Progress, Challenges, and Future Directions in Communicating Uncertainty

The Weather Enterprise has increased its creation and communication of forecast uncertainty information. Today numerous operational datasets and products are probabilistic in nature, including severe weather outlooks, hurricane tracks, the likelihood of precipitation, and storm surge. Broadcasters are using ensemble information to provide forecast scenarios on air, online, and via social media to highlight uncertainties in the forecast. Although such probabilistic information is being provided, events such as Hurricane Sandy, the Atlanta ‘snow-jam', and the 26-27 January 2015 Blizzard highlight that much more work is necessary to communicate forecast uncertainty effectively. This talk will review our community's progress, highlight key challenges, and suggest future directions in communicating uncertainty.

The challenges are many. For example, methods to estimate the uncertainty of a weather system (i.e., ensembles) often underestimate the true uncertainty, leading to overconfident communication of the forecast. Further, users have different needs and decision-making contexts. Thus the relevant probabilistic thresholds for effective warnings and messaging vary among the myriad of users, and these key thresholds are largely unknown. Moreover, there is a need to balance communicating simply and succinctly with the complexity and detail often required to convey uncertainty information. But these challenges are being recognized and there are increasing efforts to address them. Among the community there is substantial social science attention to the problem, tremendous investment in improving and calibrating ensemble information, and bold and diverse experimentation of methods to communicate uncertainty. There is recognition that coupling these social and atmospheric science efforts is important to ensure the provision of useful uncertainty information. This talk will highlight these recent advances and how they may shape the future of communicating uncertainty.