(PDF) Spatial and temporal trends in condition of Atlantic cod Gadus morhua on the Icelandic shelf

MARINE ECOLOGY PROGRESS SERIES

Mar Ecol Prog Ser

Vol. 362: 261–277, 2008

doi: 10.3354/meps07409

Published June 30

INTRODUCTION

Fish condition is influenced by the combined effects

of physiological and environmental factors (Cui &

Wootton 1988) and follows seasonal and annual cycles

in cod (Lambert & Dutil 1997a, Schwalme & Chouinard

1999, Lloret & Rätz 2000, Yaragina & Marshall 2000),

reflecting patterns of energy accumulation and deple-

tion during periods of intense feeding, maturation,

reproduction, migration and over-wintering (Eliassen

& Vahl 1982a,b, Holdway & Beamish 1984, Lambert &

Dutil 1997a). Simple metrics of condition, such as the

commonly utilised Fulton’s K and hepatosomatic index

(H), are based only on length and weight data, but

have been found to be reliable indicators of biochemi-

cal composition and energy reserves (those stored in

muscle proteins and liver lipids respectively) of

Atlantic cod (Lambert & Dutil 1997b).

Condition has been shown to influence or be corre-

lated with most life history parameters, notably those

linked with energetic processes, and has the potential

to provide important information about the productiv-

ity and reproductive potential of a stock. In Atlantic

cod Gadus morhua, condition is related to fecundity,

fertilisation success, egg size, larval quality and viabil-

ity, atresia and the probability of skipped spawning

(e.g. Kjesbu et al. 1991, Marteinsdóttir & Steinarsson

1998, Rakitin et al. 1999, Lambert & Dutil 2000, Ride-

out et al. 2000, Marteinsdóttir & Begg 2002, Rose &

O’Driscoll 2002, Yoneda & Wright 2004, 2005a). How-

ever, other studies (e.g. McIntyre & Hutchings 2003,

Koops et al. 2004) have found contrasting evidence for

such relationships. Condition was found to influence

recruitment in the Northeast Arctic cod stock (Mar-

shall et al. 1999, 2000), the risk of post-spawning mor-

tality in Atlantic cod (Lambert & Dutil 2000) and the

probability of being mature in Atlantic cod (Marteins-

dóttir & Begg 2002, Dutil et al. 2006), walleye Stizoste-

dion vitreum (Henderson & Morgan 2002) and female

American plaice Hippoglossoides platessoides (Mor-

Spatial and temporal trends in condition of Atlantic

cod Gadus morhua on the Icelandic shelf

Heidi Pardoe

1, 2,

*

, Gudmundur Thórdarson

1

, Gudrun Marteinsdóttir

2

1

Marine Research Institute, Skúlagata 4, PO Box 1390, 121 Reykjavik, Iceland

2

Institute of Biology, University of Iceland, Sturlugata 7, 101 Reykjavik, Iceland

ABSTRACT: Interannual (1993 to 2006), spatial and bathymetric trends in the hepatosomatic index

(H) and relative body condition index (Kr) of Atlantic cod Gadus morhua in Icelandic waters were

investigated from data collected during annual spring (March) groundfish surveys. Individual stom-

ach content data were analysed to assess the relationship between condition and consumption of

capelin Mallotus villosus. Liver condition was found to vary with region and depth, tending to be

highest in deeper water and in the northern and eastern regions; where temperatures were relatively

cold and capelin consumption relatively high. A continuous decline in condition of cod in waters off

the east coast was observed throughout the study period. Mature cod were found to be in signifi-

cantly better condition than immature cod of the same age, and mature females were in better condi-

tion than mature males. Different patterns emerged from the 2 condition indices; little variation in Kr

was observed whereas H was found to be a dynamic index of condition, suggesting that these met-

rics are not equivalent measures of cod bioenergetic condition.

KEY WORDS: Cod · Iceland · Liver condition · Body condition · Spatial variation · Bathymetric

variation · Capelin · Stomach content

Resale or republication not permitted without written consent of the publisher

Mar Ecol Prog Ser 362: 261–277, 2008

gan 2004). Furthermore, positive correlations between

growth rate and condition have been detected (Hold-

way & Beamish 1984, Björnsson 2002, Rätz & Lloret

2003), and poor condition may result in increased rates

of natural mortality through greater vulnerability to

disease, predation and reduced capability to catch

mobile prey (Dutil & Lambert 2000, Dutil et al. 2006).

Features of the environment which influence physio-

logical condition, such as temperature and food supply

(Krohn et al. 1997), are likely to vary within the geo-

graphic area inhabited by a stock. The hydrography of

the waters surrounding Iceland has been described

by several authors (see Vilhjálmsson 2002 for an over-

view). The system is characterised by several currents,

linked to bottom topography, which generate spatial

and temporal differences in the water masses around

Iceland. The south and west coasts are influenced by

Atlantic water of relatively high temperature, salinity

and stability, in contrast to the northern and eastern

areas, where polar water transported by the more vari-

able East Greenland and East Icelandic currents pre-

dominates (Vilhjálmsson 2002). Capelin Mallotus villo-

sus is the most important prey item of Icelandic cod

(Magnússon & Pálsson 1989, 1991), and its migrations

and distribution on the Icelandic shelf and in the Ice-

land Sea are closely linked to this system of ocean cur-

rents, and are highly sensitive to environmental fluctu-

ations (Vilhjálmsson 2002).

