Content uploaded by D. R. Bohnenstiehl
Author content
All content in this area was uploaded by D. R. Bohnenstiehl
Content may be subject to copyright.
Broadband calibration of R//V Ewing seismic sources
M. Tolstoy, J. B. Diebold, S. C. Webb, D. R. Bohnenstiehl, E. Chapp,
R. C. Holmes, and M. Rawson
Lamont-Doherty Earth Observatory of Columbia University, Palisades, New York, USA
Received 12 April 2004; revised 6 June 2004; accepted 22 June 2004; published 27 July 2004.
[1] The effects of anthropogenic sound sources on marine
mammals are of increasing interest and controversy [e.g.,
Malakoff, 2001]. To understand and mitigate better the
possible impacts of specific sound sources, well-calibrated
broadband measurements of acoustic received levels must
be made in a variety of environments. In late spring 2003 an
acoustic calibration study was conducted in the northern Gulf
of Mexico to obtain broad frequency band measurements of
seismic sources used by the R/V Maurice Ewing. Received
levels in deep water were lower than anticipated based
on modeling, and in shallow water they were higher. For
the marine mammals of greatest concern (beaked whales) the
1–20 kHz frequency range is considered particularly
significant [National Oceanic Atmospheric Administration
and U. S. Navy, 2001; Frantzis et al., 2002]. 1/3-octave
measurements show received levels at 1 kHz are 20 –33 dB
(re: 1 mPa) lower than peak levels at 5 –100 Hz, and decrease an
additional 20–33 dB in the 10–20 kHz range. INDEX
TERMS:3025 Marine Geology and Geophysics: Marine seismics
(0935); 3094 Marine Geology and Geophysics: Instruments and
techniques; 6615 Public Issues: Legislation and regulation; 6620
Public Issues: Science policy; 4259 Oceanography: General: Ocean
acoustics. Citation: Tolstoy, M., J. B. Diebold, S. C. Webb, D. R.
Bohnenstiehl, E. Chapp, R. C. Holmes, and M. Rawson (2004),
Broadband calibration of R/V Ewing seismic sources, Geophys. Res.
Lett.,31, L14310, doi:10.1029/2004GL020234.
1. Introduction
[2] As anthropogenic activity in the oceans increases, the
impact of these activities, including shipping, naval oper-
ations and seismic exploration, on the background noise
levels of the oceans is a source of growing concern
[National Research Council of the National Academies,
2003]. In addition, concerns have been raised about specific
sound sources, in particular navy mid-range active sonar
systems, having potentially physically damaging conse-
quences on marine mammals [e.g., Balcomb and Claridge,
2001; Cudahy and Ellison, 2002], and navy low-frequency
sonar impacting behavior [e.g., Miller et al., 2000]. The
response of different species to different acoustic sources
also is poorly understood, though a number of studies have
been done [e.g., Richardson et al., 1995; McCauley et al.,
2000]. For sperm whales exposed to seismic surveys, some
studies suggest indifference to low levels [e.g., Madsen et
al., 2002] or relatively short-distance aversion [Stone,
2003], whereas others suggest changes in calling patterns
or behavior at long distances [e.g., Gordon et al., 2004].
[3] Seismic research, including oil exploration and geo-
physical studies, has been ongoing for decades. Over 90
large seismic vessels are currently in operational condition
[Schmidt, 2004], with perhaps 15–20 active on any given
day. Only one stranding event with a plausible spatial and
temporal correlation has been recorded [Malakoff, 2002],
and there is no proof that this correlation indicates a causal
link. However, unequivocal behavioral and distributional
effects have been demonstrated, occasionally at distances of
20 km or more [Richardson et al., 1995, 1999]. Therefore, it
is prudent to better quantify the acoustic output of such
seismic systems. It is important also that the characteristics
of these sources be described at a broad range of frequencies,
since different marine mammals are sensitive to different
frequencies.
[4] During seismic operations subject to US jurisdiction,
increasingly strict guidelines are adhered to for minimizing
impacts on marine mammals in accordance with the Marine
Mammal Protection Act (MMPA) of 1972. These include
careful monitoring for marine mammal activity prior to and
during seismic operations, and a gradual ‘ramp-up’ of the
size of the operating seismic array over the course of 30–
60 minutes. The ramp-up is designed to provide a warning
to marine mammals that may not have been detected
acoustically or visually, and allow them time to leave the
immediate area. When mammals are seen within or near
designated safety radii, it is now a common requirement that
the airguns be powered down.
