Nature 503, 67–71 (2013)

Little is known about the effect of changing atmospheric aerosols on cloud processes or the associated radiative forcing. Large uncertainties in the impact of aerosols on cooling the climate limit our confidence in understanding climate sensitivity to greenhouse gas emissions.

To determine the uncertainty caused by anthropogenic aerosols, Ken Carslaw, of the University of Leeds, UK, and colleagues performed a sensitivity analysis on a global aerosol model. By comparing model simulations for pre-industrial (1750) to near present day (2000) aerosol distributions they quantify the sources of uncertainty.

Natural emissions, such as those from volcanos, oceans and biomass burning, are found to account for 45% of uncertainty in the forcing change since 1750. Anthropogenic emissions account for 34%. Choosing different reference years shows that natural emissions remain a large source of uncertainty. These results highlight the need to understand pristine environments, such as the pre-industrial, to improve model representation of this variable.