Modelling Archaeological potential with GIS in Northern British Columbia

Abstract

The use of site location models, whether internal intuitive or external formalized, is inherent in archaeological site prospection, discovery and management. The widespread use of Geographic Information Systems (GIS) for building archaeological site location models in the context of Cultural Resource Management (CRM) has reignited criticism and debate regarding their appropriate construction and use. This thesis demonstrates a correlative approach to modelling archaeological site location that is clear to follow, based on robust spatial statistics (Weights of Evidence) and is independently tested for performance. This thesis is important to the modelling endeavour in the Canadian boreal forest in that it can legitimately make these three claims. Additionally, the recognized importance of trails to archaeological site location is examined through an example of how least-cost paths may be used to fill-in for missing traditional trails data in the modelling endeavour. The model created in this thesis was successful with an overall performance of 90% of known sites in 41 % of the study area, modelled as moderate and high potential. The high potential areas scored a very high Kvamme Gain Statistic of 0.85. Also, the least­cost paths demonstration led to the consideration of an evidence category not previously entertained in Canadian boreal forest modelling programs; that of lakes containing aquatic resources.