Mike Williamson (
Wed, 10 Jun 1998 08:18:23 -0400 (EDT)

Comments on low frequency playback experiments to singing humpback
whales in Hawaiian waters =96 phase III of LFA marine mammal research

Peter L. Tyack
Biology Department
Woods Hole Oceanographic Institution


There has recently been publicity and interest in a series of playback
experiments conducted with singing humpback whales off the Big Island in
the Hawaiian Islands during February and March 1998. These playback
experiments were designed to evaluate responses of humpback whales on
the breeding ground to carefully controlled exposures of sounds from a
low frequency sonar, called SURTASS LFA, developed by the U.S. Navy. I
have been involved in the design, planning and field work for these
experiments, and would like to give my views of how they fit in the
general issue of low frequency noise and marine mammals.  I am a
biologist whose primary research interests focus on acoustic
communication and social behavior in marine mammals. I am personally
deeply concerned that manmade ocean noise may disrupt the behavior and
communication of marine animals. I have worked to help identify this as
a critical conservation issue and have been a member of two committees
of the National Academy of Sciences to evaluate this problem (Green et
al 1994). I have also conducted several experiments to evaluate
responses of marine mammals to noise (e.g. Malme et al. 1983, 1984).

The growing problem of ocean noise

Addressing the impacts of ocean noise on marine mammals requires a shift
of focus from earlier conservation efforts. The greatest threats to
marine mammals several decades ago occurred when humans intentionally
targeted individual animals for whaling or to set nets on tuna.
Regulations to protect animals from these threats prohibited humans from
=93taking=94 or killing them. Marine mammals now face growing unintentional
threats from habitat degradation =96 from chemical and noise pollution to
loss of food resources. These indirect effects can threaten animal
populations by reducing rates of growth or reproduction even when they
may not kill or injure animals from acute exposure. In my opinion,
protecting marine animals from these threats will demand research to
define the risks along with a new regulatory structure to minimize the
cumulative risks from all activities rather than focusing on intentional
=93takes=94 of individual animals.

The basic problem with low frequency noise and marine mammals is that
humans are introducing an ever-growing number of ever-louder noise
sources in the sea. There is no effective regulation to protect marine
animals from noise pollution, and even if there were, we do not know
what levels of sound exposure are safe. The National Academy of Sciences
Committee on Low-Frequency Sound and Marine Mammals pointed out that
this creates an urgent need for research on what acoustic exposures pose
a risk of hearing loss and behavioral disruption. For studying
behavioral disruption, especially on baleen whales, the Committee
recommended =93planned experiments in which the received level of sound
and the behavior of the animal can be studied together.=94

The risk of behavioral disruption is an important concern for the

I was introduced to the SURTASS LFA sonar when the Navy met with
representatives of environmental groups and with scientists concerned
about the potential impact of this sonar on marine animals. The unusual
feature of this sonar is not its loudness =96 there are thousands of
sonars that are as loud or louder. Rather it is the low frequency at
which it operates -- 100-500 Hz. Sound energy at the high frequencies
typical of most sonars is absorbed by seawater, but loud low frequency
sounds can travel and be detected hundreds of kilometers away. My
greatest concern about this sonar was that if it disrupted the behavior
of animals at relatively low levels, the behavioral disruption could
occur over a large proportion of a species=92 habitat. A biologically
significant disruption of behavior involving many animals over a large
area could have a greater impact on a population than a few animals
accidentally being exposed very near the sonar source to levels loud
enough to harm them. I felt that it was critical to evaluate how animals
thought to be particularly sensitive would respond to this sonar at
received levels potentially well below those thought to pose a risk of
harm. The best way to evaluate the risk of behavioral disruption is to
conduct studies relating animal responses to carefully controlled
exposures of sounds from this sonar system.

Why we need research on the effects of sound on marine animals.

