Sound Velocity Basics PDF
Uploaded by
astefanovichSound Velocity Basics PDF
Uploaded by
astefanovich. Transmission loss over ROCK is greater than that over SAND, but less than over SAND-
‘Mub.
| There is some evidence that transmission loss increases with Increasing wind force and
increasing hydrophone depth. In this connection, however, the effects were irregular, in-
‘consistent, and considerably smaller in magnitude than those of Conclusion (1). They may
not be real.
‘Transmission loss is roughly linearly dependent on R~! where R is the length of the arches,
formed by repeated reflections of the limiting ray calculated by assuming a linear tempera-
ture gradient in the upper 150 feet.
5. Effective bottom reflection losses calculated from this observed dependence were: 4 db
for SAND, 11 db for ROCK, and 16 db for SAND-MUD.
849 APPENDIX II
SHALLOW WATER TRANSMISSION AT 12 AND 24 KHZ.
Ref. 3
|. Transmission ranges at 12 kHz are on the average about 50% longer than those observed
at 24 kHz.
Over all types of bottoms, transmission is dependent on the thermal gradient in the water,
‘but more particularly on the extent of the sound field described by the ray diagram,
Transmission varies with bottom type and is best over SAND and worst over MUD.
In depths from 60-600 ft, transmission improves with depth of water when there Is down:
ward refraction
3. In depths from 600-1200 ft., transmission is poorer than in depths from 60-600 ft. when
there is downward refraction.
3. When there is an isothermal surface layer, transmission in the layer improves with increasing
thickness of the layer as long as the wind force is low. With high wind force values there is
‘no significant changa in transmission with inereasing thickness of the isothermal layer.
7. Near the surface the Rao range decreases with increasing wind force. (The Reo range was
defined as the range at which the transmission loss, relative to 100 yas., was 40 db. Relative
to 1 yd, the loss to Re» would be 80 db.)
820 APPENDIX Il
SHALLOW WATER TRANSMISSION CONTOURS.
Ref. 4
The transmission plots such as those of Fig. 3, in directions parallel to and perpendicular to the bottom slope
have the following noteworthy features:
(a) Effect of Range. The transmission tends to be better than spherical (that i, the transmission anomaly is nega
tive) out to @ range generally from 5 to 10 miles, followed by a rapid fall-off with range (when a logarithmie range
scale is used).
This characteristic is typical of all range runs in shallow water, and is due to trapping at short ranges and the
dominant effect of boundary attenuation long ranges.
(b) Effect of Frequency. The transmission tends to be best in the 100-200 Hz or 200-400 He octaves, and poorer
in octaves below and above. At lower frequencies, the transmission is worse because of a lower ratio of water depth
to wavelength and a lesser number of trapped modes and a greater attenuation; at higher frequencies, the trans
mission is worse because of a higher volume attenuation. The existence or an optimum frequency for these reasons
is characteristic of all sound channels, of whatever type, in the sea
(c) Effect of Hydrophone Depth. The transmission to the shallow (80 feet) hydrophone is better than to the
‘dveper (180 feet) hydrophone at the higher frequencies, presumably because of trapping in the near-surface sound
channel, At the lower frequencies, there is no great difference between the two hydrophone depths.
(d) Effect of Direction. In any one octave band the transmission out to a distance of about 10 miles is unaffected
by the depth of the hydrophone or by the direction of the run. The transmission may therefore be said to be essen-
tially isotropic about the receiving vessel out to 10 miles, and independent of hydrophone depth. But at longer
ranges, there are great differences in transmission in the different directions. For example, and most notably, the
transmission to the North is poorer than to the South at low frequencies where the bottom acoustic characteristics
‘must play a dominant role, but not appreciably different at high frequencies where trapping in the near surface layer
determines the transmission, However, in the direction toward shore, the high frequencies are transmitted to the
shallow hydrophone better than in other directions, possibly because of the enhancement of the mixed-layer trap
by the shallow bottom at or just below its base in this direction. Contrarywise, in the direction toward deep water
the transmission is the poorest of all at all frequencies, so much so that the run was aborted at 30 miles due to weak
fr absent shot signals in this direction,
821 CHAPTER 9
REFLECTION AND SCATTERING
BY THE SEA SURFACE CHAPTER 9
REFLECTION AND SCATTERING
BY THE SEA SURFACE
Introduction
‘The surface of the sea affects underwater sound in a number of ways, The sea surface
1). produces a reflection lass for sound incident upon it. This lose, though small
total loss for RSR paths in the Deep Sound Channel at long ranges.
2) produces back-scattering, causing sea surface reverberation
3) produces forward scattering, the dominant loss process in propagation in the mixed layer.
4) produces interference effects with the direct path in the sound field from a shallow source.
8) produces a doppler shift in narrowband reflected and scattered sound, due to its wind-driven motion.
6) has a bubbly layer of air just beneath it when it is rough, that often acoustically hides the surface itself
7) casts a shadow in @ negative gradient extending all the way up to the surface.
8) is the source of ambient noise in the deep sea at kilohertz frequencies (0.5 to 50 kHz),
9) reduces the power radiated by a sound source located a quarter-wavelength or less below it.
OF these various effects, Numbers 1 to 7 are included in what follows, leaving others to remain as topics other than
Propagation.
is a major contributor to the
Roflection and Scattering
When a plane sound wave in water strikes an infinite plane boundary (a perfectly smooth sea surface), nearly al
Of its energy Is reflected at the boundary. At normal incidence, the ratio of the intensity of the sound crossing the
boundary to the incident intensity is only 0.0002. Nearly all of the sound remains in the water, and goes into the
specular direction as a plane wave (Fig Ta) that is perfectly coherent, ie. with unity correlation coefficient, with
the incidont wave.
When the surface begins to get rough, some sound goes into other directions, and each small surface area acquires
8 directional pattern (Fig 1b). The correlation coefficient between the returned sound from the surface and the in
cident sound is less than unity, and the intensity in the vicinity of the specular direction is reduced. Finally, when.
the surface is very rough, the returned sound is entirely scattered sound having no trace of # maximum in the specu
lar direction and with no coherence with the incident sound.
