(Appendix) Oxygen isotope record and carbonate mineralogy of the top part of ODP Hole 101-633A

Hole 633A was drilled in the southern part of Exuma Sound on the toe-of-slope of the southeastern part of Great Bahama Bank during ODP Leg 101. The top 55 m, collected as a suite of six approximately 9.5-m-long hydraulic piston cores, represents a Pliocene-Pleistocene sequence of periplatform carbonate ooze, a mixture of pelagic calcite (foraminifer and coccolith tests), some pelagic aragonite (pteropod tests), and bank-derived fine aragonite and magnesian calcite. A 1.6-m.y.-long hiatus was identified at 43.75 mbsf using calcareous nannofossil biostratigraphy and magnetostratigraphy. The 43.75-m-thick periplatform sequence above the hiatus is a complete late Pliocene-Quaternary record of the past 2.15 m.y. The d18O curve, primarily based on Globigerinoides sacculifera, clearly displays high-frequency/low-amplitude cycles during the early Pleistocene and low-frequency/high-amplitude cycles during the middle and late Pleistocene. Variations in aragonite content in the fine fraction of the periplatform ooze show a cyclic pattern throughout the Pleistocene, as previously observed in piston cores of the upper Pleistocene. These variations correlate well with the d18O record: high aragonite corresponds to light interglacial d18O values, and vice versa. Comparison of the d18O record and the aragonite curve helps to identify 23 interglacial and glacial oxygen-isotope stages, corresponding to 10.5 aragonite cycles (labeled A to K) commonly established during the middle and late Pleistocene (0.9 Ma-present). Strictly based on the aragonite curve, another 11 aragonite cycles, labeled L to V, were identified for the early Pleistocene (0.9 to 1.6 Ma). Mismatches between the d18O record and the aragonite curve occur mainly at some of the glacial-to-interglacial transitions, where aragonite increases usually lag behind d18O depletion. When one visually connects the minima on the Pleistocene aragonite curve, low-frequency (0.4 to 0.5 m.y.) supercycles seem to be superimposed on the high-frequency cycles. The timing of this supercycle roughly matches the timing of the Pleistocene carbonate preservation supercycles described in the Pacific, Indian, and Atlantic oceans. Mismatches between aragonite and d18O cycles are even more obvious for the late Pliocene (1.6 to 2.15 Ma). Irregular aragonite variations are observed for the late Pliocene, although after the onset of late Pleistocene-like glaciations in the North Atlantic Ocean 2.4 m.y. ago the d18O record has shown a mode of high-frequency/low-amplitude cycles. Initiation of climatically induced aragonite cycles occurs only at the Pliocene-Pleistocene transition, 1.6 m.y. ago. After that time, aragonite cycles are fully developed throughout the Quaternary. The 11-m-thick periplatform sequence below the hiatus represents a lower Pliocene interval between 3.75 and 4.45 Ma. The bottom half (4.25-4.45 Ma) has a fairly constant, high aragonite content (averaging 60%) and high sedimentation rates (28 m/m.y.) and corresponds to the end of the prolonged early Pliocene interglacial interval (4.1-5.0 Ma), established as a worldwide high sea-level stand. The second half (3.75-4.25 Ma), in which aragonite content decreases by successive steps, paralleled by a gradual 5180 enrichment in Globigerinoides sacculifera and low sedimentation rates (10 m/m.y), corresponds to the climatic deterioration established worldwide between 4.1 and 3.8 Ma, to a decrease of carbonate preservation observed in the equatorial Pacific Ocean, and to a global sea-level decline. Dolomite, a ubiquitous secondary component in the lower Pliocene, is interpreted as being authigenic and possibly related to diagenetic transformation of primary bank-derived fine magnesian calcite. Transformation of the primary mineralogical composition of the periplatform ooze was evidently minor, as the sediments have retained a detailed record of the Pliocene-Pleistocene climatic evolution. Clear evidence of diagenetic transformations in the periplatform ooze includes (1) the disappearance of magnesian calcite in the upper 20 m of Hole 633A, (2) the occurrence of calcite overgrowths on foraminiferal tests and microclasts at intermittent chalky core levels, and (3) the ubiquitous presence of authigenic dolomite in the lower Pliocene.

