J.
Northw.
Atl. Fish. Sci., Vol. 9:
103-113
The Influence of Temperature on Some Fish Population
Parameters in the Barents Sea
Harald Loeng
Institute of Marine Research, P. O. Box 1870 Nordnes
5024 Bergen,
Norway
Abstract
At the
beginning
of this
century
it was clear that in the Barents Sea, variations in the physical
conditions
have a great influence on the
biological
conditions
of fish. The cold period
during
1977-82
initiated new investigations on the influence of
oceanographic
conditions
on recruitment,
distribution
and
growth
of
commercial
fish species in the Barents Sea, both by Norwegian and
USSR scientists. A review of some of the most
important
results achieved so far is presented.
Rich year-classes of cod
occur
only
in years with relatively
high
temperature on the spawning
grounds
and the areas of
their
distribution
during
the first half-year of
their
lives. Feeding
distributions
of cod,
haddock
and capelin depend on the
climatic
conditions
in the Barents Sea
with more easterly and
northerly
distributions
noted in warm years than in
cold
ones. The
growth
of fish also seems to depend on the environmental temperature, but the
temperature-growth
relationships are
probably
not simple. The
climatic
fluctuations
also
influence
the plankton
production
and thereby the
food
conditions
for
all plankton feeders. Temperature effects linked to
the variability of food may therefore be as
important
as the
direct
effect of temperature on the
biological
conditions
of fish.
Introduction
In a review of Norwegian cod and
herring
fisheries,
Hjort
(1914) reported
fluctuations
in fisheries back to
the early-1700s. Some attempts had been made to
explain these
fluctuations,
but
most of the theories
were valueless and
only
served "as
indications
of the
state of general
knowledge
concerning
marine
biology
at the periods in
which
they
arise" (Hjort, 1914). A more
scientific
approach was made by Helland-Hansen and
Nansen (1909)
who
bel ieved
that
variations in the phys-
ical
conditions
had great
influence
on the
biological
conditions
of various fish species, and
that
temperature
variations in the sea "are the
primary
cause of the great
and
hitherto
unaccountable
fluctuations
in the fisher-
ies".
Although
Hjort
(1914) was
critical
ofthishypothe-
sis, later investigations have
shown
that the physical
conditions
are at least
important
indicators
for
recruit-
ment,
distribution
and
growth
of the
most
important
commercial
fish species.
Du ring the last half of the 1970s there was a marked
temperature decrease in the
whole
Barents Sea,
which
resulted in great changes in the
distribution
areas of
Arcto-Norwegian
cod
(Gadus
morhua
L.),
haddock
(Melanogrammus
aeglefinus)
and capelin
(Mal/otus
vil/osus). This led to new investigations on the
influ-
ence of
climate
on the
biological
conditions
such as
recruitment,
distribution,
migration
and
growth
of
these species, both by
Norwegian
and USSR scientists.
The
intention
of this paper is to summarize and review
the results
from
investigations carried
out
mainly
dur-
ing the 1980s. The more general
ecological
features of
the Barents Sea was earlier reviewed by Dragesund and
Gj¢saeter (1988) and Loeng (1989),
while
Skjoldal and
Rey (1989) summarized the role of
plankton
production
during
the 1980s. Some
ecological
features of the adja-
cent Norwegian Sea were described by
Blindheim
(1989).
Physical Conditions
There are three main water masses in the Barents
Sea. These are related to three
different
current
sys-
tems: the Norwegian Coastal current, the
Atlantic
cur-
rent, and the
Arctic
current
system. Figure 1 shows the
main features of these
current
systems. Both the Coas-
tal and
Atlantic
water masses
flow
into
the Barents Sea
from
the
southwest
and
occupy
the southern part of the
sea. In the
northern
Barents Sea, the
Arctic
waters
flow
in
opposite
directions,
towards
the south and west.
Between the
Arctic
and
Atlantic
waters, there is an area
called the Polar
front
in
which
the
two
water masses
mix. In the western part of the Barents Sea, the
front
is
sharp and it separates
two
distinct
biological
regimes.
In the eastern part, the
front
is diffused and sometimes
difficult
to trace.
Horizontal
distribution
of temperature and
salinity
at 100 m depth (Fig. 2), also shows the
different
water
masses.
According
to Helland-Hansen and Nansen
http://journal.nafo.int
104
J.
Northw.
Atl. Fish. SeL, Vol.
9,1989
I
I
--
/
'\
,
/
/
,
\
/
/
/
/
/
,
//
Goose Bk.
