Age-depth plots for each site are presented with magnetostratigraphic and biostratigraphic events in Figures F6, F7, and F8. Biostratigraphic data is from Tables T4 and T5; polarity events and their interpretation are from Table T7. Previous magnetopolarity interpretations are also given in Table T6.
The age model for the lowermost part of the section at Site 1095 is based solely on paleomagnetic data. Magnetostratigraphy suggests an essentially complete section throughout this time interval with high but gradually decreasing sedimentation rates (average = 80 m/m.y, with higher values of ~120 m/m.y. at the base of the section), and with only minor short-term changes in sedimentation rate values. In the interval between ~150 and 420 mcd, the biostratigraphic data, despite the presence of virtually all zones for both diatoms and radiolarians (see "Biostratigraphic Summary" in "Site 1095" in "Data Sources: Magnetobiochronology of Leg 178 Rise Sites"), suggests significant differences in sedimentation rates, with somewhat higher rates of sedimentation on average than those implied by the magnetostratigraphic data and with the presence of an ~1-m.y.-long interval of low sedimentation, or a hiatus, at ~235 mcd. The biostratigraphic data is internally in good agreement; there is little scatter in the data, and there are not any systematic differences in age estimates between diatoms and radiolarians. Thus, whereas the age estimates from magnetostratigraphy and biostratigraphy are in broad agreement—both suggest ages from early late Miocene to basal Pliocene—there are significant offsets between the two types of data over much of the interval. The maximum age difference is ~1 m.y. in the Miocene/Pliocene boundary interval between ~215 and 275 mcd. There are several possible reasons for this discrepancy, including incorrect assignment of the paleomagnetic polarity data to the GPTS, inaccurate identification of the position of biostratigraphic events within the section, and incorrect calibrations of biostratigraphic events to the GPTS.
The magnetostratigraphic polarity pattern of Site 1095 is based on a coherent set of shipboard core measurements, U-channel samples, and logging results, and the proposed correlation in the pattern seen in the data and the GPTS, as plotted in Figure F6, is remarkably clear and straightforward. Although it is possible to correlate the magnetostratigraphic polarity pattern to the GPTS in such a way that the line of correlation is in agreement with the biostratigraphic data (e.g., Iwai, 2000a, 2000b) (Table T6), such interpretations require repeated major changes in sedimentation rates within the section. There is no evidence, however, from either seismic stratigraphy or sedimentology for a hiatus in this section or for any major changes in sediment characteristics that would suggest such rapid changes in sedimentation rate.
Biostratigraphic events can be mislocated within sedimentary sections for a variety of reasons, including reworking of older specimens into younger sediments, downcore contamination of older sediments by younger sediments during drilling, range truncation due to poor preservation, extension or truncation of local ranges due to local environmental conditions, and misidentification of species due to taxonomic problems, among others. We cannot rule out such processes entirely. Indeed, given the rather sporadic nature of microfossil preservation at this site, some degree of range error from this processes is to be expected. We nonetheless note that on the whole, the biostratigraphic data is remarkably internally consistent; there is very little scatter in the data. Instead, the data is consistently offset from the age implied by the magnetostratigraphic data. It is difficult to envisage how any of the above-mentioned processes could produce so consistent a pattern of offsets in the biostratigraphic data. Nor is there any evidence that any of the other above processes was particularly important at this site: reworking has not been noted as a problem in any of the original biostratigraphic reports; samples taken from within split cores are the primary source of data in this study and, in any case, do not differ from core catcher observations; the presence of all zonal marker species for both diatoms and radiolarians suggests that the local environment was not especially unusual; and most of the species used for zonation in this study are well-known species that have been employed for decades in Southern Ocean biostratigraphic work.
We thus conclude that the most probable reason for the discrepancy between the biostratigraphic and magnetostratigraphic data from Site 1095 is previous miscalibration of Southern Ocean biostratigraphic events to the GPTS in the late Miocene to basal Pliocene time interval. This interval has long been the most difficult part of the entire Neogene Southern Ocean biostratigraphic zonation to calibrate, due to the presence of hiatuses and low sedimentation rates in many, or most, of the sections used previously, plus the difficulty in synthesizing data from multiple holes in the absence of mcd scales (Gersonde et al., 1990; Barron et al., 1991; Harwood et al., 1992). Although our conclusion is tentative and needs to be confirmed by reexamination of those sections used to make the published calibrations, this result, if true, may help to significantly improve the accuracy of Neogene chronologic scales in the Antarctic region.
