Tests of artificial light for bycatch reduction in an ocean shrimp (Pandalus jordani) trawl: Strong but opposite effects at the footrope and near the bycatch reduction device

Abstract

We investigated how the addition of artificial light in the vicinity of the rigid-grate bycatch reduction device (BRD) and along the fishing line of an ocean shrimp (Pandalus jordani) trawl altered fish bycatch and ocean shrimp catch. In separate trials using double-rigged shrimp nets, with one net incorporating artificial lights and the other serving as a control, we 1) attached one to four Lindgren-Pitman Electralume® LED lights (colors green or blue) in locations around the rigid-grate BRD, and 2) attached 10 green lights along the trawl fishing line. Both experiments were conducted with rigid-grate BRDs with 19.1 mm bar spacing installed in each net. Contrary to expectations, in 12 paired hauls the addition of artificial light around the rigid-grate increased the bycatch of eulachon (Thaleichthys pacificus), a threatened anadromous smelt species, by 104% (all by weight, P = 0.0005) and slender sole (Lyopsetta exilis) by 77% (P = 0.0082), with no effect on ocean shrimp catch or bycatch of other fishes (P > 0.05). In 42 paired hauls, the addition of 10 LED lights along the fishing line dramatically reduced the bycatch of a wide variety of fishes with no effect on ocean shrimp catch (P > 0.05). Bycatch of eulachon was reduced by 91% (P = 0.0001). Bycatch of slender sole and other small flatfishes were each reduced by 69% (P < 0.0005). Bycatch of darkblotched rockfish (Sebastes crameri), a commercially important but depressed rockfish species, was reduced by 82% (P = 0.0001) while the bycatch of other juvenile rockfish (Sebastes spp.) was reduced by 56% (P = 0.0001). How the addition of artificial light is causing these changes in fish behavior and bycatch reduction is not known. However, in both experiments the addition of artificial light appears to have greatly increased the passage of fishes through restricted spaces (between BRD bars and the open space between trawl fishing line and groundline) that they typically would not pass through as readily under normal seafloor ambient light conditions.

Full text

Fisheries
Research
170
(2015)
60–67
Contents
lists
available
at
ScienceDirect
Fisheries
Research
j
ourna
l
ho
me
pa
ge:
www.elsevier.com/locate/fishres
Tests
of
artificial
light
for
bycatch
reduction
in
an
ocean
shrimp
(Pandalus
jordani)
trawl:
Strong
but
opposite
effects
at
the
footrope
and
near
the
bycatch
reduction
device
Robert
W.
Hannaha,∗,
Mark
J.M.
Lomelib,
Stephen
A.
Jonesa
aOregon
Department
of
Fish
and
Wildlife,
Marine
Resources
Program,
2040
S.E.
Marine
Science
Drive,
Newport,
OR
97365,
United
States
bPacific
States
Marine
Fisheries
Commission,
205
SE
Spokane,
Portland,
OR
97202,
United
States
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
16
October
2014
Received
in
revised
form
3
March
2015
Accepted
12
May
2015
Handled
by
Dr.
P.
He
Keywords:
Bycatch
reduction
Artificial
light
Fish
behavior
Shrimp
trawl
Eulachon
Rockfish
a
b
s
t
r
a
c
t
We
investigated
how
the
addition
of
artificial
light
in
the
vicinity
of
the
rigid-grate
bycatch
reduction
device
(BRD)
and
along
the
fishing
line
of
an
ocean
shrimp
(Pandalus
jordani)
trawl
altered
fish
bycatch
and
ocean
shrimp
catch.
In
separate
trials
using
double-rigged
shrimp
nets,
with
one
net
incorporating
artifi-
cial
lights
and
the
other
serving
as
a
control,
we
1)
attached
one
to
four
Lindgren-Pitman
Electralume®LED
lights
(colors
green
or
blue)
in
locations
around
the
rigid-grate
BRD,
and
2)
attached
10
green
lights
along
the
trawl
fishing
line.
Both
experiments
were
conducted
with
rigid-grate
BRDs
with
19.1
mm
bar
spacing
installed
in
each
net.
Contrary
to
expectations,
in
12
paired
hauls
the
addition
of
artificial
light
around
the
rigid-grate
increased
the
bycatch
of
eulachon
(Thaleichthys
pacificus),
a
threatened
anadromous
smelt
species,
by
104%
(all
by
weight,
P
=
0.0005)
and
slender
sole
(Lyopsetta
exilis)
by
77%
(P
=
0.0082),
with
no
effect
on
ocean
shrimp
catch
or
bycatch
of
other
fishes
(P
>
0.05).
In
42
paired
hauls,
the
addition
of
10
LED
lights
along
the
fishing
line
dramatically
reduced
the
bycatch
of
a
wide
variety
of
fishes
with
no
effect
on
ocean
shrimp
catch
(P
>
0.05).
Bycatch
of
eulachon
was
reduced
by
91%
(P
=
0.0001).
Bycatch
of
slender
sole
and
other
small
flatfishes
were
each
reduced
by
69%
(P
<
0.0005).
Bycatch
of
darkblotched
rockfish
(Sebastes
crameri),
a
commercially
important
but
depressed
rockfish
species,
was
reduced
by
82%
(P
=
0.0001)
while
the
bycatch
of
other
juvenile
rockfish
(Sebastes
spp.)
was
reduced
by
56%
(P
=
0.0001).
How
the
addition
of
artificial
light
is
causing
these
changes
in
fish
behavior
and
bycatch
reduction
is
not
known.
However,
in
both
experiments
the
addition
of
artificial
light
appears
to
have
greatly
increased
the
passage
of
fishes
through
restricted
spaces
(between
BRD
bars
and
the
open
space
between
trawl
fishing
line
and
groundline)
that
they
typically
would
not
pass
through
as
readily
under
normal
seafloor
ambient
light
conditions.
©
2015
Elsevier
B.V.