Foraminiferal oxygen and carbon isotopic measurements from 5 sites of southern Atlantic and Pacific

The stratigraphy and paleoceanography of the late Miocene and early Pliocene have been examined at six sites in the South Atlantic and southwest Pacific oceans: Deep Sea Drilling Project (DSDP) sites 284, 516A, 519, 588, and 590 and two piston cores from Chain cruise 115. A consistent stratigraphy was developed among sites using graphic correlation, which resulted in age models for all sites that are tied to the revised paleomagnetic time scale of Berggren et al. (1985). Applying these chronologies, we assessed latitudinal and interocean contrasts in the stratigraphic ranges of late Miocene-early Pliocene planktonic foraminiferal and nanno - fossil datums. Salient stratigraphic results include (1) The last appearance datum (LAD) of Globoquadrina dehiscens is a late Miocene (approx. 6.4 Ma) event in the subtropics and is not useful for the placement of the Miocene/Pliocene (M/P) boundary in this biogeographic province. (2) The first appearance datum (FAD) of Globorotalia crassaformis occurred at 5.1 Ma in the South Atlantic near the M/P boundary, suggesting that Gr. crassaformis may have first evolved in the South Atlantic and later migrated to other regions. (3) In the southwest Pacific, the FADs of Gr. margaritae (5.97 Ma), Gr. puncticulata (5.09 Ma), and Gr. crassaformis (4.87 Ma) are significantly time transgressive between temperate and warm subtropical regions. Time lags of 1.0 m.y. were required for these species to adapt to physical and/or biotic conditions peripheral to their endemic biogeographic provinces. (4) Between the subtropics of the South Atlantic and southwest Pacific, many planktonic foraminiferal datums (FAD of Dentogloboquadrina altispira, Gr. cibaoensis, Gr. conomiozea, Gr. margaritae, and Gq. dehiscens and LAD of Gr. cibaoensis) markedly depart from the correlation suggested by magnetostratigraphy, indicating that these datum levels are unreliable for correlation between these ocean basins. (5) In contrast, available calcareous nannofossil datum levels fall on or near the paleomagnetic correlation line, indicating synchroneity of events within the subtropics. (6) Biostratigraphic, magnetic, and 87Sr/86Sr correlation between sites 588 and 519 and the M/P neostratotype at Capo Rossello, Sicily, suggests that the base of the Zanclean stratotype occurs at 5.1-5.0 Ma in the lower reversed subchron of the Gilbert, about 2-3 * 10**5 years above the Gilbert/Chron 5 boundary. Oxygen isotopic results from DSDP sites 284, 519, and CH115 piston cores confirm a prolonged benthic d18O increase in the latest Miocene between 5.6 and 5.0 Ma, as originally proposed by Shackleton and Kennett (1975). At DSDP site 588, the benthic d18O record in the latest Miocene is marked by high-frequency fluctuations with amplitude variations of 0.5per mill, and a long-period wavelength component of 400,000 years. Maximum d18O values, however, occurred during the late Miocene (Kapitean Stage) between 5.5 and 5.1 Ma. The late Miocene d18O changes resulted from mid- and high-latitude cooling and pulses of ice sheet expansion and contraction. Glacial events were most intense during the latest Miocene (Kapitean Stage), and occurred at 5.50-5.35 Ma and at 5.10 Ma. Glacial events are estimated to have lowered sea level by 40 to 60 m and contributed to the isolation and desiccation of the Mediterranean Basin during the late Messinian. Interglacial conditions prevailed at 5.2 Ma and between 5.0 and 4.1 Ma in the early Pliocene. The beginning of the Pliocene was marked by changes in many proxy climatic indicators at all sites, suggesting a prolonged interval of warm, interglacial conditions between 5.0 and 4.1 Ma during the earliest Pliocene.

Upper Eocene to Oligocene strontium, carbon and oxygen isotope record of DSDP holes

We improved upper Eocene to Oligocene deep-sea chronostratigraphic control by integrating isotope (87Sr/86Sr, delta18O, delta13C) stratigraphy and magnetostratigraphy. Most previous attempts to establish the timing of isotope fluctuations have relied upon biostratigraphic age estimates which have uncertainties of 0.5 to over 4.0 m.y. Deep Sea Drilling Project (DSDP) Site 522 contains the best available upper Eocene to Oligocene magnetostratigraphic record which allows first-order correlations of isotope records (87Sr/86Sr, delta18O, delta13C) to the Geomagnetic Polarity Time Scale (GPTS). Empirical calibrations between the 87Sr/86Sr of foraminifera and magnetochronology at Site 522 allow more precise correlation of ,unknown' samples with the GPTS. For example, shallow water and high-latitude sections may be tied into the deep-sea record. Sr-isotope stratigraphic resolution for the latest Eocene to Oligocene is approximately 2 m.y.

