International Council for
the Exploration of the Sea
ICES C.M. 19951P: 14
Theme Session on Causes of
Observed Variation in Fish Growth
Factors influencing weight at age of cod off eastern Newfoundland (NAFO Divisions 2J+3KL)
Peter A. Shelton and George R. Lilly
Science Branch, Department of Fisheries and Oceans
Northwest Atlantic Fisheries Centre, P.O. Box 5667
St John's, Newfoundland
Canada AIC 5Xl
Abstract
•
I
{eI ~c.Jv.J "tlo.Lc.u}
/(c,'Jl.lfckn 1J; tI<f)!v> t f/
Growth increments and condition factor in the northern cod stock (~AFO Divs. 2J, 3K and
3L) measured from research vesse1 trawl sampies show considerable variabllity over the period
1978 to 1994. This variability in the data is examined through a variety 9f plots. Scatter plots of
growth and condition data against temperature, cod biomass and capelin'abundance in cod
stomachs are used to look for possible corre1ations. A near-decadel scale cycle in the area of the
cold intermediate layer (CIL) is correlated with the cycle in cod growth. The generally declining
trend in temperature is correlated with the declining trend in growth. Correlations between the
abundance of capelin in cod stomachs and cod growth or condition are not easy to interpret and
probably require more attention to temporal and spatial patterns of abundance of predator and prey
to resolve. Correlations with cod biomass are weak and, with the exception of Div. 3L, are
positive, Le. do not provide evidence of density dependent growth. It is important to take changes
in growth into account in future management of this cod stock and predictive models based on
variables such as the area ofthe.CIL would be very useful.
Introduction
Growth of Atlantic cod (Gadus morlllla) varies considerably among stocks, with much of
the differcnce attributable to variability in ambicnt temperature (Brander 1995). There is also
persistent geographie variability within some stocks, such as those around Iceland (J6nsson 1965)
and off Labrador and eastern Newfoundland (Fleming 1960; May et al. 1965). Within the latter
region there has historically been a cline in growth, from very slow growth off central Labrador to
relatively rapid growth on the southwestern Grand Bank.
Annual variability in growth has been reported for many eod stocks, with important
implieations for fisheries management. Studies of faetors responsible for this variability have
emphasized temperature, stock size and the abundance of prey. In the stock of eod occupying
Northwest Atlantie Fisheries Organisation (NAFO) Divisions 2J, 3K and 3L (eommonly ealled the
northem eod stock), ehanges in length-at-age of eod were found to be affeeted by both the biomass
ofthe eod stock (WeHs 1984, 1986; MilIar and Myers 1990) and temperature (MilIar and Myers
1990). There has also been cancern that ead growth might be affected by changcs in abundanec of
capelin, the most important prey of eod on the Northcast Newfoundland Shclf (Lilly 1987).
Declincs in eapelin abundance have becn eorrelated with reduetions in growlh rate of eod in waters
around Iccland (Steinarsson and Stcfansson 1991) and with a reduction in bath growth rate and
}
•
2
somatie eondition of eod in the Barents Sea (Mehl and Sunnanill991; Jprgensen 1992).
However, neither Akenhead et al. (1982) nor Millar et al. (1990) eould find a statistieally
significant relationship between cod growth and capelin abundance off Newfoundland. The lack
of statistical signifieance might reflect weak linkage between cod and its prey, but might also refleet
an inability to detect a link because of uncertainty in the timeseries of capelin abundance (Akenhead
et al. 1982; Shelton et al. 1991).
In 1990 the ICES Multispecies Assessment \Vorking Group undertook an analysis of
growth in cod from four Arcticlboreal systems (Barents Sea, Iceland, Greenland and
Newfoundland, Anon 1991). Length at age and length increments were examined for stock
density, capelin abundance and temperature effects. A significant residual year effect was found in
most of the data sets after accounting for these three factors. At a subsequent meeting of the
\Vorking Group (Anon 1992) cod stornach content data for the North Sea, Baltic Sea, Barents Sea,
Iceland, Newfoundland and Northeast USA were compiled into a common format and analysed in
some detail. At this same meeting further, but preliminary analysis of the cod growth data base
was carried out taking into account the information in the cod stornach content data base. It was
conc1uded (Anon 1992) that in the eomparison among ecosystems "the relationship between
growth and feeding remains confounded by the inter-dependencies of growth, temperaturedependent stornach evacuation and differences in environmental temperature".
Brander (1995) extended the analysis of the temperature effect to 17 cod stocks and
conc1uded that 92% of the variance in the logarithm of mean weight at age for ages 2 to 4 fish
among stocks can be explained by a ANCOVA model with age (dass) and temperature
(eontinuous) effects. The weight at age data used in Brander's study eame mainly from the
sampling of the commercial eatch. For the stock off Labrador and the east eoast of Newfoundland,
and possibly for other stocks, the weight at age in the commercial eatch is inferred from a eonstant
length-weight relationship. Although the comparisons among cod stocks by Anon (1991, 1992)
and Brander (1995) has much merit, we believe that more detailed intra-stock analyses are required
if the combined and probably interacting effects of temperature, food, population density and
fishing mortality on cod growth are going to be disentangled.
In this paper we return to the problem of explaining cod growth variability off southern
Labrador and the cast eoast of Newfoundland in further detail. In particular we attempt to (i)
account for the length-stratified approach to the subsampling of the research trawl catch, (ii)
consider thermal measures additional to those made at Station 27, (iii) give greater consideration to
weight at age data, (iv) evaluate condition factor data, and (iv) give further consideration to the
possible effects of capelin abundance in cod stomachs on cod growth. The approach in this study
remains primarily descriptive and exploratory. Formal hypothesis testing and statistical inference
leading to predictive models ofuse in stock assessments have been attempted in the past (e.g.
Miliar and Myers 1990) and are an objective for future studies by ourselves. Strong signals in
ocean temperature data (e.g. Colbourne 1995), eod population size (see Bishop et al. 1995), the
abundance of capelin in cod stomachs (Lilly 1993, 1994, 1995), and possibly the abundance of
capelin in the ocean (Carscadden 1994) should facilitate statistical inference, however, before
proceeding further with such analyses, wc believc that a solid base of data description and
exploration is required. Although the present paper docs not complcte this task far thc southcrn
Labrador and cast coast of Newfoundland region, wc hope it is a further step in this direction.
•
3
Data sources
Temperature measurements
Temperature profiles from surface to bottom (175m) are routincly taken throughout the year
at Station 27 just off St 10hn's (Fig. 1) at the commencement and termination of nearly every
research cruise (Colbourne et al. 1994, Colbourne 1995). For this analysis all data for the period
1978 to 1994 have been used to compute annual average water column temperature. In addition,
the bottom temperature anomaly from the longterm mean (1961-90), following MilIar and Myers
(1990), has been considered.
•
Oceanographic transects across the Hamilton Bank (Seal Island) in Div. 21, off Bonavista
Bay across the Div. 3K1 Div. 3L boundary, and in Div. 3L shorewards from the Flemish Cap along
the 47 0 N latitude have been surveyed regularly in summer since the 1950s (Colbourne et al. 1994,
Colbourne 1995). Annual estimates of the area of the cold intermediate layer (CIL), defined as
~O°C, have been caIculated as an index of thermal conditions within the northern cod habitat
(Coibourne 1995) and are used in this stu~y.
Estimates of cod abundance
•
Two sources of information on cod abundance were considered - research vessel (RV)
estimates for each division and estimates from sequential population analysis. RV surveys have
been conducted by Canada during the fall in Divs. 21, 3K and 3L since 1977, 1978 and 1981
respectively (excluding 1984 in Div. 3L). All surveys in Divs. 21 and 3K were conducted with the
74 m stern trawler R.~ 'Gadus Atlantica'. Surveys in Divs. 3L were conducted with the 51 m
side trawler R.~ 'A. T. Cameron' and the sister 50 m stern trawlers R.~ 'Wilfred Templernan'
and R.~ 'Alfred Needler'. The 'Gadus Atlantica', 'Wilfred Templernan' and 'Alfred Needler'
deployed an Engcl-145 trawl, whereas the 'A. T. Cameron' deployed a Yankee 41-5 trawl. In all
instances, a 29 mm meshliner was inserted in the codend. Tows were made at 3.5 knots for 30
min at each fishing station, and catches from the few tows of duration other than 30 min were
appropriatcly adjusted. No adjustments were made for possible between-vessel differences in
catching efficieney. Details regarding areas and 10eations of strata and changes in survey pattern
are provided by Bishop et al. (1994), Lilly and Davis (1993) and Bishop (1994). The most
notable change in survey eoverage was the addition of depths between 100 and 200 m in
northwestern Division 3K (St. Anthony Shelf and Grey Islands Shelf) in 1984 and subsequent
years. Fishing in all divisions and years was eonducted on a 24-h basis. Set alloeation within
each stratum is random, with the number of sets determined primarily by the area of the stratum.
