On the sensitivity of anthropogenic aerosol forcing to model-internal variability and parameterizing a Twomey effect

Despite efforts to accurately quantify the effective radiative forcing (ERF) of anthropogenic aerosol, the historical evolution of ERF remains uncertain. As a further step toward a better understanding of ERF uncertainty, the present study systematically investigates the sensitivity of the shortwave ERF at the top of the atmosphere to model-internal variability and spatial distributions of the monthly mean radiative effects of anthropogenic aerosol. For this, ensembles are generated with the atmospheric model ECHAM6.3 that uses monthly prescribed optical properties and changes in cloud-droplet number concentrations designed to mimic that associated with the anthropogenic aerosol using the new parameterization MACv2-SP. The results foremost highlight the small change in our best estimate of the global averaged all-sky ERF associated with a substantially different pattern of anthropogenic aerosol radiative effects from the mid-1970s (-0.51 Wm(-2)) and present day (-0.50 Wm(-2)). Such a small change in ERF is difficult to detect when model-internal year-to-year variability (0.32 Wm(-2) standard deviation) is considered. A stable estimate of all-sky ERF requires ensemble simulations, the size of which depends on the targeted precision, confidence level, and the magnitude of model-internal variability. A larger effect of the pattern of the anthropogenic aerosol radiative effects on the globally averaged all-sky ERF (15%) occurs with a strong Twomey effect through lowering the background aerosol optical depth in regions downstream of major pollution sources. It suggests that models with strong aerosol-cloud interactions could show a moderate difference in the global mean ERF associated with the mid-1970s to present-day change in the anthropogenic aerosol pattern