‘Ameasure of the acoustic roughness of a surface is the Rayleigh parameter R defi
1d by the relation
R= 2kh sing
wave number = 2n/A, h is the rms roughness height (2h is the rms “waveheight”, crest
‘rough) and ¢ is the grazing angle. R may be considered to be the rms phase difference between a reflection from
2 crest or trough at @ distance h from the mean surface and that from the mean surface, as illustrated heuristically
in Fig 2, It is found empirieslly that when R <1 the surface may be considered to be smooth; when R > 1, the
surface is rough.
‘The reflected field occurs in certain discrete directions at angles Bry to the vertical, given by the grating formula
a,
Sin Opp = sin 9 + mA m=, 31,22
mk
where K and k are the two wave numbers 2n/A and 2n/, respectively. The directions yy are shown in Fig 3b. The
‘otal number mmax of directions (“modes" or "orders"? is limited by the condition that
sin Bg <1
‘The Rayleigh method has ted to a certain amount of controversy. Rayleigh assumed that the sound field every:
where is given by the sum of the direct wave and the various contributing modal waves in the directions 8. The
validity of this assumption at points close to the corrugated surface has been theoretically questioned by several
theoreticians and has led to different alternative answers to the problem of reflection from a sinusoidal surface
(4) (8) (6) (7). The different theoretical treatments result in differences in the amplitudes of the reflected modes.
For example, in Eckart’s theory (7), the amplitude coefficients are given by
on ‘Smooth Surface a.)
Rough Surface b.)
Figure 1. (a) Reflection from smooth sea surface; nearly all the incident intensity is reflected at the
specular angle at which y, = ¢; (b) Directional patterns of the return from a rough sea surface for three
sea states or degrees of roughness.
92 xms phase difference = 2%. 2h sin 0
= 2kh sin 6
=R
Fig. 2. Conceptual view of the Rayleigh coefficient R. The reflection from the root mean-square wave crest differs
‘+ iphase by R radians from the reflection from the mean surface. The difference in path length is 2h sin y.
93 Fig. 3, Reflection from a one dimensional sinusoidal surface. In (@) a plane weve of wavelength 2 is incident
‘upon the surface of wavelength A. In (b) the reflection occurs in certain discrete directions m,
04 ete,
= 248m + Joy (e+ omnlh
An = 238 (e+ cm]
where Am is the amplitude coefficient of the mth mode, ¢ = COs ,¢m = c08 Om, In the specular direction, this
reduces 10
Ag = Jol2hk cos 0)
‘The Rayleigh method appears to be valid (1) for kh << 1. However, for many conditions of k,h, 8, the different
‘theoretical results are not significantly different from one another.
Experimental Verification
No sea-going work has been done to rlate these theories to a real quasi-sinusoidal sea surfece, such as when a
swell is running and a near-sinusoidal shape exists. However, two laboratory model experiments have been made.
LaCasce and Tamarkin (8) used a sheet of corrugated cork floating on water to give the sinusoidal pressure release
surface. Several surfaces were used, over a frequency range 80 to 300 kHz at 10 kHz intervals and at various angles
Of incidence. The data was compared with some 13 theoretical models (9). In this comparison, all of the models,
including the Rayleigh model, were found to fit the data fairly well, with nothing to suggest that any one is better
than the others. More recent measurements have been made by Barnard et al. (10) ata frequency of 100 kHz using
pieces of styrofoam cut on a milling machine so as to have \ = 4,5 em and h = 1.5 cm. Satisfactory agreement was
found with one (Uretsky’s (6)) theoretical model.
‘Surface Reflection Interference (Lloyd Mirror Effect)
When the sea surface is smooth enough to be a reflector (R << 1), an interference pattern is produced between
direct-path sound and sound reflected from the sea surface (Fig 4). This sound field may be divided into three parts,
1) a nearfield close to the source in which the image source is too far away, and the reflected sound is too weak,
‘to produce appreciable interference, 2) an interference field in which there are strong loops and nulls in the signal
reesived by @ receiver moving outward in range, and 3) a farield, in which there is an increasingly out-of-phase
condition between source and image and the intensity falls off as the inverse fourth-power of the range (TL in-
creases as 40 log r). Fig. 4c) shows the transmission anomaly (transmission loss minus spherical spreading) drawn
‘against a normalized range in these three parts of the sound field from a source and its surface image.
‘The image interference or Lloyd Mirror effect was first worked out during World War I! (11). Strange to say,
ithas not been systematically studied through at-sea measurements,
The effect of surface scattering and of bandwidth is to fill up the crests and troughs in the interference region.
Simple theory assumes straight‘line propagation paths; refraction shifts the interference pattern in range — a subject
investigated in one of the earliest papers (12) on underwater sound published in JASA.
‘The Random Surface
By “random” surface is meant an irregular stochastic surface describable only in statistical terms. Referring to
Fig a, the two parameters required for this definition are (1) the roughness height h, or root mean square devi-
tion of the surface from a plane, defined ¥F(x), where y(x) is the height of the surface at a point x rela
tive to the mean surface and the bar represents an average of y(x) over many points and (2) a correlation distance
determined by
Taye
als) he
For theoretical tractibility, p(t) is commonly taken to be of the form p($) = exp (-{2/T"), where T is a constant,
called the correlation length. Other functions like exp(-t/T) appear in the literature, but the normal form just given
‘occurs most often. Both T and h are, in general, functions of direction in the plane of the surface.
For random surfaces, the scattering is quantitatively described by a scattering coefficient s; the quantity
S-10 log s is called the scattering strength of the surface. If a plane wave of intensity Ij is incident on a small area
A of a rough surface at angle of incidence 6, (Fig 5b), the scattered intensity at distance r at angle 8 is
o5 oO
oy ™
4; ><>
Fre 2
wt a ia »
Source i ~
a
acter
= (24449) ao f8te
| q
THe
ean Fieto | THE INTERFERENCE FIELO | THe Fan FieLo
w
| 16 \ !
|
s+-— +
TRANSMISSION
ANOMALY, €8
we eA wR
aR,
Som addy
Figure 4. The Liovd Miror sound fini. In (a, receiver at depth da receives sound from the real wource
tt depth and slant distance & pls sound from an out ot phese image. (b shows the reautingdretionl
patter in the interference region (e) shows the transmission anomaly (TL:20 log +) plotted against
hormalized range inthe thee regions deserbod in the ext.
