The connected ocean: comparing inter-ocean transport at the surface and at depth
File(s)
Author(s)
McAdam, Ronan
Type
Thesis
Abstract
The oceans form a global network of inter-connected basins, with currents transporting a range
of tracers such as heat and pollution between them. Although the ocean dynamics play a key
role in determining transport between basins, the basins are de ned by arbitrary geographical
borders rather than physical connections. This thesis proposes a method of detecting dynamical
boundaries and applies it to the ocean surface and interior. To achieve this, probabilistic
models are used to summarise the transport described by vast datasets of water-following
trajectories. Such models, known as Markov Chain models, have previously been used to map
the distribution of plastic pollution yet are known to induce an artificial dispersion. First,
a sensitivity analysis of the effect of model parameters on artificial dispersion is performed,
to determine the optimal model set-up. Next, a global dataset of observed trajectories is
used to detect dynamic transport barriers in two key areas of inter-ocean surface exchange: the
Agulhas Current system and the North Atlantic inter-gyre transport barrier. Connectivity maps
are introduced as maps of tracer destinations which highlight dynamical segregation between
regions. For example, these are used to identify the source region for Agulhas Leakage and one-way
equator-ward transport across the Gulf Stream. Using a new method to extract geostrophic
motion from a trajectory dataset, the geostrophic contribution to inter-basin transport can
be identified. Finally, connectivity maps are produced for the ocean interior using virtual
particles released along isopycnals in the eddy-permitting ORCA025 ocean circulation model.
The change in connectivity, in the two study regions, between two target isopycnals is the first
step in creating a 3D border of the ocean basins. New basin definitions are then explained
by transport processes relevant to the region. Future work should establish the link between
dynamical boundaries and the distribution of heat and pollution.
of tracers such as heat and pollution between them. Although the ocean dynamics play a key
role in determining transport between basins, the basins are de ned by arbitrary geographical
borders rather than physical connections. This thesis proposes a method of detecting dynamical
boundaries and applies it to the ocean surface and interior. To achieve this, probabilistic
models are used to summarise the transport described by vast datasets of water-following
trajectories. Such models, known as Markov Chain models, have previously been used to map
the distribution of plastic pollution yet are known to induce an artificial dispersion. First,
a sensitivity analysis of the effect of model parameters on artificial dispersion is performed,
to determine the optimal model set-up. Next, a global dataset of observed trajectories is
used to detect dynamic transport barriers in two key areas of inter-ocean surface exchange: the
Agulhas Current system and the North Atlantic inter-gyre transport barrier. Connectivity maps
are introduced as maps of tracer destinations which highlight dynamical segregation between
regions. For example, these are used to identify the source region for Agulhas Leakage and one-way
equator-ward transport across the Gulf Stream. Using a new method to extract geostrophic
motion from a trajectory dataset, the geostrophic contribution to inter-basin transport can
be identified. Finally, connectivity maps are produced for the ocean interior using virtual
particles released along isopycnals in the eddy-permitting ORCA025 ocean circulation model.
The change in connectivity, in the two study regions, between two target isopycnals is the first
step in creating a 3D border of the ocean basins. New basin definitions are then explained
by transport processes relevant to the region. Future work should establish the link between
dynamical boundaries and the distribution of heat and pollution.
Version
Open Access
Date Issued
2019-03
Date Awarded
2019-09
Copyright Statement
Creative Commons Attribution NonCommercial No Derivatives Licence
Advisor
van Sebille, Erik
Czaja, Arnaud
Sponsor
Imperial College London
Publisher Department
Physics
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)