The playback experiments to humpback whales in Hawaii resulted from a
unique response by the Navy to make the LFA sonar ship available to
biologists in order to study what levels of sound exposure may evoke
behavioral disruption in marine mammals. The Navy worked with a broad
group of scientists, environmental groups, and regulators to take the
fullest advantage of the research capabilities of using this sonar
system for playback experiments to marine mammals. The sonar system
includes sophisticated methods to predict sound propagation, allowing
better control of acoustic exposure for playback subjects.

The purpose of these playback experiments is to provide data of use to
protect marine animals from adverse impacts of manmade noise. The Navy
is preparing an environmental impact statement to evaluate whether and
how the sonar can be used without biologically significant risks to
marine animals. The playback studies will add to the database used in
preparing this environmental impact statement. Many features of the
playback experiments were designed to best evaluate the impact of the
SURTASS LFA sonar system, but the data may also be useful in evaluating
impacts of other sources of ocean noise.

It is difficult to study the behavioral responses of whales at sea. Even
with the added experimental control provided by the SURTASS LFA system,
these playback experiments only have limited power to uncover some of
the responses whales might make. However, our ignorance of the impact of
this kind of sound source has been so profound, that even limited
research results are critical to informing policy decisions. A major
focus of this research program was to select the most important and
sensitive tests on behavioral effects of low frequency sound. Many
meetings were held with a broad group of interested parties in order to
select which species should be studied, to select where and when the
system would be used and to develop experimental protocols governing the
location of the ship and operation of the source. Three different
species and settings were selected for LFA playback experiments with
marine mammals. The first involved blue and fin whales feeding in the
Southern California Bight (Sept-Oct =9297); the second involved gray
whales migrating past the central California coast (Jan =9298), and the
third involved humpback whales in Hawaii.

These playback experiments themselves were subjected to extensive review
and regulation. An environmental assessment was prepared in accordance
with the National Environmental Policy Act. The National Marine
Fisheries Service reviewed the protocols and issued a scientific
research permit for these experiments under the Marine Mammal Protection
Act. The California Coastal Commission reviewed and approved the two
California phases of the research, and the State of Hawaii agreed that
the third phase was consistent with the state=92s Coastal Zone Management
Plan. Some critics suggested that these experiments required an
environmental impact statement even though the research is designed to
provide data for an environmental impact statement. This seems like a
catch 22 to me, and the extensive review and regulation for these
playbacks stand in marked contrast to most noise sources which are not
regulated by either state or federal governments.

Rationale for the playback experiments with humpback whales in Hawaii.

The SURTASS LFA sonar was designed to find submarines that may be too
silent to detect by listening for their own sounds. The playback
experiments in Hawaii faced legal challenges and were criticized for
being attempts by the Navy to target endangered whales for
anti-submarine warfare tests in a critical habitat.  From my own
involvement in this experiment, I know that this is a misrepresentation.
The source was not used as the Navy would for detecting submarines.
Rather, biologists worked with the Navy, environmental groups, and
government regulators in order to use the LFA sonar ship in playback
experiments designed to optimize detection of whale responses at the
lowest exposure level while minimizing adverse impacts. The critics were
correct regarding the sensitivity of endangered humpback whales on their
Hawaiian breeding ground, but they are mistaken about the reasons for
deciding to conduct playback experiments in such a sensitive
environment. It is not possible to study the responses of all species in
all contexts. Therefore, if the results of these studies were to be used
to establish guidelines for safe operation of this kind of sound source,
it was critical to select sensitive species in sensitive settings where
responses were most likely to be detected.

