Evaluating implicit visualization of uncertainty for public policy decision support

Decision makers increasingly rely on science to inform public policy decisionmaking. Although this integration of science and policy offers the potential to support more informed decisions, scientific results are often not provided in a manner usable to decision makers. When faced with highly uncertain conditions, such as climate change, communicating science in a usable manner becomes even more important. In decision support settings, visualization of geographic information offers a powerful means to communicate uncertain science to decision makers. However, building believable representations does not provide a complete understanding of the potential consequences of decisions. Developing uncertainty representations to reflect the processes of decision-making under uncertainty offers a means to provide insight into the relationships between decisions, uncertainty, and outcomes (consequences of policy decisions). Yet, visualizations often avoid the explicit inclusion of contextual information, such as explanations of risk and uncertainty. This research makes a distinction between explicit and implicit uncertainty for visualization in decision support. In explicit visualization, uncertainty is conceived of, and evaluated as, unique information, related to, but not the same as, the underlying data. Implicit visualizations embed uncertainty information into the representation, instead of expressing uncertainty as separate or additional data. When reframing uncertainty in this way, the relationship between uncertainty, outcomes and decisions is emphasized over explicit representation frameworks that dissociate the method from the user. This paper presents an implicit method for visualizing the impact of climate change uncertainty on policy outcomes in a water model for a hypothetical metropolitan area. The effectiveness of this method for visualizing the relationship between uncertainty and policy impacts was evaluated through a human subject test. The paper reports on the results of the pilot study and how this method compares to methods for explicitly visualizing uncertainty.