Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
ENDANGERED SPECIES RESEARCHEndang Species Res
Vol. 19: 201–221, 2013doi: 10.3354/esr00464
Published online January 25
© Inter-Research 2013 · www.int-res.com*Email: [email protected]
REVIEW
To ping or not to ping: the use of active acousticdevices in mitigating interactions between small
cetaceans and gillnet fisheries
Stephen M. Dawson1,* Simon Northridge2, Danielle Waples3, Andrew J Read3
1Department of Marine Science, University of Otago, 310 Castle Street, Dunedin, New Zealand2Sea Mammal Research Unit, University of St Andrews, St Andrews, Fife KY16 8LB, UK
3Division of Marine Science and Conservation, Nicholas School of the Environment, Duke University, 135 Duke Marine Lab Road, Beaufort, North Carolina 28516, USA
ABSTRACT: Active sound emitters (‘pingers’) are used in several gillnet fisheries to reducebycatch of small cetaceans, and/or to reduce depredation by dolphins. Here, we review studiesconducted to determine how effective these devices may be as management tools. Significantreductions in bycatch of harbour porpoise Phocoena phocoena, franciscana Pontoporia blainvillei,common Delphinus delphis and striped dolphin Stenella coeruleoalba, and beaked whales as agroup have been demonstrated. For harbour porpoise this result has been replicated in 14 con-trolled experiments in North America and Europe, and appears to be due to porpoises avoidingthe area ensonified by pingers. Two gillnet fisheries (California-Oregon driftnet fishery for sword-fish; New England groundfish fishery) with mandatory pinger use have been studied for over adecade. Bycatch rates of dolphins/porpoises have fallen by 50 to 60%, and there is no evidence ofbycatch increasing over time due to habituation. In both fisheries, bycatch rates were significantlyhigher in nets sparsely equipped with pingers or in which pingers had failed, than in nets withoutany pingers at all. Studies of pinger use to reduce depredation by bottlenose dolphins Tursiopstruncatus generally show small and inconsistent improvements in fish catches and somewhatreduced net damage. Dolphin bycatch in these fisheries is rare, but still occurs in nets withpingers. Taken together, these studies suggest that the most promising candidates for bycatchreduction via pinger use will be gillnet fisheries in developed countries in which the bycaughtcetaceans are generally neophobic species with large home ranges. We offer a set of lessonslearned from the last decade of bycatch management.
KEY WORDS: Gillnet · Bycatch · Dolphin · Porpoise · Pinger · Acoustic devices
Resale or republication not permitted without written consent of the publisher
Contribution to the Theme Section ‘Techniques for reducing bycatch of marine mammals in gillnets’ OPENPEN ACCESSCCESS
INTRODUCTION
Three fundamentally different approaches exist tomitigate deleterious interactions between humansand wildlife. The first is to mandate a change inhuman behaviour or to create a climate in which suchchanges are made voluntarily. The second is to mod-ify the nature of the interaction by the introduction of
technology. The third, and most challenging, is tochange the behaviour of the animals themselves,without requiring commensurate changes in humanbehaviour. For example, consider the problem of col-lisions between birds and aircraft, also known as birdstrikes. The first approach might entail requiring achange to flight paths or landing times, such as land-ing at night when many birds are inactive. The sec-
Endang Species Res 19: 201–221, 2013
ond could involve modifications to the intake of air-craft engines to reduce the likelihood of failure incase of a bird strike. The last option would attempt tomodify the behaviour or distribution of the birdsthemselves through harassment or other means toreduce the likelihood of a collision.
In this paper, we explore an example of the thirdapproach described above: the use of acousticpingers to modify the behaviour of dolphins, por-poises and small whales to reduce the frequency oftheir interactions with gillnet1 fisheries. The inciden-tal mortality of small cetaceans in such fisheries is themost pressing threat to the global conservation ofthese marine mammals (Brownell et al. 1989, Read etal. 2006). All 3 of the approaches described abovehave been used to address this conservation prob-lem, including: changing human behaviour throughtime−area fishery closures (e.g. Dawson & Slooten1993, Murray et al. 2000), technological modifica-tions to the fishing gear (e.g. Hembree & Harwood1987, Trippel et al. 2003) and the use of active soundemitters or acoustic ‘alarms’ (here referred to aspingers) to reduce the likelihood of entanglement bymodifying the behaviour of the animals themselves(e.g. Kraus et al. 1997). These approaches are some-times used in com bination (e.g. area closures andpingers, Palka et al. 2008; modification to float linedepth and pingers, Barlow & Cameron 2003). Never-theless, the use of pingers is attractive to fishermen(under certain conditions) because it does not requiresubstantial changes to fishing behaviour or gear and,thus, is perceived as less costly than alternativeapproaches.
Active sound emitters have been used primarily in2 contexts with marine mammals: (1) to reduce inci-dental mortality of dolphins and porpoises in fishinggear, known as bycatch; and (2) to reduce the eco-nomic cost of dolphins removing or damaging caughtfish, known as depredation. Previous reviews (e.g.Reeves et al. 2001) considered 2 broad categories ofsound emitters, the first to address the problem ofbycatch and the second to mitigate depredation.Devices in the first category emit relatively low inten-sity sounds (<150 dB re 1 µPa at 1 m)2; these areknown as pingers or acoustic deterrent devices(ADDs). The second category is comprised of rela-tively high output emitters (>185 dB), often termedacoustic harassment devices (AHDs). The latter
devices were originally designed to deter pinnipedsfrom mariculture operations and have been usedwidely by the aquaculture industry (Johnston &Woodley 1998, Quick et al. 2004). AHDs were ini-tially designed to cause discomfort or pain when ananimal approaches closely.
Widespread concerns about depredation by smallcetaceans in a variety of fisheries (reviewed in Read2008) have resulted in the use of devices with inter-mediate output levels becoming more common incommercial fisheries. Thus, the range of availableactive sound emitters no longer falls into 2 discretecategories but is more of a continuum (Table 1).Many mid-range devices are designed to deterdepredation of static fishing gear by coastal bottle-nose dolphins Tursiops truncatus (see Table 1), butsome have also been designed or used to deter cap-ture of pelagic dolphins in mid-water trawls (e.g.DDD03, Cetasaver; Table 1).
We focus here on the use of acoustic devices toreduce bycatch of small cetaceans in static fishinggear, primarily gillnets. We focus on gillnets becausebycatch in this type of gear is widely considered themost important threat to populations of small ceta -ceans (Read et al. 2006). Bycatch in gillnets is the pri-mary conservation issue for the 2 most endangeredsmall cetaceans, the vaquita Phocoena sinus (Rojas-Bracho et al. 2006) and Maui’s dolphin Cephalo -rhynchus hectori maui (Slooten et al. 2006). In addi-tion, bycatch is known to threaten several otherpopulations (e.g. franciscana Pontoporia blain villei,Secchi & Wang 2002; Hector’s dolphin Ce phalo -rhynchus hectori, Slooten et al. 2000; Baltic harbourporpoises Phocoena phocoena, Berggren et al. 2002).Considering that small cetaceans use sound both forcommunication and echolocation, it is not surprisingthat acoustic devices have been explored as a tool forreducing bycatch. Indeed, the use of pingers hasbeen considered or adopted as a conservation meas-ure for each of the above species and many others.
In this review we consider only the small, battery-powered pingers that can be attached directly to gill-nets. We have not considered other acoustic deter-rent practices such as the use of percussion tubes orpipes to keep animals away from gillnets and othergears (e.g. Kasuya 1985, Zahri et al. 2004), nor anymethods involving playbacks of deterrent noisessuch as killer whale calls (Fish & Vania 1971, ICES2010).
Our purpose is to provide a comprehensive reviewof the use of pingers as tools to reduce bycatch ofsmall cetaceans in gillnet fisheries. We review 4types of studies: (1) controlled experiments in com-
202
1Gillnet is meant here generically, including all static entan-gling nets, including trammel nets.
2All source levels in this paper are referenced to 1 µPa at 1 m
Dawson et al.: Pingers, small cetaceans, and gillnets 203
Pin
ger
Fre
q.
Sou
rce
Bro
adca
stD
ura
tion
Inte
rval
Man
ufa
ctu
rer
Ava
ilab
le?
Sp
ecie
s te
sted
Stu
die
sS
ourc
e(k
Hz)
leve
l (d
B)
(pri
ce)
Air
mar
1013
2C
onti
nu
ous
300
ms
4 s
Air
mar
Yes
Ph
ocoe
na
ph
ocoe
na
Net
tri
al,
Sea
Fis
h I
nd
ust
ry A
uth
orit
y (2
005)
, Cos
gro
ve e
t g
illn
etT
ech
nol
ogy
(€50
)en
tan
gle
men
tal
. (20
05),
Kro
g &
Lar
sen
(20
07)
pin
ger
C
orp
.A
irm
ar 7
070
132
Con
tin
uou
s30
0 m
s4
sA
irm
ar?
Pon
top
oria
bla
invi
llei
Net
tri
alB
ord
ino
et a
l. (
2004
)T
ech
nol
ogy
Cor
p.
Aq
uam
ark
FM
20−
145
Con
tin
uou
s20
0−5−
30 s
Aq
uat
ecY
esP
hoc
oen
a p
hoc
oen
aN
et t
rial
,S
ea F
ish
In
du
stry
Au
thor
ity
(200
5), C
osg
rove
et
100
160
300
ms
Su
bse
a L
td(G
B £
80)
dep
red
atio
n,
al. (
2005
), D
esp
orte
s et
al.
(20
06),
Kro
g &
Lar
sen
en
tan
gle
men
t(2
007)
, Lar
sen
& K
rog
(20
07),
Gaz
oet
al. (
2008
),
Mor
izu
r et
al.
(20
09),
Har
dy
& T
reg
enza
(20
10)
Aq
uam
ark
FM
5−
145
Con
tin
uou
s20
0−4−
21A
qu
atec
Yes
Ste
nel
la c
oeru
leoa
lba
Byc
atch
net
Le
Ber
re (
2005
),Im
ber
t et
al.
(20
07)
200
160
300
ms
Su
bse
a L
td(G
B £
80)
tria
lA
qu
amar
kF
M 5
−15
0C
onti
nu
ous
50−
5−30
sA
qu
atec
Yes
Tu
rsio
ps
tru
nca
tus
Dep
red
atio
nB
roto
ns
et a
l. (
2008
b)
210
160
300
ms
Su
bse
a L
td(G
B £
88)
Aq
uam
ark
1013
2C
onti
nu
ous
300
ms
4 s
Aq
uat
ecY
es30
0 S
ub
sea
Ltd
(GB
£88
)A
qu
amar
k5−
3016
5C
onti
nu
ous
Var
iou
sV
ario
us
Aq
uat
ecY
es84
8 S
ub
sea
Ltd
(GB
£50
00)
Aq
uat
ec20
−16
016
5C
onti
nu
ous
<1
s5−
20 s
Lou
gh
bor
oug
h/
See
Del
ph
inu
s d
elp
his
,O
bse
rvat
ion
alL
een
ey e
t al
. (20
07),
Ber
row
et
al. (
2008
)C
PA
qu
atec
Aq
uam
ark
Tu
rsio
ps
tru
nca
tus
100
Aq
uat
ec35
−16
015
7R
esp
onsi
ve30
0 m
s−5−
20 s
Lou
gh
bor
oug
h/
No
Del
ph
inu
s d
elp
his
,O
bse
rvat
ion
alL
een
ey e
t al
. (20
07),
Ber
row
et
al. (
2008
)R
P1
sA
qu
atec
Tu
rsio
ps
tru
nca
tus
Mod
ifie
d20
−16
015
7R
esp
onsi
ve10
s1−
6 s
Lou
gh
bor
oug
h/
No
Del
ph
inu
s d
elp
his
Ob
serv
atio
nal
Ber
row
et
al. (
2008
)A
qu
atec
ran
dom
Aq
uat
ecR
PC
etas
aver
FM
30−
190
Con
tin
uou
s1
s2−
5.5
sIF
RE
ME
R-
Yes
Del
ph
inu
s d
elp
his
Ob
serv
atio
nal
Ber
row
et
al. (
2008
)V
.03
150
IXT
raw
l(€
5890
)D
DD
-02F
FM
5−
174
Con
tin
uou
s0.
5−9
sR
and
omS
TM
Pro
du
cts
No
Del
ph
inu
s d
elp
his
Ob
serv
atio
nal
,B
erro
w e
t al
. (20
08),
Nor
thri
dg
e et
al.
150
dep
red
atio
n(2
008)
, Bu
scai
no
et a
l. (
2009
)D
DD
03L
5−50
017
4C
onti
nu
ous
0.5−
9 s
Ran
dom
ST
M P
rod
uct
sY
es (
€25
0)P
hoc
oen
a p
hoc
oen
aE
nta
ng
lem
ent
Mor
izu
r et
al.
(20
09),
Nor
thri
dg
e et
al.
(20
11)
DD
D 0
3N5−
500
174
Con
tin
uou
s0.
5−9
sR
and
omS
TM
Pro
du
cts
Yes
(€
250)
DD
D 0
3HF
M 3
0−17
4C
onti
nu
ous
0.5−
9 s
Ran
dom
ST
M P
rod
uct
sY
es (
€25
0)80
DD
D 0
3U5−
500
174
Con
tin
uou
s0.
5−9
sR
and
omS
TM
Pro
du
cts
Yes
(€
250)
DiD
01
FM
5−
174
Inte
ract
ive
0.5−
9 s
Ran
dom
ST
M P
rod
uct
sY
es15
0(€
345)
Du
kan
e10
a13
2C
onti
nu
ous
300
ms
4 s
Du
kan
eN
oC
eph
alor
hyn
chu
sN
et t
rial
,K
rau
s et
al.
