Lisa Sattenspiel (University of Missouri)
Amy L. Warren (University of Missouri)
Alan Swedlund (University of Massachusetts)
George Gumerman, III (Santa Fe Institute)
Abstract. The original Artificial Anasazi (AA) model included several different implementations, one of which extended the model’s household level behavior to include individual-level demographic processes such as age-specific fertility and mortality. This individual-level model, which we now call the Artificial Long House Valley (ALHV) model was never fully tested, however, and in addition, it does not generate realistic population projections. In this talk we describe the basic structure of the two models, and show the results of initial sensitivity analyses of the original AA model, a necessary step that has long needed our attention. We go on to show how simulations of the AA and ALHV models differ, and suggest reasons for these differences. We hope to generate discussion that will aid in determining how best to capture complexities associated with individual-level demographic processes so that we can better assess the impact of infectious disease, changes in fertility, and other individual-level factors on the collapse of the population and abandonment of the Long House Valley.