Content uploaded by Jochen Depestele
Author content
All content in this area was uploaded by Jochen Depestele on Jun 01, 2016
Content may be subject to copyright.
Fisheries
Research
175
(2016)
35–42
Contents
lists
available
at
ScienceDirect
Fisheries
Research
j
ourna
l
ho
me
page:
www.elsevier.com/locate/fishres
Ray
bycatch
in
a
tropical
shrimp
fishery:
Do
Bycatch
Reduction
Devices
and
Turtle
Excluder
Devices
effectively
exclude
rays?
Tomas
Willemsa,b,∗,
Jochen
Depestelea,
Annelies
De
Backera,
Kris
Hostensa
aInstitute
for
Agricultural
and
Fisheries
Research
(ILVO),
Bio-Environmental
Research
group,
Ankerstraat
1,
8400
Oostende,
Belgium
bDepartment
of
Biology,
Marine
Biology
Section,
Ghent
University
Krijgslaan
281
S8,
9000
Gent,
Belgium
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
15
January
2015
Received
in
revised
form
3
November
2015
Accepted
11
November
2015
Keywords:
Bycatch
Rays
Tropical
shrimp
trawling
Turtle
Excluder
Device
Bycatch
Reduction
Device
a
b
s
t
r
a
c
t
Worldwide,
many
species
of
elasmobranchs
(Chondrichthyes:
Elasmobranchii)
are
currently
threatened
by
marine
fisheries
activity
and
are
on
the
Red
List
of
the
International
Union
for
Conservation
of
Nature
(IUCN).
Although
Bycatch
Reduction
Devices
(BRDs)
for
teleost
fish
and
Turtle
Excluder
Devices
(TEDs)
are
now
widespread
in
tropical
shrimp
trawling,
information
on
their
ability
to
mitigate
bycatch
of
elasmo-
branchs,
particularly
rays
(Batoidea),
is
scarce
and
limited
to
only
a
few
isolated
fisheries.
The
objective
of
this
study
was
to
evaluate
the
potential
of
trawls
fitted
with
a
square-mesh
panel
BRD
and
super-
shooter
TED
in
reducing
ray
bycatch.
In
this
study,
65
catch-comparison
hauls
were
conducted
in
the
Atlantic
seabob
shrimp
(Xiphopenaeus
kroyeri)
fishery
off
Suriname.
Trawls
with
a
BRD
and
TED
combi-
nation
reduced
ray
catch
rate
by
36%.
A
21%
reduction
in
mean
size
indicated
the
preferential
exclusion
of
large
rays.
Hence,
high
escape
ratios
were
observed
for
Dasyatis
geijskesi
(77%),
a
large-sized
species,
while
exclusion
of
the
small
species
Urotrygon
microphthalmum
was
not
significant,
although
their
disc
width
is
small
enough
to
pass
through
the
meshes
of
the
BRD.
Furthermore,
a
size-dependent
escape
for
the
two
most
abundant
mid-sized
ray
species
Dasyatis
guttata
and
Gymnura
micrura
was
observed.
Exclusion-at-size
differed
for
both
species,
however,
likely
related
to
species-specific
morphology
or
behavior
in
response
to
the
TED.
This
study
shows
that
the
combination
of
BRD
and
TED
causes
an
impor-
tant
reduction
in
ray
bycatch
in
seabob
shrimp
fisheries
off
Suriname.
The
great
reduction
in
catch
of
large-sized
rays
is
positive,
but
the
mortality
of
juvenile
rays
is
likely
to
have
negative
consequences
for
their
populations.
We
therefore
recommend
gear-based
and
non-gear
adaptations
to
further
reduce
the
bycatch
of
small-sized
rays.
©
2015
Elsevier
B.V.
All
rights
reserved.
1.
Introduction
Concern
has
been
increasing
recently
regarding
the
capture
and
mortality
of
elasmobranchs
in
marine
fisheries
(Stevens
et
al.,
2000).
In
contrast
to
most
teleost
fish,
elasmobranchs
are
gener-
ally
slow-growing
and
long-lived,
with
late
attainment
of
sexual
maturity,
low
fecundity
and
low
natural
mortality
(e.g.,
Fisher
et
al.,
2013;
Goodwin
et
al.,
2002).
This
K-selected
life-history
strategy
makes
them
particularly
vulnerable
to
exploitation
in
fisheries,
implying
that
overfished
populations
have
a
low
ability
to
recover
(Graham
et
al.,
2001).
Several
species
of
elasmobranchs
have
been
∗Corresponding
author
at:
Institute
for
Agricultural
and
Fisheries
Research
(ILVO),
Bio-Environmental
Research
group,
Ankerstraat
1,
8400
Oostende,
Belgium.
Fax:
+32
59
330629.
E-mail
addresses:
tomas.willems@ilvo.vlaanderen.be,
tomaswillems@gmail.com
(T.
Willems),
jochen.depestele@ilvo.vlaanderen.be
(J.
Depestele),
annelies.debacker@ilvo.vlaanderen.be
(A.
De
Backer),
kris.hostens@ilvo.vlaanderen.be
(K.
Hostens).
decimated
and
even
brought
to
the
brink
of
local
extinction
due
to
fishing
activity
(Baum
et
al.,
2003;
Dulvy
et
al.,
2000;
Dulvy
and
Reynolds,
2002).
