ICES Journal of Marine Science, 53: 507–512. 1996
Relationship between acoustic backscattering strength and density
of zooplankton in the sound-scattering layer
Kohji Iida, Tohru Mukai, and DooJin Hwang
Iida, K., Mukai, T., and Hwang, D. J. 1996. Relationship between acoustic backscattering strength and density of zooplankton in the sound-scattering layer. – ICES
Journal of Marine Science, 53: 507–512.
Hydroacoustic sampling in the ocean is a useful technique for studying the biomass
and structure of the distribution of marine organisms, including zooplankton.
Acoustical and biological sampling of zooplankton in the sound-scattering layer (SSL)
off the east coast of Oshima Peninsula, northern Japan has been conducted during the
last 10 years to obtain the scale factor for converting the backscattering strength to
biological density. The volume backscattering strength (SV) was measured at 25, 50,
100, and 200 kHz, while an IKMT (Isaacs-Kidd Midwater Trawl) and a Norpac
(North Pacific standard net) were used to sample the biological organisms in the SSL.
During drifting observation at dusk, the maximum speed of the upward migration of
the SSL toward the surface was 3 m min "1, and the SV changed over a range from
"80 dB to "50 dB. Vertical tows of the Norpac indicated that euphausiids were the
major zooplankton species of the migrating SSL. For quantitative analysis, the IKMT
was towed horizontally about one hundred times, and the calculated density of
zooplankton showed a maximum of 10 g m "3. We assumed that the acoustic
reflectivity of the individual plankton was related to the size of the plankton squared,
and that the acoustic backscattering strength of biomass was proportional to the
distribution density. Results of regression analysis showed a linear relationship
between the log of zooplankton density ñ (mg m "3) and the acoustic volume
backscattering strength SV (dB), with correlation coefficients greater than 0.5 at all
frequencies.
? 1996 International Council for the Exploration of the Sea
Key words: euphausiid, sound-scattering layer, volume backscattering strength,
zooplankton.
K. Iida, T. Mukai, and D. Hwang: Laboratory of Instrument Engineering for Fishing,
Faculty of Fisheries, Hokkaido University, 3-1-1 Minato-cho, Hakodate, Hokkaido, 041,
Japan. Correspondence to Iida [tel: +81 138 40 8852, fax: +81 138 43 5015].
Introduction
All underwater objects scatter acoustic waves. Small
marine organisms such as zooplankton aggregate at
specific depths in the ocean, and the reflected sound
waves from these organisms can be recorded as a
scattering layer on the echogram of an echo-sounder.
This is the so-called sound-scattering layer (SSL) and
has been observed in the ocean throughout the world
(Sameoto, 1982; Sameoto et al., 1985). Generally, the
SSL appears at a depth of a few hundred metres in the
day-time and rises to near the surface at night.
Acoustical and biological sampling of the SSL using
echo-sounders and sampling nets was conducted over
the period 1983–1992. The survey area was the subarctic
coastal region in northern Japan, where nutrients are
abundant year-round, because this is where the cold
Oyashio Current from the north and the warm Kuroshio
1054–3139/96/020507+06 $18.00/0
Current from the south meet. It is known that during the
day-time the dense sound-scattering layer is often found
at a depth of 200 m in this area in spring and summer
(Suzuki et al., 1984; Takiguchi et al., 1988).
The present study aims to determine the SSL vertical
migration behaviour at dusk and to obtain a regression
equation of the acoustical backscattering strength on
the biological density of SSL in order to estimate
zooplankton biomass.
Acoustical observation and biological
sampling of the sound-scattering layer
The survey was conducted using the RV ‘‘Ushio Maru’’,
in the area off the east coast of Oshima Peninsula with
a depth of 300 m, from May to October every year
during the period when the cold Oyashio Current was
dominant. In the survey, acoustical observation and
? 1996 International Council for the Exploration of the Sea
508
K. Iida et al.
(a)
Codend
One metre ring
Flow meter
Depressing vane
Netsonde
(b)
17.7 m
2.8 m
2.9 m
Figure 1. Isaacs-Kidd Midwater Trawl (IKMT) net used to quantify the density of zooplankton in the sound-scattering layer. The
original IKMT (a) had a depressing vane to enlarge the net mouth, and since 1988 the improved IKMT has had a hard frame
supporting the mouth shape (b).
biological sampling were carried out during twilight at
dusk when the SSL migrates upward to the surface. The
acoustical properties of the SSL, namely the volume
backscattering strength at 25, 50, 100, and 200 kHz,
were measured by a scientific echo-sounder which had a
time-varied-gain (TVG) circuit of 20 log R.
