Understanding and Improving the Path of the Gulf Stram extension in moderate resolution climate models

Much of the transfer of heat from the ocean to the atmosphere in the Atlantic takes place over the Gulf Stream and the atmosphere has been known for many years to be very sensitive to the ocean surface temperatures in the region of the Gulf Stream extension. The coarse resolution ocean models with 1o grid spacing that have been used in nearly all climate simulations reported by the IPCC misrepresent the path of the Gulf Stream extension, the flow being too zonal and too far to the south. High resolution ocean models (with 1/12o grid spacing) reproduce the path well and some ocean models with 1/4o grid spacing that have recently been used in seasonal prediction also have an improved path. Scaife et al. (2011) attribute much of the improvement in the prediction of winter-time blocking events in their system to this improved path. Climate simulations and seasonal predictions are likely to use models of this resolution for at least a decade and the accuracy of their representation of the path of the Gulf Stream extension will be an important factor. The robustness of the path at this resolution is by no means assured so it is important to understand better what determines the path and to seek techniques which can provide good simulations at relatively coarse resolutions for the right reasons.

There have been a number of past contributions to progress in understanding the dynamics that determine this path: in particular Greatbatch et al. (1991), Zhang & Vallis (2007). They showed that the contributions from the vortex stretching and bottom pressure torques associated with the flow over bathymetry and the viscous torques are important in determining the direction of the depth integrated flow. Mesoscale instabilities are known to have an important role in strengthening the flow at depth and hence the steering by the bathymetry (Hurlburt & Hogan 2008). The LANS-a representation of the momentum equations, formulated using generalised Lagrangian mean theory allows, a better representation of the kinetic energy of the mesoscale when it is only partially resolved. Peterson, Hecht, and Wingate implemented a simplified form of LANS-a in the POP ocean model and found that the steering of the depth integrated flow by the bathymetry in a model with a 0.4o grid was significantly improved. Barnier et al. (2006) have also shown that the various options for the formulation of momentum advection within NEMO have a significant influence on the barotropic flow in ORCA025 in regions where there are strong interactions with bathymetry. Issues with the treatment of bathymetry in this formulation have recently come to light and one of the supervisors is developing a revised scheme in collaboration with Barnier and co-workers. Therefore new, modified, or improved combinations of parameterizations and numerical implementations of the dynamics may make an important difference to the robustness of the Gulf Stream extension's path at coarse resolution.

The first step in the PhD project will be to calculate diagnostics of vorticity tendencies and the Gulf Stream path using models of 1o, 1/4o and 1/12o resolution. Depending on the results and the interests of the student the PhD could develop in the following directions: in-depth investigation and interpretation of aspects of the vorticity diagnostics, the theory and dynamics of western boundary currents in the presence of sloping bathymetry, assessment of the impact of alternative representations of momentum advection or physical parametrisations of eddy fluxes on the Gulf Stream path. The student will be encouraged to formulate and experiment with novel ideas in numerics, diagnostics and theory for ocean physics and dynamics. Research questions to be addressed include:
1) What processes control the vorticity in the Gulf Stream extension?
2) What are the 1) roles of mesoscale fluctuations in western boundary currents?
2) How do mesoscale fluctuations strengthen the interactions with bathymetry