Boundaries to strata correspond to water depth eontours. The RV estimates of trawlable biomass
by division obtained from the program STRAP (Smith and Somerton 1981) are provided in Bishop
et al. (1995).
Aversion of s~quential population analysis, ADAPT, has been routinely applied in northern
eod assessments to eommercial eatch at age data and tuned using the RV estimates of abundanee
from STRAP. ADAPT estimates ofbiomass in Divs. 2J3KL eombined are provided in Bishop et
aI. (1993).
4
Length and weight at age sampIes from fall RV surveys
The number of cod sampled for which length and body weight measurement are available
and for which age was determined directly from the otolith (rather than an age-length key) varies
across years and among ages (Tables 1-3). During the early years of the surveys, otoliths for
aging were obtained from a sampIe of 25 cod per 3-cm length-group per division. An additional
sampIe of 5 cod per 3-cm length-group was frozen at sea and thawed in the laboratory, where
otoliths were extracted and body weight (both whole and gutted, head-on) was recorded. All fish
were measured (fork length, cm) at sea. Additional sampIes of frozen fish were collected in Div.
21 in 1984 and 1985. The number of frozen fish increased in 1988 and subsequent years, when
Divs. 21 and 3K were both subdivided into 2 areas, and a sampIe of 5 cod per 3-cm length-group
was obtained from each area. The number of cod for which weights were available increased
dramatically when weighing at sea was initiated (in 1989 in Divs. 21 and 3K and in 1990 in Div.
3L, Table 2). Length-at-age was determined for a total of 30,535 cod (Table 1) over the study
period. Body weights were obtained for 13,822 of these cod (Table 2).
•
The cod for which length and weight at age were determined represent a subsampIe of the
trawl catch at each station. The subsampIe is length stratified - where possible 25 fish are collected
from each 3 cm length interval in each of the three divisions over the course of the survey. From
1989 onwards in Divs. 2J and 3K and from 1990 onwards in Div. 3L body weights have been
recorded at sea prior to freezing and these weights have replaced the thawed weight previously
entered into the data base. Although these two weights are expected to differ systematically, a
correction has not yet been applied (although the data are avialable to do so) and in the present
study the differe~ce has bcen ignored.
In addition to the subsampIe, the length frequency of the entire catch for each set, or a
random portion of the catch if it is too large to process, is also determined. The sampie length
frcquency is transformed into a population length frequency in each division using the program
STRAP (Smith and Somerton 1981). In this study population mean length and weight at age by
division were obtained by weighting the individual measurements in the bio10gical sampIe by the
ratio of the sampIe number per 3 cm length dass to the STRAP estimate of the number in the
population in each 3 cm length dass.
The average number of length and weight sampIes by age for the period 1978 to 1994 are
summarised in Table 3. It is evident that after age 10 the number of length sampIes is considerab1y
reduced «20 per year on average). The number of sampIes for weight is 10w «20 per year) after
about age 8. For this reason the present analysis was restricted to ages 10 and 1ess for length at
age and ages 8 and less for weight at age. Although the number of sampIes are also low for ages
less than 3, there is relatively little variance in length and weight at age among sampIes within a
year, so ages 1 and 2 have been induded in the analysis.
Cod condition measurements from fall RV sampIes of length and weight at age
Condition was cxprcssed as Fulton's condition factor:
K = 100*(\V/L3)
where \V is body weight (g) and L is fork length (ern). 80th whole body (round) weight and
guUed (head on) weight were used. Round weight is of greater significance when considering the
impact of growth on stock assessments and the TAC. Gutted weight,with liver and gonad weight
measured separately, are of greater significance when considering aspects of bioenergetics and are
•
5
eurrently subjects of further study not reported here. Beeause growth in eod is not isometrie,
annual variability in eondition was eompared within speeifie eod length-groups rather than within
age-groups.
Cod stornach contents from fall RV surveys
•
Stomaehs were eolleeted from up to 3 randomly seleeted eod per lO-em length-group per
station in 1980-1982 and 3 per 9-em length group in 1983-1994. Stomaehs were not eolleeted
from fish which showed signs of regurgitation, such as food in the mouth or a flaecid stomaeh.
Stomaehs were individually tagged and excised. In 1978-1993, the stomaehs were fixed and
preserved in 4% formaldehyde solution in semvater prior to examination of their eontents in the
laboratory. In 1994, the stomaehs were frozen prior to examination.
Examination involved separation of food items into taxonomie eategories. Fish and
deeapod erustaeea were identified to species, but most other groups were assigned to higher order
taxa. Hems in eaeh taxon were plaeed brieflyon absorbent paper to remove exeess liquid, and then
eounted and weighed to the nearest 0.1 g.
The quantity of eapelin in the stomaehs of the eod from a specified sampie was expressed
as a mean partial fullness index (Fahrig et aI. 1993):
•
where Wej is the weight (g) of eapelin in fish j, Lj is the length (ern) of fish j, and n is the number
of fish in the sampie. This index is based on the assumption that stomaeh capacity is apower
funetion of length, and is analogous to Fulton's eondition faetor. Mean total fullness index was
ealeulated as
I n W
TFI=-L~·104
n'_
L3
J- 1 j
where \Vlj is the total weight of prey in fish j.
For simplieity, the present analysis was restrieted to eod within the 36-71 em length range,
and all eod within this range were pooled. Cod smaller than about 30-35 cm cannot feed on the
largest capelin and eod larger than about 70 em tend to feed to an inereasing extent on groundfish
and erabs (Lilly 1991).
Capelin abundanee estimates
Estimates of capelin biomass from hydroaeoustie surveys condueted by Canada and Russia
during the fall in Divs. 2J and 3K are available intermittently from 1974 onwards (Miller 1994).
Considerable uneertainty exists regarding some of the estimates, partieularly those in recent years
(see for example Carseadden 1994).
6
Data description
Temperature
\Vater temperatures on the shelf off Labrador and Newfoundland have been below the
longterm mean for most of the period since 1980, with two unusually cold periods in the mid1980s and early 1990s (CoIbourne et al. 1994). These recent two cold periods separated by a
warmer period ean be c1early seen in the annual average water eolumn temperature und bottom
temperature anomaly plotted in Fig. 2. Underlying this near-deeadel seale variability is longer
period variability apparent as a downward trend in temperature over the study period with eurrent
values weil below the relatively warm 1950s and 1960s (CoIbourne 1995). The near-deeadel seale
variability is c1early illustrated in the unnual variability in the area of the CIL calculated along the
three transeets within the study area Fig. 3). As weil, the upward trend in the area of the CIL,
eorresponding to the eooling trend in Fig. 2, ean be seen.
•
Capelin
Estimates of eapelin biomass from hydroaeoustie surveys eondueted by Canada and Russia
during the fall in Divs. 21 and 3K deelined to very low levels in the early 1990s with the Canadiun
series showing a very abrupt drop between 1989 and 1990. Low levels of eapelin in the acoustie
fall survey in Divs. 21 and 3K, as weH as in the spring survey in Div. 3L have persisted in
subsequent years. The 1983 und 1986 eapelin year c1asses are thought to havc been particularly
strong (Carseadden 1994).
Peaks in eapelin PFI in eod stomaehs (Fig. 5) in Div. 21 oecur in the mid 1980s and in the
late 1980s, eorresponding to the high values in the Canadian hydroacoustie survey. Subsequently
the eapelin PFI dropped to a very low level und their was a eorresponding drop in TA. In Div. 3K
eapelin PFI increased over the 1980s and was fairly eonstant over the early 1990s, dropping to a
low level only in the most reeent year for which data are available (1994). The eapelin PFI in Div.
3L increased in 1990 and has remained at a high level up to und including the most reeent survey.
It is c1ear that there is little compensation in the eontribution of other food in the diet of eod and that
the trends in TFI are therefore driven mostly by the eapelin contribution (Lilly 1991).
