Comparing Effectiveness of Four Graphical Designs for Probabilistic Hazard Information for Tornado Threat

Abstract

Finding effective measures to decrease the nation’s loss resulting from severe weather is crucial. Probabilistic Hazard Information (PHI) system is being developed for severe weather forecasting. Little work has been done to investigate which kind of graphical designs for PHI can increase users’ perception and interpretation, and elicit the best user response to the threats. This study investigates the effects of varying graphical designs for tornado threat. Four designs, “four colors”,” red scale”, “gray scale” and “contour” were tested. Radar was provided for one set of treatments, the other without radar. Analysis showed response time significantly decreased for designs without radar. “Without radar” displays helped participants make significantly more accurate decisions than “with radar” displays. Participant’s preference, the “four colors” design, was the most effective choice among “without radar” displays. For “Without radar” displays, “gray scale” design results in higher accuracy for interpreting threat information and choosing appropriate responses.