Group seminar on 11. January, 17:00 CET
Recent progress in the simulation of the MJO and tropical modes of variability using stochastic multicloud models
Boualem Khouider
A major challenge in contemporary parameterizations of cumulus convection is associated with their inability to capture the interactions across multiple temporal and spatial scales of cloud systems in the tropics. This is mainly due to the lack of an adequate representation of subgrid variability due to organized convection at the meso- and sub-meso-scales, which is inherited from the underlying quasi-equilibrium assumption (QEA). QEA is believed to be the culprit for the poor performance of climate and earth system models in simulating the tropical rainfall variability and the associated multiple synoptic and planetary scale tropical wave modes, such as convectively coupled waves, the Madden Julian Oscillation (MJO), and the Indian monsoon intra-seasonal oscillations. This is believed to be associated with the fact that the MJO and other tropical convective systems are thermodynamically as well as dynamically coupled to convective scale processes across spacial and temporal scales, without an apparent scale separation. This has been demonstrated for instance by the success of super-parametrization and the use of cloud permitting models to simulate the MJO. A not so-expensive alternative however is offered by the use of stochastic parameterization which are capable to reintroduce some subgrid scale variability into the model without any significant computational overhead. In this talk, I will summarize some recent breakthroughs achieved in this regard by the simple replacement of the Arakawa-Schubert parameterization for deep convection, in NCEP's Climate Forecasting Model, by a stochastic multicloud model (SMCM) parmeterization. The SMCM is based of the representation of multiple cloud types, which characterize organized convection in the tropics, by an order parameter that takes discrete values on a square lattice overlaid on top of each model grid box and evolve dynamically as a Markovian process with transition probabilities depending on the large-scale (model resolved) dynamics and thermodynamics. The main result is a significant improvement in the simulation of the MJO and other tropical modes of variability as well as the tropical rainfall statistics.