Unlocking the Kimberley’s past: the applicability of organic spring deposits for reconstructing late Quaternary climatic and environmental change

Abstract

There are limited high-resolution records of climatic and environmental change from the Kimberley region of northwest Australia. This has hindered the development of knowledge of climate and environmental change in Australia’s monsoonal tropics, and the ability to provide context for the area’s rich archaeological record and globally renowned rock art. The lack of high-resolution records from this region is primarily a result of the monsoonal climate which limits the presence of “classic” palaeoenvironmental archives such as perennial lakes and wetlands. Organic spring deposits are unconventional archives of past environmental change yet offer potential to provide outstanding records in arid and semi-arid regions such as the Kimberley. Despite this, the majority of existing research demonstrates complications with their use, in particular the application of standard radiocarbon (14C) techniques to build robust chronologies of their development. Because of the importance of springs as critical palaeoenvironmental archives, and the pressing need for high-resolution records from northwest Australia, this thesis utilised three organic spring deposits to develop new, high-resolution records of climate and environmental change for the Kimberley. These records span the last ~14,500 years, and were underpinned by a new protocol for establishing robust chronologies from these settings which was developed via a rigorous geochronological investigation.Initial 14C results from a core collected from Black Springs in 2005 were confusing, with convoluted ages below 94.5 cm limiting the reliability of an initial palaeoenvironmental reconstruction >9070 cal. yr BP. A new protocol for obtaining robust chronologies from organic spring deposits was therefore developed for cores collected from the three sites in 2015, utilising multiple geochronological methodologies including 14C dating of different carbon fractions (stable polycyclic aromatic carbon (SPAC), macro-charcoal, pollen concentrate, bulk sediment and roots), 210Pb dating, the application of 239+240Pu, and high-spatial resolution, luminescence techniques (natural sensitivity-corrected luminescence (Ln/Tn)). Whilst Ln/Tn demonstrated that the organic spring deposits contained a relatively uncompromised stratigraphic record, 14C dates were contaminated by roots, groundwater fluctuations, and incorporation of allochthonous “old” carbon. SPAC isolated by the hydrogen pyrolysis (HyPy) pre-treatment appeared to remove post-depositional sources of alteration, and therefore provided a viable approach for constructing chronologies in these settings. Developing more recent chronologies (i.e. the past 100 years) using 210Pb was found to be problematic due to the behaviour of the springs as an open system with regards to uranium, suggesting this may not be possible in many spring environments. Whilst this investigation was made with respect to springs, the contamination pathways are not unique to these settings and therefore the chronological developments are more widely applicable (e.g. for constructing chronologies in archives subject to deep root growth and/or significant changes in hydrology).This new dating protocol enabled the climatic and environmental reconstructions from the three new records presented in this thesis to be placed on secure timescales. These new records utilised pollen, micro-charcoal, non-pollen palynomorphs (NPPs) and elemental geochemistry (Itrax micro X-ray fluorescence (μXRF) core scanning). Principal Components Analysis (PCA) axis 1 scores demonstrated similarity between inter-site elemental geochemistry and non-pollen palynomorph datasets, although there were significant discrepancies between the pollen records which reflected predominantly local conditions. In order to isolate dominant modes of regional variability, Monte Carlo Empirical Orthogonal Function (MCEOF) analysis was applied to proxies from the geochemistry and NPP datasets, providing a regional reconstruction of climatic variability in the Kimberley for the last 14,500 years. The results of this reconstruction show increasing monsoon activity from 14,500 cal. yr BP consistent with the deglaciation, followed by peak monsoon activity during the early-Holocene (between ~11,000 – 7500 cal. yr BP). There is a trend of decreasing monsoon activity throughout the remainder of the Holocene, interrupted by a short period of wetter conditions at ~4200 cal. yr BP, and peak aridity spanning ~2600 – 1000 cal. yr BP.This thesis makes a significant contribution to understanding of climatic change in Australia’s monsoonal tropics, in particular long-term changes in monsoon activity in northwest Australia since the Last Glacial-Interglacial Transition (LGIT). It also provides a method for building reliable chronologies in organic spring deposits which are critical palaeoenvironmental archives in arid and semi-arid environments. In light of the Kimberley’s rich archaeological record, requirements have been highlighted for the need to extend high-resolution palaeoclimate reconstructions in the region to ~50,000 cal. yr BP to encompass the entire known history of likely human occupation in the region.