(19
97),
Sto
ne
et a
l. (
1997
, 200
0),
Net
mar
kh
ecto
ri, D
elp
hin
us
obse
rvat
ion
al,
Kra
us
& B
rau
lt (
1999
), T
rip
pel
et
al. (
1999
),
1000
del
ph
is, P
hoc
oen
ad
epre
dat
ion
Cox
et
al. (
2001
, 200
4), S
MR
U (
2001
), B
ord
ino
ph
ocoe
na,
Pon
top
oria
et a
l. (
2002
), C
arls
tröm
et
al. (
2002
, 200
9),
bla
invi
llei
, Sot
alia
Bar
low
& C
amer
on (
2003
), M
onte
iro-
Net
o et
al.
fl
uvi
atil
is, T
urs
iop
s(2
004)
, Bro
ton
s et
al.
(20
08b
), G
önen
er &
Bil
gin
tr
un
catu
s, L
agen
o-(2
009)
, Alf
aroS
hig
uet
o (2
010)
rhyn
chu
s ob
scu
rus,
Ph
ocoe
na
spin
ipin
nis
,
Tab
le 1
. Dif
fere
nt b
ran
ds
of p
ing
ers
wit
h a
cou
stic
sp
ecif
icat
ion
s. S
ourc
e le
vels
(as
clai
med
by
the
man
ufa
ctu
rer)
are
ref
eren
ced
to 1
µP
a at
1 m
. Pri
ces
giv
en h
ere
are
as o
fA
ug
ust
201
2 an
d f
or 1
pin
ger
on
ly; p
rice
s va
ry b
y th
e q
uan
tity
ord
ered
. Th
e A
qu
amar
k 8
48 a
nd
Cet
asav
er V
.3 a
re e
ssen
tial
ly e
xper
imen
tal
dev
ices
in
ten
ded
for
det
er-
ren
ce o
f w
hal
e sp
ecie
s (A
qu
amar
k)
or u
se i
n t
raw
l n
ets
(Cet
asav
er V
.3).
FM
: fr
equ
ency
mod
ula
ted
; D
DD
: d
olp
hin
det
erre
nt
dev
ice;
DiD
: d
olp
hin
in
tera
ctiv
e d
evic
e
Tab
le 1
con
tin
ued
on
nex
t p
age
Endang Species Res 19: 201–221, 2013204
Glo
bic
eph
ala
sp.
Pin
ger
Fre
q.
Sou
rce
Bro
adca
stD
ura
tion
Inte
rval
Man
ufa
ctu
rer
Ava
ilab
le?
Sp
ecie
s te
sted
Stu
die
sS
ourc
e(k
Hz)
leve
l (d
b)
(pri
ce)
Fis
hT
ek40
−11
0,15
4C
onti
nu
ous
400
ms
4−10
s,
Fis
hT
ekY
es
Ph
ocoe
na
ph
ocoe
na
Lim
ited
im
ple
-B
anan
ara
nd
omra
nd
om(G
B£2
0)m
enta
tion
, tri
alP
ing
eru
nd
erw
ay i
nB
P15
4H
olla
nd
Fu
mu
nd
aa3b
135
Con
tin
uou
s30
0 m
s4
sF
um
un
da
Yes
Meg
apte
raL
imit
ed i
mp
le-
(US
$150
)n
ovae
ang
liae
men
tati
on, t
rial
un
der
way
on
shar
k n
ets
inQ
uee
nsl
and
&N
ew S
outh
Wal
esF
um
un
daa
10b
132c
Con
tin
uou
s30
0 m
s4
sF
um
un
da
Yes
Tu
rsio
ps
tru
nca
tus,
En
tan
gle
men
tS
ea F
ish
In
du
stry
Au
thor
ity
(200
5), C
osg
rove
et
(US
$72)
Ph
ocoe
na
ph
ocoe
na,
dep
red
atio
n,
al. (
2005
),K
rog
& L
arse
n (
2007
),B
erg
Sot
o et
al.
S
ousa
ch
inen
sis
net
tri
al(2
009)
,Rea
d &
Wap
les
(201
0), C
arre
tta
&
Bar
low
(20
11)
Fu
mu
nd
aa70
145
Con
tin
uou
s30
0 m
s4
sF
um
un
da
Yes
Tu
rsio
ps
tru
nca
tus
Dep
red
atio
nR
ead
& W
aple
s (2
010)
(US
$100
)L
ien
2.9b
123
Con
tin
uou
s30
0 m
s30
0 m
sM
odif
ied
by
Sel
f-b
uil
tP
hoc
oen
a p
hoc
oen
aN
et t
rial
,K
osch
insk
i &
Cu
lik
(19
97),
Laa
ke
et a
l. (
1998
),
Fu
llil
ove
(199
4)ob
serv
atio
nal
Ric
hte
r et
al.
(19
99),
Gea
rin
et
al. (
2000
)M
arex
i10
132
Con
tin
uou
s30
0 m
s4
sM
arex
i?
Ph
ocoe
na
ph
ocoe
na
Byc
atch
,K
rog
& L
arse
n (
2007
),M
oriz
ur
et a
l. (
2009
)n
et t
rial
PIC
E-9
920
−16
014
5C
onti
nu
ous
300
ms
4−30
s,
Lou
gh
bor
oug
hN
oP
hoc
oen
a p
hoc
oen
a,N
et t
rial
,N
orth
rid
ge
et a
l. (
1999
),S
ton
e et
al.
(20
00),
ran
dom
Un
iver
sity
Cep
hal
orh
ynch
us
obse
rvat
ion
alC
uli
k e
t al
. (20
01)
hec
tori
Sav
eWav
e20
−80
145
Con
tin
uou
s30
0 m
s3−
5 s,
Sav
eWav
e B
VY
esH
iPro
tect
ran
dom
Sav
eWav
e5−
160
155
Con
tin
uou
s20
0−,
4−16
s,
Sav
eWav
e B
VY
esT
urs
iop
s tr
un
catu
sD
epre
dat
ion
,N
orth
rid
ge
et a
l. (
2003
),S
ea F
ish
In
du
stry
D
olp
hin
Sav
er90
0 m
sra
nd
om(€
60)
net
tri
alA
uth
orit
y(2
005)
, Kro
g &
Lar
sen
(20
07),
Bro
ton
s ra
nd
omet
al.
(20
08b
),W
aple
s et
al.
(20
13)
Sea
mas
ter
10−
9016
5C
onti
nu
ous
1.9
s15
sS
eam
aste
rY
esA
ust
rali
a(U
S$8
50)
a Nam
e si
nce
ch
ang
ed t
o F
utu
re O
cean
s; b
Incl
ud
es m
ult
iple
ult
raso
nic
har
mon
ics;
c Tes
ted
at
144
dB
(S
. M. D
awso
n u
np
ub
l. d
ata)
Tab
le 1
. (co
nti
nu
ed)
Dawson et al.: Pingers, small cetaceans, and gillnets
mercial gillnet fisheries, (2) studies of the behaviourof small cetaceans around pingers or gillnets, (3)observations of bycatch rates in fisheries in whichpingers are used as part of a mitigation strategy, and(4) studies of the use of pingers to reduce depreda-tion by dolphins. We include the latter becausedepredation can result in entanglement, and thesestudies provide insight into how dolphins react topingers. We have reviewed both published andunpublished studies. For studies not published inpeer-reviewed journals, we required that a report ofprofessional quality was available detailing thestudy’s methods and results. We attempt to sum-marise what we see as the emerging lessons andremaining uncertainties.
CONTROLLED EXPERIMENTS IN GILLNETFISHERIES
Researchers wishing to determine whether pingerswill reduce bycatches of small cetaceans in a com-mercial gillnet fishery typically employ an experi-mental protocol in which bycatch rates are comparedin nets with and without active pingers. Wheneverpossible, all other attributes of the nets (e.g. meshsize, length, depth, hanging ratio) are kept constant.Independent observers are used to monitor bycatchrates in both types of nets. Some experiments arequite elaborate, including blind protocols in whichneither fishers nor observers know which pingers areactive or silent, to eliminate any chance of intentionalor unintentional bias. In such cases, the pingers areactivated upon immersion in salt water (e.g. Kraus etal. 1997). After several studies had shown that har-bour porpoise bycatch in gillnet fisheries can be sub-stantially reduced by using pingers, several recentstudies in European waters have used less rigorousexperimental designs to test specific devices in par-ticular fisheries. In these cases, the main objectivehas been to assess whether the devices are effectiveand practical under realistic fishery conditions ratherthan in a controlled experiment.
We reviewed 19 controlled experiments in whichthe effect of pingers on bycatch rates of dolphins orporpoises in gillnet fisheries was examined. Fourteenof these focussed on harbour porpoises (Table 2).Only 3 of these studies failed to produce statisticallysignificant reductions. In 1 of these, no porpoiseswere caught in either control nets or nets withpingers (Carlström et al. 2002). In another, thepingers used had several faults which resulted intheir failure (Northridge et al. 1999), while in the
third study, sample sizes were small, and there wasalso a high rate of pinger failure (Morizur et al. 2009).Even if a bias exists, in which studies demonstratingsignificant differences are more likely to be submit-ted and accepted for publication than those that donot, this is a striking result. Taken together, they indi-cate that large reductions in harbour porpoisebycatch can be achieved in controlled experimentswith a variety of pinger types over much of the spe-cies’ range. We concur with the results of previousreviews (e.g. IWC 2000) that further experimentationto evaluate the efficacy of pingers with harbour por-poises and gillnet fisheries is unnecessary.
The other 5 controlled experiments addressed thebycatch of franciscana in Argentina (Bordino et al.2002), a variety of small cetacean species off Califor-nia (Barlow & Cameron 2003) and off Peru (AlfaroShigueto 2010), striped dolphins in the Mediterran-ean (Imbert et al. 2007), and bottlenose dolphins inAustralia (McPherson et al. 2004). The first 3 studiesused the Netmark 1000 pinger, the most extensivelytested alarm, and produced bycatch reductions>70%. The acoustic specifications of this unit arenow required for pingers used in the formal ‘TakeReduction Plans’ of the California driftnet fishery andthe Gulf of Maine ground fish gillnet fishery, but thedevice itself is no longer sold commercially.
The trials using pingers to reduce bycatch ofstriped dolphins in the Thonaille fishery for tuna inthe NW Mediterranean (Imbert et al. 2007) are diffi-cult to interpret because an initial controlled experi-ment in August 2001 was followed by 2 seasons ofsimple observation of the effects of widespread, butinconsistent, pinger use with few control operations.Nevertheless, a total of 344 pingered and unpingereddriftnet sets were monitored from 2001 to 2003. Twopinger types (Aquamark 200 and Fumunda 10 kHz)were trialled. The 2001 data showed a significantreduction (81%) in striped dolphin bycatch with theuse of Aquamark pingers, but subsequent bycatchrates with more widespread deployment were highlyvariable and not consistent with the low rates initiallyobserved. At least 4 factors may have been involved:(1) skippers apparently used the devices sparingly;(2) devices were not always used at the correct spac-ing; (3) depleted batteries were not dealt with consis-tently (which required the purchase of new pingersin the case of the Aquamark device); and (4) at timesnets correctly equipped with functioning pingerscaught dolphins at a rate similar to nets withoutpingers, and some of these dolphins seem to havebeen fatally attracted to the devices rather than displaced.
205
Endang Species Res 19: 201–221, 2013206
Sp
ecie
sP
ing
er
F
req
.
Sp
acin
gC
ontr
ol n
ets
Net
s w
ith
pin
ger
s
Byc
atch
Sta
tist
ical
Loc
atio
n
Sou
rce
(k
Hz)
(m
)
No.
of
N
o. o
f
re
du
ctio
n
sig
nif
ican
ce
k
ills
s
ets
kil
ls
set
s
(
%)
Ph
ocoe
na
Net
mar
k
1
0a
92
25
423
2
421
92
Y
es
G
ulf
of
Mai
ne,
Mai
ne,
US
A
K
rau
s et
al.
(19
97)
ph
ocoe
na
1000
Ph
ocoe
na
Net
mar
k
1
0a
92
8
2
20
0
1
80
1
00
Yes
Jef
feri
es L
edg
e,
K
rau
s &
Bra
ult
(19
99)
ph
ocoe
na
1000
Mas
sach
use
tts,
US
AP
hoc
oen
aL
U-1
40−
120
70
23
9
075
1
6523
94
Y
es
D
anis
h N
orth
Sea
L
arse
n (
1999
)p
hoc
oen
ap
roto
typ
eP
hoc
oen
aP
ice-
99
20
−16
0
9
5
9
150
4
35
−90
No
Cel
tic
Sea
Nor
thri
dg
e et
al.
(19
99)
ph
ocoe
na
Ph
ocoe
na
Lie
n
2.9
5a
9
−12
2
2
117
9
2
5
13
79
Yes
Gra
nd
Man
an I
slan
d, B
ay o
f
Ric
hte
r et
al.
