation rates do not appear to be systematically linked to thestate of global climate (Fig. 8C). For instance, broad carbon ac-cumulation minima occur during oxygen isotope interglacialstage 5 and glacial stage 8 whereas maxima are documented dur-ing interglacial stages 7 and 9, and glacial stage 6 (Fig. 8C). Asdowncore variations of density are comparatively small (Fig.8B), the reduction of the organic carbon signal in the accumula-tion record when compared to the percent profile during stages5 and 8 are mostly controlled by sedimentation rate minima (seeFig. 5). Due to irregularities of our oxygen isotope records (see

above) the implied chronostratigraphies, and thus the inferredsedimentation rate changes, are not tightly constrained. How-ever, it appears unlikely that improvements of the chronostratig-raphy will significantly alter the magnitude of the sedimentationrate variations and we believe that the structure of the organiccarbon accumulation profile (Fig. 8C) can be viewed as reasona-bly documenting the late Quaternary carbon burial rates at theupper Oman continental margin.

DISCUSSION

The Planktonic Isotope Record of Changing SSTand Upwelling

Intense upwelling of cold and nutrient-enriched subsurfacewater causes significant gradients of temperature and nutrientlevels in the surface ocean. Various studies have demonstratedthat these upwelling signals are preserved in the δ 1 8θ and, atleast to some extent, also in the δ13C values of planktonic fora-minifers which have been exposed to the upwelled waters (e.g.,

296

LATE PLEISTOCENE PLANKTONIC AND BENTHIC ISOTOPE RECORDS: SITE 724

Oxygen isotope stages ( EU3 = glacials)

Λ hi 13 11 * 1 L J L J 1*11f\ ;3.85%

ü

2 -

1 •

0 -i

l i t ' I I L

,1 1 f. .1 •tn1 111 i l l ;

)i - i l l y 1

il• i *r ii i

i t

vwK\Λ

fF!i

• i I I

— i — ' r~^—i i

IIit 1Λ • I ‰ <I

I U '1 1

1 1

1 i

1

, 1

1 if lub

i

i

i•

^—T"1

i Ii A

IΛ

Ulvb 1

ftP

ii Vi i

—i—i-r-i 1

100 200 300 400

B

* ε

i ü3 σ>

Q

2 -

1 -

n -

i

hfii

ii

ii

itΛ

Kf**^ IIIr

2H07|3H01

1

3H07|4H01

I

100 200 300 400

100 200 300

Age [1000yr B.P. ]

400

Figure 8. Organic carbon profile for Site 724. A. wt% profile.B. GRAPE wet-bulk-density analog data, corrected to dry-bulk-density.C. Carbon accumulation rates. Percent-to-accumulation rate conver-sion procedure is described in the text.

Berger et al., 1978; Prell and Curry, 1981; Duplessy, 1982; Gans-sen and Sarnthein, 1983; Kroon and Ganssen, 1989). The dis-tinct susceptibility of the foraminiferal isotope composition toupwelling processes is determined by the fractionation responseof δ 1 8 θ and δ13C to varying temperature and nutrient cycling.Prell and Curry (1981) and Duplessy (1982) studied the stableisotope variation of planktonic foraminifera in core top samplesfrom the Western Arabian Sea and found good correlations be-tween upwelling-induced gradients of seasonal sea surface tem-peratures (SST) and the δ 1 8 θ signal of shallow-dwelling plank-tonic foraminifers (G. bulloides and G. ruber). In contrast, plank-tonic δ1 3C apparently does not correlate well with theupwelling-induced nutrient gradient (Prell and Curry, 1981).That is, planktonic δ 1 8 θ appears to be a far better recorder ofupwelling in the Arabian Sea than planktonic δ 1 3C.

In an attempt to extract the potential SST signal from theplanktonic isotope record of Hole 724C, we have estimated thedifference of the δ 1 8 θ variations of G. ruber from the meanocean δ 1 8 θ s e a w a t e r variations. To do this we have constructed asynthetic δ 1 8 θ s e a w a t e r record for the past 350,000 yr by adjustingthe amplitude of δ 1 8 θ variations of the Specmap isotope stack

(Imbrie et al., 1984) (Fig. 9A) to those of the mean oceanδ 1 8 θ s e a w a t e r curve of the past 120,000 yr which was presented byLabeyrie et al. (1987) (Fig. 9B). At least 30% of the variation offoraminiferal δ 1 8 θ on a glacial-to-interglacial time scale is dueto changes of seawater temperature. The synthetic δ̂ Ogeawaterrecord shown in Figure 9C does not contain this temperaturesignal and hence, the range of variation in this record is signifi-cantly reduced. The interglacial δ 1 8 θ level of the converted Spec-map record (Fig. 9C) was then set to - 1.5‰ (PDB). This is themodern δ 1 8 θ value of G. ruber in the Oman upwelling area asimplied by core top values at shallow "inshore" sites ORKS 08(-1.58‰ (PDB), in Duplessy, 1982) and A15597 (-1 .5‰(PDB), in Prell and Curry, 1981). This mean isotope record(bold line in Fig. 10A) should now represent the δ 1 8 θ fluctua-tions of G. ruber solely as a function of mean ocean δ 1 8 θ s e a w a t e r

variations, i.e., as a function of ice volume changes. The syn-thetic record can therefore be considered to depict the expectedvariations in planktonic δ 1 8 θ if no SST changes had occurred atSite 724. Thus, deviations of the G. ruber δ 1 8 θ profile from thissynthetic record should represent the temperature signal in-cluded in the planktonic δ 1 8 θ record. The isotope difference be-

Oxygen isotope stages = glacials )

2.0-

B

-2.0

0.5

100 200 300 400

fc -0.5 -

-1.5

0.5

100 200 300 400

-0.5-

-1.5

100 200 300

Age [ 1000 yr B.P. ]

400

Figure 9. A. Specmap global stack of δ 1 8 θ (from Imbrie et al., 1984).This record has been used to construct a δ 1 8 θ s e a w a t e r record for the past350,000 yr. B. Mean-ocean δ 1 8 θ s e a w a t e r record as given in Labeyrie et al.(1987). C. Synthetic 5 ^ 0 ^ ^ ^ record constructed by adjusting the Spec-map δ 1 8 θ record to the δ 1 8 θ s e a w a t e r record shown in B. This synthetic re-cord was used as a reference to estimate possible SST and δ 1 8 θ s e a w a t e r

variations from glacial-interglacial fluctuations of planktonic and ben-thic δ 1 8 θ at Site 724.

297

R. ZAHN, T. F. PEDERSEN

CDQ

α_

Oαo

- 3 •

-2 -

- 1 •

0 -

Oxygen

•PI 3 |;i 5

11 i

ry&F °ii i i

isotope

iQ l

is**

stages

1 71

( ü iIII1

= glacials)

111 91

V I sL,

c

100 200 300

a G. ruber

Mean ocean(normalized to G.ruber)

400

B | 1-0

g 0.5-

O •* 0.0-00 (A

< E 0.5-

CD -1.5100 200 300 400

O

100 200

Age [1000yr B.P.

300 400

\ | Monsoon index |

Figure 10. A. Comparison of surface water δ 1 8 θ c a l c i t e variations as recorded by G. ruber with mean-oceanδ 1 8 θ variations (heavy line). The original G. ruber δ 1 8 θ record has been smoothed by a 12 k.y.Gaussian filter at a step size of 3 k.y. The mean-ocean δ 1 8 θ s e a w a t e r record is the same as in Figure 9C withthe exception that its interglacial isotope level is adjusted to - 1.5‰ which is the modern δ 1 8 θ value of G.ruber at the coast of Oman (see text). B. δ 1 8 θ difference between the mean-ocean record (normalized toG. ruber and the G. ruber isotope record. C. Temperature deviation from the modern SST as estimatedfrom the isotope difference (crosses) and variation of organic carbon accumulation (open circles). Alsoshown are maxima of the orbital monsoon index which may be used as an indicator for potential upwell-ing intensity. For discussion see text.

tween the measured and synthesized profiles was calculated bysampling both records at 3 k.y. intervals and subtracting themfrom each other (Fig. 10B). Finally, the isotope difference wastransformed into temperature by assuming a mean fractionationresponse of ö^C^te of 0.25‰ per 1°C temperature change.The resulting temperature record (crosses in Fig. 10C) thereforegives temperature as degree centigrade deviation from the mod-ern SST.

The magnitude of the SST deviations from the modern wouldhave been much larger (i.e., ΔSST >8°C) had we chosen themore negative δ 1 8θ value of - 2.7‰ (PDB) as the modern con-trol for the G. ruber isotope record. This value has been ob-tained by Kroon and Ganssen (1989) from living specimens col-lected in plankton tows at 12°N (plankton tow stations 112 and118; Kroon and Ganssen, 1989). However, we believe that thisnegative value records the warm nonupwelling SST typical for

tropical-subtropical surface waters and is not representative ofthe cold upwelled waters extant at Site 724.

