skip to content

Mathematical and Statistical Approaches to Climate Modelling and Prediction

Participation in INI programmes is by invitation only. Anyone wishing to apply to participate in the associated workshop(s) should use the relevant workshop application form.

11th August 2010 to 22nd December 2010
Richard Chandler University College London
Matt Collins Met Office, University of Exeter
Peter Cox University of Exeter
Kevin Horsburgh [Proudman Oceanographic Laboratory], National Oceanography Centre, Liverpool
John Huthnance [Proudman Oceanographic Laboratory], National Oceanography Centre, Liverpool
Jonathan Rougier University of Bristol
David Stephenson University of Exeter
John Thuburn University of Exeter


Programme Theme

Our best estimates of future climate are based on the use of complex computer models that do not explicitly resolve the wide variety of spatio-temporal scales making up Earth's climate system. The non-linearity of the governing physical processes allows energy transfer between different scales, and many aspects of this complex behaviour can be represented by stochastic models. However, the theoretical basis for so doing is far from complete. Many uncertainties remain in predictions derived from climate models, yet governments are increasingly reliant on model predictions to inform mitigation and adaptation strategies. An overarching aim of climate scientists is to reduce the uncertainty in climate predictions and produce credible assessments of model accuracy. This programme focuses on two key themes that both require the close collaboration of mathematicians, statisticians and climate scientists in order to improve climate models and the interpretation of their output.

The first theme is the development of improved stochastic sub-grid-scale physics models, which have the potential to improve the variability of ensemble climate simulations. Progress can be made by establishing frameworks for relating models of different resolutions and for combining stochastic Earth System Models of Intermediate Complexity (EMICs) with global climate models (GCMs). Stochastic approaches to climate modelling will benefit from improving the connection between deterministic models and statistical tools such as downscaling, emulation, reified modelling and dimensional reduction.

The second aspect of the programme concerns the use of statistical techniques to create a theoretically sound basis for probabilistic climate prediction. The vast amount of data produced by climate models needs synthesis in order to provide the predictions, and credible error estimates, required by policy makers. This theme will provide an environment in which to seek quantitative answers to fundamental questions of interpreting probabilistic output, and the reliability of climate predictions at varying space and time scales. It will also consider what measurements can be used to assess the quality of climate model predictions.

This programme will bring together world-leading researchers in climate modelling, mathematics and statistics in order to make progress in solving some of the major issues facing climate prediction.

Final Scientific Report: 
University of Cambridge Research Councils UK
    Clay Mathematics Institute London Mathematical Society NM Rothschild and Sons