All
rights
reserved.
1.
Introduction
The
limited
species
selectivity
of
trawls
is
a
continuing
concern
for
fisheries
scientists
and
managers.
Developing
new
technology
to
reduce
non-target
catch
(bycatch)
in
trawl
fisheries
is
especially
important
when
a
bycatch
species
is
considered
“threatened”
or
“endangered”.
This
is
the
case
with
eulachon
(Thaleichthys
paci-
ficus),
an
anadromous
smelt
inhabiting
the
western
coasts
of
the
United
States
and
Canada.
The
southern
distinct
population
seg-
ment
for
this
species
has
been
listed
as
threatened
under
the
U.S.
Endangered
Species
Act
(Gustafson
et
al.,
2012;
NWFSC,
2009)
and
∗Corresponding
author.
Tel.:
+1
541
867
0300x231;
fax:
+1
541
867
0311.
E-mail
addresses:
bob.w.hannah@state.or.us
(R.W.
Hannah),
MLomeli@psmfc.org
(M.J.M.
Lomeli),
Steve.a.jones@state.or.us
(S.A.
Jones).
is
being
considered
for
listing
as
“endangered”
under
the
Canadian
Species
at
Risk
Act
(http://www.dfo-mpo.gc.ca/species-especes/
species-especes/eulachon-eulakane-eng.htm#information).
Eula-
chon
are
regularly
captured
as
bycatch
in
the
small-mesh
trawl
fisheries
targeting
ocean
shrimp
(Pandalus
jordani)
operating
on
the
west
coasts
of
the
United
States
and
Canada.
Fish
bycatch,
including
the
catch
of
eulachon,
has
been
greatly
reduced
in
these
fisheries
via
the
mandatory
use
of
codend
bycatch
reduction
devices
(BRDs)
similar
to
the
Nordmøre
grate
system
(Hannah
and
Jones,
2007,
2012;
Isaksen
et
al.,
1992)
and
through
modifications
to
trawl
footropes
(Hannah
and
Jones,
2000).
However,
eulachon
are
a
small
fish
that
can
easily
fit
between
the
19.1
mm
bar
spacing
of
the
rigid-grate
BRDs
required
in
this
fishery.
Their
successful
exclusion
from
shrimp
trawls
is
behaviorally-based
and
is
most
efficient
for
larger
(>150
mm
TL)
eulachon
that
are
stronger
swimmers
(Hannah
and
Jones,
2012).
So,
when
small
eulachon
are
abundant,
eulachon
http://dx.doi.org/10.1016/j.fishres.2015.05.010
0165-7836/©
2015
Elsevier
B.V.
All
rights
reserved.

R.W.
Hannah
et
al.
/
Fisheries
Research
170
(2015)
60–67
61
bycatch
in
ocean
shrimp
trawls
can
be
large
and
has
been
increasing
as
eulachon
have
rebounded
from
very
depressed
population
levels
(Al-Humaidhi
et
al.,
2012).
Although
eulachon
population
abun-
dance
has
increased,
the
ocean
shrimp
fishery
is
still
considered
a
moderate
threat
to
eulachon
recovery
(Gustafson
et
al.,
2012),
thus,
further
reduction
of
eulachon
bycatch
in
the
ocean
shrimp
fishery
is
an
important
research
priority.
Several
studies
have
demonstrated
that
fish
encountering
trawls
or
simulated
trawl
components
respond
behaviorally
to
changes
in
visual
stimuli
(Glass
et
al.,
1995;
Glass
and
Wardle,
1995;
Ryer
and
Olla,
2000;
Ryer
et
al.,
2010),
suggesting
the
potential
to
use
color
or
artificial
light
as
a
means
to
reduce
bycatch.
However,
to
our
knowledge,
no
practical
applications
of
such
techniques
for
com-
mercial
trawl
fisheries
have
been
developed.
We
report
on
what
we
believe
to
be
the
first
successful
development
of
a
practical
bycatch
reduction
technology
for
a
shrimp
trawl
fishery
based
on
the
use
of
artificial
lighting.
Hannah
and
Jones
(2012)
analyzed
the
behavior
of
eulachon,
as
they
escaped
from
shrimp
trawls
via
BRDs,
to
evaluate
their
phys-
ical
condition
and
showed
that
excluded
eulachon
were
actively
swimming
and
mostly
avoiding
contact
with
the
rigid-grate
BRD.
That
study
utilized
underwater
video
with
bright
artificial
lighting,
bringing
into
question
how
the
presence
of
artificial
light
may
have
influenced
eulachon
escape
behavior.
Trawling
for
ocean
shrimp
is
conducted
at
depths
from
about
90–300
m
where
ambient
light
levels
are
typically
very
low.
The
video
observations
showing
that
eulachon
mostly
avoided
the
rigid-grate
BRD
that
was
illuminated
with
artificial
lights
suggested
the
possibility
that
enhancing
the
visibility
of
the
rigid-grate
with
artificial
light
under
actual
fishing
conditions
might
improve
eulachon
exclusion
efficiency
(Hannah
and
Jones,
2012).
In
the
first
field
experiment
reported
on
here,
we
tested
this
hypothesis.
The
footropes
(defined
as
the
combination
of
groundline,
fish-
ing
line
and
connecting
hardware)
used
on
ocean
shrimp
trawls
are
designed
to
keep
the
fishing
line
of
the
trawl
(where
the
net-
ting
is
attached)
elevated
about
35–70
cm
above
the
groundline
which
drags
along
the
seafloor
(Hannah
et
al.,
2011).
Recent
stud-
ies
have
shown
that
modifying
the
trawl
footrope
to
eliminate
portions
of
the
groundline
can
significantly
reduce
the
bycatch
of
eulachon,
however
the
modifications
tested
also
caused
signifi-
cant
shrimp
loss
(Hannah
and
Jones,
2013;
Hannah
et
al.,
2011).
