Cartographic Design for Improved Decision Making: Trade-Offs in Uncertainty Visualization for Tornado Threats

At present, the National Oceanic and Atmospheric Administration is developing new technologies that could offer explicit estimates of the probability that a thunderstorm could produce a tornado up to an hour ahead of the event. Such technologies could radically alter how risk spaces are represented and understood by those who must decide whether or not to take protective action. In addition, there are relatively few studies that examine mapped representations of uncertainty in weather information or the influence of this uncertainty information in weather hazard decision making. To address these gaps, this study presents research subjects with a variety of representations of uncertainty based on the precepts of cartography and information visualization. We propose and test for the existence of three geospatial framing effects that potentially influence subjective estimates of risk: distance from a hazard, warning boundary inclusion or exclusion, and symbolic color coding of uncertainty information. Using a series of computer-aided geographic experiments with a large sample (N = 5,564) of the U.S. population, we find evidence for the existence of each of the three proposed geospatial framing effects. Two of these framings are controlled by the mapmaker—in this case, the weather forecaster—and thus should be considered during the development stages of new products. We discuss the practical implications of the experimental study for current and future tornado warning practices. Key Words: cartography, risk, tornado, uncertainty, visualization.