Benjamin's new convection scheme

Centre of Excellence researcher Benjamin Möbis has developed a new convection scheme that is in the early stages of implementation in the Max Planck Institute's climate model, ICON. The scheme is one of the first to bring multi-updraught capability to climate models.

In the majority of models only one form of convection can occur at a time. This tends to lead to abrupt shifts between either shallow or deep convection leading to peculiar transitions and a poor representation of the diurnal cycle of rainfall. Observations on the other hand show the coexistence of shallow and deep convection as well as smooth transitions between shallow and deep dominated states.

Benjamin’s scheme allows multiple updraughts of different strengths and depth to occur simultaneously. This means convection processes can move gradually from shallow to deep convection, treating the process like a continuum. The new scheme also includes a congestus type, which is an intermediate mode between shallow and deep convection. Taking this approach creates a pathway for deep convection to develop out of shallow convection and even pass through the intermediate congestus stage.

Benjamin says importantly the new scheme also has a “memory” by using prognostic rather than diagnostic equations, another feature absent in almost all current convection schemes. Allowing for memory makes for a smoother growth of deep convection and suppresses unrealistic rapid transitions.

Benjamin selected ICON because of our close collaboration with the Max Planck Institute (MPI) in Hamburg, where he had previously done his PhD. This experiene meant he was familiar with the treatment of convection in the ICON model.

"It would have been very hard to modify the convection scheme of other models without extensive knowledge about the models," Benjamin said.

"One thing is certain, it is not a plug and play job."

Even at this early stage of the new convection scheme is bringing home some interesting results and Benjamin suspects it is likely to improve forecasting capabilities.

The new scheme uses the stochastic multi-cloud model (SMCM) to decide on the existence and strength of the different types of convection. The SMCM is “trained” with radar data from Darwin, so its performance is expected to be more realistic than many existing approaches. It is hoped that the new scheme will lead to improvements of the diurnal cycle, the Madden-Julien Oscillation and other equatorially trapped waves as well as the Inter Tropical Convergence Zone.

This kind of fundamental modelling work comes with a variety of challenges for those working on it. Changing model schemes is long-term hard work that produces very few papers but which can have a profound impact on the simulations and predictions produced by climate models.

The lack of publications and long-term commitment required is one of the reasons why few people work on model development in the university sector. It’s also why only few groups work on convection schemes at all even though they are often cited as a main source of systematic model biases and hence model uncertainty.

The vast majority of convection scheme development is done by government-funded institutions, such as the Met Office or the European Centre for Medium Range Forecasts, which depend far less on project money than universities and are not trapped in a publish or perish model of research.

Benjamin said there had hardly been any improvement of the systematic biases climate models show in the tropics over the past decades.

But with his change in convection scheme, Benjamin hopes to bring the focus back to this area. It’s a focus that should lead to improved results in climate simulation and, as a result, improved decision-making.

UNSW logo ANU logo Monash logo UMelb logo UTAS logo