Dating megafaunal extinction on the Pleistocene Darling Downs, eastern Australia: the promise and pitfalls of dating as a test of extinction hypotheses

A key to understanding Late Pleistocene megafaunal extinction dynamics is knowledge of megafaunal ecological response(s) to long-term environmental perturbations. Strategically, that requires targeting fossil deposits that accumulated during glacial and interglacial intervals both before and after human arrival, with subsequent palaeoecological models underpinned by robust and reliable chronologies. Late Pleistocene vertebrate fossil localities from the Darling Downs, eastern Australia, provide stratigraphically-intact, abundant megafaunal sequences, which allows for testing of anthropogenic versus climate change megafauna extinction hypotheses. Each stratigraphic unit at site QML796, Kings Creek Catchment, was previously shown to have had similar sampling potential, and the basal units contain both small-sized taxa (e.g., land snails, frogs, bandicoots, rodents) and megafauna. Importantly, sequential faunal horizons show stepwise decrease in taxonomic diversity with the loss of some, but not all, megafauna in the geographically-small palaeocatchment. The purpose of this paper is to present the results of our intensive, multidisciplinary dating study of the deposits (>40 dates). Dating by means of accelerator mass spectrometry (AMS) 14C (targeting bone, freshwater molluscs, and charcoal) and thermal ionisation mass spectrometry U/Th (targeting teeth and freshwater molluscs) do not agree with each other and, in the case of AMS 14C dating, lack internal consistency. Scanning electron microscopy and rare earth element analyses demonstrate that the dated molluscs are diagenetically altered and contain aragonite cements that incorporated secondary young C, suggesting that such dates should be regarded as minimum ages. AMS 14C dated charcoals provide ages that occur out of stratigraphic order, and cluster in the upper chronological limits of the technique (∼40–48 ka). Again, we suggest that such results should be regarded as suspicious and only minimum ages. Subsequent OSL and U/Th (teeth) dating provide complimentary results and demonstrate that the faunal sequences actually span ∼120–83 ka, thus occurring beyond the AMS 14C dating window. Importantly, the dates suggest that the local decline in biological diversity was initiated ∼75,000 years before the colonisation of humans on the continent. Collectively, the data are most parsimoniously consistent with a pre-human climate change model for local habitat change and megafauna extinction, but not with a nearly