Modeling and Decision-Making

Mathematical and computer-based simulation models have played an important role in managing the Great Lakes; indeed, the Great Lakes are one of the most modeled aquatic systems in the world (Modeling Task Force, 1987). Modeling applications include setting phosphorus limits in the Great Lakes Water Quality Agreement (Task Group III, 1978; Modeling Task Force, 1987), assessing the movement of toxic chemicals into the lakes and through aquatic food chains (Schottler and Eisenreich, 1997; Trudel and Rasmussen, 2001), predicting the impacts of invasive species (Riciardi, 2003), adjusting stocking levels in Lake Ontario (Jain and DePinto, 1996), and evaluating water regulation plans for Lake Ontario and the St. Lawrence River (Manno, 2003; Limno-Tech, Inc., 2005). Models are often used as part of decision support systems (DSS) to clarify processes, simulate change, set targets and goals or choose among management alternatives. Models aid decision-making by providing an approximation of how the system of interest will respond to change. Although models cannot predict the future with certainty, their output can inform and support management decisions by providing estimates of what can be expected to occur under a range of management actions and environmental scenarios (Hall and Day, 1977).

Good models lead to good decisions when they help improve understanding and efficiency, in other words when they increase the likelihood that choices will be made based on the best available science and when they facilitate the selection of policies that achieve environmental goals in the most efficient and effective manner. But even good models can lead to bad decisions when they constrain creativity, preclude options that are too difficult to model, produce predictions that are irrelevant to the goals of the stakeholders, when the geographic or temporal scales of the output fail to match the scale of the processes of most concern, or when the uncertainty bounds are too large or inaccurately communicated. In the past, it may have been good enough for modelers to provide technical support to technocratic managers. However, in an era of increasingly participatory decision-making models can and should provide stakeholders access to data and improve their understanding of it’s meaning, help all participants understand the analysis of the system dynamics incorporated within the model, and in so doing promote meaningful discussion among stakeholders.

(Details of a modeling project will be added this summer.)