Ocean Acoustic Tomography
123
Perhaps the most characteristic result of these early CW transmissions was
the variability in receptions: fade-outs are the rule rather than the exception. Early
attempts at understanding this inherent variability were in terms of a homogeneous
and isotropic ocean fine structure. That assumption, though mathematically convenient, has nothing to do with reality: ocean fine structure is neither homogeneous nor
isotropic. Garrett and I had developed a statistical model of internal waves, which was
found consistent with the acoustic phase and intensity variances in the Florida and
Eleuthera transmissions.
8 This work has been greatly extended by Dashen, Flatt´ e, and
others.
9 We now know that internal waves are the primary (but not the only) source
of “noise” in the measurements of mesoscale induced travel time variations.
THE BEGINNINGS (1976–1980)
At a meeting in 1976 celebrating the thirtieth anniversary (1946–1976) of the Office of Naval Research, I described the transmissions between the R/V Alexander
Agassiz and the R/VEllen B. Scripps at 25-km range. These results were the basis for
Worcester’s thesis
10 (the first of many Tomography dissertations). There is a noteworthy asymmetry between the clean arrivals from beneath the sound axis and the
complex multipaths through the upper ocean (Figure 8.3). The sum and difference
τ sum = τ A→B + τ B→A ,
τ dif = τ A→B − τ B→A
of the oppositely traveling transmission times yielded useful estimates of the deep
and shallow temperatures and currents (sound travels faster in warm water and in the
direction with the current).
By 1977 the first tomographic group was being formed. Spindel and Porter
from Woods Hole brought their experience in underwater acoustics, autonomous
instrumentation, setting deep-sea moorings, and precise acoustic position keeping.
Webb had built acoustic sources for SOFAR floats. Birdsall and Metzger from the
University of Michigan had developed the signal processing for the MIMI transmissions. Wunsch from MIT was pioneering the application of inverse theory to ocean
observations. Worcester and Munk from Scripps participated. The formation of the
group was something of a shotgun wedding under the persuasion of Hugo Bezdeck
of ONR who wanted to avoid duplicate efforts; yet the group was to work together
for a quarter-century.
In the fall of 1978 a 900-km transmission from a source southwest of Bermuda
proved adequate to track 14 ray paths for the duration of the experiment.
11 This work
came just in time to save tomography from an early demise. A reviewer of our proposal
had written that individual ray arrivals could not be resolved, and even if resolved
could not be identified, and even if identified would not be stable. The proposal was
declined. We responded by sending Figure 8.4 with the sentence: “we have resolved,
identified and tracked 13 stable arrivals for over 2 months (see figure).” The proposal
was accepted.
123
Perhaps the most characteristic result of these early CW transmissions was
the variability in receptions: fade-outs are the rule rather than the exception. Early
attempts at understanding this inherent variability were in terms of a homogeneous
and isotropic ocean fine structure. That assumption, though mathematically convenient, has nothing to do with reality: ocean fine structure is neither homogeneous nor
isotropic. Garrett and I had developed a statistical model of internal waves, which was
found consistent with the acoustic phase and intensity variances in the Florida and
Eleuthera transmissions.
8 This work has been greatly extended by Dashen, Flatt´ e, and
others.
9 We now know that internal waves are the primary (but not the only) source
of “noise” in the measurements of mesoscale induced travel time variations.
THE BEGINNINGS (1976–1980)
At a meeting in 1976 celebrating the thirtieth anniversary (1946–1976) of the Office of Naval Research, I described the transmissions between the R/V Alexander
Agassiz and the R/VEllen B. Scripps at 25-km range. These results were the basis for
Worcester’s thesis
10 (the first of many Tomography dissertations). There is a noteworthy asymmetry between the clean arrivals from beneath the sound axis and the
complex multipaths through the upper ocean (Figure 8.3). The sum and difference
τ sum = τ A→B + τ B→A ,
τ dif = τ A→B − τ B→A
of the oppositely traveling transmission times yielded useful estimates of the deep
and shallow temperatures and currents (sound travels faster in warm water and in the
direction with the current).
By 1977 the first tomographic group was being formed. Spindel and Porter
from Woods Hole brought their experience in underwater acoustics, autonomous
instrumentation, setting deep-sea moorings, and precise acoustic position keeping.
Webb had built acoustic sources for SOFAR floats. Birdsall and Metzger from the
University of Michigan had developed the signal processing for the MIMI transmissions. Wunsch from MIT was pioneering the application of inverse theory to ocean
observations. Worcester and Munk from Scripps participated. The formation of the
group was something of a shotgun wedding under the persuasion of Hugo Bezdeck
of ONR who wanted to avoid duplicate efforts; yet the group was to work together
for a quarter-century.
In the fall of 1978 a 900-km transmission from a source southwest of Bermuda
proved adequate to track 14 ray paths for the duration of the experiment.
11 This work
came just in time to save tomography from an early demise. A reviewer of our proposal
had written that individual ray arrivals could not be resolved, and even if resolved
could not be identified, and even if identified would not be stable. The proposal was
declined. We responded by sending Figure 8.4 with the sentence: “we have resolved,
identified and tracked 13 stable arrivals for over 2 months (see figure).” The proposal
was accepted.