(19
99)
ph
ocoe
na
F
un
dy,
Can
ada
Ph
ocoe
na
Net
mar
k
1
0a
100
14
26
7
3
2
49
77
Yes
Low
er B
ay o
f F
un
dy,
Can
ada
T
rip
pel
et
al. (
1999
)p
hoc
oen
a10
00P
hoc
oen
aL
ien
2
.95a
1
6
4
7
12
0
14
29
2
88
Yes
Sp
ike
Roc
k, W
ash
ing
ton
, US
A
G
eari
n e
t al
. (20
00)
ph
ocoe
na
Ph
ocoe
na
Net
mar
k
1
0a
100
10
23
7
1
1
81
87
Yes
Cel
tic
Sea
SM
RU
(20
01)
ph
ocoe
na
1000
Ph
ocoe
na
Net
mar
k
1
0a
100
0
1
84
0
1
89
0
N
o
S
kag
erra
k S
ea
Car
lstr
öm e
t al
. (20
02)
ph
ocoe
na
1000
Ph
ocoe
na
Aq
uam
ark
20
−16
0
4
55
22
41
0
24
100
Y
es
D
anis
h N
orth
Sea
L
arse
n &
Kro
g (
2007
)p
hoc
oen
a10
0P
hoc
oen
aA
qu
amar
k
20−
160
585
22
4
1
5
4
3
78
Y
es
D
anis
h N
orth
Sea
L
arse
n &
Kro
g (
2007
)p
hoc
oen
a10
0P
hoc
oen
aN
etm
ark
10a
2
00
92
20
tri
ps
2
20
trip
s
98
Y
es
B
lack
Sea
Gön
ener
& B
ilg
in (
2009
)p
hoc
oen
a10
00P
hoc
oen
aA
qu
amar
k
Var
iou
s
1 k
m,
3
4
62 k
m
2
1
50 k
m
−51
No
Iro
ise
Sea
(C
elti
c S
ea)
M
oriz
ur
et a
l. (
2009
)p
hoc
oen
a10
0, D
DD
02,
4
00 m
,M
arex
i
200
mP
hoc
oen
aD
DD
-03
5−25
0
Var
iab
le
1
6
78
0
7
9
29
6
3b
Yes
Cel
tic
Sea
Nor
thri
dg
e et
al.
(20
11)
ph
ocoe
na
t
o 8
km
Pon
top
oria
Net
mar
k
1
0a
50
45
309
7
295
84
Y
es
C
abo
San
An
ton
io, A
rgen
tin
a
Bor
din
o et
al.
(20
02)
bla
invi
llei
1000
Dep
hin
us
Net
mar
k
1
0a
91
18
314
3
295
82
Y
es
C
alif
orn
ia, U
SA
Bar
low
& C
amer
on (
2003
)d
elp
his
1000
Ste
nel
laA
qu
amar
k
5−
160
100
7
1
2c
3
2
7
81
Y
es
L
igu
rian
Sea
Im
ber
t et
al.
(20
07)
coer
ule
oalb
a20
0D
elp
hin
us
Net
mar
k
1
0a
200
27
7
1
5
4
9
73
Y
es
P
eru
A
lfar
o S
hig
uet
o (2
010)
del
ph
is,
1000
Lag
eno-
rhyn
chu
sob
scu
rus,
Tu
rsio
ps
sp.,
Glo
bic
eph
ala
sp.
Tu
rsio
ps
Not
2.
9 an
d
Var
iab
le
3
1
60
2
1
05
−2
N
o
Q
uee
nsl
and
, Au
stra
lia
M
cPh
erso
n e
t al
. (20
04)
tru
nca
tus
rep
orte
d
10
a In
clu
des
mu
ltip
le u
ltra
son
ic h
arm
onic
s; b
Byc
atch
red
uct
ion
was
95
% f
or s
pac
ing
of
<4
km
, b
ut
not
sig
nif
ican
tly
dif
fere
nt
from
con
trol
s w
ith
sp
acin
g >
4 k
m a
par
t; o
vera
ll 6
3%
red
uct
ion
; c Dat
a on
ly f
rom
Au
gu
st 2
001
—in
oth
er m
onth
s ex
per
imen
tal
con
trol
was
les
s ri
gid
bu
t th
e ap
par
ent
rate
red
uct
ion
was
mu
ch l
ess
mar
ked
Tab
le 2
. Con
trol
led
exp
erim
ents
com
par
ing
en
tan
gle
men
t fr
equ
ency
in n
ets
wit
h a
nd
wit
hou
t ac
tive
pin
ger
s
Dawson et al.: Pingers, small cetaceans, and gillnets 207
Too few dolphins were caught for McPherson etal.’s (2004) trials to be statistically powerful, but theirresults do not suggest that these pingers caused anyreduction in the bycatch of bottlenose dolphins.Observers in this study also noted that bottlenosedolphins sometimes behaved aggressively towardthe pingers, repeatedly attacking them.
STUDIES OF BEHAVIOURAL RESPONSES TOPINGERS
Controlled experiments on the effects of pingers onbycatch rate obviously provide the most direct test oftheir efficacy as a mitigation strategy, but studies ofthe behavioural response of animals to pingers canbe extremely informative. Such an approach allowsinvestigation of the mechanism of bycatch reduction,which is seldom shown by controlled experimentsalone. Behavioural studies also have broad appealbecause no animals are killed. In addition, such stud-ies are typically much less expensive than the con-trolled experiments described above.
Researchers have used several approaches toexamine the behavioural response of small cetaceansto pingers. In approximately half of the studies wereviewed, researchers used theodolites to track themovements of individuals or groups in the vicinity ofa pinger or an actual or simulated gillnet (Table 3).The theodolite is used to estimate the position of indi-vidual animals or groups of animals at the surface.This allows researchers to compare closest approachdistances when pingers are active and silent and,thus, to determine whether animals are displaced bythe sound of the device. Unfortunately, several stud-ies have considered each surfacing as an independ-ent event, when in fact they are auto-correlated. Thisinflates sample size and can result in a falsely signif-icant statistical test. The simplest way around thisproblem is to use only the closest observed approachdistance for each group of dolphins or porpoises(Dawson & Lusseau 2005).
In several studies researchers observed the re -sponse of animals to an active pinger lowered from anearby boat. Six studies supplemented or replacedvisual observations by monitoring the occurrence ofecholocation signals using click detectors (e.g. Cox etal. 2001, Desportes et al. 2006, Leeney et al. 2007,Northridge et al. 2008, Carlström et al. 2009, Hardy &Tregenza 2010). In about half of the studies wereviewed, the studies were conducted blind, so theobservers did not know whether the pinger wasactive.
All but one study of harbour porpoises showedsome degree of avoidance, fewer acoustic detections(Hardy & Tregenza 2010), or physical displacementin which animals surfaced farther away whenpingers were active, in some cases by several hun-dred metres. The remaining study (Desportes et al.2006) showed no displacement reaction to Aquamark100 pingers, despite demonstrated bycatch reduc-tions using this pinger in controlled experiments(Larsen & Krog 2007). The Desportes et al. (2006)study took place in an experimental area wherepingers had been used in previous years.
The results of studies of other species are less clear.Common dolphins in Ireland showed no obviousreaction to a variety of pingers presented from a boat(Berrow et al. 2008), nor did individuals of this spe-cies react to several pinger signals deployed from aboat in Spain (Sagarminaga et al. 2006). Commondolphins in France, however, apparently showed adramatic reaction to a DDD pinger (van Marlen2007). Another study in England showed a decreasein click detections (presumed to be common dolphin)when a DDD-02 was deployed on a gillnet comparedwith when the net and pinger were absent (North-ridge et al. 2008).
Surfacing positions of tucuxi Sotalia fluviatilis(Monteiro-Neto et al. 2004), Sousa chinensis (BergSoto et al. 2009) and Hector’s dolphin (Stone et al.1997) were not obviously affected by the presence ofactive Netmark 100 or Fumunda pingers. These spe-cies did not show the clear zone of displacementdemonstrated by harbour porpoises.
Bottlenose dolphins exhibit variable responses topingers. Cox et al. (2004) observed fewer dolphingroups within 100 m of nets with active pingers com-pared to nets with inactive pingers, but there was nosignificant difference in closest observed approach.Leeney et al. (2007) monitored the echolocationactivity of bottlenose dolphins in the vicinity ofmoored pingers using T-POD echolocation detectors.These researchers tested 2 Aquatec pinger types,one that continuously produced sound (‘continuous’pinger, CP) and one that emitted sound only afterdetecting dolphin echolocation clicks (‘responsive’pinger, RP), against dummy pingers that made nosound. The CP and RP pingers made broadband fre-quency-modulated (FM) sweeps with a source levelof 165 dB — much louder than most other pingersused to reduce the bycatch of small cetaceans. Therewere significantly fewer dolphin echolocation detec-tions when the CP was active. There was also lessecholocation activity in the presence of active RPs,but the active versus dummy comparison was non-
Endang Species Res 19: 201–221, 2013208
Sp
ecie
s
P
ing
er
Exp
erim
ent
typ
e
Res
pon
se
Loc
atio
n
S
ourc
e
Cep
hal
orh
ynch
us
N
etm
ark
100
0
T
heo
dol
ite
obse
rvat
ion
s of
Non
e ob
viou
sa
A
kar
oa H
arb
our,
S
ton
e et
al.
(19
97)
hec
tori
m
oore
d p
ing
er (
bli
nd
)
New
Zea
lan
d
Cep
hal
orh
ynch
us
P
ice,
Net
mar
k 1
000
O
bse
rvat
ion
s of
pin
ger
s
Avo
idan
ce r
eact
ion
to
Net
mar
k 1
000
A
kar
oa H
arb
our,
S
ton
e et
al.
(20
00)
hec
tori
d
eplo
yed
fro
m s
ame
boa
t
o
nly
, in
62
% o
f ca
ses
New
Zea
lan
d
Del
ph
inu
s
BIM
CP,
BIM
RP,
O
bse
rvat
ion
s of
pin
ger
s
Non
e ob
viou
s
C
ork
, Ire
lan
d
Ber
row
et
al. (
2008
)d
elp
his
Mod
ifie
d B
IM R
P,
d
eplo
yed
fro
m s
ame
boa
t
Cet
asav
er, D
DD
(
bli
nd
)
Del
ph
inu
s
DD
D
T
-PO
D o
bse
rvat
ion
s of
Dec
reas
e in
cli
ck d
etec
tion
s
C
orn
wal
l,
Nor
thri
dg
e et
al.
(20
08)
del
ph
is
pin
ger
s at
tach
ed t
o n
et
E
ng
lan
d
Ph
ocoe
na
L
ien
Th
eod
olit
e ob
serv
atio
ns
of
97
/105
gro
up
s ‘a
void
ed’
Cla
yoq
uot
Sou
nd
,
K
osch
insk
i &
Cu
lik
(19
97)
ph
ocoe
na
sim
ula
ted
gil
lnet
Bri
tish
Col
um
bia
, Can
ada
Ph
ocoe
na
L
ien
Th
eod
olit
e ob
serv
atio
ns
D
isp
lace
men
t b
y >
125
m
S
pik
e R
ock
,
L
aak
e et
al.
(19
98)
ph
ocoe
na
of g
illn
et
W
ash
ing
ton
, US
A
Ph
ocoe
na
N
etm
ark
100
0
T
heo
dol
ite
& T
-PO
D o
bse
r-
Dis
pla
cem
ent
init
iall
y 20
8 m
, few
er
G
ran
d M
anan
Isl
and
,
C
ox e
t al
. (20
01)
ph
ocoe
na
vati
ons
of m
oore
d p
ing
er
T-P
OD
det
ecti
ons
wh
en p
ing
er o
n
Bay
of
Fu
nd
y, C
anad
a
Ph
ocoe
na
P
ice
Th
eod
olit
e ob
serv
atio
ns
M
edia
n a
pp
roac
h d
ista
nce
in
crea
sed
C
layo
qu
ot S
oun
d,
Cu
lik
et
al. (
2001
)p
hoc
oen
a
of
sim
ula
ted
gil
lnet
from
150
to
530
m
B
riti
sh C
olu
mb
ia, C
anad
a
Ph
ocoe
na
M
odif
ied
Aq
uam
ark
T
heo
dol
ite
and
hyd
rop
hon
e
No
obvi
ous
dis
pla
cem
ent
effe
cts
Fyn
s H
oved
, Den
mar
k
D
esp
orte
s et
al.
(20
06)
ph
ocoe
na
10
0 (A
Q62
6 IP
),
ob
serv
atio
ns
of m
oore
d
Aq
uam
ark
100
pin
ger
s an
d s
imu
late
d g
illn
et
Ph
ocoe
na
N
etm
ark
100
0
T
heo
dol
ite
& T
-PO
D o
bse
r-
Ave
rag
e ap
pro
ach
dis
tan
ce i
ncr
ease
d
Wes
t S
cotl
and
C
arls
tröm
et
al. (
2009
)p
hoc
oen
a
va
tion
s of
sim
ula
ted
gil
lnet
b
y 30
0 m
an
d s
ign
ific
antl
y fe
wer
(bli
nd
)
T-P
OD
det
ecti
ons
wh
en p
ing
ers
on
Ph
ocoe
na
A
qu
amar
k 1
00
C
-PO
D o
bse
rvat
ion
s of
pin
ger
s
Sig
nif
ican
t d
ecre
ase
in c
lick
det
ecti
ons
Cor
nw
all,
En
gla
nd
Har
dy
& T
reg
enza
(20
10)
ph
ocoe
na
atta
ched
to
net
s
ar
oun
d n
ets
wit
h p
ing
ers
Sot
alia
N
etm
ark
100
0
O
bse
rvat
ion
s of
moo
red
N
one
obvi
ousa
Ira
cim
a B
each
,
M
onte
iro-
Net
o et
al.