If
footrope
modifications
can
be
found
that
reduce
eulachon
bycatch
with
minimal
shrimp
loss,
they
would
have
the
added
benefit
of
completely
avoiding
trawl
entrainment
of
these
fish,
thus
mini-
mizing
exhaustion
or
associated
behavioral
impairment
(Hannah
and
Jones,
2012;
Ryer
et
al.,
2004).
Footrope
modifications
also
have
a
greater
potential
than
codend
BRDs
to
reduce
the
bycatch
of
many
small
fishes,
which
may
have
enough
swimming
ability
to
escape
the
approaching
trawl
at
the
footrope,
but
be
too
fatigued
to
respond
effectively
when
they
reach
a
codend
BRD
(Hannah
and
Jones,
2013).
Bycatch
reduction
technology
in
shrimp
trawls
relies,
in
part,
on
a
fundamental
behavioral
difference
between
fish
and
shrimp.
Fish
respond
to
the
approaching
components
of
the
trawl
with
a
patterned
avoidance,
or
optomotor
(station-keeping),
response,
while
shrimp
exhibit
either
no
response
or
a
more
random,
reflex-
ive
and
unpatterned
response
(Hannah
et
al.,
2003;
Wardle,
1993;
Watson
et
al.,
1992).
However,
the
patterned
response
of
fish
that
can
be
used
to
separate
them
from
shrimp
depends
on
the
fish’s
ability
to
see
the
approaching
trawl
components
and
respond
to
them
(Kim
and
Wardle,
2003).
In
our
second
experiment,
we
tested
whether
using
artificial
lights
to
make
the
fishing
line
of
an
ocean
shrimp
trawl
more
visible
to
eulachon
and
other
fish
species
would
enhance
their
ability
to
avoid
the
net
and
escape
under
it,
generating
bycatch
reduction
with
little
or
no
shrimp
loss.
2.
Methods
2.1.
Field
methods
We
evaluated
the
effect
of
artificial
light
on
fish
bycatch
in
ocean
shrimp
trawl
nets
by
comparing
catches
from
the
port
and
starboard
nets
of
a
double-rigged
shrimp
vessel,
with
one
net
incorporating
artificial
lighting
and
the
other
acting
as
a
con-
trol.
To
generate
artificial
light
underwater
we
used
a
number
of
green
(centered
on
540
nm,
≥0.5–2.0
lx)
or
blue
(centered
on
460
nm,
≥0.5–2.0
lx)
Lindgren-Pitman
LED
Electralume®fishing
lights
attached
to
selected
portions
of
the
trawl
(detailed
below).
These
lights
were
chosen
for
several
reasons.
They
are
small,
inex-
pensive
and
use
low
amounts
of
battery
power.
They
are
also
pressure-rated
to
water
depths
greater
than
the
fishery
operates
at
and
are
rugged
enough
to
withstand
the
net
handling
procedures
used
by
vessel
operators.
Green
and
blue
lights
were
chosen
simply
because
these
colors
transmit
well
through
seawater.
Both
experiments
were
conducted
utilizing
the
21
m
double-
rigged
shrimp
trawler
F/V
Miss
Yvonne,
out
of
Newport,
Oregon,
in
July
2014.
The
trawl
nets
used
were
high-rise
box
trawls,
typical
for
the
ocean
shrimp
fishery.
Each
net
had
footrope
and
headrope
lengths
of
23
m
and
codend
mesh
size
of
35
mm
(BK,
stretched).
Each
net
was
spread
with
1.8
×
2.1
m
wood
and
steel
doors.
All
experimental
hauls
were
conducted
during
daylight
hours
which
is
also
typical
for
the
fishery,
as
ocean
shrimp
are
known
to
migrate
vertically
into
the
water
column
at
night,
becoming
unavailable
to
bottom
trawl
gear
(Pearcy,
1970).
The
study
area
chosen
was
the
shrimp
grounds
between
Cascade
Head
and
Cape
Lookout,
Ore-
gon
(45.0–45.34◦N.
latitude),
an
area
in
which
both
eulachon
and
ocean
shrimp
were
expected
to
be
found
in
moderate
abundance.
For
both
experiments,
each
net
incorporated
a
rigid-grate
BRD
with
19.1-mm
bar
spacing.
Neither
BRD
incorporated
a
guiding
panel
to
concentrate
catch
at
the
bottom
of
the
grate
(for
a
diagram
of
typ-
ical
rigid-grate
BRDs
in
this
fishery,
see
Hannah
and
Jones,
2007).
To
avoid
catches
that
were
too
large
to
sort
and
weigh
with
the
staff
available,
areas
where
moderate
levels
of
shrimp
catch
were
expected
were
targeted
and
haul
duration
was
also
limited
to
about
45–75
min.
Commercial
fishery
hauls
are
frequently
of
this
duration
but
are
sometimes
as
long
as
2–4
h,
depending
on
anticipated
catch
and
bycatch
levels.
Towing
speed
over
ground
was
typical
for
ocean
shrimp
trawling,
ranging
from
3.0
to
3.3
km
h−1(1.6–1.8
kt).
We
used
three
techniques
to
control
for
potential
differences
in
catch
efficiency
between
the
two
nets.
First,
both
nets
were
inspected
to
ensure
they
were
similarly
constructed.
Second,
the
treatment
effect
was
interchanged
periodically
between
the
two
nets
at
regular
intervals.
Lastly,
we
used
recording
inclinometers
attached
to
the
fishing
line
of
each
net
to
measure
and
equalize
fishing
line
height
(FLH)
between
the
two
nets.
FLH
has
been
shown
to
strongly
influence
both
shrimp
catch
and
fish
bycatch
in
ocean
shrimp
trawls
(Hannah
and
Jones,
2003).
The
inclinometers
were
also
used
continuously
on
both
nets
throughout
our
study
so
that
hauls
in
which
the
equality
of
FLH
between
the
nets
was
compro-
mised
by
large
debris
getting
tangled
on
the
footrope
of
one
net
could
be
excluded
prior
to
data
analysis.
