User-Centered Design and Testing of Drought Forecast Visualizations to Improve Understanding and Decision-Making

Abstract Climate forecast visualizations can often be unintuitive, complicated to use, and difficult to understand. Using evidence from multidisciplinary social science research and a robust mixed-method approach, we collaborated with scientists and forecasters at the National Oceanic and Atmospheric Administration’s (NOAA) Climate Prediction Center and National Integrated Drought Information System (NIDIS) to redesign deterministic U.S. Drought Outlook forecast graphics issued by NOAA. Results from focus groups, workshops, and literature reviews informed the coproduction of three modified drought forecast graphics that were empirically tested in a control–treatment experiment and survey of U.S. adults ( N = 813) who were randomly assigned to one of three treatment conditions or the control group. Through task-based assessments, we measured the effects of color scales and legend design on dimensions of usability such as efficacy, intuitiveness, and likeability and on accuracy—an objective measure of understanding. Results show that modified graphics significantly improved users’ understanding of the key message of the U.S. Drought Outlooks. Modified graphics were more intuitive than the control, and the modified legend was significantly better understood. Regression models showed that familiarity with geospatial visualizations, having color vision impairments, and education significantly explained usability dimensions. Accuracy was likely to be higher if efficacy was also high, but likeability, familiarity, and one of the modified graphic conditions negatively affected accuracy. We discuss the implications of these findings; present evidence-based recommendations for operationally improving the U.S. Drought Outlooks; and highlight the value of collaborative, interdisciplinary, social scientific research in improving vital decision support products. Significance Statement We collaborated with federal scientists and forecasters to identify whether certain design elements within the U.S. Drought Outlook forecast visualization maps made them less intuitive, harder to interpret correctly, and unclear in their key message. We found that the map’s colors and legend were confusing and unclear, and most users incorrectly interpreted it as a drought intensity forecast. Using multidisciplinary scientific evidence, we then coproduced three modified designs and tested them in a control–treatment experiment. All three modified designs improved understanding of the key message. Two of the redesigned graphics had more intuitive colors, and the modified legend was clearer than the current one. Our findings are intended to improve the design of national drought forecast visualizations to increase their decision-making impact and public understandability.