Introduction to Uncertainty and Sensitivity Analysis in Archaeological Computational Modeling

Abstract

Computational modeling in archaeology is approaching its 50th anniversary. Despite its longevity, it can be argued that this approach has not matured accordingly, in terms of its contributions to theory building and research design, or general archaeological inquiry. Further, this approach has not undergone the same type of rigorous, self-reflective testing as similar computational approaches have in related fields. Symptomatic of this situation is the short shrift given to model validation and verification in archaeology, particularly in terms of the origins, nature, and magnitude of embedded uncertainties. However, other disciplines routinely apply various forms of sensitivity analysis (SA) as a systematic component of research design that evaluates such uncertainties. Here, the role of and approaches to computational modeling in archaeology is explored. Use of such modeling validation techniques in related fields that target geologic, ecosystems, and social modeling are presented in order to identify translatable insight regarding the detection, acknowledgement, incorporation, and mitigation of uncertainty. The interface between model equifinality and uncertainty are then addressed, with the overall goal of determining the utility of SAs for archaeological computational modeling, as a way to identify model strengths and weaknesses. This synopsis acts as a benchmark for various exemplars of SA in archaeological computational modeling.