The
inclinometer
data
showed
that
both
nets
were
fishing
at
comparable
FLH
throughout
the
8
days
of
experiments,
with
some
normal
haul-to-haul
variabil-
ity
(Fig.
1).
No
hauls
were
excluded
due
to
abnormal
variation
in
the
FLH
of
a
particular
net.
Handling
of
data
in
the
field
was
similar
for
both
experiments.
The
catch
from
each
net
was
emptied
into
one
side
of
a
divided
hopper,
and
then
sorted
to
species
and
counted
and
weighed
at
sea.
In
a
few
cases,
hauls
were
subsampled
(approximately
30
kg)
by
weight
before
sorting
and
total
weight
by
species
was
estimated
from
the
species
composition
of
the
subsample
and
the
total
catch
weight.
Eulachon
and
juvenile
rockfish
(Sebastes
spp.)
were
placed

62
R.W.
Hannah
et
al.
/
Fisheries
Research
170
(2015)
60–67
22
27
32
37
42
47
1
2
3
4
5
6
7
8
9
10
11
12
Fishing line height (cm)
Haul number
Starboard control
Port control
Starboard treatment
Port treatment
22
27
32
37
42
47
1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
Fishing line height (cm)
Haul nu
mber
Starboard control
Port control
Starboard t
reatment
Port treatment
Fig.
1.
Mean
fishing
line
height
(cm,
measured
at
the
center
of
the
fishing
line)
in
port
and
starboard
nets,
by
treatment,
in
fishing
experiments
comparing
catches
in
ocean
shrimp
trawl
nets
with
1–4
LED
lights
attached
in
the
vicinity
of
the
bycatch
reduction
device
(upper
panel)
and
with
10
LED
lights
attached
to
the
trawl
fishing
line
(lower
panel),
by
haul
number.
into
labeled
sample
bags
and
frozen
for
later
lab
analysis.
Lengths
(TL,
mm)
were
measured
only
for
eulachon
and
were
generated
during
lab
analysis.
In
some
cases,
the
complete
eulachon
catch
for
one
or
both
nets
was
too
large
to
retain.
For
those
catches,
a
subsample
(approximately
1–2
kg)
was
bagged
and
frozen
and
the
rest
were
weighed
and
discarded
at
sea.
Light
levels
inside
the
nets
were
measured
using
Wildlife
Com-
puters
TDR-MK9
archival
tags.
Prior
to
field
sampling,
the
MK9
tags
were
calibrated
using
an
International
Light
IL1700
light
meter
and
PAR
sensor.
Both
MK9
tags
had
similar
responses
to
the
calibration.
Therefore,
the
tag
values
were
pooled
and
one
calibration
function
was
generated.
The
calibration
function
used
to
convert
the
MK9
relative
light
units
to
irradiance
units
was:
y
=
1
×
10−9e0.1472x(1)
where
x
is
the
relative
light
unit
from
the
MK9
and
y
is
the
cor-
responding
irradiance
unit
in
!mol
photons
m−2s−1.
The
R2value
from
our
calibration
curve
was
0.9867.
2.2.
Artificial
light
near
the
rigid-grate
BRD
The
initial
light
configuration
tested
was
four
green
Lindgren-
Pitman
Electralume®LED
lights
attached
with
zip-ties
directly
to
the
forward
side
of
the
rigid-grate
BRD,
spaced
evenly
around
the
edges
of
the
circular
grate.
Several
different
locations
in
the
vicin-
ity
of
the
BRD
were
also
tried,
with
either
green
or
blue
lights.
Due
to
the
difficulty
of
interpreting
effects
of
the
different
light
config-
urations
with
such
small
sample
sizes,
we
present
all
of
the
data
combined
from
12
hauls
over
two
days
of
field
trials
with
lights
on,
or
near,
the
BRD,
including
behind
the
BRD
and
arranged
around
the
escape
hole
in
front
of
the
BRD,
as
all
results
were
similar.
For
this
experiment,
we
attached
the
Wildlife
Computers
TDR-MK9
archival
tags
on
the
floor
of
each
net
facing
upward,
directly
in
front
of
the
BRD,
to
measure
light
levels
in-situ
both
with
and
without
the
LED
lights.
2.3.
Artificial
light
along
the
fishing
line
To
evaluate
the
effect
of
artificial
light
in
the
vicinity
of
the
trawl
footrope,
we
attached
10
green
Lindgren-Pitman
Electralume®LED
lights
with
zip-ties
directly
to
the
central
40%
of
the
fishing
line
of
the
trawl
(Fig.
2).
Lights
were
equally
spaced
at
about
1.2
m
apart.
We
conducted
42
hauls
evaluating
this
configuration,
switching
the
lights
from
the
port
to
the
starboard
net
periodically
over
6
days
of
field
trials.
For
this
experiment
the
MK9
archival
tags
were
attached
to
the
floor
of
the
net
directly
behind
the
center
of
the
fishing
line,
facing
upward,
to
measure
light
levels
near
the
seafloor
in
each
net.
2.4.
Data
analysis
Catch
weight
(kg)
data
were
analyzed
as
a
3-factor
ANOVA,
with
haul,
side
of
gear
(port
or
starboard)
and
the
treatment
as
main
effects
without
interaction,
following
Hannah
et
al.
(2011).
For
some
species
or
species
groups,
transformations
were
utilized
to
achieve
normality
of
model
residuals.
Length
data
for
eulachon
were
expected
to
be
multi-modal
and
therefore
length
samples
were
compared
between
treatment
and
control
nets
using
the
non-
parametric
Wilcoxon
two-sample
test
(Sokal
and
Rohlf,
1981)
and
also
evaluated
graphically.
For
graphical
comparison,
length
fre-
quency
sample
data,
by
treatment,
were
combined
across
hauls
using
a
catch-weighted
average
and
expressed
as
a
percentage
of
the
total
frequency.
