the early 1970s with programmes such as the
Mid-Ocean Dynamics Experiment (MODE), the
Coastal Upwelling Experiment (CUEA), POLYMODE (an amalgam of the Russian POLYGON
with the Mid-Ocean Dynamics Experiment) and
International Southern Ocean Studies (ISOS). In
this view, there are multiple, local, general circulations. Mean mass fluxes are perceived to be complex spatially varying fields:(x, y, z)v(x, y, z)9
where x, y, z are three space coordinates, v:
[u, v, w], is the fluid density, and the brackets
:·9 denote the true time average. But the circulation of any scalar field, C (e.g. temperature,
carbon, or potential vorticity), must be computed as :(x, y, z, t)v(x, y, z, t)C (x, y, z, t)9,
where t is time. Unfortunately,
:(x, y, z)v(x, y, z)9 :C(x, y, z)9
:(x, y, z, t)v(x, y, z, t)C(x, y, z, t)9 (2.1.1)
where the failure of the equality can be at
zero-order. An extreme case of such a failure
would be the attempt to calculate the timeaverage temperature transport of the Gulf
Stream by multiplying the time-average mass
transport (a slow, broad flow) by the timeaverage temperature field, also a broad, slowly
varying field. The true temperature transport
involves averaging the product of a very narrow
intense jet with a field having a narrow, very
large instantaneous temperature maximum. In
such a situation, the two sides of inequality
(2.1.1) can and do differ by an order of magnitude. Because the covariances between the mass
flux fields and C will be different for differing
fields C, the circulation of each will be different,
and often radically so.
Because of the regional focus, these differing
general circulations are not usually pieced
together into any sort of global picture. They
remain rather, as isolated sub-basin scale,
stand-alone pictures.
4 The high-resolution numerical modellers’ ocean.
Beginning in about 1990, basin-to-global scale
models emerged that contained variability
resembling the oceanic mesoscale eddy field.
The character of this newly emergent view of
the ocean can be seen, e.g. in the results of
Semtner and Chervin (1992), Böning et al.
(1991), Cox (1985), Smith et al. (2000) and
others. These results, when heavily averaged,
have a qualitative resemblance to some elements
of all of 1–3, but differ quantitatively from all
of them. The general circulation property fluxes
differ, as in 3, from those of mass alone,
because as in 3, the mass flux/property covariances are all different, but the scope is global.
Little communication between the apostles of these
different personalities appears to exist; nearly disjoint literatures continue to flourish.
2.1.2 The origins of WOCE
By the late 1970s, a few scientists had begun to
recognize that the various circulation personalities
listed above could be thought of as originating in
two conflicting paradigms (to use the terminology
of Kuhn, 1962), although no one seems to have
described it that way at the time. The two paradigms are:
1 The ocean is a quasi-steady, large-scale equilibrium system. I will refer to this as the ‘historical’
paradigm.
2 The ocean is a fundamentally turbulent, constantly changing, non-equilibrium system in which
no element is actually fully steady. I will call this
the ‘WOCE’ paradigm, although it would be
wrong to claim that all those who formulated and
carried out WOCE were subscribers to it.
2.1.2.1 The historical paradigm
The ocean is opaque to all forms of electromagnetic radiation at usable wavelengths, hence modern oceanography, from its beginnings in the late
nineteenth century, has been built upon an observational base acquired by physically placing an
instrument at particular positions and depths of
interest. For other technical reasons, most such
measurements have been not of the velocity field
in the ocean, but rather of scalar quantities such as
temperature, salinity, oxygen content, etc. Because
ships provided the only platform for reaching
mid-ocean locations, it took many decades to
acquire observations adequate to delineate the
bulk scalar properties of the ocean.