The volume backscattering strength (SV) is defined as
the ratio of intensity of sound scattered back in the
direction of the sound source by a unit volume to the
intensity of the incident plane wave. The SV is also
equivalent to the summation of the backscattering crosssection (óbs) of scatterers involved per unit volume.
Therefore, SV is in proportion to the numerical density
(N) of scatterers ensonified, as well as the weight density
(ñ). Using decibels this can be expressed as:
SV (dB)=10 log (N · óbs)
=10 log N+TS,
(1)
where the target strength (TS) is the decibel equivalent
of óbs. Assuming the numerical density is proportional
to weight density, Equation (1) can be rewritten as
follows:
SV (dB)=10 log ñ+A,
(2)
where the constant (A) is the mean target strength of
unit weight of scatterers. If the A value, or a regression
equation between SV and ñ are obtained, the abundance
of marine organisms can be estimated from acoustical
measurements.
As for biological sampling, a Norpac net (45 cm in
diameter, 0.334 mm in mesh size) was towed vertically
from 50 m deep to the surface at a speed of 1 m s "1
repeatedly to observe the change in species composition
with time. At night, after the migration of the SSL had
finished, an IKMT was towed horizontally through the
SSL to quantify the biological density. The net mouth of
the IKMT measured about 2.8 m by 2.9 m and the mesh
size at the codend was 2 mm.
From 1983 to 1987, an IKMT with a depressor vane
to enlarge the net mouth was used, but there were some
handling difficulties so that, in 1988, it was modified to
include a hard frame to support the net mouth, as shown
in Figure 1.
The depth of the towing net and the height of the net
mouth were monitored using a netsonde, which was
attached to the top bar. The IKMT was towed for
10 min at a speed of 3 knots and the filtered volume was
calculated from the product of the area of net mouth
and the towed distance, adjusted by reading the flowmeter attached to the centre of the 1 m ring which
connected the codend to the body net. The relationship
between biological density and the acoustic volume
Relationship between acoustic backscattering strength and zooplankton density
509
Figure 2. Echogram of upward migrating sound-scattering layer (SSL) at 50 kHz observed at dusk on 3 August 1983. The SSL
began to rise 15 min before sunset with a maximum speed of 3 m s "1; it arrived at the surface 30 min after sunset. The ‘‘V’’-shaped
strong echoes, which are indicated by encircled numbers, are due to water disturbance caused by the vertical Norpac (North Pacific
standard net) hauls.
backscattering strength of the SSL was analysed at the
four frequencies.
Vertical migration and biological density
of the sound-scattering layer (SSL)
Figure 2 shows the 50 kHz echogram, recorded on
3 August 1983, of the SSL, which is seen migrating
upward to the surface at dusk. On this day, the SSL
remained at a depth of 120 m in the day-time, and slowly
began to rise about 15 min before sunset (18:41, 1000 lx
illuminance at the surface). Five minutes before sunset
(18:49, 600 lx), the ascending speed of the SSL increased
and it arrived at the surface 30 min after sunset
(19:24, 2.8 lx) and formed a stationary strong scattering
layer near the surface. The speed of ascent of vertical
K. Iida et al.
Solar illuminance (lx) –3
Density of copepods (no. m )
18:30
1200
18:54
Local time
19:00
19:30
2.0
900
1.5
600
1.0
300
0.5
0
1
2
3
4
5
6
7 8 9 10 11 12 13 14
Haul number
Density of euphausiids–3
and chaetognaths (no. m )
510
0
Figure 3. Change in the number of zooplankton sampled by repeating vertical Norpac hauls at dusk in relation to surface
illuminance. The appearance of copepods did not change greatly, but euphausiids suddenly increased after the ninth haul.
