Wise, S. W., Jr., Schlich, R., et al., 1992Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 120

45. MIOCENE STABLE ISOTOPE STRATIGRAPHY, SITE 747, KERGUELEN PLATEAU1

James D. Wright2 and Kenneth G. Miller3

ABSTRACT

We correlated Miocene δ 1 8θ increases at Ocean Drilling Program Site 747 with δ 1 8θ increases previouslyidentified at North Atlantic Deep Sea Drilling Project Sites 563 and 608. The δ 1 8θ increases have been directly tiedto the Geomagnetic Polarity Time Scale (GPTS) at Site 563 and 608, and thus our correlations at Site 747 providea second-order correlation to the GPTS. Comparison of the oxygen isotope record at Site 747 with records at Sites563 and 608 indicates that three as-yet-undescribed global Miocene δ 1 8θ increases may be recognized and used todefine stable isotope zones. The δ 1 8 θ maxima associated with the bases of Zones Mila, Milb, and Mi7 havemagnetochronologic age estimates of 21.8, 18.3, and 8.5 Ma, respectively.

The correlation of a δ 1 8 θ maximum at 70 mbsf at Site 747 to the base of Miocene isotope Zone Mi3 (13.6 Ma)provides a revised interpretation of four middle Miocene normal polarity intervals observed between 77 and 63 mbsfat Hole 747A. Oxygen isotope stratigraphy indicates that the reversed polarity interval at 70 mbsf, initiallyinterpreted as Chronozone C5AAr, should be C5ABr. Instead of a concatenated Chronozone C5AD-C5AC withdistinct Chronozones C5AB, C5AA, and C5A (as in the preliminary interpretation), δ 1 8 θ stratigraphy suggests thatthese normal polarity intervals are Chronozones C5AD, C5AC, and C5AB, whereas Chronozones C5AA-C5A areconcatenated. This interpretation is supported by the δ13C correlations.

The upper Miocene magnetostratigraphic record at Hole 747A is ambiguous. Two upper Miocene δ 1 8 θ events atSite 747 can be correlated to the oxygen isotope records at Site 563 and 608 using the magnetostratigraphy derivedat Hole 747B. Our chronostratigraphic revisions highlight the importance of stable isotope stratigraphy in attainingan integrated stratigraphic framework for the Miocene.

PRINCIPLES OF STABLE ISOTOPESTRATIGRAPHY

Variations in foraminifer δ 1 8θ records are used widely forPleistocene stratigraphic correlations and are potentiallyuseful for pre-Pleistocene correlations (Shackleton and Op-dyke, 1973; Imbrie et al., 1984; Ruddiman et al., 1986;Williams et al., 1988). Following the pioneering work ofEmiliani (1955), Shackleton and Opdyke (1973) designated22 late Pleistocene δ 1 8 θ cycles as "stages." (In view ofchanges in stratigraphic nomenclature [Hedberg, 1976;North American Commission on Stratigraphic Nomencla-ture, 1983], these cycles should now be termed chrons, notstages, when referencing time, or chronozones when refer-encing sections.) Hays et al. (1976) established that latePleistocene (0-700 Ka) δ 1 8θ fluctuations occurred with reg-ular astronomical (Milankovitch) periodicities. Knowing theastronomical forcing allowed Imbrie et al. (1984) to correlateby "tuning" δ 1 8 θ records to Milankovitch periodicities.Imbrie et al. (1984) produced the SPECMAP δ 1 8θ time scale,which is the standard for late Pleistocene marine correla-tions. This time scale provides correlations with relativeaccuracy of 5 k.y. (Imbrie et al., 1984).

Ruddiman et al. (1986), Williams et al. (1988), and Joyce etal. (1990) extended the late Pleistocene "stages" into the earlyPleistocene, but their numbering schemes are mutually exclu-sive (Williams et al., 1988, fig. 15). The possibility of usinghigh-resolution (I04—105 yr) oxygen isotope stratigraphy for

1 Wise, S. W., Jr., Schlich, R., et al., 1992. Proc. ODP, Sci. Results, 120:College Station, TX (Ocean Drilling Program).

2 Lamont-Doherty Geological Observatory and Department of GeologicalSciences, Columbia University, Palisades, NY 10964, U.S.A.

3 Department of Geological Sciences, Rutgers University, New Brunswick,NJ 08903, U.S.A., and Lamont-Doherty Geological Observatory, ColumbiaUniversity, Palisades, NY 10964, U.S.A.

pre-Pleistocene correlations was suggested by Ruddiman et al.(1986). Following this lead, Raymo et al. (1988) used high-resolution δ1 8θ stratigraphy to correlate upper Pliocene sec-tions from Deep Sea Drilling Project (DSDP) Sites 552 and 607in the North Atlantic with sections from Ocean DrillingProgram (ODP) Site 677 in the eastern equatorial Pacific.Keigwin et al. (1987) similarly noted that high-frequency δ1 8θchanges occurred at Site 552 in the late Miocene. Pisias et al.(1985) observed that high-frequency δ1 8θ fluctuations oc-curred in the middle Miocene in the equatorial Pacific atDSDP Site 574. If these high-frequency δ 1 8θ changes areglobal, then they have potential for high-resolution strati-graphic correlations.

It has proven difficult to develop a high-resolution (104-105yr) δ 1 8θ stratigraphic scheme for the pre-Pliocene because ofproblems in sampling resolution and stratigraphic correla-tions. In fact, first-order (m.y. scale) structures are still beingidentified in the Cenozoic δ 1 8θ record. Although the generalnature of the Cenozoic δ 1 8θ record was identified by Shack-leton and Kennett (1975) and Savin et al. (1975), there aremany stable isotope events on the m.y. scale that are only nowbeing recognized. For example, Miller and Fairbanks (1985)and Miller et al. (1989) identified a distinct δ 1 8θ increase nearthe Oligocene/Miocene boundary that was previously un-known because of poor sampling in this interval.

Recent studies have shown that stable isotope records canbe used to improve pre-Pliocene correlations on the m.y. orfiner scale. The key to establishing a stable isotope strati-graphic framework is a firm chronology in the referencesection. We have suggested that magnetostratigraphy pro-vides the best chronology for Miocene reference sections, andthat stable isotope records must be tied directly to the GPTS(Miller et al., 1985, 1989, 1991a, 1991b). Correlating δ1 8θrecords from different oceans to polarity patterns establishessynchrony, while it also allows absolute ages to be assigned toidentified events.

855

J. D. WRIGHT, K. G. MILLER

We have used this philosophy to establish formal oxygenisotope zones for the Oligocene-Miocene. Miller et al. (1991b)directly correlated seven Oligocene-Miocene δ 1 8θ increases tothe GPTS on the m.y. scale, and then used them to establishformal isotope zones. They also noted that more eventsshould be found in lower Miocene sections as more detailedrecords became available.

Knowing the causes of global δ1 8θ shifts is not a prerequi-site to their use in stratigraphic correlations, but knowing thecause(s) provides a predictive tool for the timing of fluctua-tions. For example, establishing the Milankovitch forcingmechanism for the late Pleistocene allows "tuning" (i.e.,adjusting the correlations of records and their numerical ages)to this known signal (Imbrie et al., 1984). Oligocene toHolocene δ 1 8θ changes on the m.y. scale may be causallyrelated to eustatic (global sea level) changes. Miller et al.(1991b) proposed that Oligocene to Miocene δ 1 8θ increasescorrelate with the second- and third-order sea-level sequenceboundaries of Haq et al. (1987). A causal relationship betweenoxygen isotope increases and sea-level sequence boundariescan facilitate deep-sea and shallow-marine correlations andprovide a predictive method to identify isotope events that arenot yet known.

Carbon isotopes have not been used as extensively as δ1 8θrecords for stratigraphic correlation for two reasons:

1. Deep-sea δ13C values are nonconservative and representa composite of several different parameters, including globalinventory changes related to the input and output ratios oforganic and inorganic carbon, water mass mixing, and deepwater aging (Miller and Fairbanks, 1985).

2. Carbon isotopes are used to interpret paleoceanographicchanges, making it potentially circular reasoning to correlatewith δ13C records, and then to interpret paleoceanographicchanges using the same records.

Despite these limitations, several Miocene δ13C fluctua-tions have been globally correlated and provide potential forstratigraphic correlations. The most widely known of these isthe late Miocene "carbon shift" (Haq et al., 1980), which hasbeen tied not only to the Amaurolithus datum (Haq et al.,1980), but also to Chron 6 (Keigwin, 1987). Vincent andBerger (1985) identified a global δ13C increase that spanned theearly/middle Miocene boundary; Miller and Fairbanks (1985)identified this increase as part of three global Oligocene-Miocene δ13C cycles.

At present, Miocene isotope correlations rely heavily onbiostratigraphy. Uncertainties in biostratigraphic correlationare typically 0.5-2.0 m.y. in the Miocene (Miller and Kent,1987), making it difficult to distinguish short-term events (i.e.,I05—106 yr duration). Problems arise in many biostratigraphiccorrelations between the deep-sea and coastal environmentsbecause of facies changes and the absence of marker taxa. Inaddition, comparison of low- and high-latitude isotope recordsby biostratigraphy alone is difficult because (1) low- andhigh-latitude biozonations usually are based on different spe-cies, and (2) zonal markers may be diachronous across lati-tude (e.g., Miller et al., 1985). Oxygen isotope stratigraphyprovides a means for supplementing biostratigraphy, particu-larly in correlating high-latitude sections such as KerguelenPlateau with standard low-latitude stratigraphic schemes.

Stable isotope stratigraphy is not a stratigraphic panacea.As with any stratigraphic scheme that relies on ordinal pat-tern-matching, stable isotope stratigraphy is based uponmatching an unknown record to a known or standard curve.As a result, δ 1 8θ time series of the unknown section must befairly complete and represent an interval of time with an

40°S

50'

60'

70c

CrozetBasin

Crozet Island

4 0 0 0 Kerguelen Island

• ODP Leg 119 Sites

®ODP Leg 120 Sites

Australian-Antarctic

Basin-^750751

ANTARCTICA -

Measured at 55°S

50° 60° 70- 80° 90°

Figure 1. Map showing the location of Site 747 on the KerguelenPlateau. Also shown are other Leg 120 sites as well as sites from Leg119.

identifiable event if they are to be stratigraphically useful.Stable isotope records are best used in combination withbiostratigraphy or some other stratigraphic framework that isused to establish broad age estimates. Once the general age isknown, δ1 8θ events can improve stratigraphic correlations tobetter than 100 k.y., depending on the sampling interval. Inthis manner, stable isotope stratigraphy works in much thesame way as magnetostratigraphy.

Postburial diagenesis can potentially alter the original δ 1 8θsignal in foraminifer tests, limiting the stratigraphic correla-tion by stable isotopes (e.g., Killingley, 1983). Diageneticchanges, however, are associated most commonly with burialdepths >400 m; sections with burial depths <400 m areusually unaltered (Miller and Curry, 1982; Barrera et al., 1987;Miller et al., 1987). Although the shallow burial depths of theMiocene section at Site 747 (<150 m) do not preclude diage-netic alteration, the lack of optical indications of diagenesisthrough calcite overgrowths indicates that the δ 1 8θ recordedat Site 747 was primary.

In this study, we sampled the Miocene section of Site 747as part of ongoing studies of Miocene stable isotope stratigra-phy and paleoceanography. During these investigations itbecame clear that the oxygen isotope stratigraphy of Site 747was important for the interpretation of the magnetostratigra-phy of this borehole. We present here the stable isotopestratigraphy of the Miocene of Hole 747A and discuss theramifications of this stratigraphy to the interpretation of themagnetostratigraphy.

LOCATION

Site 747 was drilled in the center of the Kerguelen Plateau(54°48.68'S, 76°47.64'E) at a water depth of 1695 m (Fig. 1).Site 747 is south of the modern Polar Front and within the flowof the Antarctic Circumpolar Current (ACC). A fairly com-

856

MIOCENE ISOTOPE STRATIGRAPHY

plete Miocene calcareous ooze section was recovered fromHole 747A in Cores 120-747A-4H to -14H, with an averagesedimentation rate for the Miocene of 5 m/m.y. (Schlich,Wise, et al., 1989). Leg 120 was one of the first drillingoperations in the Southern Ocean to retrieve a completemiddle Miocene section. Previously drilled Southern OceanDSDP Legs 28 and 35 and ODP Legs 113, 114, and 119encountered unconformities or siliceous middle Miocene sed-iments instead of calcareous sediments.

METHODS

Samples were washed with tap water and dried, andCibicidoides spp. were picked from the >250-µm size frac-tion. These benthic foraminifers were ultrasonically cleanedfor 5 to 10 s and then roasted at 370°C in a vacuum. Oxygenand carbon isotopes were generated at Lamont-Doherty Geo-logical Observatory by a Carousel-48 automatic carbonatepreparation device attached to a Finnegan MAT 251 massspectrometer. The standard deviations, based on 20 NBS-19measurements made during the analysis of stable isotope datafrom Site 747, were 0.05%o and 0.02%o for δ1 8θ and δ13C,respectively. Replicates of 22 samples indicate that samplevariability was O.IO‰ and 0.11‰ for δ 1 8θ and δ13C, respec-tively. These values may overstate the true sample variabilitybecause replicates were often chosen for samples that exhib-ited large isotopic offsets from adjacent sample values. Thedata are reported in values referenced to PDB (Table 1). TheGPTS of Berggren et al. (1985) was used throughout.

