AOPP University of Oxford

Marine-ice formation occurs on a vast range of length scales: from millimetre scale frazil crystals, to consolidated sea ice a metre thick, to deposits of marine ice under ice shelves that are hundreds of kilometres long. Scaling analyses is therefore an attractive and powerful technique to understand and predict phenomena associated with marine-ice formation, for example frazil crystal growth and the convective desalination of consolidated sea ice. However, there are a number of potential pitfalls arising from the assumptions implicit in the scaling analyses. In this talk, I tease out the assumptions relevant to these examples and test them, allowing me to derive simple conceptual models that capture the important geophysical mechanisms affecting marine-ice formation. 

We are interested in finding the Probability Density Function (PDF) of high dimensional chaotic systems such as a global atmospheric circulation model. The key difficulty stems from the so called “curse of dimensionality”. Representing anything numerically in a high dimensional space seems to be just too computationally expensive. Methods applied to dodge this problem include representing the PDF analytically or applying a (typically linear) transformation to a low dimensional space. For chaotic systems these approaches often seem extremely ad-hoc with the main motivation being that we don't know what else to do.

The Lorenz 95 system is one of the simplest systems we could come up with that is both chaotic and high dimensional. So it seems like a good candidate for initial investigation. We look at two attempts to approximate the PDF of this system to an arbitrary level of accuracy, firstly using a simple Monte-Carlo method and secondly using the Fokker-Planck equation. We also describe some of the (sometimes surprising) difficulties encountered along the way.

In many places, the Antarctic and Greenland ice sheets are fringed by tongues of ice floating on the ocean, called ice shelves. Recent observations and modelling suggest that melting and disintegration of the floating ice shelves can impact ice sheet flow, and hence have consequences for sea level rise. Of particular interest are observations of channels and undulations in the ice shelf base, for which the conditions for genesis remain unclear. To build insight into the potential for melting-driven instability of the ice shelf base, this talk will consider a free boundary problem with melting at the ice-ocean interface coupled to a buoyant plume of meltwater confined below a stationary ice shelf. An asymptotic model of turbulent heat transfer in the meltwater plume reveals that melting rates depend on ice-shelf basal slope, with potentially shocking consequences for the evolving ice-shelf geometry

Ocean climate models are unlikely routinely to have sufficient

resolution to resolve the turbulent ocean eddy field. The need for the

development of improved mesoscale eddy parameterisation schemes

therefore remains an important task. The current dominant mesoscale eddy

closure is the Gent and McWilliams scheme, which enforces the

down-gradient mixing of buoyancy. While motivated by the action of

baroclinic instability on the mean flow, this closure neglects the

horizontal fluxes of horizontal momentum. The down-gradient mixing of

potential vorticity is frequently discussed as an alternative

parameterisation paradigm. However, such a scheme, without careful

treatment, violates fundamental conservation principles, and in

particular violates conservation of momentum.

A new parameterisation framework is presented which preserves

conservation of momentum by construction, and further allows for

conservation of energy. The framework has one dimensional parameter, the

total eddy energy, and five dimensionless and bounded geometric

parameters. The popular Gent and McWilliams scheme exists as a limiting

case of this framework. Hence the new framework enables for the

extension of the Gent and McWilliams scheme, in a manner consistent with

key physical conservations.