(20
04)
flu
viat
ilis
pin
ger
s (b
lin
d)
F
orta
leza
, Bra
zil
Sou
sa
F
um
un
da
10
Vio
dol
ite
obse
rvat
ion
s of
pin
ger
N
o d
iffe
ren
ce i
n c
lose
st o
bse
rved
Q
uee
nsl
and
, Au
stra
lia
Ber
g S
oto
et a
l. (
2009
)ch
inen
sis
arra
y &
ob
serv
atio
ns
of p
ing
er
ap
pro
ach
to
arra
y; n
o su
bst
anti
al
dep
loye
d f
rom
sam
e b
oat
ch
ang
e in
beh
avio
ur
Tu
rsio
ps
Net
mar
k 1
000
Th
eod
olit
e ob
serv
atio
ns
of
N
o d
iffe
ren
ce i
n c
lose
st o
bse
rved
A
tlan
tic
Bea
ch,
Cox
et
al. (
2004
)tr
un
catu
s
m
oore
d g
illn
et (
bli
nd
)
a
pp
roac
h, b
ut
sig
nif
ican
tly
few
er g
ruop
s
Nor
th C
arol
ina,
US
A
wit
hin
100
m w
hen
pin
ger
s ac
tive
Tu
rsio
ps
Aq
uat
ec C
P a
nd
RP
M
oore
d p
ing
er w
ith
T-P
OD
s,
S
ign
ific
antl
y fe
wer
T-P
OD
det
ecti
ons
S
han
non
Est
uar
y,
Lee
ney
et
al. (
2007
)tr
un
catu
s
an
d b
oat
tria
ls
wit
h a
ctiv
e C
P; a
void
ance
in
3 o
f 4
boa
t
Ir
elan
d
tri
als
of e
ach
pin
ger
typ
e
Tu
rsio
ps
IC
A S
.L A
HD
M
oore
d A
HD
att
ach
ed t
o
N
o ef
fect
on
dol
ph
in p
rese
nce
, gro
up
siz
e, S
ard
inia
, Ita
ly
Lóp
ez &
Mar
iño
(201
2)tr
un
catu
s
fi
sh f
arm
cag
e
tim
e sp
ent
in t
he
area
, dis
tan
ce t
o A
HD
a Sta
tisi
cal
resu
lt n
ot v
alid
(se
e D
awso
n &
Lu
ssea
u 2
005)
Tab
le 3
. S
tud
ies
of b
ehav
iou
ral
resp
onse
s of
fre
e-ra
ng
ing
dol
ph
ins
and
por
poi
ses
to p
ing
ers.
CP
: co
nti
nu
ous
pin
ger
; R
P:
resp
onsi
ve p
ing
er;
DD
D:
dol
ph
in d
eter
ren
t d
evic
e;A
HD
: aco
ust
ic d
eter
ren
t d
evic
e; T
-PO
D a
nd
C-P
OD
: aco
ust
ic e
ven
t re
cord
ers
(ww
w.c
hel
onia
.co.
uk
); v
iolo
dit
e: c
omb
ined
vid
eo c
amer
a/th
eod
olit
e ob
serv
atio
n s
yste
m
Dawson et al.: Pingers, small cetaceans, and gillnets
significant due to more variable reactions to the RP.It is possible that reduced acoustic detection wascaused by a reduction in vocalisation rather than displacement (e.g. Cox et al. 2001). Both types ofpingers were also used 4 times each in separate trialsfrom a boat to observe visually the responses of thenearby bottlenose dolphins (Leeney et al. 2007).While the sample size is very small, and the trial con-ditions were somewhat artificial (see subsection‘Studies of behavioural responses to pingers’), strongavoidance reactions occurred in 3 of the 4 trials foreach pinger.
IMPLEMENTATION IN COMMERCIAL FISHERIES
Gulf of Maine bottom-set gillnet fishery
Management of harbour porpoise bycatch in gill-net fisheries along the north-eastern United Statesis described in the ‘Harbor Porpoise Take ReductionPlan’ (www.nero.noaa.gov/protected/ porptrp/). Theplan, implemented in 1999, uses a mixture of ap -proaches, including time−area restrictions on fishingeffort, modifications to gear and a requirement touse pingers in certain times and areas. Under thisplan, a pinger is defined as ‘an acoustic deterrentdevice which, when immersed in water, broadcastsa 10 kHz (±2 kHz) sound at 132 dB (±4 dB) re1 micropascal at 1 m, lasting 300 milliseconds (±15milliseconds), and repeating every 4 seconds (±0.2seconds).’ During routine monitoring of by catch oncommercial vessels, observers reported that gillnetsequipped with a full complement of pingers (spacedno further than 92 m apart) caught 60% fewer por-poises per haul than nets without pingers (Palka etal. 2008). This reduction is much smaller than thatachieved in the controlled experiment conducted inthis same fishery (92%; Kraus et al. 1997). Palka etal. argue that this could be because the current fish-ery uses a range of mesh sizes, while those used inthe initial experiment have a higher bycatch rate. Itis likely that other factors also contribute to thissmaller reduction in bycatch rate (see ‘Why arepingers less effective in real fisheries than in con-trolled experiments?’).
The most interesting observation, however, is thatnets equipped with some pingers, but not a full com-plement, catch >2.5 times as many porpoises thanthose with no pingers (Palka et al. 2008). Thus, par-tial implementation of this conservation strategy canbe worse than no implementation at all. The most
important impediment to the use of acoustic pingersto reduce bycatch in this fishery is a lack of com -pliance. As noted by Cox et al. (2007), during 2003,fishermen deployed gillnets without functioning ping -ers in 155 out of 173 trips monitored by on-boardobservers — a noncompliance rate of 78%.
California-Oregon drift net fishery
A large-scale controlled experiment showed a sig-nificant reduction in the bycatch of marine mammals,particularly of common dolphins (Barlow & Cameron2003) in this fishery (Table 2). The use of pingers hasbeen mandated since 1998 as part of the ‘Pacific Off-shore Cetaceans Take Reduction Plan’ (www. nmfs.noaa. gov/ pr/ interactions/ trt/ poctrp. htm).
Since the implementation of pingers in this fishery,the entanglement rate of common dolphins has beenreduced by approximately half (Carretta & Barlow2011), and bycatch of beaked whales has been elim-inated (Carretta et al. 2008). Comparison of recentbycatch rates with those observed soon after adop-tion of pingers shows no evidence of habituation.Hence, in this fishery, like that in the Gulf of Maine,pingers have been effective in reducing cetaceanbycatch, though not to the level seen in the preced-ing controlled experiments. Pinger failure is animportant issue in this fishery too; sets with 1 or morefailed pingers had a significantly higher bycatch rateof cetaceans than sets that were fully equipped withfunctional pingers (Carretta & Barlow 2011).
On observed vessels, compliance with the require-ment to use pingers has remained >98% since theywere made mandatory (Carretta & Barlow 2011). Anincreasing fraction of the vessels in the California-Oregon fishery (11 of 34 in 2009) are regarded asbeing too small to accommodate observers, so com-pliance and bycatch in this part of the fleet is un -known. Nevertheless, compliance regarding pingeruse appears to be considerably better in this fisherythan in the Gulf of Maine. Future research shouldexamine why compliance is so different in these 2fisheries.
In European waters, where EU legislation requiresthe use of pingers to minimise harbour porpoisebycatch in certain fisheries, several other studieshave been conducted to test the robustness and prac-ticality of pingers in fishing operations (Sea FishIndustry Authority 2003, 2005, Cosgrove et al. 2005,Le Berre 2005, Lunneryd 2006, Krog & Larsen 2007).Most pinger types showed significant operationalproblems, including time taken in attachment, diffi-
209
Endang Species Res 19: 201–221, 2013
culty of checking functionality, propensity for tangling the gear and unreliability. For some pingertypes failure rates exceeded 50% (e.g. Sea FishIndustry Authority 2003, Cosgrove et al. 2005). Theseissues are important impediments to widespreadadoption of this mitigation approach, and were notall fully evident in prior scientifically controlledexperiments.
USE OF PINGERS TO REDUCE DEPREDATION
Animals involved in depredation are, by definition,seeking to remove captured fish from a net, a processwhich can damage both catches and gear. The eco-nomic losses resulting from such interactions provideone of the primary incentives to use an acousticdevice. The second incentive is provided by the riskof entanglement to the animals involved. An exam-ple showing both incentives is the inshore gillnetfishery for Spanish mackerel Scomberomorus macu-latus which operates near Cape Hatteras on the eastcoast of the United States (Read et al. 2003, Waples etal. 2013).
It is striking that reported depredation of gillnetsby cetaceans almost always involves bottlenose dol-phins; many of these reports arise from the Mediter-ranean (Lauriano et al. 2004, 2009, López 2006, Bro-tons et al. 2008a, Rocklin et al. 2009). Several studieshave addressed whether pingers reduce damage togillnets and captured fish, testing a number of differ-ent pinger types (Table 4). Interpretation of these tri-als is complicated by the fact that both low-level (ca.132 dB) and high-level (>170 dB) devices have beenused. Some of these studies have examined only asmall number of sets and, hence, had low statisticalpower to detect effects (see Dawson et al. 1998, fordiscussion of power in this context). In addition tothe studies reported here, there have been a numberof uncontrolled and poorly documented ‘experi-ments’ with acoustic pingers in attempts to addressdepredation.
In general, experiments have demonstrated a smallreduction in the damage to nets (reflected in thenumber of holes, for example) and increases in fishcatches when acoustic pingers are employed (North-ridge et al. 2003, Brotons et al. 2008b, Gazo et al.2008, Buscaino et al. 2009). Several other issues limitour ability to draw broad conclusions from this work,including the existence of numerous other marinepredators (sharks, turtles and other fishes) thatengage in depredation, and the fact that some fishlost to depredation are removed completely, i.e. they
cannot be counted. Observations of the behaviourof bottlenose dolphins around nets equipped withpingers (see Brotons et al. 2008b, Waples et al. 2013)suggest that the pingers can reduce the frequencywith which dolphins interact with fishing gear, but donot eliminate such interactions.
In at least 2 cases, however, trials with pingershave not shown any reduction in bottlenose dolphindepredation of gillnets in the Mediterranean (Cor-sica, L. Rossi et al. unpubl. data; Cyprus, M. Hadji -christophorou unpubl. data). Neither study has beenformally published, which underscores the problemthat negative results are less likely to be published.
Importantly, although bycatch rates of bottlenosedolphins are typically very low in these gillnet fish-eries (compared, for example, with those of harbourporpoises in the Gulf of Maine), there have been atleast 2 incidences of lethal entanglements in netsequipped with active pingers.
Northridge et al. (2003) investigated the effect ofSaveWave pingers on depredation of trammel nets inthe eastern Mediterranean by observing 76 sets ofstrings with inactive pingers and 70 with activepingers, where strings were randomly allocated apinger type. A bottlenose dolphin was found dead in1 string equipped with active pingers. Its stomachcontained undigested prey and a piece of 34 mmmesh from a trammel net, clear evidence of the risksof depredation to the animals concerned. A seconddolphin was killed in trials of a prototype 70 kHz145 dB Fumunda pinger undertaken off Cape Hat-teras, North Carolina, USA, in a gillnet fishery forSpanish mackerel (Read & Waples 2010). This pingerwas specifically designed to deter dolphins from ap -proaching nets; its transmitting frequency was set atthe maximum sensitivity of bottlenose dolphin hear-ing (Au 1993). Although isolated incidents, such re -ports do not inspire confidence that acoustic deter-rents are a useful approach to minimizing bottlenosedolphin bycatch.
In summary, the studies of pinger effects on depre-dation by bottlenose dolphins show relatively smalland variable reductions in net or catch damage. Asnoted above, there are issues of low statistical powerin some studies (e.g. Gazo et al. 2008, Buscaino et al.2009), but observed reductions are not consistentlysignificant or of substantial economic benefit. Per-haps most telling is the fact that we are unaware ofany case in which fishermen have voluntarily adoptedand continued the use of acoustic pingers to reducedepredation, although it is certainly true that fisher-men have experimented with such devices outsideformal experimental settings. The number of entan-
210
Dawson et al.: Pingers, small cetaceans, and gillnets 211
Pin
ger
F
req
.
Sou
rce
S
pac
ing
Bli
nd
N
o. o
f
N
o. o
f
Res
pon
se
Fis
h s
pec
ies
L
ocat
ion
S
ourc
e
(kH
z)
leve
l (d
B)
(
m)
a
ctiv
e se
ts c
ontr
ol s
ets
Sav
eWav
e
3
0−16
0
1
55
200
Yes
70
76
S
ign
ific
ant
dec
reas
e in
Var
iou
s,
Aeg
ean
Nor
thri
dg
e et
al.
(20
03)
Dol
ph
insa
ver
d
epre
dat
ion
an
d n
o. o
f h
oles
Sp
arid
ae
Sea
,
in
act
ive
net
s; 1
dol
ph
in
mos
t
Gre
ece
e
nta
ng
led
in
an
act
ive
net
com
mon
Aq
uat
ech
Aq
uam
ark
5−
160
1
30−
155
Not
Yes
743
306
a
I
nte
ract
ion
s w
ith
net
s
D
iver
se
Bal
aeri
c
Bro
ton
s et
al.
(20
08b
)21
0, D
uk
ane
stat
ed
d
ecre
ased
49
% w
ith
act
ive
I
slan
ds,
Net
Mar
k 1
000,
pin
ger
s (o
nly
AQ
UA
mar
k w
as
S
pai
nS
aveW
ave
s
ign
ific
ant)
; no
sig
nif
ican
tD
olp
hin
save
r
eco
nom
ic b
enef
it
Aq
uat
ech
20
−16
0
1
45
150
Yes
27
16
b
87%
few
er h
oles
in
net
s w
ith
Mu
llu
s
B
alae
ric
G
azo
et a
l. (
2008
)A
qu
amar
k 1
00
a
ctiv
e p
ing
ers;
dep
red
atio
n r
ate
su
rmu
letu
s
Isl
and
s,
dec
reas
ed a
bou
t 50
% i
n a
ctiv
e
Sp
ain
n
ets;
no
dif
fere
nce
in
fis
h c
atch
DD
D02
0.1
−20
0
16
0
300
No
2
9
29
31%
few
er h
oles
in
net
s an
d
V
ario
us,
F
avig
nan
a B
usc
ain
o et
al.
