18O records, which are identifiable at both sites despite relatively low sedimentation rates (Fig. F2A).
All ages given in this study are based on Cande and Kent (1992; 1995) (CK95) unless stated otherwise (i.e., the OMB is placed at 23.8 Ma). Although the astronomically tuned age of the OMB is 22.9 Ma (Shackleton et al., 2000), the older Oligocene part of the section does not have a revised age scale; thus, we prefer to stick with CK95 for the time being. Data in the older parts are available only at low resolution (identified by gray backgrounds in figures), whereas resolution in the younger parts is 75 cm for Site 1168, 20 cm for Sites 1170 and 1171, and 10 cm for Site 1172. All biomagnetostratigraphic datums (as of August 2002) and additional control points used for our age models are listed in Table T1 (see also Stickley et al., this volume).
The benthic stable isotope records of Site 1170 on the South Tasman Rise and Site 1090 on the Agulhas Ridge, both located in the Southern Ocean, suggest some similarities despite different paleolatitudes. The benthic isotope record for Site 1090 was established by Billups et al. (2002), who developed their age model by matching the excellent magnetic polarity stratigraphy to the geomagnetic polarity timescale of CK95. A comparison of their results with the astronomically tuned record of Shackleton et al. (2000) yielded a constant offset of ~900 k.y. between the two approaches. We focus on the timing of the Mi-1 and Mi-1a event in the benthic 18O records, which are identifiable at both sites despite relatively low sedimentation rates (Fig. F2A).
The Mi-1 event appears in benthic foraminiferal 18O records as a cold interval lasting from ~24.02 to 23.66 Ma, centering around a double peak of extremely heavy (cold) values from 23.89 to 23.79 Ma (cf. Billups et al., 2002). One important difference between this event at Site 1170 relative to Site 1090 (as well as 929) (Paul et al., 2000) is that the second peak appears to be as strong, if not even slightly stronger than the first peak. A second difference is that this double peak appears to begin and end later at Site 1170 (Fig. F2B). At Site 1170 timing of the Mi-1 event is constrained by four magnetostratigraphic datums across a depth of only 3.5 m, including one at 23.8 Ma, giving evidence for a strongly condensed section. It is this reversal at 23.8 Ma (onset of Subchron C6Cn.2n at 382.90 meters below seafloor [mbsf]), which is in conflict with an alignment of the Mi-1 double peak at both sites (23.88 Ma [383.36 mbsf] and 23.8 Ma [381.96 mbsf]). Matching the timing of the Mi-1a event at Site 1170 with Site 1090 provides two additional control points (Fig. F2A).
Additional information evident in the record of weight percent sand for Site 1170 indicates a change in the pattern of sedimentation (increase in the coarse fraction) between 406.69 and 405.34 mbsf (Fig. F4). This is consistent with the correlation indicated by the magnetostratigraphy (Fig. F3). The weight percent sand record further suggests placement of the Marshall Paraconformity (~32-29 Ma) (Fulthorpe et al., 1996) somewhere between 476.66 and 463.61 mbsf (no samples) at a drop in average values (see also "Site 1171," "Site 1172," and "Site 1168"). This hiatus is common in sedimentary records from the southwest Pacific Ocean and is a logical explanation for an early Oligocene hiatus in the Leg 189 sites. At Site 1170 this hiatus is constrained by three biostratigraphic markers (33 Ma [478.5-460 mbsf] and 30.62/30.24 Ma [462.3-454.61 mbsf]) and one magnetic reversal (33.058 Ma [470 mbsf]) (Fig. F3). We use the center of the youngest datum (30.24 Ma; first occurrence [FO] of Rocella vigilans) for the purpose of interpolation downward from the last control point (Fig. F3).
These data raise an interesting conundrum in correlation, as one would not expect to find the isotope signature of a significant glacial event (Mi-1) with different ages at separated locations (i.e., they should occur at the same time worldwide), but the same is also true of magnetic reversal ages. Closer examination is required. See also more recent age model information in the papers by Stickley et al. (this volume) and Pfuhl et al. (in press).
