CLIMATE
VARIABILITY
RESEARCH
Idealised Extra-Tropical Cyclone Life-Cycle Experiments in the Unified Model
The wintertime flow in the mid-latitude troposphere is characterised by the growth and decay of extra-tropical cyclones. The momentum, temperature and moisture transports by these weather systems are vitally important for the maintenance of the observed time-mean extra-tropical flow. It therefore follows that an understanding of how well the time-mean flow is captured within a climate model is crucially dependent upon how well extra-tropical cyclones are represented in that model.
A related issue is how the representation of extra-tropical cyclones in climate models is dependent upon the choice of physical parametrizations and the horizontal and vertical resolution. One particular issue arises from the use of the Unified Model for both climate research and NWP, which allows the same weather system to be modelled at a range of horizontal resolutions. The precipitation from a frontal system in a modelled extra-tropical cyclone occurs from both the sub-grid scale convective parametrization and the condensation from the dynamical lifting of warm, moist air. As the horizontal resolution of the model is increased the amount of dynamic precipitation increases at the expense of the parametrized convective precipitation.
The processes that occur in the transition from sub-grid scale parametrized convection to resolved sloping convection are difficult to interpret in a long integration of a climate model, where cause and effect are often blurred. For similar reasons the use of analysed initial states to determine the processes occurring in observed extra-tropical cyclones is also precluded, because of the significant and very different spin-ups of the differing configurations of the Unified Model from analyses.
An alternative is to investigate these processes within the simpler context of idealised extra-tropical cyclone life-cycle studies. These experiments mirror those carried out in the mid-seventies and early eighties where the non-linear development of baroclinic waves from zonally symmetric idealised mid-latitude jet streams were investigated. Understanding how baroclinic waves grew in simple models shed light on the mechanisms by which extra-tropical cyclones developed in the real troposphere. However these simple models lacked any representation of water vapour and offered nothing more complex in the way of physics than simple friction and thermal damping terms.
The UKMO atmospheric climate model, HadAM3, has a complex package of physical parametrizations, and it is of interest to see how those parametrizations influence the growth of extra-tropical cyclones. However the complexity of the physics in the model limits the choice of initial conditions, as the initial state must be in balance with the model and the model's parametrizations, e.g. an arbitrary tropical distribution of water vapour may be radically altered by the model's convective parametrization.
In these experiments the water vapour is determined by the model itself, by applying the external forcings consistent with the initial state. Perpetual wintertime radiation and the SSTs consistent with the surface temperature of the zonal mean initial state, the '1103' idealised mid-latitude jet used in Simmons and Hoskins (1978), are applied to the aqua planet configuration of HadAM3. The model is then integrated for three months to provide a relative humidity climatology that is used to specify the water vapour for the initial 1103 state.
Figure 1 shows the surface pressure five days into an integration of HadAM3 at climate resolution (96x73 grid points), from an initial 1103 zonal mean state with a 1mb circular surface pressure perturbation at 60W and 45N. This perturbation deepens exponentially at a growth rate of 0.85 1/days, reaching a maximum on day six. The hope is that the results from these experiments will provide information on the sensitivities seen in longer climate integrations, and that life-cycle studies will provide a convenient test-bed for investigating the impact of changes in the resolution and parametrizations of the model.
References
A. J. Simmons and B. J. Hoskins (1978), The life-cycles of some non-linear baroclinic waves. J. Atmos. Sci. 35 pp 414-432.
Len Shaffrey
CGAM, University of Reading
swrshaff@met.rdg.ac.uk
