Communicating uncertainty via probabilities: The case of weather forecasts

Capturing uncertainty through numerical probabilistic statements is orthodoxy in risk science—and most of science and technology. There are a wide range of views on the utility of such statements for risk communication, and they are often seen as being central to the failure to generate common understanding about risks between science and non-scientists. The extent to which probability statements are understood is unclear. If such statements are misunderstood by many, what alternatives might communicate uncertainty better?

These questions are examined in the context of daily weather forecasts. The probabilities used in such statements concern daily events experienced by everyone, unlike the extremely small probabilities about unfamiliar events often used in risk communication. If people do not understand weather forecasts, there is little hope that statements about unfamiliar events using unfamiliar language will be understood. Some jurisdictions use numerical probabilistic statements on the likelihood of precipitation, and a variety of qualitative or categorical forecasts are also used. Drawing on a range of sources including public surveys conducted by the Australian Bureau of Meteorology, the paper examines public understanding of probabilities and public and specialist understanding of verbal categorical forecast terms.

The majority of those surveyed have basic understanding about probabilities as used in weather forecasts, but significant groups do not. However, there was limited agreement among forecasters on what the probabilistic statements meant. Similarly, there was limited shared meaning between forecasters and the public on the verbal forecast expression examined.