Climate Evidence from Drilling on the Antarctic Margin

The onset of glaciation in Antarctica is not yet well constrained, largely because no cores have yet been obtained that unequivocally provide a continuous transition from no-ice to ice-sheet scale glaciation. This is because (i) most cores terminated before the base of the glacigenic sediments was reached, (ii) a hiatus exists at the base of the glacigenic strata, as in CRP-3 in the Ross Sea (Barrett and Ricci, 2001a,b) and Ocean Drilling Program (ODP) Site 1166 in Prydz Bay (Shipboard Scientific Party, 2001a,b,c,d), or (iii) the age models based on multiple criteria have been revised several times (e.g. CIROS-1; Wilson et al., 1998a). On land, there is evidence for ice proximal-fjordal sedimentation possibly dating back to Oligocene time in the Prince Charles Mountains (Hambrey and McKelvey,

2000a; McKelvey et al., 2001) and some exposures of the Sirius Group in the Transantarctic Mountains could also be this old (Sugden and Denton, 2004).

8.2.4.1. Drill cores in the western Ross Sea - CIROS-1 and Cape Roberts

Two drill cores have been recovered from the western Ross Sea that approach or even cross the E/O boundary (Fig. 8.2). Drilling was undertaken from a sea-ice platform in spring-time, and was characterized

Figure 8.2: Location of the drill sites in the western Ross Sea. CIROS-1 and CRP-3 are those that bear on the E/O question (from Hambrey et al., 2002). Reproduced with permission of The Geological Society Publishing House,

Figure 8.2: Location of the drill sites in the western Ross Sea. CIROS-1 and CRP-3 are those that bear on the E/O question (from Hambrey et al., 2002). Reproduced with permission of The Geological Society Publishing House,

by exceptionally high recovery (up to 98%). In the 702 m deep CIROS-1 hole (Barrett, 1989; Barrett et al., 1991), the lower part of the core was originally regarded as Late Eocene, with a breccia passing up into mudstone and sandstone. The boundary with the Oligocene was originally placed at about 570 m (Barrett et al., 1989), but magnetobiostratigraphic data (Wilson et al., 1998a) suggest that the E/O boundary is much higher at about 410-420 m. In either case, there is no obvious lithological transition, these finer grained facies including alternations of weakly stratified sand and mud, with intraformational conglomerate and occasional diamictite. These facies are strongly bioturbated and contain exotic clasts, as well as intraclasts, while some beds are graded. Moving up core, a major hiatus exists at 366 m, which coincides with the Early/Late Oligocene boundary. Above is a suite of fining-upwards sandstone beds, followed by alternating massive and stratified diamictites and thin interbeds of sand and mud (Hambrey et al.. 1989).

The CIROS-1 core was originally interpreted in terms of depositional setting, ice proximity and water depth (Hambrey et al., 1989; Hambrey & Barrett, 1993; Barrett, 1996). The breccia at the base of the hole is interpreted as a fault-brecciated conglomerate. The overlying sandstone/mudstone/ diamictite succession is marine, influenced to varying degrees by resedimentation and iceberg-rafting. Above the Early/Late Oligocene hiatus, the sandstones were regarded as fluvial, and the diamictites as basal glacial deposits, indicating ice overriding the site. However, Fielding et al. (1997) argued, based on a sequence stratigraphic analysis, that the Late Oligocene diamictite was also glaciomarine. In contrast, Hiemstra (1999) reverted in part to the original view of grounded ice on the basis of microstructural studies. Whichever solution is the correct one, there is no clear evidence for a major environmental shift at the E/O boundary, but there is one towards lower sea level at the Early/Late Oligocene transition.

A record of climate change through the E/O transition has also been determined from the environmental magnetic record in the CIROS-1 core (Sagnotti et al., 1998). Variations in magnetite were related to the concentration of detrital material transported into the Victoria Land Basin, influenced by climate and weathering rates on the Antarctic continent (especially of the Ferrar Group). Sagnotti et al. (1998) determined, from changes in the abundance of magnetite, that although there were some cold dry intervals (35-36 and >36.5 Ma) alternating with warm humid climates during the Late Eocene, a stable cold dry climate was not established in Antarctica until the E/O boundary, with major ice-sheet growth occurring at the Early/Late Oligocene boundary. This pattern matches clay mineral history, which shows a shift from smectite-rich to smectite-poor assemblages in Antarctica at the E/O boundary (Ehrmann and Mackensen, 1992; Ehrmann, 1997).