It is important to eonsider changing geographie patterns of eod and eapelin abundanee in
the interpretation of these data (Lilly 1994). Despite the apparent low abundanee of eapelin in the
1990s, eod in Divs. 3K und 3L had a relatively high eapelin PFI, at least partly beeause the eapelin
changed their distribution and occupied the restricted area whcre bulk of thc remaining eod
population was eneountered during the fall RV survey. The low eapelin PFI in Div. 21
corresponds to a very low abundanee of eod in this division after 1989.
Cod abundanee
The time series of eod abundance in fall RV surveys and from the ADAPT estimates are
illustrated in Figs. 6 and 7. The RV estimates fluetuated with no apparent trend up to 1988. High
estimates in 1986 in Divs.21 and 3K relative to estimates before und after are not easily interpretcd.
Thc progressive dccline in Div. 2J, Div. 3K und Div. 3L in the Jute 1980s und 1990s to the eurrent
low biomass is c1early seen. The ADAPT time series is smoother and suggests that the biomass
rose to a peak in 1984 following extension of jurisdiction and then declined monotonically to 1992,
•
7
after which sequential population analysis has not been applied. These perspectives on the changes
in cod stock size, although not independent because of the tuning exercise, are somewhat different
and both have been considered in the data exploration exercise that follows.
Cohort length at age
•
The cohort length at age for Divs. 2J, 3K and 3L are plotted in Fig.8 for ages ~ 10. Note
that because the point estimates are from the fall of each year and growth is slow over the winter,
most of the increment in length or weight is occurring in the following year. Cohorts are taken
back to age 0 at which age length is assumed to be 0 to facilitate the identification of cohorts in the
plots. Lines roughly parallel to the x-axis identify the ages whereas lines roughly parallel to the yaxis indicate cohorts. Ridges corresponding to fast growing cohorts and valleys corresponding to
slow growing cohorts can be clearly seen in the plot fer Div. 2J and to a lesser extent in the plot for
Div. 3K. The Div. 3L "surface" is relatively flat.
The ridges in the cohort growth correspond to fast growing 1976-78 cohorts and fast
growing cohorts from the early to mid 1980s. The elevated length increment does not occur in all
ages in these cohorts but by age 4 the effect of the cumulative incremcnt is clear and remains
distinct to age 6, after which the general slowing down in growth and increasing mcasurement
eITor tend to obscure the pattern. The valleys correspond to slow growing cohorts from the mid
1970s,late 1970s-early 1980s and late 1980s. The most marked declines in growth in these
cohorts (lowest points to the valleys) occurred in the mid 1980s and early 1990s, primarily within
ages 2 to 5, particularly in Div. 2J but also in Div. 3K. Year-to-year variability is much less
pronounced in the Div. 3L data.
In addition to these cohort and year effects in the growth, it is evident that there is an
overall decreasing trend in length at age in Divs. 2J and 3K illustrated by the downward slope of
the iso-age lines with time. A similar trend is not apparent in the Div. 3L data.
•
Cohort weight at age
The cohort weight at age plots (ages ~ 8) for the 3 divisions (Fig. 9) show similar pattern
of ridges and valleys to those described above for the cohort length at age plots. The overall
downward trend in weight at age in Divs. 2J and 3K is also quite marked, illustrated by the
decrease in area of the rectangles bounded by age and cohort lines in the right portion of the plots.
There is some indication of an increase in weight at age for the 1989+ cohorts in the older ages and
most recent years.
Year effects in length and weight increments
In an attempt to see the year effects in length and weight increments more clearly, the
general approach taken in the 1992 Multispecies Assessment \Verking Group (Anon 1992) was
adopted. A multiplicative model accounting for age and year effects were fitted to the logarithm of
the annual increments for each division separately,
Ln(X 1··)
= u·I + ß·J + E ,
Y
8
where Xij is the length or weight in a particular division at age i in year j, <Xi is the age effeet,
ßj is
the year effeet and e is normally distributed error. The vector ßj for both length and weight is
plotted for each division in Fig. 10. The year effects for both length and weight inerements in
Divs. 2J and 3K have a similar pattcrn and are compatible with the downward trend and neardecadel seale variability described in the cohort plots. As explained above, the effeet ascribed to
year t oeeurs mainly in year t+ 1, requiring that these year effeet estimates be lagged by one year in
the subsequent data exploration exercises described below.
Condition
Annual variability in eondition basd on round weight was similar among eod lengthgroups, but differed among divisions (Fig. 11). In Div. 2J, average condition based on round
weight inereased slowly but irrcgularly in the 1980s before deelining abruptly in the early 1990s.
There is some indication of an inerease in the most reeent two years. In Div. 3K, condition
fluetuated without trend until the mid-1980s, when it inereased for about 4 years before dropping
baek to the earlier level. In Div. 3L, there was a small inerease in the 1990s.
•
Condition based on gutted weight (Fig. 12) was less variable than condition based on
round weight. In Div. 2J, there was no trend until a rapid decline in the early 1990s, followed by
some improvement in reeent years. In Div. 3K, condition based on gutted weight did not rise as
prominently in the mid-1980s as did eondition based on round weight. In Div. 3L, eondition
based on gutted weight was variable but without trend. Similar patterns in annual variability in
eondition based on gutted wcight are apparent when the data are presented by age-group (Bishop
and Baird 1994; Taggart et al. 1994; Bishop et al. 1995).
Data exploration
A data exploration exercise was carried out to eompare the patterns described above for cod
growth and the "environmental" variables. The primary tool adopted for doing this was seatter plot
analysis and the construction of 95% bivariate normal density ellipses. Correlation in the variables
is scen by the eollapsing of the ellipses along the diagonal. If the ellipses is fairly round and not
oricntated along the diagonal then the variables are not eorrelated. In this exercise correlations are
not taken to indieate eause and effeet, but to suggest possible relationships that could be examined
further.
Correlations with eapelin
The PA for capelin in eod stomaehs is plotted against cod Iength inerements, weight
inerements and eondition faetor in Fig. 13. Correlations with length and weight inerements are
generally weak. The Div. 2J PFI for capelin show little correlation with cod length and weight
inerements but is positively correlated with condition factor in Divs. 2J and 3K and negatively
correlated with condition in Div. 3L. Div. 3K capelin PA is negatively corrclated with Iength and
weight increments in all divisions but in some cases the correlation is very weak. There is no
eorrclation with condition factor. Capelin PFI for Div. 3L is weakly negatively correlated with
Icngth and weight increments. However there is a relatively tight cIlipses indieating negative
•
9
correlation in the scatter plot with Div. 21 and 3K condition factors and a positive correlation with
Div. 3L condition.
•
The generaIly negative correlations, albeit weak, between capelin PA and cod length and
weight increments are unexpected. Increased consumption of capelin, and hence generaIly
increased total consumption, might rather be expected to translate into increased growth increment.
The positive correlation between capelin PFI in Div. 21 and cod condition in Divs. 21 and 3K is
reassuring. The differing correlations between capelin PA and condition between Divs. 21/3K and
Div. 3L might suggest that good feeding conditions in the north occur at the expense of fish in the
south, and vice versa. It is of interest that the intermediate area, Div. 3L shows virtuaIly no
correlation with any of the growth or condition data.
Correlations with cod biomass
The RV survey biomass estimates for the three divisions and the ADAPT estimate of 3+
biomass are plotted against the growth and condition data in Fig. 14. \Vith few exceptions, the
correlations are weakly positive. Correlations with Div. 3L condition are negative. The tightest
positive correlations are between weight increment estimates in Div. 21 and RV estimates in Div. 21
as weIl as ADAPT estimates. Correlations with population size are generaIly anticipated to be
negative because of density dependent effects, however in this study period the correlations are
consistent with a growth contribution to the observed biomass changes.
Correlations with Station 27 temperature data
•
Average annual water column temperature and annual bottom temperature anomaly are
plotted against the growth and condition data in Fig. 15. Correlations vary in strength but are
generaIly positive with growth and condition data for Divs. 21 and 3K but negative with the Div.
3L data. Correlations are weakest with respect to Divs. 21 and 3K length increments. \Vith the
exception of Div. 3L, these correlations are consistent with warm water conditions causing
enhanced growth.
