Carbon dioxide has exerted the largest portion of radiative forcing and surface temperature change over the industrial era, but other anthropogenic influences have also contributed. However, large uncertainties in total forcing make it difficult to derive climate sensitivity from historical observations. Anthropogenic forcing has increased between the Fourth and Fifth Assessment Reports of the Intergovernmental Panel of Climate Change (IPCC; refs , ), although its relative uncertainty has decreased. Here we show, based on data from the two reports, that this evolution towards lower uncertainty can be expected to continue into the future. Because it is easier to reduce air pollution than carbon dioxide emissions and because of the long lifetime of carbon dioxide, the less uncertain carbon dioxide forcing is expected to become increasingly dominant. Using a statistical model, we estimate that the relative uncertainty in anthropogenic forcing of more than 40% quoted in the latest IPCC report for 2011 will be almost halved by 2030, even without better scientific understanding. Absolute forcing uncertainty will also decline for the first time, provided projected decreases in aerosols occur. Other factors being equal, this stronger constraint on forcing will bring a significant reduction in the uncertainty of observation-based estimates of the transient climate response, with a 50% reduction in its uncertainty range expected by 2030.