Users survey: how to communicate level of confidence

Published on: 
February, 2014

In order to effectively communicate confidence and uncertainty in seasonal-to-decadal climate predictions (S2DCP) it is important to understand end-user needs and preferences with respect to this information. The results of a  survey, conducted with EUPORIAS stakeholders and interested organisations (totalling a sample size of 50), have just been published. The survey covered four thematic areas:

  1.  Perception of the accessibility, understandability and usefulness of S2DCP;
  2.  Information currently obtained and used by end-users;
  3.  Organisational approach to uncertainty; and
  4.  Preference for information formats.

 One of the findings of the survey is that S2DCP are perceived to be more useful than they are accessible or understandable. This emphasises the need for more user friendly formats and interfaces. The need to make information on the reliability of forecasts more salient and understandable was also highlighted as the majority of current users did not receive information about how well earlier forecasts matched observed climate. Most respondent organisations were concerned with rare but severe events although  many also indicated that they tended to focus on those events most likely to occur. Sectoral differences were noted with respect to how organisations liked to receive and utilise information on uncertainty. For instance, preference for information formats that facilitate Yes/No decision making was strong amongst those in the water and energy sectors, but not amongst those in the health sector. With respect to visualising confidence and uncertainty, a clear preference for maps and depictions of spread emerged; with preference for the latter being associated with greater comfort with statistical information. The findings of this survey highlighted the importance of:

  • Ensuring that information regarding confidence and uncertainty is presented in a manner that is readily comprehensible;
  • Taking into account the types of decision making this information will be utilised in; and
  • Fully testing visualisations to ensure that they are interpreted as intended.