DATA

Oxygen IsotopesThe benthic foraminifer δ 1 8θ record from Site 747 shows

patterns similar to those recognized at other sites (Fig. 2).Distinct δ 1 8θ increases recorded in Miocene benthic foramin-ifer records can be identified. Oxygen isotope maxima occurat 127.9, 118.4, and 94.1 m below seafloor (mbsf) in the lowerMiocene (indicated with arrows in Fig. 2). Oxygen isotopevalues are fairly constant (~1.5%o) between 107 and 97 mbsf.A pronounced δ 1 8θ decrease from 1.75%o to l.O‰ was re-corded over the interval from 94 to 90 mbsf, spanning thelower/middle Miocene boundary. Oxygen isotope values in-creased between 85 and 84 mbsf, before reaching the lowestδ 1 8θ value in the Miocene (0.85%o at 81 mbsf). The middleMiocene section shows the classic "middle Miocene δ 1 8θincrease" (Shackleton and Kennett, 1975; Savin et al., 1981;Woodruff et al., 1981) in the interval from 79.0 to 62.9 mbsf(Fig. 2). This increase is interrupted from 70 to 67.5 mbsf by asmall decrease in δ 1 8θ values, which lends a steplike appear-ance to the classic middle Miocene δ 1 8θ increase. A small(0.5%o) δ 1 8θ increase was recorded in the upper middle Mio-cene section at a level of 56.9 mbsf. Upper Miocene δ 1 8θvalues (i.e., above 52 m) continue to increase, punctuated bya δ 1 8θ increase at 45.4 mbsf.

Carbon Isotopes

Carbon-isotope records show greater variability than theδ 1 8θ records (Fig. 2). Lower Miocene δ13C values record acycle with minimum values at 117.5 and -103 mbsf, andintervening maximum values of — 1.3%o at 111 mbsf. Carbon-isotope values climb in two steps from 103 mbsf to the highestMiocene δ13C values (1.9‰) at 83.4 mbsf. This δ13C increaserepresents the early/middle Miocene increase described byVincent and Berger (1985). Middle Miocene δ13C values werehighly variable, with excursions >0.5%e from 84 to 59 mbsf.Most of the values in this interval were >l.O‰ with maximumδ13C values of 1.5%o-1.8‰. Two distinct δ13C decreases,

however, were recorded from 70 to 65 mbsf and from 56.9 to52.5 mbsf (Fig. 2). Upper Miocene δ13C values were muchlower than those in the middle Miocene and averaged aboutl.O‰.

DISCUSSION

The goal of isotope stratigraphy is to improve inter-sitecorrelations. Higher resolution sampling of pre-Pleistocenesections continues to reveal previously unidentified m.y. andk.y. features. Interbasinal correlation of these events requiresprecise correlations. Magnetostratigraphy provides this preci-sion, although sites within a geographic region may be wellcorrelated with biostratigraphy alone. Unfortunately, notenough magnetostratigraphic records are available for mostinterbasinal correlations. As a result, correlations must de-fault to biostratigraphy with possibly large uncertainties (Mill-er and Kent, 1987).

Aliasing the High-frequency Signal

Pisias et al. (1985) noted that a high-frequency signal wasembedded in the benthic foraminifer δ 1 8θ record at Site 574(central equatorial Pacific) (Fig. 3). Biostratigraphic age mod-els indicate that the high-frequency component appeared tooccur in 40-k.y. cycles. Pisias et al. (1985) were careful not tolink the δ 1 8θ cycles changes to orbital forcing without betterage control and support from additional records. Regardless ofthe cause or frequency, undersampling the high frequencysignal may alias the record and can yield spurious conclusions(Pisias and Mix, 1988). This is important to consider since ourδ 1 8θ datums are part of a lower frequency cycle (—1-2 m.y.).

We think that aliasing has not compromised our isotopezonations for two reasons: (1) These low-frequency changes(—1 m.y.) have been documented in many records of differentsampling intervals (see Miller et al., 1991a, for a summary);and, (2) although high-frequency δ 1 8θ changes are superim-posed on the long-term record (Pisias et al., 1985), we can stillidentify the lower frequency cycles (Fig. 3). Aliasing thehigh-frequency signal can alter the amplitude of low-fre-quency cycles. However, the timing of the m.y. cycles can beidentified because the high-frequency amplitude is —0.3%o incomparison to m.y. changes of 0.7%o to l.O‰. However, insome cases, aliasing can obscure the m.y. events (e.g., Mi5),making them difficult to recognize because the amplitudes aresmall (0.5%^0.6%o) (Fig. 3).

The Million-Year Signals

We have attempted to establish that m.y.-scale variationsin the deep-water δ 1 8θ and δ13C records were global bycomparing stable isotope and polarity correlations at Site 747with those at North Atlantic Sites 563 and 608 (Miller et al.,1991b). We can independently verify the paleomagnetic inter-pretation at Site 747 because we are correlating δ 1 8θ and δ13Crecords to polarity patterns at more than one site. By corre-lating isotope events to recognizable polarity patterns atdifferent locations, we establish the synchrony of theseevents. Uncertainties in the identification of magnetochrons atany site can be clarified by integrating the stable isotope andpolarity records from all three sites (see below).

Initial age estimates for the δ1 8θ and δ13C records at Site 747were obtained by using the polarity interpretations of Heider etal. (this volume) (Fig. 4). The Miocene δ1 8θ record at Site 747was correlated to standard δ1 8θ records at North Atlantic Sites563 and 608 (Miller et al., 1991a) (Table 2). These North Atlanticlocations provide the requisite standards for Miocene isotopezones (Zones Mi2 through Mi6), because they constitute theprimary source of first-order correlations of early and middleMiocene δ1 8θ events with the GPTS (Miocene isotope Zone Mil

857

J. D. WRIGHT, K. G. MILLER

Table 1. Oxygen and carbon isotope data for Cibicidoidesspp., Hole 747.

Table 1 (continued).

Core, section,interval (cm)

120-747A-

4H-4, 100-1044H-4, 100-1044H-5, 100-1044H-5, 100-1044H-6, 100-1044H-7, 10-144H-CC4H-CC5H-3, 40-445H-3, 105-1095H-4, 40-445H-4, 105-1095H-5, 40-445H-5, 105-1095H-6, 40-445H-6, 105-1095H-7, 10-145H-CC5H-CC6H-1, 40-446H-1, 105-1096H-2, 40-446H-2, 105-1096H-3, 40-446H-3, 105-1096H-4, 40-446H-4, 105-1096H-4, 105-1096H-5, 40-446H-5, 105-1096H-6, 40-446H-6, 105-1096H-7, 10-146H-CC7H-1, 40-447H-1, 102-1067H-2, 40-447H-2, 104-1087H-3, 40-447H-3, 100-1047H-4, 40-447H-4, 101-1057H-5, 40-447H-5, 100-1047H-6, 40-447H-6, 100-1047H-7, 40-447H-7, 40-447H-CC8H-1, 40-448H-1, 100-1048H-2, 40-448H-2, 100-1048H-3, 40-448H-3, 100-1048H-3, 100-1048H-4, 40-448H-4, 100-1048H-4, 100-1048H-5, 100-1048H-5, 100-1048H-6, 40-448H-6, 100-1048H-7, 12-169H-2, 40-449H-2, 100-1049H-3, 40-449H-3, 100-1049H-4, 40-449H-4, 101-1059H-4, 101-1059H-5, 40-449H-5, 100-1049H-6, 40-449H-6, 98-102

