Enhanced Visualization of Low Probability Trajectories

Forecast paths for hurricanes are commonly communicated through Cone of Uncertainty (COU) visualizations. An alternative visualization, animated risk trajectories (ARTs), uses dynamic ensembles of instances to convey an array of possible future tracks. Four experiments with simplified scenarios explored how surprising or misleading each form of visualization was for outcomes beyond the most likely forecast path, using participants with little direct experience (Colorado-based participants) versus more familiarity with hurricane risk and forecasts (Florida-based participants). For subsequent landfalls at locations farther from the central projected path, all participant types felt less misled when viewing ARTs rather than COUs. While Colorado-based participants reported less surprise when viewing discrepant outcomes following ARTs, Florida-based participants found deviant outcomes surprising regardless of the visualization used. Results suggest ARTs may help people better understand uncertainty when singular “most likely path” forecasts are wrong, although those expected to have greater experience may benefit less.