S.G. Webb⁴, P. Clark⁵, Johnston, H⁶

1. Department of Geology, The Faculties, ANU. Jwmagee@geology.anu.edu.au

Lunettes have long been known to exist along the remote downwind north eastern and northern shoreline of Lake Eyre North, between the mouth of the Cooper Creek and Lewis Bay in the north west corner of the lake. Dulhunty (1983) was the first to map the distribution of lunettes and examine them on the ground. He provided a number of surveyed cross sections and noted an average height of 35 m above the lake shoreline and surveyed a maximum height of 48 m. He noted that the lunettes were currently being eroded, and that their structure was complex, with cores consisting of pale yellow-brown argillaceous sands. He generally recognised the presence of a single lunette, which diverged to two just south east of the Warburton Creek mouth.

The current study was initiated to analyse the deflation episodes of Lake Eyre to better define the stratigraphy and chronology of those parts of the palaeohydrologic record associated with low lake levels and aridity. Previous attempts to study these features on the ground had been thwarted by the difficulties of access, and a perceived lack of good exposure of the internal stratigraphy. However, low-level aerial reconnaissance indicated that the area between the mouths of Cooper Creek and Kalaweerina Creek had multiple parallel lunettes with relatively good deflation exposures of the internal stratigraphy. Additionally, access was deemed to be possible along the margin of the elongate inlet on the north side of the Cooper mouth, which extends about 30 km into the Tirari Desert dunefield to the east (‘North Harbour’). This area was visited in 1998 and 1999 with the expectation that the multiple lunette phases represented deflation episodes of different ages.

The area was found to be more stratigraphically and topographically complex than expected. Up to 3-4 multiple lunettes, which trend NNE and occur as isolated remnants with deep deflation exposure, are embedded in a younger longitudinal dunefield with a more northerly trend. The original lunettes appear to have varied in height, and perhaps width, along their length and generally only the eroded remnants of the higher portions remain as isolated inliers within the dunefield. The presence of a number of playas, including inlets from North Harbour, the Horseshoe Lake (which extends south of the Kalaweerina mouth) and a number of isolated playas within the dunefield, adds further complexity.

Occurring commonly at the margins of the playas, particularly the Horseshoe Lake, there is a bench of grey-brown sediment with a relatively flat surface, which underlies both the lunettes and the dunefield. The only stratigraphic sections of this unit were some gully exposures of up to about 2 metres which showed a pedogenically modified zone, with relatively abundant secondary displace gypsum. Unmodified sediments, indicative of depositional origin, have not been examined. The flat low-gradient topography and elevation (- 5 m AHD) suggest either deposition as a fan delta during a high-lake phase or as vertical accretion deposits on a low-gradient alluvial fan during a low-lake phase. The secondary gypsum zones might contain either or both groundwater deposited displacive gypsum and pedogenically translocated gypsum. Genyornis eggshell and marsupial skeletal remains were both found in the upper portion of the unit.

The eroded lunette remnants often expose up to 6-8 m of fresh sediment, from the core of the dune, on the floor of deflation hollows and on the vertical sides of residuals. The stratigraphy is essentially the same in all of the multiple lunettes with two units of slightly clayey yellowish-grey sands occurring in virtually all deflation exposures. Primary sedimentary lamination in the form of long sub-horizontal sets, typical of lunette sediments, is poorly preserved in both units. The upper 0.5 — 1.0 m of the lower unit usually contains small (0.5 — 1.0 cm diameter) pedogenic rhizonodules, which are gypseous and also contain some carbonate. The rhizonodules are usually moderately abundant but vary from sparse to very abundant. The lower unit is truncated at its upper boundary, commonly at the top of the rhizonodule-zone. The upper unit is lithologically very similar, but does not contain rhizonodules. It has a well-defined clay-rich horizon up to 1.0 m thick in its upper portion which is probably due to pedogenically translocated clay. This horizon often forms an erosion-resistant bench and is characterised by abundant rilling. The upper unit is probably truncated at the top of this horizon at some sites, though exposure of the upper part of the unit is often very poor, and its relationship to the overlying sediment is not clear. Separating the two units is a thin red-brown clay-rich unit which is absent in some sites and varies from a few cm to 20 cm in thickness. The unit has sharp erosional upper and lower boundaries and clearly has a depositional origin, but its environmental implications are not yet clear. Thin sections are under preparation but not yet available. The upper lunette unit is overlain by a top unit of poorly consolidated pale-yellowish medium sand, with the boundary between the units rarely exposed. The lower unit contains abundant Genyornis eggshell and occasional emu eggshell, which are patchily distributed; marsupial and Genyornis skeletal remains also occur. The upper unit contains rare eggshell, including Genyornis. The top unit is thought to relate to the longitudinal dunefield and contains only emu eggshell.

Genyornis eggshell from upper part the basal grey-brown gypseous bench, has a D/L ratio of 0.98, indicating a stage 5 age. D/L ratios of eggshell from the two primary lunette units indicates that both are equivalent to the Williams Point aeolian unit (Magee and Miller, 1998) and date to the period 50-60 ka. The lower portion of this unit at Williams Point is dominated by sediment derived from a regressive beach phase, containing quartz, ooids and biogenic carbonates, and the upper part is dominated by sediment derived from playa deflation, with abundant clay pellets and gypsum. The upper part of this depositional phase was previously recognised by Magee and Miller (1998) as associated with the excavation of the current Lake Eyre playa basin. We relate the upper and lower units at North Harbour to the upper and lower parts of the Williams Point aeolian unit These latest results, from North Harbour, indicate the formation of multiple large lunettes at this time and support the interpretation of this phase as a major deflation episode. The stratigraphic complexity, in the North Harbour lunettes, indicates a more complex depositional history through this stage, than was evident at the Williams Point section (Magee et al, 1995). This more complex depositional sequence has also been noted in lunettes of Williams Point aeolian unit age on the eastern margin of Jackboot Bay.

A whole but partially fractured emu egg was found in aeolian sediment on the western side of the Horseshoe Lake and was retrieved with a dense mass of infilling sediment. With a D/L of 0.48, which indicates an age between 25 and 30 ka, the egg provides an excellent opportunity for a four-way dating comparison between AAR, AMS ¹⁴C and U/Th on the eggshell and OSL on the contained sediment.

Until now, megafaunal remains from the Lake Eyre region, with the exception of Genyornis eggshell, have been found reworked in fluvial sediments of greater than 60 ka age. In addition to abundant eggshell, the North Harbour lunette sediments contain Genyornis and marsupial megafaunal skeletal remains. The most spectacular find was an articulated Diprotodon cranium and mandible, almost 1 metre in length, in the upper portion of the lower lunette unit, though poorly preserved. These remains at about 55ka represent the youngest marsupial megafauna yet recovered from the Lake Eyre region. This Diprotodon will be dated by an OSL date from the surrounding sediment, an ESR date on tooth enamel, U-series on bone (by TIMS and Ge Gamma Spectrometry), and TIMS U series and AAR on stratigraphically associated Genyornis eggshell. A Procoptodon mandible was also found in a similar stratigraphic setting.

Two large meteorites were found lying on the surface of deflation hollows. These were both chondritic meteorites and found in a similar stratigraphic setting about 5 km apart. Thin-section analyses indicate they are not related.