Communicating Uncertain Weather Information

Anders Doksæter Sivle
2020-11-20
10:15 - 11:15
Visual Computing Forum

Anders works with weather forecasting and communication with users at MET Norway. In his current position, Anders has the daily responsibility at MET for the Yr weather service. This work involves coordination and follow-up of all employees and activities at MET that affect the operation and development of Yr, as well as daily follow-up with NRKs Yr team. Anders is also involved in research projects that focus on co-production of services, communication of uncertainty and the usefulness of weather warnings. His primary area of interest is the communication process between meteorologists and the general public, but also to professional users.

Abstract:
Information on future weather conditions has always been important to people in different occupations. In 1919, Vilhelm Bjerknes and the Bergen School of Meteorology invented the Polar Front Model. This enabled weather forecasts with greater accuracy. The more precise predictions were presented as deterministic information, typically without statements of uncertainty. In the 1970s and 1980s, computer-based weather forecasts became more prevalent, further improving the accuracy. In addition, in the last 10-20 years, ensemble prediction systems are more frequently used, providing probabilistic forecast information. Although most forecast information still is deterministic, there is a movement toward including more uncertainty information. Several studies show that such information is useful to end-users. However, other studies suggest that uncertainty information can also make the forecasts more demanding to understand and use. At MET Norway, and the Yr weather service, both deterministic and probabilistic information is communicated. In this presentation, I will give an overview of literature and practical examples of how complex and uncertain weather information is communicated and used by the general public.