Weighting Models Based on Several RCM-Specific Metrics: Exploring the Concept
Christensen, O.B., J.H. Christensen, F. Giorgi and Rummukainen, M.
Pages 33-34. In: T. Stocker, Qin Dahe, G.-K. Plattner, M. Tignor and P. Midgley (eds): IPCC Expert Meeting on Assessing and Combining Multi Model Climate Projections. National Center for Atmospheric Research, Boulder, Colorado, USA, 25-27 January 2010. Meeting Report 33.
An important new development within the European FP6 project ENSEMBLES has been to explore performancebased weighting of regional climate models. Until now an assumption of equal weight has been implicitly applied. At the same time, it is evident that different RCMs give results of a varying degree of realism, e.g. related to extremes. It is not straightforward to construct, assign and combine metrics of model performance. Rather, there is a considerable degree of subjectivity both to the definition of the metrics and to the way these are combined into weights. This does not mean that weighting, however exploratory, would not be meaningful. Rather, it stresses that the assumptions and choices behind the weights need to be recognized and taken into account. Here we discuss the applicability of combining a set of six specifically designed RCM weights to produce one model index producing combined climate change information from the range of RCMs used within ENSEMBLES.
|
| |
