Evaluating Southern Ocean cloud biases in ACCESS1.3 using hybrid cloud regimes


  • Cloud biases in representing current climate, and effects for predicting cloud feedback in warming climate.

  • Use of cloud regimes to aid model evaluation (identifying compensating biases)

  • Use of ISCCP-style cloud regimes to for observational studies and model evaluation.

  • Issues relating to identifying ISCCP-style cloud regimes from models:

    • Identifying inconsistent cloud regimes in each model.

    • Applying modelled clouds to observed cloud regimes.

  • In this study we present a hybrid approach, using both observed and modelled cloud to identify cloud regimes that are common to the model and observations, as well as cloud regimes found only in the models.

  • We evaluate the ACCESS1.3 GCM; a basic comparison of this model against observations using the ISCCP-simulator is given in Figure \ref{fig:hist_sim-obs}: we note that optically thick low and mid-topped cloud are strongly under-represented, while optically thin cloud is over-represented.

  • NOTE: significant under-estimate in cloud about! May be the same problem noticed by John Haynes. Shit.

\label{fig:hist_sim-obs} The difference between observed and simulated cloud in the region of interest (50–65 S) during Austral summer (DJF), based on 2 yr of model data.

In this paper we present a hybrid approach to model evaluation using cloud regimes derived from both observations and simulations, applied to the ACCESS model:

  • In section 1 we describe the satellite observations, re-analysis and climate models used herein.

  • In section 2 we detail the methodology of identifying the hybrid cloud regimes and present their key properties in observations and in the model.

  • In section 3 we continue the model evaluation by regarding the occurrence of the cloud regimes in the context of a composite extratropical cyclone.

  • In section 4 we compare the top-of-atmosphere radiative fluxes from the models against observations, and decompose the shortwave CRE bias into components related to the frequency of occurrence of the cloud regimes and the radiative properties of the cloud regimes.

  • In section 5 we use the hybrid cloud regimes to evaluate the effect of changing the microphysics parameterisation, demonstrating the utility of the hybrid cloud regimes as a tool in model development.

  • Section 6 is the discussion and conclusion.

Data and methods

Satellite observations


Global climate model


From Bi et al (2013): Uses the UK MetOffice Global Atmosphere model 1.0 (GA1.0 Hewitt et al., 2011):

  • The PC2 prognostic cloud scheme (Wilson et al., 2008)

  • Does not implement Wood et al. (2007) boundary layer solver

  • PC2 is modified according to Franklin et al. (2012); changes to ice cloud fraction and cloud area scheme.