AAAR 33rd Annual Conference
October 20 - October 24, 2014
Rosen Shingle Creek
Orlando, Florida, USA
Abstract View
Can We Tame the Aerosol Uncertainty Monster?
KEN CARSLAW, Lindsay Lee, Kirsty Pringle, Carly Reddington, Leighton Regayre, University of Leeds
Abstract Number: 446 Working Group: Invited by Conference Chair
Abstract Aerosol science has made enormous steps in understanding fundamental processes and in making measurements that probe ever more complex particle properties. But how much of this new knowledge is being translated into better models of how aerosols affect climate? From the third to the fifth IPCC assessment report, aerosols have remained the largest radiative forcing uncertainty. In this presentation I address two questions: firstly, whether we are working on the right processes and secondly, how we can use models and observations to slowly reduce the persistent uncertainty. Using fairly well established statistical techniques it is possible to perform essentially a Monte Carlo simulation with a complex global aerosol model. This enables the contribution of all important processes to the overall prediction uncertainty to be quantified and mapped. The list of key processes turns out to differ substantially depending on whether you want to understand the uncertainty in present-day aerosol or its effect on radiative forcing. The properties of biomass burning particles are important for present-day global CCN uncertainty, but are less important for the uncertainty in forcing, although big questions remain about how these particles interact with clouds. The list also depends on whether you want to understand the sources of uncertainty in forcing since the pre-industrial period or over recent decades. Natural aerosols tend to dominate forcing uncertainty when referenced back to the pre-industrial, but uncertain anthropogenic emissions are the most important factor for recent changes in forcing, with aerosol microphysical processes being less important. Regardless, the list of most uncertain parameters enables us to begin to constrain the model uncertainty using well-chosen measurements. In this direction, I will describe a methodology that could be used to define an optimum measurement strategy if the reduction in model uncertainty is the objective.