Strongly negative deviations of as much as l‰ from themodern δ 1 8θ values are observed in the G. ruber isotope profileat the core top and at the last interglacial substage 5e. Thiswould imply that temperatures at the last climatic optimum(about 6-9 k.y.B.P.) and at the last interglacial have been warmerby 3°-5°C than the modern SST. However, pollen records fromcores collected in the northern Arabian Sea (Van Campo et al.,1982; Prell and Van Campo, 1986) indicate increased southwestmonsoon winds during both these time periods which wouldhave resulted in enhanced upwelling intensity and thus increased(i.e., "colder") planktonic δ 1 8θ values.

The reliability of SST estimates made strictly on the basis ofplanktonic δ 1 8θ inherently depends on the preferred species-spe-cific temperature habitat, the assumed slope of δ 1 8θ equilibrium

298

LATE PLEISTOCENE PLANKTONIC AND BENTHIC ISOTOPE RECORDS: SITE 724

fractionation with respect to temperature, and on local varia-tions of surface water δ̂ Oseawater For instance, large-scale injec-tions of 18O-depleted meltwaters from waning Tibetan glaciersat the end of major glaciations could be one explanation for thepronounced negative excursions of planktonic δ 1 8θ. However,this suggestion begs the question as to why there is no similarmeltwater spike at the end of full-glacial stages 8 and 10. Varia-tions of the clay mineral assemblages in the Western ArabianSea have been considered to indicate increased river runoff fromthe coast of Oman during periods of enhanced humidity in theearly Holocene (Sirocko, 1989). That is, the surface waterδ̂ Oseawate,. values at the Oman Margin could have varied as afunction of continental humidity which is in turn monsoon-driven. The SST recorded in δ 1 8θ of planktonic foraminifers de-pends further on the species-specific environmental preferences,the recorded SST lying between the mean preferred temperatureand the true mean water temperature (Mix, 1987). For our δ1 8θ-SST estimates this means that a preference of G. ruber to calcifyduring the warmer nonupwelling winter season, for instance,would diminish the value of our records as an indicator for up-welling-driven SST variations. However, comparison with theorganic carbon accumulation profile (open circles in Fig. IOC)shows at least some qualitative correlation between warmer SSTand low carbon accumulation and vice versa during isotopestages 3, 5, 7, and 9 (Fig. IOC). This correlation breaks downduring stages 6 and 9 where relative SST minima correlate withcarbon accumulation minima (Fig. IOC). Furthermore, compar-ison with the orbital insolation monsoon index (MI; as definedby Rossignol-Strick, 1983) shows no correlation (r2 < 0.1, n =113) between predicted enhanced monsoon-driven upwelling andestimated SST and organic carbon anomalies (Fig. IOC).

The downcore profile of estimated SST oscillation impliesthat local temperatures were in general colder during glacialstages than during interglacial stages (Fig. IOC). This appears tobe consistent with the conclusions of ten Haven and Kroon (thisvolume) who used the L$ ratio in the lipid fraction of sedimentsfrom the Owen Ridge to infer that the SST in the northern Ara-bian Sea was 2°-3°C colder during glacial periods than today.They suggested that the lower temperatures reflected cooling bycold adiabatic winds issuing from the Tibetan glaciers as a com-ponent of the prevailing northeasterlies. Such SST changes wouldhave been superimposed on the upwelling-driven SST variationsand could thus be another explanation for the poor correlationbetween the predicted strength of the monsoonal winds and ourSST estimates.

The Benthic Intermediate Water Isotope Record

The intermediate water hydrography at the Oman margin isdetermined by the coastal upwelling regime and the advection ofRed Sea outflow waters at depth. As mentioned earlier, the com-bination of high rates of input of organic matter from the highlyproductive surface waters and subsequent oxidation below thethermocline together with low renewal rates of intermediate depthwaters result in a pronounced oxygen deficiency between 200and 1500 m. Marine organic matter and the CO2 which is re-leased during its oxidation and added to the ambient water Σ C O 2

are significantly depleted in δ13C ( - 20‰ to - 25‰ PDB). As aresult, δ13C2;cθ2 values in the mid-depth Arabian Sea are de-pleted when compared to the regional deep waters (e.g., Kroop-nick, 1985). Warm and saline outflow waters from the Red Seaalter the temperature and salinity characteristics of the local in-termediate waters (Fig. 1). In addition, Red Sea waters have in-creased δ13CECθ2 values of about +0.9‰ as revealed by the ver-tical isotope profile of Geosecs Station 407 at the northern en-trance of the Bab el Mandab Strait (Ostlund et al., 1987). Also,Red Sea δ 1 8θ s e a w a t e r values of + 1.9‰ (SMOW) (Craig, 1966) areconsiderably higher than the δ 1 8θ s e a w a t e r values of + 0.2‰ to

+ 0.4‰ (SMOW) which are encountered below the core layer ofRed Sea outflow water in the Arabian Sea (Fig. 11 A). There-fore, variations of the organic carbon flux from the surface wa-ters and of the advective flow from the Red Sea during the latePleistocene have the potential to alter considerably the benthicδ1 8θ and δ13C records at Site 724.

mQ

ü

+2.0

+1.5

+1.0

+0.5

0.5

Southern Oceanintermediate water

ΔRed Sea outflow

Φ Oman Margin

Arabian Sea water

B

0 +1 +2

δ 1 β θ s e a w a t e r (‰SMOW)

Ice effect "

O

Q.

ECD

# Modern

O Glacial

-0.2 0 +0.2 +0.4 +0.6 +0.8 +1.0 +1.2

Δ δ 1 8 O s e a w a t e r

Figure 11. A. δ18OseaWater-δ13Cj;cθ2 composition of water masses in theArabian Sea. Red Sea outflow data are from Geosecs station 407(δ^Cscc^) t o t n e north of the Bab el Mandab Strait and from Craig(1966) (á^Ose^ate,.). Arabian Sea data are from Geosecs stations 413,416, 417, and 418 at depth of about 1000 m, i.e., below the core layer ofRed Sea waters. The isotope composition of intermediate waters at theOman Margin was estimated by assuming a mixing ratio of 2:8 betweenRed Sea waters and Arabian Sea waters (Siedler, 1968). Southern Oceanintermediate water data are from Geosecs stations 427, 428, and 429 atdepths between 400 and 600 m. All Geosecs data are from Ostlund et al.(1987). A cessation of Red Sea outflow and its replacement by SouthernOcean waters would result in a depletion of δ 1 8 θ s e a w a t e r at Site 724 byabout 0.4‰ whereas the δ13Cj;cθ2 values would remain about the sameor increase slightly. B. Temperature-δ1 8O s e a w a t e r plot showing isolines ofequilibrium S ^ O c a ^ (calculated using the paleotemperature equationof Shackleton, 1974). Glacial 1: All of the glacial-interglacial benthicδ 1 8 θ shift is due to changes in δ 1 8 θ s e a w a t er (1 2‰ ice volume minus 0.4‰due to diminished Red Sea outflow). Glacial 2: All of the glacial-inter-glacial benthic δ 1 8 θ shift is due to changes in seawater temperature. As-suming an enhanced influence of cold and 18O-depleted AAIW on themid-depth hydrography at Site 724, a true glacial point in the diagrammay be somewhere between 1 and 2; δ 1 8 θ s e a w a t e r of this glacial AAIWmust have been depleted by at least 0.5‰-1.2‰ relative to the ambientwaters so as to compensate the temperature effect.

299

R. ZAHN, T. F. PEDERSEN

The mean ocean ö^O^^g,. change between the last glacialmaximum (approximately 18,000 yr ago) and the present inter-glacial has been estimated to be about 1.2-1.3‰ (Labeyrie etal., 1987; Fairbanks, 1989). This global isotope shift is of aboutthe same order as the apparent δ 1 8θ s e a w a t e r contrast between RedSea waters and Arabian Sea waters. Global sea level was lowerat the last glacial maximum by about 120 m (Chappell andShackleton, 1986) and probably even lower during glacial-maxi-mum oxygen isotope stages 6, 10, and 12. During these stagesbenthic isotope records show δ 1 8θ values which are usually morepositive than those observed at the last glacial maximum imply-ing the existence of even larger global ice volumes and thereforelower sea level stands than at 18,000 yr ago. As the sill depth be-tween the Red Sea and the Gulf of Aden is only about 140 m,the connection between the Red Sea and the Indian Ocean musthave been reduced in depth to < 20 m or even entirely closed offduring glacial periods. As a consequence, the advective flowfrom the Red Sea to the mid-depth Arabian Sea would havebeen reduced or entirely shut down during glacial maxima. AsFigure 11A shows, the net effect of a complete cessation of theRed Sea outflow and replacement by Arabian Sea waters wouldbe a decrease of δ 1 8θ s e a w a t e r and δ13Cj;cθ2 a t s i t e 7 2 4 b v respec-tively 0.4‰ and 0.3‰. Comparing our benthic δ 1 8θ record withthe synthetic δ 1 8 θ s

record indeed indicates that the δ 1 8θvalues at Site 724 during glacial stages 2-4 and 6 are more nega-tive by O.3‰-O.5‰ than expected strictly on the basis of the icevolume-δ18θseawater effect (Fig. 12A). The depleted benthic δ 1 8θvalues during glacial periods could therefore be explained bythen reduced (or absent) advection of Red Sea waters to theArabian Sea. Heavier than expected δ t 8 θ values are recordedduring interglacial substage 5e and stage 7, and during early gla-

