Communicating uncertainty in seasonal climate forecasts in Europe

INTRODUCTION Organisations in sectors as diverse as water management, agriculture, energy, health, transportation, and tourism need to adapt to future climate variability and change. Seasonal climate forecasts therefore have... [ view full abstract ]

INTRODUCTION
Organisations in sectors as diverse as water management, agriculture, energy, health, transportation, and tourism need to adapt to future climate variability and change. Seasonal climate forecasts therefore have potentially high value for these organisations. Seasonal climate forecasting for Europe is starting to outperform historical averages for predicting climate variables and impacts for future seasons. However, seasonal forecasts are inherently uncertain, with uncertainty coming from both the probabilistic nature of the predictions, and the reliability of the underlying models. Forecast providers therefore face the challenge of effectively communicating these uncertainties to their users. Here, we examine user perceptions of and preferences for such communications.
METHODOLOGY
We conducted a survey with 50 participants from the relevant sectors mentioned above. We asked about 1) how accessible, understandable and useful they found their current sources of climate information; 2) whether they received information about how well earlier forecasts have matched observed climate; and 3) their preferences for and familiarity with seven visualisations of forecast likelihoods (including bar-graphs, pie-charts, and others).
RESULTS
We present three main findings. First, participants found seasonal climate forecasts to be useful, but also rated them as relatively difficult to access and understand. Second, only 13% indicated that they received information about how well seasonal climate forecasts match observations. Third, participants preferred visualisations with which they were already more familiar.
CONCLUSION
The mismatch between the perceived usefulness and understandability of seasonal climate forecasts highlights a potential need for more user friendly formats, as does the fact that few participants reported receiving information about how well previous predictions have matched observations. Such formats may be better received if they are perceived to be familiar. We discuss the implications of these findings for both the providers and users of climate information, as well as for communication of risk and uncertainty.