***********  +++++++++++++++++++++
051396B.OAC  + Source:  ONR Asia +
***********  +++++++++++++++++++++
Contributory Categories: BIO,ENG,ENV

Country: Australia 

From:  DSTO Aeronautical and Maritime Research Laboratory
       PO Box 4331
       Melbourne Victoria 3001
       Telephone:   (03) 626 8111
       Fax: (03) 626 8999
       @ Commonwealth of Australia 1995
       AR No. 009-213
       March 1995

KEYWORDS: Australia, Spencer Bay; Acoustic survey and Range,
ambient noise, biogenic acoustic noise


Ian S.F. Jones, Douglas H. Cato, L.J. Hamilton,
Sandra Tavener an             d B.D. Scott

Maritime Operatio              ns Division
Aeronautical and      Maritime Research Laboratory


10.1 General summary
     The above figure summarizes the components of ambient noise
at the Wedge site, which may be taken as indicative of the
general characteristics of the Spencer Gulf region.  The total
noise at any time is determined by summing the intensities of the
components (not the decibel values).
     In the figure, averaged wind dependent noise spectra are
shown for wind speeds of 5, 10, and 20 knots.  These were
determined from plots of noise as a function of wind speeds, as
shown in Figs. 8.2 to 8.5 (IN ORGINIAL TEXT ONLY), by calculating
regression lines on the data.  The range of variability about the
averages shown in the figure can be seen in these figures.  At
high wind speeds, this variability is least and about  2 dB.  It
is greater at low wind speeds because of the variable
contribution of non wind dependent noise to the data from which
the regression lines were calculated.
     Wind dependent noise is the dominant and prevailing
component of the ambient noise over the mid frequency range.  It
shows a greater rate of change with wind speed than usually
observed in the open ocean, probably because of the much lower
levels of non wind dependent noise contribution at lower wind
speeds.  The shapes of the wind dependent noise spectra are
somewhat different to those measured in ocean waters around
Australia, which show evidence of two components - one with a
broad peak at about 500 Hz and one dominant below about 100 Hz
with noise level increasing with decreasing frequency (Cato 1976
and 1978b ). The latter component results in high noise levels at
low frequencies.  In Spencer Gulf this low frequency component
appears to be absent or very low, resulting in significantly
lower wind dependent noise levels at low frequencies.  For
example, it is about 15 dB lower at 30 Hz for a wind speed of 2
knots t an off Perth (Jones, Cato, Hamilton and Scott 1992).
     Non wind dependent noise is the residual back-round noise
from sources such as distant shipping and boats, biological
noise, surf noise etc., when other identifiable noise components
are absent.  It is the dominant component at low frequencies
(below 20 to 100 Hz, depending on conditions).  Because traffic
noise (the noise of distant shipping) is so low in Spencer Gulf
it may not dominate the low frequency non wind dependent noise as
it does in the open ocean.
     The noise from snapping shrimps is always present and varies
between the levels shown, dominating the ambient noise at
frequencies above 2 to 10 kHz.  The highest levels are reached
only for a short period just before sunrise and just after
sunset.  Otherwise, the highest levels are about 3 dB lower than
those shown.
     The evening chorus was observed on all evenings at the Wedge
site for about one hour, peaking at about 2130 hours.  The peak
levels varied from day to day by about 3 dB, the maximum values
being shown in Fig. 10.1. Measurements at other sites around
Australia show that choruses of this type occur for long periods,
in some cases throughout the year.  No such chorus was observed
at the Thistle or the Berry sites.
     The spectral peaks marked "fish" in the figure are
indicative of a chorus typical of those produced by fish
strumming their swim bladders.  Such choruses are usually
seasonally dependent since they are related to activities such as
spawning.  They may occur for weeks and in some cases months at a
time at various times of day.  Very much higher noise levels have
been observed from such choruses in other areas, extending over
frequencies from 50 Hz to 1 kHz.  The present observations are
too limited to give any indication of the likely occurrence of
these choruses in Spencer Gulf other than to show that they do
occur and it is possible that much higher levels over a broader
frequency band may occur at other times of year.  Although this
chorus was observed at the Wed ae site only, there is no reason
to expect that such choruses would be confined to Wedge.  They
should be expected at all sites.
     The contributions from passing boats or ships are not
included in the figure and can be expected to produce high noise
levels.  These were, on average, audible on less than 2% of the
noise samples, which is consistent with the numbers of ships
passing through the area per day.
     Significant intermittent noise from transients was evident
at low frequencies at certain times.  The times of occurrence
were similar at all three sites.  These are considered to be the
result of movement at the bottom in the vicinity of the
hydrophones, and caused by impact of bottom debris on the
hydrophones and the underwater canister, and possibly movement of
the hydrophones themselves.  It is possible that some noise
results from motion of shells and shell fragments on the bottom. 
These transients are not considered to be part of the ambient
noise but interaction between the recording system and the
environment, and so are not included in the summary figure. The
possible exception would be sounds generated by the motion of
debris on the bottom.  The sea floor at the sites appears to be
relatively hard comprising compacted shell fragments.  The
hydrophone system would have been more vulnerable to movement
than in many other areas where the bottom is muddy and the system
becomes fixed in the mud.  The design of a system to record
sounds in Spencer Gulf needs to take into consideration the
effects of the motion near the bottom, and the nature of the sea
floor, adequately isolating the hydrophones.  Further
investigation of the nature of transient generation at the chosen
site is required to provide the information needed to design the