Correlations with area of the CIL
Annual estimates of the area of the CIL for the three seetion as weIl as the average area for
all three sections combined are plotted against growth and condition data in Fig. 16. Correlations
are generally negative, with the exception of Div. 3L , and in some cases are quite strong. For
example, weight increments in Divs. 21 and 3K are considerably smaIler when the area of the CIL
is large. Div. 21 length increment data are also negatively correlated with area of the CIL but the
relationship with Div. 3K length increment data is not dear.
Further exploration oLthe correlation with area of the CIL
To examine the correlation in growth increments with the average area of the CIL from the
three seetions more closely, the totallength and weight increment in each cohort from the beginning
of age 2 to the end of age 5 was plotted against the average area of the CIL experienced by the
cohort over that period (Fig. 17). For the CIL data this is equivalent to a 3 year running mean and
brings out the near-decadel scale cycle very clearly. Evaluating the growth increment over the three
10
years of the cohort's life when most of the growth is occurring and for which most of the data arc
available brings out the near-decadel scale of variability in growth more clearly than can be seen by
plotting year effects in the increments. This is probably at least in part because the ycar to year
errors in the sampIe estimates canceI. \Varren (1993) applied the KaIman Filter technique to the
cod growth data for Divs. 21, 3K and 3L for a longer time period. This technique appears to be
purticularly useful in picking out both the downward trend and the near-decadel eyele in growth.
For the time periods that overlap the pattern is very similar in the study by \Varren (1993) and in
the present study.
In the case of both length and weight increments, but most cIearly with respect to weight,
the increments decline when the arca of the CIL increases and increase when the area of thc CIL
declines. The overall declining trend in growth, particularly strong with respect to weight, is also
apparent, corresponding to the increasing trend in CIL, most easily seen in the plots in Fig.3.
Diseussion and Conclusions
•
Cod growth has varied considerably over the time period considered in this study. There
appears to be at least two components to this variabiIity: (i) a downward trend over the entirc
period and (ii) an approximately decadel eycle in growth. The combined effect, as illustrated in
Fig. 17 is that the weight increment by cohorts between the ages of 2 and 5 reached a minimum in
Divs. 21 and 3K in the 1986 and 1987 cohorts. Tbe decline in weight increment is quitc
substantial and could perhaps be referred to as a "collapse in growth". The 1976 and 1977 cohorts
benefited from being at both the high end of the trend and the high point in thc approximately
decadel eycIe, and increased by nearly 1.6 kgs over ages 2 to 6. In comparison, the 1986 and
1987 cohorts were at the low end of the trend and the low point in the cycle and increased by only
about 400 g in Div. 21 and 700 g in Div. 3K. Thus thc weight increment over these ages decreased
by more than 1 kg for cod in Div. 21. The timeseries for Div. 3L is less complete but shows much
less evidence of a similar trend or eycle.
Isolating the causes for the trend and eycle in cod growth in Divs. 21 and 3K would have"
important implications for management of the northem cod stock from both thc TAC-fishing
mortality aspeet and from a spawner stock per recruit point of view. Several variables cxamined in
this study show weak to strong corrc1ation with these changes, but causation is still a matter of
speculation. Relationships with temperature, particularly the area of the CIL, may have the most
promise with respect to cxplaining thc near-decadel eycle in cod growth. The downward trend in
cod growth is correlated with the overall downward trend in ocean temperaturc over the period of
thc study. Howcver, other factors, such as thc generally incrcasing fishing mortality over the
study period mayaIso be important. Fishing mortality could progressively wced out the fast
growing genotypes from thc population leaving only genotypes with a low growth capacity.
The effect of capelin on cod growth, whilc anticipated to occur, is not easy to isolate in thc
current data. It is thought that the 1983 and 1986 year cIasses werc strong (Carscadden 1994,
leading to biomass peaks about two years Iater in the hydroacoustic surveys (Fig. 4). Cod growth
was in fact low following the 1983 year class, however two years after the 1986 eapelin yearclass
eod growth rate was high and/or inereasing. Capelin in eod stomachs in Div. 21 inereased in the
late 1980s when cod growth rate and condition werc high and then decreased sharply to very low
levels after 1990, eoineident with the abrupt decline in the weight inerement und the eondition
faetor, although the overall eorrelation of eapelin PFI with weight is weak. The general increasc in
capelin in eod stomachs in Divs. 3K and 3L over the time period is not compatible with changes in
•
ll
growth and condition in Div. 3K, but is consistent with the trend of improving condition in Div.
3L. At this stage the capelin effect on cod growth and condition should be interpreted as
equivocal. It seems almost certain that resolution of the temporal and spatial patterns of abundance
of cod and capelin are an essential prerequisite to the interpretation of capelin effects on cod (see
Lilly 1994 for more insight into this).
.
•
•
Density dependent effects on cod growth rate have been suggested in the past, particularly
when the pattern in the residuals are examined after removing the predominant temperature effect
(e.g. MilIar and Myers 1990). In this data visualization and exploration exercise there is no
evidence of density dependent growth in northern cod in Divs 21 and 3K, although there are
correlations compatible with density dependent growth in Div. 3L. In Divs. 21 and 3K cod growth
and condition are in fact generally positively correlated with biomass estimates supporting the
interpretation of an accelerated growth component to the increase in biomass following extension
of jurisdiction in 1977.
Difference in the sign of the correlations between Divs. 2113K and Div. 3L with respect to
the variables considered in this study suggest that fish on the Labrador and north east
Newfoundland shelves are responding differently to those on the northern Grand Bank. This is
particularly evident with respect to the increasing condition factor in Div. 3L in the late 1980s and
early 1990s at a time when condition factor was dropping precipitously in Divs. 21 and 3K.
Reasons for the spatial differences over the range of northern cod stock need to be examined in
more detail.
The cod stock in NAFO Divs. 2J3KL are presently under a fishery moratorium following
the very large decline in biomass in the late 1980s and early 1990s. This decline coincided with
rapidly increasing fishing mortality which must be considered to be the major cause. However
what amounts to a collapse in growth in cohorts entering the fishery at this time amplified the
fishing mortality effect considerably and reduced the biomass and yield per recruit. While it is
important to resolve the various contributors to the decline in the northern cod stock, the isolation
of predictive relationships between cod growth and, for example, temperature and capelin
consumption, could have considerable value in adjusting future fisheries management for changes
in cod growth rate.
Acknowledgements
\Ve acknowledge the considerable efforts ofDFO personnel who collected the sampIes on
which this analysis is based. Our recently retired colleague Claude Bishop and his predecessors in
the Gadoids Section in particular have done much to ensure that a comprehensive biological data
base is available for northern cod. Our colleague, Eugene Colbourne, is thanked for providing
Station 27 and CIL data from the oceanographic data base and for advice on interpretation.
References
Akcnhcad, S. A., J. Carscadden, H. Lcar, G. R. Lilly, and R. Wclls. 1982. Cod-capcIin interactions off northeast
Newfoundland and Labrador, p. 141-148. In M. C. Mercer (cd.) Multispecies approaches to fisherics
management advicc. Can. Spec. Publ. Fish. Aquat. Sei. 59.
Anon. 1991. Report ofthe Multispecics Assessment Working Group, Woods Hole, 4-13 Decembcr, 1990. ICES
C.M. 19911ASSESS:7, 246p.
12
Anon. 1992. Report of the Multispeeies Assessment Working Group, Copenhagen, 16-25 June 1992. ICES C.M.
19921ASSESS:16, 152p.
Bishop, C.A., E.F. Murphy, M.B. Davis, J.W. Baird and G.A Rose. 1993. An assessment of the eod stock in
NAFO Divisions 2j+3KL. NAFO SCR Doe., No. 86, Serial No. N2271 , 51p.
Bishop, C. A 1994. Revisions and additions to stratifieation sehemes used during research vessel surveys in
NAFO Subareas 2 and 3. NAFO SCR Doe. 94143, Serial No. N2413. 23 p.
Bishop, C. A., J. Anderson, E. Dalley, M. B. Davis, E. F. Murphy, G. A. Rose, D. E. Stansbury, C. Taggart, and
G. Winters. 1994. An assessment of the eod stock in NAFO Divisions 2J+3KL. NAFO SCR Doe.
94/40, Serial No. N241O. 50 p.
Bishop, C. A, and J. W. Baird. 1994. Spatial and temporal variability in eondition faetors of Divisions 2J and
3KL eod (Gadus morlzua). NAFO Sei. Coun. Studies 21: 105-1I3.
