Will Harbour porpoises (Phocoena phocoena) habituate to pingers?

Abstract

Large bycatches of harbour porpoises (Phocoena phocoena) occur in gillnet fisheries throughout the Northern Hemisphere. Several mitigation measures, including acoustic deterrent devices or 'pingers', have been used in efforts to reduce this bycatch. The potential exists for harbour porpoises to habituate to pingers, thus reducing their effectiveness over time. A field experiment was conducted to test the hypothesis that porpoises habituate to the sound produced by pingers. Porpoise echolocation and movements were monitored around a mooring equipped with a pinger (Dukane NetMark™ 1000) for three months in summer 1998 in the Bay of Fundy. Using a mean-shift model it was estimated that porpoises were initially displaced 208m from the pinger (p = 0.019), but this displacement diminished by 50% within four days (p = 0.019). Using a probability model it was demonstrated that the probability of porpoises within 125m of the pinger initially decreased when the pinger was turned on, but then increased to equal the control in 10-11 days. Echolocation rate (p < 0.001) and occurrence (p < 0.001) were significantly reduced in the vicinity of the pinger. These results indicate that porpoises habituated to the Dukane NetMark™ 1000 pinger and are not alerted to echolocate in the presence of nets by pingers.

Full text

Will harbour porpoises (Phocoena phocoena) habituate to
pingers?
Tara M. Cox
*
, Andrew J. Read
*
, Andrew Solow
+
And Nick Tregenza
#
Contact e-mail: tara.cox@duke.edu
ABSTRACT
Large bycatches of harbour porpoises (Phocoena phocoena) occur in gillnet fisheries throughout the Northern Hemisphere. Several
mitigation measures, including acoustic deterrent devices or ‘pingers’, have been used in efforts to reduce this bycatch. The potential exists
for harbour porpoises to habituate to pingers, thus reducing their effectiveness over time. A field experiment was conducted to test the
hypothesis that porpoises habituate to the sound produced by pingers. Porpoise echolocation and movements were monitored around a
mooring equipped with a pinger (Dukane NetMark™ 1000) for three months in summer 1998 in the Bay of Fundy. Using a mean-shift
model it was estimated that porpoises were initially displaced 208m from the pinger (p=0.019), but this displacement diminished by 50%
within four days (p=0.019). Using a probability model it was demonstrated that the probability of porpoises within 125m of the pinger
initially decreased when the pinger was turned on, but then increased to equal the control in 10-11 days. Echolocation rate (p<0.001) and
occurrence (p< 0.001) were significantly reduced in the vicinity of the pinger. These results indicate that porpoises habituated to the Dukane
NetMark™ 1000 pinger and are not alerted to echolocate in the presence of nets by pingers.
KEYWORDS: BEHAVIOUR; ECHOLOCATION; FISHERIES; GILLNETS; INCIDENTAL CAPTURE; NOISE; BYCATCH
INTRODUCTION
Large numbers of dolphins and porpoises die in gillnets
worldwide, posing serious threats to several populations and
species (Jefferson and Curry, 1994; Perrin et al., 1994).
Acoustic alarms or ‘pingers’ are currently used in several
fisheries to reduce these bycatches (Kraus et al., 1997;
Cameron, 1998; Trippel et al., 1999; Gearin et al., 2000). As
the use of pingers spreads, concerns have been raised about
their long-term effectiveness (Dawson et al., 1998). This
issue of acoustic alarms has recently been reviewed by the
Scientific Committee of the International Whaling
Commission (IWC, 2000).
One of the most intensive efforts to reduce small cetacean
