Behaviour of convective updraughts: A challenging problem for weather and climate models

The aim of the project is: to improve the understanding of the physical processes that determine the time evolution of convective updraughts. For example, we believe that clouds are composed of thermals; see the schematic diagram below and compare with the photo of a real cloud. But how can this schematic model be represented as an idealised model that could form the basis of the representation in weather and global climate models? The PhD student will likely find this question to be interesting and challenging. Perhaps there is a different way of representing the clouds, as say plumes? Specific questions that need to be considered include: what is the size of the cloud core region and the amount of liquid in it, and how do these change with time? What is the interaction between this core region and the descending air at the edge of the cloud? What are the characteristics of the turbulent eddies that cause mixing, as well as the circulation of the whole cloud. These questions will be addressed by using simulations of both single clouds and collections of clouds.
The simulation of convective updraughts is particularly sensitive to model resolution. The PhD research will involve numerical experiments at 25 m resolution. At this resolution, mixing between convective clouds and their environment is well resolved. The availability of a new high-resolution model (the Met Office NERC cloud model, MONC) that can efficiently run on the latest generation of supercomputers, now makes it possible to perform such simulations on large domains, such that a collection of clouds can be studied. This is a very exciting development: the figure below shows an example of the detailed cloud structures that can be simulated with such a model. Numerical experiments will likely include studies of idealized single clouds, as well as multiple clouds in a setting which is based on field campaigns. The use of these field observations provides an important check on the realism of the model simulations.