Depth(mbsf)

33.5033.5035.0035.0036.5037.1037.4037.4040.9041.5542.4043.0543.9044.5545.4046.0546.6046.9046.9047.4048.0548.9049.5550.4051.0551.9052.5552.5553.4054.0554.9055.5556.1056.4056.9057.5258.4059.0459.9060.5061.4062.0162.9063.5064.4065.0065.6065.6065.9066.4067.0067.9068.5069.4070.0070.0070.9071.5071.5073.0073.0073.9074.5075.1275.9076.5077.4078.0078.90 (79.5179.5180.4081.00 (81.90 182.48 1

δ 1 8 θ

2.2962.6982.4272.5252.4352.3292.5112.4502.5092.2102.5072.3562.3962.3722.5322.4082.4232.0802.2212.3102.1982.3382.1912.2092.2102.2071.9892.1542.2492.0812.3622.1842.1352.2092.5532.0392.2972.4902.4492.1062.2282.0642.1982.1362.3541.9582.0792.0012.218L.9291.6201.7321.7271.742.733.792.503

1.4021.5491.466.692.477.286.166.182.110.132.027

).929.301.437.117

).846.015.201

« 1 3 C

1.0310.6491.3441.2361.0221.1761.2841.2991.0350.9591.1060.8901.0020.8721.0581.0741.0631.0311.0551.0770.9131.0890.9890.8251.0040.9110.5440.6750.7481.1461.4731.3761.1310.9861.4351.4541.3261.4531.3971.1951.0930.9391.1380.9300.9450.8421.2851.1261.0421.0811.0091.1961.2581.4951.7311.8271.2311.1181.1621.5021.4981.5891.6231.3761.5761.4781.5741.5241.3571.6981.7091.4821.0961.4981.674

Core, section,interval (cm)

9H-7, 40-449H-7, 100-1049H-7, 100-1049H-8, 10-149H-8, 10-149H-8, 40-449H-CC9H-CC

10H-1, 40-4410H-1, 100-10410H-2, 40-4410H-2, 100-10410H-3, 40-4410H-3, 100-10410H-4, 40-4410H-4, 100-10410H-5, 100-10410H-6,40-4410H-6, 100-10410H-6, 100-10410H-7, 10-1410H-7, 40-4411H-1,40-4411H-1, 100-10411H-2, 40-4411H-2, 102-10611H-3, 40-4411H-3, 100-10411H-4, 40-4411H-4, 100-10411H-5, 40-4411H-5, 101-10411H-6, 40-4411H-6, 101-10411H-7, 10-1411H-7, 40-4412H-1, 40-4412H-1, 100-10412H-2, 40-4412H-2, 100-10412H-3, 40-4412H-3, 100-10412H-4, 40-4412H-4, 100-10412H-4, 100-10412H-5, 40-4412H-5, 100-10412H-5, 100-10412H-6, 40-4412H-6, 100-10412H-6, 100-10412H-7, 10-1412H-7, 10-1412H-CC13H-1, 40-4413H-1, 101-10513H-2, 40-4413H-2, 101-10513H-3, 40-4413H-3, 101-10513H-4, 40-4413H-4, 101-10513H-4, 101-10513H-5, 40-4413H-5, 100-10413H-5, 100-10413H-6, 40-4413H-6,100-10413H-7, 10-1413H-7, 40-4414H-1, 40-4414H-1, 102-10614H-2, 40-4414H-2, 102-10614H-3, 40-4414H-3, 102-10614H-3, 102-10614H-4, 40-44

Depth(mbsf)

83.4084.0084.0084.6084.6084.7584.9084.9085.4086.0086.9087.5088.4089.0089.9090.5092.0092.9093.5093.5094.1094.4094.9095.5096.4097.0297.9098.5099.40

100.00100.90101.51102.90103.01103.60103.90104.40105.00105.90106.50107.40108.00108.90109.50109.50110.40111.00111.00111.90112.50112.50113.10113.10113.40113.90114.51115.40116.01116.90117.51118.40119.01119.01119.90120.50120.50121.40122.00122.60122.90123.40124.02124.90125.52126.40127.02127.02127.90

δ 1 8 θ

1.2621.4791.3501.1181.1491.0561.4621.3921.0530.9721.1101.0981.1591.1151.1291.0871.5851.6031.3591.5151.7131.4361.2591.5751.6491.3201.4281.4641.4311.3161.3911.3881.6411.3711.3291.4341.4471.1991.4831.5001.4861.535

s 1.3751.3931.3351.6731.6321.5661.5991.6101.6111.4211.5621.5141.6371.3521.4101.5501.5341.6391.7351.5191.4741.7671.5171.5181.4171.3801.2921.5841.6331.6351.7411.6421.8421.8251.4781.882

δ 1 3 C

1.8901.8071.6621.6461.3171.6331.8011.7121.5221.3421.3231.4861.2741.4051.2161.3001.3141.3391.4161.3331.3821.3331.0301.1861.1030.8871.0370.8460.9150.7950.7990.8840.8300.7750.8160.9550.9470.8440.9220.9771.0961.0180.8910.8661.0441.1071.2970.9311.2900.9200.9271.2910.9671.1141.1450.7720.7680.8640.8940.6851.1130.8300.8491.0231.0340.8890.8100.7870.7320.8291.1861.3771.1861.2001.0921.0041.0401.432

858

MIOCENE ISOTOPE STRATIGRAPHY

Table 1 (continued).

Core, section,interval (cm)

14H-4, 102-10614H-4, 102-10614H-5, 40-4414H-5, 101-10514H-5, 101-10514H-6, 40-4414H-6, 101-10514H-7, 12-1614H-7, 40-4415H-1, 40-4415H-1, 100-10415H-2, 40-4415H-2, 100-10415H-3, 40-4415H-4, 40-4415H-5, 40-4415H-6, 40-4415H-7, 40-44

Depth(mbsf)

128.52128.52 1129.40130.01130.01 1130.90 1131.51 1132.12132.40132.90 1133.50134.40135.00135.90137.40138.90140.40141.90

518O

.453

.558

.687

.513

.527

.530

.627

51 3C

.176

.192

.063

.175

.253

.213

.239.430 0.865.331 0.773.295 0.963.377 0.963.358.388.227

.106

.116

.183.389 0.911.2% 0.923.650 0.912.449 0.873

is based upon correlations to South Atlantic Site 522; Miller etal., 1991b). Isotope Zones Mi2-Mi6 have been documented inother oceans, but have been directly tied to the GPTS only atSites 563 and 608. We provide the depth at which each corre-sponding δ1 8θ maximum occurred at Site 747 (Table 2); themaximum was chosen because the isotope zonal boundaries aredefined upon the maximum values (Miller et al., 1991b). Agesbased upon correlation with Sites 563 and 608 were comparedwith initial age estimates established from the initial magneto-stratigraphic age estimates of Heider et al. (this volume) (Table 2and Fig. 4).