Traditionally, biological sampling schemes for stock

assessments have, in addition to catch and effort data,

only required information on length and age (Sparre &

Venema 1998). The additional effort and resources

required to obtain weight data means that the informa-

tion necessary to systematically or even accurately

estimate trends in condition using established simple

metrics is lacking for most fish stocks, particularly for

long-term studies. In Iceland, such data for cod have

been routinely collected during spring groundfish sur-

veys. Here we analyse these data to produce a compre-

hensive report on the potential variation in condition

that has existed within the Icelandic cod stock during

the period 1993 to 2006, and to assess the influence of

water temperature, bathymetry and feeding (by analy-

sis of stomach content data) on the emergent patterns.

An important part of the study was to examine the

effect of age, sex and maturity status on H and the

relative body condition index (Kr).

MATERIALS AND METHODS

Data were obtained from the annual (1993 to 2006)

spring groundfish survey conducted by the Marine

Research Institute, Iceland. The surveys were con-

ducted in March, overlapping with the beginning of

the spawning season for cod (Pálsson et al. 1989). More

than 550 fixed-location stations, covering the area of

the continental shelf of Iceland occupied by cod, were

sampled each year using a bottom trawl with 40 mm

mesh size in the codend (our Fig. 1; Pálsson et al. 1989).

At most stations a minimum of 5 and a maximum of 25

fish were sampled randomly, for which total length (L)

was measured to the nearest 1 cm, and body, gutted

(GW) and liver weights (LW) were determined to the

nearest 1 g. Otoliths were collected for identification of

individual age, sex was recorded and maturity stage

(1: immature, 2: ripening, 3: spawning, 4: spent) was

determined macroscopically according to the stages

defined in Powles (1958). In the present study, we

chose to focus on those age classes where both imma-

ture and mature cod are represented. Therefore data

for immature (n = 9907) and mature cod (n = 7354)

aged 5 yr (5I, 5M respectively) and 6 yr (6I, 6M respec-

tively) were selected.

The data were divided geographically into 4 regions;

south (S), west (W), north (N) and east (E) (Fig. 1) based

on hydrographical and ecological differences, as well

as the approximate location of the 2 frontal zones: west

of Iceland’s northwest peninsula (between regions W

and N) and in the southeast (between regions E and S).

Tagging studies (Pálsson & Thorsteinsson 2003) have

revealed at least 2 groups of Icelandic cod based on

their foraging strategies: those which almost exclu-

sively inhabit shelf waters above 200 m, and deep-

water cod which spend variable periods of time outside

of the spawning season in deeper, cooler waters and

increase their vertical movement. Therefore, the data

were separated at the 200 m depth contour and classi-

fied as either shallow or deep. Sample sizes by age,

maturity, region and depth are provided in Appendix 1.

262

Fig. 1. Station locations of Icelandic spring groundfish surveys

(1993 to 2006) for 4 statistical regions: W, N, E and S. Marked

are the 200 and 500 m depth contours

Pardoe et al.: Condition of Icelandic cod

Condition. Trends in condition were investigated

using H and Kr (Le Cren 1951), calculated as:

(1)

(2)

where GW

pred

is the predicted GW from a

length/weight relationship of the form:

GW = exp (α + β log L) (3)

where GW and L are measurements from individual

cod sampled during the period 1993 to 2006. Rather

than the custom linear regression on log-transformed

length/weight data, the relationship was estimated

using a generalised linear model (GLM) with a log link

function and gamma distribution, as GLMs avoid the

use of correction factors when backtransforming. The

gamma distribution was chosen partly because of

increasing variation in the data with length (Gud-

mundsdóttir & Steinarsson 1997, Begg & Marteins-

dóttir 2003). A single length/weight relationship (Eq. 3)

was estimated for the whole dataset so that Kr would

be comparable across ages, sexes, maturity, regions

and years. Cod of length >90 cm (n = 179) were

excluded from the analyses due to their disproportion-

ate increase in weight (Begg & Marteinsdóttir 2003),

resulting in a poor fit of the estimated length/weight

relationship to these large individuals. The condition

indices were calculated using GW rather than total

weight to ensure a more accurate assessment of trends

in condition as feeding intensity and gonad maturation

can vary significantly between individuals, locations

and years (Lambert & Dutil 1997a).

We did not use the more familiar Fulton’s K as a met-

ric of body condition as it showed an increasing trend

with length, and more importantly, the assumption of

isometric growth (b

3

in the formula K = W/L

3

) was not

satisfied despite values for slopes of L/GW relation-

ships in the range 2.95 to 3.13 (t-test; H

0

slope (b) = 3:

p > 0.05 females 5M and 6M, p < 0.05 5I, 6I, males 5M

and 6M). To test for dependency of the chosen condi-

tion indices on size, linear regressions with length

were performed. For H, the regressions were signifi-

cant (p < 0.001) for 5I (F = 44.67, 6655 df), 6I (F = 110.6,

3248 df), 5M (F = 94.36, 3093 df) and 6M (F = 108.1,

4257 df); however, all slopes were ≤0.016, and in fact

negative for 5I and 6I, along with all values of the coef-

ficient of determination being very low (r

2

< 0.03).

Length–Kr regressions were significant (p < 0.001) for

5I (F = 32.07, 6655 df) and 6I (F = 103, 3248 df) with

slopes of 0.0004 and 0.0009 and r

2

values of 0.005 and

0.03 respectively, but nonsignificant for 5M (F = 0.003,

3093 df, p = 0.95) and 6M (F = 1.735, 4257 df, p = 0.19).

The large amount of data (Appendix 1) and resultant

high statistical power likely explains the statistical sig-

nificance of these otherwise very weak relationships.

Consequently, the use of H and Kr as metrics of condi-

tion for the present study was deemed acceptable.

