7.5 Modelling of Oceanic General Circulation
223
• the subpolar gyres in the North Atlantic and Pacific basins, and
• the flow in the Greenland-Norwegian Sea.
The OGCM velocities computed for these currents are too weak and their patterns are too broad compared with reality. However, the total transport of
the currents is similar to that observed in the oceans. A discussion on other
solution features can be found in a review article by McWilliams (1996).
Ocean models used in simulation of the Earth's climate do not resolve mesoscale eddies because of the computational cost. Therefore in the short term,
eddy-resolving global ocean models cannot be used in climate simulations, even
on modern supercomputers. On the other hand, energetic mesoscale eddies are
important in the transport of heat, salt, and passive tracers such as radiocarbon and freon in the oceans. Their importance has been documented following observations in the Antarctic Circumpolar Current, equatorial Pacific and
North Atlantic oceans. Thus, there is a need for parameterization of mesoscale
eddies for climate models. Recently, such parameterization of the effects of
these eddies, based on an adiabatic down-gradient diffusion of the thickness
between neighbouring isopycnal surfaces has been proposed by Danabasoglu et
al. (1994).
A very important motivation for developing ocean general circulation models is their use in coupled atmosphere-ocean models to study climate and its
changes. More than fifty years ago, Sverdrup (1945) wrote: It is not yet possible to deal with the system atmosphere-ocean as one unit, but it is obvious that,
in treating separately the circulation of the atmosphere, a thorough consideration of the interaction between the atmosphere and the oceans is necessary.
The first simulation of climate with a combined ocean-atmosphere model was
that discussed by Manabe (1969) and Bryan (1969). A nine-level atmospheric
model, known as the GFDL (Geophysical Fluid Dynamics Laboratory) oceanatmosphere model, was used to calculate values of atmospheric variables on a
grid with space dimensions of about 500 km, and a five-level oceanic model
(Bryan, 1969). In spite of many simplifying assumptions, the model accurately
simulates seasonal patterns of rainfall in the tropics and associated wind fields.
The locations of rainbelts and associated disturbances are determined primarily
by the distribution of the sea surface temperature.
Another coupled atmosphere and ocean model was developed at the National
Center for Atmospheric Research (NCAR), the United States (Washington et
al., 1980). This model links separate existing models of the atmosphere, ocean
and sea ice. The atmospheric part of the model uses a generalized vertical
coordinate with eight layers, each ~ 3 km thick, and 5° horizontal grid spacing
over the entire Earth. The ocean model is a modification of the GFDL model
mentioned above, developed by Bryan (1969). In the sea ice part of the model,
a simplified calculation of heat flux through sea ice is used. As the density
of the atmosphere is about 1000 times smaller than that of sea water, the
coupled atmosphere-ocean model should deal with two different time scales,
for the atmosphere and the ocean environments. In the NCAR model, an
223
• the subpolar gyres in the North Atlantic and Pacific basins, and
• the flow in the Greenland-Norwegian Sea.
The OGCM velocities computed for these currents are too weak and their patterns are too broad compared with reality. However, the total transport of
the currents is similar to that observed in the oceans. A discussion on other
solution features can be found in a review article by McWilliams (1996).
Ocean models used in simulation of the Earth's climate do not resolve mesoscale eddies because of the computational cost. Therefore in the short term,
eddy-resolving global ocean models cannot be used in climate simulations, even
on modern supercomputers. On the other hand, energetic mesoscale eddies are
important in the transport of heat, salt, and passive tracers such as radiocarbon and freon in the oceans. Their importance has been documented following observations in the Antarctic Circumpolar Current, equatorial Pacific and
North Atlantic oceans. Thus, there is a need for parameterization of mesoscale
eddies for climate models. Recently, such parameterization of the effects of
these eddies, based on an adiabatic down-gradient diffusion of the thickness
between neighbouring isopycnal surfaces has been proposed by Danabasoglu et
al. (1994).
A very important motivation for developing ocean general circulation models is their use in coupled atmosphere-ocean models to study climate and its
changes. More than fifty years ago, Sverdrup (1945) wrote: It is not yet possible to deal with the system atmosphere-ocean as one unit, but it is obvious that,
in treating separately the circulation of the atmosphere, a thorough consideration of the interaction between the atmosphere and the oceans is necessary.
The first simulation of climate with a combined ocean-atmosphere model was
that discussed by Manabe (1969) and Bryan (1969). A nine-level atmospheric
model, known as the GFDL (Geophysical Fluid Dynamics Laboratory) oceanatmosphere model, was used to calculate values of atmospheric variables on a
grid with space dimensions of about 500 km, and a five-level oceanic model
(Bryan, 1969). In spite of many simplifying assumptions, the model accurately
simulates seasonal patterns of rainfall in the tropics and associated wind fields.
The locations of rainbelts and associated disturbances are determined primarily
by the distribution of the sea surface temperature.
Another coupled atmosphere and ocean model was developed at the National
Center for Atmospheric Research (NCAR), the United States (Washington et
al., 1980). This model links separate existing models of the atmosphere, ocean
and sea ice. The atmospheric part of the model uses a generalized vertical
coordinate with eight layers, each ~ 3 km thick, and 5° horizontal grid spacing
over the entire Earth. The ocean model is a modification of the GFDL model
mentioned above, developed by Bryan (1969). In the sea ice part of the model,
a simplified calculation of heat flux through sea ice is used. As the density
of the atmosphere is about 1000 times smaller than that of sea water, the
coupled atmosphere-ocean model should deal with two different time scales,
for the atmosphere and the ocean environments. In the NCAR model, an