The second core that contains the E/O transition is the Cape Roberts Project Core CRP-3; Barrett (2008; Chapter 3) has synthesized the wealth of data and numerous papers from this and other Cape Roberts cores. The strata were deposited in the same rift basin, the Victoria Land Basin, as CIROS-1. The basin floor comprises Early Devonian sandstone, above which is about 1,500 m of Cenozoic sediment. CRP-3 records a dolerite conglomerate and a basal sandstone breccia at the base of the Cenozic succession, believed to be of latest Eocene age (34 Ma). Above lies nearly 800 m of sandstone with thin beds of conglomerate, all of Early Oligocene age. Diamictite and sandstone occur towards the top of the hole, while outsized clasts are scattered through much of the core. Core CRP-2A/2A almost follows on directly above CRP-3 and spans the Late Oligocene/Early Miocene interval. The CRP-1 core overlaps with the top of CRP-2/2A core and extends further up into the Miocene. Alternating sequences of mudstone, sandstone, diamictite and conglomerate occur within this part of the Cenozoic record.

The facies in CRP-3 and CRP-2A/2A represent a marine sedimentary record (Barrett, 2007). Conglomerate, sandstone and mudstone are typical of the coastal margin of a subsiding sedimentary basin, influenced by iceberg rafting. The diamictite beds additionally record tidewater glaciers that extended periodically beyond the coast. In these respects, the combined Cape Roberts cores represent an expanded record of that from CIROS-1, although there is little evidence of ice-grounding. Nevertheless, the facies fluctuations are thought to reflect glacioeustatic changes in sea level on a wave-dominated coast, in parallel with tidewater glacier fluctuations. In this context, diamicton and sand represent nearshore sedimentation, and this association grades upwards into shelf mud and then to inner shelf/shoreface sand.

A conceptual depositional model for the Late Oligocene/Miocene interval was developed by Powell et al. (2000), based on facies associations and comparison with modern glaciomarine environments, such as those in Alaska and Greenland. This shows that during an advance and still-stand, a grounding-line fan develops, and this is followed by rapid recession until another fan develops. The sequence becomes even more complex when the glacier overrides previously formed fans. Figure 8.3 is a simplified version of this model.

The huge volume of palaeoecological, mineralogical and geochemical data generated by these drilling projects enables us to gain insight into the broader environmental and climatic evolution through the early Cenozoic (Hambrey et al., 2002; Barrett, 2007). A temperate glacier regime is suggested for the Early Oligocene, followed by cooling into the Miocene, typified by

Figure 8.3: Grounding-line fan model of glaciomarine sedimentation for Late Oligocene/Early Miocene time (from Hambrey et al., 2002; Simplified from Powell et al., 2000). Reproduced with permission of The Geological Society Publishing House, Bath, UK.

Figure 8.4: Cartoon depicting the Victoria Land coast in Early/Late Oligocene time, with glacier- and river-influenced coast, and vegetated mountainsides and lowlands. This scenario is based on a combination of sedimentological evidence and floral data from CIROS-1 and Cape Roberts cores (from Hambrey et al., 2002). Reproduced with permission of The Geological Society Publishing House, Bath, UK.

Figure 8.4: Cartoon depicting the Victoria Land coast in Early/Late Oligocene time, with glacier- and river-influenced coast, and vegetated mountainsides and lowlands. This scenario is based on a combination of sedimentological evidence and floral data from CIROS-1 and Cape Roberts cores (from Hambrey et al., 2002). Reproduced with permission of The Geological Society Publishing House, Bath, UK.

polythermal glaciers. The cold frigid regime of today only began at the end of Pliocene time. The Early Oligocene landscape was characterized by temperate glaciers flowing from the early East Antarctic Ice Sheet, some terminating in the sea, and others on braided outwash plains (Fig. 8.4).

0 0

Post a comment