Capturing convective cloud population dynamics with simple spectral models


Convective cloud fields are populations of evolving objects, interacting with their environment but also with each other. Such population-internal dynamics can play a key role in the immediate response of a convective cloud field to an external forcing, and its subsequent interaction with the larger-scale flow and climate. Population-internal dynamical feedbacks have been identified to play a role in shallow cumulus equilibration, cold pool dynamics and convective aggregation. Spectral cloud modeling, a method that was pioneered decades ago, offers opportunities for gaining insight into such population dynamics. Multiple degrees of freedom are purposely retained to independently model individual classes of convective elements, allowing for simple interactions within the population. This can potentially introduce negative feedback mechanisms that control the response of the convective cloud field to the larger-scale forcing. In addition, size-filtering can be applied in spectral models which creates opportunities for introducing scale-adaptive and stochastic modeling of convective transport. Conceptually the cloud scale fits well with spectral modeling; the closure problem then moves from bulk assumptions to predicting the object size distribution and its evolution in time. In this presentation such potential benefits of the spectral approach in convective modeling will be briefly reviewed and explored. The focus lies on recent results with the ED(MF)^n spectral cloud model as implemented in the Dutch Atmospheric Large Eddy Simulation (DALES) model, replacing the existing subgrid-scale parameterization for transport and clouds. The LES model is thus used as a simple non-hydrostatic larger-scale circulation model, creating a testing ground for investigating the behavior of simple spectral convection models and their interaction with the resolved flow. Using this system, the impacts of size-filtering on the behavior of the spectral scheme in the grey zone of moist boundary layer convection will be assessed. In addition, opportunities for introducing stochastic effects through the size distribution of cloud number will be discussed, making use of super-large LES results to gain insight. Finally, the possible merits of combining the spectral approach with microgrid or lattice techniques for representing convective memory are explored.