Ensembles and uncertainty in climate change impacts

There is an increasing use of large numbers of climate simulations (known as multi-member ensembles) for assessing the impact of future climate change on society, often through model intercomparison projects (MIPs). This presents both opportunities for understanding uncertainties and confidence in our projections, and challenges for interpreting the results. We outline current approaches to assessing uncertainties in climate impacts, statistical methods for assessing uncertainties, issues regarding model integration and complexity, and ways in which uncertainty frameworks can be used to inform adaptation decisions, with case studies focused on agriculture. Finally, we highlight future research needs and provide recommendations for making further progress.

The paper suggests that:

·         there is a need to consider how model intercomparison projects can be better used to improve and develop models, and to synthesize knowledge more effectively;

·         A more consistent approach should be taken to assessing the effect of model integration;

·         There is a need to assess the benefits and tradeoffs between model complexity, resolution and ensemble size for making impact assessments;

·         The use of statistical methods to assess uncertainty should be promoted;

·         The usability of impact assessments for decision making could be improved through:

o   Greater clarity in the methods and assumptions used

o   Appropriate selection of assessment approaches

o   Common, clear ways of reporting and describing uncertainties

o   The use of models as tools from which information is extracted, rather than as competing attempts to represent reality

o   Assessment of methods which go directly from climate model to decision parameter, removing intermediate steps and potentially reducing embedded uncertainty

The full paper is available through this link: