Contributions to Global Ocean Observations
57
on moorings and, at the same time, was studying if basin-scale subsurface general
circulation, then the province of hydrographers and scattered float studies, could be
measured by large numbers of floats that infrequently rose to the surface to be located
by the same satellite tracking systems used by surface drifters. Bretherton (1980) had
already outlined the sampling problem in averaging energetic mesoscale variability
to extract mean flow and it looked feasible if the observation cost per year of float
observation could be significantly reduced. Because satellite tracking and data relay
would make a vertically cycling float autonomous from acoustic tracking networks
and significantly reduce the cost of low-density sampling arrays, I suggested to Webb
that such a float would be tremendously useful. We agreed that I would finish up the
ADCP work and obtain funding for the float project while he began work on it.
The main problems to overcome in a self-contained ADCP were finding a
suitable data recorder and fashioning acoustic beam patterns that were concentrated
enough that they could be made to scatter primarily off weak volume scatterers rather
than stronger discrete targets like mooring components or the surface. RDI considered an electronically steered directional array but settled on four mechanically
shaded transducers. Data storage was through a streaming magnetic tape recorder
with good energy efficiency. Considerable work and extensive tank testing went into
adjusting the transducer beam patterns to have low side-lobes. Even though developed
for SIO under an ONR grant, the prototypes were delivered elsewhere; Schott et al.
(1993) report their first use. Similar units, operating at 300 kHz and using simple
narrow-band Doppler processing, were delivered in time for use in the Ocean Storms
experiment (D’Asaro, et al., 1995). By chance, Ocean Storms included a wind event
that resonantly generated strong inertial currents in the mixed layer that then propagated downwards (and laterally), providing a great example of what the ADCP could
describe. Comparisons with VMCMs showed that, when ADCPs pinged infrequently
enough to last 9 months (90 pings per 15-minute average), mooring motion would
affect their accuracy as noise rather than the bias seen in rotor-vane meters. ADCP–
VMCM differences were approximately 10% of the 10–30 cm/s ambient currents and
70% of this difference was noise. The noise was substantially stronger from an ADCP
mounted in the surface-buoy bridle than from one in the mooring line at 115 m. With
the exception of the “fish” problem encountered on TAO moorings (Plimpton et al.,
1997), the self-contained ADCP has proven versatile and accurate and has nearly
replaced discrete current meters.
The autonomous float project started slowly. The main problem was to repetitively change float buoyancy so it could rise from its depth of neutral buoyancy to
the surface for satellite communication and then return to depth. Webb initially concentrated on novel techniques for varying buoyancy including individual capsules of
chemicals that would generate gas when exposed to seawater. We also experimented
with a flapper valve mechanism to rectify wave-induced vertical motion into buoyancy at the surface, but failed to make it function. After 3 years of frustration and with
concern for the continuation of ONR funding in the absence of progress, I convinced
Webb to shift to a more conventional electric motor driven buoyancy pump such as
57
on moorings and, at the same time, was studying if basin-scale subsurface general
circulation, then the province of hydrographers and scattered float studies, could be
measured by large numbers of floats that infrequently rose to the surface to be located
by the same satellite tracking systems used by surface drifters. Bretherton (1980) had
already outlined the sampling problem in averaging energetic mesoscale variability
to extract mean flow and it looked feasible if the observation cost per year of float
observation could be significantly reduced. Because satellite tracking and data relay
would make a vertically cycling float autonomous from acoustic tracking networks
and significantly reduce the cost of low-density sampling arrays, I suggested to Webb
that such a float would be tremendously useful. We agreed that I would finish up the
ADCP work and obtain funding for the float project while he began work on it.
The main problems to overcome in a self-contained ADCP were finding a
suitable data recorder and fashioning acoustic beam patterns that were concentrated
enough that they could be made to scatter primarily off weak volume scatterers rather
than stronger discrete targets like mooring components or the surface. RDI considered an electronically steered directional array but settled on four mechanically
shaded transducers. Data storage was through a streaming magnetic tape recorder
with good energy efficiency. Considerable work and extensive tank testing went into
adjusting the transducer beam patterns to have low side-lobes. Even though developed
for SIO under an ONR grant, the prototypes were delivered elsewhere; Schott et al.
(1993) report their first use. Similar units, operating at 300 kHz and using simple
narrow-band Doppler processing, were delivered in time for use in the Ocean Storms
experiment (D’Asaro, et al., 1995). By chance, Ocean Storms included a wind event
that resonantly generated strong inertial currents in the mixed layer that then propagated downwards (and laterally), providing a great example of what the ADCP could
describe. Comparisons with VMCMs showed that, when ADCPs pinged infrequently
enough to last 9 months (90 pings per 15-minute average), mooring motion would
affect their accuracy as noise rather than the bias seen in rotor-vane meters. ADCP–
VMCM differences were approximately 10% of the 10–30 cm/s ambient currents and
70% of this difference was noise. The noise was substantially stronger from an ADCP
mounted in the surface-buoy bridle than from one in the mooring line at 115 m. With
the exception of the “fish” problem encountered on TAO moorings (Plimpton et al.,
1997), the self-contained ADCP has proven versatile and accurate and has nearly
replaced discrete current meters.
The autonomous float project started slowly. The main problem was to repetitively change float buoyancy so it could rise from its depth of neutral buoyancy to
the surface for satellite communication and then return to depth. Webb initially concentrated on novel techniques for varying buoyancy including individual capsules of
chemicals that would generate gas when exposed to seawater. We also experimented
with a flapper valve mechanism to rectify wave-induced vertical motion into buoyancy at the surface, but failed to make it function. After 3 years of frustration and with
concern for the continuation of ONR funding in the absence of progress, I convinced
Webb to shift to a more conventional electric motor driven buoyancy pump such as