By great good fortune, the ocean was early on
perceived to display large-scale, temporally stable,
contourable fields of these tracers (‘tongues’, etc.;
see particularly Wüst, 1935; some of these pictures
are reproduced in Wunsch, 1996). It was thus possible to combine the scalar observations spanning
SECTION 2 OBSERVATIONS AND MODELS
48
Mid-Ocean Dynamics Experiment (MODE), the
Coastal Upwelling Experiment (CUEA), POLYMODE (an amalgam of the Russian POLYGON
with the Mid-Ocean Dynamics Experiment) and
International Southern Ocean Studies (ISOS). In
this view, there are multiple, local, general circulations. Mean mass fluxes are perceived to be complex spatially varying fields:(x, y, z)v(x, y, z)9
where x, y, z are three space coordinates, v:
[u, v, w], is the fluid density, and the brackets
:·9 denote the true time average. But the circulation of any scalar field, C (e.g. temperature,
carbon, or potential vorticity), must be computed as :(x, y, z, t)v(x, y, z, t)C (x, y, z, t)9,
where t is time. Unfortunately,
:(x, y, z)v(x, y, z)9 :C(x, y, z)9
:(x, y, z, t)v(x, y, z, t)C(x, y, z, t)9 (2.1.1)
where the failure of the equality can be at
zero-order. An extreme case of such a failure
would be the attempt to calculate the timeaverage temperature transport of the Gulf
Stream by multiplying the time-average mass
transport (a slow, broad flow) by the timeaverage temperature field, also a broad, slowly
varying field. The true temperature transport
involves averaging the product of a very narrow
intense jet with a field having a narrow, very
large instantaneous temperature maximum. In
such a situation, the two sides of inequality
(2.1.1) can and do differ by an order of magnitude. Because the covariances between the mass
flux fields and C will be different for differing
fields C, the circulation of each will be different,
and often radically so.
Because of the regional focus, these differing
general circulations are not usually pieced
together into any sort of global picture. They
remain rather, as isolated sub-basin scale,
stand-alone pictures.
4 The high-resolution numerical modellers’ ocean.
Beginning in about 1990, basin-to-global scale
models emerged that contained variability
resembling the oceanic mesoscale eddy field.
The character of this newly emergent view of
the ocean can be seen, e.g. in the results of
Semtner and Chervin (1992), Böning et al.
(1991), Cox (1985), Smith et al. (2000) and
others. These results, when heavily averaged,
have a qualitative resemblance to some elements
of all of 1–3, but differ quantitatively from all
of them. The general circulation property fluxes
differ, as in 3, from those of mass alone,
because as in 3, the mass flux/property covariances are all different, but the scope is global.
Little communication between the apostles of these
different personalities appears to exist; nearly disjoint literatures continue to flourish.
2.1.2 The origins of WOCE
By the late 1970s, a few scientists had begun to
recognize that the various circulation personalities
listed above could be thought of as originating in
two conflicting paradigms (to use the terminology
of Kuhn, 1962), although no one seems to have
described it that way at the time. The two paradigms are:
1 The ocean is a quasi-steady, large-scale equilibrium system. I will refer to this as the ‘historical’
paradigm.
2 The ocean is a fundamentally turbulent, constantly changing, non-equilibrium system in which
no element is actually fully steady. I will call this
the ‘WOCE’ paradigm, although it would be
wrong to claim that all those who formulated and
carried out WOCE were subscribers to it.
2.1.2.1 The historical paradigm
The ocean is opaque to all forms of electromagnetic radiation at usable wavelengths, hence modern oceanography, from its beginnings in the late
nineteenth century, has been built upon an observational base acquired by physically placing an
instrument at particular positions and depths of
interest. For other technical reasons, most such
measurements have been not of the velocity field
in the ocean, but rather of scalar quantities such as
temperature, salinity, oxygen content, etc. Because
ships provided the only platform for reaching
mid-ocean locations, it took many decades to
acquire observations adequate to delineate the
bulk scalar properties of the ocean.
By great good fortune, the ocean was early on
perceived to display large-scale, temporally stable,
contourable fields of these tracers (‘tongues’, etc.;
see particularly Wüst, 1935; some of these pictures
are reproduced in Wunsch, 1996). It was thus possible to combine the scalar observations spanning
SECTION 2 OBSERVATIONS AND MODELS
48