•
Bishop, C. A, D. E. Stansbury, and E. F. Murphy. 1995. An update of the stock status of Div. 2J3KL eod. DFO
Atlantie Fisheries Res. Doe. 95134. 38 p.
Brander, K. M. 1995. The effeet oftemperature on growth of Atlantic eod (Gadus morlzua L.). ICES J. mar. Sei.
52: 1-10.
Carseadden, J. 1994. Editor. Capelin in SA2 + Div. 3KL. DFO Atlantie Fisheries Research Doeument 94118, 164p.
Colbourne, E., S. Narayanan, and S. Prinsenberg. 1994. Climatie ehanges and evironmental eonditions in the .
Northwest Atlantic, 1970-1993. ICES mar. Sei. Symp., 198:311-322.
Colbourne, E. 1995.' Oeeanographie eonditions and climate change in the Newfoundland region during 1994. DFO
Atl. Res. Doe. 9513, 36p.
Fahrig, L., G. R. Lilly, and D. S. Miller. 1993. Predator stomaehs as sampling tools for prey distribution: Atlantic
eod (Gadus morlzua) and eapelin (Mallotus villosus). Can. J. Fish. Aquat. Sei. 50: 1541-1547.
Fleming, AM. 1960. Age, growth and sexual maturity of eod (Gadus morhua L.) in the Newfoundland area, 19471950. J. Fish. Res. Board Can. 17: 775-809.
J6nsson, J. 1965. Temperature and growth of eod in Icelandie waters. ICNAF Special Publieation 6: 537-539.
Jprgensen, T. 1992. Long-term ehanges in growth of North-cast Aretic eod (Gadus morlzua) and some
environmental influenees. ICES J. mar. Sei. 49: 263-277.
Lilly, G. R. 1987. Interactions betwcen Atlantie eod (Gadus morhua) and eapelin (Mallotus villosus) off Labrador
and castern Newfoundland: a review. Can. Teeh. Rep. Fish. Aquat. Sei. 1567: 37 p.
Lilly, G. R. 1991. Interannual variability in predation by eod (Gadus morhua) on eapelin (Mal/olus villosus) and
other prey off southern Labrador and northeastern Newfoundland. ICES mar. Sei. Symp. 193: 133-146.
Lilly, G.R. 1994. Predation by Atlantie eod on eapelin on the southern Labrador and Northeast Newfoundland
shelves during aperiod of changing spatial distributions. ICES mar Sei. Symp. 198:600-611.
•
13
LiIly, G.R. 1995. Did the feeding level of the cod off southern Labrador and eastern Newfoundland decline in the
1990s? DFO Atlantic Fisheries Research Document 95174, 25p.
LiIly, G. R., and D. J. Davis. 1993. Changes in the distribution of capelin in Divisions 21, 3K and 3L in the
autumns of recent years, as inferred from bottom-trawl by-catches and cod stornach examinations. NAFO
SCR Doc. 93/54, Serial No. N2237. 14 p.
May, A. W., A. T. Pinhorn, R. WeHs, and A. M. Fleming. 1965. Cod growth and temperature in the
Newfoundland area. ICNAF Special Publication 6: 545-555.
•
Mehl, S., and K. Sunnanä. 1991. Changes in growth ofNortheast Aretie cod in relation to food eonsumption in
1984-1988. ICES mar. Sei. Symp. 193: 109-112.
MilIar, R. B., L. Fahrig, and P. A. Shelton. 1990. Effeet of eapelin biomass on eod growth. ICES C.M.
1990/0:25. 10 p.
MilIar, R. B., and R. A. Myers. 1990. ModeHing environmentaHy induced change in size at age for Atiantic
Canada eod stocks. ICES C.M. 1990/0:24. 13 p.
MilIer, D.S. 1994. Results from an acoustic survey for eapelin (Mallotus villosus) in NAFO Divisions 2J3KL in
the autumn of 1993, p.91-98. In I. Carseadden [ed.] Capelin in SA2 + Div. 3KL. DFO Atiantic Fisheries
Research Document 94/18.
Sheiton, P. A., L. Fahrig, and R. B. MilIar. 1991. Uneertainity associated with cod-capelin interactions: how much
is too much? NAFO Sei. Coun. Studies. 16: 13-19.
Smith, S. I., and G. D. Somerton. 1981. STRAP: A user-oriented computer analysis system for groundfish
research trawl survey data. Can. Tech. Rep. Fish. Aquat. Sci. 1030: 66 p.
•
Steinarsson, B. iE., and O. Stefansson. 1991. An attempt to explain cod growth variability. ICES C.M.
1991/0:42.20 p.
Taggart, C. T., I. Anderson, C. Bishop, E. Colbourne, I. Hutchings, G. LiIly, J. Morgan, E. Murphy, R. Myers, G.
Rose and P. Sheiton. 1994. Overview cf cod stocks, biology, and environment in the northwest Atiantic
region ofNewfoundland, with emphasis on northern eod. ICES mar. Sci. Symp. 198: 140-157.
Warren, W.O. 1993. Some applieations of the Kaiman Filter in Fisheries Research. ICES C.M. 19931D:57, 19p.
WeHs, R. 1984. Orowth of cod in Divisions 21, 3K and 3L, 1971-1983. NAFO SCR Doc. 84/90, Serial No.
N881. 6 p.
WeHs, R. 1986. Declines in the average length at age of eod in Divisions 21 and 3K during 1977-85. NAFO SCR
Doe. 86/83, Serial No. N1205. 2 p.
14
Table 1. Number of fish sampled from RV surveys by NAFO division for which both length and
age have been determined.
YEAR
NAFO
FREQUENCYI2j
13k
131
---------+--------+--------+--------+
78 1
453 I
457 1
0 1
---------+--------+--------+--------+
79 I
460 1
510 I
0 1
---------+--------+--------+--------+
80
600 1
706 I
0 1
---------+--------+--------+--------+
81 1
650 1
687 I
614 1
---------+--------+--------+--------+
82 1
844 1
747 I
882 I
---------+--------+--------+--------+
83 I
725 I800 I
797 I
---------+--------+--------+--------+
84 I
1055 1
844 I
0 I
---------+--------+--------+--------+
85 1
987 1
659 I
801 I
---------+--------+--------+--------+
86 1
569 1
670 I
699 I
---------+--------+--------+--------+
87 1
819 1
681 I
743 I
---------+--------+--------+--------+
88 1
852 1
1000 I
742 I
---------+--------+--------+--------+
89 1
891 1
1055 I
713 I
---------+--------+--------+--------+
90 1
853 1
971 I
706 1
---------+--------+--------+--------+
91 1
546 1
764 I
578 1
---------+--------+--------+--------+
92 I
263 I
540 1
494 I
---------+--------+--------+--------+
93 I
95 I
355 I
378 I
---------+--------+--------+--------+
94 I
62 I
92 1
126 I
---------+--------+--------+--------+
TOTAL
10724
11538
8273
I_
TOTAL
910
970
1306
1951
•
2473
2322
1899
2447
1938
2243
2594
2659
2530
1888
1297
828
280
30535
•
15
Table 2. Number of fish sampled from RV surveys by NAFO division for which both weight and
age have been determined.
YEAR
NAFO
FREQUENCY\2j
•
•
13k
131
I
---------+--------+--------+--------+
78 I
132 I
120 I
0 I
---------+--------+--------+--------+
79 I
113 I
119 I
0 I
---------+--------+--------+--------+
80 I
140 I
156 I
0 1
---------+--------+--------+--------+
81 I
145 I
141 I
138 I
---------+--------+--------+--------+
82 I
135 I
160 I
51 I
---------+--------+--------+--------+
83 1
173 I
156 I
152 I
---------+--------+--------+--------+
84 I
532 I
167 I
0 I
---------+--------+--------+--------+
85 I
506 I
143 I
147 I
---------+--------+--------+--------+
86 I
119 I
130 I
142 I
---------+--------+--------+--------+
87 I
104 I
13~ I
161 I
---------+--------+--------+--------+
88 I
200 I
249 I
156 I
---------+--------+--------+--------+
89 I
890 I
1055 I
142 I
---------+--------+--------+--------+
90' I
852 I
970 I
706 I
---------+--------+--------+--------+
91 I
546 I
764 I
576 I
---------+--------+--------+--------+
92 I
263 I
538 I
494 I
---------+--------+--------+--------+
93 I
95 I
355 I
377 I
---------+--------+--------+--------+
94 I
62 I
92 I
126 I
---------+--------+--------+--------+
TOTAL
5007
5447
3368
TOTAL
252
232
296
424
346
481
699
796
391
397
605
2087
2528
1886
1295
827
280
13822
16
Table 3. Range and average annual number of fish' sampled for Iength and weight at age from RV
surveys by age, alI three divisions combined.