/
/
/
I
I
I
I
,
I
I
t
/
/
I
I
~
Skolpen Bk.
/
I
/
I
/
,
.:
Great Bk.
/
po-/
/
I
/
/
I
.:
68°
~""""'--'-......L..-&--.a.....&-...&-~.a.&.:lL.8.lo---L""""'......L.......L..-&-"""'~~.&....I"---I~--'---&..-.a...
.......
__
~",---L...-a.;;..a.&:'-'--'-......L..-&-.....IlI.:....L-"""-"'---""""~;..;....,a,
~
5° 10° 15° 20° 25° 30° 35° 40° 45° 50° 55°
Fig. 1. Main features of surface currents in the Barents Sea. Atlantic currents
(~),
Coastal currents
(---~)
and
Arctic
currents
(--------~).
The hatched line indicates the mean position of the Polar
front
(Loeng, 1989).
Fig. 2.
LOENG: Influence of Temperature on Fish Populations
105
(1909), the Atlantic water is defined by salinities
higher
than 35.0 %0. The salinity decreases
gradually
east-
ward, and water with salinity
higher
than 34.9%0 and
temperature above 2° C may be considered as
Atlantic
water. The Coastal water has
almost
the same tempera-
ture as the
Atlantic
water.
but
is characterized by
lower
salinity (S <34.7
%0).
The
Arctic
water generally has a
temperature belowO° C, and
below-1
° C in many areas,
while the salinity varies between
34.4-34.8%0.
In long
cold periods, the eastern part of the Barents Sea may be
filled with
cold
dense
bottom
water
(t<-1.7°
C)
which
is
formed on the bank areas
off
Novaya Zemlya and on the
Central Bank (Midttun, 1985).
The
climatic
variations in the Barents Sea depend
mainly on the activity and properties of the
inflowing
Atlantic
water
(Midttun
and Loeng, 1987).
Climatic
vari-
ations therefore can be recorded in sections crossing
the
Atlantic
current. Figure 3 shows temperature
anom-
alies in the Kola-section
after
1930 (along 33° 30' E)
based on data
from
Bochkov
(1982) and
Midttun
et al.
(1981). Before 1940, the section was
occupied
irregu-
larly, wh iIe after 1945 the observations have been car-
ried
out
once a month.
After
a warm period in the 1930s,
the years after1945 were characterized by
fluctuations
of
3-5
year duration. Since 1970, iarge variations have
been observed in the Barents Sea. The period 1970-76
was warm, while the period 1977-82 was the longest
continuous
cold period observed since 1920 (Loeng,
1989). During 1982 there was a strong temperature
increase followed by a
couple
of warm years, while the
subsequent years have had temperatures
below
the
1921-80 mean.
The same variations in temperature as in the Bar-
ents Sea were found along the whole Norwegian coast
(Blindheim
et al., MS 1981),
indicating
that these varia-
tions were large-scale. These temperature variations
are as a result of advection as first suggested by
Helland-Hansen and Nansen (1909), thus a time lag is
indicated of about 6
months
between temperature vari-
at.ons in Lofoten,
which
is the spawning
ground
for
cod, and that of the central/eastern Barents Sea (Loeng
and Sundby, 1986; Ellertsen
et al., 1987.)
Variations in the
Atlantic
inflow
influence the ice
conditions, especially
during
winter
(Loeng, 1989).
Loeng (1979) introduced an ice
index
where the ice
coverage in the central Barents Sea was integrated over
the year: the results are
included
in Fig. 3. Negative
values indicate heavy ice conditions, while positive
values mean there was
little
ice. The close relationship
between the temperature of the
inflowing
Atlantic
water and the ice
cond
itions is apparent
for
the period
1970-88. The
discrepancy
during
1984 and 1985 was
due
to
extremely
high
melting
of ice
during
the
summers, when solar radiation plays the
dominant
role
by
directly
melting ice and heating the surface layer
(Vinje: 1984).
Recruitment of Cod
The year-class strength of cod is
mainly
deter-
mined
during
the first year of life (Hjort, 1914).
Sundby
et
al
(1989)
concluded
that year-class abundance is
determined
during
the first 6 months. There are several
factors that may be responsible
for
the variations of the
year-class strength. These
include
starvation at the
start of exogenous feeding (Hjort, 1914; Wiborg, 1957;
Kislyakov, 1961; Ellertsen et al., MS 1976, 1980 and
1984), predation on eggs and larvae (Murphy, 1961;
Melle and Eliertsen, 1984) and physical factors acting
directly
on egg and larval
populations
(Garrod and
Colebrook, 1978; Koslow, 1984;
Sinclair
et al., 1985). At
the active feeding stages of larvae, variable
contact
rate
between larvae and prey, induced by
wind
mixing
in the
surface layer, may be an
important
regulatory mecha-
nism
for
the formation of year-class strength (Rotschild
and Osborn, 1988;
Sundby
and Fossum, MS 1989).
