100–200 m from the instrument, and each relative
to the unknown velocity of the instrument. These
unknown velocities are removed by differentiating
the profiles in the vertical. The resulting overlapping shear profiles are then interpolated to a uniform depth grid and averaged to give a composite
shear profile. Integrating this shear profile in depth
gives a velocity profile relative to a single
unknown constant of integration. If the vertical
mean of the relative velocity profile is subtracted
out, then the constant that remains to be determined is just the depth-averaged velocity. This can
be calculated by a method closely analogous to
that used in shipboard ADCP work (Fischer and
Visbeck, 1993). The depth-averaged absolute
water velocity is the time-average of the velocity of
the water relative to the instrument, plus the timeaverage of the ship velocity as calculated from the
position difference between the start and end of
the cast, minus a small correction (usually less
than 1 cm s
91
), calculated from the time-integral of
the relative velocity profile. If the vertical velocity
were a constant during the downcast, and another
constant during the upcast, then the time-integral
would be equivalent to a depth-integral – which is
of course zero for the de-meaned relative velocity
profile. Hence the calculation of the depthaveraged velocity is very insensitive to the accuracy of the relative velocity profile.
At the start of the WOCE Hydrographic Programme (WHP) in 1990, shipboard ADCP installations were common but the technique was not yet
entrenched in the mainstream of physical oceanography. Shipboard ADCP data were not archived by
the US National Oceanographic Data Center
(NODC). The shipboard ADCP was generally
viewed as an ancillary tool for the WHP, of secondary priority; but it was cheap and readily available, so it survived – and eventually thrived. By the
end of the WHP, good shipboard ADCP installations were almost universal on blue-water research
ships, the data were archived at the US NODC and
the Japanese Oceanographic Data Center (JODC),
and the shipboard ADCP was widely viewed as
a standard and highly valuable tool. Most WHP
cruises resulted in shipboard ADCP data sets
(Fig. 3.1.4b, see Plate 3.1.4b, p. 172); many nonWHP cruises did also, yielding an extensive global
data set available for the WOCE analysis phase.
Lowered ADCP profiling was in its infancy at
the start of the WHP: the method was developed
in 1989 (Firing and Gordon, 1990; Fischer and
Visbeck, 1993). Early LADCP hardware was cumbersome, and its use on WHP cruises was initially
restricted to regions where it was expected to
be particularly valuable. A prime example is the
near-equatorial band, where the complex current
structure is poorly estimated by geostrophic calculations from typical synoptic sections (Moum
et al., 1987). With hardware improvements making the LADCP easier to use during the Pacific
phase of the WHP, current profiling on complete
CTD sections became routine; coverage was particularly good in the Indian Ocean and reasonable
in the Atlantic (Fig. 3.1.4c, see Plate 3.1.4c, p. 172).
3.1.3.2 Evolution and status of shipboard
ADCP systems and the impact of GPS
Shipboard and lowered ADCP systems critically
depend on more than the Doppler sonar itself;
equally important are the attitude and position sensors. During the WHP, shipboard system improvement was almost entirely in the navigational
component: Global Positioning System (GPS)
measurement of position and heading.
Although it did not officially reach Initial Operational Capability until December 1993 (Standard
Positioning Service) and Full Operational Capability until April 1995 (fully tested military functionality), research ships started to obtain fixes from
GPS receivers much earlier. The first satellite was
launched as early as 1978, but numbers grew
slowly at first, with 10 satellites launched up until
1985, followed by a 4-year gap. In the mid- to
late-1980s, many research ships were fitted with a
GPS receiver to enable them to obtain fixes during
the few hours per day when sufficient satellites
were visible above the horizon. With the aid of
a rubidium or caesium standard clock, a twodimensional position fix could be obtained with just
two satellites. Fixes obtained in this way gave useful relative positions for the calculation of ship
speed, but were susceptible to large apparent
jumps in position (sometimes several hundred
metres) when there was a change in the combination of satellites used to calculate the fix.