Stomach content analysis. Cod stomach content data

have been collected in Icelandic waters since 1979

(Pálsson & Björnsson 1993). After 1996, stomachs were

removed from most individuals for which otoliths were

sampled and therefore we examined data for the

period 1996 to 2006. Because capelin is the most

important prey item of Icelandic cod (Magnússon &

Pálsson 1989, 1991) and has been estimated to con-

tribute 80 to 90% by weight to the diet of cod in March

(Pálsson 1997), analyses of cod stomach content data

were restricted to this prey species.

In accordance with Stefánsson & Pálsson (1997),

empty stomachs were considered separately to the

amount of capelin observed in non-empty stomachs.

To account for differences in fish size, a fullness index

(FI; Schwalme & Chouinard 1999, Björnsson 2002) was

calculated for each non-empty stomach as:

(4)

where W

cap

is the weight (g) of capelin in the stomach of

an individual cod. For the annual analysis of FI, ages

5 and 6 yr were combined in order to increase the num-

ber of data points. The proportion of capelin in cod stom-

achs (P

cap

) for each region and year was estimated as:

(5)

where W

prey

is the total stomach content weight (g) and

n is the total number of non-empty stomachs in a given

region and year (Yaragina & Marshall 2000).

Statistical analysis. Individual condition data were

square-root transformed as this provided the most suc-

cessful normalisation, an assumption of ANOVA. For

these data, 1-way ANOVAs, followed by Tukey’s hon-

estly significant difference (HSD) multiple comparison

tests (where applicable), were used to identify differ-

ences in H and Kr between (1) regions, (2) depth strata,

(3) sexes of immature (5I, 6I) and mature cod (5M, 6M)

and (4) maturity status of age 5 and 6 yr cod. Tests were

performed separately for the appropriate combinations

of age, maturity status, sex and region, on combined

data for all years.

In addition, ANOVAs were used to examine the

combined influence of depth strata, region, year, bio-

logical characteristics and 4 first-order interaction

terms on average condition. Another assumption of

ANOVA is independence of observations. Intrahaul

P

W

cap

n

prey

n

=

⎛

⎝

⎜

⎞

⎠

⎟

∑

100

FI

W

GW

cap

=

⎛

⎝

⎜

⎞

⎠

⎟

Kr

GW

pred

=

⎛

⎝

⎜

⎞

⎠

⎟

H

LW

GW

=

⎛

⎝

⎞

⎠

100

263

Mar Ecol Prog Ser 362: 261–277, 2008

correlation associated with fisheries data is likely to

invalidate this to some extent (Stefánsson & Pálsson

1997 and references therein). In order to help alleviate

this problem, and reduce the likelihood of obtaining

significant effects due to the large amount of data

alone (Appendix 1), mean condition was modelled

rather than the individual observations. This data is

essentially that presented below (see Figs. 2 & 3) with

the exception that, because of their high similarity,

only combined data for immature males and females

are shown. The final models were:

mean H = μ + M

i

+ D

j

+ R

k

+ S

n

+ Y

p

+ A

q

+ (DR)

jk

+

(MR)

ik

+ (RY)

kp

+ (RA)

kq

(6)

mean Kr = μ + D

j

+ A

q

+ R

k

+ S

n

+ Y

p

+ M

i

+ (DR)

jk

+

(MR)

ik

+ (RY)

kp

+ (RA)

kq

(7)

where H and Kr are the indices of condition, μ is the

mean effect, M

i

is the maturity effect, D

j

is the depth

strata effect, R

k

is the region effect, S

n

is the sex effect,

Y

p

is the year effect, A

q

is the age effect and DR, MR,

RY and RA are first-order interactions between the

main effects. Temperature was not included in Eqs. (6)

and (7) as the combination of depth, region and year

can be thought of as a proxy for this environmental

characteristic (see Fig. 4). An additional measure of

model fit was based on a pseudo-coefficient of deter-

mination (pR

2

) which was the fraction of the total ‘vari-

ation’ explained by the model:

(8)

where deviance was analogous to the residual sums of

squares (Swartzman et al. 1995). The contribution of

each effect to the total pR

2

was also calculated.

The relationship between mean annual P

cap

and con-

dition of immature and mature cod (ages 5 and 6 yr

combined) in shallow (<200 m) and deep (>200 m)

water by region was examined using linear regression

analysis. Prior to 2002, mean weight-at-age of cod

aged 5 to 8 yr in the winter was positively related to

abundance of adult capelin in the previous summer

(Vilhjálmsson 2002). Therefore linear regression

analysis was used to determine whether similar rela-

tionships between annual mean condition of age 5 and

6 yr cod (in March) and annual adult capelin stock bio-

mass and abundance (in both August of the previous

year and January of the same year) exist.

RESULTS

Interannual variation in condition

The H and Kr time series (1993 to 2006) showed

rapid fluctuations which were often synchronous

across ages, maturity and sexes (Figs. 2 & 3). In region

W, mean annual H was relatively stable during 1993 to

2000 with values in the approximate ranges of 2–5,

4–6 and 6–8% for 5I, male 5M/6M and female 5M/6M

respectively. There was no consistent trend in Kr dur-

ing this period. A significant peak in H and Kr was

observed in region W in 2001; H reached >6% and

>10% for most immature and mature cod respectively.