Additional records of bulk 13C and benthic 18O and 13C presented in Figure F4 will be referred to in "Site 1171" below and "Site 1172".
The biomagnetostratigraphy for Site 1171 suggests a transition from a heavily condensed Oligocene (sedimentation rates < 0.5 cm/k.y.) to slightly increased sedimentation rates in the Miocene (0.5-1 cm/k.y.) (Fig. F3). Unfortunately, magnetic reversals are less frequently observed than at Site 1170 and we fail to detect a complete Mi-1 event in the benthic 18O record (Fig. F5), but we do observe a number of characteristic markers in the records obtained.
Our oldest observation is a dramatic (~60%) and instantaneous (between 273 and 272.8 mbsf [279.68 and 279.48 meters composite depth (mcd)]) increase in CaCO3 accompanied by a drop in weight percent sand similar to the one observed at Site 1170 (Fig. F4). We once again suggest that this transition marks the Marshall Paraconformity. This is in good accordance with six biostratigraphic datums (Fig. F3). We use the same diatom datum as at Site 1170 to mark the top of the hiatus at 30.24 Ma (FO of R. vigilans) and our suggested depth of 272.9 mbsf (error of datum = 274.32-272.74 mbsf). It is followed by a rapid transition to heavier values in both benthic and bulk 13C (269.36-269.16 and 269.16-268.56 mbsf) (Fig. F5). The lack of variability and the transition over 60 cm in the case of the bulk record suggests a strongly condensed section rather than a few hiatuses. We observe similar shifts in the 13C records at Site 1170 at ~401 mbsf (~26.9 Ma) and 398 mbsf (~26.34 Ma) followed by a strongly condensed interval ending at ~24.781 Ma (Fig. F4). Three biostratigraphic control points at Site 1171 suggest the same pattern for this site. Unfortunately, this interval is also marked by poor recovery. However, in order to relate these observations to the age model we decided to add control points centering on the two 13C increases and ignore the biostratigraphic datums.
The magnetic reversal at 268 mbsf, assigned an age of 24.118 Ma (onset of Subchron C6Cn.3n), implies a very strong increase in sedimentation rate, which appears most unlikely based on our present understanding of the sedimentary regime at all Leg 189 sites. Considering both poor recovery and low sedimentation rates before and after this datum, we suggest this datum also be ignored. A further control point, which we place at 258.56 mbsf (23.88 Ma), is based on the increase in benthic 18O, which we identify as the onset of the Mi-1 event (Fig. F5). The remainder of this event was not recovered during drilling of Hole 1171C (core gap at 258.80-253.84 mbsf). Connection of this and the previous control points are consistent with the biostratigraphy (Fig. F3).
Above this gap we observe a low in benthic 13C and matched increases in bulk 13C and CaCO3 (251.90 and 251.06-250.86 mbsf) (Fig. F5). Based on similarities with Site 1170, we assign ages of 23.15 and 22.8 Ma (increase at Site 1170 between samples at 377.46 and 376.50 mbsf across a core gap) (Fig. F4). The ensuing low sedimentation rates lasting ~1 m.y. following the OMB at Site 1171 are consistent with biostratigraphic datums at this site and a similar trend at Site 1170. An additional control point is placed at 22.4 Ma (249.36 mbsf) based on correlation with the benthic 18O record at Site 1170. A last control point is placed at 21.7 Ma (240.23 mbsf) based on the benthic 13C record.
Two more gaps in recovery in Hole 1171C are evident at 268.10-264.88 and 248.90-247.38 mbsf.