2J
3K
3L
Lenoth al aoe samoies
Mean N iMin N
Years
Aoe
Division
1
2
1 3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
I
0
I 1
2
I
1 3
4
5
6
7
8
9
10
11
12
13
I 14
I
15
16
17
18
19
20
21
22
I
23
I
0
I
1
I
2
3
4
5
6
7
8
9
10
11
12
13
1 14
15
I
16
I
17
I
18
I
20
I
21
I 22
I
24
12
17
171
17
17
17
17
141
14
14
14
14
'14
10
10!
9
6
4
1
1
1
1
2
161
17
17
17
17
17
17
17
15
141
141
14
13
13'
131
101
5
• 5
2
3
3
1
11
51
111
131
131
131
131
131
131
131
111
121
11
11
10
9
91
91
61
31
21
11
11
11
14.24
68.12
88.47
83.88
84.82
70.881
61.35
52.651
41.29
29.41
15.12
9.82
4.59
3.12
1.181
0.71
0.59
0.35
0.06
0.06
0.06
0.061
0.291
23.941
87.82
98.24
84.531
80.18
70.651
58.59
55.71
42.94
30.591
18.881
11.181
5.291
4.531
2.291
1.531
0.591
0.29
0.12
0.18
0.24
0.12
0.061
2.711
12.141
71.431
92.141
81.211
67.50
61.71
58.29
47.43
37.57
19.57
14.29
10.21
6.361
3.861
2.291
1.001
0.501
0.211
0.291
0.211
0.071
0.071
Weiohl al aoa samoies
IMeanN MinN
Years
'MaxN
0
7
15!
13
81
2\
11
01
0
0
0
0
0
01
01
01
01
01
01
01
0
0
0
0
101
381
21
11
2
4
1
0
01
01
01
01
0
0
0
0
0
01
01
01
01
01
01
01
01
01
0
0
0
0
01
01
01
01
01
01
01
01
01
01
01
01
01
01
01
73
193
202
190
265
195
133
125
113
74
36
23
11
18
5
3
3
3
1
1
1
1
4
72
211
175
151
164
130
121
121
100
77
51
30
14
14
13
5
3
1
1
1
2
2
1
21
46
203
154
148
120
99
147
104
77
53
44
28
21
13
6
3
2
1
3
3
1
1
12
17
17
17
17
17
17
141
14
14
14
14
14
10
10
9
6
4
1
1
1
1
2
16
17
17
17
17
17
17
17
15
14
14
14
13
13
13
10
5
5
2
3
3
1
1
5
11
13
13
13
131
13
13
13
11
12
11
11
10
9
9
9
6
3
2
1
1
1
3.47
33.41
44.47
45.94
42.06
29.47
25.12
22.76
16.06
12.18
7.29
5.24
2.88
1.94
1.00
0.47
0.47
0.18
0.06
0.00
0.06
0.00
0.24
10.94
45.18
48.41
43.941
37.06
31.41
24.35
22.88
17.88
14.06
8.29
5.47
3.18
3.00!
1.47
1.24
0.531
0.24
0.12
0.18
0.24
0.12
0.061
0.071
2.361
24.711
38.431
38.57
29.00
27.43
17.71
14.71
15.29
8.50
7.36
5.57
4.14
2.57
2.14
0.93
0.43
0.21
0.291
0.211
0.071
0.001
I
I
:
I
I
!
Years = number of years with sampies. N = number of samoies Der vear.
I
I
I
I
I
IMaxN
0
7
7
9
4
2
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
5
8
6
1
2
4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
01
I
I
I
1
I
I
I
I
18
192
202
187
176
121
113
125
72
37
26
11
8
6
4
2
2
1
1
0
1
0
4
72
211
168
151
163
93
109
121
79
50
30
13
9
8
9
4
3
1
1
1
2
2
1
1
13
79
117
116
91
98
64
75
66
31
19
16
12
10
6
3
2
1
3
3
1
0
•
•
60'
SB'
SB'
52'
50'
54'
o
Il)
52'
::: .:
: rtEMIS
,:;: ~ ;,_:;::?:;
50'
.,.....;, oall'-'
'
' .._.01''3
..
1------,r-.......--~.7'; -'-t!':
,:,,"
CAP
:
:
.
••.••,J._. ••••••_•••••__.
3L
Co.
,..
,.,
'"
4B'
,~
........
,
f
"'1..1
~"\
~
""Q
~
,
I
4B'
60'
SB'
54'
52'
50'
48'
48'
~-1-------1---_----:L...-+60
55
50
-L.._+-__-"'_-.
45
40
LONGITUDE
'-.1
Fig. 1. Maps showing NAFO division boundaries, the loeation of Station 27 and the position of
the three seetions. The Seal Island section erosses the Hamilton Bank and is referred to in the text
as the Hamilton Bank seetion.
1.5
(J
18
~
1
CI
Q)
"C
Q)
....
1.......-+
~
+
....ca
o0
Q)
Q.
E
Q)
I-
- D - AVETEMP
0.5
-0.5
-1
+t{
I
I
0
I
I
co
I
I
I
I
I
I
co ! ' co 0>
0
co co co co co co co co co co
..-
0>
l'--
l'--
- + - - BOTANOM
C\J
C')
'<t
LO
I
I
I
0
..-
C\I
C')
'<t
0>
0>
0>
0>
0)
I
Year
Fig. 2. Time series of Station 27 average annual water column temperature and the bottom (175
m) temperature anomaly from the 1961 - 1990 mean.
50 45
40
E
::.:: 35 I
Q)
....
30 I
1
1
1
~
+
-... 20V
C'"
l/)
ca
Q)
<
25 I
--0-
+
+--r
15
10
5
0
+
+/'
\C16
1-/+ .
ca
BONCIL
- + - FLEMCIL
- - D - HAMCIL
I
I
co
l'--
0>
!'
0
co
..-
co
I
I
C\I
C')
co
co
'<t
co
I
I
I
LO
co
co
!'
co
co
I
I
I
I
I
co 0>
co co
I
0
..-
C\I
C')
0>
0>
0>
0>
'<t
0>
Year
Fig. 3. Annual area of the cold intermediate layer (CIL) on the Hamilton Bank, Bonavista and
FIemish Cap sections.
•
4.-------------------__--,
c:J Capelin
_
2J
Otherprey
3
2
2000 -r------~-----------__,
-...
g
o
e-e Canada
78
0··0 Russia
84
86
88
90
92
E 1000
>< 3
Q)
"0
C
\
o
1).
94
3K
<\.. .
cn
cn
(lj
82
•
1500
'-'
80
~ 2
Q)
.~
Q).
0..:
(lj
.f:
:i
C!......
500
u..
\
()
o
o +--r---r-..-~O'~..~9_r__r__,__r_,_.,.__r_,.__...,:_'tq=:+_;
~
ro
n
00
~
M
M
M
00
~
78
80
82
84
86
88
90
92
94
4,........!.~---:~-.:::::..-~-.....:::.:::.....-......::;=---=-=----==--=-:.....,
~
Year
3L
-
3rr-, ..
r-
2r-
Fig. 4. Canadian and Russian fall capelin hydroacoustic estimates.
r-
-
-
_r-
-
1 r1
o 1.-.1-...L..-....I1_
78
80
LL
82
84
86
Year'
88
90
92
II~
94
\0
,
Fig. 5. Partial fullness indices (PFI) for capelin in cod stomaehs in the fall RV survey, together
with the PFI for other prey, by division.
.
20
450000 -
UI
400000 I
350000
-
300000 -;-
- - D - - RV2J
250000 ~
200000 ~
- - 0 - RV3K
C'll
s::::
0
UI
UI
E
0
m
i
- + - - RV3L
150000 :),
•
i
100000 ,
I
50000
""'+""'=-+
0
CX)
t--
0)
0
T"""
C\I
"'"
CX)
CX)
(Xl
('I)
(Xl
~
(Xl
1.0
(Xl
(0
(Xl
"'"
(Xl
(Xl
(Xl
0>
0
T"""
C\I
('I)
(Xl
0>
0)
0>
0>
~
0)
Year
Fig. 6. Annual biomass estimates by division from the fall RV trawl survey.
