18
Francis Bretherton
bottles spaced along a nonconducting cable lowered by winch from research ships on
the surface. Currents within and above the thermocline were inferred from assumed
geostrophic balance between the gradient of pressure in a horizontal direction and
the Coriolis force associated with motion relative to a rotating earth. The horizontal
pressure gradient was based upon differences of temperature and salinity between
sounding stations that were typically spaced hundreds of kilometers apart, converted
to density using laboratory determinations of the equation of state for seawater, and
hence to pressure (or equivalently dynamic height) using hydrostatic balance. This
inferential process gives only vertical differences in horizontal velocity, unless in
each water column there is a reference level at which the velocity itself can be independently determined. Unfortunately, the obvious candidate for reference level, the
ocean surface, was precluded by wind drift, tidal fluctuations, and uncertainties of
navigation, and no proven alternative methodology existed.
However, in the cold, dense, abyssal water below the main thermocline the
temperature and salinity seemed remarkably uniform, with horizontal differences
that in most places were measurable only on the scale of an ocean basin. It was also
known that the carbon-14 age of water deep in much of the Atlantic Ocean was at
least several hundred years, and in the Pacific it was over one thousand. These facts
suggested that the abyssal water was relatively quiescent, so horizontal velocities from
deep hydrographic soundings were typically reported relative to a “level of no motion”
that was supposed to be representative of the entire layer below the thermocline. To
account for the carbon-14 data, water that had sunk or mixed beneath the surface in
the North Atlantic had to be moving southward through the ocean basin. However, the
average velocity required was about 1 mm/s, which was well below the precision of
measurement of the remainder of the sounding. Thus, it was widely supposed that the
deep water in the open ocean was effectively quiescent and there was an identifiable
“level of no motion” below the main thermocline.
Swallow (1955) had devised an ingenious float consisting of a scaffolding tube
plugged at both ends, equipped with batteries and an acoustic transducer that pinged
at intervals. Because the effective compressibility of such a device was less than that
of sea water, it could be ballasted to sink to a predetermined depth where it became
neutrally buoyant. Thereafter it would drift with the current at that depth while its
position was determined by repeated triangulation relative to a research ship equipped
with a directional acoustic antenna at the ocean surface. In consultation with Stommel,
he and Worthington (1957) had tracked several such floats, in a region where they
expected the water below the main thermocline to be effectively stationary. It was not!
Instead they found sustained velocities of up to 10 cm/s in several different directions.
It was clear that some unanticipated dynamical process was operating, but a much
larger sample would be needed to determine what that was. Unfortunately, directional
tracking from a dedicated ship was prohibitively cumbersome and expensive, so it
was impractical to use this methodology on the scale required.
By the early 1970s several technological developments had progressed to the
point at which reconsideration seemed worthwhile.
Francis Bretherton
bottles spaced along a nonconducting cable lowered by winch from research ships on
the surface. Currents within and above the thermocline were inferred from assumed
geostrophic balance between the gradient of pressure in a horizontal direction and
the Coriolis force associated with motion relative to a rotating earth. The horizontal
pressure gradient was based upon differences of temperature and salinity between
sounding stations that were typically spaced hundreds of kilometers apart, converted
to density using laboratory determinations of the equation of state for seawater, and
hence to pressure (or equivalently dynamic height) using hydrostatic balance. This
inferential process gives only vertical differences in horizontal velocity, unless in
each water column there is a reference level at which the velocity itself can be independently determined. Unfortunately, the obvious candidate for reference level, the
ocean surface, was precluded by wind drift, tidal fluctuations, and uncertainties of
navigation, and no proven alternative methodology existed.
However, in the cold, dense, abyssal water below the main thermocline the
temperature and salinity seemed remarkably uniform, with horizontal differences
that in most places were measurable only on the scale of an ocean basin. It was also
known that the carbon-14 age of water deep in much of the Atlantic Ocean was at
least several hundred years, and in the Pacific it was over one thousand. These facts
suggested that the abyssal water was relatively quiescent, so horizontal velocities from
deep hydrographic soundings were typically reported relative to a “level of no motion”
that was supposed to be representative of the entire layer below the thermocline. To
account for the carbon-14 data, water that had sunk or mixed beneath the surface in
the North Atlantic had to be moving southward through the ocean basin. However, the
average velocity required was about 1 mm/s, which was well below the precision of
measurement of the remainder of the sounding. Thus, it was widely supposed that the
deep water in the open ocean was effectively quiescent and there was an identifiable
“level of no motion” below the main thermocline.
Swallow (1955) had devised an ingenious float consisting of a scaffolding tube
plugged at both ends, equipped with batteries and an acoustic transducer that pinged
at intervals. Because the effective compressibility of such a device was less than that
of sea water, it could be ballasted to sink to a predetermined depth where it became
neutrally buoyant. Thereafter it would drift with the current at that depth while its
position was determined by repeated triangulation relative to a research ship equipped
with a directional acoustic antenna at the ocean surface. In consultation with Stommel,
he and Worthington (1957) had tracked several such floats, in a region where they
expected the water below the main thermocline to be effectively stationary. It was not!
Instead they found sustained velocities of up to 10 cm/s in several different directions.
It was clear that some unanticipated dynamical process was operating, but a much
larger sample would be needed to determine what that was. Unfortunately, directional
tracking from a dedicated ship was prohibitively cumbersome and expensive, so it
was impractical to use this methodology on the scale required.
By the early 1970s several technological developments had progressed to the
point at which reconsideration seemed worthwhile.