Radiative forcing in the CMIP5 models

Erik J.L. Larson (1), Robert W. Portmann (2)

Abstract
Effective radiative forcing (ERF) is calculated as the flux change at the top of the atmosphere, after allowing fast adjustments, due to a forcing agent such as greenhouse gasses or volcanic events. Accurate estimates of the ERF are necessary in order to understand the drivers of climate change. ERF cannot be observed directly and is difficult to estimate from indirect observations due to the complexity of the climate response to individual forcing factors. Using CMIP5 data we estimate the ERF from historical simulations using two methods, regressions and kernel functions. The regression method is sensitive to the number of years chosen for the regression due to the nonlinearity of the climate feedback parameter. The kernel method is sensitive to the variable (temperature, imbalance, etc) used to create the kernel. We quantify these sensitivities and discuss ramifications for estimating earth’s energy budget. We also reconcile the differences in volcanic forcing calculated from the CMIP5 and those reported in the IPCC AR5. The IPCC AR5 uses the adjusted forcing for stratospheric volcanic aerosol resulting in a larger forcing than calculated in transient runs in the CMIP5. We calculate an ERF efficacy of the CMIP5 simulations of 0.80. This leads to a difference in the integrated forcing between the IPCC AR5 report and the CMIP5 mean of 4.2 x1022 J, about 5% of the total since 1870, just from the 5 largest volcanic events.