Site 1172 has a greater number of magnetostratigraphic control points than Site 1171 across the interval studied (Fig. F3). However, the Oligocene interval is marked by very few additional biostratigraphic datums. We estimate that sedimentation resumes above the Marshall Paraconformity at 30.098 Ma (magnetic reversal onset of Subchron C11n.2n [358.6 mbsf]), but sedimentation rates remain very low, as suggested by three closely spaced magnetic reversals. As at Sites 1170 and 1171, we observe a sharp drop in weight percent sand at the Marshall Paraconformity (i.e., between 358.66 and 358.16 mbsf) (Fig. F6). This constrains placement of the hiatus to within 6 cm (358.66-358.60 mbsf). Our records suggest two additional control points based on correlation with Sites 1170 and 1171. At Site 1172 the bulk 13C record shows two strong increases (347.36-347.26 and 342.96-342.76 mbsf) (Fig. F6) that we match with the respective control points for the other two sites. Whereas these datum levels are close to the error given for two biostratigraphic datums, they are inconsistent with the magnetostratigraphy. However, in a strongly condensed section with potential hiatuses, identification of reversals is severely compromised.
We are confident in the placement of these two 13C datums and suggest two additional points based on similarities between the benthic 18O (onset of the Mi-1 event) and benthic 13C records from 344.47-344.27 and 342.56-342.46 mbsf, respectively (Fig. F6). The second control point is more difficult to define at Site 1170 where the decrease in benthic 13C occurs over a number of samples from 375.56 to 375.06 mbsf (Fig. F4). However, we place our control point based on the turning point in the 13C record. Two final control points are based on the benthic and bulk 13C records at 23.15 Ma (343.26 mbsf) and 21.79 Ma (339.86 mbsf).
Although the general trends observed in the records at Sites 1170-1172 reveal numerous similarities, we have difficulty in detecting matching patterns at Site 1168 (Fig. F7). In this we are also constrained by lack of comparable data (i.e., the only stable isotopes available were measured on benthic foraminifers at very low resolution) (Dr. S. Schellenberg, pers. comm., 2002; but see also Pfuhl et al., in press). CaCO3 was measured on the fine fraction only, but a comparison with bulk values at this site (Dr. C. Kelly, unpubl. data) and between bulk and fine values at Sites 1170-1172 suggest that the records can be overlaid (Fig. F7).
Our data reveal a strong transitional interval from ~765 to 735 mbsf marked by decreasing weight percent sand and organic carbon content, as well as increasing carbonate content (Fig. F8). We suggest that this interval marks the Eocene-Oligocene transition similar to Sites 1170-1172, possibly without the development of the Marshall Paraconformity at this rather protected site on the continental margin of Tasmania. Following this transition we note a number of changes in the biogenic and lithogenic composition of the coarse fraction (>63 µm) at this site (Fig. F9). The first two occur at ~699 and 538 mbsf. At ~429 mbsf we note the end of a strongly condensed interval (shipboard observation on benthic foraminifers). Between 405.22 and 404.98 mbsf the carbonate content increases (Fig. F8) followed by an increase in dissolution and weight percent sand at 386.23 mbsf (Fig. F9). All these datums are matched closely by changes in sedimentation suggested by the biomagnetostratigraphy based on magnetic reversal interpretation MR1 (Fig. F3).
If we assume the timing of the Mi-1 event recognized isotopically to be identical at all Leg 189 sites, we would have to tentatively place the peaks of this event at 23.9 Ma (420 mbsf) and 23.75 Ma (414 mbsf). This is consistent with the biomagnetostratigraphy (based on magnetic reversal interpretation MR1) within the error resulting from very low resolution (1.50 m) of measurements available at present. In comparison with the other three sites the changes in sedimentation rates across the Oligocene-Miocene transition are relatively strongest at this site. We suggest ignoring the planktonic foraminiferal datum at 22.8 Ma (379.60 mbsf; last occurrence [LO] of Dentoglobigerina globularis) because of its proximity to an event at 22.6 Ma (373.97 mbsf; FO of Globoturborotalita woodi) and place the FO of Globoquadrina dehiscens (planktonic foraminifers) at its upper limit of error (407.64 mbsf). The extreme increase in sedimentation rate at Site 1168 prior to the Oligocene-Miocene transition is fixed by two magnetic reversals and has to be accepted at this point.
In Figure F10 we show the results of our measurements of lithologic and isotopic parameters at Sites 1170-1172 against their revised ages on the timescale of Cande and Kent (1992, 1995). Figure F11 shows a histogram of the sedimentation rates.