96 vo) vetr)
Figure 5. (a) Cross-section in the x direction of @ random surface, (b) a plane wave is incident on a
small area of the surface. The scattering coefficient s and the scattering strength 10 log s relate |, to |,
o7 In torms of the angles 0, and 02, Beckmann and Spizzichino (1) derived a theoretical expression for x that was
made more useful to sonar by Clay and Medwin (13) as
F 00878 scot? (sind -sinda)?
Pa ne ea ot “|
(cos + cost)? (cos, + cos0,)® a
where
14 cos) co992 - sind sind2
0x0 (e030 + cos)
‘and where B is @ slope angle of the surface such that the rms slope is Yétan?#. Cox and Munk (14) made studies of
the glitter pattern of the sun on the sea surface and found that the slope angle Bis approximately related to the wind
speed by
‘stan’ = 0,003 + §.12W x 10°*
\Where W is the wind speed in meters per second.
For back scattering,
where y is the angle of the incident wave with the horizontal. We may note that there is no dependence on fre:
‘quency; the formulas apply only for a very rough surface such that
2h
Lleos8 + coxd2) > 1
and so are restricted to high frequencies, high grazing angles and very rough surfaces.
'At the other extreme, for low frequencies, low angles, and slightly rough surfaces, a result given by Eckart (7)
predicts a variation of s with the fourth power of frequency.
“The scattering of sound by rough surfaces is a vast and challenging subject to theoreticians. As long ago as 1954,
‘paper by Miles (15) on rough surface scattering was introduced by references to as many as 16 earlier papers;
‘many more could have been added and many more have appeared in subsequent years. Existing summaries of the
subject are references (9) and (16).
Fig 6 gives the back scattering strength s for three wind speeds at angles of O° and 60°. Notice from the curves
cof Fig 7 that, as the wind speed increases, the amount of sound going into angles near the specular direction de
creases, while that going into other angles increases, as one would expect. This behavior is consistent with the
view that, at angles near the specular direction, the sea surface behaves as if it consisted of numerous small mirrors
‘or wave facets which become smaller and more inclined to the horizontal with increasing wind speed.
Frequency Effects
In addition to distributing an incident plan wave in space, the vertical motion of the rough sea surface modu-
lates the amplitude of the incident wave, and superposes its own spectrum as upper and lower sidebands on the spec-
‘trum of the ineidant sound. This is shown in Fig 8. Any horizontal motion of the surface, such as would be caused
by wind or currents, also will appear in the sound scattered from the surface and cause a Dopplershifted and Dop-
pler-smeared spectrum. Such spectra were observed in surface back-scattering at 6OkHz by Igarashi and Stern (17)
and are illustrated in Fig 9 for the downwave and upwave directions.
‘Another observation of rea-sea back scattering at 85 kHz showed a frequency shift that indicated a surface veloc:
ity of 0.6 knot in @ 20 knot wind (18). A frequency spread was also observed, due to the fact that the surface scat
terers do not all have the same velocity. A frequency spread has also been measured for shallow water transmission
(19) due probably to the same cause, as well as turbulent water motion of the whole water column; over 5-16 miles,
the spreading amounted to + O.5Hz for signals in the range 350-2400 Hz. Frequency effects caused by surface
oe : —
* | E-
15kn
54° 58° 62 66 70 74 78 82 86 90°
Grazing Angle, ¢
Fig. 7. Sea surface back-scattering strength at three wind speeds computed from the Clay-Medwin formula (13)
910 LOW SEA STATE
f
SCATTERED
SPECTRUM
HIGH SEA
INCIDENT
SPECTRUM
; b)
WAVE SPECTRUM
4)
O .1 12 44 65
Hz.
Fig. 8, In (a), a wave incident on a moving surface becomes amplitude-modulated by the surface mation. The
narrow-band incident spectrum shown in (b) acquires sidebands as shown in (c), which have the same
shape as the motional wave spectrum (d). With inereasing sea state, the sidebands increase at the expense
of the spectral peak at the center frequency fo,
on DOWNWAVE
TRANSMITTED
110 MS, PULSE
UPWAVE
100 50 0 50 100
FREQUENCY SHIFT, Hz
Figure 9, Backscattering spectra from the sea surface at 60 kHz at a grazing angle of 30° in the
dowmwave and upwave directions of the incident 110 ms pulse. Ref. 17. motion were predicted from theory and observed in a laboratory tank by Roderick and Cron (20). They have been
Tepeatedly observed in atsea experiments, such as by Brown and Fisk (21), Williams in shallow water (22), Urick
(23) and Shooter and Witliams (24). Three examples of motional spectra are shown in Fig 10, In short, the motion
of the sea surface produces amplitude fluctuations of a steady signal, with upper and lower sidebends having the
same spectrum as the motion of the surface.
Surface Loss (Forward Scattering)
The loss suffered by sound after encountering the sea surface and being reflected and/or scattered in the specular
direction has been investigated at sea over the years by many different investigators using different methods. Un-
fortunately, the results are made ambiguous in many cases bv a mixture of reflected (coherent) and scattered (i
coherent) sound in the surface return, Three methods have been used:
|. Comparing the amplitude or energy of pulses returned from the surface with that of the direct arrv
II, Using the Lloye-Mirror effect and observing the depth of the minima as the frequeney is varied,
IIL, Measuring the attenuation in the surface duct,
Table | is a tabulation of surface loss measurements taken from the literature. From this compilation it appears
that the surface loss is less than 1 db at frequencies less than 1 kHz and rises to about 3 db at 25 and 30 kHz. For
long-range propagation over RSR paths at frequencies of @ few hundred Hz and less, the surface loss in computer
‘models is commonly taken to be zero.
