The Feature Paper can be either an original research article, a substantial novel research study that often involvesseveral techniques or approaches, or a comprehensive review paper with concise and precise updates on the latestprogress in the field that systematically reviews the most exciting advances in scientific literature. This type ofpaper provides an outlook on future directions of research or possible applications.
Previous examples of weighting or selection on the basis of realism have argued that it may be more justifiable for specific regions or applications where the key aspects of model behaviour can be identified and assessed (e.g. Overland et al. 2011; McSweeney et al. 2012). We explore whether it is either practical or justifiable to extend a regional approach in order to identify a subset of models which are suitable for use across multiple regions. Such an approach would yield practical benefits by reducing the overhead of interfacing between GCMs and RCMs for modelling groups involved in providing downscaled information for more than one region, as well as reducing the need for multiple selection studies. Further, projects with a global scope such as ISIMIP (Intersectoral-Impacts model inter-comparison project) (Warszawski et al. 2013) have expressed interests in the use of a consistent set of GCMs for downscaling in order to generate datasets with consistently generated uncertainty ranges globally.
Our selection of evaluation criteria includes key aspects of the large-scale climate of the regions (i.e. relevant to driving the high resolution downscaling) and also takes advantage of specific CMIP5 regional assessments documented in the existing literature. However, we note that for many of these existing studies the analysis is limited to only a subset of the LBC-avail models due to the lag in data availability in the CMIP5 archive. An example of this is the thorough evaluation of the Asian summer monsoon of Sperber et al. (2013) in 25 of the CMIP5 GCMs. This limits our scope to make use of this potential very valuable source of information. We therefore note the outcomes of other studies where relevant, and use them as supporting evidence for our model assessment, but we cannot rely on this evidence alone to influence the selection of a model if the assessment does not extend to all models. As the body of literature on the assessment of CMIP5 models becomes more complete over coming years and an increasing volume of well-documented processed-based evaluation results becomes available, in future it might be possible to undertake a thorough selection exercise based solely on information available in existing literature.
Our assessment has also necessarily been limited to a restricted number of criteria as it is not feasible to undertake full assessments of all CMIP5 models. As an increasing body of literature on both the performance and projections of CMIP5 models emerges, others will be able to draw on a wider range of well documented evidence for selection. Efforts to gather and share well-documented metrics from CMIP5 models, such as those currently being undertaken by the WCRP Climate Model metrics Panel, could provide a valuable basis for the informed selections which are likely to become an increasingly important stage in the development of regional climate change projections. 1e1e36bf2d