Elasmobranchs
are
also
often
of
low
economic
value
in
fisheries
that
target
teleost
fish
or
invertebrates,
and
are
hence
discarded
as
unwanted
bycatch
(Stevens
et
al.,
2000).
Fur-
thermore,
elasmobranch
discards
often
remain
unreported
(Worm
et
al.,
2013),
resulting
in
insufficient
information
on
their
occur-
rence
and
population
sizes
worldwide.
This
is
a
major
impediment
for
effective
conservation
measures
(Bonfil,
1994;
Stevens
et
al.,
2000).
Many
species
of
elasmobranchs
are
known
to
occur
as
bycatch
in
tropical
shrimp
trawling
(Shepherd
and
Myers,
2005;
Simpfendorfer,
2000).
Nonetheless,
efforts
to
reduce
bycatch
in
shrimp
trawls
have
so
far
focused
mainly
on
teleost
fish
and
sea
turtles
through
the
development
of
Bycatch
Reduction
Devices
(BRDs)
and
Turtle
Excluder
Devices
(TEDs)
(Broadhurst,
2000).
Sev-
eral
types
of
BRDs
have
proven
to
cause
significant
reductions
in
the
bycatch
of
non-commercial
teleost
fish
(e.g.,
Broadhurst,
2000;
Heales
et
al.,
2008;
Rogers
et
al.,
1997;
Rulifson
et
al.,
1992).
TEDs,
http://dx.doi.org/10.1016/j.fishres.2015.11.009
0165-7836/©
2015
Elsevier
B.V.
All
rights
reserved.
36
T.
Willems
et
al.
/
Fisheries
Research
175
(2016)
35–42
on
the
other
hand,
are
highly
effective
in
reducing
sea
turtle
bycatch
(Eayrs,
2007,
2012;
Robins
and
McGilvray,
1999).
Moreover,
they
act
as
sorting
grids,
and
exclude
any
organism
larger
than
the
TED’s
bar
spacing
(typically
10
cm)
from
the
trawl,
including
large-sized
elasmobranchs
(Brewer
et
al.,
2006,
1998;
Griffiths
et
al.,
2006).
In
the
Atlantic
seabob
shrimp
(Xiphopenaeus
kroyeri)
fishery
off
Suriname,
trawls
are
required
by
law
to
be
equipped
with
two
widely-used
devices:
square-mesh
panel
BRD
and
super-shooter
TED.
In
this
fishery,
these
trawl
adaptations
have
proven
effective
in
reducing
bycatch
of
non-target
teleost
fish
(Polet
et
al.,
2010)
and
sea
turtles
(S.
Hall,
pers.
comm.),
respectively.
Average
bycatch
lev-
els
have
now
been
reduced
to
20–30%
of
the
total
catch
by
weight,
and
most
bycatch
species
in
this
fishery
are
assumed
to
be
within
safe
biological
limits
(Polet
et
al.,
2010;
Southall
et
al.,
2011).
These
efforts
have
contributed
to
the
certification
of
the
Suriname
seabob
shrimp
fishery
by
the
Marine
Stewardship
Council
(MSC)
in
2011.
Nevertheless,
the
MSC
assessment
team
raised
particular
concerns
over
mortality
of
rays
(Elasmobrachii:
Batoidea),
which
were
iden-
tified
as
the
most
vulnerable
bycatch
species.
Ray
bycatch
remains
a
key
issue
to
be
tackled
by
the
fishery
in
order
to
pass
future
MSC
reassessments
(Southall
et
al.,
2011).
The
Suriname
seabob
shrimp
fishery
is
known
to
capture
sev-
eral
ray
species
which
are
globally
endangered
and
are
listed
on
the
IUCN
Red
List
of
Threatened
Species,
including
Dasyatis
gei-
jskesi
and
Rhinoptera
bonasus
(‘near
threatened’),
Dasyatis
guttata
and
Gymnura
micrura
(‘data
deficient’)
(IUCN,
2015).
Because
these
species
commonly
grow
to
80–100
cm
disc
width
(Léopold,
2005),
we
could
expect
them
to
escape
through
the
TED.
A
fifth
frequently
caught
ray
species,
Urotrygon
microphthalmum
(‘least
concern’;
IUCN,
2015)
is
much
smaller
with
a
maximum
disc
width
of
25
cm
(Léopold,
2005),
and
might
escape
through
the
square-mesh
panel
BRD
because
of
its
small
size.
On
the
other
hand,
due
to
their
flat-
tened
body
shape
and
high
flexibility,
even
large
rays
might
still
be
able
to
pass
between
the
bars
of
a
TED
and
end
up
in
the
codend.
With
the
exception
of
very
small
rays,
their
size
and
morphology
would
also
prevent
escape
through
the
BRD.
It
remains
unclear
how
frequently
these
rays
occur
in
the
bycatch
of
this
fishery,
and
to
what
degree
the
current
trawl
adaptations
(i.e.,
BRD
and
TED)
reduce
their
capture.
In
the
present
study,
we
have
assessed
the
effectiveness
of
the
combination
of
BRD
and
TED
in
reducing
bycatch
of
rays
in
the
Atlantic
seabob
shrimp
fishery
off
the
coast
of
Suriname.
We
present
the
results
of
a
catch-comparison
study
in
which
we
have
focused
on
ray
bycatch
and
analyzed
ray
catches
in
trawls
with
and
without
the
combination
of
BRD
and
TED.
The
aims
were
to
assess
whether
these
devices
are
effective
in
excluding
rays
from
the
trawls,
and
whether
exclusion
of
rays
is
species-
and
size-
dependent.
2.
Materials
and
methods
2.1.
Study
area
The
study
was
conducted
on
commercial
fishing
grounds
for
seabob
shrimp
(6.17◦N
to
6.25◦N
and
55.39◦W
to
55.84◦W)
on
the
continental
shelf
off
Suriname
(FAO
Statistical
area
31).
This
area
is
characterized
by
mud
and
sandy
mud
substrates
and
water
depth
is
typically
20–25
m
(Fig.
1).
Commercial
shrimp
fishing
activity
occurs
year-round
in
this
area.
2.2.
Gear
specifications
Hauls
were
done
onboard
FV
Neptune-6,
a
typical
20-m,
425-hp
‘Florida-type’
outrigger
trawler
used
in
the
seabob
shrimp
trawling
fleet.
The
vessel
was
equipped
for
quad-rig
bottom-trawling,
which
involves
dragging
two
trawls
attached
to
two
steel-footed
wooden
doors
and
a
sledge
at
either
side
of
the
vessel,
resulting
in
two
port-
and
two
starboard-codends.
Mesh
size
of
each
trawl
was
57
mm
in
the
body
and
wings
of
the
trawl
and
45
mm
in
the
codend.
Each
trawl
was
fitted
with
an
aluminum
super-shooter
TED.
Bar
spacing
was
10
cm
and
each
was
installed
in
a
downward-excluding
con-
figuration
in
an
angle
of
approximately
50◦from
the
horizontal.
A
single
net
flap
covered
each
bottom
escape
opening,
and
there
was
no
guiding
funnel
in
front
of
the
TED.
Each
trawl
was
also
fitted
with
a
square-mesh-panel
(11
×
11
meshes,
15
cm
stretched
mesh
size)
BRD
inserted
ca.
40
cm
behind
the
TED
in
the
upper
side
of
the
codend
(Fig.
2).
2.3.
Sea
trials
and
catch
sampling
A
total
of
65
experimental
catch-comparison
hauls
were
con-
ducted
on
eight
commercial
seabob
fishing
trips
between
February
2012
and
April
2013.
During
each
trip,
seven
to
ten
experimen-
tal
hauls
were
conducted
to
compare
ray
bycatch
in
trawls
with
a
BRD
and
TED
combination
(‘wBT
net’)
versus
trawls
without
a
BRD
and
TED
combination
(‘noBT
net’).
In
the
noBT
net,
both
codends
with
BRD
and
TED
were
removed
and
replaced
by
codends
with-
out
any
devices.
The
side
of
the
vessel
dragging
the
wBT
and
noBT
net
was
alternated
every
trip
to
exclude
port
and
starboard
effects.
Hauls
were
done
under
commercial
fishing
circumstances,
except
for
a
shortened
dragging
time
(avg.
1h16±SD
0h16versus
3–4
h
normal
dragging
time),
to
reduce
the
risk
of
injury
or
mortality
of
vulnerable
species
in
the
noBT
net.
Although
the
fishery
normally
operates
day
and
night,
experimental
hauls
were
done
during
day-
time
only
for
practical
reasons.
The
wBT
net
and
noBT
net
were
dragged
alongside
each
other
at
a
speed
of
2.5–3.5
knots,
in
accor-
dance
with
normal
fishing
practice
(Pérez,
2014).
To
ensure
that
the
catches
from
the
wBT
and
noBT
nets
remained
separate,
the
two
wBT
codends
were
unloaded
separately
from
the
two
noBT
codends
on
deck.
Per
net,
the
catch
from
the
two
codends
was
combined.
All
rays
were
sorted
out
from
the
catches,
identified
to
species
level
and
measured
(disc
width)
to
the
nearest
centimeter.
The
catch
was
subsequently
processed
as
usual
by
the
crew
and
could
not
be
analyzed
further
for
practical
reasons.
2.4.
Data
analysis
Ray
catches
were
recalculated
to
a
standardized
catch
rate
(indi-
viduals
h−1).
Differences
in
mean
catch
rate
between
the
wBT
and
noBT
net
were
analyzed
using
Wilcoxon
signed
rank
tests.
Differ-
ences
in
mean
ray
size
between
wBT
and
noBT
net
were
analyzed
with
Mann–Whitney
U
tests.
Both
analyses
were
done
per
ray
species
and
for
all
rays
combined.
Differences
in
mean
size
among
ray
species
were
tested
using
the
Kruskal–Wallis
test
and
Nemenyi-post-hoc
pairwise
compar-
isons
(Pholert,
2014).
For
these
analyses,
only
data
from
noBT
net
catches
were
used
because
size-selection
was
expected
in
the
wBT
net.
Non-parametric
tests
were
used
because
the
assumptions
for
(paired)
t-tests
and
ANOVA
were
not
met.
The
relationship
between
ray
size
and
escape
from
the
trawls
was
explored
using
Generalized
Linear
Mixed
Models
(GLMM).
To
do
so,
size
classes
(originally
1
cm)
were
lumped
and/or
hauls
with
sufficient
individuals
per
size
class
were
selected
to
obtain
enough
data-points
per
size
class.
The
proportion
retained
by
the
wBT
net
at
size
class
S
can
be
expressed
for
each
size
class
and
each
haul
as:
(S)=NS,wBT
NS,wBT +
NS,noBT
T.
Willems
et
al.
/
Fisheries
Research
175
(2016)
35–42
37
Fig.
1.
Study
area
with
location
of
the
experimental
hauls.
Fig.
2.
Sketch
of
a
wBT
trawl
codend
fitted
with
Bycatch
Reduction
Device
(BRD)
and
super-shooter
TED.
where
(S)
is
the
probability
of
catching
an
individual
at
size
class
S
in
the
wBT
net.
NS,wBT and
NS,noBT are
the
number
of
rays
at
size
class
S
measured
for
the
wBT
net
(with
a
BRD
and
TED
combina-
tion
in
both
trawls)
and
the
noBT
net
(without
BRDs
and
TEDs),
respectively.
A
value
of
=
0.5
indicates
that
there
are
no
differ-
ences
in
catch
in
numbers
between
the
two
nets
at
size
class
S.
The
catch-at-size
proportion
(S)
was
modeled
using
the
GLMM
with
binomial
distribution
and
logit
link
function,
according
to
the
method
described
by
Holst
and
Revill
(2009).
The
expected
pro-
38
T.
Willems
et
al.
/
Fisheries
Research
175
(2016)
35–42
portion
of
the
catch
retained
by
the
wBT
net
at
size
class
S
was
expressed
as:
logit [(S)] =
ˇ0+
ˇ1S1+
ˇ2S22
where
ˇ0is
the
intercept
coefficient,
ˇ1and
ˇ2the
model
coef-
ficients
for
respectively
the
linear
and
quadratic
effects
of
the
explanatory
variable
‘size
class
S’.
The
catch
comparison
curves
vary
among
hauls,
potentially
in
a
size-specific
manner.
In
addition
to
the
fixed
effects,
inter-haul
correlation
was
incorporated
into
the
models
by
the
inclusion
of
random
intercept
and/or
slope
effects
(Venables
and
Dichmont,
2004).
Escape-at-size
was
modeled
for
all
ray
species
combined
and
for
species
frequently
caught,
i.e.,
present
in
≥20
hauls
with
a
min-
imum
of
20
individuals.
This
was
the
case
for
D.
guttata
and
G.
micrura.
Size
classes
of
10
cm
were
used
to
make
a
model
of
all
ray
species
combined
over
a
large
size
range
(20–90
cm).
For
D.
guttata
and
G.
micrura
a
finer
resolution
(3-cm
size
classes)
were
used
in
a
more
restricted
size
range
based
on
24
and
25
hauls
with
>20
individuals
per
haul,
respectively.
The
D.
guttata
model
was
fitted
between
20
and
72
cm
and
the
G.
micrura
model
between
18
and
57
cm.
All
analyses
were
carried
out
using
R
statistical
environment
(R
Core
Team,
2013).
3.
Results
Rays
were
caught
in
every
experimental
haul
performed.
A
total
of
3181
individuals
were
captured,
comprising
of
five
different
species.
Smooth
butterfly
ray
(G.
micrura)
and
Longnose
stingray
(D.
guttata)
were
the
most
abundant
species,
contributing
45%
and
37%
to
the
total
ray
catch
by
number,
respectively.
Smalleyed
round
stingray
(U.
microphthalmum;
11%),
Sharpsnout
stingray
(D.
geijskesi;
6%)
and
Cownose
ray
(R.
bonasus;
1%)
were
less
abun-
dant
(Fig.
3).
Mean
catch
rate
of
rays
in
the
noBT
net
ranged
from
6.3
±
3.1
to
45
±
19.6
ind.
h−1(average
±
SD
of
May
resp.
April
2012)
corresponding
to
a
mean
density
of
0.6
±
0.3
to
4.3
±
1.9
rays
ha−1
trawled
in
the
study
area.
Overall,
mean
catch
rate
of
rays
(over
all
hauls)
was
significantly
reduced
by
36.1%
in
the
wBT
net
(15.3
±
13.2
ind.
h−1)
compared
to
the
noBT
net
(23.9
±
19.2
ind.
h−1;
p
<
0.001).
Significant
reduction
in
catch
rate
in
the
wBT
net
was
observed
for
D.
geijskesi
(−76.6%),
D.
guttata
(−40.2%)
and
G.
micrura
(−32.1%;
all
p
<
0.001).
Catch
rate
reductions
in
R.
bonasus
and
U.
microphthalmum
were
not
signifi-
cant
(Fig.
3).
Size
of
rays
captured
during
the
experiment
ranged
from
3
to
116
cm
with
a
mean
of
29.6
±
16.8
cm.
Mean
sizes
of
rays
caught
in
the
noBT
net
were
statistically
different
among
species
(2(4)
=
737.2;
p
<
0.001).
Post-hoc
tests
revealed
that
all
species
differed
significantly
in
mean
size
(p
<
0.001)
except
for
R.
bonasus,
which
did
not
differ
from
any
other
species
(Fig.
4).
Rays
caught
in
the
wBT
net
(avg.
25.5
±
12.4
cm)
were
on
average
20.6%
smaller
than
rays
caught
in
the
noBT
net
(avg.
32.2
±
18.6
cm;
p
<
0.001).
Size
reduction
in
the
wBT
net
was
significant
for
D.
geijskesi
(37.8%;
p
<
0.001)
and
D.
guttata
(22.7%;
p
<
0.001)
(Fig.
4).
The
modeled
proportion
of
rays
retained
by
the
wBT
net
was
always
<0.5,
indicating
an
overall
exclusion
from
the
wBT
net.
Fur-
thermore,
the
proportion-at-size
of
rays
caught
in
the
wBT
net
was
size-
and
species-dependent
(Fig.
5).
Catch
rate
of
all
species
com-
bined
declined
with
increasing
size,
following
a
quadratic
curve
in
the
modeled
size-range.
Total
exclusion
from
the
wBT
net
was
approached
at
90
cm
disc
width
(Fig.
5;
Table
1).
A
similar
response
was
found
for
D.
guttata,
although
the
curve
was
steeper,
reaching
total
exclusion
near
50
cm.
Catch
rate
reduction
for
G.
micrura
was
linear
and
did
not
approach
zero
in
the
modeled
size-range
(Fig.
5;
Table
1).
Table
1
Coefficient
values
and
significance
(p-value)
from
generalised
linear
mixed
mod-
elling
(GLMM)
of
the
proportion
()
of
the
catch
excluded
by
the
wBT
net
in
relation
to
size
(S),
where
logit
[(S)]
=
ˇ0+
ˇ1S
+
ˇ2S2×
ˇ0=
intercept,
ˇ1=
size,
ˇ2=
size2.
Species
Parameter
Estimate
SE
p-value
All
ray
species
ˇ2−0.0006
0.0001
<0.001
Dasyatis
guttata ˇ10.0700
0.0261
0.0073
ˇ2−0.0035
0.0012
0.0022
Gymnura
micrura
ˇ1−0.0145
0.0057
0.0104
4.
Discussion
The
combined
use
of
BRD
and
TED
in
the
Suriname
seabob
shrimp
fishery
caused
a
significant
36.1%
reduction
in
the
over-
all
catch
rate
of
rays.
In
one
of
the
few
other
studies
that
quantified
the
effect
of
BRDs
and
TEDs
on
ray
catch
rate
(Brewer
et
al.,
2006),
a
remarkably
similar
36.3%
reduction
in
Australia’s
northern
prawn
trawl
fishery
was
found.
This
reduction
was
assigned
to
the
effect
of
the
TED,
as
no
significant
reduction
in
ray
bycatch
was
found
in
trawls
exclusively
equipped
with
a
BRD
(bigeye
or
square-mesh
panel).
Although
the
effect
of
BRD
and
TED
cannot
be
evaluated
sep-
arately
in
the
present
study,
the
observed
catch-rate
reductions
are
likely
to
be
caused
by
the
TED
rather
than
the
BRD
for
the
following
reasons:
No
significant
reduction
was
observed
for
U.
microphthal-
mum,
the
only
species
which
could
theoretically
escape
through
the
meshes
of
the
BRD
due
to
its
small
size.
Moreover,
rays
caught
in
the
trawls
with
a
BRD
and
TED
combination
were
on
average
20.6%
smaller
than
those
in
the
trawls
without
devices,
indicating
a
tendency
for
larger
rays
to
escape.
If
small-sized
rays
would
be
escaping
from
the
trawl
through
the
BRD,
this
would
theoretically
cause
a
relative
size
increase
instead
of
the
observed
decrease.
We
further
quantified
the
effect
of
body
size
on
escape
ratio
and
confirmed
that
escape
was
size-dependent,
with
high
escape
ratios
(>80%)
for
large
individuals
(>50
cm).
Still,
factors
other
than
size
may
affect
escape
ratio
as
well.
Exclusion-at-size
was
clearly
dif-
ferent
between
the
two
modeled
species
D.
guttata
and
G.
micrura.
Looking
at
their
morphology,
D.
guttata
has
a
thick
and
rigid
disc,
in
contrast
to
the
more
flexible
and
smooth
body
of
G.
micrura.
G.
micrura
might
more
easily
bend
and
slip
in
between
the
bars
of
the
TED,
while
a
similar-sized
individual
of
D.
guttata
is
more
likely
to
escape
upon
interaction
with
the
TED.
A
TED
is
classified
as
a
mechanical
excluder,
separating
species
according
to
size
and
morphology
rather
than
behavior
(Broadhurst,
2000).
Neverthe-
less,
behavioral
differences
between
species
are
known
to
influence
escape
from
trawls
(e.g.,
Hannah
and
Jones,
2012)
and
could
be
of
importance
here
also.
Fish
escaping
from
trawls
may
suffer
delayed
mortality
due
to
injury
or
stress
caused
by
the
catch-and
escape-process
(Suuronen,
2005).
The
survival
of
rays
escaping
the
trawls
through
the
TED
remains
unclear.
Likewise,
although
discarded
rays
might
have
higher
chances
of
survival
than
teleost
fish
(Depestele
et
al.,
2014),
the
fate
of
rays
that
are
brought
on
deck
and
subsequently
discarded
is
not
well
understood.
The
performance
of
BRD
and
TED
was
assumed
constant
dur-
ing
the
study.
Even
though
the
gear
was
inspected
before
each
trip,
including
monitoring
of
the
grid
angle,
wear
and
damage
of
the
gear
might
well
have
affected
BRD
and
TED
performance
(e.g.,
Eayrs,
2007),
and
hence
exclusion
of
rays.
Nevertheless,
our
results
reflect
the
conditions
encountered
over
a
long
period
of
time,
under
normal
commercial
fishing
conditions.
A
very
high
escape
ratio
(77%)
was
observed
for
D.
geijskesi,
linked
to
the
fact
that
most
individuals
of
this
species
were
rather
large.
Escape
ratios
for
D.
guttata
(40%)
and
G.
micrura
(32%)
were
lower.
In
both
species,
the
dominant
catches
were
small-sized
indi-
viduals
that
were
unable
to
escape
from
the
trawls.
Nevertheless,
the
models
for
both
species
showed
that
larger
specimens
did
T.
Willems
et
al.
/
Fisheries
Research
175
(2016)
35–42
39
Fig.
3.
Mean
(+SD)
catch
rate
of
rays
in
noBT
net
(dark
grey)
and
wBT
net
(light
grey).
Percentages
denote
reduction
in
mean
catch
rate
in
the
wBT
net.
Asterisks
indicate
significant
differences
(Wilcoxon
signed
rank
tests;
***
=
p
<
0.001;
ns
=
not
significant);
n
=
number
of
individuals.
Fig.
4.
Box-and-whisker-plots
showing
minimum,
maximum,
0.25
percentile,
0.75
percentile
and
median
size
(disc
width)
of
the
different
ray
species
in
noBT
net
(dark
grey)
and
wBT
net
(light
grey).
Open
circles
indicate
the
mean
size
and
percentages
denote
reduction
in
mean
size
in
the
wBT
net.
Asterisks
indicate
significant
reductions
(Mann–Whitney
U
tests;
***
=
p
<
0.001;
ns
=
not
significant).
escape
efficiently
from
the
trawls
equipped
with
a
BRD
and
TED
combination.
Because
fecundity
tends
to
increase
with
body
size,
the
protection
of
large-sized
individuals
is
essential
to
maintain
productive
populations
(Stevens
et
al.,
2000).
Furthermore,
recruit-
ment
of
cartilaginous
fishes
to
the
adult
population
is
very
closely
linked
to
the
number
of
breeding
females
(Taylor
et
al.,
2013).
Females
of
D.
guttata
are
mature
from
50
to
55
cm
onwards
(Yokota
and
Lessa,
2007).
Our
results
show
a
nearly
complete
exclu-
sion
from
the
trawls
at
this
size,
allowing
for
potential
survival
of
breeding
females.
Still,
as
has
been
shown
for
Dasyatis
dipterura
in
the
Gulf
of
Mexico,
survival
of
both
adult
and
juvenile
stages
strongly
influences
population
growth
rates
(Smith
et
al.,
2008).
40
T.
Willems
et
al.
/
Fisheries
Research
175
(2016)
35–42
Fig.
5.
Size
distribution
and
GLMM
of
size
for
all
ray
species
combined
and
for
Dasyatis
guttata
and
Gymnura
micrura.
Left-hand
plots
present
size-frequency
distributions
in
wBT
net
(dashed)
and
noBT
net
(solid).
Right-hand
plots
present
the
GLMM
modeled
proportion
(shaded
area
=
95%
CI)
of
the
total
catch
in
the
wBT
net.
Interpretation:
a
value
of
0.5
(dashed
line)
indicates
an
even
split
between
the
two
trawls,
whereas
a
value
of
0.2
indicates
that
20%
of
all
rays
at
that
size
were
caught
in
the
wBT
net
and
80%
were
caught
in
the
noBT
net.
T.
Willems
et
al.
/
Fisheries
Research
175
(2016)
35–42
41
For
G.
micrura,
first
maturity
of
females
occurs
at
34–36
cm
(Yokota
and
Lessa,
2007),
a
size
at
which
exclusion
from
trawls
with
TEDs
was
low.
Due
to
its
relatively
early
maturity,
the
species
could
be
more
resilient
than
D.
guttata
(Walker
and
Hislop,
1998),
and
bet-
ter
able
to
cope
with
a
reduced
exclusion
rate.
Nevertheless,
G.
micrura
appeared
as
a
vulnerable
elasmobranch
species
in
the
Gulf
of
Mexico,
where
it
has
undergone
a
99%
decrease
since
the
early
1970s
due
to
shrimp
trawling
(Shepherd
and
Myers,
2005).
Both
D.
guttata
and
G.
micrura
are
red-listed
as
‘data
deficient’
(Grubbs
and
Ha,
2006;
Rosa
and
Furtado,
2004),
and
any
population
estimates
for
the
study
area
are
lacking.
No
reduction
in
catch
rate
was
observed
for
U.
microphthal-
mum.
Although
the
species
is
currently
assessed
as
‘least
concern’
(Rosa,
2004),
the
TED
caused
no
reduction
in
bycatch
of
this
species
because
of
its
small
size
(max.
25
cm;
Léopold,
2005)
and
it
did
not
appear
to
escape
through
the
BRD
either.
This
species
might
there-
fore
be
prone
to
high
fishing
mortality
in
shrimp
trawls.
Insufficient
data
were
collected
to
make
any
conclusions
on
R.
bonasus
from
the
current
study
(‘near
threatened’;
Barker,
2006).
The
current
study
shows
that
TEDs
cause
a
significant
reduction
in
the
bycatch
of
rays,
although
reduction
was
highly
dependent
on
size
and
species-specific
morphology.
Whilst
larger
rays
were
able
to
escape
at
a
relatively
high
rate,
rays
with
a
disc
width
of
approximately
20
cm
were
most
common,
a
size
at
which
escape
ratio
was
lower
(<60%).
This
is
still
a
positive
result,
given
that
in
the
pre-TED
days
a
much
higher
percentage
of
small
rays
would
not
survive.
However,
improvement
in
the
escapement
of
small
rays
is
required.
Smaller
rays
are
less
likely
to
survive
the
discard
process
than
larger
ones
(Benoit
et
al.,
2013;
Davis,
2002;
Depestele
et
al.,
2014),
adding
to
their
effective
mortality
relative
to
larger
rays.
Although
few
direct
estimates
have
been
generated
for
elasmo-
branch
fishes
(e.g.,
Gruber
et
al.,
2001;
Simpfendorfer,
1999),
their
natural
mortality
is
assumed
to
be
low
(Cailliet
et
al.,
2005;
Cortes,
2007).
Any
fisheries-induced
mortality,
even
of
juvenile
rays,
is
thus
likely
to
significantly
affect
the
ray
populations.
In
conclusion,
we
have
shown
that
the
BRD
and
TED
combination
causes
an
important
reduction
in
ray
bycatch
in
the
seabob
shrimp
fishery
off
Suriname.
Despite
the
large
reduction
in
catch
rates
of
large-sized
rays,
the
relatively
high
rate
of
mortality
of
juvenile
rays
is
likely
to
have
negative
consequences
for
their
populations.
As
very
little
information
is
currently
available,
a
precautionary
approach
in
fisheries
management
is
advisable
until
assessments
of
the
population
sizes
and
status
of
the
rays
in
these
fishing
grounds
become
available.
Future
gear
adaptations
and
efforts
should
focus
on
reducing
bycatch
of
small-sized
rays.
To
our
knowledge,
no
trawl
modifications
have
been
developed
to
specifically
tackle
ray
bycatch;
we
therefore
suggest
an
assessment
of
the
ability
of
sort-
ing
grids
with
reduced
bar
spacing
to
exclude
small-sized
rays
while
still
catching
shrimp.
Assessing
the
potential
of
super-shooter
TEDs
with
smaller
bar
spacing
seems
to
be
a
logical
next
step.
Another
option
could
be
Nordmøre-grids,
as
they
have
shown
not
to
affect
shrimp
catches
in
a
Brazilian
seabob
shrimp
fishery,
even
when
the
bars
are
spaced
only
17
mm
apart
(Silva
et
al.,
2012).
Finally,
square-
mesh
panel
BRDs
with
larger
meshes
to
reduce
small-sized
rays
could
also
be
tested.
Non
gear-related
solutions
can
include
spatial
and
temporal
restrictions
to
fishing
effort,
changes
in
fishing
prac-
tices
(e.g.,
move-on
rules;
Auster
et
al.,
2011)
and
modifications
in
catch
handling
on
deck
to
increase
post-capture
survival
(Depestele
et
al.,
2014).
Acknowledgments
We
thank
captain
Stephen
Hall
and
crew
from
FV
Neptune-
6,
fisheries
observers
from
the
Suriname
Ministry
of
Agriculture,
Livestock
and
Fisheries
(LVV)
and
people
from
the
Anton
De
Kom
University
of
Suriname
for
their
help
in
data
collection
on
board.
The
Heiploeg
Group
and
Heiploeg
Suriname
N.V.
are
acknowl-
edged
for
financial
and
logistic
support
of
the
research.
The
first
author
acknowledges
a
PhD
scholarship
from
VLIR-UOS
(VLADOC
2011–06).
We
are
grateful
to
three
anonymous
reviewers
for
their
helpful
comments
and
suggestions,
which
significantly
improved
the
quality
of
the
manuscript.
Thanks
also
to
Miriam
Levenson
for
English-language
review
of
this
manuscript.
References
Auster,
P.J.,
Gjerde,
K.,
Heupel,
E.,
Watling,
L.,
Grehan,
A.,
Rogers,
A.D.,
2011.
Definition
and
detection
of
vulnerable
marine
ecosystems
on
the
high
seas:
problems
with
the
move-on
rule.
ICES
J.
Mar.
Sci.
68,
254–264.
Barker,
A.S.,
2006.
Rhinoptera
bonasus.
The
IUCN
Red
List
of
Threatened
Species.
Version
2014.3.
<www.iucnredlist.org>
(accessed
28.11.14.).
Baum,
J.K.,
Myers,
R.A.,
Kehler,
D.G.,
Worm,
B.,
Harley,
S.J.,
Doherty,
P.A.,
2003.
Collapse
and
conservation
of
shark
populations
in
the
Northwest
Atlantic.
Science
299,
389–392.
Benoit,
H.P.,
Plante,
S.,
Kroiz,
M.,
Hurlbut,
T.,
2013.
A
comparative
analysis
of
marine
fish
species
susceptibilities
to
discard
mortality:
effects
of
environmental
factors,
individual
traits,
and
phylogeny.
ICES
J.
Mar.
Sci.
70,
99–113.
Bonfil,
R.,
1994.
Overview
of
world
elasmobranch
fisheries.
FAO
Fisheries
Technical
Paper.
341.
Brewer,
D.,
Heales,
D.,
Milton,
D.,
Dell,
Q.,
Fry,
G.,
Venables,
B.,
Jones,
P.,
2006.
The
impact
of
turtle
excluder
devices
and
bycatch
reduction
devices
on
diverse
tropical
marine
communities
in
Australia’s
northern
prawn
trawl
fishery.
Fish.
Res.
81,
176–188.
Brewer,
D.,
Rawlinson,
N.,
Eayrs,
S.,
Burridge,
C.,
1998.
An
assessment
of
bycatch
reduction
devices
in
a
tropical
Australian
prawn
trawl
fishery.
Fish.
Res.