‘Scattering Strength (Back Scattering)
‘The other direction for which measurements have been made is back toward the source. The methods for deter
‘mining the scattering strength of the sea surface are:
|. Using a directional source/receiver and short pulses, beamed up to the surface at different angles.
I. Using explosive sound sources and a non-diractional receiver.
‘An example of Method | is the use by Urick and Hoover (25) of 60 kHz pulses sent and received by a tiltable
narrow-beam transducer. Method I! is the only method feasible for low-frequencies, and depends on the geometry
‘to sort out the vatious grazing angles. A good example is the work of Chapman and Harts (26).
‘The measured data tend to show a strong variation with frequency at low frequencies and low grazing angles
(R-<1), and no variation with frequency at high frequencies and high grazing angles (R > 1).
Several attempts to summarize existing data have been made. One, by Chapman and Scott (27), gave an expres
sion based on Eckart’s theory:
s
010g s = -10 log 8x6? - 2.178 “Fcot?y
Which is said to fit their own data at grazing angles greater than 60°. Here f is the mean square surface slope and
vis the grazing angle.
‘Another summaty by Schulkin and Shaffor (28), relates s to the Rayleigh parameter R according to
(aR)?
where a and b are empirical constants determined by the data. Individual series of measurements were fitted by
values of b between 1.03 and 2.03, The data as @ whole (several series by differont workers) gave
099
fhsind
3.65x10"
‘The data to which this relation applies extended from fh sin 0 = 0.6 to th sin @ = 200 kHe-ft,, corresponding to R.
between 0.6 and 260. The expression for S was said to fit all the data of a series of four measurements with a stan
dard deviation of 5.4 db.
AA third expression, given by Chapman and Harris (26) is
= 3.30 log, 039-4.24 log. oA + 2.6
A= 158(uf)"/>
913 a)
b ‘ a
iT Naa
“iat aeI
rmequency wa)
oo
BAND LeveL
“0 db re CARRI eR j
% TABLE 1
Compilation of
Measurements of Loss
METHOD,
At The Sea Surface
Conditions
‘Measured Loss Freq. Wave Height Grazing Angle Ref.
db kHz ft degrees
3 25 449 30
3 30 0.2.08 8 31
(interdecile range
=10 to +3 db}
o 04.64 Wind Speeds 1058 32
5-20 knots
3 5 2 3
i: 1 5
1 35 7 34
a it es
on 53 — 04st 35
1.03
on 8 ‘Sea State 2 2252 36
“Calculated from Loss = 0.1(fh)*”. Applies to surface duct propagation only.
948 whore ¢ Is the grazing angle in degrees, v is the wind speed in knots and f is the frequency in Hz. This empirical
expression was based on octave band analyses of explosive shot returns from the sea surface (Method II) over the
range 0.4 and 6.4 kH2 and wind speeds zero to 30 knots. Fig 11 shows curves computed from this expression for
several frequencies.
The Bubbly Layer
{At wind speeds of 10 knots or to, a layer of scattering has been observed to exist (29) just below the surface
Oscilloscope photos of returns using a vertical transducer showed a return blob extending 25 feet below the sur-
face that appeared suddenly at a wind speed of 9 knots. This return was likely to have been caused by alr bubbles
‘whipped into the sea by the waves and carried by wind mixing to a depth below the surface.
Physical Processes of Back Scattering
‘The peculiar measured variation of s with angle at 60 kHz suggest three processes in different regions of angle
‘0 account for the return (Fig 12). At low grazing angles (Region 1) the scattering is due to the bubbly layer just
mentioned, causing the curves to be flat. At high angles (Region I)), the reflection from small flat wave facets
{R > 1) was said to be effective, while at intermediate angles roughness-caused scattering was postulated to occur.
‘There is approximate quantitative agreement between the measurements in Region Ill and the Beckmann-Spicci
chino formula (Eq, 1)
216 e
s,
8
sori soci
= asa eee 1 Lo
6
5
w+ gt L
a a L
0-4 L
©
nt 6 L
or L
20-4 L
— 1
2 | 0 | a
20002
20 a es
8
4 , J oo! |
ot »
ois
® r | 2 +
4 Bed , L
40-4 » Lot L
6
4 Lo. L
TOT T T T T T T T T T T T
o | 2m | 9% | 4 mw |) | wf
GRAZING ANGLE, DEG.
Figure 11. Sea surface backscattering. strongths
formula
947
GRAZING ANGLE, DEG.
vs. grazing angle computed from Chapman-Haris REGION REGION REGION
r x =
1S KNOTS
SCATTERING STRENGTH (4b)
TO KNOTS,
30405060 30
GRAZING ANGLE (deg)
20
Fig, 12. Measured sea surface back scattering strengths vs. grazing angle at 60 kHz. Different processes of scatter-
ing are suggested in three regions of grazing ancl.
9.18 REFERENCES
1. P. Beckmann and A. Spizzichino, Scattering of Electromagn
Press, 1963, Chaps. 2, 3, 4.
2 J.-L. Jones, C. B. Leslie and L. E. Barton, Acoustic Characteri
‘1964,
ie Waves from Rough Surfaces, Porgamon
of Underwater Bottoms, JASA 36, 184,
3. Lord Rayleigh, Theory of Sound, Dover Publications, New York, 1945, Vol. Il, pp 89-94,
4. W.G. Meechan, Variational Method for the Calculation of the Distribution of Energy Reflected from a Peri
‘odie Surface, Jour. Appl. Phys. 27, 361, 1956.
5. J. L, Uretsky, Scattering of Plane Waves from Periodic Surfaces, Annals of Physics 33, 400, 1965.
6. _H.S. Heaps, Reflection of Plane Waves of Sound from a Sinusoidal Surface, Jour. Appt. Phys. 28, 815, 1957.
7. C. Eckart, Seattering of Sound from the Sea Surface, JASA 25, 666, 1953,
8. E.0. LaCasce and P, Tamarkin, Underwater Sound Reflection from a Corrugated Surface, Jour. Appl. Phys.
27, 138, 1956.
9, R.F, Meyer and B. Romberg, Acoustic Scattering in the Ocean Project Trident Report 1960863, A. D. Little
Co, 1963,
10. G. R. Barnard and others, Underwater Sound Reflections from a Pressure Release Sinusoidal Surface, JASA
39, 1162, 1966.