Paleomagnetic properties of ODP Site 119-744 (Appendix)

Paleomagnetic studies conducted on board JOIDES Resolution during Leg 119 indicate that the cores collected at Site 744 range from Quaternary through Eocene in age. Initial studies of the sediments completed on board the ship measured the magnetization of the archive halves of the sedimentary cores, using the pass-through cryogenic magnetometer. Stratigraphic plots of the declination and inclination derived from these measurements displayed numerous long intervals with essentially constant magnetic directions. Further study of these intervals led to a discovery that the background signal had been incorrectly computed due to faulty software on the ship. Because this background signal was not recorded in the data-processing system, corrections could not be made. Therefore, subsequent shorebased studies have been made on the individual samples collected at approximately 30-cm intervals in the cores in order to verify the initial magnetostratigraphy reported in the Initial Reports volume for Leg 119 (Barron, Larsen, et al., 1989, doi:10.2973/odp.proc.ir.119.1989). Numerous reversals were identified and correlations were suggested with the seafloor magnetic anomaly sequence of Berggren et al. (1985, doi:10.1130/0016-7606(1985)96<1407:CG>2.0.CO;2) back to anomaly number 17.

Chronostratigraphic summary and raw counts of selected dinoflagellate cysts and architarch taxa of ODP Hole 151-907A

A detailed dinoflagellate cyst investigation of the almost continuous Middle Miocene through Pliocene of Ocean Drilling Program Hole 907A in the Iceland Sea has been conducted at 100-kyr resolution. The investigated section is well constrained by magnetostratigraphy, providing for the first time an independent temporal control on a succession of northern high-latitude dinoflagellate cyst bioevents. Based on the highest/lowest occurrences (HO/LO) and highest common occurrence (HCO) of 20 dinoflagellate cyst taxa and one acritarch species, 26 bioevents have been defined and compared with those recorded at selected DSDP, ODP, and IODP sites from the North Atlantic and contiguous seas, and in outcrops and boreholes from the onshore and offshore eastern U.S.A., and the North Sea and Mediterranean basins. Comparisons reveal near-synchronous HOs of the dinoflagellate cysts Batiacasphaera micropapillata (3.8-3.4 Ma, mid-Pliocene) and Reticulatosphaera actinocoronata (4.8-4.2 Ma, Lower Pliocene) across the Nordic Seas and North Atlantic, highlighting their value on a supraregional scale. This probably applies also to Hystrichosphaeropsis obscura (upper Tortonian), when excluding ODP Hole 907A where its sporadic upper stratigraphic range presumably relates to cooling in the early Tortonian. Over a broader time span within the upper Tortonian, the HO of Operculodinium piaseckii likely also permits correlation across the Nordic Seas and North Atlantic, and the HO of Labyrinthodinium truncatum appears useful in the Labrador and Nordic Seas. Biostratigraphic markers useful for regional rather than supraregional correlation are the HOs of Batiacasphaera hirsuta (c. 8.4 Ma, upper Tortonian) and Unipontidinium aquaeductus (c. 13.6-13.9 Ma, upper Langhian), the HCO of the acritarch Decahedrella martinheadii (c. 6.7-6.3 Ma, Messinian), and possibly the LO of Cerebrocysta irregulare sp. nov. (c. 13.8 Ma, uppermost Langhian) across the Nordic Seas. Since Habibacysta tectata, B. micropapillata, R. actinocoronata and D. martinheadii have been observed in the Arctic Ocean, they are potentially useful for high latitude correlations in the polar domain. The LOs of Habibacysta tectata and Unipontidinium aquaeductus suggest a mid- to late Langhian age (15.1-13.7 Ma) for deposits at the base of Hole 907A, thus providing new constraints on the age of basalts at the base of ODP Hole 907A. The stratigraphically important dinoflagellate cysts Cerebrocysta irregulare sp. nov., and Impagidinium elongatum sp. nov. are formally described.