Communicating Uncertainty in Weather Forecasts and Warnings: Challenges and Visualization Examples

The mobile weather app nowadays represents the most important channel for MeteoSwiss to communicate its forecasts and warnings to the general public. The forecast presentation on the app has evolved from an initially deterministic format to now also include probabilistic elements. However, these first steps towards including uncertainty information have provoked mixed user feedback, highlighting the challenge of effectively communicating uncertainty information to our stakeholders. If communicated effectively, forecast uncertainty information leads to better decisions thus illustrating the added value of such information (Joslyn and LeClerc, 2013). 

To address this challenge, we have started an effort to explore new ways effectively communicating forecasts and their uncertainty and will present examples of new visualizations. We start from the existing and well-established forms of forecast visualizations including time series of forecasts and spatial maps and extend these to convey forecast uncertainty in an intuitive and thus effective manner. To gauge the effectiveness of these new visualizations, we will perform end-users evaluations. Such end-user evaluations are seen as key element but represent a substantial effort: There is a risk of rejecting promising concepts if the prototype visualization is not of sufficient visual quality, as graphical details may decide on acceptance or rejection by end-users. Our ultimate goal is to enable the general public to take well-informed and thus better decisions by improving their understanding of the uncertainty associated with weather forecasts and warnings. 

Susan Joslyn and Jared LeClerc, Decisions With Uncertainty: The Glass Half Full, Current Directions in Psychological Science, Vol. 22, 4, 2013, https://doi.org/10.1177/0963721413481473