By 2002–2003 condition had returned to values similar

to those observed previously, followed by a general

improvement in the most recent years. In region N,

there was an overall decline in liver condition between

1993 and 1999–2001, despite a peak of variable magni-

tude in 1996–1997. In the latter part of the time period,

an improvement in Kr was seen across most age/matu-

rity groups, whereas H was more variable and even

decreased after an additional peak in 2002–2003. In

region E there was an overall decline in condition from

1993 to 2006. Mean H of female and male 6M

decreased from 13–14 and 9–11%, respectively, to

nearer age-specific average values of 6–8%, accompa-

nied by comparable declines in the other age/maturity

groups (Fig. 2). Fish with GW > GW

pred

, GW = GW

pred

or GW < GW

pred

at a given L will have a Kr > 1, Kr = 1

and Kr < 1 respectively. In region E, mean Kr was most

often >1 prior to 2001, but values <1 were more fre-

quent in the latter part of the time period (Fig. 3). Con-

dition was most variable in region S, with the excep-

tion of 6M in shallow waters, for which H was

relatively stable during the time period (excluding the

peak in 2002) at approximately 6% and 8.5% for males

and females respectively. The greater fluctuation in

this region may be explained in part by the low sample

sizes in some years, particularly for immature cod. This

inconsistency in interannual variation across the

regions is supported by the significant interactions

between region and year in ANOVAs which model

mean H and Kr as a function of spatial, temporal and

biological characteristics (Table 1).

Variation in condition with sex, age and

maturity status

Although condition (H) of immature and mature cod

overlapped greatly, H of mature female and male cod

was significantly (p < 0.05; Fig. 2) higher than that of

immature cod for all ages and regions. Moreover,

maturity was found to be the most important factor in

an ANOVA on modelled averages (Table 1). The effect

of maturity on H was dependent to a small degree on

the region in which the cod were located, shown by the

significant interaction term (Table 1) and through com-

parison of, for example, differences in mean H of

immature and mature age 6 yr cod in shallow waters of

pR

2

1=−

⎡

⎣

⎢

⎤

⎦

⎥

Residual deviance

Null deviance

264

Pardoe et al.: Condition of Icelandic cod

regions N and E (Fig. 2). For immature cod, no signifi-

cant differences in H or Kr were detected between the

sexes (p > 0.05). However, age 5 and 6 yr mature

female cod had significantly (p < 0.001; Fig. 2) higher H

than mature males of the same age in all regions. This

was reflected in the multiple-factor ANOVAs; sex

significantly affected mean H and Kr, although it

accounted for a low proportion of the variation relative

to the other factors. Age did not significantly explain

any of the variation in mean H (Table 1).

It is important to note that despite the high degree of

significance of most individual factors in both ANOVA

models, the fraction of variation in mean Kr accounted

for by the total model (pR

2

= 0.39) was approximately

265

2

3

4

5

6

7

8

9

10

11

12

13

14

15

93 95 97 99 01 03 05

(a)

Shallow

93 95 97 99 01 03 05 93 95 97 99 01 03 05 93 95 97 99 01 03 05

2

3

4

5

6

7

8

9

10

11

12

13

14

15

93 95 97 99 01 03 05

(b)

Deep

93 95 97 99 01 03 05 93 95 97 99 01 03 05 93 95 97 99 01 03 05

2

3

4

5

6

7

8

9

10

11

12

13

14

15

93 95 97 99 01 03 05

(c)

93 95 97 99 01 03 05 93 95 97 99 01 03 05 93 95 97 99 01 03 05

2

3

4

5

6

7

8

9

10

11

12

13

14

15

93 95 97 99 01 03 05

(d)

Deep

93 95 97 99 01 03 05 93 95 97 99 01 03 05 93 95 97 99 01 03 05

Yea r

Mean H (%)

Region W Region N Region E Region S

Age 5

Age 6

Fig. 2. Gadus morhua. Interannual (1993 to 2006) and spatial variation of mean hepatosomatic index (H, %) of; age 5 yr immature

(

s sexes combined) and mature (females M and males ✳) cod in (a) shallow (<200 m) and (b) deep (>200 m) water, and age 6 yr

immature (

s sexes combined) and mature (females M and males ✳) cod in (c) shallow and (d) deep water. The horizontal dashed

line shows the overall mean for the age present in each plot

Mar Ecol Prog Ser 362: 261–277, 2008

half that explained in mean H (pR

2

= 0.78, Table 1). In

contrast to mean H, maturity did not explain any of the

variation in mean Kr (p > 0.05, Table 1). On a regional

and age basis, however, Kr of mature female and male

cod in regions N and E was significantly (p < 0.05)

higher than that of immature cod. In regions S and W,

several comparisons were also significant (region W: 6I

and male 6M; region S: 5I and male 5M, 6I and female

and male 6M) but were actually the result of immature

cod with higher mean Kr than mature cod (Fig. 3). Kr

was significantly different among the sexes of age 5 yr

mature cod in regions E and S (p < 0.05) and age 6 yr

266

0.90

0.95

1.00

1.05

1.10

1.15

1.20

1.25

0.90

0.95

1.00

1.05

1.10

1.15

1.20

1.25

0.90

0.95

1.00

1.05

1.10

1.15

1.20

1.25

0.90

0.95

1.00

1.05

1.10

1.15

1.20

1.25

93 95 97 99 01 03 05 93 95 97 99 01 03 05 93 95 97 99 01 03 05 93 95 97 99 01 03 05

Region W Region N Region E Region S

(a)

Shallow

(b)

Deep

(c)

(d)

Deep

Yea r

Mean Kr

Age 5

Age 6

Fig. 3. Gadus morhua. Interannual (1993 to 2006) and spatial variation of mean relative condition factor (Kr) of age 5 yr immature

(

s sexes combined) and mature (females M and males ✳) cod in (a) shallow (<200 m) and (b) deep (>200 m) water, and age 6 yr

immature (

s sexes combined) and mature (females M and males ✳) cod in (c) shallow and (d) deep water. The horizontal line

shows Kr = 1.0

Pardoe et al.: Condition of Icelandic cod

mature cod in regions W (p < 0.001), N and S (p < 0.05).