Oxygen isotope sta ( IZU = glacials )

. Site 724

— Mean ocean

B

Site 724

V19-30

0 100 200 300 400

Age [1000yr B.P. ]

Figure 12. A. Comparison of the smoothed benthic δ 1 8 θ record at Site

724 (solid squares) with the mean-ocean variation of δ I 8 O s e a w a t e r (bold

line). The δ 1 8 θ s e a w a t e r record is the same as in Figure 9C with the excep-

tion that its interglacial isotope level was adjusted to the benthic core

top value of +1.9‰ (PDB). B. The smoothed epibenthic δ 1 3 C record

from Site 724 (open squares) is compared to the deep-ocean δ 1 3C record

of core V19-30 (solid line) (Cibicidoides scale; Shackleton and Pisias,

1985). We interpret the combined signal of depleted benthic δ 1 8 θ and

constant or enriched epibenthic δ 1 3 C as documenting the replacement of

Red Sea outflow waters by intermediate waters low in δ 1 8 θ s e a w a t e r and

similar or increased in δ13Cj;cθ2 These waters could have originated

from the Southern Ocean (see text and Fig. 11).

cial stage 8. At this point we cannot determine whether theseisotope features reflect real hydrographic signals or whether theyare due to inaccurate δ 1 8θ s e a w a t e r estimates and/or downslope ad-vection of sediments from shallower ("warmer") water depths.

The epibenthic δ13C record is far too incomplete to permit adetailed evaluation of intermediate water δ^Cjxx^ variations (Fig.12B). The few data available do not show a systematic glacial-interglacial pattern of variation. δ13C values are more positivethan the modern value (which is about + 0.2‰; see Fig. 11Aand also core top value of Cibicidoides spp. in Fig. 7) duringmid-stage 5 and at the last glacial maximum. From the last in-terglacial back to oxygen isotope stage 7, the δ13C values showonly minor variations and remain close to their modern level.The combined pattern of δ 1 8θ and δ13C variation at Site 724(Fig. 12) suggests that glacial intermediate-depth waters at theOman Margin were lower in δ 1 8θ s e a w a t e r than Red Sea waters butsimilar or occasionally even enriched in δ13Ci;cθ2

Modern sources for intermediate water in the Northern In-dian Ocean are manifold and include Pacific waters which areadvected by way of the Banda Sea (Godfrey and Golding, 1981),south equatorial intermediate waters (Luyten et al., 1980; Quad-fasel and Schott, 1982), and possible upwelling of waters fromdeep western boundary currents (Warren, 1981). The isotopiccompositions of these water masses are largely unknown andtherefore we cannot define the ultimate origin of the benthicδ18O/δ13C variations at Site 724. However, glacial-interglacialbenthic isotope variations at intermediate water sites were con-siderably different from those at pelagic core sites throughoutthe northern Indian Ocean (Kallel et al., 1988). This may pointto the existence of large-scale convection such as occurs in theSouthern Ocean. There, large negative δ1 8θ anomalies have beenobserved in glacial sections of planktonic δ 1 8θ records whichhave been interpreted as documenting enhanced meltwater inputduring successive surges of the ice shelves (Labeyrie et al., 1986).This meltwater contribution could enhance the low-δ18θseawater

character of Antarctic Intermediate Water (AAIW) and thuspartly explain the benthic δ 1 8θ depletion at Site 724 during gla-cial times.

Penetration of AAIW to the north is currently restricted toabout 15°S by hydrographic fronts associated with the equato-rial current system (e.g., Warren, 1981). In the absence of theRed Sea source for intermediate waters, AAIW could have spreadfarther north, thus influencing the ambient water hydrographyat Site 724 (see, e.g., Kallel et al., 1988). The δ ^ O , ^ ^ signalof this glacial AAIW must have been more negative by at least0.4‰-1.2‰ so as to compensate the effect that its low tempera-ture would have had on the benthic δ1 8θ values at Site 724 (Fig.11B). Because a further depletion in dissolved oxygen at theOman Margin appears inconceivable, the low epibenthic δ13Cvalues prior to stage 8 (Fig. 12C) most likely imply a shift fromlow-δ13CEcθ9 to high-δ13QCo2 source waters, i.e., a shift in thepreformed δ*3Cεcθ2 value.

Uvigerina spp. δ13C and Organic Carbon Variations:Records of Pore Water δ^Cεα>j a n d Paleoproductivity

The δ13C record obtained from endobenthic Uvigerina spp.appears to vary independently from that of epibenthic Cibici-doides spp. (Fig. 7). The interspecific difference in δ13C (Δδ13C)has already been proposed to be driven by varying burial ratesof organic carbon and oxidation of organic matter in the sedi-ments (Zahn et al., 1986; McCorkle et al., 1989). This relation-ship has been confirmed by the comparison of in-situ measure-ments of pore water Ò13CΣCO2

a n d organic carbon fluxes at therespective sites (McCorkle, 1987). The preservation of the δ13Q;cθ2gradient of pore water to bottom water in turn appears to beprimarily controlled by the thickness of the stagnant sublayerabove the sediment surface (Archer et al., 1989). In this view, in-

300

LATE PLEISTOCENE PLANKTONIC AND BENTHIC ISOTOPE RECORDS: SITE 724

creased Δδ13C values would suggest that the burial rate of or-ganic matter was generally enhanced during glacial times (Fig.13A). However, intermediate water δ1 3CE Co2

a t Site 724 appearsto be highly variable through time as a function of various sourcewater δ13CECo2

a n d hence, one could argue that the Δδ13C varia-tions through time at Site 724 are also a function of the pre-formed intermediate water δ13Cj;cθ2 It is especially important tonote that the variation seen in the interspecific Δδ13C values atSite 724 reflects primarily δ13C fluctuations of endobenthic Uvi-gerina spp. (compare Fig. 13A with the benthic δ13C profiles inFig. 7), which appears to record the organic carbon flux morefaithfully than the epibenthic Cibicidoides spp. In fact, the Uvi-gerina spp. δ13C profile appears to trace the organic carbon ac-cumulation profile during the past 200 k.y. in that high carbonaccumulation rates coincide with more negative Uvigerina spp.δ13C values, and vice versa (Fig. 13). Prior to this period, bothparameters are not well correlated; whereas carbon accumula-tion varies at about the same or an even larger order, Uvigerinaspp. δ13C remains comparatively stable.

As with the organic carbon accumulation rates, we do notobserve any clear correlation between the orbital monsoon in-dex (MI) which may be used as an indicator for enhanced mon-soonal winds and more intense upwelling, and the Uvigerinaspp. δ13C profile (r2 < 0.1, n = 146). Maximum strengths ofmonsoonal winds (as indicated by high MI values) which corre-late with organic carbon accumulation maxima during earlystage 1, mid and late stage 7, and late stage 9, do not corre-spond with corresponding Uvigerina spp. δ13C minima. The lack

of correlation between the organic carbon accumulation rate,Uvigerina spp. δ13C and assumed monsoonal wind strength aspredicted from the insolation monsoon index, begs the ques-tion: why do productivity indicators such as the abundance ofthe planktonic foraminifer Globigerina bulloides (Prell and VanCampo, 1986) presumably trace the history of upwelling at theOman Margin more accurately than more direct tracers such asorganic carbon? This apparent conflict cannot be resolved atpresent. Based on the data available to date it seems that theOman Margin upwelling system is more complex than previ-ously assumed, unless we infer that most of the signals recordedat Site 724 are disturbed by downslope sediment transport. In-dependent proxy records for SST and productivity, e.g., fromtransfer functions and the nutrient-related tracer Cd/Ca areneeded to better constrain variations of paleoproductivity at Site724, and to support (or reject) the hypothesis that upwellingvariations at the Oman Margin during the Quaternary shouldbe primarily controlled by variations in strength of the mon-soonal winds. Such records are not yet available.

SUMMARY AND CONCLUSIONS

Time series of benthic and planktonic δ 1 8 θ and δ13C, and ofCo r g have been obtained for the past 350,000 yr at Site 724. Theδ13C record of endobenthic Uvigerina spp. varies by up to l‰over the last three glacial-interglacial cycles. At the same time,δ13C values recorded by epibenthic Cibicidoides spp. remain con-stant or are slightly increased. The Uvigerina spp. δ13C profiletraces variations in organic carbon accumulation during the past

A

CDaQ_

ü

Oxygen isotope stages ( glacials )

Cibicidoides spp.minus

Uvigerina spp.

