Human—environment interactions in Australian drylands: exploratory time-series analysis of archaeological records

Abstract

Exploratory time-series analysis of radiocarbon data from archaeological contexts is used to reconstruct the population history of arid Australia, allowing this to be read in concert with records of climatic variability over the last 20 000 years. Probability distribution plots of 971 radiocarbon ages from 286 sites in five dryland regions (the arid west coast, Pilbara and Murchison, Nullarbor, arid interior and the southeastern arid zone) provide a proxy record of prehistoric population fluctuations in these areas. There is regional variation, but the radiocarbon density plots suggest a step-wise pattern of population growth and expansion, with significant thresholds at 19, 8 and 1.5 cal. kyr BP. Within this, the plots suggest a saw-tooth pattern of rapid population growth and decline on a 1—3 kyr frequency, with a marked collapse of dryland hunter-gatherer populations around 3—2.5 cal. kyr BP affecting most regions. Comparison with climate data shows broad correlations with past temperature and rainfall variability, sea-level change and ENSO activity, but the interaction of prehistoric populations and these environmental changes is not well resolved. High amplitude environmental changes appear to have triggered stadial changes in population, rather than smooth transitions. Dryland populations may also have become more sensitive to small environmental changes in the late Holocene, as population density increased. A large increase in population around 1.5 cal. kyr BP is associated with small changes in regional palaeoecology, which are not otherwise represented in palaeoclimatic data sets. Spectral analysis identifies two cyclical periodicities of 1340 and 175 years within the population histories, also suggesting responses to millennial and submillennial climatic variability, a pattern most marked in the late Holocene.

The Impact of Enso in the Atacama Desert and Australian Arid Zone: Exploratory Time-Series Analysis of Archaeological Records

“…Similarly, Cosgrove et al (2007) argued that ENSO and its effects on fire regimes and vegetation were important catalysts in allowing permanent Aboriginal occupation of tropical rainforests in northeastern Australia. The present study develops ideas and approaches in Smith et al (2008). In that study, we (AW and MAS) used time-series analyses of radiocarbon data to identify the impact of ENSO in Australian drylands.…”

“…The method we use here treats radiocarbon density as a proxy for population history, on the premise that the number of radiocarbon ages is a measure of occupation intensity. In practice, a number of other factors also determine radiocarbon density distributions: the size of the dataset; the reduced visibility of older sites; the geomorphic history of sedimentary units; differential preservation of organics; and the selection of samples for radiometric dating (see Smith et al 2008 for a review of methodology). The effects of these can be estimated and offset to some extent, but uncertainty must remain about the robustness of radiocarbon data as a proxy for trends in prehistoric population.…”

“…Figure 6 breaks this down by region, concentrating on the past 15,000 years. Overall, there is a stepped pattern of population growth, with major steps at 8 and 1.5 cal kyr BP, though these thresholds are not uniformly expressed across the arid zone (see Smith et al 2008 for more details). Within this pattern, our time-series data show a distinct saw-toothed pattern of population expansion and collapse, with a sharp decline in population around 3 cal kyr BP evident across all regions.…”

Climate and Demography in Early Prehistory: Using Calibrated14C Dates as Population Proxies

“…In addition, it would be desirable to develop more rigorous means of statistically evaluating the properties of the summed probability distribution frequencies, with respect to both their departure from randomness and their correlation with climatic proxies [see M. A. Smith et al (2008) for an attempt]. However, many climatic proxies may not be perfectly suitable when the goal is to investigate the relationship of human demography to the environment.…”

“…Taphonomic and sampling concerns have to be, and can be, addressed case by case (e.g., M. A. Smith et al 2008), and care has to be taken when interpreting summed radiocarbon dates (e.g., Buchanan et al 2008). Although the method as a whole remains "explicitly exploratory" (M. A.…”