3.
Results
3.1.
Artificial
light
near
the
rigid-grate
BRD
The
first
four
hauls
with
green
LED
lights
attached
directly
to
the
rigid-grate
BRD
(hauls
1–4)
showed
a
strong
and
unexpected
result
of
greatly
increased
eulachon
bycatch
in
the
net
incorporating
the
lights
(Fig.
3).
Subsequent
hauls
with
1
green
light
attached
to
the
BRD
escape
opening
(hauls
5–8),
4
green
lights
behind
the
rigid-
grate
(hauls
9–10)
and
3
blue
lights
attached
to
the
edges
of
the
escape
opening
(hauls
11–12)
provided
similar
results
(Fig.
3).
Blue
lights
were
used
on
the
last
two
hauls
to
see
if
difference
in
color
would
strongly
alter
the
results
being
observed.
For
these
12
hauls,
fishing
line
height
(cm)
was
well
equalized
between
the
port
and
starboard
nets,
averaging
(±SE)
39.6
(±0.3)
and
38.4
(±0.5)
cm
for
the
port
and
starboard
nets,
respectively,
(Fig.
1).
After
12
hauls,
further
experimentation
with
lights
near
the
BRD
was
abandoned
in
favor
of
using
the
remaining
vessel
time
to
investigate
the
effects
of
artificial
light
at
the
footrope,
a
change
that
was
also
necessitated
by
the
need
to
limit,
to
the
extent
practicable,
total
eulachon
catch
mortality
in
these
two
experiments.
The
mean
ambient
light
level
(±SE)
measured
in
front
of
the
rigid-grate
BRD
in
the
control
net
during
this
experiment
was
3.11
×
10−5(±1.05
×
10−5)
!mol
photons
m−2s−1and
ranged
from
5.70
×
10−7to
1.30
×
10−4!mol
photons
m−2s−1(Fig.
4).
The
LED
lights
in
the
vicinity
of
the
BRD
increased
the
average
light
level
measured
at
this
location
to
3.86
×
10−3!mol
photons
m−2s−1
(±1.00
×
10−3),
or
about
1–2
orders
of
magnitude
(Fig.
4).
Considered
together,
these
12
hauls
showed
that
artificial
light
in
the
vicinity
of
the
rigid-grate
BRD
increased
eulachon
bycatch
by
104%
(all
by
weight
unless
noted,
P
=
0.0005,
Table
1),
but
had

R.W.
Hannah
et
al.
/
Fisheries
Research
170
(2015)
60–67
63
Fig.
2.
Image
of
a
green
Lindgren-Pitman
Electralume®LED
light
zip-tied
to
the
fishing
line
of
an
ocean
shrimp
trawl
(lower
left);
image
of
the
placement
of
a
green
LED
in
relation
to
the
drop-chains
and
groundline
of
an
ocean
shrimp
trawl
(upper
left);
image
depicting
green
LED
lights
attached
to
the
fishing
line
of
an
ocean
shrimp
trawl
prior
to
trawl
deployment
(right).
no
effect
on
shrimp
catch
or
the
bycatch
of
darkblotched
rock-
fish
(S.
crameri)
or
other
juvenile
rockfish
(P
>
0.05).
Interestingly,
the
bycatch
of
slender
sole
(Lyopsetta
exilis)
was
also
increased
by
77%
when
artificial
light
was
present
in
the
vicinity
of
the
BRD
(P
=
0.0082,
Table
1)
while
bycatch
of
other
small
flatfishes
was
not
influenced
(P
>
0.05).
Eulachon
captured
in
the
treatment
net
were
slightly
larger
than
in
the
control
net,
with
the
treatment
and
con-
trol
catch
samples
averaging
(±SE)
123.1
(±0.5)
and
121.1
(±0.5)
mm,
respectively
(P
=
0.0158).
The
graphical
comparison
of
length
frequency
(Fig.
5,
upper
panel)
suggests
that
some
of
the
larger
eulachon,
which
typically
would
have
escaped
the
net
via
the
BRD
in
the
absence
of
artificial
light,
passed
through
the
rigid-grate
BRD
more
frequently
when
artificial
light
was
present.
Given
the
large
increase
in
eulachon
bycatch
with
artificial
lights
near
the
rigid-
grate
(Table
1)
and
the
very
modest
shift
in
eulachon
length
(Fig.
5,
upper
panel),
the
influence
of
artificial
light
on
eulachon
behavior
near
the
rigid-grate
was
not
considered
to
be
strongly
size-based.
3.2.
Artificial
light
along
the
fishing
line
Along
the
fishing
line,
the
effect
of
introducing
green
artifi-
cial
light
in
42
comparison
hauls
was
nearly
opposite
of
what
we
observed
at
the
rigid-grate
BRD.
The
LED
lights
reduced
eulachon
bycatch
by
91%
(Fig.
6,
Table
2,
P
=
0.0001).
The
lights
also
reduced
S
PP
S
S
P
P
SS
P
S
P
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
0
20
40
60
80
100
120
140
1
2
3
4
5
6
7
8
9
10
11
12
Ratio control/treatment
Eulachon catch (kg)
Haul
LED ligh
ts (1-4)
near BRD
Control (no
lights)
Ratio
cont
rol/tre
atme
nt
Fig.
3.
Haul-by-haul
comparison
of
the
catch
of
eulachon
(kg)
in
the
two
nets
of
a
double-rigged
shrimp
trawl
vessel
with
one
side
incorporating
1–4
LED
lights
near
the
bycatch
reduction
device
(see
text
for
light
configuration
by
haul
number)
and
the
other
acting
as
a
control
(no
lights).
The
ratio
of
control/treatment
catch
is
also
shown
(solid
line).
Label
“P”
or
“S”
denotes
the
side
of
trawl
gear
(port
or
starboard)
used
as
the
control
net.

