Abstract/Summary

The investigation of palaeofire provides the basis for testing the reliability of recognised controls on fire in the present day, but in environmental and social conditions that differ markedly. Whilst present day analysis can make use of comprehensive data at high spatial resolution, palaeodata enables the investigation of controls both spatially and temporally, on time scales that are relevant for longer term processes. Although there are exceptions, the majority of palaeofire research has tended to analyse relationships by visually assessing the synchronicity of time series, and with a focus on a single or limited selection of controls. In this thesis, a number of statistical approaches are tested to identify relationships between palaeofire and its drivers, making use of diverse data proxies. The relationship between people and palaeofire was first investigated in Iberia during the Holocene. Results from superposed epoch analysis (SEA) show no clear relationship between periods of high population growth or on the time-transgressive spread of farming and fire. To extend this research to the European scale, and to ensure that controls representing humans, vegetation and climate were included in a holistic analysis of palaeofire at a site-level, a new statistical method, with broad-scale applicability, was developed to reconstruct tree cover using sedimentary and modern pollen data. At an aggregate level, the resulting reconstructions broadly matched those published in the literature. These reconstructions were then included within cross-cutting analyses using regression techniques, Granger causality and SEA. Limited evidence of consistent relationships were found at the scale of analysis, emphasising the challenges of drawing regional-level conclusions, and helping to explain contrasting findings from the research literature. Overall, despite the challenges associated with palaeodata, this research highlights the need for statistically robust, multi-variable analyses of the drivers of palaeofire to improve our understanding of fire in the past, present and future.