Implications of Oxygen Isotope Stratigraphy

We have identified Miocene isotope Zones Mil throughMi6 at Site 747 following the framework by Miller et al.(1991b) (Fig. 5). At Site 747, δ 1 8θ maxima at 127.9, 84.9, and56.9 mbsf are correlated with the Mil, Mi2, and Mi5 maximarecorded at Sites 563 and 608. The initial age estimates of theMil, Mi2, and Mi5 events at Site 747 are 23.7, 16.1, and 11.3Ma, respectively, and they correspond well with age estimatesfrom Sites 563 and 608 (Table 2 and Fig. 4). This supports thecorrelation of magnetozones at Site 747 to the GPTS: ChronsC6C, C5C, C5B, and C5r, as determined by Heider et al. (thisvolume).

There were additional lower Miocene δ 1 8θ maxima identi-fied at Site 747 at 118.4, 112.5, and 94.1 mbsf (Fig. 2). Two ofthese maxima were also noted at Sites 563 and 608, but theywere not interpreted as global events by Miller et al. (1991b)because of ambiguous polarity records in this interval. Forexample, the original interpretation of the polarity record atSite 563 by Miller et al. (1985) placed an unconformity at alevel of 270 mbsf, separating Chronozones C6n and C5Dr. Theisotope data, however, indicate that the polarity change at 270mbsf at Site 563 is the conformable boundary between Chro-nozones C5En and C5Dr. Stable isotope evidence indicatesthat the original identification of Magnetochronozones C6Cthrough C6n at Site 608 by Clement and Robinson (1987) wasincorrect (see below).

At Site 747, δ 1 8θ maxima at 118.1 and 94.1 mbsf have ageestimates of 21.8 and 18.1 Ma, respectively, based on corre-lations with Chronozones C6Ar and C5Dr (Fig. 2). At Site 563,apparently coeval δ 1 8θ maxima occur in the upper part ofChronozone C6Ar at a level of —283 mbsf and in ChronozoneC5Dr at a level of -268 mbsf (Fig. 5). The age estimates ofthese δ 1 8θ maxima from Site 563 are 21.8 and 18.3 Ma,respectively. Using the revised interpretation of the polarity

record at Site 608 (this study; see also Miller et al., 1991a),δ1 8θ maxima at 379 mbsf and 349 mbsf correspond to Chro-nozones C6Ar and C5Dr and have age estimates of 21.8 and18.1 Ma, respectively. Based on the documentation of in-creases in benthic oxygen isotope records at Sites 563, 608,and 747, we formally recognize the maximum δ 1 8θ valuesassociated with these increases as the base of two zones: Milaand Milb (Appendix).

In the middle Miocene at Site 747, a sequence of fourdistinct normal polarity intervals from 77 to 62 mbsf wasinterpreted as Chronozones C5AC/C5AD concatenated (77-72mbsf), C5AB, C5AA, and C5A (Heider, et al., this volume).Identifying the δ 1 8θ maximum at 70.0 mbsf at Site 747 ascorrelative with the δ 1 8θ increase at the base of Zone Mi3(13.6 Ma, Table 2 and Fig. 5) yields a different interpretation.At Site 608, the maximum δ 1 8θ values associated with thebase of Zone Mi3 clearly occur within Chronozone C5ABr(Fig. 5) at an estimated age of 13.6 Ma (Table 2). If the level ofZone Mi3 is 70.0 mbsf at Site 747, then the reversed polarityinterval at this level must be Chronozone C5ABr and notC5AAr, as interpreted by Heider et al. (this volume). Usingthis oxygen isotope correlation, Chronozones C5AD, C5AC,and C5AB are distinguished, whereas Chronozones C5AAand C5A are concatenated and are represented by one normalpolarity interval between 65 and 61 mbsf. This interpretationis strongly supported by the identification of the δ 1 8θ increaseassociated with Zone Mi4 (12.6 Ma) at 64.4 mbsf at Site 747.At Sites 563 and 608, the Mi4 δ 1 8θ increase occurs withinChronozone C5Ar at an estimated age of 12.6 Ma (Fig. 4). AtSite 747, however, the δ 1 8θ increase falls within a normalinterval that must represent the concatenation of Chrono-zones C5AA and C5A (Fig. 2).

Our interpretations of Chronozones C5-C5A are similar tothe reinterpretation of Harwood based on diatoms (pers,comm., 1990). Harwood similarly identifies Chronozones C5Bthrough C5AB from Cores 9 through 8, and places a hiatusduring Chron C5AA. Although we cannot confirm a hiatus forthis chron, it is consistent with our suggestion of the concat-enation of C5AA and C5A.

The initial age estimates of the upper Miocene δ 1 8θ in-creases at Site 747 do not agree with the ages estimated bycorrelation of these increases to Sites 563 and 608. A δ 1 8θincrease with maximum values at 9.6 Ma was defined as thebase of Zone Mi6 (Miller et al., 1991b); we note another δ1 8θincrease at 8.5 Ma and use this to recognize Zone Mi7informally (Appendix). This increase was discussed by Milleret al. (1991b) in records from North Atlantic Sites 563 and 608and by Wright et al. (1991) for Southern Ocean Site 704. AtSite 747, δl8O maxima occurred at 9.0 and 8.0 Ma accordingto the preliminary paleomagnetic age model, whereas Sites563 and 608 recorded maxima at 9.6 and 8.5 Ma (Table 2). Oneexplanation for the discrepancy is that the long normal ofChronozone C5 is not fully represented at Hole 747A and thatthe interpretation of Chronozone C4An at 42 mbsf is incor-rect. The magnetostratigraphy of this interval appears to bemore complete at Hole 747B, and the upper Miocene polaritypatterns from Hole 747B better resemble the GPTS than thosefrom Hole 747A (Heider et al., this volume). Applying themagnetochronology of Hole 747B to the oxygen isotoperecord at Hole 747A, the Mi6 and Mi7 events have estimatedages of 9.3 and 8.4 Ma. These age estimates are in closeragreement with the magnetochronologic ages derived fromSites 563 and 608 (9.6 and 8.5 Ma, respectively). It implies thatthe magnetochronology of the upper Miocene of Hole 747B iscorrect, and that there is a small offset (< 1 m or so) betweenHoles 747A and 747B.

859

2 0 O 0 U^ o

> ~ °

I I I 1 2 . 5 2.0 1.5 1.0 0.5 0 0.5 1.0 1.5 2.03 0 ~• ^ ^ C3 ' * ' * ' ' ' ' ' ' * ' * ' ' ' ' r

• 4 i J O•^ rip^—O _• β.— ^ ^

I ^ ^ ^ C6AA —Y-y^yti ^ , ^ Ad^H| »

H I I C6B r>SK) ^Λ--^~~^"*

. ^ I . I . I . I I I I , I I I I I ,

Figure 2. Oxygen and carbon isotope values from Hole 747A plotted vs. depth. Arrows indicate δ 1 8θ and δ13C levels discussed in the text. Themagnetic polarity column and the chronozones are reinterpreted from Heider et al. (this volume).

