Abstract

Slightly raised landforms have greater archaeological potential than their surroundings even in floodplains. In the muskeg plains and plateaux of the northern Canadian boreal forest, sites are almost invariably located on raised landforms such as knolls, ridges, rises, and terraces. Some of these can be very subtle features. Numerous methods have tried to incorporate these landforms into predictive models by identifying them from conventional maps, remote sensing, and GIS. None of these was successful or inexpensive enough to consider for modelling large areas. A method to model for raised landforms developed by Millennia Research works well to identify medium to large landforms, but lacked a sufficiently detailed DEM to model for microtopographic landforms. LIDAR (LIght Detection And Ranging) is a remote sensing method that will revolutionise many areas of archaeological investigation. LIDAR provides data with sub-metre accuracy even under trees. Data collected for every square metre means that incredibly detailed maps and models can be generated over hundreds of square kilometres. Formerly "featureless" landscapes become highly varied. Millennia Research has further developed the landform identification routines and algorithms to work with LIDAR data to identify a wide range of landforms, down to microtopographic features.Years of archaeological research in north-eastern British Columbia has shown hat topographic features can be the single most important component when modelling for archaeological sites, as these are strongly correlated with large and small landforms. Normally, forest cover is used as a proxy to identify these landforms. However, the sole use of forest cover can result in assigning an area with high potential when it is actually low. In this paper I will present a novel method to model for landforms based on the use of rectangular running windows, and present the results of the ground-truthing of the Archaeological Potential Model developed by Millennia Research for north-eastern British Columbia.