These playback experiments involve a delicate balance between testing
what sound exposures may start to disrupt behavior vs minimizing the
potential risk of biologically significant disruption to animals exposed
to the playback. Many safeguards were built into the playback
experiments to reduce the risk of adverse impact.  None of our
observations suggested that animals were harmed by these playbacks, but
we cannot rule out the possibility that these playbacks might have
exposed some animal not under observation to some risk. Any experiment
designed to assess risk must involve some small but nonzero chance of

Some critics of the Hawaii playback experiments argued against them
because of the uncertain risk to animals exposed to sounds of playback.
The most compelling reason to allow the small risk incurred by carefully
controlled playback experiments is the large risk posed by the large and
increasing number of sources of sound in the sea that are similar or
louder than the sound broadcasts used in the Hawaiian playbacks. These
sources include the propulsion noises from motorized ships, noise from
underwater oil exploration (e.g. air guns) and production (e.g. drilling
rigs), sounds from oceanographic research (e.g. ATOC), underwater
explosions, and from sonars used by the military, by the fishing
industry, and for depth sounding. We are profoundly ignorant of what
effects this low frequency noise has on marine animals, and this
ignorance makes it almost impossible to develop a conservation policy
for this issue (Green et al. 1994).  In the long run, marine animals may
receive a substantial benefit from playback experiments that can provide
data of use in developing policy to protect marine life from the adverse
impacts of manmade noise.

Selection of singing humpback whales as playback subjects

Humpbacks in Hawaii were chosen for the third phase of LFA experiments
on marine mammals because of the following reasons:

*    It is important to study the breeding season because breeding and
calving behavior is sensitive.
      Disruption of breeding behavior could have biologically
significant effects.
*    The behavior of humpbacks on the Hawaiian breeding ground is well
enough understood to
      allow interpretation of the biological significance of any
observed disruption.
*    There are good baseline data on the behavior and distribution of
humpbacks whales in Hawaiian

In addition, singing male humpback whales were selected for playback
subjects rather than females for the following reasons:

*     The frequency range of humpback song is centered in the frequency
range of the SURTASS
        LFA sonar. This means that humpback whales are likely to have
hearing that is sensitive to
        SURTASS LFA signals. The playback stimuli may also have more
disruptive effects than if
        they were outside of the frequencies of the animal's own
*      The vocal behavior and movements of singing humpbacks can be
followed continuously, even
         when a singer is not at the surface.
*      The behavior of singers is sufficiently regular and
well-documented to facilitate detection of
         behavioral responses to playback
*      Temporary responses of offshore singing males to playback were
deemed less likely to have a
         biologically significant effect than potential responses of
females with young calves.

Features of the playbacks designed to reduce risk to marine animals.

These experiments were designed to optimize our chance of detecting
behavioral responses at the lowest exposure levels while minimizing the
risk of exposure. Many protocols were used to minimize the risk that the
playbacks could have an adverse impact either on the subjects of the
experiments or on other animals that might be in the vicinity.  For
example, the SURTASS LFA ship can predict acoustic propagation for each
playback. Once a singing whale was selected as a subject for playback,
this allowed the loudness of the playback to be adjusted so that the
subject experienced a specific received level. In addition, the playback
protocols called for starting with a goal exposure level at the whale
near 125 dB, which is about the same loudness experienced by a whale
about 30 m from a singing whale or a typical fishing or whale-watch boat
while they are underway (all three tend to have source levels in the 155
dB range). This 125 dB received level is also just above the 120 dB
level at which some baleen whales have been observed to show behavioral
responses to some continuous noise sources operating at relatively low
source levels (Malme et al. 1983, 1984; Richardson et al 1989). If after
several days of playbacks, consistent responses were not observed at the
goal received level, the playback protocols called for increasing the
goal exposure level by a 10 dB increment.

The playback protocols also called for not exposing whale subjects to
sound levels above 155 dB, the average source level (with respect to 1
microPascal at 1 m) reported for humpback song (range =3D 144-174 dB;
Richardson et al. 1994).  Humpback whales are sighted within meters of
singers for longer periods of exposure than these playbacks, so this
exposure is within the normal range of loudness which humpback whales
experienced on the breeding ground before the introduction of manmade
noise. No whale subjects were exposed to levels above this, and most
whales were exposed to levels well below this maximum exposure.

Each playback consisted of the LFA sonar ship playing a 42 second sound
sequence every 6 minutes for ten times over an hour. Two different 42
second sound stimuli were alternated, one with energy between approx.
160-230 Hz and one between 260-330 Hz. These sounds were selected
because they closely matched LFA sounds used in typical LFA operations,
and because these sounds appeared similar to humpback song.