(20
09)
2
8%
mor
e fi
sh i
n n
ets
wit
h
Sp
arid
ae
Isla
nd
,
act
ive
pin
ger
s
m
ost
It
aly
com
mon
Fu
mu
nd
a
1
0
132
1
00
N
o
196
179
‘A
t ri
sk’ i
nte
ract
ion
s d
ecre
ased
S
com
ber
omor
us
Hat
tera
s,
Rea
d &
Wap
les
(201
0)
fro
m 8
1 to
50
% i
n a
ctiv
e n
ets;
mac
ula
tus
Nor
th
1 d
olp
hin
en
tan
gle
d i
n a
C
arol
ina,
c
ontr
ol n
et
U
SA
Fu
mu
nd
a
7
0
145
1
00
Y
es
50
44
No
sig
nif
ican
t d
iffe
ren
ce i
n b
e-
Sco
mb
erom
oru
s H
atte
ras,
R
ead
& W
aple
s (2
010)
h
avio
r ar
oun
d a
ctiv
e vs
. con
trol
m
acu
latu
s
N
orth
n
ets;
1 d
olp
hin
en
tan
gle
d i
n a
n
Car
olin
a,
act
ive
net
U
SA
Sav
eWav
e
5−16
0
1
55
100
No
8
3
68
Sig
nif
ican
tly
few
er i
nte
ract
ion
s
Sco
mb
erom
oru
s H
atte
ras,
W
aple
s et
al.
(20
13)
Dol
ph
insa
ver
and
mor
e ec
hol
ocat
ion
wit
h
mac
ula
tus
Nor
th
net
s w
ith
act
ive
pin
ger
s; n
o
Car
olin
a,
dif
fere
nce
in
fis
h c
atch
US
A
a An
ad
dit
ion
al 1
44 s
ets
wit
h n
o p
ing
ers
wer
e ob
serv
ed; b
an a
dd
itio
nal
12
sets
wit
h n
o p
ing
ers
wer
e ob
serv
ed
Tab
le 4
. Tu
rsio
ps
tru
nca
tus.
Stu
die
s u
sin
g p
ing
ers
to d
eter
bot
tlen
ose
dol
ph
in d
epre
dat
ion
. Sou
rce
leve
ls a
re r
efer
ence
d t
o 1
µP
a at
1 m
Endang Species Res 19: 201–221, 2013
glements recorded in these experiments is too smallto address quantitatively the question of whetherpingers reduce the entanglement risk associated withdepredation. Nevertheless, we know of 2 studies inwhich dolphins were fatally entangled in alarmednets, illustrating that pingers do not eliminate en -tanglement risk. Thus, for the animals concerned, de -predation is not a ‘free lunch’. There is no evidencethat pingers which are specifically de signed toreduce depredation (and hence are louder) are anymore effective in reducing entanglement than thequieter devices designed to reduce entanglement ofporpoises.
Several authors have commented on the possibilitythat acoustic pingers could function as an uninten-tional dinner bell (e.g. Mate & Harvey 1987, Dawson1991, Kraus 1999). Most of these cautions clearly hadpinnipeds in mind (Jefferson & Curry 1996), but bot-tlenose dolphins involved in depredation could easilyuse pingers to enhance their ability to find nets (e.g.Read et al. 2003). The behavioural flexibility charac-teristic of this species facilitates opportunistic forag-ing. It seems to us that these animals are likely to tol-erate even high-output pingers if there is a food‘reward’. We note also that high-output acousticpingers do not produce sounds that are nearly as loudas bottlenose dolphins’ own sounds (Au et al. 1974).We also note that there are several cases in whichdolphins have been entangled close to active pingersand that there are also reports of aggressive be -haviour by bottlenose dolphins towards pingers (seesubsection ‘Studies of behavioural responses topingers’).
REMAINING QUESTIONS
Will habituation result in a decrease in the effectiveness of pingers?
Habituation is usually thought of as a gradual less-ening of behavioural response to a repeated stimulusand seems to be a general feature in the animal king-dom. Several authors have raised the question ofwhether habituation will cause reduced reaction topingers over time and compromise effectiveness inreducing bycatch.
The most salient evidence for a lack of habituationcomes from temporal trends of bycatch rates in the 2longest running implementations of pingers in gillnetfisheries: the Gulf of Maine and California-Oregon.In neither case is there evidence of a long-termincrease in bycatch rates (as would reflect a diminu-
tion in their efficacy) when the recommended num-bers of functioning pingers are used on each string ofnets (Palka et al. 2008, Carretta & Barlow 2011).
Interestingly, however, several observational stud-ies of harbour porpoises have shown evidence ofhabituation in an experimental context. Cox et al.(2001) made theodolite observations of harbour por-poises in the vicinity of a moored Netmark 1000pinger and noted that the initial mean displacement(208 m) was reduced by half in 4 d, and by 11 d wasnot significantly different from the control. Cox et al.also noted that echolocation rate and occurrence,measured with a T-POD acoustic event recorder,were significantly lower when the pinger was activethan when it was silent. A broadly similar result wasobtained by Carlström et al. (2009), who used a simi-lar methodology but multiple T-PODs at differentranges from a simulated gillnet equipped with multi-ple Netmark 1000 pingers. Like Cox et al. theyobserved a clear initial displacement and lowerecholocation rate with active pingers. Echolocationrate near pingers, however, increased over time atsome locations, which was interpreted as evidence ofhabituation (Carlström et al. 2009). Following 2 sea-sons of trials of alarmed versus non-alarmed nets, allgillnets used in the 1997 trials by Gearin et al. (2000)were alarmed, in part to study habituation. No por-poises were taken in the first 18 d of the 1997 season,and 11 of the 12 porpoises taken were taken in thelast 2 wk of the fishery. While not conclusive, thissuggests that habituation may have occurred, withthe consequence of higher entanglement rate. Takentogether, these studies suggest that some degree ofhabituation to the sounds of pingers could occur ininshore areas where porpoises are at least seasonallyresident.
Habituation will only increase the risk of entan-glement if the resulting displacement is less thanhalf of the distance between adjacent pingers on thenets. Hence, reduction of displacement distance perse might not be a problem. The key issue is whetherhabituation results in approaches close enough toresult in entanglement. Further research is requiredto tease apart the effect of pingers on the fine-scalebehaviour of porpoises and other species aroundgillnets.
It seems likely that habituation would compromisethe efficacy of acoustic pingers used to deter interac-tions such as depredation, in which animals gain areward from ignoring the stimulus. Indeed, North-ridge et al. (2003) showed a decline in effectivenessof the pingers in reducing bottlenose dolphin depre-dation as the trial continued into a second month.
212
Dawson et al.: Pingers, small cetaceans, and gillnets 213
Why are pingers less effective in real fisheries thanin controlled experiments?
In both cases in which pingers have been formallyrequired as a mitigation strategy in commercial fish-eries, the real-world reductions in bycatch rate havebeen considerably less than those achieved under ex-perimental conditions. In the California-Oregon ex-periment common dolphin bycatch was reduced by82% in driftnets with pingers versus control nets (Bar-low & Cameron 2003). In the New Hampshire gillnetexperiment the corresponding reduction in bycatchrate of harbour porpoises was 92% (Kraus et al. 1997).After pinger use was mandated, entanglement ratesof common dolphins dropped by about 50% in Cali-fornia-Oregon driftnets, while those of harbour por-poises in New England gillnets dropped by 60%.
In both cases, several potentially confounding factorsmay have reduced apparent effectiveness. There havebeen shifts in the distribution of fishing effort in bothfisheries, and other management measures (includingtime-area closures and gear modifications) have beenimplemented to address the catch of both target andnon-target species. In addition, education and outreachprograms have varied in their scope and efficacy.
The single largest confounding factor, however,seems to be compliance with the regulations. Compli-ance in the New England fishery has been highly vari-able, and, as noted above, at times remarkably poor.In New England the proportion of hauls using the re-quired number of pingers fell from between 70 to 95%(depending on management area) in 1999 to 2000 to alow of 0 to 38% between 2003 and 2005. In 2007, afterworkshops explaining the need to use pingers, theseproportions climbed to approximately 50 to 80% (C. D.Orphanides & D. L. Palka unpubl. data). The lack ofcom pliance with the requirement to use pingers inboth fisheries (even when federal observers are aboarda fishing vessel) is a caution for fisheries in other areas.
It should be noted that there is a singular exceptionto the general observation that pingers are less effec-tive in real fisheries than in controlled experiments.Since pingers have been adopted as part of the man-agement strategy there have been no beaked whalestaken in the California-Oregon driftnet fishery (Car-retta et al. 2008). The likelihood of this having oc -curred by chance alone is vanishingly small.
How do pingers work?
Four main hypotheses have been proposed toexplain how pingers could reduce the bycatch rates
of small cetaceans (e.g. Dawson 1994, Kraus 1999,IWC 2000):(1) The sounds of acoustic pingers are generally
aversive and act to displace animals from thevicinity of the pinger.
(2) Pinger sounds encourage echolocation or other-wise alert the animals to the presence of the netand hence make avoidance of entanglementmore likely.
(3) Pinger sounds interfere with the animals’ sonar,causing them to leave the area.
(4) Pinger sounds act by altering the distribution ofprey.
Of these, only the ‘aversive’ hypothesis (1) hasstrong support and only for 1 species, the harbourporpoise. As described above, studies using theodo-lites to record surfacing positions of harbour por-poises have generally shown a clear zone of displace-ment around active pingers (e.g. Gearin et al. 2000,Culik et al. 2001; for an exception see Desportes et al.2006). Such displacement zones do not appear tooccur with bottlenose dolphins; indeed Cox et al.(2004) found no significant difference in the closestpoint of approach of dolphins to active or inactivepingers. No clear zone of exclusion was observed instudies of Hector’s dolphin (Stone et al. 1997) or oftucuxi (Monteiro-Neto et al. 2004). Pseudo-replica-tion problems in both these studies invalidate theirapparently significant statistical results (Dawson &Lusseau 2005).
The ‘alerting’ hypothesis (2) has also been studiedin harbour porpoise and bottlenose dolphins. Studiesof reactions to moored pingers, combining visualobservations with T-POD acoustic detections ofecholocation, have shown that both species usuallyecholocate less frequently in the vicinity of an activepinger than near an inactive one (Cox et al. 2001,Leeney et al. 2007, Carlström et al. 2009, for excep-tion see Desportes et al. 2006). Further, if pingerswork by stimulating echolocation and hence increas-ing the chance of net detection, one would expectavoidance only at ranges of 10s of metres, consistentwith studies of the detection ranges of gillnets bycaptive dolphins and porpoises (Kastelein et al.2000). Yet when pingers are avoided by harbour por-poises, the displacement ranges are typically on theorder of 100s of metres (Anderson et al. 2001, see ourTable 3). This implies that the pinger sounds them-selves are aversive.
We are not aware of any published studies specifi-cally addressing Hypothesis 3, whether pingers func-tion by ‘jamming’ echolocation or making it lesseffective. Most current pingers mimic the acoustic
Endang Species Res 19: 201–221, 2013
output of the discontinued Dukane Netmark 1000which proved effective at reducing bycatch in sev-eral experiments (Table 2). These pingers emit a10 kHz tone every 4 s that is relatively rich in har-monics, but there is very little energy near 130 kHz,the dominant frequency of harbour porpoise echo -location signals (Au et al. 1999). Thus, it seemsunlikely that these pingers act by making echo -location more difficult. We note, however, that somepingers marketed to reduce depredation are at leastintended to work by jamming echolocation (e.g.SaveWave ‘high-protect dolphin saver’).
Hypothesis 4 was raised by Kraus et al. (1997), whonoted significantly lower catch rates of Atlantic her-ring Clupea harengus, a primary prey species of har-bour porpoises, in the New Hampshire experiment.Thus, these authors reasoned, the lower porpoisecatch could have been due to indirect effects on theirprey. A springtime test of the same pingers in thesame fishery confirmed the effectiveness of pingers inreducing porpoise bycatch, but did not show signifi-cant effects on fish catches (catches in control and ex-perimental nets differed by <10%; Kraus & Brault1999). Likewise, Trippel et al. (1999) found no differ-ences in catches of harbour porpoise prey in nets withand without Dukane pingers. Tests of 3 differentpinger types (Lien, Dukane and PICE) in a commercialherring fishery showed that the Lien pinger (2.9 kHz)was associated with a significantly higher catch rateof herring, but that nets with the other 2 pingers hadvery similar catch rates to nets without pingers (Culiket al. 2001). These studies, in combination with Wilson& Dill’s (2002) finding that a Dukane pinger had noobservable effect on the behaviour of adult herring,indicate that effects on prey species are not responsi-ble for the observed effects of pingers on porpoises.
It is clear that the aversive hypothesis best explainsthe effects on harbour porpoises, but pingers in gill-nets have also significantly reduced catch rates ofcommon dolphins (Barlow & Cameron 2003), beakedwhales (Carretta et al. 2008), franciscana (Bordino etal. 2002) and, at least temporarily, striped dolphins(Imbert et al. 2007). It is possible that the aversivehypothesis might also explain these results, but atpresent there is no evidence to support or refute thishypothesis.