1200000 -
UI
-
•
1000000
s::::
0
UI
UI
C'll
E
800000 600000
0
m
I
I
400000
1
200000
-
+
M
0
I
('I)
"'"
-.:t
t--
1.0
"'"
<0
"'"
t-t--
I
I
(Xl
0)
0
T"""
N
"'"
(Xl
(Xl
(Xl
"'"
I
('I)
CX)
-.:t
(Xl
I
I
1.0
(0
ro
(Xl
t--
ro
CX)
CX)
0>
ro
0
0>
T"""
C\I
0>
0)
Year
Fig. 7. ADAPT estimates of the 3+ beginning of year biomass in NAFO Divs. 2J3KL combined.
Div 2J cohort length at age
Div 3K cohort ·length at age
90
Div 3L cohort length at age
90
90
80
70
s
.........
60
S
~ 50
~ 50
30
20
10
78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94
78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94
Year
Year
Fig. 8. Cohort lengths. at age for each division for ag~s 10 and less. Lines orientated roughly
parallel to the the x-aXlS denote ages. Length at age 0 IS taken to be 0 so that the Iines orientated
roughly parallel to the y-axis denoting cohorts intercept the x-axis to give the year in which the
cohort arose.
Year
Div
2J cohort weight at age
Div 3K cohort weight at age
Div 3L cohort weight at age
4000
3000
.....
~
--'
'-""
'-""
+.J
~ 2000
~ 2000
~
~
~
....
~
50
~
0.0
0.0
2000
+.J
.....
.....
~
1)
;;.
;;.
1000 I
o~~~z~z~z~~~
78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94
Year
nnwM~~M~M~MMOO~~~W
Year
78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 9·
Year
N
N
Fig. 9. Cohort weights at age for each division for ages 8 and less.
23
0.4
0.2
0
u
Q) -0.2
Q) -0.4
Q)
-0.6
.~
ca -0.8
Q)
-1
a:
-1.2
-1.4
-1.6
--
- D - - 2J Une
--0--
-
3K Une
- - + - - 3L Une
co
0)
0
T"""
C\I
f'-
f'-
CX)
CX)
CX)
C')
ro
~
CX)
10
ro
CD
f'-
CX)
CX)
co
ro
0
0)
co
0)
Year
0.8
0.6,
f
-- O.~ It
0.4
0
- - - 0 - 2J Wine
Q)
•
Q)
-0.2 T
I
-0.4 +
ca
Q) -0.6 t
a: -0.8 -r
--0--
Q)
-
3K Wine
.~
- - + - - 3L Wine
I
-1
-1.2 J
-1.4 .
co
0)
0
T"""
f'-
f'-
CX)
CX)
I
I
I
I
I
C\I
C')
CX)
oo::t
10
CD
CX)
CX)
CX)
ro
f'-
ro
I
I
I
I
CX)
CX)
O'l
0
T"""
C\I
C')
~
CX)
O'l
O'l
O'l
O'l
O'l
Year
Fig. 10. Relative year effects in cohort length and weight increments in each division, aIl ages
combined.
1.1 . . . . . - - - - - - - - - - - - - - - - - ,
0.85
2J
1.0
--r-----------2J
--,
0.80
0.9
0.75
36-44cm
- 6 - - 45-53 cm
54-62cm
- ....- 63-71 cm
-0-
0.8
\
\
..
78
80
82
84
86
88
90
92
94
;:1.1--r------------------,
-g>
Ci)
.c
o. 0.65
o
78
80
82
84
86
_x 0.85 - - r - - - - - - - - -
Q)
-g>
Q)
Ci)
-
0.9
'Q)
~
0.8
90
92
94
----,
0.80
...J
-
-.. 0.75
.c
Cl
88
3K
.c
1.0
4---r---"T""---r--r-,r--r-T-.,.-J...,.-...,...-.--r--'r---r-1~r--r-i
~
--J
_
-6--
\
o O.7 +-~,---,-----r----r~=r=::;:::::r==;:..-.-_r_.--l_r--r-l
~
36-44cm
45-53 cm
54·62cm
- ....- 63·71 cm
-0-
0.70
.c
Cl
'Q5 0.70
0.7 -1--r-,--r-""1c----r----.--.----r---.....---.--.--.--.--,~__"T--I
:J
78
80
82
84
86
90
92
94
88
o
"C
C
~ 1.1 - - r - - - - - - - - - - - - - - - - - ,
3L
1.0
3:
-g::::
0.65 -+--,--.--.--.-----r-r---.-.---,-,,---r-r-r--r-'-"T-".---t
:::s
78
80
82
84
86
88
90
92
94
~ 0.85 - , - - - - - - - - - - - - - - - - - - - ,
3L
0.80
.6
0.9
-------~-----...
lt
0.75
•
0.8
-------~----/~~
0.70
78
80
82
84
86
88
90
92
94
78
Year
80
82
84
86
88
90
92
94
Year
?
Fig. 11. Average condition (round weight) of cod by division in .em
. Fig. 12. Average eondition (gutted weight, head on) of eod by
length groups from sampies taken in the fall RV surveys.. A eondltlon faetor
jivision in 9 cm length groups from sampIes taken in the fall RV
of 0.9, which is sometimes assumed to be generally appheable to eod (eg.
mrveys.
A eondition faetor ofO.77, whieh is the overall average
Brander 1995), is shown for referenee.
reported by Taggart et al. (1994), is shown for referenee.
25
(Scatter Plot Matrix
)
I. "
2.0 Ip2JC I
1.0
,:,,~
·
.,
.. ··.
0.0
3.0
2.0
1.0
0.0
"-..
/": \
'';
:\
\
..
"
,'.
-1.0
0.2
.-/
,:- I • :--....-.
-0.2 1--:'
-0.8
0.4
I~
J
-.
• .--
... I,'
F2
~
"~
.:.
0\
:.. •
I
• _',t'
I'- .
i"
.,
'" ·. .
'.
.......
'I
.'
--
(
..
3KLINC
i\..
'\
.
·,
1"'- '. .... \
V·" ., '\~
....
!~.'
V
\
. \
"·.... ~
. - .·· ..· ., . ·..
· ~.
K":' .'. K·
~.
"
'
. ·... .~ .
["-....
••
.
r-.-.",,'.
'
e
@
'
J
.
.. : . - ..
"""-
/
.
.
·.'.
·. . Ip3KC I '.. . ,. .
,.
','
·.' ":/
.../ ~
·
:
Ip3LC I
1.5
.
\ J.. .· \
·.
!I
\ . .
1.0
..
0.5
:'
·
.
..
1\
..
\
\ ...·
~'.
......... 2JLINC
.......... If:~
0.0 v,.
·.
·.
-0.5
...... .
·
'
'.'
.,
.:.>e
-· :-..
.' ~ ·
r--..:. ••
':' ~
'
&·
'
V:",
... :.', i · :"/
. I(~ I~OO:>
/.
· , . . .I .'\. V' •
· ;,
'\.
1/'
,
'.'.
,,
,{
"
.'
~ K~ ' . /
/
\
.':\
....
. ·..
I~.
V·, v··. v·
. : .'. . ..
. ..' . · .',
.'
I~>
:
~
,
.'
."\
~ ../
1(.)
"
2
,"
~/ f'>... '..-/ ~
V
•
k,
'. · .. ~
~ · .....
.
"
.
..
~ '. -.-.:.... ~.".- ;.~
·
.,11 ..'
.
I
3LLINC V
'.
'1/
l/.'
·
·
·
. . .. . I· ·.
,
!/.'· · . ..
:
.'
0.0
·
·
-0.4
· ./' ..)
/\
· .) '1/..':'/ .':/
/ I\·.~· ~ · \ " J \. ··
·2JWINC
.
.. \ '.· .
.. . ·I/. , . · 11..·. . .,' .\
0.0 ·.
.· -'.11;:·' .
'"
·. · .,
·.
-0.5
·
.
:~)
:
.·.
..
:.)
..
·1.0
.
..
....
)
·
.. /
·. · . . /1'\ . .
·
1"-.. ·· .