11. Physics of Sound in the Sea, National Defense Research Committee Div. 6 Summary Technical Report 8,
1946, p. 32 and p. 95.
12. RW. Young, Image Interference in the Presence of Refraction, JASA 19, 1, 1947.
13. C. S. Clay and H. Medwin, High Frequency Acoustical Reverberation from a Rough Sea Surface, JASA 36,
2131, 1964,
14, C. Cox and W. Munk, Measurements of the Roughness of Sea Surface from Photographs of the Sun's Glitter,
J. Opt. Soc. Am. 44, 838, 1954.
18. J.W. Miles, Non-specular Reflection at a Rough Surface, JASA 26, 191, 1954,
16. L. Fortuin, A Survey of Literature on Reflection and Scattering of Sound at the See Surface, JASA 47, 1209,
1970,
17. Y. larashi and R. Stern, Wind-Wave Generated Doppler Shifts in Surface Reverberation, JASA 49, 802,
1971.
18, R.H. Mellen, Doppler Shift of Sonar Backscatter from the Surface, JASA 36, 1395, 1964.
19. _K. V. Mackenzie, Long Renge Shallow Weter Signal Level Fluctuations and Frequency Spreading, JASA
34, 67, 1962.
20, W. J. Roderick and B. F. Cron, Frequency Spectra of Forward Scattered Sound from the Ocean Surface,
JASA 48, 759, 1970,
21, M, V. Brown and G. V. Fisk, Frequency Smearing of Sound Forward Scattered from the Ocean Surface,
JASA 56, 744, 1974,
1g Ocean Wind Wave Spectra by Means of Underwater Sound, JASA 53, 910, 1973.
23. RJ. Urick, Amplitude Fluctuations of the Sound from a Low-Frequency Moving Soures in the Deep Ses,
NOLTR 74:43, 1974,
24. J, Shooter and S. Mitchell, “Acoustic Sidebands in CW Tones Received at Long Ranges” JASA 60, 829, 1976.
25. RJ. Urick and R, M. Hoover, Backscattering of Sound from the Sea Surface, etc., JASA 28, 1038, 1956,
26. A. P. Chapman and H. H. Harris, Surface Backscattering Strengths Measured with Explosive Sound Sources,
JASA 34, 1592, 1962.
27. RP. Chapman and H. D. Scott, Backscattering Strengths Measured over an Extended Range of Frequencies
and Grazing Angles, JASA 36, 1735, 1964.
28. M. Schulkin and R. Sharer, Backscattering of Sound from the Sea Surface, JASA 36, 1699, 1964.
29, R. J. Urick, Processes of Sound Scattering at the Ocean Surface and Bottom, Jour. Mar. Res. 15, 134, 1956.
30, R. J. Urick and H. L. Saxton, Surface Reflection of Short Supersonic Pulses in the Ocean, JASA 79, 8, 1947.
31. LN. Liebermann, Reflection of Underwater Sound from the Sea Surface, JASA 20, 498, 1948.
32. _R. H. Addiington, Acoustic Reflection Losses at the Sea Surface Measured with Explosive Sources, JASA.
35, 1834, 1962
33. H.W. Marsh, Sound Reflection and Scattering from the Sea Surface, JASA 35, 240, 1963.
34. R. H, Ferris and W. Kuperman, An Experiment on Acoustic Reflection from the Sea Surface, NRL Rept.
7075, 1970.
35, M.A. Pedersen, Comparison of Experimental and Theoretical Image Interference in Deep-Water Acoustics,
JASA, 34, 1197, 1963.
36. R. B. Patterson, Intensity Fluctuation of Direct and Surface Reflected Signals from a Deep Source, Jour.
Underwater Acoust. 12, 41, 1962.
219 CHAPTER 10
REFLECTION AND SCATTERING
BY THE SEA BOTTOM CHAPTER 10
REFLECTION AND SCATTERING BY THE SEA BOTTOM
Introduction
‘The sea bottom has many of the same effects on sound propagation as does the sea surface. It reflects and scatters
sound, making bottom-bounce sonars possible, as well as producing bottom reverberation. It casts a shadow in the
positive gradient water overlying it, when the water is deep. It produces an interference pattern in the radiation from
2 source; image-interference directivity patterns caused by bottom reflection have been computed by Mackenzie (1),
for rock, sand and silt bottoms.
But the return of sound from the sea bed is vastly more complex than from the sea surface for several reasons.
First, the bottom is more variable in its acoustic properties, inasmuch as it may vary in composition from hard rock
to a soft mud. Second, it is often layered, with a density and sound velocity thet change gredually or abruptly with
depth. Third, it is more likely to be laterally inhomogeneous, with different characteristics over relatively short
distances. Fourth, sound can readily enter a sedimentary bottom and be reflected back into the sea by sub>bottom
layers, or be refracted back by the steep velocity gradient in sediments. For these reasons, the loss of intensity
suffered by sound encountering the sea bed is less easily predictable than the loss at the sea surface.
Reflection versus Scattering from the Sea Bottom
The ratio of the intensity of the returned sound from a rough bottom to the intensity of an incident plane wave
. ~arising \? agg
Where Hp is the reflection coefficient that would exist if the surface were smooth, and the exponential factor is the
effect of the surface roughness, where R is the Rayleigh parameter. This expression is quoted by Clay (2) and
appears in a paper by Lysanov (3). The reflection coefficient sg depends upon the density, compressibility, rigidity,
absorption, and layering of the bottom materials, while the exponential term depends on its roughness character
istics, which, incidentally, for real world bottoms, are almost totally unknown. In other words, the exponential
term expresses the reduction in the intonsity of the reflected wave caused by scattering at the rough water bottom
interface.
‘The sound received at @ point Q above the sea bottom consists of the following contributions, shown digram-
matically in Fig 1
1. areflection from P, spreading as if originating from the image O! in the bottom of the real source O.
21 scattered sound from the bottom in the vicinity of P.
3, scattered sound from other portions of the bottom such as A, and Az , each such small area radiating spherical
waves to the receiver at Q.
