Skip to content

Sensitivity of the aerosol indirect effect to subgrid variability in the cloud parameterization of the GFDL Atmosphere General Circulation Model AM3

July 1st, 2011


Key Findings

  • Conducted a physics perturbation experiment by constructing three alternate atmospheric model configurations with different, but justifiable, cloud assumptions.
  • Alternate configurations predict similar present-day climates, yet differ markedly in their magnitude of the aerosol cloud indirect effects.
  • This study is a reminder of the impact of uncertainties in current cloud parameterizations.

The recently developed GFDL AM3 model (Donner et. al 2011) incorporates a prognostic treatment of cloud drop number to simulate the aerosol indirect effect. The present work explores formulation sensitivities by constructing three alternate model configurations (S1, S2, S3). These alternate configurations exhibit only small differences in their present day climatology (Figure 1).

Figure 1: Comparison of zonal averages between the reference (REF) AM3 model and three alternate configurations (S1, S2, S3). Differences are generally small, especially when compared to available observations.
Figure 1: Comparison of zonal averages between the reference (REF) AM3 model and three alternate configurations (S1, S2, S3). Differences are generally small, especially when compared to available observations.

However, the total anthropogenic radiative flux perturbation (RFP), which measures the impact of greenhouse gases, aerosol direct and indirect effects between present-day (PD) and preindustrial (PI) conditions, varies by +/-50% from the reference (REF). This difference is attributed to a large sensitivity of the aerosol indirect effect to specific formulation choices. In particular, we show a linear correlation between the choice of the autoconversion threshold radius in the cloud parameterization (a standard “tuning” parameter) and the RFP (Figure 2).

Figure 2: Correlation between the total anthropogenic radiative flux perturbation (RFP) and the autoconversion threshold. The change in RFP is caused by a change in the aerosol indirect effect.
Figure 2: Correlation between the total anthropogenic radiative flux perturbation (RFP) and the autoconversion threshold. The change in RFP is caused by a change in the aerosol indirect effect.

Our results serve as a reminder that uncertainties in the formulation of individual components of a cloud parameterization can translate into significant uncertainties in the aerosol indirect effect. Reducing these uncertainties will necessitate more realistic physical representation of cloud processes and better observational constraints.

For more information: Golaz, J-C, M Salzmann, Leo J Donner, Larry Horowitz, Yi Ming, and Ming Zhao (2011), Sensitivity of the aerosol indirect effect to subgrid variability in the cloud parameterization of the GFDL Atmosphere General Circulation Model AM3. Journal of Climate, 24(13), doi:10.1175/2010JCLI3945.1.