a SOFAR float that included ALACE technology
to cycle between a pair of isotherms in the main
thermocline. While ALACEs are pre-ballasted to
drift approximately at a selected depth, the more
modern autonomous floats used toward the end of
WOCE, such as the SOLO (Davis et al., 2000) and
APEX, have full buoyancy control and can be programmed to follow the flow at a different depth
than the bottom of their profiles, or even to track
isotherms. These floats correspond exactly to
the category of floats that ‘keep in trim in the
manner of a submarine’ that Stommel had in mind
in 1955.
The aspect of float development that made possible the bold WOCE attempt to directly measure
velocity fields on basin scales was the cost reduction made possible by the development of RAFOS
and vertical cycling technologies that make it
possible to measure velocity at a cost near two
thousand dollars per year. A RAFOS float costs
about $4500, autonomous floats twice this, and
MARVOR floats about three times as much.
RAFOS floats typically operate for 1–2 years,
autonomous floats can report every 10 days
for over 5 years, and Ollitrault (1999) has shown
that 50% of his MARVOR floats were still
alive after 5 years. Sound sources (approximately
$22 000 each) and costs of analysis of acoustic
time delays add to the cost of RAFOS and
MARVOR operations.
The scientific utility of ALACE floats has been
significantly increased by adding temperature and
conductivity sensors so that temperature and salinity profiles from transits between the surface and
the target depth can be reported (cf. Davis et al.,
2000). Because a Profiling ALACE (P-ALACE)
executes 200 depth cycles, the per-profile cost
is competitive with expendable probes so long as
the conductivity sensor remains stable enough to
yield good salinities. Even if the float’s entire construction and communication cost is charged to
profiling (neglecting the value of velocity observations), the full operational cost of a temperature
profile is about $50, with $25 to add a salinity
profile.
To complete this historical perspective, let us
note that the community is now implementing
Henry Stommel’s last vision for observing the
ocean. He was aware that early autonomous floats
could, in addition to tracking currents, be adept at
measuring profiles, but that there were many reasons for wanting to control where these profiles
were taken. A simple solution was to fit wings to
a buoyancy-changing float to produce a simple
autonomous underwater glider. In a visionary
article, Stommel (1989) combined this with the
revolutionary idea of generating the necessary
buoyancy forcing from the ocean’s thermal stratification to propose a new observational tool, which
he called ‘Slocum’ in honour of the first global
circumnavigator, Captain Joshua Slocum (1900)
and enjoyed demonstrating a prototype thermal
buoyancy engine in the workshop behind his Falmouth home. Technologists from University of
Washington, Scripps Institution of Oceanography
and Woods Hole Oceanographic Institution are
adapting autonomous float technology to develop
simple electric-powered underwater gliders, while
Webb Research Corporation is implementing
Stommel’s full vision of thermal-powered gliders.
In each implementation, wings are used to efficiently convert buoyancy into forward motion of
the O(30 cm s
91 ) as the vehicle cycles between
shallow and deep levels. This forward motion
can be used to hold station to gather a time series
of profiles (a virtual mooring) or to sample
autonomously a hydrographic section of several
thousand kilometres in length.
3.2.3 The WOCE Float Programme
There was, at the start of WOCE, great optimism
that inverse analyses based on conserving transport
of tracer water properties measured in the WOCE
Hydrographic Programme could accurately estimate the absolute mean velocity of the general circulation. But for two reasons it was felt that direct
observations of absolute subsurface flow should be
included in the measurement programme. First, the
precision that could be achieved by inverse analysis
procedures was uncertain, but it was clear that
their accuracy would be improved by including
direct velocity observations. Second, these inverse
methods, like their more traditional predecessors,
use the distribution of tracers to infer absolute flow
and therefore depend on knowing how well tracers
are conserved along streamlines. Since absolute
velocities and mixing are both linked to tracer
distributions (cf. Toole and McDougall, Chapter
5.2), if absolute mean velocities were measured,
3.2 Subsurface Lagrangian Observations during the 1990s
127
Davis and Zenk
to cycle between a pair of isotherms in the main
thermocline. While ALACEs are pre-ballasted to
drift approximately at a selected depth, the more
modern autonomous floats used toward the end of
WOCE, such as the SOLO (Davis et al., 2000) and
APEX, have full buoyancy control and can be programmed to follow the flow at a different depth
than the bottom of their profiles, or even to track
isotherms. These floats correspond exactly to
the category of floats that ‘keep in trim in the
manner of a submarine’ that Stommel had in mind
in 1955.