4, reflection and scattering from sub-bottom layers.
5. refracted sound, entering the bottom at R, turned upward back into the water by the velocity gradient in the
bottom.
The result is that, when a short pulse strikes a rough bottom, it becomes, after reflection and scattering, along
drawn-out blob, often with a sharp high amplitude beginning followed by a decaying tal of scattering
‘The scattered and reflected components are affected differently by the distance of source and receiver above the
bottom, as illustrated in Fig 2. When the bottom is 2 reflector, the component of the transmission loss due to
spreading is 20 log(2r) in the absence of refraction; but, whon itis a scatterer, each portion of the bottom radiatos
spherical waves, and the spreading loss becomes 20 log 1? = 40 log r. The difference betwoen these two losses can
be very great.
The partition between reflection and scattering is important for bottom-bounce sonars. The partition
a} determines the fall-off of the bottom raturn with distance from the bottom and
b) determines the correlation loss of correlation sonars.
Reflection Models
Various theoretical models have been used to relate jg to the physical properties of natural bottoms. The models
range from the single two-fluid Rayleigh model (4) to models that include bottom absorption, rigidity (ie. shear
waves) and layering. The theory and the resulting expressions for Hg will not be repeated here, but may be found
104 Fig. 1. A receiver at Q receives reflected and scattered sound from a variety of sources on and in a sedimentary
bottom. A square-topped incident pulse becomes a longtailed blob at Q if the initial source O and the
receiver at Q.are non-directional
102 a)
REFLECTION
»
SCATTERING
Lene
Fig.2. In a) the bottom is a reflector and the spreading loss from the source O to Q via the bottom is 20 log
(21). In b) the bottom isa seatterer and the spreading loss is 20 log r*
103 in Brekhovskikt’s book (5), and are listed in Appendix 1. The theoretical expressions have appeared in the U.S.
literature in various forms and they are always computerized for computation.
For two fluids separated by a plane boundary, the reflection loss depends on the density and sound velocity in
the two media, Fig 3 shows diagrammatically how the loss varies with grazing angle for different combinations of
density ratio, m=p2/9; and n=c:/e2, where p, and ¢, are the density and sound velocity of the upper medium
(water) and p2 and C3.are the density and sound velocity of the lower medium (bottom). OF the four combinations
shown, a) and ch are most characteristic of sedimentary bottoms. In combination c} there exists @ critical angle,
while in a) and d) there is an angle of intromission where the loss reaches a maximum. When attenuation exists in
the lower medium, as it always does for the real ocean bottom, these angles become smeared out and obscured, as
shown by the dashed curves in the figure.
‘Some specific examples showing the reflection loss 10 log 4, in dB, computed using Brekhovskikh’s equation
3.25 (Appendix 1), as a function of grazing angle for different values of the two ratios and a dimensionless atten.
uation coefficient, are given in Fig 4. The parameter of the curves in this figure is the dimensionless
quantity e=a/2n where a! is the amplitude attenuation coefficient in units of reciprocal length and is the
wavelength,
‘Acoustics of Sediments
1) Density
In a simple additive mixture of two components, such as sediment, which is a suspension of mineral particles
in water, any property of the mixture equals the sum of the properties of the two components separately, weighted
by the proportion of each in the mixture. For example, in @ mixture of medium 1 and medium 2, let medium 1
have density p1 and let it occupy a faction f of the volume of the mixture (Bis the volume concentration). Let
medium 2 have density p2 with a volume concentration equal to (1 ~ 8), The additive “law” states that the density
of the mixture is
Pac” 80, + (1 8),
This “law has been found to hold closely fr sediments and sedimentary rocks by Nafe and Drake (35), who found
ftom numerous measurements on sediments that Pix = 2.68~ 1.669, where medium 1 is taken to be weter(p, = 1)
and medium 2, the sedimentary particles (o, = 2.68). Here f is the porosity of the sediment, equal to the volume
concentration of water inthe mixture
2) Compressional Wave Velocity
Hamilton, eta. (6), reported measurements on a variety of shallow-water sediments off San Diego using both
an initu probe method and a laboratory resonant chamber method. Additional measurements were made by
Shumway (7). Suton, etal (8) investigated the velocity in core samples of deep ocean sediments by placing par
of transducers on the sides of the cor and measuring the travel time of short pulses diametrically through the core
This same method was desried later by Winokur (9)
‘A compilation (10) of data has shown that the mixture formula, cited by Wood (11) and verified exper:
mentally in the lab on kaolr-water mixtures by Urick (12), isnot a bad approximation to the measured daa, The
mixture formula gives forthe sound velocity
¥° Wang Kix! = Cow Ps (1~ Bld yf + kg (1-91 -%
where Py, ky and py, ky are the density and compressibility of the water and solid particles, respectively. Measured
velocities ‘ae'somewhat higher than those given by this formula for low porosity partl-indurated sediments, where
‘an-to-grain contact of the sediment particles exists.
The mixture formula predicts that the sound velocity in high porosity sediments should be slightly ess than
that of water. The reason for this is that as sediment is added to water, the increase of density of the mixture is
‘more rapid than the decrease of compressibility, resulting in an initial decrease in sound velocity. The existence of
low velocity bottoms, at least for the top few feet of the bottom material, have been repeatedly inferred (13) (14)
from an observed phate reversal of the bottom reflection. Also, in stagnant inshore waters, alow velocity condition
in the bottom can be caused by gas (methane) in the bottom produced by organic decomposition (15).