Age was a relatively important explanatory factor for

the variation in mean Kr (p < 0.0001, Table 1).

Spatial and bathymetric trends in condition

Spatial variability in H of Icelandic cod was evident

throughout the time period (1993 to 2006) (Fig. 2),

although annual variability obscured the differences

between regions in some years, an interaction that was

also detected statistically (Table 1). Together with

maturity and depth, region explained 55% of the total

variation in mean H (Table 1). Differences in Kr

between regions were more poorly defined (Fig. 3). In

general, cod were found to be in best condition in

region E despite the overall decline in condition in this

region, with H and Kr of both immature and mature

cod being above the age-specific average or 1.0

respectively for the majority of years. On the whole,

the poorest condition was observed in region W. Dif-

ferences in H between the west and east were signifi-

cant for all age/maturity groups (Tukey’s HSD, p <

0.05). Nonsignificant comparisons for H were those

between regions N and S for 6I, 5M and 6M and

between regions W and S for 6I. This may be due to

low numbers of 6I cod on the main spawning grounds

in the south and southwest. However, regions W and S

did show the most overlap in condition (Fig. 2). Spatial

differences in Kr were also statistically less well

defined than H; region explained a small fraction of the

total variation in mean Kr (Table 1) and significant

region comparisons showed no consistent trend be-

tween age/maturity groups. When all ages were com-

bined, however, Kr of immature fish was significantly

different between all regions whereas for mature cod

this was only true between regions W and E (Tukey’s

HSD, p < 0.05).

Cod in deep water (>200 m) had significantly better

mean H than those at depths <200 m (p < 0.05; Fig. 2),

with the only exception being 6I in region S, for which

few data were available. However, for Kr the trend was

the opposite: mean Kr was significantly higher in shal-

low water (p < 0.05; Fig. 3). Nonsignificant compar-

isons of mean Kr between the depth strata were found

for 5I in region N and 6I in region E. Additionally,

depth strata was an important contributing factor to

the explanation of total variation in mean H, and to a

lesser extent mean Kr (Table 1). The effect of depth on

mean H and Kr was somewhat dependent on the

region, as shown by the significant interaction terms

(Table 1).

Environmental effects

From measurements made during the survey, annual

mean near-bottom temperature was calculated and

showed an increasing temporal trend in all regions

(Fig. 4). All regions, particularly in the east, were char-

acterised by interannual fluctuations in temperature,

which were more or less synchronous between the

depth strata. In region W, the deep water (>200 m) was

warmer than the shallow water (<200 m) throughout

the study period, whereas in region E the opposite was

true. Temperature differences between the depth

strata in regions N and S were less well defined, with

the exception of the most recent years in the north

when the deep water remained approximately 1°C

cooler. In accordance with the known hydrographic

conditions of Icelandic waters, mean temperature was

highest in region S and decreased in a clockwise direc-

tion. Comparison of the time series of mean condition

(Figs. 2 & 3) and bottom temperature (Fig. 4) did not

reveal any obvious overlaps in interannual trends.

Prominent differences in average P

cap

between years

and particularly regions were observed. This variation

was generally similar between ages 5 and 6 yr; there-

fore the averaged trends are presented (Fig. 5). In a

given region, year and depth stratum, the magnitude

of P

cap

was generally quite similar for immature and

267

Source df SS F pR

2

H

Maturity 1 1192.13 831.90** 0.232

Depth 1 513.83 358.57** 0.100

Region 3 1131.98 263.31** 0.221

Sex 1 173.83 121.30** 0.034

Year 13 206.52 11.09** 0.040

Age 1 0.91 0.64

ns

0.000

Region × Depth 3 137.58 32.00** 0.027

Region × Maturity 3 133.95 31.16** 0.026

Region × Year 39 472.51 8.45** 0.092

Region × Age 3 13.89 3.23* 0.003

Residual error 806 1155.01 0.775

Kr

Depth 1 0.15 98.83** 0.074

Age 1 0.06 36.57** 0.028

Region 3 0.08 17.85** 0.040

Sex 1 0.01 6.16* 0.005

Year 13 0.09 4.45** 0.044

Maturity 1 0.00 1.35

ns

0.001

Region × Depth 3 0.10 21.70** 0.049

Region × Maturity 3 0.08 17.89** 0.040

Region × Year 39 0.22 3.74** 0.110

Region × Age 3 0.00 0.70

ns

0.002

Residual error 806 1.24 0.393

Table 1. Gadus morhua. Multiple-factor ANOVA examining

the influence of main effects and first-order interaction terms

on average condition (hepatosomatic index, H and relative

condition factor, Kr) of Icelandic cod from 1993 to 2006. *p <

0.05, **p < 0.0001, ns = not significant

Mar Ecol Prog Ser 362: 261–277, 2008

mature cod. Feeding on capelin, at the

time of the survey, was highly variable

in region W. For immature and mature

cod in shallow (<200 m) water, good

capelin feeding years in this region

(‘good’ defined here as P

cap

≥ 40%)

were 1997, 2001 and 2002 along with

1996 for mature individuals. In deep

(>200 m) water, this can only be said

for immature cod in 2001. For the

majority of cod in region W,

1998–1999 and 2003–2005 were very

poor (P

cap

< 10%) capelin feeding

years, along with 2006 in deep water.