[5] Prior to operation an authorization to ‘‘harass’’ marine
mammals must be obtained from the National Marine
Fisheries Services. This requires a detailed environmental
assessment as well as extensive review and a public
comment period. Once an authorization has been granted,
and the cruise takes place, marine mammal activity in the
vicinity of seismic operations is monitored closely. At
present, National Marine Fisheries Service defines the radii
with received levels of 190 dB and 180 dB re 1 mPa (rms) as
safety radii for pinnipeds and cetaceans, respectively. The
radii with received levels 170 dB and 160 dB re 1 mPa (rms)
are considered to be distances within which some marine
mammals are likely to be subject to behavioral disturbance.
For seismic experiments conducted by the R/V Maurice
Ewing the sizes of these radii previously have been based on
modeling but here we present results from the first well-
calibrated broadband measurements of the R/V Maurice
Ewing’s airgun array.
[6] In this paper, received sound pressure is expressed as
the root-mean-square (rms) pressure level measured in mPa
(re 1 mPa), which is a measure of the average pressure over
the (variable) duration of the pulse. The calculation is done
within a window length sufficient to capture the entire pulse
(for this study, 0.5 and 1.0 s window lengths were used for
GEOPHYSICAL RESEARCH LETTERS, VOL. 31, L14310, doi:10.1029/2004GL020234, 2004
Copyright 2004 by the American Geophysical Union.
0094-8276/04/2004GL020234$05.00
L14310 1of4
the deep and shallow sites respectively). The rms pressure is
the preferred measurement reported in virtually all marine
mammal studies. Note dB values reported in water differ
significantly from those reported in air due to different
reference systems and differing densities and sound speeds
between the two mediums.
2. Experiment
[7] The calibration work was conducted aboard the
R/V Maurice Ewing in late May and early June of 2003
in the Gulf of Mexico (Figure 1). Calibration measurements
were conducted at shallow (30 m) and deep (3200 m)
water sites. Plans to calibrate the array in slope water sites
could not be carried out due to moderate winds, which
prevented confident monitoring for marine mammals
(specifically beaked-whales) at the necessary ranges. The
seismic sources calibrated were a 20-airgun array, which
contained subsets closely resembling the 6-, 10-, 12-, and
20-gun arrays to be used during future seismic programs, as
well as a 2 GI gun array.
[8] A spar buoy was adapted to include two broadband
hydrophones hung beneath the buoy, with a near-real time
telemetry link to the ship. The depths of the hydrophones
were adapted by altering the length of cable. At the deep
water site, the two hydrophones were deployed with 18 and
500 m of cable. At the shallow water site, both hydrophones
were deployed at 18 m depth. The hydrophones were
sampled at rates up to 50 kHz, to allow sound levels to
be characterized as high as 25 kHz.
[9] The two hydrophones used with the LDEO spar buoy
were based on Benthos Company Model AQ-1 hydro-
phones. The hydrophones were calibrated after the cruise
in the U.S. Navy TRANSDEC facility in San Diego. They
have a specified acoustic sensitivity of 202.5 ± 1 dB
relative to 1 V/mPa with a flat frequency response (±1.5 dB)
in a frequency band from 1 Hz to 10 kHz. The buoy
hydrophones are essentially omni-directional (±1 dB) below
5 kHz; however they become more directional, introducing
up at a 10 dB uncertainty in the spectra at the highest
frequencies. The post-recording processing corrected for all
filter and instrument responses to give accurate records of
the airgun pressure signal at all relevant frequencies up to
25 kHz.
[10] Airgun arrays are designed to focus energy down-
ward rather than to the sides, and the design of the R/V
Maurice Ewing arrays (athwart ship) leads to the highest
received levels astern and forward of the ship relative to the
port and starboard received levels at equivalent distances.
Therefore, while shots were measured from a range of
azimuths, only results from in-line shots were used to
estimate radii (Table 1 and Figures 2 and 3), as these
Figure 1. The study area for the May – June 2003 acoustic
calibration study in the northern Gulf of Mexico, showing
ship tracks at the three planned calibration sites. Calibra-
tions were only conducted at the deep and shallow sites due
to weather constraints at the slope site.
Figure 2. Received levels at deep calibration site for 6, 10,
12 and 20-gun arrays. Received levels are shown for the
shallow (18 m) hydrophone (gray squares) and the deep
(500 m) hydrophone (black circles). Note there is a paucity
of data for the deep hydrophone due to clipping at the close
ranges, and instrument related problems at the far ranges.
