Modelling uncertainty refers to an incomplete understanding of Earth system processes and their imperfect representation in climate models.
Models provide a mathematical representation of many of the processes in the climate system, which are based on a mixture of theory, observations and experimentation, and are inevitably uncertain.
Different modelling groups seek to represent climate processes in the best possible way in their models and, because this is to an extent a subjective judgement, this leads to different groups adopting different representations. Not surprisingly, this leads to different strengths (and even, in the case of clouds, directions) of feedbacks in the models, and hence different projections of future changes. This is the structural error component of modelling uncertainty.
It is not possible to explicitly incorporate every physical process operative in the climate system, into a climate model. This is often because the processes involved operate at too fine a spatial resolution, to be incorporated into a model, due to computational constraints. Consequently, parameters are assigned to represent these processes, involving a mixture of theory, observations, experimentation and expert judgement. Definition of parameter values is seldom precise enough to be able to assign accurate values. This is the parameter error component of modelling uncertainty.
- An overview of the uncertainties associated with climate modelling is described in Section 3.1 of the UKCP09 Briefing report.
Find out more
- IPCC Fourth Assessment Report, Working Group 1 report The Physical Science Basis , Chapter 10