Phytoliths, Late Quaternary Environments and Archaeology in Tropical Semi-Arid Northwest Australia

Abstract

As a consequence of poor preservation and a lack of depositional sites, palaeoecological records are almost entirely non-existent for the tropical semi-arid region of northwest Australia. Subsequently, our knowledge relating to late Quaternary environments and the interrelationship between people and the northwest landscape is impoverished. This thesis presents the results of investigations of phytoliths in archaeological and other late Quaternary sediments in the Kimberley region as a means of partially rectifying this problem. Three specific research goals were addressed. Firstly, to examine the nature of phytolith production in extant Kimberley flora. Secondly, to examine the nature of phytolith assemblages preserved in non-archaeological sediments. Finally, to apply the derived knowledge to the study of prehistoric plant exploitation and palaeoenvironment at an archaeological site.  In order to examine aspects of variability in phytolith production by modern plants, a comparative reference collection of 338 specimens (representing 54 families) was constructed. This revealed a range of patterns, from a total absence of phytoliths through to the production of massive quantities, across both monocotyledonous and dicotyledonous plants. At the family level these results are largely in line with those from similar studies conducted elsewhere. Sufficient morphological variation has been demonstrated to exist to enable the use of phytoliths as a microfossil system for broad-level vegetation reconstruction in the savannah regions of northern Australia. Whether phytoliths provide a sufficient level of detail to address specific archaeological research questions is an issue that requires further investigation. A number of important economic plants in the Kimberley have been shown to produce distinctive phytoliths. However, in many more phytoliths are absent, or only redundant types are produced.  Following the baseline floristic study, the focus of the research then shifted to the analysis of phytoliths recovered from modern sediments associated with a range of different ecological and vegetational settings. The results indicate there is a degree of homogeneity between samples, reflecting the dominance of grasses in most vegetation communities of the Kimberley. However, a more positive outcome is that it is possible to identify a number of distinct vegetation communities and environments on the basis of their associated sedimentary microfossil assemblages. Additionally, the examination of phytoliths from a small number of mud nests and tufa formations demonstrates the potential of such novel sediment sources to serve as chronological and environmental data traps in this region.  The application of phytolith analysis to an archaeological site completed the study. Carpenter’s Gap 1 is a rockshelter located along the Napier Range in the inland southwest Kimberley, with an occupation sequence spanning c.43,000 years. A detailed assessment of the formation processes at the site suggests the phytolith assemblage contained therein comprises primarily of two components, a largely naturally derived grass element and a predominantly culturally derived non-grass element. The phytolith record is interpreted as indicating that significant changes have occurred in both the local vegetation, and people’s behaviours, throughout the period of occupation of the site. The presence of certain grass and palm phytoliths at the base of the site suggests there had been a greater availability of water in the landscape c.43,000 BP. By 34,000 BP the local grassland vegetation had apparently shifted to a spinifex-dominated community, probably as a result of increasing aridity. Additionally, by the onset of the Last Glacial Maximum (LGM) palms had disappeared from the sequence, providing further evidence of decreased precipitation in the site vicinity. However, despite the wider climatic and vegetation shifts, the presence of vine thicket type phytoliths throughout the sequence indicates the maintenance of such communities along the range at all times until at least 17,000 BP, although the extent of these patches cannot be gauged. These vine thicket patches would have provided a valuable focus for Aboriginal plant exploitation activities, as would have (semi-) permanent water sources represented by Cyperaceae type phytoliths. The occurrence of substantial quantities of sponge spicules and diatoms in the LGM levels of the site are interpreted as representing a need of people to transport water to the site in response to environmental stress associated with increased glacial aridity. The evidence from the archaeological phytolith analysis was examined within the wider regional context and found to fit reasonably well with the other evidence for vegetational and climatic change in the area.  Overall, the study has demonstrated the suitability of phytolith analysis to questions of palaeoenvironmental interest in the tropical semi-arid areas of northern Australia. The application of the technique to Carpenter’s Gap 1 further demonstrates the potential of the approach in the archaeological discipline, providing researchers with glimpses into aspects of human plant exploitation. However, much remains to be discovered about the usefulness of phytolith analysis for addressing specific research questions in Australian archaeology and broader issues in palaeoenvironmental studies, and this thesis serves as a basis for such future studies in the region.