Investigating aerosol–convection interactions on a global scale using the Convective Cloud Field Model (CCFM) within the aerosol–climate model ECHAM–HAM

Convection plays an important role in the climate system through its effects on radiation, precipitation, large-scale dynamics and vertical transport of aerosols and trace gases. The effects of aerosols on the development of convective cloud and precipitation are a source of considerable uncertainty in current climate modeling. A fuller understanding of aerosol–convection interactions is therefore crucial if we are to reduce this uncertainty. It is only feasible to run convection-resolving simulations on small domains, which limits their ability to represent the interactions between convective systems via larger-scale dynamics which are thought to play an important role. We therefore consider how the parameterization of convection in global models might be refined in order to better capture aerosol–convection interactions on the global scale.

Despite the inherent nonlinearity in the development, behavior and interactions of convective clouds, most global general circulation models (GCMs) use "mass-flux" convection parameterizations, which assume that the ensemble of convective clouds within each GCM column by a single "mean" convective updraft. In addition to averaging out spatial and temporal variability in convective precipitation, such an approach prevents a meaningful representation of the microphysical processes underlying aerosol–convection interactions. In particular, the relationship between aerosol and the droplet size distribution via Köhler theory depends on the vertical velocity distribution, about which little or no information is available in a mass-flux parameterization. In addition, the entrainment, transport, detrainment and scavenging of aerosol may vary nonlinearly over the ensemble of convective clouds within a GCM column.

The Convective Cloud Field Model (CCFM) addresses these limitations of traditional mass-flux convection schemes by simulating a spectrum of different convective cloud types within each column of the GCM, according to their horizontal radius at cloud base. The parameterization combines the quasi-equilibrium hypothesis of Arakawa and Schubert with an entraining Lagrangian parcel model for each cloud type. This parcel model is initiated by a surface buoyancy perturbation for each cloud type, producing a robust determination of vertical velocity at cloud base which can be used in the parameterization of microphysical processes within the cloud. Convective tracer transport and scavenging are calculated independently for each cloud type, recognizing their nonlinear character. The distribution of these cloud types within a GCM column is determined by their competition for the convective available potential energy (CAPE) at the resolved scale.

By using CCFM embedded within a current-generation global GCM with interactive size-resolved aerosol (ECHAM6–HAM2), we show how this approach to convective parameterization can improve the spatiotemporal distribution of simulated convective precipitation events, using both satellite observations and simulations with a limited-area convection-permitting model for evaluation.

We also demonstrate the ability of CCFM to represent aerosol indirect effects on convective cloud and precipitation within a global model, and present an estimate of their magnitude in comparison to those on resolved-scale stratiform cloud.