Feeding on capelin, particularly in

deep water, was markedly lower in

region S than regions N and E in

almost all years. In region S, P

cap

was

consistently ≥30% and <12% in shal-

low and deep water, respectively. In

region N, P

cap

of immature cod in shal-

low and deep water ranged between

8–100% and 32–100%, respectively,

and P

cap

of mature cod varied between

268

93 94 95 96 97 98 99 00 01 02 03 04 05 06

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

Region W

Mean temperature (°C)

Shallow

Deep

93 949596 97 989900 010203 0405 06

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

Region N

93 94 95 96 97 98 99 00 01 02 03 04 05 06

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

Region E

93 949596 97 989900 010203 0405 06

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

8.0

Region S

Yea r

Fig. 4. Annual (1993 to 2006) mean (± 95% CI) near-bottom water temperature

(°C) of 2 depth strata (shallow: <200 m, deep: >200 m) in 4 statistical regions of

the Icelandic shelf. Note different scales on y-axes

1996 1998 2000 2002 2004 2006

0

10

20

30

40

50

60

70

80

90

100

(a) Immature, shallow

1996 1998 2000 2002 2004 2006

0

10

20

30

40

50

60

70

80

90

100

Region W

Region N

Region E

Region S

(b) Immature, deep

1996 1998 2000 2002 2004 2006

0

10

20

30

40

50

60

70

80

90

100

(c) Mature, shallow

*

1996 1998 2000 2002 2004 2006

0

10

20

30

40

50

60

70

80

90

100

(d) Mature, deep

Year

P

cap

(%)

Fig. 5. Gadus morhua. Time series (1996 to 2006) of average proportion of capelin (P

cap

) in stomachs of immature Icelandic cod in

(a) shallow (<200 m) and (b) deep (>200 m) water, and mature Icelandic cod in (c) shallow and (d) deep water in 4 regions

(W, N, E and S) of the Icelandic shelf.

*

Missing data for 1996 (c)

Pardoe et al.: Condition of Icelandic cod

1–83% and 27–98%, respectively. Similarly, P

cap

of

immature cod in shallow and deep water of region E

ranged between 17–71% and 12–70%, respectively,

and P

cap

of mature cod varied between 20–64% and

12–72%, respectively. Annual mean capelin feeding in

region N was quite stable, with good capelin feeding

predominating in the majority of years, particularly in

deep water. Overall, P

cap

in the north was lowest from

1999–2001, but was particularly good (>50%) in 1996

in both depth strata, along with 2005–2006 in shallow

water. A general decline in capelin feeding in region E

was observed, perhaps with the exception of mature

cod in shallow water. In 1996–1997, P

cap

of immature

cod in shallow water was >63%, and in deep water

was >70% and >49% for immature and mature cod

respectively. However, by 2005–2006 P

cap

had fallen to

<27% for immature cod in shallow water and <32%

for immature and mature cod in deep water.

Interestingly, significant (p ≤ 0.05) positive relation-

ships between mean annual condition (H and Kr) and

mean annual P

cap

of immature and mature cod in shal-

low (<200 m) and deep (>200 m) water existed only in

regions N and E (Table 2). Such a relationship was not

found between the capelin stock and overall condition

of Icelandic cod; all regressions of mean annual condi-

tion (H and Kr, all regions combined) of age 5 and 6 yr

cod and annual adult capelin stock biomass or abun-

dance (for both August of the previous year and Janu-

ary of the same year) were nonsignificant (p > 0.4).

Analysis of relative capelin biomass in stomachs (FI,

%) revealed pronounced interannual and spatial varia-

tion in capelin consumption by immature and mature

cod (Fig. 6). Although unequal sample coverage makes

it difficult to make systematic comparisons, it could be

concluded that on the whole FI of immature cod was

greater than that of mature cod. Feeding on capelin

appeared to be most intensive at or

near the 200 m contour line in the

northwest and beyond this depth in

the north and east. For both immature

and mature cod, FI was particularly

high in the northwest from 2005–2006,

coinciding with a ‘good’ P

cap

year in

region N as well as a general improve-

ment in Kr after 2003–2004 (Figs. 3, 5

& 6). Patterns of FI in the east did not

clearly follow those observed for con-

dition and P

cap

; it could be interpreted

that the number of cod with empty

stomachs increased in region E during

the time period (Fig. 6), but as with all

interpretation of this stomach content

data, interannual and spatial variabil-

ity in the distribution of data points

would need to be taken into account

before such a conclusion could be made reliably. Cod

feeding on prey species other than capelin did not

show any notable spatial or temporal trends.

High densities of cod with empty stomachs or those

that had not eaten any capelin (Fig. 6) might explain

the relatively poor condition in the west throughout the

time series. FI was particularly low in 1998, 1999, 2004

and 2005; Kr of mature 6 yr old cod also reached low-

est values in shallow water of region W in 1999, but

there were no other distinctive correlations between FI

and condition in these years. The peak in condition

(Figs. 2 & 3) and P

cap

(Fig. 5) in region W in 2001 does

correspond to an atypical number of samples with FI >

10% in depths of <200 m. However, a similar event oc-

curred in 1997 but H and Kr did not improve markedly

in this year.

DISCUSSION

We observed high interannual variability in H of Ice-

landic cod, along with variation associated with geo-

graphic location, depth and temperature. Surprisingly,

H and Kr showed very little agreement. Consumption

of capelin was found to be related to some of these

temporal and spatial trends. In addition, significant

relationships between H and biological characteristics

were detected; mature cod were found to be in signifi-

cantly better condition than immature cod of the same

age, and the mean H of mature females was greater

than that of mature males. Conversely, many of these

trends were much weaker or not detected by Kr.