The LFA sonar ship started a playback several kilometers from the whale
subject and slowly approached the whale, playing back the sound stimuli
at a constant source level. In order to generate the desired exposure at
the expected closest point of approach, often a kilometer or so, the
ship had to broadcast the playback stimuli at a source level ranging
from 185-200 dB re 1 microPascal at 1 m. At the start of a series of
playbacks, the source would start at a source level of 155 dB. The
source level was increased by 10 dB with each successive 42 second sound
until the final goal source level was reached.

The source level and timing of the playbacks were designed so that the
exposure of whale subjects would not be much higher than the same whales
might be expected to encounter near other sound sources in their
environment such as large ships or vocalizing whales. While the playback
source levels in Hawaii were louder than the propulsion noise of most
ships in the study area, whales migrating to and from the breeding area
might encounter shipping noise in the loudness range of the playbacks.
Large commercial ships such as supertankers produce source levels while
underway in about the same 185-200 dB range as the playbacks (Richardson
et al. 1995). A tug pulling a fully loaded barge into Kawaihae Harbor in
our study area might have a source level of about 170 dB (Richardson et
al. 1995). A whale 100 m from such a tug would be exposed to a received
level of about 130 dB if sound energy propagated in all directions.
Under similar conditions, a whale at 1 km from the LFA ship operating at
a 190 dB source level would have the same 130 dB exposure. Source level
is not the only issue for sound exposure; one must also examine timing
of exposure. Noise from ships is continuous, while the playbacks only
occurred for at most 420 sec or 7 minutes every hour.

The playbacks were conducted over a period of about 30 days. No more
than three playbacks were conducted on a given day, yielding a maximum
of 21 minutes of actual sound broadcasts in 24 hours. Cumulative
exposure of individual whales was not expected to be significant for
periods of more than several days, because humpback whales off the
island of Hawaii are thought to move through the island chain, with
residence times of 4-5 days (Gabriele 1991, 1992). However, each focal
whale was photo-identified, and every effort was made to ensure that
each playback involved a different individual. The photo-identification
also allows us to test for any repeated playbacks to the same individual
or for any unexpected delayed reactions of individuals to playback.

The playbacks were designed to provide a controlled and limited exposure
for the whale subjects, but other marine mammals or potentially
vulnerable animals such as sea turtles might also be in the vicinity of
the playback vessel. In order to guard against inadvertent exposure of
animals to high levels of exposure, a team of observers watched for
animals surfacing near the LFA ship and bioacousticians on board
monitored hydrophone arrays being towed by the ship to monitor animal
vocalizations. The visual watch was posted forward and aft on both port
and starboard sides of the ship from an elevated deck with good
sighting. This watch was designed to sight and track surfacing animals
that might come within the sound field at levels near 160 dB near the
ship. The acoustic monitoring commenced at least four hours before any
playback, and the visual monitoring commenced at least 30 minutes before
any playback and continued for at least 30 minutes after each playback.

Acoustic and visual monitors were in immediate contact with personnel
running the playback. They were ready to request cessation of playback
if any animal were sighted such that it could come closer than the
maximum sound exposure level of 160 dB, or if any behavioral reactions
were observed that raised concern that the sound playback might be
having an adverse impact.  At the loudest source level of 200 dB at 1 m
from the source, sound levels were reduced to 160 dB at ranges of
approx. 100 m from the ship. In practice, this meant that sounds were
not broadcast if marine mammals or sea turtles were sighted within
several hundred meters of the ship. Several playbacks were not started
on schedule because of dolphins riding the bow of the ship or because of
a sighting of a sea turtle. Special attention was dedicated to ensuring
that any such animals were left well behind before any sounds were