bycatch has occurred in the Gulf of Maine. Between 1992
and 1996, an average of 2,100 harbour porpoises (Phocoena
phocoena) died annually in Gulf of Maine sink gillnets -
approximately 4% of the estimated population of 54,300, a
rate that greatly exceeded allowable removal levels set under
USA legislation (Waring et al., 1999). Kraus et al. (1997)
demonstrated that pingers caused a significant reduction in
the bycatch rate of harbour porpoises in the Gulf of Maine.
Fishermen have taken an active role in the development and
testing of pingers and are supportive of their widespread use
in this fishery. Consequently, the use of pingers was
recommended as an integral component of the management
plan designed to reduce incidental mortality to sustainable
levels (Federal Register, 1998).
In addition to recommending the use of pingers in the Gulf
of Maine, the management plan recommended that research
be conducted on several aspects of their use, including the
potential for habituation. Habituation is defined as ‘the
relatively permanent waning of a response as a result of
repeated stimulation which is not followed by any kind of
reinforcement’ (Thorpe, 1966). Participants at a workshop
sponsored by the US National Marine Fisheries Service and
the Marine Mammal Commission also noted the possibility
that the effectiveness of pingers could decline due to
habituation (Reeves et al., 1996). As more and more pingers
are used in the Gulf of Maine, the avoidance response of
harbour porpoises to these pingers could wane, reducing the
efficacy of this management tool.
The purpose of this study was to evaluate the potential for
porpoises to habituate to pingers. This experiment,
conducted in the summer of 1998, forms part of a larger
research programme designed to address the question of
habituation. Another important aspect of this overall
programme is to monitor the observed bycatch rate of
porpoises over time in areas where pingers are used, to
determine whether or not habituation is occurring. In the
field experiment described here, a technique similar to that
employed by Koschinski and Culik (1997) is used, in which
shore-based observers used a theodolite, or surveyor’s
transit, to track the movements of porpoises in the vicinity of
active pingers. In a study of six days duration, Koschinski
and Culik noted that porpoises avoided an experimental net
equipped with pingers. Similar findings have been reported
by Kastelein et al. (1997) for porpoises in a captive setting.
In this study, patterns of harbour porpoises were monitored
in relation to pingers over longer periods to assess the
potential for habituation.
METHODS
Study area and experimental design
Porpoises were observed from a cliff on Grand Manan
Island, New Brunswick, Canada between 26 June 1998 and
14 September 1998. This area has a high density of harbour
porpoises during the summer months (Waring et al., 1999).
A single Dukane NetMark™ 1000 pinger was attached 10m
below the surface to a mooring at 44°47.7’N, 66°48.2’W
(Fig. 1). The mooring was approximately 1,000m offshore
and was set in 75m of water. The Dukane NetMark™ pinger
emits a regular interval pulsed, broad-band signal with a
fundamental frequency of 10kHz and a minimum sound
pressure level of 132dB re 1mPa at 1m, which meets the
*
Duke University Marine Laboratory, Beaufort, NC 28516 USA.
+
Marine Policy Center, Woods Hole Oceanographic Institution, Woods Hole, MA 02543 USA.
#
Cornwall Wildlife Trust, Penzance, Cornwall TR20 8JE, UK.
J. CETACEAN RES. MANAGE. 3(1):81–86, 2001 81

7DEOH
7LPLQJRIKDELWXDWLRQWULDOV7ULDOZDVWHUPLQDWHGHDUO\GXHWRSRRU
ZHDWKHUFRQGLWLRQV
 %HJLQ (QG 3LQJHU 32'
7UDLQLQJ -XQ -XQ 2II 
7ULDO 