However radii estimates must be based on the deep
hydrophone measurements since the shallow hydrophone
is subject to Lloyds Mirror effects. Clipped measurements
are indicated by open symbols, and provide a lower limit on
the possible dB level at that range.
Table 1. Measured Values for 160 – 190 dB re 1 mPa (RMS)
Radii
a,b
Site/Array
Measured
190 dB
Measured
180 dB
Measured
170 dB
Measured
160 dB
Deep 20 NC NC NC 2.5 km
Deep 12 NC NC NC 2.5 km
Deep 10 NC NC NC >2 km
Deep 6 NC NC NC 1.5 km
Shallow 20 NC 3.5 km 7 km 12 km
c
Shallow 12 NC 2km 5.5km 9km
Shallow 10 NC 2km 4km 9km
Shallow 6 NC 1.5 km 4 km 7 km
Shallow 2 GI NC NC >0.5 km 1.5 km
a
Deep site hydrophone may have been as shallow as 330 m, and so larger
values may exist at greater depths.
b
NC indicates that results for the dB level were not constrained by the
available data, mainly because measurements made close to the array were
clipped. The proximity of the closest measurement to the airguns can be
determined by looking at Figures 2 and 3.
c
This value may extend beyond 12 km, but no measurements were made
beyond 11.7 km where a value of 160 dB was received – see text.
L14310 TOLSTOY ET AL.: CALIBRATION OF SEISMIC SOURCES L14310
2of4
show the highest received levels at any given distance,
and therefore represent the most conservative (i.e.,
precautionary) radii. Using these estimates provides the
maximum protection for marine mammals in the area and
simplifies the observation procedures by eliminating the
need to consider azimuth in the observations.
[11] During operations with the 20-gun array, the number
of airguns active varied from 6 to 20. The 20-gun array was
discharged every 2 min in the following sequence: 6 guns
(two shots), 10 guns (2), 12 guns (2) and 20 guns (2).
The 2 GI guns were discharged every 30 s. While towing
the arrays, the R/V Maurice Ewing approached the spar
buoy from 10 km away, passed 100 m to the side of the
buoy, and continued until it was 10 km beyond the buoy.
[12] The deep site calibration, conducted on 30 May
2003, recorded approximately 145 shots from the 20-gun
array, equally divided among the four subset arrays with
6, 10, 12 and 20 airguns. The shallow site calibration,
conducted on 2 June 2003, recorded a total of 290 shots,
using the 20-gun array and its smaller subsets as well as a
2 GI gun array.
3. Results
[13] The records of the airgun pulses when the ship was
closest to the buoy are clipped and underestimate the
received levels. We examined all records before correcting
for the instrument response and have labeled the data points
from those records when some clipping occurred. Open
symbols on Figures 2 and 3 represent clipped data. Clipping
occurs when the signal exceeds the dynamic range of the
digitizer, which leads to a ‘‘squaring off’’ of the peaks and
troughs in the signal. On a large subset of the clipped
records, the data are clipped only on one side (either just the
positive or just the negative values) because of a nonzero
mean due to pressure variations from buoy heave.
[14] For the deep site, only the 160 dB radii were clearly
documented, given the clipping of records at the closest
ranges. The 160 dB distances observed via the deep
hydrophone suggest that the previously-predicted 160 dB
radii tend to overestimate actual 160 dB distances in deep
water (see Figure 2 and Table 1 versus Table 2). We can
infer from the unclipped data, based on either spherical or
cylindrical transmission loss, that the 180 dB radii for all
arrays should occur at less than 1 km, and likely signifi-
cantly less than 1 km. These results will need to be
confirmed in future experiments with a larger number of
observations at the closer distances.
[15] The shallow hydrophone recordings show signifi-
cantly lower dB levels than recordings from the deep phone,
due to a Lloyds Mirror effect (destructive interference
between the direct arrivals and the reflections from the sea
surface). This serves as a reminder that marine mammals at
shallow depths in deep water areas can be exposed to levels
considerably lower than the maximums received at deeper
depths. However, this also reminds us that to make accurate
estimates of maximum received levels at a given range
measurements must be made at several depths. Ideally,
future deep water measurements should utilize a vertical
hydrophone array to ensure that peak levels at a given range
are recorded regardless of depth.
[16] A related caveat for the deep site is that the deep
hydrophone was at a maximum depth of 500 m set by the
cable length, but may have been shallower due to drift of the
buoy in a strong current. Analysis of reflected arrivals
indicates it may have been as shallow as 330 m. Modeling
of the effect of the free surface reflection suggests the peak
signal strength at close ranges may be lower at 330 m than
at deeper depths.