Capelin is the most important prey item of Icelandic

cod (Magnússon & Pálsson 1989, 1991), although its

contribution by weight to their diet varies between 80

and 90% during the capelin spawning season in

269

Condition Maturity Region Depth Regression r

2

p

index

H (%) I N S H = 2.17 + 0.05 × P

cap

0.68 <0.01

H (%) I E S H = 4.66 + 0.05 × P

cap

0.57 <0.01

H (%) I N D H = 4.14 + 0.07 × P

cap

0.64 <0.01

H (%) M N S H = 4.82 + 0.02 × P

cap

0.40 0.05

H (%) M N D H = 7.13 + 0.04 × P

cap

0.51 <0.05

Kr I N S Kr = 0.95 + 0.001 × P

cap

0.74 <0.001

Kr I E S Kr = 0.95 + 0.001 × P

cap

0.71 <0.01

Kr I E D Kr = 0.95 + 0.001 × P

cap

0.35 0.05

Kr M N S Kr = 0.97 + 0.001 × P

cap

0.62 <0.01

Kr M E S Kr = 0.98 + 0.001 × P

cap

0.37 <0.05

Kr M N D Kr = 0.94 + 0.001 × P

cap

0.52 <0.05

Table 2. Gadus morhua. Regression equations for significant relationships

between the mean annual (1996 to 2006) proportion of capelin in stomachs (P

cap

)

and mean annual condition (hepatosomatic index, H and relative condition

factor, Kr) of Icelandic cod (I = immature, M = mature) in 2 depth strata (S =

shallow, <200 m and D = deep, >200 m)

