GIS-Based Spatial Analysis for Archaeological Site Prediction and Evaluation

Application of GIS-based spatial analysis for archaeological site prediction and evaluation Predictive modelling of Aboriginal archaeological sites has been a common thread in Victorian archaeological practice for many decades. With the advance of GIS-based mapping methods, it is now possible to employ more sophisticated and automated techniques to characterise the available archaeological data, and to develop analytical methods to assist in the identification, investigation and evaluation of the archaeological record. The use of spatial data is now embedded in the Aboriginal heritage legislation, with specific geographic criteria to determine "Aboriginal cultural heritage sensitivity" as one trigger for a Cultural Heritage Management Plan. By describing the landscape features of recorded cultural heritage sites, criteria for mapping cultural heritage sensitivity have been derived, such as distance from waterways, coastlines, and existing archaeological places, and the presence of sand ridges, swamp margins, and ancestral streams. However, it is often the finer resolution geographic features that are crucial to predicting and understanding Aboriginal archaeology. The influence of micro-topography fine resolution landscape, topographic and environmental factors in campsite selection can be helpful in narrowing down the potential locations of archaeological sites and improving the effectiveness of field survey programs. By testing and characterising the site types, we can better understand the human behavior behind the choices made in selecting campsites, and the way the landscape was used and inhabited by people. For example, along the Maribyrnong Valley, a number of silcrete outcrops have been identified as ancient sources of stone for tool making. The various sites appear to occur consistently at about the 37m contour on slopes greater than 25 degrees. At the same time, nearby occupation sites appear to be restricted to the few areas of flat land less than 15 degrees slope, near the river, or on the more distant high ground on the change of slope at the edge of the valley escarpment. Investigation and refinement of the patterns of Aboriginal site distribution using GIS-based analysis and remotely acquired datasets such as LiDAR can therefore help us better predict where new sites will be found, and understand how Aboriginal people used the landscape. While the current use is broad scale and coarse in its predictive capacity, there is potential for far more targeted use of the techniques if more accurate and specific data on environmental conditions, landform and ethnographic behaviour can be compiled and built into the models. While an intuitive approach to understanding the landscape from a human perspective might still be superior to most algorithms so far developed for predictive modelling of Aboriginal sites, there will be greater roles of GIS in managing cultural heritage in the future.