Spatial Variations in the Uncertainty of Future Tropical Precipitation Change

This study aims to map, and begin to understand, the substantial uncertainties in future changes in local precipitation due to differences in model formulation. This is motivated by the need, first, to provide a benchmark that can be used to evaluate the success of techniques to constrain regional uncertainty, and second, to provide knowledge about the uncertain physical processes in order to help develop such constraints. The focus is primarily on the vulnerable tropical regions. A metric suitable for mapping uncertainty in modelling precipitation change is therefore developed, removing the contribution due to natural variability, and scaling it by natural variability. This is applied to each of four multi-model ensembles (MMEs): a 268-member perturbed physics slab-model ensemble, a 17-member perturbed physics coupled-model ensemble, a 17-member ensemble with perturbed carbon cycle parameters, and a 16-member coupled-model ensemble including uncertainty in model structure, viz. CMIP3 models. Substantial spatial variations are found in the uncertainty of modelling local precipitation change, with rich structure evident on a range of scales. Key large-scale features are broadly similar between different MMEs, suggesting that more detailed investigation of the large slab model ensemble may have wider applicability. For example, in all MMEs, uncertainty due to model formulation is larger in equatorial regions than in the subtropics, and in the former always markedly exceeds that due to natural variability. Possible reasons for this contrast, and other large-scale features and seasonalities, are explored.