100 200 300

Uvigerina spp.

400

B 300

U

- ^ 200 -

oσ>E 100 -

»38.8

l^i44l45ll^l^l^l5il62l46ll25ll58l45ll5θTi8l •< 1 Monsoon index |

100

Age

200

1000 yr B.P.]

300 400

Figure 13. Comparison of the benthic δ13C productivity index (A) to the estimated variations of organic carbon ac-cumulation (B). Uvigerina spp. δ13C traces the variations of organic carbon accumulation, at least during the past200 ky. However, there is no clear correlation between the δ13C data, organic carbon accumulation and the orbitalmonsoon index.

301

R. ZAHN, T. F. PEDERSEN

200,000 yr and supports the hypothesis that endobenthic δ13Cshould vary primarily as a function of the supply and oxidationof organic carbon in the upper few centimeters of the surfacesediments (Zahn et al., 1986; McCorkle et al., 1989). Prior tothis period, however, the correlation between Uvigerina spp.δ13C and accumulation of organic carbon is poor.

The planktonic δ1 8θ record has been used to infer that gla-cial SSTs at the Oman Margin were lower than they are today.We have observed some graphical correlation between negative(colder) SST anomalies as deduced from planktic δ 1 8θ and en-hanced carbon accumulation at Site 724. This implies an influ-ence of up welling intensity on paleoproductivity levels, as hasbeen previously observed at other upwelling regions. However,the combined SST-Corg signals do not correlate with the orbitalmonsoon index (MI) and with pollen indices which have bothbeen used as proxies for the strength of monsoonal winds, andpotential upwelling intensity (Van Campo et al., 1982; Rossi-gnol-Strick, 1983; Prell and Van Campo, 1986). In view of ournew data it appears quite surprising that faunal upwelling indi-ces which are based on the relative abundance of G. bulloidespresumably show a better correlation to the orbital monsoon in-dex, and hence to potential upwelling intensity (Prell, 1984; Prelland Van Campo, 1986), than the more direct productivity proxyCorg. Without independent SST and productivity estimates, e.g.,from transfer function estimates and benthic-planktonic Cd/Cadata, we cannot resolve this conflict.

The benthic δ 1 8θ record varies at a lower amplitude than theestimated δ 1 8θ s e a w a t e r record. In respect to the constant or ele-vated epibenthic δ13C levels we conclude that the source of inter-mediate waters systematically changed between glacial and in-terglacial periods. During interglacials the mid-depth hydrogra-phy at the Oman Margin is significantly influenced by outflowwaters from the Red Sea. Red Sea waters are increased in bothδ18θseawater and δ13C£co2 During glacial periods the Bab el Man-dab Strait was reduced to < 20 m in depth or even completelyclosed off so that advection of Red Sea waters to the ArabianSea was drastically reduced if not completely switched off. Thebenthic isotope data imply that there was advection of interme-diate waters to the Oman Margin which were still enriched inδ13Cj:cθ2 but depleted in δ 1 8θ s e a w a t e r . These waters could haveoriginated from the Southern Ocean where modern intermedi-ate waters are enriched in δ13CEco2 with respect to the underly-ing deep and bottom waters. Glacial depletion in δ 1 8θ s e a w a t e r ofSouthern Ocean intermediate waters could have been caused bysporadic surges of the ice shelf (Labeyrie et al., 1986).

In general, the constantly enriched epibenthic δ13C data atSite 724 appear to be consistent with similar observations fromthe mid-depth North Atlantic and North Pacific Oceans andsupport the hypothesis (Boyle, 1988b) that the hydrochemicalcycles of the mid-depth ocean during glacial-interglacial timesare decoupled from those of the abyssal ocean. Comparison ofsuitable geochemical proxy records from shallow and deep HPCcores which have been obtained during Leg 117 may eventuallyprovide the first long time series for studying the long-term be-havior of the ocean's vertical asymmetry and its relation to globalclimatic change.

ACKNOWLEDGMENTS

We are very grateful to the shipboard technical staff duringLeg 117 for their assistance with sampling. In particular, the ef-forts of Bob Wilcox are highly appreciated. We would like tothank Bente Nielsen for her careful operation of the mass spec-trometer at UBC, and Shirley French and Mark Noyon for theirassistance in the preparation of the foraminifer samples. We aregrateful to Shirley French and Maureen Soon for carefully run-ning the carbon analyses. Laurent Labeyrie provided us with a

digitized version of his mean-ocean δ 1 8θ s e a w a t e r record, and BillBusch expeditiously supplied bulk density data on short notice.The manuscript benefitted from constructive reviews by WarrenPrell, Alan Mix, Dick Kroon, William Curry, and an anony-mous reviewer. TFP acknowledges the financial support of theNatural Sciences and Engineering Research Council of Canada.

REFERENCES

Archer, D., Emerson, S., and Smith, C , 1989. Direct measurement ofthe diffusive sublayer at the deep sea floor using oxygen microelec-trodes. Nature, 340:623-626.

Balsam, W. L., and Posmenier, E. S., 1981. Comments on spectral anal-ysis of core data. Palaeogeogr., PalaeoclimatoL, PalaeoecoL, 45:369-372.

Berger, W. H., Diester-Haas, L., and Killingley, J. S., 1978. Upwellingoff Northwest Africa: the Holocene decrease as seen in carbon iso-topes and sedimentological indicators. Oceanol. Acta, 1:1-7.

Berger, W. H., and Heath, G. R., 1968. Vertical mixing in pelagic sedi-ments. J. Mar. Res., 26:134-143.

Bethoux, J. P., 1987. Variabilité climatique des échanges entre la MerRouge et 1'Océan Indian. Oceanol. Acta, 10:285-291.

, 1988. Red Sea geochemical budgets and exchanges with theIndian Ocean. Mar. Chem., 24:83-92.

Boyle, E. A., 1988a. Cadmium: chemical tracer of deepwater paleocean-ography. Paleoceanography, 3:471-490.

, 1988b. The role of vertical chemical fractionation in control-ling late Quaternary atmospheric carbon dioxide. J. Geophys. Res.,93:15701-15714.

Chappell, J., and Shackleton, N. J., 1986. Oxygen isotopes and sealevel. Nature, 324:137-140.

CLIMAP Project Members, 1984. The last interglacial ocean. Qµat.Res., 21:123-224.

Corliss, B. H., 1985. Microhabitats of benthic foraminifera within deep-sea sediments. Nature, 314:435-438.

Craig, H., 1966. Isotopic composition and origin of the Red Sea andSalton Sea geothermal brines. Science, 154:1544-1548.

Curry, W. B., Duplessy, J. C , Labeyrie, L. D., and Shackleton, N. J.,1988. Changes in the distribution of δ1 3 C of deep water ECO2 be-tween the last glacial and the Holocene. Paleoceanography, 3:317-341.

Cushing, D. H., 1973. Production in the Indian Ocean and the transferfrom the primary to the secondary level. In Zeitschel B. (Ed.), TheBiology of the Indian Ocean. New York (Springer).

Düing, W., and Leetmaa, A., 1980. Arabian Sea cooling: a preliminaryheat budget. Bull. Am. Meteorol. Soc, 10:307-312.

Düing, W., and Schwill, W.-D., 1968. Ausbreitung und Vermischungdes salzreichen Wassers aus dem Roten Meer und aus dem Persis-chen Golf. "Meteor"Forschungsergeb. Reihe A, 3:44-66.

Duplessy, J. C , 1982. Glacial to interglacial contrasts in the northernIndian Ocean. Nature, 295:494-498.

Duplessy J. C , Shackleton, N. J., Fairbanks, R. G., Labeyrie, L., Oppo,D., and Kallel, N., 1988. Deepwater source variations during the lastclimatic cycle and their impact on global deepwater circulation. Pa-leoceanography, 3:343-360.

Fairbanks, R. G., 1989. A 17,000-year glacio-eustatic sea level record:influence of glacial melting rates on the Younger Dryas event anddeep-ocean circulation. Nature, 342:637-642.

Ganssen, G., and Sarnthein, M., 1983. Stable-isotope composition offoraminifers: the surface and bottom water record of coastal upwell-ing. In Suess, E., and Thiede, J. (Eds.), Coastal Upwelling, Its Sedi-ment Record. New York (Plenum), 10A:99-121.

Godfrey, J. S., and Golding, T. J., 1981. The Sverdrup relation in the In-dian Ocean, and the effect of Pacific-Indian Ocean throughflow onIndian Ocean circulation and on the East Australian Current. /.Phys. Oceanogr., 11:771-779.

Imbrie, J., Hays, J. D., Martinson, D. G., Mclntyre, A., Mix, A. C ,Morley, J. J., Pisias, N. G., Prell, W. L., and Shackleton, N. J.,1984. The orbital theory of Pleistocene climate: support from a re-vised chronology of the marine delta δ 1 8 θ record. In Berger, A., Im-brie, J., Hays, J., Kukla, G., and Saltzman, B. (Eds.), Milankovitchand Climate (Pt. 1): Dordrecht (D. Reidel), 269-305.