=chaetognaths; =euphausiids; # =illuminance; =copepods.
migration started at 1 m min "1, increasing to 3 m
min "1. The acoustical volume backscattering strength
of the central part of the SSL measured "80 dB at the
beginning of the ascent and increased to "50 dB at
the end of the migration.
The ‘‘V’’-shaped strong echoes in the echogram,
which appeared every 3 min, were due to water disturbance caused by the vertical Norpac hauls. The encircled
figures in the echogram show the net haul number. The
sampled organisms from the Norpac included copepods
(Acartia clausi, Clausocalanus arcuricornis, Paracalanus
parvus, Pseudocalanus sp., and Calanus tenuicornis), all
less than 1 mm, a chaetognath (Sagitta elegans), and a
euphausiid (Euphausia pacifica), both about 10 mm in
length. Euphausiids and chaetognaths caught by the
Norpac were smaller than those caught by the IKMT,
presumably due to net avoidance caused by the slow
towing speed and the narrow net opening.
Biological sampling with the IKMT has been carried
out more than 100 times every year during May to
October since 1983. Sampled organisms caught by the
IKMT were generally larger than those from the Norpac
hauls. The species collected included a euphausiid
(Euphausia pacifica), a salp (Aglantha digitale), a
copepod (Calanus cristatus), a chaetognath (Sagitta
elegans), an amphipod (Parathemisto japonica), a
gastropod (Limacina helicina helicina), and larvae of
walleye pollock (Theragra chalcogramma). Each sample
was fixed in 5% formalin, and later the individual
organisms were identified, measured, and weighed. The
biological density was calculated by dividing the wet
weight of the sampled organisms by the net-filtered
volume.
The regression equations obtained describing the
relationship between the mean volume backscattering
strength SV (dB) at the four frequencies and the
biological density ñ (mg m "3) were as follows:
25 kHz: SV=6.4 log ñ"94.2 (r=0.48)
50 kHz: SV=17.2 log ñ"148.4 (r=0.84)
100 kHz: SV= 5.3 log ñ"91.4 (r=0.51)
200 kHz: SV= 8.8 log ñ"111.0 (r=0.61).
Discussion
The change in species composition of zooplankton
sampled by continuous vertical hauls of a Norpac at
dusk revealed the biological structure of the SSL. Figure
3 shows the species composition change during the
Norpac hauls. The occurrence of copepods did not
change greatly with time, but the peak of chaetognaths
occurred in the 11th haul and the abundance of
euphausiids suddenly increased after the 9th haul. The
appearance of the peak of chaetognaths corresponded to
the time when the top of the SSL reached the towing
range of the Norpac. The sudden increase in euphausiids
corresponded to the time when the bottom of the SSL
entered the towing range of the Norpac. These facts
suggest that euphausiids, not copepods, play the most
important role in acoustic backscattering of the upward
migrating SSL. Chaetognaths located in the upper part
of the SSL and euphausiids located in the lower part of
the SSL may also be important.
The backscattering strength of the SSL increased from
"80 dB at the beginning of the ascent to "50 dB at the
end of the vertical migration. This can be explained by
the fact that micro-organisms in deeper waters successively join the SSL at dusk and form dense zooplankton
aggregations. The size of zooplankton sampled by the
Norpac was smaller than those of the IKMT. This was
probably due to net avoidance caused by the small net
Relationship between acoustic backscattering strength and zooplankton density
–40
–60
200 kHz
511
100 kHz
SV = 8.8 log ρ – 111.0 (r = 0.61)
a
–80
b
SV (dB)
–100
SV = 5.3 log ρ – 91.4 (r = 0.51)
–120
–40
50 kHz
25 kHz
–60
a
–80
SV = 17.2 log ρ – 148.4 (r = 0.84)
b
–100
–120
SV = 6.4 log ρ – 94.2 (r = 0.48)
101
2
10
3
10
4
10
101
2
10
3
10
4
10
Biological density, ρ (mg m–3)
Figure 4. The relationship between the mean volume backscattering strength SV (dB) at the four frequencies and the biological
density ñ (mg m "3) calculated from IKMT hauls. The regression lines and their 95% confidence intervals are shown in each figure.