&RQWURO -XQ -XO 2II 
([SHULPHQWDO -XO $XJ 2Q 
7ULDO

&RQWURO $XJ 6HS 2II 2Q
([SHULPHQWDO 6HS 6HS 2Q 2Q
current regulatory specification for a pinger in the Gulf of
Maine (Federal Register, 1998). During an initial two-week
training period, porpoises were tracked by the study team to
improve proficiency in the use of the theodolite (see below).
Porpoises were then tracked for two weeks around the
mooring while the pinger was attached but not turned on -
Control 1 (Table 1). On 11 July, the pinger was turned on and
porpoises were tracked for four weeks - Experimental Trial
1. On 7 August, the pinger was turned off, and tracking
began again on 19 August - Control 2. At this time a porpoise
echolocation detector, POD, was also attached (see below).
On 2 September the pinger was turned back on, and tracking
continued for four weeks - Experimental Trial 2.
The sound pressure level and frequency of the Dukane
pinger decrease with decay in battery voltage (Trippel et al.,
1999), so the pinger batteries were changed once a week and
the voltage of each battery was tested after it was
removed.
Tracking
Two researchers tracked porpoises using a Geodolite 404
total station and a Husky FS/GS data collector from a 100m
cliff approximately 1,000m from the mooring. The
observational area encompassed a 500m radius around the
mooring. One researcher, the surveyor, used Fujinon 7 3 50
binoculars to scan the observational area for porpoises. The
surveyor looked in concentric circles around the mooring,
extending out to 500m. This individual reported sightings of
porpoises to the tracker, the researcher stationed at the
theodolite. The tracker used the theodolite to track
surfacings of the lead porpoise in a group until: (1) the
animals left the study area; or (2) the tracker lost sight of the
porpoises or could not confirm that it was the same group.
The tracker then began tracking the next group of porpoises
identified by the surveyor. During the training period, the
researchers tested their ability to estimate 500m from the
mooring in all directions. The theodolite was used to
measure the distances and ground-truth the estimates. After
two weeks of training, both researchers were able to estimate
500m to within 10m in all directions.
Echolocation
On 20 August 1998 a POD was attached to the mooring. The
POD continuously logged the number of echolocation clicks
in 10s intervals. The POD was programmed to record several
Fig. 1. Study area at Grand Manan Island, New Brunswick, Canada. The star represents the position of the
pinger mooring. Track of satellite tagged animal from 6 August 1998 to 16 September 1998 (Westgate
and Read, unpublished data). Individual points represent best position per day.
COX et al.: WILL HARBOUR PORPOISES HABITUATE TO PINGERS?82