<—i

p

axH

p

E

MIOCENE ISOTOPE STRATIGRAPHY

2.5

3.0

Site 574

80 100 120 140 160 180 200 220

Depth (mbsf)Figure 3. Oxygen isotope record vs. depth from Site 574 (central equatorial Pacific). The isotope data are from Pisiaset al. (1985), Shackleton (unpubl. in Woodruff and Savin, 1989), and Woodruff and Savin (in press). The "Mi"maxima are noted on the figure as well as on the planktonic foraminifer FADs and LADs. Age estimates for thebiostratigraphic datums are from Berggren et al. (1985).

We revised the initial age estimates by using our oxygenisotope correlations to reinterpret the magnetochronology asdiscussed above. We then used this revised magnetochronol-ogy to estimate revised ages. The revised chronology showsonly minor (0.5 m.y. or so) differences from the initial ageestimates (Fig. 4). Still, these differences are significant in adetailed interpretation of the stratigraphic correlations of Site747.

Implications of Carbon-Isotope Records

Although no formal stratigraphic scheme has been basedon δ13C variations, many studies recognized the global natureof some carbon isotope fluctuations. For example, Vincentand Berger (1985) described a late early Miocene global δ13Cincrease. We established the age of this increase by tying itinto the GPTS at Sites 563 and 608 (Fig. 6); from this, an ageestimate of 19-16 Ma can be assigned to the increase, basedon its position within Chronozones C5E-C5C (top). The δ13Cincrease at Site 747 begins at 103 mbsf in an interval correlatedwith Chronozone C5En (age estimate = 18.9 Ma) and ends at84 mbsf (Chronozone C5Br, age estimate = 16.1 Ma) (Figs. 2and 5). This excellent agreement documents that this δ13Cincrease has not been locally overprinted at Site 747.

A distinct middle Miocene δ13C peak can be identified andcorrelated to the GPTS at Sites 563 and 608. This peakprecedes a sharp δ13C decrease and occurs within ChronozoneC5ABr at Site 608, but it falls within a concatenated Chrono-zone C5AI‰C5AB interval at Site 563. At Site 747, the peak

δ13C value is recorded at 70 mbsf. The preliminary magne-tochronology indicates that this is Chronozone C5AAr and notC5ABr (Heider et al., this volume). Both oxygen and carbonisotope interpretations, however, indicate that the level of 70mbsf is best correlated to Chronozone C5ABr and not C5AAr.This implies that the normal polarity intervals from 77 to 63mbsf should be interpreted as Chronozones C5AD, C5AC,and C5AB, with the last normal representing a concatenationof Chronozones C5AA-C5A.

Another δ13C maximum occurred near the top of the middleMiocene within Chronozone C5r (Fig. 6). Age estimates forthis maximum are 11.0 at Sites 563 and 608. At Site 747, a δ13Cincrease near the end of the middle Miocene peaked at 56.4mbsf within a reversed polarity interval recognized as Chro-nozone C5r. Both the δ13C and δl8O isotope correlations(Mi5) firmly establish that the position of Chronozone C5r atSite 747 is as in Heider et al. (this volume).

Reinterpretation of Previous MagnetostratigraphicRecords

The initial interpretation of the lower Miocene polarityrecord at Site 608 was complicated by the uncertain identifi-cation of Magnetochrons C6A-C6C. A large δ 1 8θ increase at411 mbsf at Site 608 was identified as Mil (Miller et al., 1991a),and Clement and Robinson (1987) identified this interval ofnormal polarity as Chron C6B. The position of Zone Mil,however, establishes that this interval must represent Chro-nozone C6C. The δ13C records confirm the misidentification

861

18

CD

<

δ l ö 0

α. ü3.0 2.0

24

26

isotopetime scale

13δ 1 J C isotopetime scale

1.0 0.0 0.0

3Tiδx

o

3.0 2.0 1.0 0.0 0.0 0.5 1.0 1.5 2.0

1 8

magnetictime scale

O13 A0 V-/ magnetic

time scale

MIOCENE ISOTOPE STRATIGRAPHY

Table 2. δ 1 8 θ event stratigraphy and associated ages.

Event

MilbMileMi2Mi3Mi4Mi5Mi6Mi7

GPTScorrelation

C6nC5DnC5BrC5ABrC5AnC5rC5nbase C4n

Estimatedagea(Ma)

(21.7)(18.1)16.113.612.611.39.6

(8.5)

Depthat

Site 747

117.5094.1084.6070.0063.5056.0446.6037.40

Initialage

estimate

21.417.616.113.212.111.19.08.0

Note: GPTS = Geomagnetic Polarity Time Scale from Berggren etal. (1985). Zone Mi7 is not formally recognized as an isotopechronozone, but it is useful in upper Miocene correlations(Wright et al., 1991).

aEstimated ages are derived by δ 1 8 θ correlation of Sites 563 and608 stable isotope events to the GPTS (Miller et al., 1991a)except for the ages indicated in parentheses, which are fromFigure 2 (this study).

by establishing the position of the C6AA/C6B boundary. Thelower Miocene polarity patterns are clear at Sites 563 and 747.Carbon isotope values decreased across the C6AA/C6Bboundary in both records. At Site 608, a δ13C decrease in theinterval from 395 to 385 mbsf straddles a polarity change at390 mbsf. Thus, this change must be the C6AA/C6B boundaryinstead of the C6A/C6AA boundary, as interpreted initially byClement and Robinson (1986). This reinterpretation of thelower Miocene polarity record is consistent with the 5180 andδ13C records and it does not violate the biostratigraphicdatums at Site 608. We use the revised correlations of the Site608 record in this study and that of Miller et al. (1991a).

CONCLUSIONS

Oxygen isotope stratigraphy provides a correlation toolindependent of biostratigraphy and magnetostratigraphy. Likeany ordinal pattern matching technique, stable isotope stratig-raphy requires (1) that limits be placed upon the general timeinterval (e.g., early Miocene), and (2) that the time series besufficiently long and complete to recognize global patterns.We used our oxygen isotope stratigraphy to evaluate thepublished preliminary magnetostratigraphy of Hole 747A. Therecords at Site 747A were correlated to the δ1 8θ and δ13Crecords at North Atlantic Sites 563 and 608, which have goodmagnetostratigraphic control. Hence, we were able to evalu-ate correlations of Hole 747A to the GPTS in two ways: direct(using the preliminary paleomagnetic record) and indirect (bycorrelation through Sites 563 and 608). This allowed us tocheck the magnetostratigraphic interpretations at Site 747.

The complete lower Miocene section at Site 747 providesevidence for two early Miocene δ 1 8θ isotope increases thatwere not formally recognized; the maxima associated withthese increases are used here to recognize formally the base oftwo new stable isotope zones (Mila and Milb).

