Vegetation cover was highly dynamic over these periods, responding to broad controls like climate change, soil development and postglacial species migration and establishment. It is therefore an interesting but complex period to understand human responses to this varied and changing landscape. When the first anthropogenic impacts on vegetation appear is a long-standing topic of interest and debate, particularly where archaeological evidence is absent. Although burning, erosion and woodland opening are the main palaeoecological indicators used to infer disturbance, there are no unambiguous indicators of human disturbance (Tipping 1994). Natural factors and deliberate modification can produce the same signal in the palaeoecological record, for example charcoal associated with short-lived woodland canopy reduction may originate from enhanced flammability of biomass under climatic aridity or through deliberate fire-setting.
In Scotland, emphasis has focused on fire as a potential indicator of Mesolithic activity. Although Mesolithic fire-setting has been proposed (Hulme and Shirriffs 1986), separating human and natural ignition at single sites is difficult (Caseldine 1980; Milburn 1997). Several analyses have focused on multi-site comparisons, including sites in neighbouring Fife, to differentiate climate-driven burning, which may be expected to be more widespread and synchronous in flammable vegetation types, from human modification, which may be more localised and diachronous in nature. Multiple sites show elevated charcoal abundance in the late glacial (Upper Palaeolithic) and the later Mesolithic (around 7000 cal BP) (Edwards and Whittington 2000; Edwards et al 2000; Tipping and Milburn 2000). Support for human ignition of fires remains uncertain in most cases, however, with the authors emphasising the need to interpret fire signals in the context of climate regimes. Most analyses in Perth and Kinross have been single site studies, and authors repeatedly recognise that human impacts may be within or below the range of variability of natural disturbances such as erosion and canopy opening (Tipping 1995; Milburn 1997; Edwards and Whittington 1998a). As a result, evidence to support or refute Mesolithic impacts on vegetation cover remains equivocal. These issues remain relevant into the Neolithic, when causality for burning remains unclear, particularly at sites where cereal pollen is absent (eg Milburn 1997).
Research Priorities
Priority 1: Make fuller use of advances in modelling to understand landscapes of the Neolithic. Uncertainties about the extent and palynological visibility of landscape change during this period can be further explored using pollen modelling to test ideas about the spatiality of the Neolithic. As has been shown by work from the Neolithic ‘timber hall’ at Crathes in Aberdeenshire (Tipping et al 2009b) and the earliest clearances and elm decline in central Ireland (Caseldine and Fyfe 2006). These scenarios help to facilitate dialogue between palaeo and archaeological communities, and can be used to communicate landscape dynamics and the challenges involved in reconstructing our past to wider, non-specialist audiences (Bunting et al 2018).