10.2 Comparison of sites
The ambient noise at the Thistle site was found to be generally
similar to that at Wedge, while Berry was found to be
significantly noisier.  Wind dependent noise was similar at all
sites for the same wind speeds, but while winds speeds were
comparable at Thistle and Wedge, there were significant periods
of substantially higher wind speeds at Berry (due to sea
breezes).  Consequently, wind dependent noise can be expected to
be higher at Berry than at the other sites for significant
periods of tii-ne (particularly in the afternoon).
     Non wind dependent noise was comparable at Wedge and Thistle
but significantly higher at Berry due to clustering of boats
around Corny Point, resulting in noise levels at low frequencies
being 5 to 10 dB higher.  Since this effect depends on boat
movements, there may be significant seasonal effects.
     Shrimp noise was similar at Wedge and Thistle, but about 3
dB higher at Berry.  This difference is too small to be of
practical significance.  The evening chorus and the fish chorus
were observed only at the Wedge site.  There is no reason to
suppose that fish choruses would be confined to Wedge.  They are
seasonal and may occur at the other sites at other times of year. 
The presence of the evening chorus is, however, significant
because data from other areas suggest that it is less seasonally
dependent.  On the data sample available, it is not possible to
state the relative occurrence of fish choruses at the sites and
all should be considered to be equally likely to show significant
choruses at certain times of year.  It seems likely that the
evening chorus is more prevalent at Wedge but it may also occur
at the other sites.  The evening chorus is significant for little
more than one hour and its timing is very predictable.