Under what circumstances will pingers be effective?
This question has 2 components — for which spe-cies and in which fisheries will pingers be effective?Pingers have been tested in the field on harbour
porpoises, franciscana, common dolphins, beakedwhales, Hector’s dolphins, tucuxi, striped dolphins,and bottlenose dolphins (see Tables 2, 3 & 4). Fourspecies (harbour porpoises, franciscana, striped andcommon dolphins, and 1 species group (beakedwhales) have shown unequivocal reductions in by -catch and/or clear avoidance of pinger sounds (Krauset al. 1997, Bordino et al. 2002, Barlow & Cameron2003, Carretta et al. 2008).
For other species, the evidence is somewhat con-tradictory. With bottlenose dolphins, Cox et al. (2004)and Waples et al. (2013) found that 2 types of pingersboth resulted in fewer animals approaching within100 m of the net. Some studies have shown at leasttemporary reduction in net damage by bottlenosedolphins when pingers are used (Brotons et al.2008b, Gazo et al. 2008). Other studies suggest noreduction in depredation (L. Rossi et al. unpubl. data)or a diminution of the effect over several weeks(Northridge et al. 2003). Entanglements in pingerednets in relatively small-scale trials (Northridge et al.2003, McPherson et al. 2004, Read & Waples 2010)suggest that pingers are not effective in reducingentanglement risk for this species. The apparentlyaggressive response of bottlenose dolphins to somepingers is disconcerting (McPherson et al. 2004).
There is no evidence that Hector’s dolphins arephysically displaced from moored Netmark 1000pingers (Stone et al. 1997, Dawson & Lusseau 2005).However, avoidance reactions were observed in 21 of32 nearby dolphin groups when this pinger wasimmersed from a drifting boat (Stone et al. 2000).Boat-based trials have the advantage of allowingclose observation, but they have important disadvan-tages, including the possibly confounding effect ofthe boat, and the potential for dolphins to be ‘star-tled’ (sensu Teilmann et al. 2006) by the sudden onsetof pinger sounds at close range. Furthermore, they donot mimic the behavioural context associated withactively fishing nets. Thus, boat-based trials mayprovide poor measures of responses to pingers. Evenif taken at face value, however, the reactions of Hec-tor’s dolphins were clearly less extreme than those ofharbour porpoises, which are typically displacedsome 100s of metres from functioning pingers.Acoustic pingers were used by some gillnet fishers inCanterbury, New Zealand, under a voluntary ‘Codeof Practice’ (SE Finfish Management Companyunpubl. data), but low levels of observer coveragemade it impossible to determine whether thisreduced entanglement rates. Additionally, compli-ance with the code of practice has been low in thisfishery (Dawson & Slooten 2005).
214
Dawson et al.: Pingers, small cetaceans, and gillnets
We conclude that it is unreasonable to expectpingers to work with all small cetaceans. For harbourporpoises, at least, pingers work by displacing ani-mals away from the net, so we recommend that awell-designed observational study of the behaviouralreactions to pingers be conducted, in as realistic a situation as feasible (e.g. Cox et al. 2001), beforeemploying pingers in full-scale field trials with anynovel species.
Can we predict which species pingers might workfor? Experience suggests that pingers are most likelyto be successful in reducing bycatch of species whichare generally neophobic or easily startled, such asharbour porpoises. In addition, low levels of philo -patry (e.g. Read & Westgate 1997) might reduceencounter rates with static fishing gear and hencedelay habituation, should it occur. On this basis, gen-erally we might expect phocoenid species to be moreappropriate targets for bycatch reduction via pingersthan coastal delphinids. The elimination (to date) ofbeaked whale bycatch in pinger-equipped Californ-ian driftnets suggests a very high effectiveness forthis group, which is also wary of vessels and has anoceanic distribution. Delphinids, particularly thoseshowing very flexible behaviour, coastal distributionand high site fidelity, such as bottlenose dolphins,would seem among the least appropriate candidatespecies.
In some cases it may be impractical, or even uneth-ical, to attempt to demonstrate the effectiveness ofpingers in reducing bycatch. This may be becausebycatch rates are very low, resulting in a require-ment for extremely large (and costly) sample sizes(Dawson et al. 1998), or because the mortality of evena few individuals in a controlled experimental trial isunwarranted or unacceptable (see Read 2010). Thelatter situation is certainly true with Critically Endan-gered species, such as the vaquita or Maui’s dolphin.Statistical power curves given by Dawson et al.(1998) provide some guidance on the sample sizesneeded to demonstrate a target level of reduction,given knowledge of the entanglement rate in netswithout pingers. Thus, efficacy will be very difficult(or impossible) to demonstrate for some species, nomatter how susceptible those species are to pingersounds.
In addition, pingers are likely to be employedeffectively only under a certain set of socio-economicconditions. Many small-scale fisheries in the devel-oping world are unlikely to have the economicresources to employ this mitigation approach. Evenwith an initial subsidy to purchase pingers and trainfishery participants, it is unlikely that the devices will
be maintained and used effectively long-term. And,as described in ‘Implementation in commercial fish-eries’, partial ensonification of gillnets may be worsethan not using pingers at all. Further, implementa-tion of pingers is not straightforward, even in valu-able fisheries in developed nations. A number ofEuropean trials have demonstrated problems withthe durability and functionality of several of theavailable pinger models, and the difficulty of inte-grating them into mechanized hauling and setting.Fewer operational issues appear to be encounteredwith smaller coastal vessels fishing a smaller numberof nets (Sea Fish Industry Authority 2005, Lunneryd2006).
In more affluent countries with consumer-focusedfish marketing, the development of accreditationschemes that purport to ensure sustainable pro -duction methods and management procedurescould provide a means of ensuring higher levels of compliance with measures to minimize cetaceanbycatch.
A wide variety of pingers has been produced(Table 1). Most use constant-frequency (CF) tones(with multiple harmonics). Several are modelled onthe acoustic output of the Dukane Netmark 1000(10 kHz, ca. 132 dB pulses lasting 300 ms repeatingevery 4 s). Several (e.g. Aquamark 200) use FMsweeps extending from audio to ultrasonic frequen-cies. Some broadcast randomized signals (e.g. FishtekBP154) or signals at randomized intervals (e.g. Save-Wave Dolphin Saver), in an attempt to minimize thechance of habituation. CF pingers have been testedin controlled experiments more often than any othertype. At least 12 trials have shown significant re -ductions in bycatch due to their use (Table 2). FMpingers have been associated with bycatch re -ductions in 3 trials (Table 2). It is not clear whetherrandomizing the broadcast signal is more or lesseffective.
Quality control in pinger manufacture needs toimprove. In addition to the reliability issues addressedearlier, there can be substantial variation in soundpressure level among pingers of the same brand andmodel. For example, 26 identical pingers, made by aleading manufacturer, and tested in August 2009before first use, varied from 139 to 156 dB (re 1 µPa at1 m; Dawson & Nowacek unpubl. data). The claimedsound pressure level (SPL) of these pingers was147.5 dB. Likewise, Kraus et al. (1995) tested a ran-dom sample of 25 Dukane pingers, finding that SPLat the stated dominant frequency of 10 kHz variedfrom 105 to 139 dB. An independent accreditationsystem might improve confidence in pinger quality.
215
Endang Species Res 19: 201–221, 2013
Could pingers exclude animals from importanthabitat?
If pingers are effective, will marine mammals bedisplaced into less favourable habitat? Studies of har-bour porpoises in areas where high-intensity AHDsare deployed have shown that this species can be dis-placed considerable distances by high levels ofsound (Johnston 2002, Olesiuk et al. 2002). Also, thevery clear exclusion zone shown for this species inmost studies of pingers (e.g. Gearin et al. 2000, Culiket al. 2001, Carlström et al. 2009) shows that muchlower sound levels can create a similar effect on asmaller scale. If this effect was permanent and if gill-nets were set in preferred habitats, this could denyanimals access to important habitat, especially if theanimals have small home ranges. Indeed, deliberateexclusion from an entire bay has been proposed bythe installation of a string of pingers across the ca.17 km wide entrance of Puck Bay, in Poland, to miti-gate bycatches of harbour porpoises in the bay,which has high levels of gillnet effort (ICES 2010).The success of this approach is as yet unknown.
At larger scales, however, displacement seemsunlikely to be problematic. For example, Larsen &Hansen (2000) estimated that pinger signals from theentire Danish gillnet fleet could potentially ensonify<1% of the porpoises’ habitat. Similarly, Northridgeet al. (2011) showed that if the EU regulation onpinger use was fully implemented in SW England,<1% of habitat would be lost if DDD-03 devices wereused and exclusion was assumed to be complete towithin a 2 km radius of each device. Given the largeranges of individual porpoises in these waters (e.g.Sveegaard et al. 2011) and the variety of habitatsavailable, it seems unlikely that exclusion fromimportant habitat will be a serious problem at a pop-ulation level. Of greater concern would be the imple-mentation of acoustic pingers to reduce bycatch of aneophobic species with very restricted ranges, par-ticularly if nets were set repeatedly in importanthabitat.
How many pingers are needed?
Acoustic pingers are expensive, particularly whenmany nets are fished. In addition, they can hinderhauling and setting of nets. There is, therefore, con-siderable interest among fishers in reducing thenumber of pingers to the minimum required to main-tain their effect (we note also that this may reducethe amount of sound unnecessarily broadcast into the
ocean, unless fewer but much louder devices areused). Several pinger types produce signals louderthan the 132 dB ‘standard’ (Table 1) to increase thedistances over which they may be effective.
Carlström et al. (2009) found that the mean effec-tive range of displacement of harbour porpoises by10 kHz Dukane pingers was 752 m. This is muchgreater than the pinger spacing used in most experi-ments (e.g. Kraus et al. 1997) and considerablygreater than the distance required under currentregulations in the Gulf of Maine (92 m). Likewise,Larsen & Krog (2007) found that increasing spacingbetween Aquamark 100 pingers to 455 m and even to585 m produced bycatch reductions of similar magni-tude to those found in other studies with less spacingbetween pingers.
These observations appear to conflict with theresults of large-scale operational pinger use in theGulf of Maine (see ‘Gulf of Maine bottom-set gillnetfishery’), which show that reliable bycatch reductionoccurs only if the usual (ca. 92 m) spacing betweenpingers is used. If fewer pingers are used thanrequired, or if some pingers are not effective due tofailure, entanglement rates increase. It is possiblethat when the animal can hear multiple pingers atirregularly spaced intervals, it interprets a large gapbetween functioning pingers as a potential escaperoute (IWC 2000, Palka et al. 2008).
We believe that this apparent conflict points to theneed for further quantitative evaluation of the effectof incomplete ensonification of gillnets under opera-tional conditions. Such an evaluation should include:measurements of received levels along and aroundgillnets equipped with full and partial numbers ofpingers, an assessment of the behavioural responseof porpoises to such sound fields, and a rigorousexamination of the effect of partial pinger coverageusing bycatch data collected by observer programs.We note that the sound field generated by pingerswill vary across locations (Shapiro et al. 2009), andbehavioural responses will not be uniform.
LESSONS FROM THE LAST DECADE
We draw the following conclusions from the pastdecade of implementation, experimentation andobservation.(1) Pingers have been shown to be effective in
reducing bycatch of harbour porpoises, beakedwhales, common dolphins and franciscana. Theiruse does not appear to reduce the bycatch ofbottle nose dolphins, nor is there strong evidence
216
Dawson et al.: Pingers, small cetaceans, and gillnets 217
that they are effective in deterring depredationby this species. At the present time, we do nothave enough information to assess their efficacywith other species.
(2) Effective implementation of pingers in gillnetfisheries is difficult and compliance is likely to bevariable, even in relatively sophisticated fisheriesin developed countries. Education, outreach andenforcement are all critical components of effec-tive implementation plans. Based on past experi-ences, voluntary compliance will be the excep-tion, rather than the rule, but may be fosteredwhen included as part of a certification scheme.Post-implementation monitoring is critical to as -sess temporal trends in compliance and efficacy.
(3) Consistency of use and pinger durability are veryimportant. Bycatch rates appear to increase whennets are only partially ensonified due to equip-ment failure or the use of fewer pingers than rec-ommended (Palka et al. 2008, Carretta & Barlow2011). Thus, seemingly minor lapses in compli-ance and pinger reliability may result in higherbycatch rates.
(4) In the 2 longest-running programs that employpingers, other mitigation approaches, such astime−area closures and gear modification, arealso employed. Schemes which rely entirely onpingers, such as those enacted in Europe, are ofunknown effectiveness.
(5) Pingers are unlikely to be adopted and/or usedappropriately unless their use is mandated.Where pingers have been shown to be effectivebut their use is not mandated by regulation (suchas in the Bay of Fundy, Canada — see Trippel etal. 1999), the devices are not used by fishermen.In Canterbury, New Zealand, they have beenused voluntarily, but only 28% of observed setsshowed compliance with deployment instruc-tions (Dawson & Slooten 2005).
(6) In programs implementing pinger use, ensuringcompliance and ongoing effectiveness are bothdifficult and expensive. In some cases the costof the monitoring required to demonstrateongoing effectiveness may exceed the value ofthe fishery. To date, the costs of most monitor-ing programs have been borne by governmentsupport — effectively a subsidy of fishing opera-tions. One notable exception is the NewZealand Conservation Services Programme,which is administered by the Department ofConservation and levies fishers for funds tosupport quantification and mitigation of theirimpacts (West et al. 1999).
(7) The requirement of a ‘hard’ target for bycatchreduction (e.g. total allowable bycatch) is critical(i.e. not just reduction per se, but reduction to x)for determining how pingers should be employedand how they should be integrated with othermanagement approaches. Without a quantitativegoal, it is impossible to assess efficacy. Setting atarget for bycatch reduction also has a strong sci-entific benefit, as it facilitates using power analy-ses to design experiments that can detect mean-ingful effects with appropriate statistical power(Dawson et al. 1998).
(8) Pinger reliability has improved, but failure ratesremain significant with some models. Severalfishery trials have been conducted to examinepracticality and reliability issues associated withusing pingers (e.g. Sea Fish Industry Authority2003, 2005, Cosgrove et al. 2005, Lunneryd 2006,Krog & Larsen 2007). Failure rates around 50% orhigher were reported for Aquamark and Save-wave pingers in short-term trials by Cosgrove etal. (2005). Failure rates are also significant in fish-eries with mandatory pinger use. In the Gulf ofMaine bottom gillnet fishery, 64% of pingers inuse in 2003 were not working when tested; in2006/2007 this improved to 13% (Palka et al.2008). In the California-Oregon driftnet fishery,nets sampled for failed pingers showed a failurerate of about 18% (Carretta & Barlow 2011). Highfailure rates and the difficulty of practical imple-mentation are important impediments to pingeradoption in many fisheries.
NOTE ADDED IN PROOF
Recent analyses of data from the New England gill-net fishery show that compliance with the require-ment to use pingers continues to be very poor.Despite considerable pressure to improve compli-ance (including the threat of large time-area closuresif by-catch targets were exceeded), data from inde-pendent observers show that only 62% of sets in2010–2012 had a full complement of functioningpingers on each string of nets (Orphanides 2012).Harbour porpoise bycatch levels in the Gulf of Mainehave exceeded the Potential Biological Removallevel in 6 of the past 7 yr, and bycatch rates werealmost twice the target level in 2010 to 2012(Orphanides 2012), reflecting a fundamental failureof fishermen to comply with these regulations. NewEngland fishers were involved in the development ofthe pinger program and there is clear evidence that
Endang Species Res 19: 201–221, 2013
pingers can reduce bycatch in this fishery to sustain-able levels when they are used appropriately (Palkaet al. 2008), but systematic non-compliance has nowlasted for more than a decade. Continued lack ofcompliance has led scientists on the Take ReductionTeam to conclude that it is time to look for alternativemitigation strategies in New England. In addition,we suggest that it would be useful to understandwhy compliance is so poor in this fishery comparedwith that in the California-Oregon drift net fishery.Such information could help design future efforts touse pingers as a mitigation strategy in other gillnetfisheries.
Acknowledgements. The bulk of this paper was writtenwhile S.M.D. was on sabbatical at Duke University, theLighthouse Field station (University of Aberdeen), and theSea Mammal Research Unit (University of St Andrews). Weare grateful to those agencies, and to Otago University, fortheir support. For conversations about pingers, and for send-ing us their papers we are also grateful to Jim Carretta, JayBarlow, Liz Slooten, Debi Palka, Scott Kraus, ChrisOrphanides, Ed Trippel, Geoff McPherson, Ron Kasteleinand Tara Cox. We also thank the many fishermen in Europeand North America who have worked with us to test thesedevices in their fisheries.
LITERATURE CITED
Alfaro Shigueto J (2010) Experimental trial of acousticalarms to reduce small cetacean bycatch by gillnetsin Peru. Available at www. ruffordsmallgrants. org/ rsg/projects/joanna_alfaro_shigueto_0 (accessed 15 Dec2012)
Anderson RC, Barlow J, Bowles AE (2001) Pingers areacoustic harassment devices. Working paper AND/INFO3 presented to ‘Mitigation of interactions between dol-phins and fisheries through the use of acoustic harass-ment devices: effectiveness, impact and possible alterna-tives’ workshop. ICRAM, Rome
Au WWL (1993) The sonar of dolphins. Springer, HeidelbergAu WWL, Floyd RW, Penner RH, Murchison AE (1974)
Measurement of echolocation signals of the Atlantic bot-tlenose dolphin, Tursiops truncatus Montagu, in openwaters. J Acoust Soc Am 56: 1280−1290
Au WWL, Kastelein R, Rippe T, Schooneman N (1999) Trans-mission beam pattern and echolocation signals of a har-bor porpoise (Phocoena phocoena). J Acoust Soc Am 106: 3699−3705
Barlow J, Cameron GA (2003) Field experiments show thatacoustic pingers reduce marine mammal bycatch in theCalifornia drift gillnet fishery. Mar Mamm Sci 19: 265−283
Berg Soto A, Marsh H, Everyingham Y, Parra G, Noad M(2009) The acoustic and surface behaviour of coastal dolphins in Queensland: implications for management.In: Marsh H (ed) Marine and tropical sciences research -facility milestone report. Available at www. helene marsh.com/ publications/ Unpublished/ Marsh MTSRF_ Final Report300609. pdf (accessed 16 Dec 2012)
Berggren P, Wade PR, Carlström J, Read AJ (2002) Potentiallimits to anthropogenic mortality for harbour porpoises inthe Baltic region. Biol Conserv 103: 313−322
Berrow S, Cosgrove R, Leeney RH, Obrien J, Mcgrath D,Dalgard J, Gall YL (2008) Effect of acoustic deterrents onthe behaviour of common dolphins (Delphinus delphis).J Cetacean Res Manag 10: 227−233
Bordino P, Albareda D, Palmerio A, Mendez M, Botta S(2002) Reducing incidental mortality of franciscana dol-phin Pontoporia blainvillei with acoustic warningdevices attached to fishing nets. Mar Mamm Sci 18: 833−842
Bordino P, Kraus S, Alibareda D, Baldwin K (2004) Acousticdevices help to reduce incidental mortality of the Fran-ciscana dolphin (Pontoporia blainvillei) in coastal gill-nets. IWC SC/56/Sm12: 1–10. International WhalingCommission, Cambridge
Brotons JM, Grau AM, Rendell L (2008a) Estimating theimpact of interactions between bottlenose dolphins andartisanal fisheries around the Balearic Islands. MarMamm Sci 24: 112−127
Brotons JM, Munilla Z, Grau AM, Rendell L (2008b) Dopingers reduce interactions between bottlenose dolphinsand nets around the Balearic Islands? Endang SpeciesRes 5: 301−308
Brownell RL Jr, Ralls K, Perrin WF (1989) The plight of the‘forgotten’ whales: it’s mainly the small cetaceans thatare now in peril. Oceanus 32: 5−11
Buscaino G, Buffa G, Sarà G, Bellante A and others (2009)Pinger affects fish catch efficiency and damage to bottomgillnets related to bottlenose dolphins. Fish Sci 75: 537−544
Carlström J, Berggren P, Dinnetz F, Borjesson P (2002) Afield experiment using acoustic alarms (pingers) toreduce harbour porpoise bycatch in bottom-set gillnets.ICES J Mar Sci 59: 816−824
Carlström J, Berggren P, Tregenza NJC (2009) Spatial andtemporal impact of pingers on porpoises. Can J FishAquat Sci 66: 72−82
Carretta JV, Barlow J (2011) Long-term effectiveness, fail-ure rates, and ‘dinner bell’ properties of acoustic pingersin a gillnet fishery. Mar Technol Soc J 45: 7−19
Carretta JV, Barlow J, Enriquez L (2008) Acoustic pingerseliminate beaked whale bycatch in a gillnet fishery. MarMamm Sci 24: 956−961
Cosgrove R, Browne D, Robson S (2005) Assessment ofacoustic deterrent devices in Irish gillnet and tangle netfisheries. Marine Technical Report: Project 05MT07,Bord Iascaigh Mhara (IBM), Irish Sea Fisheries Board,Dublin
Cox TM, Read AJ, Solow A, Tregenza NC (2001) Will har-bour porpoises (Phocoena phocoena) habituate topingers? J Cetacean Res Manag 3: 81−86
Cox TM, Read AJ, Swanner D, Urian K, Waples D (2004)Behavioral responses of bottlenose dolphins, Tursiopstruncatus, to gillnets and acoustic alarms. Biol Conserv115: 203−212
Cox TM, Lewison R, Žydelis R, Crowder LB, Safina C, ReadAJ (2007) Comparing effectiveness of experimental andimplemented bycatch reduction measures: the ideal andthe real. Conserv Biol 21: 1155−1164
Culik BM, Koscinski S, Tregenza N, Ellis GM (2001) Reac-tions of harbor porpoises Phocoena phocoena and her-ring Clupea harengus to acoustic alarms. Mar Ecol ProgSer 211: 255−260
218
Dawson et al.: Pingers, small cetaceans, and gillnets
Dawson SM (1991) Modifying gillnets to reduce entangle-ment of cetaceans. Mar Mamm Sci 7: 274−282
Dawson SM (1994) The potential for reducing entanglementof dolphins and porpoises with acoustic modifications togillnets. Rep Int Whal Comm 15(Spec Issue): 573−578
Dawson SM, Lusseau D (2005) Pseudoreplication problemsin studies of dolphin and porpoise reactions to pingers.Mar Mamm Sci 21: 175−176
Dawson SM, Slooten E (1993) Conservation of Hector’s dol-phins: the case and process which led to establishment ofthe Banks Peninsula Marine Mammal Sanctuary. AquatConserv 3: 207−221
Dawson SM, Slooten E (2005) Management of gillnetbycatch of cetaceans in New Zealand. J Cetacean ResManag 7: 59−64
Dawson SM, Read AJ, Slooten E (1998) Pingers, porpoisesand power: uncertainties with using pingers to reducebycatch of small cetaceans. Biol Conserv 84: 141−146
Desportes G, Amundin M, Larsen F, Bjørge A, Poulsen LR,Stenback J, Petersen N (2006) NIPPER: nordic interactivepinger for porpoise entanglement reduction. Final reportto Nordic Council of Ministers Fjord & Bælt, Kerteminde.www.kolmarden.com/PageFiles/1233/NIPPER %20 final%20 report.pdf
Fish JF, Vania JS (1971) Killer whale, Orcinus orca, soundsrepel white whales, Delphinapterus leucas. Fish Bull 69: 531−535
Fullilove J (1994) How to make a gillnet ‘pinger.’ Natl Fish-erman 75(1):29–30
Gazo M, Gonzalvo J, Aguilar A (2008) Pingers as deterrentsof bottlenose dolphins interacting with trammel nets.Fish Res 92: 70−75
Gearin PJ, Gosho ME, Laake JL, Cooke L, DeLong RL,Hughes KM (2000) Experimental testing of acousticalarms (pingers) to reduce bycatch of harbour porpoise,Phocoena phocoena, in the state of Washington.J Cetacean Res Manag 2: 1−9
Gönener S, Bilgin S (2009) The effect of pingers on harbourporpoise, Phocoena phocoena, bycatch and fishing effortin the turbot gill net fishery in the Turkish Black Seacoast. Turk J Fish Aquat Sci 157: 151−157
Hardy T, Tregenza N (2010) Can acoustic deterrent devicesreduce bycatch in the Cornish inshore gillnet fishery?Field studies. Cornwall Wildlife Trust, Truro
Hembree D, Harwood MB (1987) Pelagic gillnet modifica-tion trials in northern Australian seas. Rep Int WhalComm 37: 369−373
ICES (International Council for the Exploration of the Sea)(2010) Report of the study group for bycatch of protectedspecies (SGBYC), 1−4 February 2010, Copenhagen,Denmark. ICES CM 2010/ACOM: 25
Imbert G, Laubier L, Malan A, Gaertner JC, Dekeyser I(2007) La thonaille ou courantille volante: rapport final àla région Provence-Alpes-Côte D’azur. Rapport final auConseil Régional Provence-Alpes-Côte d’Azur. Centred’Océanologie de Marseille, Marseille
IWC (International Whaling Commission) (2000) Report ofthe scientific committee. Annex I. Report of the sub-com-mittee on small cetaceans. J Cetacean Res Manag2(Suppl): 235−257
Jefferson TA, Curry BE (1996) Acoustic methods of reducingor eliminating marine mammal−fishery interactions: Dothey work? Ocean Coast Manag 31: 41−70
Johnston DW (2002) The effect of acoustic harassmentdevices on harbour porpoises (Phocoena phocoena) in
the Bay of Fundy, Canada. Biol Conserv 108: 113−118Johnston DW, Woodley TH (1998) A survey of acoustic
harassment device (AHD) use in the Bay of Fundy, NB,Canada. Aquat Mamm 24: 51−61
Kastelein RA, Au WWL, De Haan D (2000) Detection dis-tances of bottom-set gillnets by harbour porpoises (Pho-coena phocoena) and bottlenose dolphins (Tursiops trun-catus). Mar Environ Res 49: 359−375
Kasuya T (1985) Fishery−dolphin conflict in the Iki Islandarea of Japan. In: Beddington J, Beverton RJH, LavigneDM (eds) Marine mammals and fisheries. George Allenand Unwin, London, p 253−272
Koschinski S, Culik B (1997) Deterring harbour porpoises(Phocoena phocoena) from gillnets: observed reactions topassive reflectors and pingers. Rep Int Whal Comm 47: 659−668
Kraus S (1999) The once and future ping: challenges for theuse of acoustic deterrents in fisheries. Mar Technol Soc J33: 90−93
Kraus S, Brault S (1999) A springtime field test of the use ofpingers to reduce incidental mortality of harbor por-poises in gillnets. IWC SC/51/SM/WP10, InternationalWhaling Commission, Cambridge
Kraus S, Read A, Anderson E, Baldwin K, Solow A, SpradlinT, Williamson J (1995) A field test of the use of acousticalarms to reduce incidental mortality of harbor porpoisesin gill nets. IWC SC47/SM17, International WhalingCommission, Cambridge
Kraus S, Read A, Anderson E, Baldwin K, Solow A, SpradlinT, Williamson J (1997) Acoustic alarms reduce incidentalmortality of porpoises in gill nets. Nature 388: 525
Krog C, Larsen F (2007) Anvendelse af pingere i dansk gar-nfiskeri—overvågning, håndtering og effekt [Use ofpingers in Danish gillnet fishery — monitoring, handlingand effectiveness]. Report to the European Commission.Danish Ministry of Food Agriculture and Fisheries, Copen-hagen (in Danish). http:// 2.naturerhverv. fvm. dk/ Files/Filer/ Fiskeri/Projektdatabase/ Kollektive %20 foranstaltninger/ 3704-2-05-0120_pingere_i_dansk_ garn fiskeri . pdf(accessed 16 Dec 2012)
Laake J, Rugh D, Baraff L (1998) Observations of harbor por-poise in the vicinity of acoustic alarms on a set gill net.NOAA Tech Memo NMFS-AFSC-84
Larsen F (1999) The effect of acoustic alarms on the bycatchof harbour porpoises in the Danish North Sea gill netfishery. IWC SC/51/SM41, International Whaling Com-mision Cambridge
Larsen F, Hansen JR (2000) On the potential effects of wide-spread use of pingers in the North Sea. IWC SC/-52/ -SM28, International Whaling Commission, Cambridge
Larsen F, Krog C (2007) Fishery trials with increased pingerspacing. Paper presented to the Scientific Committee ofthe International Whaling Commission. IWC SC/ 59/ -SM2, International Whaling Commission, Cambridge
Lauriano G, Fortuna CM, Moltedo G, Notarbartolo di SciaraG (2004) Interaction between common bottlenose dol-phin (Tursiops truncatus) and the artisanal fishery in Asi-nara Island National Park (Sardinia): assessment of catchdamage and economic loss. J Cetacean Res Manag 6: 165−173
Lauriano G, Caramanna L, Scarnò M, Andaloro F (2009) Anoverview of dolphin depredation in Italian artisanal fish-eries. J Mar Biol Assoc UK 89: 921−929
Le Berre N (2005) Impact de l’introduction de repulsifsacoustiques à cetaces (<<Pingers>>) sur la sécurité et les
219
Endang Species Res 19: 201–221, 2013220
conditions de travail à bord des filayeurs de manch-ouest. Institut Maritime de Prévention, Lorient
Leeney RH, Berrow S, Mcgrath D, O’Brien J, Cosgrove R,Godley BJ (2007) Effects of pingers on the behaviour ofbottlenose dolphins. J Mar Biol Assoc UK 87: 129−133
López BD (2006) Interactions between Mediterranean bot-tlenose dolphins (Tursiops truncatus) and gillnets off Sar-dinia, Italy. ICES J Mar Sci 63: 946−951
López BD, Mariño F (2012) A trial of acoustic harassmentdevice efficacy on free-ranging bottlenose dolphins inSardinia, Italy. Mar Freshw Behav Physiol 44:197–208
Lunneryd SG (2006) Trials with fishing nets equipped with‘pingers’. Fiskeriverket Swedish Board of Fisheries,Göteborg
Mate BR, Harvey JT (eds) (1987) Acoustical deterrents inmarine mammal conflicts with fisheries. Oregon SeaGrant report ORESU-W-86-001, Oregon State University,Corvallis, OR
McPherson GR, Ballam D, Stapley J, Peverell S and others(2004) Acoustic alarms to reduce marine mammalbycatch from gillnets in Queensland waters: optimisingthe alarm type and spacing. Proc Acoustics 2: 1−6
Monteiro-Neto C, Avila FSC, Alves TT Jr, Araujo DS andothers (2004) Behavioral responses of Sotalia fluviatilis(Cetacea, Delphinidae) to acoustic pingers, Fortaleza,Brazil. Mar Mamm Sci 20: 145−151
Morizur Y, Le Niliot P, Buanic M, Pianalto S (2009) Expéri-mentations de répulsifs acoustiques commerciaux sur lesfilets fixes à baudroies en mer d’Iroise. IFREMER, Issy-les-Moulineaux. Available at http:// archimer. ifremer. fr/doc/ 2009/ rapport-6864.pdf (accessed 19 Dec 2012)
Murray KT, Read AJ, Solow AR (2000) The use of time/areaclosures to reduce bycatches of harbour porpoises: les-sons from the Gulf of Maine sink gillnet fishery.J Cetacean Res Manag 2: 135−141
Northridge S, Tregenza N, Rogan E, Mackey M, HammondP (1999) A sea trial of acoustic pingers in Celtic Shelf gill-net fisheries. IWC SC/51/SM43, International WhalingCommission, Cambridge
Northridge S, Vernicos D, Raitsos-Exarchopolous D (2003)Net depredation by bottlenose dolphins in the Aegean: first attempts to quantify and to minimise the problem.IWC SC/55/SM25, International Whaling Commission,Cambridge
Northridge S, Kingston A, Murphy S, Mackay A (2008) Mon-itoring, impact and assessment of marine mammalbycatch. Final report to Defra, Project MF0736, Univer-sity of St. Andrews, Sea Mammal Research Unit, St.Andrews
Northridge S, Kingston A, Mackay A, Lonergan M (2011)Bycatch of vulnerable species: understanding the pro-cess and mitigating the impacts. Final report to Defra,Project MF1003. Available at http:// randd. defra. gov.uk/Document.aspx?Document=MF1003-FINALRevisedAugust 2011.pdf (accessed 19 Dec 2012)
Olesiuk PF, Nichol LM, Sowden MJ, Ford JKB (2002) Effectof the sound generated by an acoustic harassment deviceon the relative abundance and distribution of harbor por-poises (Phocoena phocoena) in Retreat Passage, BritishColumbia. Mar Mamm Sci 18: 843−862
Orphanides CD (2012) New England harbor porpoisebycatch rates during 2010-2012 associated with Conse-quence Closure Areas. US Department of Commerce,Northeast Fisheries Science Center Reference Docu-ment 12-19. Available at www. nefsc. noaa. gov/ nefsc/
publications/crd/ (accessed 16 Dec 2012Palka DL, Rossman MC, Vanatten AS, Orphanides CD
(2008) Effect of pingers on harbour porpoise (Phocoenaphocoena) bycatch in the US Northeast gillnet fishery.J Cetacean Res Manag 10: 217−226
Quick NJ, Middlemas SJ, Armstrong JD (2004) A survey ofantipredator controls at marine salmon farms in Scot-land. Aquaculture 230: 169−180
Read AJ (2008) The looming crisis: interactions betweenmarine mammals and fisheries. J Mammal 89: 541−548
Read AJ (2010) Conservation biology. In: Boyd IL, BowenWD, Iverson SJ (eds) Marine mammal ecology and con-servation. A handbook of techniques. Oxford UniversityPress, Oxford, p 340−359
Read AJ, Waples D (2010) A pilot study to test the efficacy ofpingers as a deterrent to bottlenose dolphins in the Span-ish mackerel gillnet fishery. Bycatch reduction of marinemammals in Mid-Atlantic fisheries. Final report, Project08-DMM-02, Duke University, Beaufort, SC
Read AJ, Westgate A (1997) Monitoring the movements ofharbour porpoises (Phocoena phocoena) with satellitetelemetry. Mar Biol 130: 315−322
Read AJ, Waples D, Urian KW, Swanner D (2003) Fine-scalebehaviour of bottlenose dolphins around gillnets. Proc RSoc Lond B Biol Sci 270: S90−S92
Read AJ, Drinker P, Northridge S (2006) Bycatch of marinemammals in the US and global fisheries. Conserv Biol 20: 163−169
Reeves RR, Read AJ, Notarbartolo-di-Sciara G (2001) Reportof the workshop on interactions between dolphins andfisheries in the Mediterranean: evaluation of mitigationalternatives. IWC SC/53/SM3, International WhalingCommission, Cambridge
Richter C, Hood C, Lien J (1999) Mitigation of harbour por-poise bycatch with active acoustic devices in a non-experimental gillnet fishery. IWC SC/51/SM3, Interna-tional Whaling Commission, Cambridge
Rocklin D, Santoni MC, Culioli JM, Tomasini JA, Pelletier D,Mouillot D (2009) Changes in the catch composition ofartisanal fisheries attributable to dolphin depredation ina Mediterranean marine reserve. ICES J Mar Sci 66: 699−707
Rojas-Bracho L, Reeves RR, Jaramillo-Legorreta A (2006)Conservation of the vaquita Phocoena sinus. MammalRev 36: 179−216
Sagarminaga R, Cañadas A, Brotons JM (2006). Initiativesabout fisheries−cetaceans interactions in Spanish Medi-terranean waters. IWC SC/58/SM13, International Whal-ing Commission, Cambridge
Sea Fish Industry Authority (2003) Trial of acoustic deter-rents ('porpoise pingers') for prevention of porpoise (Pho-coena phocoena) bycatch — Phase 1 Deployment trial.Sea Fish Industry Authority, Hull. Available at www.seafish. org/ media/ Publications/ CR201.pdf (accessed 19Dec 2012)
Sea Fish Industry Authority (2005) Trial of acoustic deterrent('porpoise pingers') for prevention of porpoise (Phocoenaphocoena) bycatch. Sea Fish Industry Authority, Hull(accessed 16 Dec 2012)
Secchi ER, Wang JR (2002) Assessment of the conservationstatus of a franciscana (Pontoporia blainvillei) stock inthe franciscana management area III following the IUCNred list process. Lat Am J Aquat Mamm 1: 183−190
Shapiro AD, Tougaard J, Jørgensen PB, Kyhn LA and oth-ers (2009) Transmission loss patterns from acoustic
Dawson et al.: Pingers, small cetaceans, and gillnets
harassment and deterrent devices do not always followgeometrical spreading predictions. Mar Mamm Sci 25: 53−67
Slooten E, Fletcher D, Taylor BL (2000) Accounting foruncertainty in risk assessment: case study of Hector’sdolphin mortality due to gillnet entanglement. ConservBiol 14: 1264−1270
Slooten E, Dawson SM, Rayment WJ, Childerhouse SJ(2006) A new abundance estimate for Maui’s dolphin: What does it mean for managing this critically endan-gered species? Biol Conserv 128: 576−581
SMRU (Sea Mammal Research Unit) (2001) Reduction ofporpoise bycatch in bottom set gillnet fisheries. Reportto the European Commission Study Contract 97/095,SMRU, University College Cork, Cornish Fish Producer’sOrganisation, and Irish South and West Fishermen’sOrganisation, Cork
Stone GS, Kraus S, Hutt A, Martin S, Yoshinaga A, Joy L(1997) Reducing bycatch: Can acoustic pingers keepHector’s dolphins out of fishing nets? Mar Technol Soc J31: 3−7
Stone GS, Cavagnaro L, Hutt A, Kraus S, Baldwin K, BrownJ (2000) Reactions of Hector’s dolphins to acoustic gillnetpingers. New Zealand Department of ConservationTechnical Report Series, Wellington
Sveegaard S, Teilmann J, Tougaard J, Dietz R, MouritsenKN, Desportes G, Siebert U (2011) High density areas forharbor porpoises (Phocoena phocoena) identified bysatellite tracking. Mar Mamm Sci 27: 230−246
Teilmann J, Tougaard J, Miller LA, Kirketerp T, Hansen K,Brando S (2006) Reactions of captive harbor porpoises
(Phocoena phocoena) to pinger-like sounds. Mar MammSci 22: 240−260
Trippel EA, Strong MB, Terhune JM, Conway JD (1999) Mitigation of harbour porpoise (Phocoena phocoena)bycatch in the gillnet fishery in the lower Bay of Fundy.Can J Fish Aquat Sci 56: 113−123
Trippel EA, Holy NL, Palka DL, Shepherd TD, Melvin GD,Terhune JM (2003) Nylon barium sulphate gillnetreduces porpoise and seabird mortality. Mar Mamm Sci19: 240−243
van Marlen B (2007) Nephrops and cetacean species selec-tion information and technology. NECESSITY Final Pub-lishable Activity Report. IMARES, Wageningen. Avail-able at http://news.ices.dk/sites/eurepository/SharedDocuments/FP6_PROTECT/PROTECT Publishable finalreport.pdf (accessed 11 Jan 2013)
Waples DM, Thorne LH, Hodge LEW, Burke EK, Urian KW,Read AJ (2013) A field test of acoustic deterrent devicesused to reduce interactions between bottlenose dolphinsand a coastal gillnet fishery. Biol Conserv 157:165–171
West IF, Molloy J, Donoghue MF, Pugsley C (1999) Seabirdand marine mammal bycatch reduction through fishingindustry funded research: the New Zealand Conserva-tion Services Levy program. Mar Technol Soc J 33: 13−18
Wilson B, Dill LM (2002) Pacific herring respond to simu-lated odontocete echolocation sounds. Can J Fish AquatSci 59: 542−553
Zahri Y, Abid N, Elouamari N, Abdellaoui B (2004) Étude del’interaction entre le grand dauphin et la pêche à lasenne coulissante en Méditerranée marocaine. l’InstitutNational de Recherche Halieutique (INRH), Nador
221
Editorial responsibility: William Perrin,La Jolla, California, USA
Submitted: June 6, 2012; Accepted: September 6, 2012Proofs received from author(s): December 21, 2012