Middle Miocene stable isotope correlations indicate thatthe preliminary interpretation of Chronozones C5A throughC5AD at Hole 747A should be reexamined. A distinct two-tiered step in the "middle Miocene increase" has been iden-tified at Site 747 and in other complete middle Miocene δ 1 8θrecords (e.g., Sites 563 and 608). The base of one of thesesteps (= base of Zone Mi3, 13.6 Ma) occurred at 70 mbsf atHole 747A, implying that Chronozone C5ABr, not C5AAr, isat a level of 70 mbsf. This interpretation of the normal polarityevents is supported by the presence of a δ13C maximum withinChronozone C5ABr. The position of the base of Chronozone

C5 at Hole 747A is well supported by both δl8O and δ13Crecords.

We were able to reinterpret the lower Miocene polarityrecord at Site 608 by integrating the δ 1 8θ, δ13C, and polarityrecords from Sites 563, 608, and 747. Both Sites 563 and 747have clear lower Miocene polarity patterns. The identifica-tion of Zone Mil at Site 608 is linked to Chron C6C, whereasa distinct δ13C decrease correlates with the C6AA/C6Bboundary.

ACKNOWLEDGMENTS

We thank E. Thomas, D. Williams, W. Berggren, R. Poore,W. Ruddiman, and D. Kent for reviews, and the shipboardparty of Leg 120 for samples from Site 747 for stable isotopeanalysis. Samples were provided by ODP. This study is part ofthe Ph.D. thesis of J. D. Wright at Lamont-Doherty Geolog-ical Observatory and the Department of Geological Sciencesat Columbia University. Research was supported by NationalScience Foundation Grant No. OCE-88-11834 to K. G. Miller.This is Lamont-Doherty Geological Observatory ContributionNo. 4793.

REFERENCES

Barrera, E., Huber, B. T., Savin, S. M., and Webb, P.-N., 1987.Antarctic marine temperatures: late Campanian through earlyPaleocene. Paleoceanography, 2:21-47.

Berggren, W. A., Kent, D. V., and Van Couvering, J. A., 1985. TheNeogene: Part 2. Neogene geochronology and chronostratigraphy.In Snelling, N. J. (Ed.), The Chronology of the Geological Record.Geol. Soc. London Mem., 10:211-260.

Clement, B. M., and Robinson, F., 1987. The magnetostratigraphy ofLeg 94 sediments. In Ruddiman, W. F., Kidd, R. B., Thomas, E.,et al., Init. Repts. DSDP, 94, Pt. 2: Washington (U.S. Govt.Printing Office), 635-650.

Emiliani, C , 1955. Pleistocene temperatures. J. Geol, 63:539-578.Haq, B. U., Hardenbol, J., and Vail, P. R., 1987. Chronology of

fluctuating sea levels since the Triassic. Science, 235:1156-1167.Haq, B. U., Worsley, T. R., Burckle, L. H., Douglas, R. G., Keigwin,

L. D., Opdyke, N. D., Savin, S. M., Sommer, M. A., Vincent, E.,and Woodruff, F., 1980. Late Miocene marine carbon-isotopicshift and synchroneity of some phytoplanktonic biostratigraphicevents. Geology, 8:427-431.

Hays, J. D., Imbrie, J., and Shackleton, N. J., 1976. Variations in theEarth's orbit: pacemaker of the ice ages. Science, 194:1121-1132.

Hedberg, H. D. (Ed.), 1976. International Stratigraphic Guide: NewYork (Wiley-Interscience).

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 arevised chronology of the marine delta δ 1 8 θ record. In Berger, A.,Imbrie, J., Hays, J., Kukla, G., and Saltzman, B. (Eds.), Milank-ovitch and Climate (Pt. 1): Dordrecht (D. Reidel), 269-305.

Joyce, J. E., Tjalsma, L.R.C., and Prutzman, J. M., 1990. High-resolution planktic stable isotope record and spectral analysis forthe last 5.35 m.y.: Ocean Drilling Program Site 625, northeast Gulfof Mexico. Paleoceanography, 5:507-529.

Keigwin, L. D., 1987. Toward a high-resolution chronology for latestMiocene paleoceanographic events. Paleoceanography, 2:639-660.

Keigwin, L. D., Aubry, M.-P., and Kent, D. V., 1987. North Atlanticlate Miocene stable-isotope stratigraphy, biostratigraphy, andmagnetostratigraphy. In Ruddiman, W. F., Kidd, R. B., Thomas,E., et al., Init. Repts. DSDP, 94, Pt. 2: Washington (U.S. Govt.Printing Office), 935-963.

Kennett, J. P., 1986. Miocene to early Pliocene oxygen and carbonisotope stratigraphy of the Southwest Pacific, DSDP Leg 90. InKennett, J. P., von der Borch, C. C , et al., Init. Repts. DSDP, 90:Washington (U.S. Govt. Printing Office), 1383-1411.

Killingley, J. S., 1983. Effects of diagenetic recrystallization of1 8 1 6 values of deep-sea sediments. Nature, 301:594-597.

863

J. D. WRIGHT, K. G. MILLER

Kroopnick, P., 1985. The distribution of 13C of Σ C O 2 in the worldoceans. Deep-Sea Res. Part A, 32:57-84.

Miller, K. G., Aubry, M.-P., Khan, K. J., Melillo, A. J., Kent, D. V.,and Berggren, W. A., 1985. Oligocene-Miocene biostratigraphy,magnetostratigraphy and isotopic stratigraphy of the westernNorth Atlantic. Geology, 13:257-261.

Miller, K. G., and Curry, W. B., 1982. Eocene to Oligocene benthicforaminiferal isotopic record in the Bay of Biscay. Nature,296:347-350.

Miller, K. G., and Fairbanks, R. G., 1985. Oligocene to Miocenecarbon isotope cycles and abyssal circulation changes. InSundquist, E. J., and Broecker, W. S. (Eds.), The Carbon Cycleand Atmospheric CO2: Natural Variations Archean to Present.Am. Geophys. Union Geophys. Monogr., 32:469-486.

Miller, K. G., Fairbanks, R. G., and Mountain, G. S., 1987. Tertiaryoxygen isotope synthesis, sea-level history, and continental mar-gin erosion. Paleoceanography, 2:1-19.

Miller, K. G., Feigenson, M. D., Wright J. D., and Clement, B. M.,1991a. Miocene isotope reference section, Deep Sea DrillingProject Site 608: an evaluation of isotope and biostratigraphicresolution. Paleoceanography, 6:33-52.

Miller, K. G., and Kent, D. V., 1987. Testing Cenozoic eustaticchanges: the critical role of stratigraphic resolution. CushmanFound. Foraminiferal Res. Spec. Publ., 24:51-56.

Miller, K. G., Wright, J. D., and Brower, A. N., 1989. Oligocene toMiocene stable isotope stratigraphy and planktonic foraminiferbiostratigraphy of the Sierra Leone Rise (DSDP Site 366 and ODPSite 667). In Ruddiman, W., Sarnthein, M., et al., Proc. ODP, Sci.Results, 108: College Station, TX (Ocean Drilling Program),279-294.

