A comparison of projected future precipitation in Wisconsin using global and downscaled climate model simulations: implications for public health

Motivated by the documented linkage between water-borne disease outbreaks and heavy rainfall, we compare simulations of precipitation in Wisconsin from two different downscaling procedures (statistical and dynamical) and global climate models (GCMs) for the late 20th and middle 21st centuries (SRES A2 greenhouse emissions scenario). In the inter-model mean, all the three methods produce reasonably accurate simulations of seasonal and annual precipitation amounts during the historical period of 1971-2000 (yearly biases < 5%), but the GCMs severely underestimate extreme precipitation compared with downscaled output and observations. The modelling methodologies agree that Wisconsin should experience a modestly wetter future climate (annual increase < 10%) by the middle 21st century (2041-2070), comprised of more precipitation during winter, spring, and autumn but an equivocal summertime signal. The future simulations also exhibit robust increases in the frequency and intensity of extreme daily precipitation, consisting of larger relative changes in the return periods of heavy events (up to -50%) than in the accumulations (<30%). Although the modelling procedures vary substantially in their projected absolute differences in future precipitation, they agree surprisingly well on the simulated relative differences (percent change) of both mean and extreme precipitation. Unfortunately, there is little consistency in the simulated spatial patterns of future precipitation change across Wisconsin, complicating societal adaptation measures to enhanced extremes. Nevertheless, the composite projections presented here suggest that impending hydrological changes in Wisconsin represent a public health threat, by virtue of increasingly extreme precipitation promoting water-borne disease outbreaks.