ICES Symposium on Acoustics in Fisheries, Montpellier June 2002
Paper No 123
The effect of water temperature and salinity
on echo sounder measurements
by
Helge Bodholt
Simrad, 2.May 2002
1 The sound speed
Acoustic measurements are performed at many different places and at
different times around the year. The water temperature and the water
salinity vary, and with this the characteristics of the echo sounder
transducer. How does this affect the accuracy of split-beam target
strength measurement and biomass abundance estimation? The speed
of sound is the fundamental parameter here, and both temperature and
salinity have influence on the speed of sound (Del Grosso 1972).
2 Important transducer parameters
In Simrad Scientific Sounders, EK500, EY500 and EK60, there are a
number of transducer parameters, which can be entered. Four of these
are important for the discussion here:
Angle sensitivity, Q
Beamwidth, X
Transducer gain, J
Two way beam angle, \
Simrad has specified default values for each transducer type. The
default values are based upon transducer measurements in our water
tank with fresh water at the temperature 18 degrees Celsius. The sound
speed is here FR =1476 m/s.
The default values are named QR , X R , J R and \ R .
In the sea, the sound speed varies, and consequently the transducer
parameters vary.
Simrad P.O.Box111, N-3191 Horten, Norway.
E-mail: helge.bodholt@simrad.com
1
2.1 Angle sensitivity
We are looking for the first order effect only, and the angle measurement
in one plane is simplified to a measurement of the phase difference
between the signals received on the centre points of the transducer
halves.
G is the distance between the
centre points
M is the measured phase difference
Y is direction of the incoming waves
(angle from the beam axis)
I is the frequency
O is the wavelength
F OI is the sound speed in the
water in front of the transducer
2S 2SI
M= G sin Y = G sin Y
O F
For small angles, this formula may be approximated to
2SIG
M= Y
F
The coefficient in front of Y is called the angle sensitivity, Q . It is the
ratio between the measured phase angle and the geometrical angle to
the fish.
2SIG 2SIG F R F
Q = = = QR R
F FR F F
2.2 Beamwidth
For a transducer where the active area is circular and has a diameter ',
the beamwidth (in radians) is approximately
X =
O
=
F =
FR
F
= XR
F
' 'I 'I FR FR
2
2.3 Transducer gain
If the transducer has
an active area $
and an electro-acoustic efficiency K
then the transducer gain can be written as
4S$ 4S$I 2 F R
2 2
4S$I 2 FR
J = K 2 = K 2
= K = J R
O F FR
2
F2 F2
2.4 Two way beam angle
The equivalent two-way beam angle is the solid angle of an ideal conical
beam, which would produce the same echo of a randomly distributed
biomass as the real transducer (MacLennan and Simmonds 1992). This
equivalent ideal beam has a flat response inside the beam and zero
outsides. The equivalent beam angle is related to the transducer
beamwidth by
\ D X2
where X is beamwidth in radians.
and D is approximately 0.00017 (Urick 1983)
The variation with sound speed will be
F2 F2
\ D XR 2 2
\R 2
FR FR
3 Target strength measurement
3.1 Target on axis
We shall first consider the situation that the fish is on the beam axis.
Linear quantities are used. They give a better physical understanding
then the logarithmic dB-quantities.
Vbs is the backscattering cross section of the fish (m 2 )
TS=10 logVbs is the target strength (dB)
S is the transmitter power (W)
The source level, expressed by the sound intensity at a point UR metres
in front of the transducer, is
S
L 6/ J 2
(W/m )
4SUR
2
UR is usually set to 1 m.
3
The echo level received at the transducer face is (H.Bodholt 1990)
2
§ UR · V EV SJV EV
L( L6/ ¨ ¸ 2
©U¹ U 4SU 4
2
UR / U represents one way geometrical spreading loss. The absorption
loss is ignored here. It would disappear later, as does the spreading
loss. The equation allows calculation of the fish backscattering cross
section
4SU 4 L (
V EV
SJ
The echo sounder executes this calculation. If the default value J R of the
transducer gain is used, the measured V 0 may be different from the real
V bs.
4SU 4 L ( 4SU 4 L ( J J
2
FR
V0 V EV V EV 2
SJ R SJ J R JR F
An acoustic calibration, with a metal sphere of known target strength
suspended below the transducer, gives a correct value of J at the
current sound speed. It is therefore recommended to perform the
calibration in order to obtain a correct target strength measurement.
However, a successful calibration can only be performed in a sheltered
area under calm sea condition. If the sound speed during the calibration
differs from the sound speed at the survey area, the transducer gain
should be set according to the previous shown formula
2
FR
J JR
F2
with J R now being the transducer gain found during calibration, and F0
the sound speed there.