Recently, relationships between temperature
con-
ditions
and
the
year-class
variations
of
Arcto-
1930
35 40
45
50
55
60 65 70 75 80 85
2.0
20
D
•
DD
DD.
D
DD
••
•
ti
ti
.'(f'-fr
1.5
15
1.0
~~
10
~
0.5
5
~
x
Q)
::J
"0
«i 0
o .E
Q)
Q)
Q.
~
~
-0.5
" I
~
I
1\
-5
.....
-1.0
" :
'~.
-10
-1.5
-15
-2.0
-20
1930
35
40
45
50
55 60 65
70
75
80 85
Fig. 3.
Temperature
anomalies
in the
Kola-section
(along
33
0
30/E)
during
the
period
1930-88 (solid line)
together
with
the ice
index
during
the
period 1970-88 (dashed line). (
..
indicate
year-class of cod
with
high
abundance;
Q
indicate
year-class
with
medium
abundance;
no
arrows
indicate
low year-class
abundance)
(Bochkov,
1982;
Midttun
et a/., 1981, Saetersdal and Loeng, 1987).
106
J.
Northw.
Atl. Fish. ScL, Vol. 9, 1989
TABLE
1.
Occurrence
of year-classes of
low,
medium
and
high
abun-
dance
in
cold
and
warm
years in the
Barents
Sea, 1902-87.
The
effect
of
temperature
at the
spawning
ground
on year-class
strength
has been studied by Ellertsen et
al. (1987). Figure 4A
shows
the relation between year-
class
strength
and the mean
temperature
in
Lofoten
(Vestfjorden)
during
the
spawning
period in
March-
April.
Good
year-classes never
occur
in
cold
years,
while
they
may be
produced
in warm years.
The
"trian-
gle
plot"
indicates
that
high
temperature
is a necessary,
but
not
a
sufficient
condition
for
the
production
of
year-classes of
high
abundance
(Ellertsen et al.,
1987).
A
similar
"triangle
plot"
occurs
when using
tempera-
tures
from
the
Kola-section
in
August-September,
which
coincides
with
the
time
when the year-class
appears as O-group fish in
that
area (Fig. 4B). However,
there are some
differences
between Fig. 4A and 4B in
that
each year
does
not
exactly
have the same
position
along
the temperature-axis. As a rule, a
cold
year in
Lofoten
corresponds
to a
temperature
below
the mean
value in the Kola-section,
while
a warm year on the
spawning
ground
usually
is also above the mean in the
4.0
6.0
3.5
e83
5.5
e54
e59
e 70
e 50
e83
e64
e 49
e 48
e8e467
e54
e 59
e51
e72
e55
60
e 46
e
e74
e61
e52
e
:~;
59
e60
e50
e 70
2.5 3.0
Temperature
(0 C)
Mean
4.5
58
e
2.0
e64
I e 49
e48:
I
I
I 84
e62 :47
ee
57
:e82
e72 e51
e
53:
e55
:
6~
e74
I
e52
7:
e65
: e67 e76
e80
I
I
e 71
e58
e 62
47
e56
e 71
e82
e68
e53
67
e79
e78
e
e80
4.0
e81
I !
1.5
B. Kola Section
5.0
Temperature
(0 C)
The
relationship
between
the
year-class
strength
of
cod
and
the mean
temperature
(0 C) at (A) the
spawning
ground
Lofoten
(Vestfjorden)
in
March-April
(Ellertsen et al.,
1987)
and
(8)
year-class
strength
of
cod
and the mean
temperature
(0 C) at
the
Kola-section
(along
33°
30/E)
in
August-
September.
78 e79
e66
Fig. 4.
~
1500
x
u
o
o
""0
(5
~
1000
(1)
:;--
C')
'0
Q)
.0
§ 500
z
Mean
e81
Distribution and Growth of Young Cod
A.
Lofoten
(Vestfjorden)
1mmatu re cod feed in the Barents Sea and in the
Svalbard area, and make seasonal east-west and
north-
e66
~
1500
x
u
o
o
u
(5
~
1000
(1)
:;--
C')
'0
Q)
.0
§ 500
z
Since the
climatic
variations
are large-scale, one
would
expect
that
other
stocks
which
reproduce
in
these ecosystems
would
be
similarly
affected. Drage-
sund (1971) and Saetersdal and
Loeng
(1987)
found
several years of
common
high
recruitment
in cod,
had-
dock
and
herring,
which
gave a
convincing
demonstra-
tion
of a close
temporal
relationship
for
larval survival.