3) Compressional Wave Attenuation
‘A number of investigations have been made of the attenuation of compressional waves in natural (7) (16)
(17) and artificial (18) sediments by methods involving (a), a resonant cylinder of the materials, and (b), @ source
104 3) Loss,
8 /
n>t
m>n
bo}
eo
mli this becomes,
‘where the term in parenthesis is the loss in array gain due to the interference. In the experiment described, 1/l
‘was the ratio of the intensity carried by surface-bottom multipaths to the intensity of the direct-path sound,
ith Theory
1nd roceiver are shallow and the range is not too great, the meager available data suggest the validity
wuls for coherence in a randomly inhomogeneous medium. For example, the envelope cross
it between vertically separated hydrophones for 25 and 60 kHz pulees at ranges of several
hhundrod yards was found (16) to fall to 1/e in a vertical separation of 8 feet, a distance to the patch size of the
microstructure observed independently. In Russian work (16), the transverse correlation of two shallow receivers
spaced horizontally from 20 meters to 3000 meters for 7.5 kHe over distances out to 12km was ssid to agree with
Chernov’s theoretical work. In another Russian experiment on surface reflected sound (17), the horizontal amp
‘tude coherence of 2.6 and 4 kHz CW sound was of the same order of magnitude as the horizontal coherence of the
amplitude of the surface waves — again suggesting that the acoustic coherence is of the same scale os the inhomo-
‘gonsities of the medium.
(On the other hand, when the propagation paths are deop, the coherence is maintained over much grester dis
‘tances, and the correlation coefficient falls only slowly to zero with increasing separation. An example isthe vertical
coherence of transmission over 24 mi to a deep receiver, as already plotted in Fig. 6. Another (18) is the measure:
‘ment of coherence with a distributed array on the sloping bottom near Bermuda. Using 400 Hz sound transmitted
‘over several hundred miles in the Deep Sound Channel, individual ray arrivals (making 14 to 16 loops) were found
to have a transverse phase correlation coefficient falling to 1/e at a longitudinal separation of 700 feet and a trans-
verse separation of 1300 feet. Various other observations show good coherence between hydrophones soparatad by
long distances in the Deep Sound Channel, suggesting the presence of much larger patch sizes — a “macrostruc:
ture” — in the deep sea. The dominant scale of this deep macrostructure has not yet been determined, but is doubt:
less larger by at least an order of magnitude than that already messured near the surface.
1243 9,500 FT—>[TSos5
PROJECTOR
D
TRIDENT
VERTICAL ARRAY
7
|
| ie
| 14,400 FT= + + :
|
<4 mies ___. |
HYD, NO. 1
vi 0
v3 A.V FT
v7 14.0
vi2 30.8
vig 62.6
vo 120.8
v40 300.2
$100 FT
VISITS TTTTTT:
Fig.8 An experiment to measure coherence over a transmitted path. The two pathé shown could be separated by
the time delay between them and were reflected at the sea surface at a grazing angle of 4°, Reference 9
1214 44-88
Hz
ees ee ee ee eee eee
tT 2 3 4 5 6 FB 9 OT
HORIZONTAL RANGE (KYDS)
Fig, 9. Clipped correlation of ship noise vs. horizontal range in 7000 feet of water between two hydrophones 500
feet apart. The circles and dots show the measured correlation for the recession and the approach of the
approach of the ship source.
1245 Data Summary
Table 2 i a summary of measurements of coherence length fq for transmitted signals, as reported inthe unclas:
fied literature, together with the experimental conditions under which they were obtained. The coherence length
here is taken to be the length defined by rq = {6° ptrldr, where r is the separation of the receivers being correlated
and pr) isthe phase, or clipped, correlation coefficient. ta is, in effet, the area under the function pir) if p de-
creased linearly with, then rg isthe separation for which p falls to 0.8; if p decreases exponentially with r, then
fo is the separation at which p = e”! = .37. rp gives a rough idea of the maximum useful length of an array for
achieving array gain; arrays in incoherent noise longer than about 2rg realize no further increse in aray gain. Many
of the entries in Table 2 are, because of brevity, not entirely meaningful; the references themselves should be con
Sulted in eases of special interest. Most noticeable is the wide variety of conditions applying for the measured data
‘A study of these coherence results, augmented by an understanding of propagation of sound in the se, indicates
that the coherence of transmitted sounds, and therefore the gain of deepwater arrays in incoherent noise, should
be greatest under conditions of (a) narrow frequency bands, (b) low frequencies, (c) short averaging times, () propa:
gation such that a single dominane propagation path exists, and (e} horizontal rather than vertical separation. In the
design of receiving arrays, these conditions can never be all achieved atthe same time, while others wil be incom
patible with system requirements, so that the usual trade-off procedure will be necessary In the search for the
cptimum array design.
1246 (eng eOnI8A ANSGIeL) ePnui9g se L904 O
ye wonog uo Buus ae.
"epmusiog 200 see |2uUE\o Uo uRdep 18) GOP - 0002 wored UO P
ane suor2euet 2H Lovyse
we ‘wong eH psec
‘ men, a "4 000% asst seayng vez ammo BH BBY
ai soro
2H 6O-LO
eusonuel, vug9 wo, seg 010m vuonsing THA CEST
z rrivoz0H wy “seven MOUEUS yous, Hooz HAE LE
: 22051 4.0008
z — moo. = stwoy Seve = mouse aH coe 7 O08
2H oor 002
poseuon uonaing 2H 002-001
we oman, 009-001 wea "4 008 an yous eH OOTOS—SBUE|NEMg
ting
oz omen, wpe AVAL 40086 wg 7H 008 24 006
paomyae, mo nBuajonem g
6 sure ee¥AL 330058 soaps swank moueN, aH ose wep
one
. wore sstweuy oe _ sosing 2H 007 “O68
9001
o web = “v0 sna 2H 00% "4 0000
sane “u og wide vonseyeu soaing su mast sexs
a avoauon -w 008 ‘anten MOUS ‘aig OF pure sosiea "Oy aia 9
(wx 08
ssmsu0 sens son ‘eo-0)
ot wore = wyzigo = mous aero | MOUIN MOZHNS'L wee
BONSU343Y —-NOLLVINGIUO BONE, wugza aaa isitava aw. Huai = Aowanosus = HLONST
SON3UZI3H — BNOH4ONGAH —-H3AI393U © -UBAIZO3H DUNS ©—NOLLVDVdONd © ONIDVU3AY «NV. aNaH3HOO
‘3ounos
SLN3W3YNSV3W 3ONIYIHOO 4O AUVINWNS HV INEVL
zatavL
247 16.