3.2 Beam pattern compensation
The two-way beam pattern in one plane is close to the mathematical
formula:
2
§ Y ·
% 6¨ ¸ (dB)
©X/2¹
As shown earlier, the beamwidth varies with the sound speed, and the
beam pattern varies accordingly
2 2
§ Y · § FR ·
% 6 ¨¨ ¸¸ ¨ ¸
© XR / 2 ¹ © F ¹
The software in the echo sounder estimates the direction to the fish from
Y M / Q and compensates the beam pattern with the formula
4
2 2
§M /Q· § M ·
%0 6¨ ¸ 6¨ ¸
© X/2¹ © QX / 2 ¹
The parameters Q and X vary with the sound speed, and the accuracy
of the target strength measurement depends on the values of Q and X
entered by the user. Two cases are, both leading to a correct result, are
presented here:
Case I. The default values are used.
The phase angle M is connected with the angle to the fish by
the current angle sensitivity: M QY
2 2 2
§ QY · § Y · § FR ·
%0 6 ¨¨ ¸¸ 6 ¨¨ ¸¸ ¨ ¸
© QR X R / 2 ¹ © XR / 2 ¹ © F ¹
We see here that the beam pattern used for compensation
varies with the sound speed exactly like the real beam pattern.
Case II. Simrad provides a PC-program “Lobe” for calibration.
It stores a great number of TS-values, when the calibration
sphere is moved through the beam, and finds a best-fit value
for the beamwidth. The angle sensitivity is left at its default
value. The angle sensitivity is difficult to check, and since the
formula for beam pattern compensation has the product QX it is
not really necessary to adjust Q and X separately. This means
that the beamwidth obtained in “Lobe” may not be the true
beamwidth, but it is calculated so that the product QX gives
correct beam pattern compensation.
3.3 Example
A research vessel is equipped with Simrad EK500 Scientific Sounder and
transducer ES38B. The default values for this transducer is:
Q =21.9
X =7.1 degrees
J =450, G=10log J =26.5 dB
Calibration is performed in a sheltered bay with a sea water temperature
of 10 degrees Celsius, corresponding to a sound speed of FR 1490 m/s.
A transducer gain of 440 (26.4 dB) is obtained during the calibration and
entered into the echo sounder. The PC-program “Lobe” shows a
beamwidth of 7.2 degrees, which is also entered into the echo sounder.
The angle sensitivity is left unchanged. The echo sounder is now set for
accurate target strength measurements in seawater at 10 degrees
Celsius.
5
The vessel goes to arctic seas with 0 degrees Celsius and a sound
speed of F 1450 m/s. It is impossible to calibrate in open sea, so a new
transducer gain is calculated manually:
2
F 1490 2
J J R R2 JR J R 1.056 , corresponding to an increase of 0.2 dB
F 1450 2
If the transducer gain obtained during the calibration remains unchanged
in the echo sounder, the measured backscattering cross section V 0 of a
fish is 1.056 higher then the real cross section. The target strength is
overestimated with 0.2 dB.
If the new transducer gain, 26.6 dB is entered into the echo sounder, the
target strength measurements will be correct.
The angle sensitivity and the beamwidth are not changed. As shown
earlier in case I, the beam pattern compensation will still be correct.
In this example an error of 0.2 dB could occur. This is not a large error,
but other examples, with greater deviations in water temperature and
salinity, could show larger errors. An adjustment of the transducer gain
according to the current sound speed eliminates this error.
Correction curve for transducer gain
after calibration at 10 deg.C
0.4
0.2
G-Go (dB)
0.0
-0.2
-0.4
0 10 20 30
Sea surface temperature (deg.C)
Salinity 3.5%
4 Volume backscattering strength
The biomass abundance is estimated from the measurement of the
volume backscattering strength s V and integration of this quantity. The
measurement of s V and how it is affected by changes in the speed of
sound, is the topic of this chapter.
The echo level (sound intensity) received at the transducer face is
6
U0 2
L( L6/ 4 V9 9
U
FUW
V is the sampling volume, 9 \ U2
2
where \ is the equivalent ideal solid beam angle of the transducer
F U is the sound speed at the depth of the biomass
W is the pulse duration
Simrad scientific echo sounders utilise these formulas to calculate s V .
L( U 2 1 L( 4SU 2
V9 2
L6/ U0 \FUW / 2 SJ\F UW / 2
If the default values J R and \ R are used, the measured V0 could be
expected to differ from the real V9
2 2 2
L( 4SU J R\ R F FR
V0 V9 2 2
V9
SR J R\ R F UW /2 J\ FR F
but luckily, the variations in J and \ cancel each other.
5 Other transducer parameters
The variations of Q , X , J and \ discussed above, are caused by the law
of physics and are therefore common for all transducers. In addition
each transducer type may show changes in impedance and sensitivity,
when the temperature changes.
References
x Del Grosso, 1972 Tables of the speed of sound, J.acoust.Soc.Am.52
x R.Urick, 1983 Principles of underwater sound, McGraw Hill, New York
x D.MacLennan and J.Simmonds, 1992 Fisheries Acoustics, Chapman
& Hall, London
x H.Bodholt, Fish density derived from echo-integration and in-situ
target strength measurements, ICES, C.M.1990/B:11 Sess.R
