Australian Journal of Earth Sciences (2001) 48, 621–632
Stratigraphy, sedimentary development and palaeoenvironmental context of a naturally accumulated pitfall cave deposit from southeastern Australia A. M. KOS Department of Earth Sciences, Monash University, Vic. 3800, Australia (
[email protected]). McEachern’s Deathtrap Cave (G-49/50) is located in the Lower Glenelg region of southeastern Australia and records a Late Pleistocene to Holocene sedimentary record that has been directly influenced by surface processes during its formation. The sedimentary sequence contained within the cave is divided into lower, middle and upper sequences consisting of eight facies. The lower sequence represents the earliest phase of sedimentation, and groundwater fluctuations during the Last Interglacial period resulted in its erosion and redistribution deeper into the cave system. A decrease in the magnitude and frequency of flood events in the cave during the formation of the middle sequence indicates increasingly drier surface conditions prior to the Last Glacial Maximum. The middle sequence has a minimum age of 9840 290 a BP. Moving sand sheets during the Last Glacial blocked the entrance to the cave allowing flowstones to develop on the cave floor. The surface environment surrounding the cave was probably not as dry as contemporaneous inland sites because sedimentation continued to be dominated by flowing water during this period. Holocene sedimentation is represented by the upper sequence and reflects wetter cave conditions between 7680 160 a BP and 5700 110 a BP. A major phase of sediment accretion occurs after 5700 a BP and correlates to a phase of dune instability in the Lower Glenelg region. Flowing water remodelled the sediment cone sometime after 2240 100 a BP, which represents a period of increased surface runoff, although it is not clear whether this is due to climatic or anthropogenic influences. KEY WORDS: caves, McEachern’s Deathtrap Cave, palaeoenvironment, sedimentary history, stratigraphy, Victoria.
INTRODUCTION The stratigraphy, sedimentary history and palaeoenvironmental record of a sedimentary deposit in McEachern’s Deathtrap Cave (G-49/50), one of several caves located in the Lower Glenelg region of southwest Victoria (Figure 1), is described in this study. In the early part of the 20th century graziers had known of the cave’s existence, although knowledge of its exact whereabouts had been lost. In 1987 National Park rangers rediscovered the site, thereby providing a new opportunity for scientific study of the natural history and palaeoenvironmental record of the Lower Glenelg region.
Geological setting The Lower Glenelg region of southwest Victoria constitutes part of the southeastern extension of a Pleistocene strandline dune system that extends across southeast South Australia and southwest Victoria (Figure 2). The dune system is a result of Pleistocene sea-level fluctuations and crustal upwarping (Sprigg 1952, 1979; Schwebel 1984; White 1994). Calcareous dunes form the Bridgewater Formation and are derived from reworked Oligo-Miocene Mt Gambier Limestone Formation (Boutakoff 1963). Pleistocene dunes form consolidated, massive dune limestones, whereas younger dunes remain unconsolidated. Wind winnowing
of the calcareous dunes has produced siliceous residual sands known as the Malanganee Formation (Boutakoff 1963). The Malanganee Formation began forming in the Early Pleistocene and today is represented by draping sand sheets and dune-like bodies that overlie some truncated Bridgewater Formation dune sequences.
MCEACHERN’S DEATHTRAP CAVE Ackroyd (1994) and Kos (1998) have both provided geomorphological descriptions of McEachern’s Deathtrap Cave (G-49/50). The cave is a classic pitfall trap and consists of a long, narrow passage connected to the surface by two cylindrical entrance shafts that are 2 m in diameter (Kos 1998). At the bottom of each entrance shaft coneshaped deposits of terrigenous sediment have formed. In this study, only the sedimentary deposit associated with the G-49 entrance (southeast section of the cave) will be presented (Figure 3). The G-49 deposit is asymmetric in longitudinal section and its shape and distribution have been controlled by the 3-D confines of the cave. Several test pits have been excavated and the stratigraphy described from Trench A represents the type section for the deposit. The lateral extent of stratigraphic layers in the G-49 deposit was determined by hydraulic coring (Figure 3).
622
A. M. Kos
It is worth noting that the cave system is larger than the surveyed area shown in Figure 3, but access is restricted due to significant roof fall in the southeast cavern and sediment accumulation (G-50 deposit) in the northwest section.
