over fully turbulent elements; or as seems likely,
(c) some complex combination of both. (A particularly striking example of the complexity of the
processes that maintain the tongue-like features
can be seen in the Brazil Basin float and tracer
results of Hogg and Owens (1999), where some
large-scale, laminar flow may well coexist, but
which is almost imperceptible in the data.)
But given what we see in the WOCE data and
elsewhere, and what the high-resolution models
are telling us, it seems a vast leap of faith to assert
that one can simply integrate coarse, laminar, representations of the ocean circulation for hundreds
to thousands of years and to expect that the
Lagrangian property transports of heat, fresh
water, carbon, etc., are being computed with useful skill. If one can do so, it is either a result of
extraordinary good fortune that for this particular
fluid, the turbulent fields are purely passive, or
that the General Circulation Model (GCM) builders
are so clever they have solved the problem of fully
parameterizing the turbulent elements, boundary
jets and other unresolved elements of a threedimensional rotating, stratified fluid in a complex
geometry. If the latter is actually true, it is a
remarkable achievement of computational skills, a
landmark in the history of fluid dynamics.
2.1.4 The need for global-scale
observations
2.1.4.1 Unobserved regions
When adequate observations have been unavailable to depict the world, the human race has generally reacted in one of two extreme forms. The
first form is represented by the map of New
Guinea shown in Fig. 2.1.2. With no observations,
the cartographer felt the need to show nonexistent mountains and monsters, providing an
exciting metaphor for the general unknown. The
opposite extreme is to assume that ‘the absence of
evidence is evidence of absence’ and to infer that
nothing interesting exists where nothing has been
observed. Examples are the nineteenth century
inference of a biological desert in the abyss (Mills,
1983), the modern (until about 1965) assumption
that the deep Pacific Ocean was devoid of any significant flow, and the inference of the existence of
oceanic abyssal plains in the absence of sounding
data. It is easy to make a list of phenomena that
were so widely and plausibly believed to be absent
or irrelevant that little effort was made to test the
hypothesis by observation:
¥ Things everyone ‘knew’:
¥ No interesting currents at depth on the equator (1950)
¥ Vertical temperature and salinity profiles are
smooth (1965)
¥ No interesting abyssal flow structures in the
Pacific Ocean (1965)
¥ Ocean heat transport is negligible compared
to the atmosphere (1970)
¥ The deep interior ocean moves at climatological speeds (1958)
¥ Ocean mixing is geographically uniform
(1970).
The date given is (very approximately) when the
hypothesis started to become untenable because
someone became convinced it didn’t have to be
correct, and set out to test it.
There is a parallel problem in the inference of
overly simplified depictions of the ocean circulation, a depiction with its roots in personality 1
described in Section 2.1.1 above. Broecker’s
(1991) well-known ‘conveyor belt’ has often been
used to describe the role of the ocean in climate.
The picture is a visually powerful metaphor for
the circulation; trouble occurs, however, when
scientists begin to take it literally, forgetting its
metaphoric origin. The conveyor belt has been
invoked to claim that observing the ocean climate
state is cheap and easy – because observations can
be confined to a few upper ocean XBTs in the
northern North Atlantic. As one eminent meteorologist once insisted to me, one ‘need only keep
track of the upper branch to know what the whole
ocean is doing’. The picture has also been invoked
frequently to confirm the old prejudice that the
deep ocean does nothing of interest; it has kept
alive the completely incorrect view that water
flowing through the Indonesian Passages connects
immediately and directly across the Indian Ocean
to flow around the Cape of Good Hope, among
other myths.
Schematic pictures of the circulation are of
course, very helpful, both as mnemonics and as
aids in conveying to the non-scientific community
how the system ‘works’. But when employed out
of context, they can be extremely misleading.