The favourable physical
conditions
seem to be related
to increased heat
transport
in the
Atlantic
current
(Sae-
tersdal and Loeng, 1987;
Mukhina
et al.,
1987).
Barents Sea. An
important
exception
was
found
in
1948, when the
temperature
at the
spawning
field was
high,
while
the
temperature
in the Barents Sea was
slightly
below
the mean. Favourable
conditions
at the
spawning
ground
may be an
explanation
for
the
good
year-class
that
year.
1%
12%
High
7%
14%
Medium
35%
31%
Low
Cold
years
Warm years
Norwegian
cod have been examined. Saetersdal and
Loeng (1987)
concluded
that
the
major
part
of the
year-
classes of
high
and
medium
abundance
are
either
asso-
ciated
directly
with
positive
temperature
anomalies
in
the
early
part of a warm period in the Barents Sea, when
the feeding areas are
expanding,
or
they
occur
imme-
diately
prior
to a
shift
to a
warmer
regime,
for
example,
1948 and 1958 (Fig. 3).
Only
the year-classes 1963 and
1985 do
not
fit
into
the pattern of
high
recruitment
with
a regime of increasing
temperature
or positive
temper-
ature
anomalies
since 1930.
The
results
for
the
whole
period 1902-87 may be summarized as
shown
in
Table
1. Year-classes of
low
abundance
are
most
frequent,
occurring
in
two-thirds
of all years, and
they
are evenly
distributed
between
cold
and warm years. Year-classes
of
medium
strength
occurred
twice
as often in warm
years as in
cold
ones,
while
year-classes of
high
abun-
dance
have
occurred
12
times
more
often in warm than
in
cold
years. If we take
into
account
that
a
strong
year-class on the average
corresponds
to
two
medium
or ten weak year-classes (Saetersdal and Loeng, 1987),
the
influence
of
warm
years
on
cod
recruitment
becomes even
more
important.
As an
example
the
mean
yearly
production
during
the
warm
period
1970-76 was
3-4
times
higher
than in the
cold
period
1977-82.
Compared
to the
climatic
variations, the years
since 1970 (Fig. 3) had
strong
year-classes in 1970 and
1983,
while
the year-classes of 1973, 1975, 1984 and
1985 were of
medium
abundance.
During
the
cold
period
1977-82,
all
year-classes
were
of
low
abundance.
LOENG:
Influence
of
Temperature
on Fish
Populations
107
.........
+
/
/
/
/
_---c
1983
~.-----.cr
-
....
x1980
-_.-
~
",
..
~.
-
/x"..
,~-'
_---
......
·'7-::···--
---
.... 1984
",
,,,.-....-
....-Y
Mean
"
,,-
1981
~"
x
-4'~
-x-
tJ
- - II'
~.
-.
.:--:
1982
- -
~
- 1" _.
v: I
••..
'0':; .
...:
.
..;....~
.
,/
.9
'..:-'
.-
-'
..........
- ._. -.
-.
1979
•
~
-r
....-
.",.,......
.......+
1978
Q~
••
-..
-:-:0'-/-'
.
·0'
. "
,
, .
. ,
'Y.
"
.-.
-. -.-
..
"
I
I
,
I
I
+/
.'-..
~
::J
~
3
0,)
a.
E
0,)
f-
o
L--'--
__
....l-.-
__
"'---_---'-
__
---'-
__
-Jo-
__
~
_
__I
Nakken and Raknes (1987) also investigated the
extent
to
which
variations in the
environmental
temper-
ature of the fish
influenced
the
growth
rates. Figure 6
6.---r-----..------.----....--------.r------r----"T"""----,
4 5 6
Age (years)
Fig. 5. Mean
environmental
temperatures
for
various
age-groups
of
cod
in
February
for
the
years
1978-84
(Nakken
and Raknes,
1987).
The
mean
temperatures
for
the
period
are
indicated
with
a
solid
line.
Midttun
et
al. (1981) also
found
a westward
dis-
placement
of cod
with
increasing
age. Shevelev
et
al.
(1987) suggested that the cod year-class
migrated
westwards
with
an average of 80
nautical
miles-per-
year. Since the
temperature
decreases
from
west to
east (Fig. 2), the
younger
age-groups
inhabit
water
of
lower
temperatures
than
older
ones
found
farther
west.
