Diagnosing Instantaneous Forcing and Feedbacks of Downwelling Longwave Radiation at the Surface: A Simple Methodology and Its Application to CMIP5 Models

Climate models predict large increases in downwelling longwave radiation (DLR) at Earth's surface as atmospheric CO2 concentrations increase. Here we introduce a novel methodology that allows these increases to be decomposed into direct radiative forcing due to enhanced CO2 and feedbacks due to subsequent changes in atmospheric properties. For the first time, we develop explicit analytic expressions for the radiative forcing and feedbacks, which are calculable from time-mean fields of near-surface air temperature, specific humidity, pressure, total column water vapor, and total cloud fraction. Our methodology captures 90%-98% of the variance in changes in clear-sky and all-sky DLR in five CMIP5 models, with a typical error of less than 10%. The longwave feedbacks are decomposed into contributions from changes in temperature, specific humidity, water vapor height scale, and cloud fraction. We show that changes in specific humidity and height scale are closely linked to changes in near-surface air temperature and therefore, in the global average, that 90% of the increase in all-sky DLR may be attributed to a feedback from increasing near-surface air temperature. Mean-state clouds play a major role in changes in DLR by masking the clear-sky longwave and enhancing the temperature feedback via increased blackbody radiation. The impact of changes in cloud cover (the cloud feedback) on the DLR is small (similar to 2%) in the global average, but significant in particular geographical regions.

Automated summary

This study introduces a novel methodology to decompose the impacts of increasing CO2 concentrations on downwelling longwave radiation (DLR) at Earth's surface, including direct radiative forcing and subsequent feedback mechanisms. By decomposing the feedbacks, it is found that the changes in near-surface air temperature play a dominant role in the increase of all-sky DLR, while cloud cover changes have significant impacts on DLR in specific geographical regions.