Shortwave heating response to water vapor as a significant source of uncertainty in global hydrological sensitivity in CMIP5 models
Tuesday, 16 December 2014
The hydrological cycle is expected to undergo substantial changes in response to global warming, with all climate models predicting an increase in global-mean precipitation. There is considerable spread among models, however, in the projected increase of global-mean precipitation, even when normalized by surface temperature change. In an attempt to develop a better physical understanding of the causes of this intermodel spread, we investigate the rapid and temperature-mediated responses of global-mean precipitation to CO2 forcing in an ensemble of CMIP5 models by applying regression analysis to pre-industrial and abrupt quadrupled CO2 simulations, and focus on the atmospheric radiative terms that balance global precipitation. The intermodel spread in the temperature-mediated component, which dominates the spread in total hydrological sensitivity, is highly correlated with the spread in temperature-mediated clear-sky shortwave (SW) atmospheric heating among models. Upon further analysis of the sources of intermodel variability in SW heating, we find that increases of upper atmosphere and (to a lesser extent) total column water vapor in response to 1K surface warming only partly explain intermodel differences in the SW response. Instead, most of the spread in the SW heating term is explained by intermodel differences in the sensitivity of SW absorption to fixed changes in column water vapor. This suggests that differences in SW radiative transfer codes among models are the dominant source of variability in the response of atmospheric SW heating to warming. Better understanding of the SW heating sensitivity to water vapor in climate models appears to be critical for reducing uncertainty in the global hydrological response to future warming. Current work entails analysis of observations to potentially constrain the intermodel spread in SW sensitivity to water vapor, as well as more detailed investigation of the radiative transfer schemes in different models and how they relate to the sensitivity of SW heating.