All information in the earth sciences has some attendant uncertainty, and when information is obtained by process modelling, statistical prediction or some combination of both then this uncertainty may be substantial. It is important that users of information are fully aware of this uncertainty so that it can be taken into account in decision-making. There are various ways in which the uncertainty in information can be quantified, through statistical models and through methods to track the propagation of error in the processing of data through process models. These typically produce descriptors of uncertainty in terms of statistical moments (eg. variances) or distributions. When uncertainty pertains to environmental variables that vary in space and/or time, then spatial and/or temporal correlations in uncertainty is typically described by variograms or correlograms. Some uncertainty cannot be treated statistically, for example our uncertainty about future soil behaviour under climate change may depend on future scenarios which may not be completely characterized and which cannot be reasonably assigned probabilities.

Problems of how to characterize uncertainty and to analyse it in particular cases have received much attention. The focus of this session is on the next step: how this mathematically rigorous quantification of uncertainty can be most effectively communicated to data users, including policy makers, regulators? and members of the general public. The explanation of uncertainty presents a growing challenge, in particular to an audience with a wide range of scientific and mathematical expertise. The aim of this session is:

(i) to provide a forum to discuss experiences in the communication of uncertainty associated with information in the earth sciences;
(ii) to discuss novel ideas in the formulation of descriptors of uncertainty to facilitate communication including both textual and graphical representation;
(iii) to gain understanding on issues of audience including understanding audience types and their needs; and
(iv) to hear insights into the problem of communicating uncertainty from other areas of science (e.g. medicine) where this is an important challenge.
(v) When uncertainty is spacially distributed, how to comunicate it in maps where values and uncertainties are combined.