Graduate Research and Discovery Symposium (GRADS)

Advisor

Donald House, PhD

Document Type

Poster

Department

Computer Science

Publication Date

Spring 2013

Abstract

Although the past 30 years have seen major advances in the scientific understanding of hurricane forecasting, there has been a lack of systematic research on people’s comprehension of displays used to show these forecasts. A primary visual aids is the error cone. The center line represents the predicted hurricane track for a five day period. The width of the cone is determined by considering historical forecast errors over a five year sample, and represents a 67% likelihood region for the hurricane track. A primary challenge of this model is that that most people have difficulty in understanding the probabilistic concepts that are used to communicate uncertainty. For example, it tends to give the impression to those inside the cone that they have an exaggerated chance of being in the hurricane's path, while those outside of the cone tend to feel a false sense of security. We have developed a new method of visualizing the possible projected paths of hurricanes using the projected path of a given hurricane as well as the historical data of previous hurricanes. The goal is to maintain a display that shows a range of possible outcomes, while maintaining the statistical characteristics of the error cone.

Download
COinS