| dc.description.abstract | Radiative-convective equilibrium (RCE) is an idealized representation of the tropical atmosphere in which net radiative cooling fluxes are compensated by convective heat fluxes.
In large domain RCE experiments with cloud resolving models, or coarser models, convection can show large scale spontaneous organisation referred to as 'convective self-aggregation'.
This type of convective organisation happens despite uniform initial and boundary conditions and an absence of large scale forcing.
This phenomenon is important to understand, because it has a significant impact on the domain-mean climate.
It is not very well understood yet, partly due to generally poor agreement among model studies.
A contributor to this poor agreement is the sensitivity of convective self-aggregation to almost any model parameter.
In an attempt to further investigate this phenomenon, a model intercomparison project, RCEMIP, has been set up; each participating modelling group performs the same experiment with their own model.
Coarse models are not expected to accurately resolved near-surface turbulence.
One way of parametrizing the resulting underestimation of surface fluxes and prevent numerical issues in the free convection limit, is by enforcing a minimum value for the wind speed that is put into the surface layer formulation scheme.
In the proposed simulations of RCEMIP, a value of \umins = 1 m s-1 is used for models that apply this solution.
Limited domain large eddy simulations (LES) are performed using MicroHH to see what the effect of \umins = 1 m s-1 is on the evolution of the RCE state, compared to a value closer to zero.
We find that the majority of the domain in these experiments have winds below 1 m s-1 at any time during the 20 to 30 days of simulation.
Averaged over regions where the lowest-level wind speed is below 1 m s-1, latent heat fluxes are increased by as much as 20%.
This would mean that enforcing a minimum wind speed of 1 m s-1 reduces the flux contrast between the calm environment and convective regions, thereby reducing the strength of the surface flux feedback as a self-aggregation mechanism.
The spatial variance in the near-surface moisture field increases with a higher value for \umin, however.
Performing the same kind of experiment, but with a set-up that would allow self-aggregation to occur, is suggested to further understand the impact of changes to the surface flux feedback mechanism. | |
