Modeling Ocean Circulation
35
worked on a study of the Southern Ocean (Gill and Bryan, 1971) that provided some
key insights on dynamical constraints imposed by the Drake Passage.
Perhaps the first ocean-only calculation to have an immediate impact on current
oceanographic research was a study of the Indian Ocean (Cox, 1970; Bryan, 1991).
During the Indian Ocean Program Cox used the newly developed Weather Bureau
ocean model to simulate the monsoon-driven seasonal cycle of surface currents in
the geometry of the Indian Ocean. The calculation was an extremely ambitious one
for the primitive computers of the day. To allow for the long adjustment times of
the ocean’s main thermocline, the calculation was carried out in several stages in
which the numerical grid was refined at each stage. Cox showed how the newly
discovered, large-scale, equatorially trapped waves interacted with coastal upwelling
along the coast of Africa. This study illustrated how a numerical simulation could be a
powerful tool to link ocean theory and observation. In retrospect, the Cox study makes
another important point. Synoptic eddies in the ocean, the equivalent of atmospheric
weather, vary greatly in horizontal scale with respect to latitude. At the equator they
are quite large, of the order of hundreds of kilometers. In higher latitudes differences
in stratification and the local Coriolis force make synoptic scales quite small, of the
order of 5–50 kilometers. With limited computer power ocean circulation models
could be much more successful in simulating the synoptic scales of tropical oceans
than synoptic scales at higher latitudes. This point is borne out by the relatively early
success of numerical simulations of the El Ni˜ no phenomenon by Philander (1990),
while models are only recently reproducing the correct path of the Gulf Stream and
the Gulf Stream extension in the North Atlantic (Smith et al., 2000).
By the early 1970s numerical models of the ocean circulation were recognized
in some circles, at least, as an important new field of research. A conference was organized by the U.S. National Academy of Sciences (NAS, 1975) at the University of
New Hampshire in October 1972. The conference was small, but included scientists
from the former Soviet Union, Sweden, and Germany. Each invited talk was accompanied by an invited review, and the discussion was documented. The conference
report published in 1975 is a remarkable time capsule of the state of ocean circulation
modeling, and what scientists thought about the subject in that evolutionary period.
Sarkisyan from the Soviet Union used diagnostic calculations based on geostrophic
balance, which suggested the importance of bottom pressure torques, where strong
currents flow over bottom topography. At the time this was a very new idea, since
most existing theories of ocean circulation only dealt with wind forcing at the surface.
However, the difficulties of making accurate calculations of this effect aroused skepticism. Even today, four decades later, bottom pressure torques and their role in guiding
the path of western boundary currents are controversial. The first attempts to model
the entire World Ocean were shown in papers by Cox (NAS,1975) and Takano et al.
(NAS,1975). Cox presented a computation in which hydrographic data were inserted
as an initial condition. The model was integrated forward for a few years, allowing a
simulation of the World Ocean circulation to develop in response to surface winds and
the pressure forces implied by data. Cox’s model used a 2
◦
× 2
◦ latitude–longitude
35
worked on a study of the Southern Ocean (Gill and Bryan, 1971) that provided some
key insights on dynamical constraints imposed by the Drake Passage.
Perhaps the first ocean-only calculation to have an immediate impact on current
oceanographic research was a study of the Indian Ocean (Cox, 1970; Bryan, 1991).
During the Indian Ocean Program Cox used the newly developed Weather Bureau
ocean model to simulate the monsoon-driven seasonal cycle of surface currents in
the geometry of the Indian Ocean. The calculation was an extremely ambitious one
for the primitive computers of the day. To allow for the long adjustment times of
the ocean’s main thermocline, the calculation was carried out in several stages in
which the numerical grid was refined at each stage. Cox showed how the newly
discovered, large-scale, equatorially trapped waves interacted with coastal upwelling
along the coast of Africa. This study illustrated how a numerical simulation could be a
powerful tool to link ocean theory and observation. In retrospect, the Cox study makes
another important point. Synoptic eddies in the ocean, the equivalent of atmospheric
weather, vary greatly in horizontal scale with respect to latitude. At the equator they
are quite large, of the order of hundreds of kilometers. In higher latitudes differences
in stratification and the local Coriolis force make synoptic scales quite small, of the
order of 5–50 kilometers. With limited computer power ocean circulation models
could be much more successful in simulating the synoptic scales of tropical oceans
than synoptic scales at higher latitudes. This point is borne out by the relatively early
success of numerical simulations of the El Ni˜ no phenomenon by Philander (1990),
while models are only recently reproducing the correct path of the Gulf Stream and
the Gulf Stream extension in the North Atlantic (Smith et al., 2000).
By the early 1970s numerical models of the ocean circulation were recognized
in some circles, at least, as an important new field of research. A conference was organized by the U.S. National Academy of Sciences (NAS, 1975) at the University of
New Hampshire in October 1972. The conference was small, but included scientists
from the former Soviet Union, Sweden, and Germany. Each invited talk was accompanied by an invited review, and the discussion was documented. The conference
report published in 1975 is a remarkable time capsule of the state of ocean circulation
modeling, and what scientists thought about the subject in that evolutionary period.
Sarkisyan from the Soviet Union used diagnostic calculations based on geostrophic
balance, which suggested the importance of bottom pressure torques, where strong
currents flow over bottom topography. At the time this was a very new idea, since
most existing theories of ocean circulation only dealt with wind forcing at the surface.
However, the difficulties of making accurate calculations of this effect aroused skepticism. Even today, four decades later, bottom pressure torques and their role in guiding
the path of western boundary currents are controversial. The first attempts to model
the entire World Ocean were shown in papers by Cox (NAS,1975) and Takano et al.
(NAS,1975). Cox presented a computation in which hydrographic data were inserted
as an initial condition. The model was integrated forward for a few years, allowing a
simulation of the World Ocean circulation to develop in response to surface winds and
the pressure forces implied by data. Cox’s model used a 2
◦
× 2
◦ latitude–longitude