channels of echolocation clicks of varying duration and
frequency. The frequencies were fixed at 50kHz, 93kHz and
132kHz. Porpoises produce distinctive narrow band sonar
clicks from 110-150kHz (Mohl and Andersen, 1973;
Kamminga and Wiersma, 1981) and so only clicks at
132kHz were used in the analysis. Single click durations for
harbour porpoises are typically 100ms (Mohl and Andersen,
1973), so the POD was programmed to capture any click that
lasted up to 400ms in duration.
Response variables
From the results of previous studies, a change in porpoise
behaviour was expected when the pinger was first activated.
Then, if habituation occurred, a gradual waning of this
response was expected to occur over the experimental
period. Three variables that have direct relevance to
entanglement were examined: the point of closest approach
to the pinger, echolocation rate and echolocation
occurrence. The point of closest approach was defined as the
minimum distance between the pinger and a surfacing
porpoise. Echolocation rate was the number of clicks
recorded per unit time and echolocation occurrence was the
proportion of 10 second intervals in which clicks were
detected.
Sound field
The sound field radiated by the pinger was measured on 26
September 1998. The day was overcast and the Beaufort Sea
State was 2, diminishing to 1. Researchers drifted past the
mooring in a small boat while the position of the boat was
recorded from shore using the theodolite. The observers in
the boat monitored the sound produced by the pinger with a
Bruel and Kjaer 8100 calibrated hydrophone and a 2635
charged coupled pre-amplifier, which included a reference
signal (160Hz, 174dB re1mPa@1m) generator. The
weighted hydrophone was deployed 10m below the drifting
boat. The calibration signal and hydrophone signal were
recorded on a Sony TCD-D8 DAT recorder. Using
Syntrillium Software Corporation’s Cool Edit Pro version
1.1, the recordings were uploaded using 16-bit, single-track
settings. A power spectrum (FFT 1,024 points;
Blackman-Harris window) was then created to estimate the
sound pressure level of the pinger in relation to the reference
signal. By comparing the relative decibel level of the pinger
to the known decibel level of the reference signal, the
absolute decibel level of the pinger could be calculated.
Analysis
Two models were used to examine the data. First, a
mean-shift model was used to test the hypothesis that
porpoises were initially displaced from the pinger and then
gradually moved closer to the pinger:
E(Y
j
) = b
0
+ b
1
(–b
2
(t
j
–t
a
))
I
t
a
(t
j
)
where:
E(Y
j
) is the expected distance of closest approach for group
j (j = 1, 2, 3, …., n)
b
0
, b
1
, and b
2
are unknown parameters:
b
0
is the control mean
b
1
is the mean shift due to the pinger
b
2
is the rate at which the pinger effect decays to 0
t
j
is the day on which group j was observed
t
a
is the day the pinger was turned on
I
ta
(t
j
) = 1 if t
j
> t
a
, otherwise I
ta
(t
j
) = 0.
Under this model, mean distance (b
0
) is constant prior to
activation. Following activation, there is an immediate
increase (b
1
) in the mean distance. This increase declines
with time at rate b
2
. The time after t
a
at which the mean shift
has been reduced by 50% can then be defined as:
T
50
= –log 0.5 / b
2
To test whether there was an initial response when the pinger
was turned on, the null hypothesis H
0
: b
1
= 0 was tested
against the one-sided alternative hypothesis H
1
: b
1
> 0.
Using a randomisation test (Manly, 1991) samples were
generated under H
0
by randomising the assignment of the
observed values of Y
j
to observation dates and fitting the null
model by least squares. The significance level was estimated
by the proportion of randomised datasets for which the
residual sum of squares was less than that from fitting the
model to the non-randomised data.
To test whether there was a significant waning of response
over time, the null hypothesis H
0
: b
2
= 0 was tested against
the one-sided alternative H
1
: b
2
> 0. In this case, only those
values of Y
j
for which t
j
≥ t
a
were permuted.
Porpoises can probably not detect the pinger out to 500m
(Kraus et al., 1997) and so a second model was used to test
the probability that the proportion of sightings within 125m
changed over time in response to the pinger. The distance of
125m was chosen based on sound field analysis (see below)
and Laake et al.’s (1998) published displacement distance.
The general model is:
Prob(X
j
= 1) = p
0
0 ≤ t
j
< t
a
p
1
t
a
≤ t
j
< q
p
2
t
j
≥ q
where t
a
is the known day of activation of the pinger and q
corresponds to the beginning of the habituation period. The
binary random variable was chosen as X
j
= 1 if the closest
approach of group j is within 125m of the pinger and 0
otherwise, and t
j
was chosen as the day on which this group
was observed. Under this model the probability of a sighting
within 125m prior to activation (p
0
) is constant. Following
activation this probability falls to p
1
. However, beginning on
day q, this probability rises to p
2
.
Each of the following null hypotheses was tested against
the general model:
Hp p p
H
Hp p
0
1
012
0
2
0
3
02
:
:
:
==
=•
=
q
Under H
1
0
, there is no effect of the pinger on proportion of
porpoises within 125m. Under H
2
0
there is an effect, but no
habituation. Under H
3
0
, there is full habituation. In each case
the likelihood ratio statistic was used. In testing H
1
0
and H
2
0
against the general model, randomisation tests were used.
The former case involved randomising the full set of
observed distances; the latter case involved randomising
only the post-activation distances.
A univariate factorial analysis of variance was used to
examine variation in echolocation rate as a function of the
state of the pinger (on or off) and time of day. Day was
defined as occurring between 07:00 and 18:59 and night
occurred between 19:00 and 06:59 (Westgate et al., 1995). A
Chi-squared test was also used to compare the proportion of
10s intervals in which echolocation clicks occurred when the
pinger was off and on. Means are presented with their
associated standard deviations.
RESULTS
The closest observed approach of the porpoises to the active
pinger decreased over time (Fig. 2). Poor weather forced
truncation of the second trial. Thus the sample size was
J. CETACEAN RES. MANAGE. 3(1):81–86, 2001 83