64
R.W.
Hannah
et
al.
/
Fisheries
Research
170
(2015)
60–67
1.0E-0
7
1.0E-0
6
1.0E-0
5
1.0E-0
4
1.0E-03
1.0E-0
2
1.0E-0
1
1.0E+00
234567
8
9
10
11
12
13
µmol photons m-2 s-1
Haul number
LED lights
(1-4)
near BRD
Control
(no
ligh
ts)
1.0E-07
1.0E-06
1.0E-05
1.0E-04
1.0E-03
1.0E-02
1.0E-01
1.0E+0
0
1357
9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
µmol photons m-2 s-1
Haul number
LED lights
(10)
on fishin
g line
Control
(no
li
ght
s)
Fig.
4.
Light
levels
(!mol
photons
m−2s−1),
by
treatment
and
haul
number,
mea-
sured
in
the
fishing
experiments
testing
the
effect
of
1–4
LED
lights
in
the
vicinity
of
the
bycatch
reduction
device
(upper
panel,
see
text)
and
10
LED
lights
attached
to
the
trawl
fishing
line
(lower
panel,
see
text).
juvenile
darkblotched
rockfish
bycatch
by
82%
(Table
2,
P
=
0.0001)
and
bycatch
of
other
juvenile
rockfishes
by
56%
(Table
2,
P
=
0.0004).
Bycatch
of
slender
sole
and
other
small
flatfishes
were
both
also
reduced
69%
(Table
2,
P
=
0.0001).
The
presence
of
the
LED
lights
at
the
footrope
had
no
measurable
effect
on
shrimp
catches,
with
shrimp
catches
in
the
net
with
lights
reduced
on
average
by
just
0.7%,
a
difference
that
was
non-significant
(Table
2,
P
>
0.05).
The
mean
ambient
light
level
measured
at
the
con-
trol
net
during
the
footrope
experiment
was
4.84
×
10−5
(±2.30
×
10−5)
!mol
photons
m−2s−1and
ranged
from
2.88
×
10−7
to
8.56
×
10−4!mol
photons
m−2s−1,
indicating
similar
levels
of
ambient
light
on
the
seafloor
in
the
two
experiments
(Fig.
4).
The
LED
lights
on
the
fishing
line
increased
the
average
light
level
measured
to
1.47
×
10−4(±2.40
×
10−5)
!mol
photons
m−2s−1.
This
difference
probably
understates
the
increase
in
light
available
directly
under
the
net
because
the
MK9
archival
tags
were
oriented
upwards
such
that
any
added
light
from
the
artificial
lights
secured
to
the
fishing
line
would
have
reached
the
sensor
only
indirectly.
The
addition
of
lights
to
the
fishing
line
did
not
alter
the
mean
size
of
eulachon
captured
(P
>
0.05),
but
did
alter
the
shape
of
the
distribution.
The
graphical
comparison
of
length
frequency
in
the
treatment
and
control
nets
(Fig.
5,
lower
panel)
shows
a
pattern
that
is
consistent
with
a
weak
density-dependent
escapement
response.
The
largest
relative
proportional
decrease
in
eulachon
capture
was
between
116
and
134
mm,
the
length
range
in
which
eulachon
were
also
most
abundant,
with
relatively
reduced
effects
for
both
smaller
and
larger-sized
eulachon.
The
mean
fork
length
of
eulachon
captured
in
the
control
net
in
this
experiment
was
127.7
(±0.4)
mm,
slightly
larger
than
in
the
first
experiment.
4.
Discussion
The
addition
of
LED
lights
along
the
fishing
line
of
an
ocean
shrimp
trawl
was
highly
effective
at
reducing
bycatch
of
all
sizes
of
eulachon,
an
important
result
for
a
species
of
current
high
con-
servation
concern
(Gustafson
et
al.,
2012),
with
negligible
loss
of
ocean
shrimp.
The
lights
also
caused
a
large
percentage
reduction
in
the
bycatch
of
juvenile
darkblotched
rockfish,
a
depressed
species,
as
well
as
large
reductions
in
bycatch
of
other
small
fishes.
These
results
illustrate
the
increased
effectiveness
of
bycatch
reduction
technologies
for
small
fishes
when
implemented
near
the
front
of
the
trawl,
where
these
fish
retain
more
swimming
ability.
Facilitat-
ing
escapement
at
the
front
of
the
trawl
may
also
minimize
adverse
effects
on
escaping
fish
from
their
interactions
with
the
trawl.
These
fish
are
spared
the
exhaustion,
crowding
and
physical
contact
with
trawl
components
that
can
occur
prior
to
exclusion
via
rigid-grate
BRDs
(Hannah
and
Jones,
2012;
Soldal
and
Engås,
1997)
or
escape-
ment
through
trawl
meshes
(Ryer
et
al.,
2004;
Suuronen
et
al.,
1996,
2005).
We
would
expect
such
a
brief
encounter
with
the
trawl
to
have
minimal
impact
on
subsequent
survival.
It
is
worth
noting
also,
that
our
results
are
based
on
measuring
the
residual
bycatch
in
nets
with
fully
functioning
rigid-grate
BRDs
with
19.1
mm
bar
spacing.
Thus,
we
could
not
have
sampled
large
fish
that
would
typically
be
excluded
by
the
BRD
and
are
uncertain
how
many
of
these
fish
may
have
also
completely
avoided
trawl
entrainment.
Our
two
experiments
obtained
strong
but
opposite
effects
on
eulachon
bycatch
from
adding
artificial
lights
in
the
vicinity
of
the
rigid-grate
BRD
and
along
the
fishing
line.
Although
the
effect
on
Table
1
Comparison
of
mean
catch
by
species
or
group
(weight,
kg
haul−1except
for
darkblotched
rockfish,
other
juvenile
rockfishes
and
other
small
flatfish,
which
are
expressed
as
g
haul−1)
between
ocean
shrimp
trawl
nets
equipped
with
artificial
LED
lights
in
the
vicinity
of
a
rigid-grate
bycatch
reduction
device
(BRD)
with
19.1
mm
bar
spacing.
