Ocean-atmosphere interactions in cyclones and their predictability

Abstract

The first chapter investigates sea surface temperature (SST) as a predictor of western North Pacific tropical cyclones intersecting a domain over southern and central China. Models exist for seasonal prediction of tropical cyclones, but generally they are basin-wide and issued ~3 months before the storm season. This study has revealed a robust empirical signal relating SST off the east coast of mainland China, in the East China Sea and in the Bay of Bengal ~18 months prior to the tropical cyclone season, to activity within the selected domain; however, failed to produce a skilful empirical forecast. The following two chapters focus on the impact of the Gulf Stream on North Atlantic extra-tropical cyclones. The primary diagnostic analysed is turbulent kinetic energy (on scales of 10 to 100 kms), motivated by Sheldon et al., 2017. They proposed a mechanism for the interaction between ocean and atmosphere, the ‘warm path’, in which the Gulf Stream warm tongue influences atmospheric instability within an overlying cyclone. A single storm in an 11 member ensemble hindcast, a control forecast, and ERA-Interim reanalysis were analysed. However, even with a very similar warm tongue set up in all ensemble members, the magnitude of turbulent kinetic energy was variable by more than 60%. For the final chapter, mesoscale signatures of the North Atlantic Oscillation (NAO) in the atmosphere and ocean were identified. In positive NAO winters, an extension of the Gulf Stream warm tongue is observed, along with a displacement and acceleration of surface currents. Mesoscale activity in the atmosphere appears to follow this extension of the warm tongue, which is a new feature of NAO air-sea interactions embedded within the basin scale NAO patterns. This result highlights the importance of model resolution and the presence of unresolved processes in many climate models, which may have downstream impacts and climate implications.