     
'D\VDIWHUSLQJHURQ
&ORVHVWREVHUYHGDSSURDFKP
7DEOH
5HVXOWVRIPHDQVKLIWPRGHO6LJQLILFDQFHYDOXHVDUHLQSDUHQWKHVHVDIWHUWKHWHVWSDUDPHWHUV

&RQWURO
3

([SHULPHQW
3

&RQWUROPHDQ

-
4

0HDQVKLIW

-


5DWHRIGHFD\

-


7LPHWRGHFD\

%


7ULDO   P P  GD\V
7ULDO   P P  GD\V
3RROHG   P P  GD\V
small, so both experiments were pooled to increase the
power and test the mean-fit model. Results of the mean-shift
model for both experimental trials and the pooled trials are
presented in Table 2.
For the probability model all three null hypotheses were
tested for Trial 1. Trial 2 was truncated, so only an effect of
the pinger was tested. In Trial 1, the null hypotheses of no
effect (p=0.02) and no habituation (p=0.0) were rejected;
however, the null hypothesis of full habituation (p=0.39)
was not rejected. The maximum-likelihood estimates of the
parameters under full habituation were:
ˆˆ
.
ˆ
.
ˆ
p p p days
02 1
0 30 0 00 10 11== = =-
q
It is not possible to distinguish between habituation times of
10 or 11 days because no observations were made on Day 11
following activation. For Trial 2, there was no significant
difference between the p
0
(0.19) and p
1
(0.08) (p=0.12).
A time of 30 minutes was chosen for the analysis of
echolocation rate because only one group of porpoises
remained in the area for more than this period (31 minutes).
Therefore, independence among measurements of the
number of echolocation clicks per half hour was assumed.
Echolocation detection rate for the control (516±2,062;
n=288) was significantly greater than when the pinger was
active (82±366; n=496) (p<0.001). In addition,
echolocation detection rate was higher at night (377±1,699;
n=432) than in the day (75±409; n=352) (p<0.001) for
both control and active periods. The proportion of 10 second
intervals in which clicks were detected decreased after the
pinger had been activated (control=0.174;
experimental=0.041) (c
2
=9,241; p < 0.001).
Received sound pressure levels of the 10kHz signal
became indistinguishable from background noise at
approximately 125m from the pinger (Fig. 3). Battery
voltages averaged 5.85V ± 0.06V when removed from the
pinger.
DISCUSSION
Habituation
The analysis suggests that porpoises habituated to the
presence of the pinger. Both models indicated an initial
response and then a waning of that response. When the two
trials were pooled in the mean-shift model, porpoises were
initially displaced from the pinger, but this displacement
waned over time. Despite a small sample size and truncation
of the second trial, similar patterns were observed in each of
the two individual trials in the mean-shift model. In addition,
the probability that porpoises approached within 125m of the
pinger initially decreased, then increased after 10-11 days.
Thus, porpoises habituated to the pinger and approached it
more closely over time.
Demonstration of habituation typically relies on repeated
observations of known individuals (Richardson et al., 1995).
Fig. 2. Closest observed approach for Trials 1 and 2 pooled. (-) = Control, Trial 1; (/) = Experimental, Trial 1;
(8) = Control, Trial 2; (◊) = Experimental, Trial 2.
COX et al.: WILL HARBOUR PORPOISES HABITUATE TO PINGERS?84

