Identifying controls of the intensity of extra-tropical cyclones using a massive ensemble of baroclinic life cycle simulations.

Idealised simulations of extra-tropical cyclones (ETCs), often referred to as baroclinic life cycles, have previously been extensively used to understand ETC dynamics. More recently, baroclinic life cycle simulations have proved useful to investigate how ETCs may respond to climate change. However, these simulations are often difficult to set up and previous studies that have used baroclinic life cycle simulations to understand cyclone dynamics often include only a handful of different initial background states. Here we introduce a flexible set up for both dry and moist baroclinic life cycles that can be run with and without physics that has been implemented successfully into OpenIFS. A key strength of this set-up is that via the namelist a user can control many aspects of the background state: the height, width and strength of the jet, the average temperature, the environmental lapse rate, the surface relative humidity and the surface roughness. Using this set-up we have performed a huge ensemble (more than 5000 members) of baroclinic lifecycle simulations with the aid of climateprediction.net, a volunteer computing, climate modelling project coordinated by the University of Oxford. Using machine learning techniques we investigate how the intensity, as quantified by a range of metrics, depends on the prescribed background state. Furthermore, we attempt to physically understand the statistical relations between cyclone intensity and the background state by computing energy budgets and the energy efficiency of the simulations. These results concerning controls on ETC intensity will be presented and interpreted with regards to how ETC intensity may change in the future as the climate warms.