A FRAME-BASED PARADIGM FOR DYNAMIC ECOSYSTEM MODELS

By linking the ecological concept of a system state with the AI constr uct of a frame, one obtains a new paradigm for constructing models of ecosystem dynamics. The key aspect of this paradigm is that it partiti ons the temporal dynamics of the system; only one simple model is oper ational at any time. When the assumptions underlying that model are vi olated, a demon invokes rules for switching to a new frame and, hence, new model. A model to investigate interactions among rainfall, elepha nts, and fire in a Brachystegia woodland in Zimbabwe illustrates this approach. It shows how the paradigm provides a structure that simplifi es the construction of a model. In particular, the introduction of qua litative variables, such as cool or hot fires, offers a functional app roach that facilitates model construction. The example shows how a rel atively simple model can lead to complex and realistic results, with u nanticipated and interesting implications for management. This paradig m could lead to better communication between scientists and resource m anagers. It could also provide an appropriate syntax for modeling, at a certain level of resolution, the effects of climate change.