1"
18,
20.
a
22,
23,
REFERENCES
R. J. Urick, Principles of Underwater Sound, 2 ed., McGraw-Hill, New York, 1975, pp. 33:37.
R. J. Niolson, Array Gain, Correlation Coefficients and Patch Size, JASA 67, 60, 1977.
B. F. Cron and C. H. Sherman, Spatial Correlation Functions for Various Noise Models, JASA 34, 1732,
1962.
J. J. Faran and R. Hills, Corvelators for Signal Reception, Acoustics Research Lab., Harvard Univ., Tech.
Memo 27, 1952.
L.A. Chernov, Wave Propagation in a Random Medium, McGraw-Hill Book Co., New York, 1960, Chap.
vi
R. Bourret, Directurly ofa Linear Array in a Random Transmission Medium, JASA 33, 1793, 1961.
R. J. Urick and G. R. Lund, Vertical Coherence of Explosive Reverberation, JASA 36, 2164, 1964.
R. J. Urick, Corcelative Properties of Ambient Noise at Bermuda, JASA 40, 1108, 1966.
R. J. Urick, T. J. Tulko and I. I. Abrams, Vertical Coherence of Sound Transmitted over @ 24-mile Path,
JASA 46, 1308, 1969.
RJ. Urick and G. R. Lund, Coherence of Convergent Zone Sound, JASA 43, 723, 1968.
R. J. Urick and G. R. Lund, Vertical Coherence of Explosive Reverberation, JASA 47, 342, 1970.
R_ J. Urick and G. R. Lund, Horizontal Coherence of Explosive Reverberation, JASA 47, 909, 1970,
R. J. Urick, Coherence of Ambient Noise and the Signal from a Steady Source at Different Depths at a Deep
‘See Location, Naval Ordnance Laboratory Tech. Rep. 73-68, 1973.
R. J. Urick, Measurements of the Vertical Coherence of the Sound from a Near-Surface Source in the Sea,
JASA 54, 115, 1973,
D. C. Whitmarsh, E, Skudryzk and R. J. Urick, Forward Scatter
with the Temperature Microstructure, JASA 29, 1124, 1957.
'S. G, Gershman and Y. |. Tuzhilkin, Measurement of the Transverse Cortelation of a Continuous Signal in
the Sea, Soviet Physics-Acoustics 6, 291, 1961.
EP. Gulin and K. |. Malyshew, Spatial Correlation of Amplitude Fluctuations of a Continuous Tone Signal
with Reflections from Ocean Surface Waves, Soviet Physics Acoustics 11, 428, 1966. See also, same authors,
SPA 10, 365, 1964.
H. W. Brock, Fluctuations in Single Path Underwater Sound Transmission over a §40-mi path, Bell Tele
phone Labs Ocean Science Report 1, 1968,
R. M. Kennedy, “A Statistical Study of the Perturbation of an Acoustic Wavefront”, USN Underwater Sound
Lab. Report 868, (1968).
R. M. Kennedy, “Phase and Amplitude Fluctuations in Propagation in @ Layered Ocean", JASA 46, 737,
(1969).
R. J. Urick and G. R. Lund, “Low Frequency Coherence of Long Range Explosive Sounds and Ambient Noise
in the Deep Sea”, NOLTR 70-111, (1970).
B. E. Parkins and G. R. Fox, “Measurement of the Coherence and Fading of Long Range Acoustic Signals",
IEEE Trans, Audio and Electro. Acoust. 19, 158, (1971)
P. Wille and R. Thiele, “Transverse Horizontal Coherence of Explosive Signals in Shallow Water", JASA
50, 348, (1971)
Cf Sound in the Sea and its Correlation
qzige CHAPTER 13
MULTIPATHS IN THE SEA CHAPTER 13
MULTIPATHS IN THE SEA
Introduction
‘This chapter is intended to give a synoptic view of the paths by which sound reaches a distant point in the sea
under different conditions of depth and distance. These paths, generally speaking, become more numerous, and
‘acquire greater differences in travel time and transmission loss, as the range fram source to receiver increases. Excopt
very near the source, multipath transmission is the rule rather than the exception in sound transmission in the sea
Sound propagation in the sea has its above-water analog in electromagnetic propagation in the atmosphere, where
‘many of the same propagation phenomena are found.
Deep Water — Shallow Source and Receiver
Fig 1 attempts to show in a diagrammatic way the transmission paths botween 2 shallow source and a shallow
receiver approximately in the order of increasing range from source to receiver. Starting at a short range at the
Upper left, the direct path A is a nearly straightine path between source and receiver, over which spherical spreed:
ing, plus attenuation, determines the transmission loss. Even this path can be considered as multiple if we consider
the effects of microstructure, which cause fluctuation of the transmitted sound and the decorrelation of separated
receivers. When this path becomes equal in length to the geometric mean of the depths of source and receiver, the
surface reflection B begins to be a significant contributor to the total sound field, and coherently interferes with the
direct-path sound under conditions of calm seas, high frequencies and small grazing angles, such that the Rayleigh
Parameter is less than unity and the sea surface may therefore be said to be acoustically smooth.
When a negative velocity gradient extends upward all the way to the surface, the direct and surface paths become
bent downward into arcs of circles (for a linear gradient) and a shadow is then cast by the sea surface on the waters
below (Paths C). A receiver within the shadow receives sound only by scattering (x), diffraction (yl, and from the
bottom (2); within the shadow in deep water, 8 short pulse is received as a weak, irregular elongated blob. During
World War Il, 24 kHz pulses received well within the shadow were found (1) to be some 40 db weaker than they
would be if the shadow were absent, perhaps representing forwardscattered sound from the biological scatterers
in the Deep Scattering Layer.
Wind-mixing often creates isothermal water on top of the negative gradient, or thermocline, so as to form a
surface duct D. At frequencies well above cutoff, and for a thick layer with a calm see, the surface duct becomes
‘an effective trap, preventing the loss of sound to the abyssal depths of the sea. Under other conditions, the duct,
is s0 lossy as to be useless as @ sonar trap. There is an optimum frequency for duct transmission, such that poorer
transmission occurs at frequencies much higher or much lower than the optimum, A point well below a duct receives
You might also like