In the remaining part of the Pliocene section of Site 1095 (~4.5-1.8 Ma; ~150-25 mcd) biostratigraphic and magnetostratigraphic data are in agreement and indicate a (relatively) low sedimentation rate of ~50 m/m.y.
The age model for Site 1095 in the Pleistocene suggests continuous sedimentation, albeit at a relatively low rate of ~10 m/m.y. An unconformity was based on seismic stratigraphy, coincident with a prominent lithostratigraphic boundary (53.54 mcd = 57.5 mbsf in figure F4a of the "Site 1095" chapter in the Leg 178 Initial Reports volume; Shipboard Scientific Party, 1999b). During Leg 178, Chron C1r.1n. through C2An.2r. could not be confidently identified and the reversal at ~58.8 mbsf (= 54.8 mcd) was interpreted as the termination of the Olduvai (1.77 Ma). The onset of the Olduvai and all of Chron C2r were considered to be lost in a hiatus at the termination of the Olduvai (see "Magnetostratigraphy" in the "Site 1095" chapter; Shipboard Scientific Party, 1999b). Postcruise magnetostratigraphic interpretation using the undisturbed U-channel samples and the composite depth scale suggests that all polarity events are present, and thus no hiatus (within the resolution of the magnetostratigraphic data) needs be proposed (Acton et al., Chap. 37, this volume). Postcruise observation of core photographs (digital images are available from www-odp.tamu.edu/publications/178_IR/VOLUME/CORES/COR_1095.PDF) by the first author suggests that there is an erosional contact in Section 178-1095D-7H-2, 75 cm (54.55 mcd), at the bottom of a massive coarse sediment layer. All of the other turbidite layers in this section are parallel to the coring direction (perpendicular to the core liner walls). The irregular contact seen at 75 cm in this section is therefore interpreted as an original sedimentary structure, although it was described as core disturbance in the initial report (see "Lithology" in the "Site 1095" chapter; Shipboard Scientific Party, 1999b). This possible erosional contact in Hole 1095D may be correlated to the seismic unconformity observed at this site. This unconformity is considered to be missing in Hole 1095A between Cores 178-1095A-7H and 8H. However, the presence of all expected magnetostratigraphic events and the absence of any noticeable sedimentation rate change suggest that the time gap of the seismic unconformity (possible hiatus) is small and is probably <200 k.y. (Acton et al., Chap. 37, this volume).
The age-depth plot for Site 1096 (Fig. F7) is in most respects similar to the age equivalent interval in Site 1095, with higher rates of sedimentation (~180 m/m.y) in the Pliocene, giving way to lower rates (~80 m/m.y) in the Pleistocene. Biostratigraphic and magnetostratigraphic data are in good agreement with each other except for several biostratigraphic events, all with age calibrations between 3.2 and 3.8 Ma, which are scattered over a wide depth interval at Site 1096. We have no explanation for this, although one of the events—the top of the radiolarian Lampromitra coronata—is only an informal marker which has not been yet studied for biostratigraphic consistency.
Site 1101 is the shortest of the sections studied here and reached only the mid-Pliocene at the base of the hole (217.7 mbsf). The magnetostratigraphic age model fits the biostratigraphic data reasonably well throughout and suggests a nearly constant rate of sedimentation over the entire 0- to 3.0-Ma time interval. Only one event—the top of the radiolarian Cycladophora pliocenica—is noticeably offset from the remainder of the data. This event is usually quite reliable in Antarctic stratigraphy. However, the morphology of this species is similar to several other species in the same genus, and it is possible that the discrepancy is due to a taxonomic misidentification, particularly as the shipboard preparations were characterized by poor sample breakdown, which tends to obscure taxonomically important morphologic detail (Lazarus, Chap. 13, this volume).