There are two data groups at 25 kHz and 100 kHz obtained by the original IKMT (a) and the improved IKMT (b). Since the
original data possibly included the underestimation of swept volume caused by net avoidance, they were removed from
the regression analysis.
aperture, slow towing speed, and small mesh size of the
Norpac.
Figure 4 shows the relationship between the biological
density calculated from IKMT sampling and the
measured SV. It indicates roughly a linear correlation at
each frequency. At 25 kHz and 100 kHz, there are
apparent different data groups, which have higher values
of SV compared to the others. As we described before,
the IKMT was improved to include a hard frame in
1988. It is likely that the sampling efficiency of the
IKMT increased after the net improvement, resulting in
the different SV values.
It is known that some plankton are capable of
avoiding nets, especially large zooplankton such as
euphausiids, which have strong swimming ability.
Therefore the depressing vane and the top bar leading
the net of the original IKMT may have frightened them
away from the net more than the hard-frame type of the
improved IKMT.
If the plankton do avoid the net, the biomass per
filtered volume of net towing will be underestimated.
Indeed, at 100 kHz and 25 kHz, there are differences of
about 20 dB in the SV at the same density between the
two nets. This indicates that the sampling efficiency of
the improved net is 100 times better than the original
one. Unfortunately, it is impossible to correct the
efficiency of the original net because the biological
sampling was conducted under different conditions.
Accordingly, nothing else can be determined by comparing the SV and the plankton densities from the original
IKMT, and the data from the original net should not be
included in the regression calculation.
Generally, an organism in the ocean with no air
bladder, which is small in size compared to an acoustic
wavelength, will cause Rayleigh or resonant scattering
of the acoustic waves. Its backscattering cross-section
varies with the increase of the body-size-to-wavelength
ratio. However, independent of the backscattering crosssection of individual zooplankton, the volume backscattering theorem is basically valid. The volume
backscattering strength is therefore proportional to
both the numerical density and the weight density of
organisms. Because SV is proportional to biological
density, the slope of the dB regression line should equal
10.
We now look more carefully at this point. Figure 4
shows the regression lines and their 95% confidence
intervals. Since the regression coefficients at 200 kHz
(8.8&4.8), 100 kHz (5.3&2.3), and 25 kHz (6.4&2.9)
overlap each other, there is no significant difference
512
K. Iida et al.
between these frequencies. However, only at 200 kHz
does the confidence interval include the theoretical
value. The regression coefficients at 100 kHz and 25 kHz
are significantly lower than the theoretical one.
Furthermore, at 50 kHz the confidence interval
17.2&4.8 excludes the theoretical value and is significantly higher than the other frequencies. The value of
SV at 50 kHz is apparently also lower than at any other
frequency. This suggests that there is no linearity
between biological density and acoustic backscattering
strength at 50 kHz. The most likely explanation is that
the net has a non-linear sampling efficiency for small
organisms, which have a low response at 50 kHz.
Another possible explanation of the low and high
slopes of the regression lines is that they are due to the
heterogeneous backscattering cross-section of the
scatterers caused by the non-uniformity of the species
composition of the SSL.
Acknowledgements
We thank Professors N. Sano and T. Suzuki for promoting the study on plankton acoustics. We also thank the
students of Hokkaido University who helped to conduct
the experiments.
References
Sameoto, D. D. 1982. Zooplankton and micronekton abundance in acoustic scattering layers on the Nova Scotian slope.
Canadian Journal of Fisheries and Aquatic Sciences, 39:
760–777.
Sameoto, D. D., Cochrane, N. A., and Herman, A. W. 1985.
Response of biological acoustic backscattering to ship’s
lights. Canadian Journal of Fisheries and Aquatic Sciences,
42: 1535–1543.
Suzuki, T., Sano, N., and Iida, K. 1984. Spatiotemporal
structure of acoustic scattering layer off the east coast of
Oshima Peninsula. Bulletin of the Faculty of Fisheries,
Hokkaido University, 35: 243–253.
Takiguchi, N., Suzuki, T., Sano, N., and Iida, K. 1988.
Acoustic estimation of zooplankton density of acoustic
scattering layer off the east coast of Oshima Peninsula.
Bulletin of the Faculty of Fisheries, Hokkaido University, 39:
115–123.