The main operational series of so-called Block-II
satellites were launched at the rate of about five
per year from early 1989 onwards, and a complete
constellation of 24 satellites was reached for the
first time in mid-1993. In favourable locations,
coverage for position fixes had already grown to
SECTION 3 NEW WAYS OF OBSERVING THE OCEAN
112
to the unknown velocity of the instrument. These
unknown velocities are removed by differentiating
the profiles in the vertical. The resulting overlapping shear profiles are then interpolated to a uniform depth grid and averaged to give a composite
shear profile. Integrating this shear profile in depth
gives a velocity profile relative to a single
unknown constant of integration. If the vertical
mean of the relative velocity profile is subtracted
out, then the constant that remains to be determined is just the depth-averaged velocity. This can
be calculated by a method closely analogous to
that used in shipboard ADCP work (Fischer and
Visbeck, 1993). The depth-averaged absolute
water velocity is the time-average of the velocity of
the water relative to the instrument, plus the timeaverage of the ship velocity as calculated from the
position difference between the start and end of
the cast, minus a small correction (usually less
than 1 cm s
91
), calculated from the time-integral of
the relative velocity profile. If the vertical velocity
were a constant during the downcast, and another
constant during the upcast, then the time-integral
would be equivalent to a depth-integral – which is
of course zero for the de-meaned relative velocity
profile. Hence the calculation of the depthaveraged velocity is very insensitive to the accuracy of the relative velocity profile.
At the start of the WOCE Hydrographic Programme (WHP) in 1990, shipboard ADCP installations were common but the technique was not yet
entrenched in the mainstream of physical oceanography. Shipboard ADCP data were not archived by
the US National Oceanographic Data Center
(NODC). The shipboard ADCP was generally
viewed as an ancillary tool for the WHP, of secondary priority; but it was cheap and readily available, so it survived – and eventually thrived. By the
end of the WHP, good shipboard ADCP installations were almost universal on blue-water research
ships, the data were archived at the US NODC and
the Japanese Oceanographic Data Center (JODC),
and the shipboard ADCP was widely viewed as
a standard and highly valuable tool. Most WHP
cruises resulted in shipboard ADCP data sets
(Fig. 3.1.4b, see Plate 3.1.4b, p. 172); many nonWHP cruises did also, yielding an extensive global
data set available for the WOCE analysis phase.
Lowered ADCP profiling was in its infancy at
the start of the WHP: the method was developed
in 1989 (Firing and Gordon, 1990; Fischer and
Visbeck, 1993). Early LADCP hardware was cumbersome, and its use on WHP cruises was initially
restricted to regions where it was expected to
be particularly valuable. A prime example is the
near-equatorial band, where the complex current
structure is poorly estimated by geostrophic calculations from typical synoptic sections (Moum
et al., 1987). With hardware improvements making the LADCP easier to use during the Pacific
phase of the WHP, current profiling on complete
CTD sections became routine; coverage was particularly good in the Indian Ocean and reasonable
in the Atlantic (Fig. 3.1.4c, see Plate 3.1.4c, p. 172).
3.1.3.2 Evolution and status of shipboard
ADCP systems and the impact of GPS
Shipboard and lowered ADCP systems critically
depend on more than the Doppler sonar itself;
equally important are the attitude and position sensors. During the WHP, shipboard system improvement was almost entirely in the navigational
component: Global Positioning System (GPS)
measurement of position and heading.
Although it did not officially reach Initial Operational Capability until December 1993 (Standard
Positioning Service) and Full Operational Capability until April 1995 (fully tested military functionality), research ships started to obtain fixes from
GPS receivers much earlier. The first satellite was
launched as early as 1978, but numbers grew
slowly at first, with 10 satellites launched up until
1985, followed by a 4-year gap. In the mid- to
late-1980s, many research ships were fitted with a
GPS receiver to enable them to obtain fixes during
the few hours per day when sufficient satellites
were visible above the horizon. With the aid of
a rubidium or caesium standard clock, a twodimensional position fix could be obtained with just
two satellites. Fixes obtained in this way gave useful relative positions for the calculation of ship
speed, but were susceptible to large apparent
jumps in position (sometimes several hundred
metres) when there was a change in the combination of satellites used to calculate the fix.
The main operational series of so-called Block-II
satellites were launched at the rate of about five
per year from early 1989 onwards, and a complete
constellation of 24 satellites was reached for the
first time in mid-1993. In favourable locations,
coverage for position fixes had already grown to
SECTION 3 NEW WAYS OF OBSERVING THE OCEAN
112