Consider the schematic in Fig. 2.1.3a (see Plate
2.1.3a, p. 76) of communication links between the
SECTION 2 OBSERVATIONS AND MODELS
52
(c) some complex combination of both. (A particularly striking example of the complexity of the
processes that maintain the tongue-like features
can be seen in the Brazil Basin float and tracer
results of Hogg and Owens (1999), where some
large-scale, laminar flow may well coexist, but
which is almost imperceptible in the data.)
But given what we see in the WOCE data and
elsewhere, and what the high-resolution models
are telling us, it seems a vast leap of faith to assert
that one can simply integrate coarse, laminar, representations of the ocean circulation for hundreds
to thousands of years and to expect that the
Lagrangian property transports of heat, fresh
water, carbon, etc., are being computed with useful skill. If one can do so, it is either a result of
extraordinary good fortune that for this particular
fluid, the turbulent fields are purely passive, or
that the General Circulation Model (GCM) builders
are so clever they have solved the problem of fully
parameterizing the turbulent elements, boundary
jets and other unresolved elements of a threedimensional rotating, stratified fluid in a complex
geometry. If the latter is actually true, it is a
remarkable achievement of computational skills, a
landmark in the history of fluid dynamics.
2.1.4 The need for global-scale
observations
2.1.4.1 Unobserved regions
When adequate observations have been unavailable to depict the world, the human race has generally reacted in one of two extreme forms. The
first form is represented by the map of New
Guinea shown in Fig. 2.1.2. With no observations,
the cartographer felt the need to show nonexistent mountains and monsters, providing an
exciting metaphor for the general unknown. The
opposite extreme is to assume that ‘the absence of
evidence is evidence of absence’ and to infer that
nothing interesting exists where nothing has been
observed. Examples are the nineteenth century
inference of a biological desert in the abyss (Mills,
1983), the modern (until about 1965) assumption
that the deep Pacific Ocean was devoid of any significant flow, and the inference of the existence of
oceanic abyssal plains in the absence of sounding
data. It is easy to make a list of phenomena that
were so widely and plausibly believed to be absent
or irrelevant that little effort was made to test the
hypothesis by observation:
¥ Things everyone ‘knew’:
¥ No interesting currents at depth on the equator (1950)
¥ Vertical temperature and salinity profiles are
smooth (1965)
¥ No interesting abyssal flow structures in the
Pacific Ocean (1965)
¥ Ocean heat transport is negligible compared
to the atmosphere (1970)
¥ The deep interior ocean moves at climatological speeds (1958)
¥ Ocean mixing is geographically uniform
(1970).
The date given is (very approximately) when the
hypothesis started to become untenable because
someone became convinced it didn’t have to be
correct, and set out to test it.
There is a parallel problem in the inference of
overly simplified depictions of the ocean circulation, a depiction with its roots in personality 1
described in Section 2.1.1 above. Broecker’s
(1991) well-known ‘conveyor belt’ has often been
used to describe the role of the ocean in climate.
The picture is a visually powerful metaphor for
the circulation; trouble occurs, however, when
scientists begin to take it literally, forgetting its
metaphoric origin. The conveyor belt has been
invoked to claim that observing the ocean climate
state is cheap and easy – because observations can
be confined to a few upper ocean XBTs in the
northern North Atlantic. As one eminent meteorologist once insisted to me, one ‘need only keep
track of the upper branch to know what the whole
ocean is doing’. The picture has also been invoked
frequently to confirm the old prejudice that the
deep ocean does nothing of interest; it has kept
alive the completely incorrect view that water
flowing through the Indonesian Passages connects
immediately and directly across the Indian Ocean
to flow around the Cape of Good Hope, among
other myths.
Schematic pictures of the circulation are of
course, very helpful, both as mnemonics and as
aids in conveying to the non-scientific community
how the system ‘works’. But when employed out
of context, they can be extremely misleading.
Consider the schematic in Fig. 2.1.3a (see Plate
2.1.3a, p. 76) of communication links between the
SECTION 2 OBSERVATIONS AND MODELS
52