[17] For the shallow site, a larger number of measurements
were obtained, providing empirical data on the 180, 170 and
160 dB radii for most of the airgun configurations (Figure 3
and Table 1). Due to clipping of close range arrivals, no
measurements of the 190 dB radii were made. The 20- and
12-gun 180 dB radii were estimated based on measured levels
that were close to 180 dB, but no measurements were made
above 180 dB that were not clipped. The 170 dB radii were
well documented for all but the 2 GI gun array, and 160 dB
radii were documented for all arrays. These measurements
Figure 3. Received levels at shallow calibration site for
2 GI gun (bottom left), 6-, 10-, 12- and 20-gun arrays.
Received levels are shown for the two phones, both at 18 m.
Table 2. Predicted Values From Ray-Based Modeling for 160 –
190 dB re 1 mPa (RMS) Radii
a
Array
Predicted
190 dB
Predicted
180 dB
Predicted
170 dB
Predicted
160 dB
20 gun 0.400 km 0.95 km 3.42 km 9 km
12 gun 0.3 km 0.88 km 2.68 km 7.25 km
10 gun 0.25 km 0.83 km 2.33 km 6.5 km
6 gun 0.05 km 0.22 km 0.7 km 2.7 km
2 GI gun 0.015 km 0.05 km 0.155 km 0.52 km
a
Note that the same predicted values were used regardless of depth of
site. These were the numbers used for the 2003 Gulf of Mexico Incidental
Harassment Authorization (IHA) application, Environmental Assessment
(EA) and fieldwork, as safety radii and potential harassment criteria. Some
of these values are different from values quoted in more recent IHA
Applications and EAs, which are based on re-interpretation of model
outputs. Future IHA applications and EAs from LDEO will incorporate the
results detailed in this paper.
L14310 TOLSTOY ET AL.: CALIBRATION OF SEISMIC SOURCES L14310
3of4
suggest that, for shallow water, previously-estimated 180,
170 and 160 dB radii were underestimates of the actual
distances where such levels occur (see Figure 3 and Table 1
versus Table 2), and should be extended, particularly for the
180 dB radii. This result may vary with seafloor properties as
signals over less acoustically reflective seafloors will decay
more rapidly.
[18] Figure 4 shows energy spectral density, appropriate
for pulsed or transient signals, of a 20-gun array shot at the
deep and shallow sites. 1/3-octave levels also are shown,
since these are the most relevant values for marine mammal
hearing. At the deep water site (for the deep phone), the
energy peaks between 5 and 20 Hz and falls off rapidly
above 100 Hz. At the shallow water site, the spectrum peaks
between 30 and 80 Hz, with apparent attenuation of the
lowest frequencies, but also falls off rapidly at frequencies
above 100 Hz. For both sites, levels at 1kHzare
approximately 40 dB less than the peak values for the
energy spectral density, and 20 and 33 dB less for the deep
and shallow sites respectively using 1/3-octave levels.
Energy levels continue to drop at progressively higher
frequencies, with 10 – 20 kHz levels being 30–40 dB
lower than at 1 kHz for the energy spectral density, and
again 20 and 33 dB less for the deep and shallow sites
respectively using 1/3-octave levels.
4. Summary
[19] Results from the 2003 field program show that, for
deep water, the previously utilized 180 dB and 160 dB radii
may be conservative (overestimated), based primarily on the
measured 160 dB levels for the 20- and 12-gun arrays. For
the shallow water, 180– 160 dB radii previously used
should be expanded as detailed in Table 1. Note, for all
these estimates, we have endeavored to use the maximum
received levels at any given range rather than the average
received level, to ensure the values used are conservative.
[20] These results indicate that in shallow water, rever-
berations play a significant role in received levels. Previous
modeling to estimate radii for permit applications had not
accounted for bottom reverberations. Future modeling of
seismic energy propagation should account for this effect,
especially in shallow waters. The definition of what con-
stitutes shallow water, and what constitutes deep water is a
problem that should be tackled through both modeling
incorporating reverberations, and through continued calibra-
tion measurements. In the meantime, caution should be taken
to maintain appropriately large safety radii in shallow water
operations, and consideration of these concerns should be
incorporated into future seismic cruise planning.
[21] Spectra show that, as expected, the majority of the
energy from the seismic arrays is in the 5–100 Hz range.