Species
were
captured
off
the
Oregon
coast
in
12
hauls
employing
double-rigged
nets,
one
incorporating
artificial
lights
near
the
BRD,
during
July
2014.
SE
=
standard
error.
Artificial
lights
Control
net
(no
lights)
Percent
reduction
with
lights
(%)
Species
or
group
Mean
catch
(SE)
Mean
catch
(SE)
P-value1
Ocean
shrimp
Pandalus
jordani
117.05
(26.13)
117.08
(27.56)
0.0
ns
Pacific
eulachon
Thaleichthys
pacificus
33.48
(2.42)
16.40
(2.42)
−104.2
0.0005
Slender
sole
Lyopsetta
exilis
1.49
(0.25)
0.84
(0.16)
−77.4
0.0082
Other
small
flatfish
291.06
(54.70)
287.28
(89.70)
−1.3
ns
Darkblotched
rockfish
Sebastes
crameri
389.88
(109.80)
428.42
(135.07)
9.0
ns
Other
juvenile
rockfish
Sebastes
spp.
71.50
(17.25)
109.34
(34.83)
34.6
ns
13
factor
ANOVA
(see
text).

R.W.
Hannah
et
al.
/
Fisheries
Research
170
(2015)
60–67
65
0.0%
2.0%
4.0%
6.0%
8.0%
10.0%
12.0%
14.0%
80
89
98
107
116 12
5
134
143 15
2
161
170 17
9
188
197 20
6
215 22
4
Percent frequency
Total length (mm)
LED lig
hts (1-4)
near
BRD
Control (no lig
hts)
0.0%
2.0%
4.0%
6.0%
8.0%
10.0%
12.0%
80
89
98
107
116 12
5
134
143 15
2
161
170 17
9
188
197 20
6
215 22
4
Percent frequency
Total length
(m
m)
LED lights
(10)
on
fi
shing
li
ne
Control
(n
o lights
)
Fig.
5.
Percent
length
frequency
of
eulachon
(total
length,
mm)
captured
in
ocean
shrimp
trawl
nets
with
and
without
1–4
LED
lights
attached
in
the
vicinity
of
the
bycatch
reduction
device
(upper
panel)
and
with
and
without
10
LED
lights
attached
to
the
trawl
fishing
line
(lower
panel).
S
SS
P
P
P
PP
P
PS
S
S
S
P
P
P
P
S
SSSSSS
P
P
P
P
P
P
P
S
SS
SS
S
S
PP
P
0
10
20
30
40
50
60
70
0
10
20
30
40
50
60
135
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
Ratio control/treatment
Eulachon catch (kg)
Haul
LED ligh
ts (10
) on fi
shin
g
line
Contr
ol (no li
ghts
)
Ratio control/t
reatme
nt
Fig.
6.
Haul-by-haul
comparison
of
the
catch
of
eulachon
(kg)
in
the
two
nets
of
a
double-rigged
shrimp
trawl
vessel
with
one
side
incorporating
10
LED
lights
on
the
fishing
line
and
the
other
acting
as
a
control.
The
ratio
of
control/treatment
catch
is
also
shown
(solid
line).
Label
“P”
or
“S”
denotes
the
side
of
trawl
gear
(port
or
starboard)
used
as
the
control
net.

66
R.W.
Hannah
et
al.
/
Fisheries
Research
170
(2015)
60–67
Table
2
Comparison
of
mean
catch
by
species
or
group
(weight,
kg
haul−1except
for
darkblotched
rockfish,
other
juvenile
rockfishes
and
other
small
flatfish,
which
are
expressed
as
g
haul−1)
between
ocean
shrimp
trawl
nets
equipped
with
artificial
LED
lights
attached
to
the
fishing
line
(at
the
footrope,
see
text)
and
a
control
net
with
no
lights.
Species
were
captured
off
the
Oregon
coast
in
42
hauls
employing
double-rigged
nets,
one
incorporating
the
artificial
lights,
during
July
2014.
SE
=
standard
error.
Artificial
lights
Control
net
(no
lights)
Percent
reduction
with
lights
(%)
Species
or
group
Mean
catch
(SE)
Mean
catch
(SE)
P-value1
Ocean
shrimp
Pandalus
jordani
203.68
(24.19)
205.15
(23.69)
0.7
ns
Pacific
eulachon
Thaleichthys
pacificus
1.12
(0.20)
11.77
(1.68)
90.5
0.0001
Slender
sole
Lyopsetta
exilis
0.72
(0.17)
2.29
(0.35)
68.6
0.0001
Other
small
flatfish
171.18
(28.24)
559.97
(60.25)
69.4
0.0001
Darkblotched
rockfish
Sebastes
crameri
95.44
(21.63)
537.23
(91.01)
82.2
0.0001
Other
juvenile
rockfish
Sebastes
spp.
55.09
(22.40) 126.13
(29.73) 56.3 0.0004
13
factor
ANOVA
(see
text).
bycatch
was
opposite,
the
mechanism
behind
the
changes
in
behav-
ior
may
be
similar.
For
eulachon,
our
initial
hypothesis
of
“increased
avoidance”
of
the
rigid
grate
or
fishing
line
with
artificial
lighting
cannot
account
for
these
results;
it
is
inconsistent
with
increased
bycatch
with
an
illuminated
rigid-grate
BRD.
In
each
experiment,
the
addition
of
artificial
light
appears
to
have
encouraged
eulachon
to
pass
through
a
restricted
open
space
with
much
greater
consis-
tency,
either
between
the
bars
of
the
rigid-grate
BRD
or
between
the
fishing
line
and
groundline
of
the
trawl,
with,
of
course,
opposite
effects
on
escapement.