This
is
clearly
shown
in Fig. 5,
although
large
year-to-
year variations
appear
for
all
age-groups.
On average,
there appears to be a systematic increase in
tempera-
ture by age of
about
0.35° C per year
for
fish
>3-years-
old (Nakken and Raknes, 1987).
This
indicates
that
these
age-groups
maintain
their
distribution
within
the
temperature
field, relative to each other,
more
or less
independent
of the
absolute
values of
temperature
dur-
ing the period of observation. For fish 1- and 2-years-
old the
picture
is
more
obscure, but
this
may be
explained
by the life
history
of the fish
prior
to the
first
year (Nakken and Raknes, 1987).
TheO-group
cod des-
cend
towards
bottom
at the end of thei r fi rst
year
of life,
and the areas
where
the
1-group
is observed in the
beginning
of the
next
year must, to a
great
extent,
coincide
with
the areas where the fish settled
(Nakken
and Raknes, 1987).
The
location
of
1-group
therefore
depends
mainly
on the
drifting
process of larvae in the
previous year, i.e. the
current
conditions
of the
upper
layer.
Therefore
the
location
of
1-year-olds
within
the
temperature
field is
determined
by processes
which
may
differ
from
those
determining
the
temperature
conditions
in the
environment
of the
older
age-groups.
Since 1977, the
Institute
of
Marine
Research,
Ber-
gen,
Norway,
has carried
out
combined
acoustic-trawl
surveys in
order
to investigate the
abundance
and
dis-
tribution
of cod and
haddock
in the Barents Sea
during
early
winter
(Dalen
et
al., 1977;
Hylen
et
al.,
1986).
Midttun
e{
al. (1981) used data
from
these surveys to
study
the
geographic
distribution
of each
age-group
of
immature
cod
during
a period (1977-81)
with
low
temperatures
in the
whole
Barents Sea (Fig. 3).
Their
results
demonstrate
a
clear
westward
shift
in the
distri-
bution
of
young
cod year-classes in the period
with
decreasing temperature, causing the
restriction
of cod
to the western
half
of the Barents Sea. Later, Shevelev
et al. (1987) and
Boytsov
et al. (1987) came to the
conclusion
that
the
distributions
of
both
immature
cod
and
haddock
were
determined
primarily
by the heat
content
of the water masses. In
addition,
Shevelev
et
al.
(1987)
concluded
that
cod responds to
water
tempera-
ture
variations faster than
haddock.
The
intensity
of
Atlantic
inflow
to the Barents Sea
also
influences
the
distribution
of O-group
cod
and
haddock.
The
distribution
is
determined
by
the
strength
of the
current
transporting
them
from
the
spawning
grounds.
An easterly
distribution
of O-group
of
both
species is
probably
caused by
high
activity
of
the
Atlantic
inflow
as
indicated
by Randa (1984) and
Mukhina
et
al. (1987).
They
used
temperature
as an
indicator
but
it is
most
likely
that
a
stronger
Atlantic
inflow
does
coincide
with
higher
temperatures.
Unfor-
tunately
there are no
observations
that
relate variations
on
current
conditions
and
temperature
changes,
how-
ever, a
newly
developed
wind-d~iven
numerical
current
model
for
the Barents Sea (Adlandsvik, MS 1989)
showed a remarkable
coincidence
between
volume
flux
of
Atlantic
water
and
temperature
anomalies in the
Kola-section. In warm years, the favourable feeding
areas
for
cod larvae in the eastern Barents Sea are
expanded (Ssaetersdal and Loeng, 1987), and in an
early part of a warm period, there
will
be very few
predators in
that
area.
That
should
make the
conditions
in the eastern Barents Sea even better
for
O-group cod
than
in the rest of the Barents Sea.
south
migration
(Maslov, 1968;
Midttun
etal.,
1981).ln
addition,
there are
temperature-related
displacements
of
concentrations
both
on small and large
time
and
space scales. Eggvin (1938) reported
more
westerly
fishing
areas in the Barents Sea in
cold
years than in
warm years. Hylen
et
al. (1961)
concluded
that
a
con-
nection
exists between the
environmental
temperature
and the
distribution
of
young
cod. When the
tempera-
ture
is
high
the fish are easterly
distributed,
while
they
tend to move westward
with
low
temperatures.
Kon-
stantinov
(1967, 1969) and
Mukhin
(MS 1979) came to
the same
conclusion
on the basis of USSR
trawl
catches in the Barents Sea.