Regional Biases in Droplet Activation Parameterizations: Strong Influence on Aerosol Second Indirect Effect in the Community Atmosphere Model v5.

Thursday, 18 December 2014
Ricardo Morales, Universidad de los Andes, Civil and Environmental Engineering, Bogotá, Colombia and Athanasios Nenes, Georgia Institute of Technology, Atlanta, GA, United States
Aerosol-cloud interactions constitute one of the most uncertain aspects of anthropogenic climate change estimates. The magnitude of these interactions as represented in climate models strongly depends on the process of aerosol activation. This process is the most direct physical link between aerosols and cloud microphysical properties. Calculation of droplet number in GCMs requires the computation of new droplet formation (i.e., droplet activation), through physically based activation parameterizations. Considerable effort has been placed in ensuring that droplet activation parameterizations have a physically consistent response to changes in aerosol number concentration. However, recent analyses using an adjoint sensitivity approach showed that parameterizations can exhibit considerable biases in their response to other aerosol properties, such as aerosol modal diameter or to the aerosol chemical composition. This is a potentially important factor in estimating aerosol indirect effects since changes in aerosol properties from pre-industrial times to present day exhibit a very strong regional signature. In this work we use the Community Atmosphere Model (CAM5) to show that the regional imprint of the changes in aerosol properties during the last century interacts with the droplet activation parameterization in a way that these biases are amplified over climatically relevant regions. Two commonly used activation routines, the CAM5 default, Abdul-Razzak and Ghan parameterization, as well as the Fountoukis and Nenes parameterization are used in this study. We further explored the impacts of Nd parameterization biases in the first and second aerosol indirect effects separately, by performing simulations were droplet number was not allowed to intervene in the precipitation initiation process. The simulations performed show that an unphysical response to changes in the diameter of accumulation mode aerosol translates into extremely high Nd concentrations over South-East Asia, producing a dramatic effect on the second aerosol indirect effect. The strength of the indirect effect over South-East Asia was shown to be a dominant factor in the global estimate of aerosol indirect effect in the CAM5 model.