Most marine animals are not visible from the surface and many
air-breathing marine animals still may not always be sighted. Few marine
species are thought to be as sensitive as whales to the playback sound
frequencies, but some species of fish, for example, are known to hear
well in the frequency range of these playbacks. In order to give animals
that were not sighted a chance to reduce their exposure to loud low
frequency sound, each series of playbacks was started at a source level
of 155 dB. The source level of successive 42 second playbacks was
increased by 10 dB steps until the desired source level for playback was

Singing humpback whales at least several miles offshore were selected
for playback subjects in Hawaii. Mother-calf pairs, which tend to be
sighted farther inshore, were not selected as playback subjects. Since
the source ship operated offshore and mother-calf pairs prefer inshore
habitat, most mother-calf pairs would have experienced sound levels from
the playbacks of < 125 dB. However, we were concerned that mother-calf
pairs might be particularly sensitive or vulnerable to disturbance, so
we established a shore station just inshore of the primary playback area
in order to monitor mother-calf pairs. The goal of these shore station
observations was to test for subtle changes in behavior in response to
the distant source, as well as to guard against responses that could be
recognized immediately. The shore station was blind to playback
condition, but was able to call the LFA ship if observers noted any
potential behavioral reactions of concern.


These playback experiments were motivated by the recognition that loud
underwater noise may cause adverse impacts on whales and other marine
life at some exposure level. Marine live cannot be protected from noise
pollution without studies that define the relationship between acoustic
exposure and behavioral disruption. The federal government has recently
formed an Interagency Coordinating Group on Acoustics in order to gather
data and develop policy on this issue. The goal of the LFA playback
experiments is to gather data useful for establishing safe operating
guidelines for the SURTASS LFA sonar system, and also to gather data
that is critical for establishing a more general policy to protect
marine life from adverse impacts of manmade noise. Given the economic
and technical forces for increasing manmade noise in the sea, I believe
that the best way to protect animals from the adverse impact of ocean
noise is to establish exposure guidelines based upon the kind of study
just completed with humpbacks off Hawaii.


Gabriele CM (1992) The behavior and residence characteristics of
reproductive classes of humpback whales (Megaptera novaeangliae) in the
Hawaiian Islands. MA Thesis, University of Hawaii.

Gabriele CM and LM Herman (1991) Residence patterns of wintering
humpback whales in the Hawaiian Islands (Abstr.). In: Ninth Biennial
Conference on the Biology of Marine Mammals, Chicago, IL, December 5-9,

Green DM, DeFerrari HA, McFadden D, Pearse JS, Popper AN, Richardson WJ,
Ridgway SH and PL Tyack (1994) Low-frequency sound and marine mammals:
current knowledge and research needs. National Academy Press, Washington

Malme CI, PR Miles, CW Clark, P Tyack and JE Bird (1984) Investigations
of the potential effects of underwater noise from petroleum industry
activities on migrating gray whale behavior. Phase II: January 1984
migration.  Bolt Beranek and Newman Report No. 5586 submitted to
Minerals Management Service, U. S. Dept. of the Interior.

Malme CI, PR Miles, CW Clark, P Tyack and JE Bird (1983) Investigations
of the potential effects of underwater noise from petroleum industry
activities on migrating gray whale behavior.  Bolt Beranek and Newman
Report No. 5366 submitted to Minerals Management Service, U. S. Dept. of
the Interior.

Richardson WJ, B W=FCrsig, and CR Greene Jr (1990) Reactions of bowhead
whales, Balaena mysticetus, to drilling and dredging noise in the
Canadian Beaufort Sea. Mar. Environ. Res. 29, 135-160.

Richardson WJ, Greene CR Jr, Malme CI, Thomson DH (1996) Marine mammals
and noise. Academic Press, New York.

-To submit a message to MARMAM, send it to:
-Please include your name and e-mail address in the body of the
text of all submissions, and ensure your message has an appropriate
subject heading (ie., not "Message for MARMAM")
-To subscribe to MARMAM, send a message to:
saying: subscribe marmam Yourfirstname Yourlastname
-To contact the MARMAM editors, write to: