|
The latest multi-model approach, derived from the IPCC AR4, allows, to an unprecedented scale, latest generation coupled GCMs to be analyzed together and compared. Associated analysis shows that El Niño is now an emergent mode of variability in complex models. However, the diversity of their simulations of El Niño contributes to a large uncertainty in projections of future tropical climate and the associated teleconnections and impacts. Part of this uncertainty is due to the model shortcomings and it is key to document them in a common way. It is important to assess ENSO characteristics in terms of theoretical/mechanistic understanding of the phenomena, not just looking at local statistics (e.g. Niño 3 SST anomalies ), which may have the correct value for the wrong reasons (i.e. as a result of bias compensation). Moreover, multi-model analyses should rely upon common diagnostics. The definition of a set of “metrics” to assess a phenomenon can have great value to the wider community engaged in model development and/or analysis. Metrics are now under discussion in preparation for future CMIPs (Gleckler et al. 2008, Guilyardi et al. 2009), and the CLIVAR Pacific Panel is charged with devising metrics for ENSO and for the wider tropical Pacific climate. Here we use the term “metric” as a measure of the “distance” of the model to some observational reference, usually computed as a single scalar value (Gleckler et al. 2008) while other more complex or qualitative analyses where observations do not provide an easy reference are called “diagnostics”. The
analysis presented here builds on community discussion that led to a
proposal presented at the Perth March 2009 Pacific Panel meeting (see
associated mindmap).
In devising such metrics, several goals are pursued:
|
 |