100
110
120
130
140
150
0 50 100 150 200 250 300
Distance (m)
dB
It was not possible to identify individual porpoises as their
movements were tracked with the theodolite. However,
previous studies of the movements of porpoises in the Grand
Manan area using satellite and VHF telemetry have shown
that individual animals are present in particular areas for
weeks or months (Read and Westgate, 1997). For example,
a porpoise tagged with a satellite-linked radio transmitter
was tracked around the mooring on 1 September 1998. This
porpoise had been in the area for several weeks (Fig. 1).
Thus, individual porpoises likely experienced multiple
exposures to the pinger over the course of the experiment.
The estimate using the mean-shift model of an initial
displacement of 208m is larger than the 125m displacement
used for the probability model. In addition, the time to
response decay in the mean-shift model is considerably
faster than the estimate of 10-11 days using the probability
model. However, the precision of estimates of initial
displacement and rate of decay in the mean-shift model is
relatively low, since few observations were made
immediately after the pinger had been activated. The results
show a relatively large initial displacement, followed by a
relatively rapid habituation. However, no observations were
made immediately following activation of the pinger, so the
possibility of a smaller initial displacement and a longer
period of habituation can not be ruled out. Nevertheless, this
imprecision does not affect the conclusions that porpoises
were initially displaced by the pinger and then approached it
more closely over time.
The experimental protocol involved only a single pinger
on a mooring, so it is not possible to say with certainty that
porpoises will habituate to pingers attached to a gillnet. In
fact, even if habituation occurs, it may not lead to an increase
in bycatch rate if there is enough residual effect to keep
porpoises away from nets. Another plausible scenario is one
in which as porpoises habituate and approach the pinger
more closely, the sound may stimulate them to investigate
their surroundings more thoroughly and thus avoid the net.
This scenario assumes that the porpoise will perceive the net
as a barrier or danger. Ultimately, a monitoring programme
is necessary to ensure bycatches do not increase as porpoises
habituate to pingers used in a commercial setting (IWC,
2000).
The decrease in battery voltage would not have resulted in
a significant decay in frequency or amplitude of the pinger
(Trippel et al., 1999). However, it is unlikely that fishermen
will replace batteries every week. Thus, battery decay and
resultant changes in pinger function could lead to a decay in
their effectiveness in a commercial fishery. Pingers are
currently being developed that regulate the voltage supply so
that frequency and sound pressure do not decay with falling
battery voltage (A.D. Goodson, pers. comm.).
The experimental protocol also only involved a single
type of pinger - the Dukane NetMark™ 1000. Other types of
pingers with different sound characteristics, including
frequency sweeps as opposed to tonal pulses, varied
inter-pulse intervals and randomised frequency over time,
are currently being developed (A.D. Goodson, pers. comm.).
As these pingers become commercially available, they
should be thoroughly tested to determine if they will reduce
the likelihood of habituation.
Echolocation
Elucidating the mechanism by which pingers work will
further aid in determining if porpoises will habituate to
pingers on gillnets (see below). For example, if the sound of
pingers is aversive to porpoises, they are likely to habituate
to it. However, if pingers alert porpoises to the presence of a
barrier which they perceive as dangerous, they may be less
likely to habituate.
Kraus et al. (1997) hypothesised that pingers might
stimulate porpoises to echolocate and thus detect a gillnet.
This hypothesis was tested here by examining echolocation
rates of porpoises in relation to the moored pinger. The
reduction in echolocation rate (number of clicks per unit
time) when the pinger was activated demonstrated that
porpoises were either echolocating less frequently in the
vicinity of the pinger, using shorter click trains, or directing
their sonar away from the pinger. If porpoises emitted a
Fig. 3. Sound power spectrum level (dB re 1mPa) versus distance from the pinger. (5)= Drift 1; (U) = Drift 2;
(D)= Drift 3. At approximately 125m, the 10kHz peak in the power spectrum became indistinguishable from
background noise.
J. CETACEAN RES. MANAGE. 3(1):81–86, 2001 85

