Evaluating Advanced Aerosol Activation Treatments in a Coupled Climate/Mixing-State Resolving Aerosol Model

A major source of uncertainty in projections of climate change derives from our inability to accurately simulate interactions between aerosol, clouds and climate. In the context of a climate model, these many, complex interactions are instigated by a parameterization of aerosol activation, where cloud condensation nuclei (CCN) and the cloud droplet number concentration (CDNC) are estimated from grid cell prognostic aerosol and meteorology. Thus, any attempts to improve representation of aerosol-cloud interactions in climate models through more detailed aerosol or atmospheric physics and chemistry are slave to the accuracy of the activation parameterization.

Using off-line calculations with a detailed numerical parcel model and output from a coupled climate-aerosol model which resolves mixing between sulfate and both black/organic carbon species, it has been shown that existing activation parameterizations can over-predict the CDNC by 50% in highly anthropogenically-polluted conditions. We re-evaluate these error metrics using a new iteration of advanced activation parameterizations which have been corrected for kinetic/inertial limits on droplet growth and tuned to use a more realistic condensation coefficient. In addition, we present a new activation parameterization and framework built by emulating a detailed numerical parcel with globally-approximating polynomials; against our parcel model as a metric, the emulator performs better than other complex activation parameterizations and reduces biases in CDNC in the global average.

Finally, the impact each parameterization has on the coupled model's total aerosol indirect effect is evaluated.