36,
195–215.
Broadhurst,
M.K.,
2000.
Modifications
to
reduce
bycatch
in
prawn
trawls:
a
review
and
framework
for
development.
Rev.
Fish
Biol.
Fish.
10,
27–60.
Cailliet,
G.M.,
Musick,
J.A.,
Simpfendorfer,
C.A.,
Stevens,
J.D.,
2005.
Ecology
and
Life
History
Characteristics
of
Chondrichthyan
Fish.
Sharks,
Rays
and
Chimaeras:
the
Status
of
the
Chondrichthyan
Fishes.
IUCN
SSC
Shark
Specialist
Group.
IUCN,
Gland
Switzerland
and
Cambridge,
UK.
Cortes,
E.,
2007.
Chondrichthyan
demographic
modelling:
an
essay
on
its
use,
abuse
and
future.
Mar.
Freshw.
Res.
58,
4–6.
Davis,
M.W.,
2002.
Key
principles
for
understanding
fish
bycatch
discard
mortality.
Can.
J.
Fish.
Aquat.
Sci.
59,
1834–1843.
Depestele,
J.,
Desender,
M.,
Benoit,
H.P.,
Polet,
H.,
Vincx,
M.,
2014.
Short-term
survival
of
discarded
target
fish
and
non-target
invertebrate
species
in
the
eurocutter
beam
trawl
fishery
of
the
southern
North
Sea.
Fish.
Res.
154,
82–92.
Dulvy,
N.K.,
Metcalfe,
J.D.,
Glanville,
J.,
Pawson,
M.G.,
Reynolds,
J.D.,
2000.
Fishery
stability,
local
extinctions,
and
shifts
in
community
structure
in
skates.
Conserv.
Biol.
14,
283–293.
Dulvy,
N.K.,
Reynolds,
J.D.,
2002.
Predicting
extinction
vulnerability
in
skates.
Conserv.
Biol.
16,
440–450.
Eayrs,
S.,
2007.
A
Guide
to
Bycatch
Reduction
in
Tropical
Shrimp-Trawl
Fisheries,
Revised
ed.
FAO,
Rome,
108
p.
Eayrs,
S.,
2012.
Comparative
Testing
of
Bycatch
Reduction
Devices
in
Tropical
Shrimp-trawl
Fisheries—A
Practical
Guide.
FAO,
Rome,
pp.
122.
Fisher,
R.A.,
Call,
G.C.,
Grubbs,
R.D.,
2013.
Age,
growth,
and
reproductive
biology
of
cownose
rays
in
Chesapeake
Bay.
Mar.
Coast.
Fish.
5,
224–235.
Goodwin,
N.B.,
Dulvy,
N.K.,
Reynolds,
J.D.,
2002.
Life-history
correlates
of
the
evolution
of
live
bearing
in
fishes.
Philos.
Trans.
R.
Soc.
B-Biol.
Sci.
357,
259–267.
Graham,
K.J.,
Andrew,
N.L.,
Hodgson,
K.E.,
2001.
Changes
in
relative
abundance
of
sharks
and
rays
on
Australian
South
East
Fishery
trawl
grounds
after
twenty
years
of
fishing.
Mar.
Freshw.
Res.
52,
549–561.
Griffiths,
S.P.,
Brewer,
D.T.,
Heales,
D.S.,
Milton,
D.A.,
Stobutzki,
I.C.,
2006.
Validating
ecological
risk
assessments
for
fisheries:
assessing
the
impacts
of
turtle
excluder
devices
on
elasmobranch
bycatch
populations
in
an
Australian
trawl
fishery.
Mar.
Freshw.
Res.
57,
395–401.
Grubbs,
R.D.,
Ha,
D.S.,
2006.
Gymnura
micrura.
The
IUCN
Red
List
of
Threatened
Species.
Version
2014.3.
<www.iucnredlist.org>
(accessed
28.11.14.).
Gruber,
S.H.,
de
Marignac,
J.R.C.,
Hoenig,
J.M.,
2001.
Survival
of
juvenile
lemon
sharks
at
Bimini,
Bahamas,
estimated
by
mark-depletion
experiments.
Trans.
Am.
Fish.
Soc.
130,
376–384.
Hannah,
R.W.,
Jones,
S.A.,
2012.
Evaluating
the
behavioral
impairment
of
escaping
fish
can
help
measure
the
effectiveness
of
bycatch
reduction
devices.
Fish.
Res.
131,
39–44.
Heales,
D.S.,
Gregor,
R.,
Wakeford,
J.,
Wang,
Y.G.,
Yarrow,
J.,
Milton,
D.A.,
2008.
Tropical
prawn
trawl
bycatch
of
fish
and
seasnakes
reduced
by
yarrow
fisheye
bycatch
reduction
device.
Fish.
Res.
89,
76–83.
Holst,
R.,
Revill,
A.,
2009.
A
simple
statistical
method
for
catch
comparison
studies.
Fish.
Res.
95,
254–259.
IUCN,
2015.
The
IUCN
Red
List
of
Threatened
Species.
Version
2015.2.
<www.
iucnredlist.org>
(accessed
25.06.15.).
Léopold,
M.,
2005.
Poissons
de
mer
de
Guyane.
Ifremer.
Pérez,
A.R.,
2014.
Ecology
and
dynamics
of
the
Atlantic
seabob
shrimp
(Xiphopenaeus
kroyeri)
in
Suriname
based
on
fisheries
data
and
local
ecological