10.3 Forecast
     The following forecast is made in the absence of rain over
the hydrophones which would contribute significant noise at
higher frequencies.  It also excludes the presence of close
shipping since this is relatively infrequent.  Since the main
variation in noise level in the summary graph of Fig. 10.1 is
wind dependent, a forecast of the ambient noise can be made from
the wind speed statistics for Neptune Island and the Wedge site
(see analysis in Chapter 5).  For more than 90% of the time, the
noise level will be less than that indicated by the "20 knots"
curve in Fig. 10.1. For about 40% of the time it will be less
than that shown by the "10 knots" curve, and less than the "5
knots" curve for about 12% of the time.  At the Berry site the
wind noise would be less than the curves shown for a
significantly smaller proportion of time than at the Wedge or
Thistle sites.
     Shrimp noise can be expected to be in the lower half of the
range shown except for about two hours near dawn and again near
dusk when it would reach a peak near the highest levels shown.
     The evening chorus at Wedge peaked at about 2130 in
November.  Its timing may be related to the time of dusk, and
once established for a particular time of year would be expected
to show relatively little variation from evening to evening.  It
is possible that a similar chorus could be evident at the other
     Insufficient is known about fish choruses in the region to
predict their occurrence or their acoustical characteristics.  It
is likely that they occur at all sites at certain times of year,
and could produce noise levels up to 15 dB higher than those
shown in some part of the frequency band 50 Hz to 1 kHz.  These
choruses are intrinsically predictable in both their diurnal and
seasonal occurrence from measurements made once in the
appropriate season.
     Dolphin and fish sounds can be expected to contribute to the
noise but generally this contribution is covered by the non wind
dependent noise.  Whale sounds are likely to be significant for a
sufficiently small proportion of the time to be negligible.
     The Berry site is likely to be significantly noisier than
the other two sites over most of the measured frequency band for
much of the time.

10.4 Comparison with the ambient noise off Perth
     Spencer Gulf is substantially quieter than the region
investigated off Perth (Jones, Cato, Hamilton and Scott 1992) for
a number of reasons.  Noise as a function of wind speed is
significantly less at low frequencies (less than about 100 Hz)
because of the absence of the low frequency wind noise component. 
At 20 knots the noise at 30 Hz is about 15 dB lower.  Winds in
Spencer Gulf are on average significantly lower than off Perth. 
Traffic noise is very low and non wind dependent noise is
substantially less than off Perth.
     The Gulf wind noise at high frequencies is comparable with
Perth up to wind speeds of 10 knots and a few decibels higher at
20 knots.  It is difficult to determine if the difference at 20
knots is a real effect or due in part to the uncertainties of
determining wind dependence given the spread of the data and the
contributions from non wind dependent sources.  Part of the
difference may be real and due to differences in the acoustical
characteristics of the sites.  However the substantially greater
occurrence of low winds speeds at this site compared to the
waters off Perth ensure that, on average, wind dependent noise at
high frequencies will be lower.  The wind speed is equal to or
less than 10 knots for 40% of the time in Spencer Gulf compared
with 14% of the time off Perth.
     The evening- chorus at Wedge is comparable in level to that
observed off Perth but about an octave lower in frequency and
evident for a shorter period of time.
     Whale sounds are likely to be a rare occurrence in Spencer
Gulf whereas off Perth humpback whale sounds can be expected to
contribute significantly for about four months of the year.
     Snapping shrimps are significant in Spencer Gulf but not in
deep water sites off Perth (where they would be expected to be of
comparable levels in shallow water).  This would result in higher
levels in Spencer Gulf above 5 kHz compared with the noise for 10
knots wind off Perth.
     At the Wedge site, the "noise floor" - the lowest levels of
ambient noise measured (ie. very low winds speeds, lowest levels
of non wind dependent noise) is 10 to 15 dB less than off Perth
for frequencies between 20 Hz and 2 kHz, the greatest difference
being at frequencies around 500 Hz.  For wind speeds less than 10
knots, the noise at Wedge is from 7 to 20 dB lower at frequencies
between 20 and 100 Hz, the variation depending on relative levels
of both wind dependent and non wind dependent noise.  It is 0 to
15 dB lower between 200 Hz and 5 kHz, the difference increasing
as wind speed decreases below 10 knots.
End Part II/III

CMR Disclaimer==================================================
This document could contain information all or part of which is
or may be copyrighted in a number of countries.  Therefore,
commercial copying and/or further dissemination of this text is
expressly prohibited without obtaining the permission of the
copyright owner(s) except in the United States and other
countries for certain personal and educational uses as
prescribed by the "fair copy" provisions of that countries
Copyright Statues.