Mar Ecol Prog Ser 362: 261–277, 2008270

E

EE

E

EE

E

EEEEEEEEEEEEEEEE

E

EE

EEEEE

E

EE

E

*

****

*

**

*

****

**

*

***

*

****

*

**

*

***

*

**

*

**

*

**

*

******

**

***

1996

E

EEE

E

EE

E

EE

E

EE

E

EE

E

EEEE

E

EE

E

EE

E

EE

E

EEEE

E

EEEEEE

E

EEE

EE

E

EE

E

EEE

EE

E

EE

E

EEE

E

EEE

E

*

***

*

**

*

***

**

*

**

*

**

*********

*

**

*

**

***

*

***

*

***

*

***

*

***

**

***

*

***

*

**

*

***

**

*

**

*

**

*

**

*

**

*

**

***

*

***

**

*

***

*

**

*

**

*

**

*

**

*

1997

E

EE

E

EE

E

EE

E

EEE

E

EE

E

EEEE

EE

E

EE

E

EEE

E

EEE

EE

E

EEE

EEEE

EE

E

EE

E

EE

E

EE

EEE

E

EE

EEEEEE

E

EE

E

EE

EEEEEEEE

EE

E

EEEEE

E

EEE

E

EEEEE

E

EEE

E

EE

E

EE

E

EE

E

EEEEEE

E

EEE

EEEE

E

EEEEEEEEE

E

EE

E

EEE

*

**

*

***

*

**

*

***

*

**

*

**

*

****

*

**

*

***

**

*

**

*

***

**

*

*********

*

**

*

**

*

***

*

**

*

***

*

**

***

****

**

*

****

*

**

*

**

*

****

**

*

**

***

**

*

**

***

*

**

*

**

*

**

***

***********

****

**

*********

**

*

*****

***********

*

****

*

**************

***********

*

**

*

*****

*

**

*

****

*

**

*

**

*

1

5

1998

EE

EEE

E

EEE

E

EEE

E

EE

E

EE

E

EEE

E

EE

E

EE

E

EEE

E

EE

E

EEE

E

EE

EEE

E

EE

E

EE

E

EEE

EE

E

EE

E

EE

E

*

**

****

*

**

*

**

*

***

*

***

*

***

****

*

***

*

***

*

**

*

**

*

**

***

*

***

**

*

**

*

****

*

*******

***

*

****

*

******

***

*

***

**

****

*

*****

*

**

*

**

*

***********

*

**

*

**

*

******

**

*

**

*

**

*

**

******

*

******

*

**

****

**

*

**

****

*

***

*

1999

E

EE

E

EE

E

EE

E

EEEEE

EEE

E

EE

EEE

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EEEE

EEEEEEEEEEEE

EE

EEEE

EE

EEEE

EE

E

EE

E

EEE

E

EE

E

EE

E

EE

EEEE

E

EEE

E

EEE

E

EE

E

**

*

**

*

***

*

**

*

**

*

**

*

**

*

**

*

**

***

**

***

*

**

***

*

****

*

**

*****

*

***

**

***

**

*

**

*

**

*

***

*

**

*

**

*

**

*

****

*

***

*

**

*

**

*

**

****

**********

**

****

********

**

****

**

**********

****

**

****

**

****************

************

**********

**

****

**

**********

**

****

**

*

**

****

**

*****

******

*

****

*

**

****

*

****

***

**

*

**

*

**

****

**

*

**

****

*

*******

***

*

****

*

**

***

**

*

**

*

****

*

****

**

*

****

*

**

***

**

*

****

*

**

***

*

***

**

*

***

*

2000

E

EE

E

EE

E

EE

E

*

**

*

***

*

**

*

***

*

**

*

***

**

*

**

*

**

*

**

*

2001

EE

E

EE

E

EE

E

EEE

E

EE

E

EEE

E

EE

E

EEE

EEEEE

E

EE

E

EEEE

E

EE

E

EEE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

**

*

**

*

**

*

**

***

**

*

**

*

**

***

**

***

*

**

*

**

*

**

*

**

*

***

**

*

**

*

**

*

**

*

**

*

**

*

**

*

**

*

*******

****

**

*

**

***

*

****

*

*****

*

**

*

**

***

*

**

*

***

**

*

**

2002

E

EE

E

EE

EEEE

E

EE

E

EE

E

EE

E

EE

E

EEEE

EE

E

EE

E

EEE

EE

E

EE

EEE

EE

E

EE

EEE

E

EE

E

EE

E

EEEE

E

EE

E

EEEE

E

EE

EEEEEEEE

EEEEE

EEEE

E

EEEEEEEEEEEE

EEEEEEEEEEEEEE

EEEE

EE

EEE

EEEE

EE

EEEE

E

EE

E

EE

E

*

**

*

***

**

*

**

*

*****

*

**

*

**

*

**

*

***

*

**

*

**

*

***

**

*

**

*

***

*

**

*

***

*

**

*

**

*

**

*

**

*

**

*

**

*

***

*

**

*

**

*

***

*

******

*

**

*

**

***

**

***

*

******

*

**

*

**

*

**

*

*******

**

*

**

*

**

*

******

*

**

*

******

*

**

*

***

*

***********

*

**

*

****

*

**

*

***

**

*

**

*

***

*

***

*

**

*

2003

E

EE

E

EE

E

EE

EEE

E

EEEEE

EEE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

EEE

E

EE

E

EE

E

EEEEEE

E

EEE

E

EEE

E

EE

EEEEE

E

EE

E

EE

E

EEE

EE

E

EEE

E

EE

E

EE

E

EEE

E

EE

E

EEE

EE

E

EE

E

EE

E

EEE

E

EEE

E

EE

E

EE

E

*

***

*

***

*****

*

**

*

**

*

****

*

***

*

**

*

**

*

**

*

**

*

**

*

*****

*

**

*

***

*

**

*

**

*****

*

***

*

**

*

***

*

**

***

*

**

*

**

*

**

*

***

*

**

*

**

*******

*

***

*

*****

**

*

****

**

*

**

*

*****

*

***

*

**

*

**

*******

*************

*

*******

**

****

**

*******

**

*

**

*

**

*

***

*

**

*

**

****

*

****

*

**

***

**

*

**

***

*****

*

**

*

*******

***

***********

**

*

2004

24°W 18° 12°

EE

EEE

E

EEE

E

EEEE

E

EE

E

EEE

E

EE

E

EE

E

EE

E

EE

E

EEE

E

EE

E

EE

E

EE

EEEE

EEEEEEE

EE

EEE

E

EE

E

EEEE

EE

E

EEE

E

EEE

E

EE

EEE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EEE

EE

E

EE

EEEE

E

EE

E

EEEE

E

EEEEEE

EE

EEEEEEE

EEE

E

EE

E

EEE

E

EEE

E

EE

E

EEEE

E

EE

E

EEEEEEEEE

EE

EEE

E

EEEE

E

EEE

EE

EEE

EEEEEE

E

EE

E

EEEEEEEEEEE

EE

EEEEEE

EEE

EE

EEE

E

EEEEEEEE

EEEE

EEE

E

EEEEE

E

EEEE

E

EEEE

E

EEE

E

EEEEEE

EEEEE

EEEEEE

EE

E

*

****

*

**

*

**

*

**

*

**

*

**

*

***

*

**

*

**

*

***

*

**

****

*

***

*

**

*

**

*

**

*

**

*

**

*

***

*

************

*****

*

******

*

**

*

**

*

**

****

*

***

*****

**

*

**

*

***

**

*

********

*

****

*

**

*

***

*

****

*

**

*

****

*

**

***

*

**

*

**

*

*******

**

*

**

*

*****

*

*****

*

****

*

**

*

****

*

***

*

**

*

**

****

*

***

*****

***

*

**

*

2005

24°W 18°

12°

E

EEE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EE

E

EEE

E

EE

E

EE

E

EE

E

EE

E

EEE

E

EEE

E

EEE

E

EEEEEEEEE

E

EE

EEEEEEE

E

EE

EEEEEEEE

E

EE

E

EEEEE

E

EEE

E

EE

E

EEE

E

EE

E

EE

E

EEEEEEEEEE

E

EEEE

EEE

E

*

**

*

**

*

***

*

***

*

**

*

**

*

**

*

**

*

**

*

**

*

*****

*

**

*

**

***

*

***

*

***

*

******

*

**

*

**

*

**

*

***

*

**

*

**

*

***

*

**

*

***

*

***

*

*******

*****

*

***

****

**

********

**

******

***

****

*

**

*

**

*

****

*

**

******

*

***

*

**

*

**

2006

24°W 18° 12°

(a) Immature

10

50

63°

64°

65°

66°

67°

63°

64°

65°

66°

67°

63°

N

64°

65°

66°

67°

Fig. 6. Gadus morhua. Annual relative capelin biomass (FI: fullness index, %) in non-empty stomachs of (a) immature and (b) mature Icelandic cod. *Cod with non-empty

stomachs but FI = 0. E = cod with empty stomachs. The 200 m contour line is marked

Pardoe et al.: Condition of Icelandic cod

271

E

EEEE

E

EE

EEE

E

EE

EEEE

E

EE

E

EE

EEEE

EE

EEEEE

E

EEEEE

EEEEEEE

E

EEE

E

**

*

**

*******

**

*

**

*

***

*

**

*

**

****

*****

****

**

****

***

**

*

**

*

**

*

**

*

**

*

**

1996

63°

64°

65°

66°

67°

E

EE

E

EE

E

EEE

E

EE

E

EE

E

EE

E

EEE

E

EE

E

EEE

EE

E

EE

E

EE

E

*

**

*

**

*

**

*

**

*

**

***

*****

*

**

*

**

*

**

*

***

**

*

**

*

Spatial and temporal trends in condition of Atlantic cod Gadus morhua on the Icelandic shelf

Abstract and Figures