SEDIMENTARY SEQUENCE The stratigraphical succession can be divided into upper, middle and lower sequences consisting of eight depositional and erosional facies (Figure 4). A summary of their main features is given in Table 1. Stratigraphic logging, identification and description of sedimentary beds were carried out primarily on exposed sections in Trenches A and B. Facies classification was based on a combination of bedding and textural characteristics that dominated individual sedimentary horizons (Reading 1996). Textural descriptions followed the terminology of Folk et al. (1974). Granulometry and basic geochemistry were undertaken in the laboratory following the method of Folk (1974).
Radiometric dating A suite of charcoal samples was analysed using the AMS 14C method of absolute age determination. Charcoal fragments were sampled from various stratigraphic levels in Trench A and provide a chronological framework for the upper part of the middle sequence and the upper sequence (Figure 4). Uncalibrated radiocarbon ages are shown in Table 2. The charcoal dates do not indicate any form of age reversal and are consistent with the sequence of sedimentary deposition and erosion interpreted from stratigraphic relationships (see discussion section). Two samples of layered calcium carbonate flowstone from Trenches A and B were dated using conventional 230 Th/234U analysis. The results were inconclusive and ages could not be reliably calculated due to low uranium
content and significant 232Th contamination of the precipitated calcium carbonate (R. Chisari pers. comm. 1995).
Facies description and interpretation Facies A consists of limestone clasts supported by a muddy sand matrix and is the dominant facies in the G-49 sedimentary sequence (Figure 5, Table 1). In bed 1, facies A is chaotically bedded and contains extinct megafaunal remains. The upper boundary is erosional and characterised by channel incision, although this is only apparent in Trench A. Beds 10, 12 and 15 of the upper sequence are poorly sorted and diffusely stratified. Bedding thickness ranges from 10 mm to 50 mm. Mass-flow processes have dominated the mode of sediment deposition and redistribution in beds 10, 12 and 15, demonstrated by the finegrained, muddy characteristic of the matrix (Bull 1972; Hampton 1975; Tucker 1991). Scour and fill structures characterise facies B in beds 2 and 14. Scouring features appear as channel-like structures that are the product of open channel flow. Scour marks vary between 100 mm and 300 mm in depth and between 150 mm and 300 mm in width (Figure 6a, b). Scour-fill deposits consist of massively to finely laminated, clean silty quartz sands with basal gravelly layers. In bed 14, the lower channel fill deposits are massive, becoming laminated towards the top where laminations are delineated by fine organic material (Figure 6b). Soft sediment deformation and slumping that is characterised by rotated slip blocks are also common features in facies B (Figure 6b). Facies C is characterised by a thinly bedded, cone-like pile of very fine to fine quartz sand, which has accumulated beneath a small solution tube that is presently choked with sediment. Bedding dips 16° northeast and 16–18° southwest around a central apex in the southeast section (Figure 5), and 14° northwest in the southwest section, where bedding also thins out. The pile has a minimum height of ~0.5 m and boundaries are sharp and erosive, representing local erosional discordances. The upper boundary is sharp
Figure 1 Location of McEachern’s Deathtrap Cave G-49/50 in the Lower Glenelg region, southeastern Australia.
Palaeoenvironments in a pitfall cave deposit and irregular, and acted as a cave floor for some period of time. Facies D consists of diffusely bedded, texturally homogeneous, very fine to fine quartz sand (Figure 5). Individual beds have a massive internal structure and contain small sand and clay intraclasts. Boundaries between beds are both sharp and gradational. Sedimentation in facies D appears to have been dominated by rapid deposition through repeated pulses of sediment-laden water. Several of the flow events appear to have been laterally extensive and were probably confined by the walls of the cave. In beds 5 and 11 wispy lenticular laminations containing clean, silty, very fine to fine quartz sands represent facies E (Figure 6a). The lenticular laminations are up to 20 mm thick and average ~50 mm in length. In bed 5, the laminations appear to be more prevalent in its middle to upper sections. Lenticular laminations indicate a lower to transitional flow regime with very shallow water depths in a confined setting. They represent intermittent flooding events that were at times turbulent and erosive as indicated by curled flame structures and sharp bedding boundaries (Gillieson 1996). Discrete intervals of massive sand (
