Abstract
A common signal in climate model projections is a decline in average summer rainfall over midlatitude continents due to anthropogenic warming. Most models suggest this rainfall decline will be less severe over North America than over Europe. This study aims to understand this difference in continental response and make inferences about its reliability. Data are primarily derived from a “perturbed physics” ensemble of models [Quantifying Uncertainty in Model Predictions project, subensemble S4 (QUMP-S4)] and are also compared with data from a multimodel ensemble [the Coupled Model Intercomparison Project phase 3 (CMIP3)]. A description of the uncertainty of predicted summer rainfall decline over both continents and its broad similarity between the two ensembles suggests the possibility that the QUMP-S4 ensemble may include many of the mechanisms that cause the differential continental response in the CMIP3 ensemble. Analysis of the QUMP-S4 mechanisms and their variability across the ensemble lead to the following conclusions. Over western North America, it is judged that the change in summer rainfall is more uncertain than models suggest, with a decline that could be either more or less severe than that over Europe. This is due to the western North American region’s dependence on uncertain modeling of high-elevation winter–spring surface hydrology. Over eastern North America, it seems likely that summer rainfall will decline. In particular, this decline is likely to be less severe than that over continental Europe since this difference primarily depends on reliable aspects of the models. However, a further, but speculative, conclusion is that these mechanisms could also lead to a larger increase in extreme rainfall events over eastern North America than over Europe.
Corresponding author address: Dr. David P. Rowell, Met Office Hadley Centre, FitzRoy Road, Exeter EX1 3PB, United Kingdom. Email: dave.rowell@metoffice.gov.uk