Kabanova, Yu. G., 1968. Primary production of the northern part ofthe Indian Ocean. Oceanography, 8:214-225.

302

LATE PLEISTOCENE PLANKTONIC AND BENTHIC ISOTOPE RECORDS: SITE 724

Kaliel, N., Labeyrie, L. D., Juillet-Leclerc, A., and Duplessy, J.-C,1988. A deep hydrological front between intermediate and deep-wa-ter masses in the glacial Indian Ocean. Nature, 333:651-655.

Krey, J., 1973. Primary production in the Indian Ocean. In Zeitzschel,B., and Gerlach, S. A. (Eds.), The Biology of the Indian Ocean:New York (Springer-Verlag), 115-126.

Kroon, D., and Ganssen, G., 1989. Northern Indian Ocean upwellingcells and the stable isotope composition of living planktonic fora-minifers. Deep-Sea Res. Part A, 36:1219-1236.

Kroopnick, P., 1980. The distribution of 13C in the Atlantic Ocean.Earth Planet. Sci. Lett., 49:469-484.

Labeyrie, L., Pichon, J.-J., Labracherie, M., Ippolito, P., Duprat, J.,and Duplessy, J.-C, 1986. Melting history of Antarctica during thepast 60,000 years. Nature, 322:701-706.

Labeyrie, L. D., Duplessy, J. C , and Blanc, P. L., 1987. Variations inmode of formation and temperature of oceanic deep waters over thepast 125,000 years. Nature, 327:477-482.

Lutze, G. R, 1986. Uvigerina species of the eastern North Atlantic. Invan der Zwaan, G. J., et al. (Eds.), Atlantic-European Oligocene toRecent Uvigerina. Utrecht Micropaleontol. Bull., 35:21-46.

Lutze, G. R, and Thiel, H., 1987. Cibicidoides wuellerstorß and Planu-lina ariminensis, elevated epibenthic foraminifera. In Altenbach,A. V., et al. (Eds.), Beobachtungen an Benthos-Foraminiferen. Re-port Sonderforschungsbereich 313, Univ. of Kiel, Federal Republicof Germany: 17-30.

Luyten, J. R., Fieux, M., and Gonelia, J., 1980. Equatorial currents inthe western Indian Ocean. Science, 209:600-602.

McCorkle, D. C , 1987. Stable carbon isotopes in deep sea pore waters:modern geochemistry and paleoceanographic applications. [Ph.D.dissert.]. Univ. Washington.

McCorkle, D. C , Emerson, S. R., and Quay, P., 1985. Stable isotopesin marine pore waters. Earth Planet. Sci. Lett., 74:13-26.

McCorkle, D. C , Keigwin, L. D., Corliss, B. R., and Emerson, S. R.,1989. The δ13C of Uvigerina and the pore water δ1 3 C gradient at thesediment-water interface. Third Inter. Conf. on Paleoceanography,London (Blackwell Scientific Publications), 63.

Mix, A. C , 1987. The oxygen isotope record of glaciation. In Ruddi-man, W. R, and Wright, H. E. (Eds.), North America and AdjacentOceans during the Last Deglaciation, The Geology of North Amer-ica (Vol. K-3): Geol Soc. Am., 111-135.

Nair, R. R., Ittekkot, V., Manganini, S. J., Ramaswamy, V, Haake, B.,Degens, E. T., Desai, B. N., and Honjo, S., 1989. Increased particleflux to the deep ocean related to monsoons. Nature, 338:749-751.

Ostlund, H. G., Craig, C , Broecker, W. S., and Spencer, D. (Eds.),1987. Geosecs Atlantic, Pacific, and Indian Ocean expeditions. Shore-based data and graphics. Geosecs Atlas Series (Vol. 7): Washington(U.S. Govt. Printing Office).

Prell, W. L., 1984. Monsoonal climate of the Arabian Sea during thelate Quaternary: a response to changing solar radiation. In Berger,A. L., Imbrie, J., Hays, J., Kukla, G., and Saltzman, B. (Eds.). Mi-lankovitch and Climate (Pt. 1): Dordrecht (D. Reidel), 349-366.

Prell, W. L., and Curry, W. B., 1981. Faunal and isotopic indices ofmonsoonal upwelling: western Arabian Sea. Oceanol. Acta, 4:91-98.

Prell, W. L., and Van Campo, E., 1986. Coherent response of ArabianSea upwelling and pollen transport to late Quaternary monsoonalwinds. Nature, 323:526-528.

Premchand, K., Sastry, J. S., and Murty, C. S., 1986a. Watermass struc-ture in the western Indian Ocean—Part III: the spreading and trans-formation of Red Sea watermass. Mausam, 37:317-324.

, 1986b. Watermass structure in the western Indian Ocean—Part II: the spreading and transformation of Persian Gulf water.Mausam, 37:179-186.

Qasim, S. Z., 1982. Oceanography of the northern Arabian Sea. Deep-Sea Res. Part A, 29:1041-1068.

Quadfasel, D. R., and Schott, F., 1982. Water mass distributions at in-termediate layers off the Somali Coast during the onset of the south-west monsoon 1979. J. Phys. Oceanogr., 12:1358-1372.

Resig, J., 1981. Biogeography of benthic foraminifera of the northernNazca Plate and adjacent continental margin. In Kulm, L., Dy-mond, J., Dasch, E. J., and Hussong, D. (Eds.), Nazca Plate: CrustalFormation and Andean Convergence. Mem. Geol. Soc. Am., 154:619-666.

Rossignol-Strick, M., 1983. African monsoons, an immediate climateresponse to orbital insolation. Nature, 304:46-49.

Schemainda, R., Nehring, D., and Schulz, S., 1975. OzeanologischeUntersuchungen zum Produktionspotential der nordwestafrikani-schen Wasserauftriebsregion 1970-1973. Geod. Geophys. Veröff,16:3-84.

Schott, R, and Quadfasel, D. R., 1982. Variability of the Somali Cur-rent system during the onset of the southwest monsoon of 1979. /Phys. Oceanogr., 12:1343-1357.

Shackleton, N. J., and Opdyke, N. D., 1973. Oxygen isotope and pa-laeomagnetic stratigraphy of Equatorial Pacific core V28-238: oxy-gen isotope temperatures and ice volumes on a 105 year and 106 yearscale. Quat. Res., 3:39-55.

Shackleton, N. J., and Pisias, N. G., 1985. Atmospheric carbon diox-ide, orbital forcing, and climate. In Sundquist, E. T., and Broecker,W. S. (Eds.), The Carbon Cycle and Atmospheric CO2:. NaturalVariations Archean to Present. Am. Geophys. Union Monogr., 32:303-317.

Siedler, G., 1968. Schichtungs und Bewegungsverhàltnisse am Süd-ausgang des Roten Meeres. "Meteor"Forschungsergeb. Reihe A, 4:1-76.

Sirocko, R, 1989. Accumulation of eolian sediments in the northwestIndian Ocean: record of the climatic history of Arabia and India.Rep. Geol. Inst. Kiel, 27.

Souvermezoglou, E., Metzl, N., and Poisson, A., 1989. Red Sea bud-gets of salinity, nutrients and carbon calculated in the Strait of Bab-El-Mandab during the summer and winter seasons. J. Mar. Res., 47:441-456.

Swallow, J. C , Molinari, R. L., Bruce, J. G., Brown, O. B., and Evans,R. H., 1983. Development of near-surface low pattern and watermass distribution in the Somali Basin in response to the southwestmonsoon of 1979. /. Phys. Oceanogr., 13:1398-1415.

Van Campo, E., Duplessy, J. C , and Rossignol-Strick, M., 1982. Cli-matic conditions deduced from a 150-k.y. oxygen isotope-pollen re-cord from the Arabian Sea. Nature, 296:56-59.

Warren, B. A., 1981. Transindian hydrographic section at Lat.l8°S:property distributions and circulation in the southern Indian Ocean.Deep-Sea Res. Part A, 28A:759-788.

Wyrtki, K., 1971. Oceanographic Atlas of the International Ocean Ex-pedition: Washington (Nat. Sci. Found.).

, 1973. Physical oceanography of the Indian Ocean. In Zeit-schel, B., and Gerlach, S. A. (Eds.), The Biology of the IndianOcean (Vol. 3): New York (Springer-Verlag), 18-36.

Zahn, R., Sarnthein, M., and Erlenkeuser, H., 1987. Benthic isotopeevidence for changes of the Mediterranean outflow during the lateQuaternary. Paleoceanography, 2:543-559.

Zahn, R., Winn, K., and Sarnthein, M., 1986. Benthic foraminiferalδ13C and accumulation rates of organic carbon: Uvigerina peregrinagroup and Cibicidoides wuellerstorfi. Paleoceanography, 1:27-42.

