Stratigraphy from topography II. The practical application of the Harris Matrix for the GIS-based spatio-temporal archaeological interpretation of topographical data

Traces of human activity preserved in ground surface relief can be documented using airborne laser scanning (ALS). Various visualization techniques for ALS-based digital terrain models help to enhance the micro relief and display abundant information about the earthworks of settlements, pathways, field systems, burial grounds and the like. Such remains can express a complex pattern of intersecting and overlapping relief features produced by millennia of human activity. To ‘read’, or better decipher, this palimpsest or ‘messiness’, archaeological features must be classified, and their temporal relationship determined during interpretative mapping. While much interpretation of relief features is based on the relatively straightforward analysis of parameters like shape, morphology, topographical location or patterning, chronological sequencing of relief features can be very challenging. In this paper we propose a solution for the compilation of relative chronological sequences when mapping relief features from topographic data. We combine an interactive GIS-based archaeological interpretation with the creation of a stratigraphic sequence known as a Harris Matrix, which is extended by an interval-based hierarchical time model. This allows individual features and groups of features to be assigned to userdefined chronological periods and phases. The features extracted from the topographic data are grouped in a final Harris Matrix according to their temporal relations and can be translated into period or phase maps within the GIS environment. The value of this approach is demonstrated in a case study from Lower Austria, a complex archaeological landscape within which more than 1,450 archaeological relief features have been mapped into a coherent spatio-temporal model. The results give a detailed insight into the development of an archaeological landscape over at least 2,500 years, broken down into 10 periods, and have helped to answer specific historical questions. The approach presented here represents a starting point for further targeted analysis and investigation to provide an absolute chronological framework.