7.2.1 Remarks on the history
of ocean modelling
When the WOCE programme commenced in
the early 1980s, development of ocean models was
heading in two different and almost mutually
exclusive directions. On the one hand, models
based on the Quasi-Geostrophic (QG) approximation with a small number of adiabatic layers and an
idealized geometrical configuration were used to
study dynamical aspects of ocean circulation on the
basin scale, focusing on explicitly resolved instability processes and mesoscale eddies but excluding
diabatic processes such as the thermohaline circulation (e.g. Holland, 1978). On the other hand, models based on the full thermodynamic Primitive
Equation (PE) system were used to simulate the
global-scale circulation in a realistic geometry, and
to study the ocean’s role in global climate (e.g.
Bryan and Lewis, 1979). The coarse resolution of
PE models prevented them from properly resolving
western boundary currents and other dynamically
important, sub-basin-scale processes. There was
reason to expect that both lines of development
would eventually converge, and a ‘marriage’ of
both model types was actually predicted (WCRP,
1983). That prediction was subsequently realized
with the completion of the WOCE Community
Modelling Experiment (Bryan and Holland, 1989).
Over the last decade or so the status of ocean
modelling has changed dramatically. Increased
computer power has permitted gains in model
resolution and integration times, the WOCE
programme has provided better observations for
validation, and observational and theoretical studies have led to a better understanding and parameterization of unresolved processes. These
improvements have been accompanied by a substantial increase in the size of the modelling community. Again, however, two lines of modelling
exist that are fairly distinct, although both are
based on the fully thermodynamic PE formulation.
Models of high resolution (around 10 km) now
exist that resolve energetic boundary currents and
associated instability processes on basin to global
scales, and which describe many aspects of the
ocean circulation with a considerable degree of
realism. While these models still have deficiencies
(see Böning and Semtner, Chapter 2.2), they begin
to give a dynamically consistent description of the
ocean circulation down to spatial scales of order
of the first baroclinic deformation radius. Due to
the immense computing requirements, they are
however restricted to relatively short integration
times and/or to single ocean basins, and are thus
not immediately suitable for use in climate studies.
In contrast, ocean climate models have changed
only moderately. They differ from their ancestors
in having somewhat better resolution (typically
100–400 km, cf. Wood and Bryan, Chapter 2.3)
and moderately improved parameterizations of
sub-grid-scale mixing and the effects of mesoscale
eddies. The applications of these models have
however greatly expanded. They are used in very
long integrations for climate purposes, are often
7.2
Numerical Ocean Circulation Modelling:
Present Status and Future Directions
J.Willebrand and D. B. Haidvogel
547
OCEAN CIRCULATION AND CLIMATE
Copyright © 2001 Academic Press
ISBN 0-12-641351-7
All rights of reproduction in any form reserved
CHAPTER
of ocean modelling
When the WOCE programme commenced in
the early 1980s, development of ocean models was
heading in two different and almost mutually
exclusive directions. On the one hand, models
based on the Quasi-Geostrophic (QG) approximation with a small number of adiabatic layers and an
idealized geometrical configuration were used to
study dynamical aspects of ocean circulation on the
basin scale, focusing on explicitly resolved instability processes and mesoscale eddies but excluding
diabatic processes such as the thermohaline circulation (e.g. Holland, 1978). On the other hand, models based on the full thermodynamic Primitive
Equation (PE) system were used to simulate the
global-scale circulation in a realistic geometry, and
to study the ocean’s role in global climate (e.g.
Bryan and Lewis, 1979). The coarse resolution of
PE models prevented them from properly resolving
western boundary currents and other dynamically
important, sub-basin-scale processes. There was
reason to expect that both lines of development
would eventually converge, and a ‘marriage’ of
both model types was actually predicted (WCRP,
1983). That prediction was subsequently realized
with the completion of the WOCE Community
Modelling Experiment (Bryan and Holland, 1989).
Over the last decade or so the status of ocean
modelling has changed dramatically. Increased
computer power has permitted gains in model
resolution and integration times, the WOCE
programme has provided better observations for
validation, and observational and theoretical studies have led to a better understanding and parameterization of unresolved processes. These
improvements have been accompanied by a substantial increase in the size of the modelling community. Again, however, two lines of modelling
exist that are fairly distinct, although both are
based on the fully thermodynamic PE formulation.
Models of high resolution (around 10 km) now
exist that resolve energetic boundary currents and
associated instability processes on basin to global
scales, and which describe many aspects of the
ocean circulation with a considerable degree of
realism. While these models still have deficiencies
(see Böning and Semtner, Chapter 2.2), they begin
to give a dynamically consistent description of the
ocean circulation down to spatial scales of order
of the first baroclinic deformation radius. Due to
the immense computing requirements, they are
however restricted to relatively short integration
times and/or to single ocean basins, and are thus
not immediately suitable for use in climate studies.
In contrast, ocean climate models have changed
only moderately. They differ from their ancestors
in having somewhat better resolution (typically
100–400 km, cf. Wood and Bryan, Chapter 2.3)
and moderately improved parameterizations of
sub-grid-scale mixing and the effects of mesoscale
eddies. The applications of these models have
however greatly expanded. They are used in very
long integrations for climate purposes, are often
7.2
Numerical Ocean Circulation Modelling:
Present Status and Future Directions
J.Willebrand and D. B. Haidvogel
547
OCEAN CIRCULATION AND CLIMATE
Copyright © 2001 Academic Press
ISBN 0-12-641351-7
All rights of reproduction in any form reserved
CHAPTER