Year
2003
Degree Name
Doctor of Philosophy
Department
School of Geosciences - Faculty of Science
Recommended Citation
Ward, Ingrid A, Hidden in the sands of time: geoarchaeology of sandstone landscapes in the Keep River region, Northern Territory, Australia, PhD thesis, School of Geosciences, University of Wollongong, 2003. http://ro.uow.edu.au/theses/352
Abstract
This geoarchaeological study aims to establish the geomorphic context of Aboriginal cultural landscapes and archaeological sites in the Keep River region, Northern Territory, over the Late Quaternary. The geomorphic focus of the thesis is concentrated on the sand sheets, which occur at the base of the sandstone escarpments. Sample locations include the occupation (rockshelter and sand sheet) sites of Goorurarmum, Jinmium and Karlinga, and non-occupation(creek) sites at Karlinga and Sandy Creek Gorge. The thesis presents five interrelated studies, including (i) an assessment of the theoretical relevance of geoarchaeology in northern Australia; and three studies at different timescales, evaluating (ii) long-term landscape processes over timescales of millenia using in situ osmogenic dating, (iii) sedimentary processes over timescales of centuries to millenia using luminescence dating, and sedimentary processes over decadal timescales. A fifth study integrates the results of the above four studies with existing archaeological data. Measurement of in-situ cosmogenic 10Be and 26Al concentrations from the local escarpment bedrock has revealed denudation rates of 4 - 7 mm.ka-1 over 103 – 105 year timescales, consistent with similar studies in other parts of semi-arid Australia. Calculated bedrock denudation rates were used to model burial profiles up to 6 m deep from the Jinmium sand sheet. Measured concentrations of 10Be and 26Al in two profiles, provided vertical accretion rates of ~10 - 20 mm.ka-1 over the past few hundred thousand years. Grain size, micromorphology, mineralogy and geochemistry indicate that the sand-sheet sediments are locally sourced. The rock-shelter sediments have higher relative concentrations of CaO, P205 and greater LOI, than the surrounding sand-sheet sediments, reflecting higher levels of charcoal, guano and organic matter. Post-depositional reddening of the sediments reflects groundwater variation rather than any proxy for depositional age. A total of 33 thermoluminescence (TL) and 15 optically stimulated luminescence (OSL) dates were obtained from the rock-shelters, sand sheets and creek embankments. U- and Th-series analyses indicate relative equilibrium, with slightly higher dose rates (3.0 ± 0.83 Gy.ky-1) for mottled sediments than elsewhere (1.3 ± 0.29 Gy.ky-1). Foreshortened TL plateaux in some sand sheet sediments at Goorurarmum and Jinmium, and stepped TL plateaux in the sediments alongside Sandy Creek are indicative of episodic rapid deposition events. Basal OSL ages for the Goorurarumum rock shelter and adjacent sand sheet excavation are 0.3 ± 0.07 ky BP and 14.3 ± 0.4 ky BP respectively, and near-basal OSL ages for the Karlinga rock shelter and more distant sand sheet excavation are 18.4 ± 5.4 ky BP and 18.0 ± 0.6 ky BP respectively. The 18 ka age for the Karlinga rock shelter is uncertain, and a much younger basal radiocarbon age of ~ 4 ka is preferred. The deepest OSL ages for the Karlinga creek profile (KR99CP) and Sandy Creek Gorge profile is 8.6 ± 0.3 ky BP and 13.9 ± 0.4 ky BP respectively. An increase in the sand sheet accumulation rates from ~ 100 mm.ka-1 in the late Pleistocene to over 200 mm.ka-1 in the Holocene is interpreted to reflect enhanced monsoonal activity following postglacial marine transgression rather than human activity. The preceding results are integrated with information from radiocarbon analyses, regional archaeology, and local palaeobotanical, stone artefact and rock-art studies. Chronological results support initial evidence of seed processing, stone point production, and rock art around 3 – 4 ka. However, it is questioned whether these trends reflect ‘intensification’ or preservation and sorting of the archaeological records. The combined chronological data from the Keep River region indicates that although there is evidence of occupation from the Last Glacial Maximum (LGM), the majority of occupation deposits in the Keep River region are Holocene in age. However, this is biased by the greater number of rock shelter sequences sampled and by the poor preservation potential of the semi-arid monsoonal climate which favours younger sediment sequences. Better preservation of older sequences is found in open sites. The overall geoarchaelogical history describes a relatively insensitive sedimentary record of the comparatively sensitive Aboriginal occupation of the sandstone landscapes in the Keep River region. Scope remains for further multi-disciplinary geoarchaeological nvestigations in similar semi-arid monsoonal environments.
Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.