Aims

Given that the co-locations of the paleo-simulations and future projections is a novelty in CMIP5, the successful use of this kind of information does not have much of a track record. Thus a workshop whose aims are to a) ensure that the community is aware of how suitable techniques should be assessed right from the start, b) highlight the rich range of possible analyses, and c) produce timely and informative summaries of the state of these analyses, is both welcome and necessary.

We envisage a directed and productive workshop that will be charged with assessing what has been done so far, where are the missing gaps, and what can be done better in the future. We want to highlight the unique aspects of the paleo-climate simulations for evaluating the models in time for further analyses to be done in time for AR5 (although this kind of approach is not solely tied to the IPCC process or questions deemed interesting by those authors).

We will aim to produce a number of products. Some of these will be concrete - such as a 'best practice' review paper by participants to help guide the wider community in these kinds of analyses, but also some less tangible products - such as fostering a greater appreciation from the CLIVAR/WGCM/WCRP communities of the benefits of considering paleo-climate data and models, and encouraging a greater discipline i the paleo-climate community when it comes to making the links and encouraging a greater discipline in the paleo-climate community when it comes to making the links to future changes, the importance of synthesis of paleo data and a better appreciation for what questions can be usefully posed to, and answered by, the modellers.

In order to start the workshop off on a productive note we will be asking participants to have looked at some key fields/diagnostics ahead of time. In particular, various aspects of the regional hydrological changes in the paleo-climate simulations and in the future projects will be processed ahead of time and made available for participants. These metrics will involve rainfall, lake levels, soil moisture, forward models for water isotopes etc. for which there is plentiful evidence (and gridded datasets) of change during the periods covered by the paleo simulations (i.e. LGM, Mid-Holocene and Last Millennium). Dealing with these diagnostics and comparison with available paleo-data will serve as a template for any proposed approaches and techniques. Key issues will be getting as wide a spread of models as possible, ensuring that the forcings for each model are appropriately characterised, making clear demonstrations that any chosen skill metric does actually correlate to differences in projections, and the establishment of 'perfect model' test protocols (i.e. does any technique for creating an improved projection actually work if one assumes that one of the models is 'truth'). Each of these issues were highlighted in the 'Good Practice Guide' arising from last year's IPCC Expert Meeting (Knutti et al., 2010).

By creating a standard in one particular case, we will hope to maintain these standards in studies that workshop participants will be doing individually, and that might get done elsewhere.