This
reasoning
suggests
that
the
successful
exclusion
of
most
eulachon
by
rigid-grate
BRDs
depends,
to
some
degree,
on
the
BRD
being
only
poorly
illuminated
under
typical
seafloor
ambient
light
conditions
in
this
fishery.
If
this
is
correct,
it
follows
that
for
some
small
fishes
that
can
pass
through
the
bars
of
a
rigid-grate
BRD
but
retain
some
swimming
ability
as
they
encounter
the
BRD,
modifications
to
the
grate
to
make
it
less
vis-
ible
to
fishes,
such
as
changing
the
color
of
the
grate
or
even
the
shape
of
the
vertical
bars,
may
improve
exclusion
efficiency.
Such
modifications
would
be
expected
to
be
most
effective
in
situations
where
typical
seafloor
ambient
light
levels
are
similar
to
or
higher
than
in
the
ocean
shrimp
fishery.
The
exact
mechanism
behind
these
divergent
effects
from
artificial
lighting
is
unknown.
We
speculate
that
the
increased
movement
of
fishes
through
restricted
spaces
in
both
experiments
may
have
to
do
with
illuminating
the
area
behind
the
“threaten-
ing”
object,
either
the
rigid-grate
BRD
or
the
trawl
groundline.
In
both
instances,
the
effect
likely
encouraged
some
species
to
also
move
downwards,
perhaps
exploiting
a
natural
tendency
to
move
towards
the
seafloor
when
threatened.
It
is
also
possible
that
artificial
illumination
simply
increases
the
contrast
between
the
trawl
components
and
the
background,
facilitating
fish
navigating
between
the
trawl
components,
or
possibly
giving
fish
more
time
to
react
to
the
approaching
threat.
Glass
and
Wardle
(1995)
and
Glass
et
al.
(1995)
showed
that
for
some
species
trawl
mesh
escape-
ment
behavior
could
be
modified
by
changing
the
relative
contrast
of
light
and
dark
trawl
components.
In
our
first
experiment,
there
were
statistically
non-significant
reductions
in
bycatch
of
juvenile
rockfishes
with
artificial
lights
near
the
rigid-grate
BRD
(Table
1),
suggesting
the
effects
of
altering
the
contrast
or
visibility
of
the
BRD
may
also
be
variable
between
species.
The
effect
of
adding
artificial
lights
is
also
likely
to
vary
with
changes
in
ambient
light
on
the
seafloor,
and
thus
with
depth
and
time
of
day,
as
well
as
fish
den-
sity
and
other
factors
(Godø
et
al.,
1999;
Walsh
and
Godø,
2003).
The
comparison
of
length
frequency
data
from
the
nets
with
and
without
LED
lights
on
the
fishing
line
(Fig.
5,
lower
panel)
suggests
that,
for
eulachon,
escapement
between
the
groundline
and
fishing
line
of
an
ocean
shrimp
trawl
involves
a
weak
density-dependent
component.
This
is
also
supported
by
the
apparent
association
of
large
percentage
reductions
in
eulachon
catch
with
LED
lights
on
the
fishing
line
with
larger
eulachon
catches
in
the
control
net
(Fig.
6).
Although
we
were
surprised
by
these
results,
they
are
consistent
with
partial
results
from
some
studies
of
fish
behavior
under
differ-
ent
light
intensities.
Weinberg
and
Munro
(1999)
noted
increased
escapement
of
flathead
sole
(Hippoglossoides
elassodon)
under
a
survey
trawl
footrope
in
the
presence
of
artificial
light,
but
no
effect
on
other
species.
In
a
Pacific
hake
(Merluccius
productus)
mid-
water
trawl,
Lomeli
and
Wakefield
(2012)
noted
Chinook
salmon
(Oncorhyncus
tshawytscha)
had
a
stronger
tendency
to
exit
an
open
escape
window
that
artificial
light
was
directed
towards.
How-
ever,
this
behavior
was
not
exhibited
by
widow
rockfish
(Sebastes
entomelas).
In
contrast,
in
our
study,
adding
artificial
light
along
the
fishing
line
of
an
ocean
shrimp
trawl
greatly
increased
escapement
for
a
wide
variety
of
fishes
(Table
2).
Our
results
from
adding
artificial
lights
in
the
vicinity
of
the
BRD
conflict
somewhat
with
the
prior
behavioral
analysis
of
eulachon
escaping
via
a
rigid-grate
BRD
as
detailed
by
Hannah
and
Jones
(2012).
In
that
study,
most
large
eulachon
were
observed
avoid-
ing
contact
with
the
grate
and
swimming
upwards
and
out
of
the
exit
hole,
just
in
front
of
the
grate,
while
a
small
percentage
were
observed
swimming
directly
aft
through
the
grate.
In
our
current
study,
artificial
light
near
the
BRD
greatly
reduced
the
exclusion
efficiency
of
the
rigid-grate
BRD
for
eulachon,
causing
large
num-
bers
to
swim
aft
through
the
grate.
Since
artificial
lighting
was
present
in
both
studies,
the
two
findings
are
difficult
to
reconcile.
However,
there
were
three
notable
differences
in
the
lighting
used
in
these
two
studies,
the
color
of
the
lighting,
the
intensity
and
its
orientation.
Hannah
and
Jones
(2012)
used
a
single
white
Deep
Sea
Power
and
Light
LED
Mini-Sealite®(50
W,
3000 ◦K,
950
lm)
aimed
across
the
rigid-grate,
while
in
this
study
we
used
1–4
weaker,
more
diffuse,
green
or
blue
LED
Lindgren-Pitman
lights
in
several
loca-
tions
on
or
near
the
rigid
grate.
It’s
possible
that
the
diffuse
LED
lights
used
in
this
study
were
more
effective
at
illuminating
the
area
behind
the
rigid-grate
than
the
Mini-Sealite®that
was
pointed
directly
across
the
surface
of
the
grate.
To
our
knowledge,