Adjustment of GCM Precipitation Intensity over the United States

Abstract

A regression equation is developed to adjust the simulated monthly averaged intensity of hourly precipitation over the continental United States using air temperature at the first model level (about 80 m above ground) simulated by a revised version of NCAR?s Community Climate Model version 2 coupled with the Biosphere?Atmosphere Transfer Scheme. The adjusted precipitation intensity is in much closer agreement with the locally observed precipitation intensity, both in pattern and magnitude, than is the simulated intensity produced by climate models. The multiresponse randomized block permutation analysis shows that the authors? adjustment method is statistically significant. The equation is robust with respect to seasonal and interannual variations, as demonstrated by applying it to independent months of different years. The relationship between precipitation intensity and near-surface air temperature likely reflects a linkage among precipitation, precipitable water, and vertical velocity. It can be applied to estimate fractional precipitation coverage with temporal and spatial variability over the continental United States, and to help statistically construct a time series of local precipitation events both for the present climate and for the studies of climate change.