Date of initial receipt: 2 October 1989Date of acceptance: 8 May 1990Ms 117B-162

303

R. ZAHN, T. F. PEDERSEN

APPENDIX AStable isotope date for Site 724, hole C

Core,section

1H-11H-11H-11H-11H-11H-11H-11H-21H-21H-21H-21H-21H-22H-2H-2H-2H-2H-2H-2H-12H-12H-22H-22H-22H-22H-22H-22H-22H-32H-32H-32H-32H-32H-32H-32H-32H-42H-42H-42H-42H-42H-42H-42H-52H-52H-52H-52H-52H-52H-52H-52H-62H-62H-62H-62H-62H-62H-62H-72H-72H-72H-73H-3H-13H-13H-13H-3H-3H-23H-23H-23H-23H-23H-23H-:

Depthin section

(cm)

22-2442-4462-6482-84

102-104122-124142-144

12-1432-3452-5472-7492-94

108-1107-9

22-2442-4462-6482-84

102-104122-124142-144

12-1432-3452-5472-7492-94

112-114132-134

2-422-2442-4462-6482-84

102-104122-124142-144

12-1432-3452-5472-7492-94

112-114132-134

2-422-2442-4462-6482-84

102-104122-124142-144

12-1432-3452-5472-7492-94

112-114132-134

2-422-2442-4462-6442-4462-6482-84

102-104122-124142-144

12-1432-3452-5472-7492-94

112-114132-134

Sub-bottomdepth

(m)

0.220.420.620.821.021.221.411.621.822.022.222.422.582.873.023.223.423.623.824.024.224.424.624.825.025.225.425.625.826.026.226.426.626.827.027.227.427.627.828.028.228.428.628.829.029.229.429.629.82

10.0210.2210.4210.6210.8211.0211.2211.4211.6211.8212.0212.2212.4212.6212.8213.0213.2213.4213.6213.8214.0214.2214.4214.6214.8215.02

Age(1000 yrB.P.)

6.1337.4678.800

10.13311.46712.80014.06715.46716.80018.13319.46720.80021.86723.80025.19327.18229.17031.15933.14835.13637.12539.11441.10243.09145.08047.06849.05751.04553.03455.02357.01159.00061.00063.00065.00067.00069.00071.00076.75182.14387.71493.28698.857

104.429110.000116.000122.000123.182124.364125.545126.727127.909129.091130.273131.455132.636133.818135.000137.769140.538143.308146.077148.846151.615154.385157.154159.923162.692165.462168.231171.000172.500174.000175.500177.000

Uvigerina spp.

(‰ PDB)

1.180a2.160a2.630

1.7601.9902.0702.4502.5102.5402.2302.3402.3702.4902.2902.2522.280

2.0102.2002.3102.2602.3102.4402.2402.3601.8802.0902.1702.0102.1902.1902.3502.5102.2502.2602.3602.3302.0401.9501.9502.2101.9501.9902.0402.0602.0901.9902.0901.9302.1702.3102.1202.4102.3502.1602.4402.7702.8202.5602.6002.710

b2.82O/2.37O2.3802.2602.5702.5902.4302.5602.4102.1502.1202.2702.2002.3302.440

(‰ PDB)

-0.490a - 0 . 6 0 0a - 0 . 6 5 0

-0.410-0.850-0.630-0.730-0.720-0.710-1.02-0.830-0.710-0.610-0.700-0.718-0.600

-0.790-0.740-0.820-0.720-0.630-0.670-0.900-0.900- 0.840-0.940-1.140-0.330-0.590-0.130-0.220-1.14-0.340-0.280-0.430-0.280-0.250-0.380-0.410-0.330-0.280-0.400

0.000-0.440-0.380-0.442-0.380-0.478-0.520-0.940-0.570-0.750-1.140-1.080-1.230-1.130-1.220-1.270-1.160-1.210

b -1 .220/-1 .300-1.340-1.110-0.910-1.230-0.800-1.24-1.40-1.17-0.440-0.790-0.620-0.450-0.640

Cibicidoides spp.

δl8θ(‰ PDB)

1.040b 1.370/0.926

a1.435a1.589

1.677a1.618

1.8311.697

a2.484a1.289a1.997

1.6461.702

a1.655

a1.595

1.496a1.006a1.104

bl.820/1.898a1.602

1.637

1.7001.592

"0.9511.2481.465

a1.6331.4661.4211.5261.371

1.4591.436

^.700/0.930

a1.2161.933

"0.789

a1.562a1.697a1.652

a2.102a2.216

a l .970/2.013a1.749

a1.881a1.872a1.958

1.979

(‰ PDB)

0.230b0.210/0.133

a0.260a0.349

0.745a0.6320.6740.389

aO.8O3a0.264^.486

0.5410.605

a0.403

a0.259

0.310a0.285^.311

b0.430/0.472

0.432

0.4290.264

^.1580.8410.724

a0.5560.5770.3010.2090.392

0.1890.360

a-O.O3O/O.O3O

a0.2110.509

''o.osi

^.498a0.106^.239

a0.030a0.332

a0.220/0.150a - 0.063

^.036a0.241"0.2130.191

G. ruber (white)

(‰ PDB)

-3.010-2.130

*-2.110-1.910-0.993-0.750-0.300-0.540-1.210-0.510-0.939-0.673-0.573-0.887-0.646-0.463-0.612-0.478-0.520

-0.561-1.024-0.676-1.019-0.912-0.978-1.837-0.580-0.890-0.562-1.599-1.199-0.217-0.682-0.824-0.732-0.684-1.421-1.471-1.267-1.488-1.359-1.498-1.347-1.592-2.046-2.670-1.562-1.487

-0.839-0.931-0.772-0.572-0.394-0.653-0.245

-0.654-0.571-0.301-0.521-0.484-0.566-0.380-0.728-0.520-0.526-0.410-0.552-0.555-0.788-0.651-0.912-0.621

(%o PDB)

1.2300.9500.8400.7900.5040.4300.6200.6700.0300.2570.7440.4330.4330.5230.3770.6400.4900.6510.460

0.4990.6400.4520.5010.4560.299

-0.0790.2760.4640.3520.5170.6820.0340.5140.4870.5280.5080.6570.4710.5750.6540.6940.7530.0130.7820.4270.6200.4260.011

0.0390.0530.141

-0.1670.075

-0.377-0.219

0.019-0.161-0.225-0.189

0.2010.188

-0.016-0.078

0.0840.1170.185

-0.0210.068

-0.0370.079

-0.001-0.108

304

LATE PLEISTOCENE PLANKTONIC AND BENTHIC ISOTOPE RECORDS: SITE 724

Appendix A (continued).

Core,section

3H-33H-33H-33H-33H-33H-33H-33H-33H-43H-43H-43H-43H-43H-43H-43H-53H-53H-53H-53H-53H-53H-53H-53H-63H-63H-63H-63H-63H-63H-63H-73H-73H-73H-74H-4H-4H-4H-4H-4H-4H-4H-4H-24H-24H-24H-24H-24H-24H-24H-34H-34H-34H-34H-34H-34H-34H-34H-44H-44H-44H-44H-44H-44H-44H-54H-54H-54H-54H-54H-54H-54H-54H-64H-64H-f

Depthin section

(cm)

2-422-2442-4462-6482-84

102-104122-124142-144

12-1432-3452-5472-7492-94

112-114132-134

2-422-2442-4462-6482-84

102-104122-124142-144

12-1432-3452-5472-7492-94

112-114132-134

2-422-2442-4462-6410-1222-2442-4462-6482-84

102-104122-124142-144

12-1432-3452-5472-7492-94

112-114132-134

2-422-2442-4462-6482-84

102-104122-124142-144

12-1432-3452-5472-7492-94

112-114132-134

2-422-2442-4462-6482-84

102-104122-124142-144

12-1432-3452-54

Sub-bottomdepth

(m)

15.2215.4215.6215.8216.0216.2216.4216.6216.8217.0217.2217.4217.6217.8218.0218.2218.4218.6218.8219.0219.2219.4219.6219.8220.0220.2220.4220.6220.8221.0221.2221.4221.6221.8021.8221.9222.1222.3222.5222.7222.9223.1223.3223.5223.7223.9224.1224.3224.5224.7224.9225.1225.3225.5225.7225.9226.1226.3226.5226.7226.9227.1227.3227.5227.7227.9228.1228.3228.5228.7228.9229.1229.3229.5229.72

Age(1000 yrB.P.)

178.500180.000181.500183.000184.859186.718188.577190.437192.296194.155196.014197.873199.732201.592203.451205.310207.169209.028210.887212.746214.606216.465218.324220.183222.042223.901225.761227.620229.479231.338233.197235.056236.915238.589238.775239.704241.563243.423245.282247.141249.000254.429259.857265.286270.714276.143281.571287.000289.200291.400293.600295.800298.000300.200302.400304.600306.800309.000311.200313.400315.600317.800320.000322.200324.400326.600328.800331.000332.667334.333336.000337.667339.333341.000342.667

Uvigerina spp.

