Paper long abstract:

The use of GIS models of palaeolandscapes to investigate the relationship between past peoples and environments has been an important are of landscape archaeology in recent years. The requirements of data resolution for such GIS models are in part dictated by the nature of the questions being asked and the themes being investigated, but we propose that certain data have inherent spatial and temporal limits of resolution - uncertainty which introduces a 'fuzziness' that is problematic for the production of robust and useful models. In this paper we investigate how the inherent 'instability' of such data including that relating to chronology, hydrology (sea level) and vegetation (derived from pollen and other palaeoenvironmental proxies) is intimately tied to our ability to produce useful 'fine grained' models of palaeolandscapes. We present a series of case studies to illustrate how these issues become magnified as the scale of investigation becomes smaller. Our focus will move through the landscape scale - the inundation of the North Sea basin and the spread of peatlands in east England - through to the local scale - the nature of change on a site specific scale and the perception of such change by humans. We consider some of the methodological and theoretical issues related to the incorporation of such 'unstable' data in GIS modelling and argue that any multi-scalar approach to archaeological landscapes must be tied to the particular, scale dependent limitations of different datasets.