The aspect of float development that made possible the bold WOCE attempt to directly measure
velocity fields on basin scales was the cost reduction made possible by the development of RAFOS
and vertical cycling technologies that make it
possible to measure velocity at a cost near two
thousand dollars per year. A RAFOS float costs
about $4500, autonomous floats twice this, and
MARVOR floats about three times as much.
RAFOS floats typically operate for 1–2 years,
autonomous floats can report every 10 days
for over 5 years, and Ollitrault (1999) has shown
that 50% of his MARVOR floats were still
alive after 5 years. Sound sources (approximately
$22 000 each) and costs of analysis of acoustic
time delays add to the cost of RAFOS and
MARVOR operations.
The scientific utility of ALACE floats has been
significantly increased by adding temperature and
conductivity sensors so that temperature and salinity profiles from transits between the surface and
the target depth can be reported (cf. Davis et al.,
2000). Because a Profiling ALACE (P-ALACE)
executes 200 depth cycles, the per-profile cost
is competitive with expendable probes so long as
the conductivity sensor remains stable enough to
yield good salinities. Even if the float’s entire construction and communication cost is charged to
profiling (neglecting the value of velocity observations), the full operational cost of a temperature
profile is about $50, with $25 to add a salinity
profile.
To complete this historical perspective, let us
note that the community is now implementing
Henry Stommel’s last vision for observing the
ocean. He was aware that early autonomous floats
could, in addition to tracking currents, be adept at
measuring profiles, but that there were many reasons for wanting to control where these profiles
were taken. A simple solution was to fit wings to
a buoyancy-changing float to produce a simple
autonomous underwater glider. In a visionary
article, Stommel (1989) combined this with the
revolutionary idea of generating the necessary
buoyancy forcing from the ocean’s thermal stratification to propose a new observational tool, which
he called ‘Slocum’ in honour of the first global
circumnavigator, Captain Joshua Slocum (1900)
and enjoyed demonstrating a prototype thermal
buoyancy engine in the workshop behind his Falmouth home. Technologists from University of
Washington, Scripps Institution of Oceanography
and Woods Hole Oceanographic Institution are
adapting autonomous float technology to develop
simple electric-powered underwater gliders, while
Webb Research Corporation is implementing
Stommel’s full vision of thermal-powered gliders.
In each implementation, wings are used to efficiently convert buoyancy into forward motion of
the O(30 cm s
91 ) as the vehicle cycles between
shallow and deep levels. This forward motion
can be used to hold station to gather a time series
of profiles (a virtual mooring) or to sample
autonomously a hydrographic section of several
thousand kilometres in length.
3.2.3 The WOCE Float Programme
There was, at the start of WOCE, great optimism
that inverse analyses based on conserving transport
of tracer water properties measured in the WOCE
Hydrographic Programme could accurately estimate the absolute mean velocity of the general circulation. But for two reasons it was felt that direct
observations of absolute subsurface flow should be
included in the measurement programme. First, the
precision that could be achieved by inverse analysis
procedures was uncertain, but it was clear that
their accuracy would be improved by including
direct velocity observations. Second, these inverse
methods, like their more traditional predecessors,
use the distribution of tracers to infer absolute flow
and therefore depend on knowing how well tracers
are conserved along streamlines. Since absolute
velocities and mixing are both linked to tracer
distributions (cf. Toole and McDougall, Chapter
5.2), if absolute mean velocities were measured,
3.2 Subsurface Lagrangian Observations during the 1990s
127
Davis and Zenk