(‰ PDB)

2.3902.6002.5002.7002.6102.070

b2.280/1.9152.4601.603

b2.310/1.404b2.290/2.147b2.140/1.951bl.720/2.452bl.960/2.165b2.130/2.009b2.210/2.11

b2.57O/3.125b2.300/2.864b2.230/2.734

2.193b2.430/2.492

2.239b2.420/2.595

2.3901.7002.0202.430

b2.340/2.8382.1102.0002.1202.2502.1802.2602.5102.3902.5302.6702.5502.8002.9002.7802.4702.7102.4802.3402.2401.9202.3302.1702.4002.4802.5902.6102.500

2.3302.2202.0702.1602.2102.0501.8902.030

1.9401.6201.7602.3902.0902.0602.640

2.110

(‰ PDB)

-0.590-0.580-1.070-1.060-0.720-0.780

b -1.230/-1.232-1.070-1.609

b -1.440/-1.888b -1 .030/-1.461b -1 .310/-1.463b -0 .950/-1.013b - 1.070/ -0.793b—1.110/—1.112b -1 .040/-0.868b - 0.720/ -0.686b -0 .810/-0.561b -0 .820/-0.779

-0.805b -0 .920/-0.918

-0.704b -0.940/-0.857

-0.850-0.600-0.600-0.450

b -0 .540/-0 .493-0.550-0.420-0.240-0.630-0.580-0.420-0.500-0.710-0.600-0.630-0.790-0.890-0.850-0.870-0.780-0.690-0.600-1.130-1.340-1.260-0.960-0.740-0.560-0.500-0.440-0.510-0.610

-0.540-0.440-0.690-0.470-0.280-0.550-0.450-0.460

-0.340-0.230-0.530-0.580-0.340-0.810-0.940

-1.520

Cibicidoides spp.

δl8θ(‰ PDB)

1.9471.948

a1.813

a1.826a1.694

1.658a1.935

a1.8951.6301.518

a1.915

a1.258a1.495a1.366a1.955"2.030

1.530a2.304

1.688a1.537

a2.535

1.190a2.838a2.056a1.924a1.892a1.854

1.972a1.310a1.38Oa1.470a2.162a1.870a1.977

al.570/1.482a1.55Oa1.609

al.580/1.293

a2.241a1.648

b1.880/1.816a2.001

1.9401.8602.0402.020

a2.0702.090

a

a

1.430.580.680.620.600.750.570.370.750.234.610.000.260

a2.2001.4501.580

a1.5402.040

.470

δ l 3 C

(‰ PDB)

0.2150.188

a0.189

a0.204"0.1870.242

"0.185

"0.0950.2530.166

"0.141

a - 0 . 1 2 6^.066"0.646a0.217"0.0500.149

"0.5390.099

a -0 .048

a0.711

"-0.110"0.669"0.315"0.174"0.127a0.4100.226

"0.150"0.070^ . H O"0.407"0.163^.312

"0.010/0.081"-0.040

"0.159"0.000/-0.024

"0.182"0.131

b0.120/-0.040^ . H l0.3600.4000.2800.170

"0.1400.0600.0000.0700.2100.0400.1100.1800.120

-0.090a - 0 . 1 7 0

-0.150a - 0.060

-0.480-0.360

a - 0 . 1 5 0-0.160

0.180"0.240

-0.080-0.050

G. ruber (white)

δ ^ O(‰ PDB)

-0.487-0.313-0.586

-0.615-1.092-0.704-0.688-0.682

-0.997-0.722-0.733-0.742-0.632-0.838-0.799-0.696-1.052-1.207-0.869-1.168-0.816- ;_- ;__

1.3971.1191.8711.570L.059

-1.359-0.939-0.974-1.225-1.219-0.898-1.135-0.939-1.112-1.013-0.564-0.772-0.899

-0.763-0.752-0.816-0.944-0.711-0.756-0.316-0.461-0.657-0.852-0.970-0.779-0.612

-

.782

.265

.154

.163

.063-1.390-1.182-0.921-1.047-1.406-1.396-1.667-1.120-0.771-0.617-0.272-0.459-0.188-0.266

δ ^ C(‰ PDB)

0.0290.062

-0.146

0.101-0.703-0.190-0.174-0.368

-0.3120.0790.036

-0.828-0.328

0.3680.3020.242

-0.044-0.020

0.1610.1550.1730.4820.7220.6590.5400.3320.3600.3180.2880.4330.6590.4690.3610.1300.0750.5750.2080.2230.112

0.185-0.152-0.143-0.193-0.129

0.1710.2050.2460.1560.1890.3400.3470.4500.2310.5070.5370.4950.5920.6040.4780.8040.5820.5740.3770.1710.1510.080

-0.038-0.026-0.207-0.025-0.122

305

R. ZAHN, T. F. PEDERSEN

Appendix A (continued).

Core,section

4H-64H-64H-64H-74H-74H-74H-7

Depthin section

(cm)

72-7492-94

132-13418-2038-4058-6078-80

Sub-bottomdepth

(m)

29.9230.1230.5230.8831.0831.2831.48

Age(1000 yrB.P.)

344.333346.000349.333352.333354.000355.667357.333

Uvigerina spp.

δ 1 8 θ(‰ PDB)

2.7702.9702.810

(‰ PDB)

-0.680-0.800-0.860

Cibicidoides spp.

δ!8θ

(‰ PDB)

a2.480a2.5502.0602.430

δ»C(‰ PDB)

a - 0 . 2 9 0a - 0 . 3 2 0

-0.070-0.330

G. ruber (white)

δ!8θ(‰ PDB)

-0.270-0.477-0.408-0.433-0.472-0.495-1.141

(‰ PDB)

-0.0580.0830.1110.0060.141

-0.089-0.137

Value has been rejected (see text).Mean of replicates has been used.

APPENDIX BDensity and Organic Carbon Data for Hole 724C

Core,section

1H-1H-1H-1H-1H-1H-1H-11H-21H-21H-21H-21H-22H-2H-2H-2H-2H-2H-2H-2H-2H-22H-22H-22H-22H-22H-22H-22H-32H-32H-32H-32H-32H-32H-32H-32H-42H-42H-42H-42H-42H-42H-42H-52H-52H-52H-52H-52H-52H-52H-52H-62H-62H-62H-62H-62H-(

Depthin section

(cm)

25-2745-4765-6785-87

105-107125-127145-147

15-1735-3755-5775-7795-9710-1225-2745-4765-6785-87

105-107125-127145-147

15-1735-3755-5775-7795-97

115-117135-137

5-725-2745-4765-6785-87

105-107125-127145-147

15-1735-3755-5775-7795-97

115-117135-137

5-725-2745-4765-6785-87

105-107125-127145-147

15-1735-3755-5775-7795-97

115-117

Depthin hole

(m)

0.250.450.650.851.051.251.451.651.852.052.252.452.903.053.253.453.653.854.054.254.454.654.855.055.255.455.655.856.056.256.456.656.857.057.257.457.657.858.058.258.458.658.859.059.259.459.659.85

10.0510.2510.4510.6510.8511.0511.2511.45

Age(k.y.)

6.337.679.00

10.3311.6713.0014.3315.6717.0018.3319.6721.0024.0825.4927.4829.4731.4633.4535.4337.4239.4141.4043.3945.3847.3749.3651.3453.3355.3257.3159.3061.3063.3065.3067.3069.3071.8477.4182.9888.5594.1299.69

105.27110.90116.90122.18123.36124.54125.72126.90128.09129.27130.45131.63132.81134.00

^org(wt%)

1.7531.5450.8940.9091.1510.4360.5750.5270.6831.2341.0331.1800.7810.7681.1931.6020.6661.9773.8490.8291.0311.6701.6531.7211.0951.4371.1760.5380.7230.6540.5110.7760.6640.8571.5881.1071.5531.8071.1401.2691.6421.1971.0731.3261.1900.9150.6820.4840.5390.4400.6840.5910.6520.7930.6420.866

Wet bulkdensity

(g c m ~ 3 ) a

1.4731.5611.6801.7061.7371.8661.9051.8881.8491.7441.7251.7491.7151.8251.7751.7701.8051.7471.6981.7961.8151.7631.7491.7511.7921.7681.7961.816

:

:

1.8311.8291.8301.8351.862.869.822.781.752.724.774.779.756

1.7441.7221.7541.7741.8301.8601.8661.9291.9321.8911.8861.9131.9121.925

.898

Dry bulkdensity

( g c m - 3 ) b

0.6430.7830.9701.0131.0611.2651.3271.3011.2381.0721.0431.0811.026

(

:

1.2001.1211.1131.169.077

).9991.1551.1851.102.080

1.083.149.110.155.186.210.207.208.217.259

1.2711.1961.1311.0861.0411.1201.128

.091

.072

.038

.089

.119

.208

.256

.266

.365

.371

.305

.297

.3401.3381.3601.316

Sedimentationrates

(cm k.y. - ) c

15.0015.0015.0015.0015.0015.0015.0015.0015.0015.0015.0015.0015.0010.0910.0910.0910.0910.0910.0910.0910.0910.0910.0910.0910.0910.0910.0910.0910.0910.0910.0910.0010.0010.0010.0010.0010.003.533.533.533.533.533.533.533.533.53

11.7211.7211.7211.7211.7211.7211.7211.7211.7211.72

COrg( g c m ~ 2 k.y. ~ 1 )

16.90818.14613.00813.81218.3188.273

11.44510.28412.68319.84316.16119.13412.0209.299

13.49417.9917.856

21.48438.798

9.66112.32718.56918.01318.80612.69516.09413.7056.4388.8277.9656.2289.4448.360

10.89218.99212.52016.8666.6404.5075.0536.3244.5303.9325.0974.7013.902

10.0397.1818.6237.070

10.4628.984

10.24012.43510.23313.357

306

LATE PLEISTOCENE PLANKTONIC AND BENTHIC ISOTOPE RECORDS: SITE 724

Appendix B (continued).