Miller, K. G., Wright, J. D., and Fairbanks, R. G., 1991b. Unlockingthe icehouse. J. Geophys. Res., 96:6829-6848.

North American Commission on Stratigraphic Nomenclature, 1983.North American Stratigraphic Code. AAPG Bull., 67:841-875.

Pisias, N. G., and Mix, A. C , 1988. Aliasing of the geologic recordand the search for long-period Milankovitch cycles. Paleoceanog-raphy, 3:613-619.

Pisias, N. G., Shackleton, N. J., and Hall, M. A., 1985. Stable isotopeand calcium carbonate records from hydraulic piston cored Hole574A: high-resolution records from the middle Miocene. In Mayer,L., Theyer, F., Thomas, E., et al., Init. Repts. DSDP, 85:Washington (U.S. Govt. Printing Office), 735-748.

Raymo, M. E., Ruddiman, W. F., Backman, J., Clement, B. M., andMartinson, D. G., 1989. Late Pliocene variation in NorthernHemisphere ice sheets and North Atlantic deep water circulation.Paleoceanography, 4:413-446.

Ruddiman, W. F., Raymo, M., and Mclntyre, A., 1986. Matuyama41,000-year cycles: North Atlantic Ocean and Northern Hemi-sphere ice sheets. Earth Planet. Sci. Lett., 80:117-129.

Savin, S. M., Douglas, R. G., and Stehli, F. G., 1975. Tertiary marinepaleotemperatures. Geol. Soc. Am. Bull., 86:1499-1510.

Savin, S. M., Keller, G., Douglas, R. G., Killingley, J. S., Shaugh-nessy, L., Sommer, M. A., Vincent, E., and Woodruff, F., 1981.Miocene benthic foraminiferal isotope records: a synthesis. Mar.Micropaleontol., 6:423-450.

Schlich, R., Wise, S. W., Jr., et al., 1989. Proc. ODP, Init Repts., 120:College Station, TX (Ocean Drilling Program).

Shackleton, N. J., and Kennett, J. P., 1975. Paleotemperature historyof the Cenozoic and the initiation of Antarctic glaciation: oxygenand carbon isotope analyses in DSDP Sites 277, 279 and 281. InKennett, J. P., Houtz, R. E., et al., Init. Repts. DSDP, 29:Washington (U.S. Govt. Printing Office), 743-755.

Shackleton, N. J., and Opdyke, N. D., 1973. Oxygen isotope andpalaeomagnetic stratigraphy of equatorial Pacific core V28-238:oxygen isotope temperatures and ice volumes on a I05 year and I06

year scale. Quat. Res., 3:39-55.Vincent, E., and Berger, W. H., 1985. Carbon dioxide and polar

cooling in the Miocene: the Monterey Hypothesis. In Sundquist,

E., and Broecker, W. S. (Eds.), The Carbon Cycle and Atmo-spheric CO2: Natural Variations Archean to Present: Am. Geo-phys. Union, Geophys. Monogr. 32:455-468.

Williams, D. F., Thunell, R. C , Tappa, E., Rio, D., and Raffi, I.,1988. Chronology of the Pleistocene oxygen-isotope record:0-1.88 m.y.B.P. Palaeogeogr., Palaeoclimatol., Palaeoecol.,64:221-240.

Woodruff, F., and Savin, S. M., 1989. Miocene deepwater oceanog-raphy. Paleoceanography, 4:87-140.

, in press. Mid-Miocene isotope stratigraphy in the deep sea:high-resolution correlations, paleoclimatic cycles, and sedimentpreservation. Paleoceanography.

Woodruff, F., Savin, S. M., and Douglas, R. G., 1981. Miocene stableisotope record: a detailed deep Pacific Ocean study and itspaleoclimatic implications. Science, 212:665-668.

Wright, J. D., Miller, K. G., and Fairbanks, R. G., 1991. Evolution ofmodern deep-water circulation: evidence from the late MioceneSouthern Ocean. Paleoceanography, 6:275-290.

Date of initial receipt: 19 January 1990Date of acceptance: 6 February 1991Ms 120B-193

APPENDIX

Definition of Bases of Miocene Isotope Zones Mila, Milb, and Mi7

Zone MilaType level. Hole 563, Sample 82-563-14-2, 129-135 cm (282.79

mbsf)Age estimate. 21.8 MaCorrelation.

First order: Chronozone C6Ar; near the top of N4b; LO ofGloborotalia kugleri in Section 82-563-14-1; Miller et al. (1985)

Second order: NN2Locations observed. Benthic δ 1 8 θ records: Sites 563 (Miller and

Fairbanks, 1985; Wright, unpubl. data), Sites 366 and 667 (Miller etal., 1989), and Site 704 (Wright, unpubl. data)

Zone MilbType level. Hole 608, Sample 94-608H-38H-1, 66-70 cm (349.16

mbsf)Age estimate. 18.1 MaCorrelation.

First order: Chronozone C6DrSecond order: Planktonic foraminifer Zone N6

Locations observed. Benthic δ 1 8 θ records: Sites 563 (Miller andFairbanks, 1985; Wright, unpubl. data), Sites 366 and 667 (Miller etal., 1989), Site 704 (Wright, unpubl. data), and Site 588 (Kennett,1986)

Zone Mi7*Type level. Hole 608, Sample 94-608H-22-4, 80-84 cm (200.20

mbsf)Age estimate. 8.5 MaCorrelation.

First order: Chronozone C4ArSecond order: Planktonic foraminifer Zone N16

Locations observed. Benthic δ 1 8 θ records: Sites 563 (Miller andFairbanks, 1985; Wright, unpubl. data), Site 704 (Wright and Miller,unpubl. data), and Site 289 (Woodruff et al., 1981; Savin et al., 1981)

*Zone Mi7 is not formally recognized as an isotope zonation, but it is usefulfor upper Miocene correlations.

864

Site 608 Site 563 Site 747

140

160

180

200

220

240

260

280

300

320

34C

36C

38C

40(

420

mbs

f)D

epth

180 -

200 -

220 -

240 -

260 •

280 •

300

320

340

360

380

400

^ ^ J C4A•~^ • C5a -1 |C5A

J — \ C5AA

p ) U | C5AD

T™JC5B

ö_J H | C 5 E

| π C 6 A A

100

110

- 120

- 130

140

CΛ

E

Q.CD

Q

Figure 5. Oxygen isotope events identified at DSDP Sites 563 and 608. Correlative levels at Site 747 are indicated by arrows. Chronozones Mil and Mi2-Mi6 are defined byMiller et al. (in press), whereas Chronozones Mila, Milb, and Mi7 are discussed in this study.

Site 608 Site 563 Site 747P

E

CLΦQ 100

110

120

- 130

140

E

CLQ)Q

Figure 6. Carbon isotope events identified at DSDP Sites 563 and 608. Correlative levels at ODP Site 747 are also indicated. The lower Miocene δ13C increase is represented byan interval at both Sites 563 and 608. The interval from 109 to 103 mbsf at Site 747 represents a zone of uncertain correlation (see text).