Levels at 1 kHz are 20–40 dB lower than those at the
frequencies with peak energy, and levels continue to diminish
significantly as frequency increases above 1 kHz. This is
particularly noteworthy because of recent concern over
the sensitivity of beaked whales to seismic sources [e.g.,
Malakoff, 2002]. Beaked whales are believed to be sensitive
to frequencies in the 1 – 20 kHz range and higher, and so it is
important to realize that seismic sources have significantly
reduced energy at those frequencies.
[22]Acknowledgments. We thank P. Tyack and R. P. Dziak for
constructive and thoughtful reviews, and W. J. Richardson for much
valuable input. We thank the Captain, crew and science party aboard the
R/V Ewing. This work was support by the NSF (OCE03-17888). LDEO
contribution 6626.
References
Balcomb, K. C., III, and D. E. Claridge (2001), A mass stranding of ceta-
ceans caused by naval sonar in the Bahamas, Bahamas J. Sci.,2, 2 – 12.
Cudahy, E., and W. T. Ellison (2002), A review of the potential for in vivo
tissue damage by exposure to underwater sound, report for Department of
the Navy, 6 pp., Dept. of the Navy, Washington, D. C.
Frantzis, A., J. C. Goold, E. K. Skarsoulis, M. I. Taroudakis, and V. Kandia
(2002), Clicks from Cuvier’s beaked whales, Ziphius cavirostris,
J. Acoust. Soc. Am.,112 , 34 – 37.
Gordon, J., D. Gillespie, J. Potter, A. Frantzis, A. P. Simmonds, R. Swift,
and D. Thompson (2004), A review of the effects of seismic surveys on
marine mammals, Mar. Technol. Soc. J.,37, 16– 34.
Madsen, P. T., B. Møhl, B. K. Nielsen, and M. Wahlberg (2002), Male
sperm whale behaviour during exposures to distant seismic survey pulses,
Aquat. Mammals,28, 231– 240.
Malakoff, D. A. (2001), Roaring debate over ocean noise, Science,291,
576 – 578.
Malakoff, D. (2002), Suit ties whale death to research cruise, Science,298,
722 – 723.
McCauley, R. D., J. Fewtrell, A. J. Duncan, C. Jenner, M.-N. Jenner, J. D.
Penrose, R. I. T. Prince, A. Adhitya, J. Murdoch, and K. McCabe (2000),
Marine seismic surveys: A study of environmental implications, APPEA
J.,40, 692 – 708.
Miller, P. J. O., N. Biassoni, A. Samuels, and P. Tyack (2000), Whale songs
lengthen in response to sonar, Nature,405, 903.
National Oceanic Atmospheric Administration and U. S. Navy (2001),
Bahamas marine mammal stranding event of 15 – 16 March 2000, joint
interim report, Silver Spring, Md.
National Research Council of the National Academies (2003), Ocean Noise
and Marine Mammals, 192 pp., Natl. Acad. Press, Washington, D. C.
Schmidt, V. (2004), Seismic contractors realign equipment for industry’s
needs, Offshore,64, 36– 44.
Stone, C. J. (2003), The effects of seismic activity on marine mammals in
UK waters, 1998 – 2000, Rep. 323, 78 pp., Joint Nature Conserv. Comm.,
Aberdeen, UK.
Richardson, W. J., C. R. Greene, C. I. Malme, and D. H. Thomson (1995),
Marine Mammals and Noise, Academic, San Diego, Calif.
Richardson, W. J., G. W. Miller, and C. R. Greene Jr. (1999), Displacement
of migrating bowhead whales by sounds from seismic surveys in shallow
waters of the Beaufort Sea, J. Acoust. Soc. Am.,106, 2281.
D. R. Bohnenstiehl, J. B. Diebold, E. Chapp, R. C. Holmes, M. Rawson,
M. Tolstoy, and S. C. Webb, Lamont-Doherty Earth Observatory of
Columbia University, 61 Route 9W, Palisades, NY 10964, USA.
(tolstoy@ldeo.columbia.edu)
Figure 4. Energy spectral density (solid lines) and
1/3-octave levels (lines with black dots) of a 20-gun shot
at the deep site (gray) and shallow site (black) at ranges of
2.828 km and 3.716 km respectively. Note that peak energy
occurs in the 5– 100 Hz frequency range, with levels
dropping off by about 20–40 dB from peak at 1 kHz and
continuing to drop rapidly thereafter.
L14310 TOLSTOY ET AL.: CALIBRATION OF SEISMIC SOURCES L14310
4of4