Core,section

2H-62H-72H-72H-72H-73H-3H-3H-3H-3H-3H-3H-13H-23H-23H-23H-23H-23H-23H-23H-33H-33H-33H-33H-33H-33H-33H-33H-43H-43H-43H-43H-43H-43H-43H-53H-53H-53H-53H-53H-53H-53H-53H-63H-63H-63H-63H-63H-63H-63H-73H-73H-73H-74H-14H-14H-14H-]4H-14H-14H-14H-14H-24H-24H-24H-24H-24H-24H-24H-34H-34H-34H-34H-34H-34H-34H-34H-4

Depthin section

(cm)

135-1375-7

25-2745-4760-6225-2745-4765-6785-87

105-107125-127145-147

15-1735-3755-5775-7795-97

115-117135-137

5-725-2745-4765-6785-87

105-107125-127145-14715-1735-3755-5775-7795-97

115-117135-137

5-725-274-47

65-6785-87

105-107125-127145-14715-1735-3755-5775-7795-97

115-117135-137

5-725-2745-4765-6713-1525-2745-4765-6785-87

105-107125-127145-14715-1735-3755-5775-7795-97

115-117135-137

5-725-2745-4765-6785-87

105-107125-127145-14715-17

Depthin hole

(m)

11.6511.8512.0512.2512.4012.4512.6512.8513.0513.2513.4513.6513.8514.0514.2514.4514.6514.8515.0515.2515.4515.6515.8516.0516.2516.4516.6516.8517.0517.2517.4517.6517.8518.0518.2518.4518.6518.8519.0519.2519.4519.6519.8520.0520.2520.4520.6520.8521.0521.2521.4521.6521.8521.8321.9522.1522.3522.5522.7522.9523.1523.3523.5523.7523.9524.1524.3524.5524.7524.9525.1525.3525.8525.7525.9526.1526.35

Age(k.y.)

135.42138.18140.95143.72145.80146.49149.26152.03154.80157.57160.34163.11165.88168.65171.23172.73174.23175.73177.23178.73180.23181.73183.28185.14187.00188.86190.72192.58194.43196.29198.15200.01201.87203.73205.59207.45209.31211.17213.03214.89216.74218.60220.46222.32224.18226.04227.90229.76231.62233.48235.34237.19239.05238.87239.98241.84243.70245.56247.42249.81255.24260.67266.10271.53276.96282.39287.33289.53291.73293.93296.13298.33303.83302.73304.93307.13309.33

COrg(wt%)

1.4160.6970.8771.2520.8971.1881.0740.9080.8100.8730.7130.6810.7430.5020.4810.4390.4910.5010.6450.8210.9511.3050.8181.6500.7340.8620.6170.7491.7740.6840.6991.6610.8701.2320.9191.0240.9720.7240.5880.8561.1621.4081.5351.3821.7561.4511.0720.8360.7430.6221.3141.3300.6490.9060.6320.6931.0311.5321.4971.1521.1711.6710.8031.2680.9490.6850.5640.4790.7680.4100.3440.4791.0131.0031.1801.2301.286

Wet bulkdensity

(gcm~ 3 ) a

1.8351.8491.7011.4541.4541.6401.8251.8861.9201.8961.9341.9551.9401.9201.9561.9701.9761.9921.9771.9471.9021.8781.8701.8851.8251.8911.9531.9531.9371.8521.9011.9551.9441.8701.7651.8721.9031.8991.9091.8971.8961.8861.8211.7341.7451.7471.7331.8041.8561.9021.6181.4631.6341.6341.7971.8671.8741.8701.8481.8441.8671.8461.8711.8851.9101.9081.9051.9001.9141.9161.9151.8901.8601.8551.8631.8751.875

Dry bulkdensity

(g c m - 3 ) b

1.2171.238

1.004 d(l.238)0.612 d(l.238)0.612 d(1.238)0.908 d(l.238)1.200 d(l.238)

1.2971.3521.3131.3731.4061.3831.3511.4091.4311.4401.4651.4411.3941.3221.2841.2721.2961.2011.3051.4031.4041.3791.2441.3221.4061.3891.2711.1061.2751.3251.3171.3331.3161.3131.2981.1941.0571.0751.0761.0551.1681.2491.323

0.873 d(1.323)0.628 d(l.323)0.897 d(l.323)0.897 d(l.323)1.157 d(1.323)

1.2671.2791.271

e

1.2371.2311.2671.2341.2741.296L.3351.3321.327.320.341.344.344.304.256.248.261.281.281

Sedimentationrates

(cm k.y. ~ ' ) c

11.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7211.7210.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.6810.685.865.865.865.865.865.865.865.865.865.865.865.865.865.865.865.865.869.449.44

Corg(g cm""2 k .y . - 1 )

20.19710.113

10.320 d(12.725)8.980 d(18.166)6.434 d(13.O15)

12.642 d(17.237)15.105 d(15.583)

13.80212.83513.43411.47311.22212.0437.9497.9437.3638.2878.602

10.89313.41314.73519.63812.16525.06210.3329.2279.245

11.23126.1279.0889.869

24.94212.90616.72410.85513.94413.75510.1838.371

12.03116.29519.51919.57415.60120.16116.67412.07910.4289.9118.789

12.251 d(18.566)8.920 d(18.792)6.217 d(9.170)8.679 d(12.801)7.809 d(8.930)

9.37714.08320.79610.8518.3108.694

12.0835.9959.6307.4245.3474.3863.7056.0353.2292.7093.6607.4567.3358.720

14.87415.551

307

R. ZAHN, T. F. PEDERSEN

Appendix B (continued).

Core,section

4H-44H-44H-44H-44H-44H-44H-54H-54H-54H-54H-54H-54H-54H-54H-64H-64H-64H-64H-64H-64H-64H-74H-74H-7

Depthin section

(cm)

35-3755-5775-7795-97

115-117135-137

5-725-2745-4765-6785-87

105-107125-127145-147

15-1735-3755-5775-7795-97

115-117135-137

15-1735-3755-57

Depthin hole

(m)

26.5526.7526.9527.1527.3527.5527.7527.9528.1528.3528.5528.7528.9529.1529.3529.5529.7529.9530.1530.3530.5530.8531.0531.25

Age(k.y.)

311.53313.73315.93318.13320.33322.53324.73326.93329.13331.25332.92334.58336.25337.92339.58341.25342.92344.58346.25347.92349.58352.08353.75355.42

COrg(wt%)

1.2391.0491.1061.1161.5921.6411.9321.1270.7850.7500.4230.7840.3220.3750.3440.3360.4110.6590.8690.9060.8340.9150.6770.833

Wet bulkdensity

( g c m - 3 ) a

:

:

1.8751.8751.8751.8751.8751.7911.7981.821.880.904.920.932.950.981.975.976.986

1.9861.9741.9451.9461.9401.9221.911

Dry bulkdensity

( g c m - 3 ) b

e1.281eeee

1.281.281.281.281.147.157.195.288

1.3261.3511.3701.3991.4471.439.441.455.455.437.390.393.382.354

1.336

Sedimentationrates

(cm k.y. - ! ) c

9.449.449.449.449.449.449.449.449.449.449.449.449.449.449.44

12.0012.0012.0012.0012.0012.0012.0012.0012.00

C org( g e m " 2 k.y.~1)

14.98312.68513.37413.49519.25117.76821.10112.7139.5459.3885.395

10.1394.2535.1224.6735.8107.178

11.51014.98515.11213.94115.17411.00013.355

a GRAPE data.GRAPE data corrected to dry bulk density measured.

c Mean values calculated by linear interpolation between oxygen isotope boundaries.d Measured wet bulk density decreases sharply at core break; constant value from above the break has been used to calcu-

late organic carbon accumulation.e No GRAPE density data were available for Section 728C-4H-4; here the value from the bottom of Section 728C-4H-3 has

been used to calculate organic carbon accumulation.

308