similar number of shorter trains, the proportion of 10s
intervals containing clicks would be expected to be similar in
control and experimental treatments. However, the
proportion of 10s intervals in which echolocation events
occurred was significantly reduced when the pinger was
activated, suggesting that porpoises echolocate less
frequently in the vicinity of an active pinger.
It is possible, and perhaps likely, that many porpoises
were displaced from the pinger and the POD did not detect
their echolocation signals. Preliminary studies estimate the
range of the POD to be 50-100m (T. Cox, unpublished data).
This distance is considerably greater than the distance
(2-9m) at which porpoises can detect nets with floatlines
using echolocation (Hatakeyama and Soeda, 1990). None of
these explanations supports the hypothesis of Kraus et al.
(1997) that the Dukane NetMark™ 1000 pinger stimulates
porpoises to echolocate, as the echolocation frequency of
porpoises around the pinger did not increase when the device
was activated.
Even during the control period, echolocation clicks were
recorded only 17% of the time. Porpoises were tracked
around the mooring at this time, and three times porpoises
were oriented towards the mooring within 50m of the pinger,
but no echolocation clicks were recorded. Thus, it is likely
that porpoises are not echolocating constantly. This finding
has relevance for the development of other acoustic means of
reducing bycatch, particularly those which rely on a passive
approach.
Because Trial 2 was truncated due to poor weather
conditions, changes in echolocation response to the pinger
over time were not monitored. Future studies should monitor
echolocation rate and frequency as additional response
variables that could wane over time. Investigating these
responses over time would further elucidate the potential for
porpoises to habituate to the presence of a pinger.
CONCLUSION
The results suggest that the effects of habituation need to be
considered when pingers are used to reduce the bycatch of
small cetaceans. Long-term monitoring of bycatch using
observers is necessary to ensure the effectiveness of pingers
in gillnet fisheries (IWC, 2000). This study was not designed
to test hypotheses of the mechanism by which pingers reduce
harbour porpoise bycatch, but was able to reject the
hypothesis that pingers stimulate harbour porpoises to
echolocate and thus detect a gillnet. Monitoring harbour
porpoise echolocation around gillnets equipped with pingers
could further elucidate the mechanism by which pingers
reduce bycatch.
ACKNOWLEDGEMENTS
This research was conducted at the Grand Manan Whale and
Seabird Research Station. Special thanks go to Jeremy
Rusin, Andrew Westgate, Dave Johnston, Heather
Koopman, Krystal Tolley, Rob Ronconi and Sarah Wong for
assistance in the field. Additional assistance was provided by
the fishermen of Grand Manan, especially Jeff Foster and
Steven Bass. Equipment and aid for mapping the sound field
were provided by Dr Jack Terhune and Dave Johnston. The
experimental design was improved by comments from Jay
Barlow and the US Marine Mammal Commission and its
Committee of Scientific Advisors. We thank Dave Johnston,
Finn Larsen and Dave Goodson for their thoughtful reviews
of this manuscript. This project was funded by the US
National Marine Fisheries Service, Northeast Fisheries
Science Center under Co-operative Agreement
NA77FL0373.
REFERENCES
Cameron, G. 1998. Cetacean mortality in California gillnet fisheries:
Preliminary estimates for 1997. Paper SC/50/SM2 presented to the
IWC Scientific Committee, April 1998 (unpublished). 15pp. [Paper
available from the Office of this Journal].
Dawson, S.M., Read, A. and Slooten, E. 1998. Pingers, porpoises and
power: uncertainties with using pingers to reduce bycatch of small
cetaceans. Biol. Conserv. 84:141-6.
