User understanding of hurricane wind potential graphics

As part of a project to help NOAA communicate hurricane wind potential, forty participants were tested on their ability to understand a graphical hurricane forecast. The participants viewed the graphics on computer screen in the eye-tracking laboratory on the Mississippi State University campus. The sample consisted of 20 members of the local public, 9 experts, and 11 students. The graphic depicted a fictitious Hurricane Inga that was forecast to make landfall along the North Carolina Coast. Wind potential was only depicted over land surfaces, and a forecast cone was used without a center track line as is current National Hurricane Center practice. One graphical variable was manipulated, the position of the legend. Ideal legend placement will vary based on the specific subject matter of a map and the other graphical elements used. The authors believed that splitting the graphical elements into multiple locations would decrease the effectiveness of the map.

Participants were asked four questions about the map including, “What do you believe is the potential for damaging winds from Hurricane Inga at 8 A.M. Saturday?,” Is Hurricane Inga forecast to strengthen, weaken, or stay the same from 8 A.M. Saturday to 8 P.M. Saturday?,” “Between 8 P.M. Saturday and 8 P.M. Sunday, does the potential for damaging winds from Hurricane Inga increase, decrease, or stay the same?” and “Is Charlotte, North Carolina within the forecast error cone?” The questions were designed so that participants would most likely need to use the legend in order to answer the questions. At the same time, they were questions a typical forecast user may want to know. The results indicate that no legend produced a statistically significant difference in understanding. The best performance was recorded when the map legend was located at the top of the graphic. The average scores ranged from 59.6% for the map with the legend at the top to 48.6 for the map with a split legend. Another interesting result was the apparent difficulty most participants had in answering the questions about the potential for damaging winds. The average score for all participants was only 54%. In the first task, participants had to locate one hurricane position on the map (8AM Saturday) and determine what the current wind category would be. This involved extrapolating the wind category from land to the position of the hurricane over water. Fewer than 30% answered correctly. A second task also missed by over two-thirds of the sample required participants to correctly understand that a forecast where the hurricane icon changes from an ‘M,' (or a major hurricane) to an ‘H' is forecast to weaken. Participants performed much better on tasks 3 and 4, which involved differentiating between wind potential and location within the forecast cone as well as damaging wind potential when indicated by color. The results indicated that legend placement may not contribute significantly to understanding of the forecast graphics; however, the ability of users to interpret the graphics correctly may be lower than one might assume.