Ocean general circulation models (OGCMs) frequently operate at a grid scale larger than
the size of oceanic eddies, therefore eddying effects must be parameterised. One such example is that of eddy-induced Lagrangian transport. The current method of parameterising eddy-induced
transport assumes isotropic and dffusive behaviour. However several studies employing
numerical models (Kamenkovich et al. 2015, 2009, Berloff et al. 2002), or using satellite altimetry derived velocity fields or by taking real float data (Rypina et al. 2012) have demonstrated Lagrangian behaviour contradicting these assumptions.
We proceed by running a numerical model simulating two differing meandering oceanic
jets. Two novel spatial interpolation methods motivated by the non-divergence property of the numerical model will be tested and incorporated into a Lagrangian transport model.
Eddying Lagrangian statistics for the two jet regimes will be compared and will verify that
a new non-diffusive, anisotropic parameterisation method will need to be considered. A new flow based dispersion measure with the aim of capturing sharper diffusivity estimates at the jet
core will be motivated and introduced.
We then run three Stochastic Markov models with the aim of reproducing the observed
Lagrangian statistics and outline the applicability of some further Stochastic models. In the fi nal chapter, we decompose the
ow into Empirical Orthogonal Functions and establish that eddy-induced meridional transport is largely influenced by small-scale flow variability that cannot be realistically be captured kinematically. The eddy-induced zonal transport however can be explained by a Stokes' drift.