Federal Register. 1998. Taking of marine mammals incidental to
commercial fishing operations; Harbour porpoise take reduction plan
regulations. Department of Commerce, National Oceanic and
Atmospheric Administration 50CFR Part 229. Federal Register
63(231):66,464-490.
Gearin, P.J., Gosho, M.E., Laake, J., Cooke, L., Delong, R.L. and
Hughes, K.M. 2000. Experimental testing of acoustic alarms
(pingers) to reduce bycatch of harbour porpoise, Phocoena
phocoena, in the state of Washington. J. Cetacean Res. Manage.
2(1):1-10.
Hatakeyama, Y. and Soeda, H. 1990. Studies of echolocation of
porpoises taken in salmon gillnet fisheries. pp. 269-82. In: J.A.
Thomas and R.A. Kastelein (eds.) Sensory Abilities of Cetaceans:
Laboratory and Field Evidence. Plenum Press, New York. 702pp.
International Whaling Commission. 2000. Report of the Scientific
Committee. Annex I. Report of the Sub-Committee on Small
Cetaceans. J. Cetacean Res. Manage. 2 (Suppl.):235-57.
Jefferson, T.A. and Curry, B.E. 1994. A global review of porpoise
(Cetacea: Phocoenidae) mortality in gillnets. Biol. Conserv.
67(2):167-83.
Kamminga, C. and Wiersma, H. 1981. Investigations of cetacean sonar
II. Acoustical similarities and differences in odontocete sonar
signals. Aquat. Mamm. 8(2):41-62.
Kastelein, R.A., de Haan, D., Goodson, A.D., Staal, C. and Vaughan, N.
1997. The effects of various sounds on a harbour porpoise (Phocoena
phocoena). pp. 367-83. In: A.J. Read, P.R. Wiepkema and P.E.
Nachtigall (eds.) The Biology of the Harbour Porpoise. De Spil
Publishers, Woerden, The Netherlands. 710pp.
Koschinski, S. and Culik, B. 1997. Deterring harbour porpoises
(Phocoena phocoena) from gillnets: observed reactions to passive
reflectors and pingers. Rep. int. Whal. Commn 47:659-68.
Kraus, S.D., Read, A.J., Solow, A., Baldwin, K., Spradlin, T.,
Anderson, E. and Williamson, J. 1997. Acoustic alarms reduce
porpoise mortality. Nature, Lond. 388:525.
Laake, J., Rugh, D. and Baraff, L. 1998. Observations of harbor
porpoise in the vicinity of acoustic alarms on a set gill net. NOAA
Tech. Mem. NMFS-AFSC- 84:40.
Manly, B.F.J. 1991. Randomization and Monte Carlo Tests in Biology.
Chapman and Hall, London. 281pp.
Mohl, B. and Andersen, S. 1973. Echolocation: high-frequency
component in the click of the harbour porpoise (Phocoena phocoena
L.). J. Acoust. Soc. Am. 54:1,368-72.
Perrin, W.F., Donovan, G.P. and Barlow, J. (eds.). 1994. Report of the
International Whaling Commission (Special Issue 15). Gillnets and
Cetaceans. International Whaling Commission, Cambridge, UK.
629pp.
Read, A.J. and Westgate, A.J. 1997. Monitoring the movements of
harbour porpoises (Phocoena phocoena) with satellite telemetry.
Mar. Biol. 130:315-22.
Reeves, R.R., Hofman, R.J., Silber, G.K. and Wilkinson, D. 1996.
Acoustic deterrence of harmful marine mammal-fishery interactions:
Proceedings of a workshop held in Seattle, Washington, 20-22
March 1996. NOAA Tech. Memo. NMFS-OPR-10. 70pp.
Richardson, W.J., Greene Jr, C.R., Malme, C.I. and Thomson, D.H.
1995. Marine Mammals and Noise. Academic Press, San Diego.
576pp.
Thorpe, W.H. 1966. Learning and Instinct in Animals. Harvard
University Press, Cambridge, MA. 558pp.
Trippel, E.A., Strong, M.B., Terhune, J.M. and Conway, J.D. 1999.
Mitigation of harbour porpoise (Phocoena phocoena) by-catch in the
gillnet fishery in the lower Bay of Fundy. Can. J. Fish. Aquat. Sci.
56:113-23.
Waring, G.T., Palka, D.L., Clapham, P.J., Swartz, S., Rossman, M.C.,
Cole, T.V.N., Bisack, K.D. and Hansen, L.J. 1999. US Atlantic
marine mammal stock assessments - 1998. NOAA Technical
Memorandum NMFS-NE 116:151-9.
Westgate, A., Read, A., Berggren, P., Koopman, H. and Gaskin, D.
1995. Diving behaviour of harbour porpoises (Phocoena phocoena)
in the Bay of Fundy. Can. J. Fish. Aquat. Sci. 52(5):1064-73.
COX et al.: WILL HARBOUR PORPOISES HABITUATE TO PINGERS?86