1.0
3KWINC
...... l;>---V
v ..0.5
~
.. '" . .1 ~
")
~
I~
..
k
'.
·
.
0.0 ·...
.
.'·
·
·
.
.,
.
.
.' . ./K····
r---.... •
y
-0.5
'./
i'-.
.I~'
I'-- ·
0.4 V:
...............
/(0' ~ :
·v . / ."\ 3LWINC V .--"'. " L/ .............v· ...'\
0.2
·'. ~ )
0.0
· .. ·.· : · .
-. 1\ . .:. .' . ·.
·· · ..
.
-0.2
..
·
:. J
·
·. '" . . 1\ ": J · . ·
-0.4 ·.
··
V
( '.: \ /· ..' \ /
O
1.05 '/.
. . · .. ·. / ...! \ · '.. · \· 2JCOND
'. ,/
·
.
..
.
.
.:
·.
.'
·
·
.
·
0.95
}.
- \·
~ . ~") \".,·./
I\.\. ,
./
0.85
' / .'
·
1.05
3KCOND
V
.
'
...
\
·.
li '.
I
.'
....~ (
.
· ... \
\
·.
1.00 V '..
·
..
·
') V"
0.95
~ ... J
.' . k"
0.90 :.j/
. .. 1\"
· . \. ::) . :.,:-; / ' ... / " ;\ ..· :.... ~.,: J
v. ~ 3LCOND
/ .. \
1.00
v· .~
'.~
I/'
.
"
"
Ir
0.95
.. .
~."> .·
·- ..
~:.:'. ~.. ~ ..
0.90 · .
"-...:
f',........ •1(· ." / "./
:
./
1\\
·
, ,
..
0.90 1.00
0.90 1.00
1.5
-1.0 0.0 -0.8 0.0 0.&0.4 0.0 0.4 -1.0 0.0 -0.5 0.5 -0.4 0.00.30.85 1.00
0.0 1.0 2.0 0.0 1.5 3.00.0
. ·
,
·. "\
'"
1\'.')' '"
.'
I
\
.' \
"
(:
,'
"
",
-'s
~
I--.
•
• I
'---",
..........
•
'
,;
16 ·u v.
'
•
_,I
'
,
I'-
1(:·':
~ .:.
v· \~
'
;/
... J
9
\l
'
....
",1
•
'
•
si
,
?;K:>
J
~
•
'
'..-/
y'
/
>./
.y
'.'
,
I
".
•
./
,\
Fig, 13. Scatter plot matrix of the partial fullness index of capelin in each division (P21C, P3KC,
P3LC) against the relative annuallength increments for all ages combined (21 LINC, 3K LINC,
3K LINC), relative annual weight increments for all ages combined (21 WINC, 3K WINC, 3L
WINC) and round weight condition factor for cod in the 54 - 62 cm length dass (21 COND, 3K
COND, 3L COND),
26
Scatt.r Plot Matrix )
M .<.-/-. r.,.:,. V; .:VS.i/7~·
. l(:.
"/,01/ ..;/". '.' '.'-'.
,y' ....~10::
/I~ . . -I .. ,
::1[§J <.~~.:
o
0
;:: 1/..
1= .:".)
250000
150000
50000
IRV3K
i/.. V
",.' /
I
.
.. . IRV3L I
•
:' /
V
.. •
.
0
.'
:/
0.4..:
.::: . . )'
(j
1/ .
"
j': . J...j
'. /
•
'. / .
"/. /
3KLINC
(..
:. ;..'
••••
H '<:J0--:~k
'0:5~/;:'./~"'
... /
0.4
0.2
.' ' \ .
~.~
.... )
•0:4
••••
..
v
r::
:
'
.
.)V. . v. .. v:. .
~::0YYU'"
•
1.05
1.00
0.95
0.90
•
•
\
'J'
_
: .':
•
.(t
..
3L UNC
=. . ... ; /
• : • '. / ' , , •
•
..)
.,?:,
~.
':.j
.
.h~ .
•
\"
I.:
( :-:.
[,;(, ----- ~ ._ I~
'1./ . . /. ... 11 •":)
. .) :''" / .'::J f .:.j ....: ;/. ..
~WINC f:: J '·1 f.:·..: 1/'· .' ... :.\
~
.y
i/.
-:
...;
•
.. '
- ::,./:3K·WINC
.../
'\'"\ 1/'/
. : . .... '..
~
(
1'\ .' ).
1(, ' ../
~IC:~:'~·'I\).
···/I~···~~·:
3LWINC
"\
1('.
'."17
'.'
,.'(.)"
I'.)
....
..
.
• 'I\. '
•,
..
V. '"V:- ..... ':,'1 . , .; ~COND I/;r.,.:. )
.1~··0vD·IQ~
"
.
v '. v
'
.'
o'
....
••: " : /
".
•
'/
.'.~.
' . , / ' '.. ' ·,,--/,·.I\.·
0'0.
1(;"
1/':' ,
3KCOND
')
1\'"
...."
r---..' /
,'\
R\~IQI~
~:~~ ~:.~~~
~I,,('"
\
• • • • • ~'l'
• •~..
•,v:-. .....
.
'.
•
••
090""'.
. I "'-.....
. ' ./ 1'-.. .: ~ '.
3OOOoom<xxi 2oo()(JC)JO(lOOI000000
lDIOOOO rollHllllO
I
•.
·1.0
'.\
"':J
"'Ir
v ..." .
" .,"/ . \ .
..
..
i
(1 ('7'"
J
1.05
Ir ~':J
'.' .
~··.:::010S>0~'-:~~.:0.
''''\ . .
."
"I;: : l( . I~'
1 '.
;'\
.. /" .:'
.~:~ 1(·:' JV:'~ '.' V. '.:.. I/:~ , ·/.. 1/·"
·1.0
I~\\
·::,V.·I/}/.::)V: ·1 /~.I~· ~
./. ./ .. /
./
/ I( . >J . \ :. ,
I..
.'
~ '.~ I'-~"
/ .
'..
;:~r0~P:"j;. !r
·0.8
/
........,
y;
1=1//~:.: 'V:'.:': ~3KLB
J. j
400000
. .\
1\)
.... / .
V .... 1/.... .~\ . ".\ V.' . . ~ ,.1: .'
:'J "~J . "·1' .\ ~':I'
.... .'.'
:..
~"/
o'
'•
0.0 -0 8
.' ..... f'.-.:.
0.0 0.&0.4 0.0 0.4 ·1.0
.:.
:.-/
0.0 ·0.5
."0• •
•'.
0.5 -0.4 0.0 0.3085 1.00
:
I":
'
•
3LCOND
".
.
0.90 1.00
0.90 1.00
Fig. 14. Scatter plot matrix ofbiomass estimates by division from the fall surveys (RV2J, RV3K,
RV3L) and the ADAPT estimate of 3+ biomass for all divisions combined (2J3KL B) against
growth and condition data (same abbreviations as in Fig. 13).
•
\
.
27
)
(Scatter Plot Matrix
1.5
I
!AVET
0.5
D
~
-0.5
0.3
0.1
-0.1
-0.3
-0.5
0.0
-0.5
-1.0
0.2
-0.2
-0.8
0.4
0.0
-0.4
0.0
-0.5
-1.0
1.0
0.5
0.0
-0.5
0.4
0.2
0.0
-0.2
-0.4
1.05
0.95
0.85
1.05
1.00
0.95
0.90
1.00
0.95
0.90
.
.
. . CJ)
· .',
IBOTAN
--:
-,'
.n
·-.
~
~
. . V·.'.. V· . ...
.. ':/ -./ .., :/ \):
~~
·
... :.. ,~
(!)
·. @
~L:--Y
-
V
I
..·, ,..---
'."",
"
~
'1'
· .V
11(8
· ·.
. ~'. · . V
.. -/ . 10
V. . "'~
..
..
.
V·"
·.
·
.
.:.;;;
·
() ~ r--..:./ I~
0' "B'... .: - ()
·
~ ·..
@ D (j)
69
~ -. -~ ~
.
. '.
/...
.
. V·' ·
.
.
I:':) \ .... .
·
. '.:/ :/ / ..)
·· '/
..
()
. . .
,
2JLrnC
~:'
~."'"
.....
....
"
..
'
V
"
:. ".
/
.:
.'
..-........
.'
'
. -,
;.
.. ' .
"
~:
\
'. ,
'. :
"
/
..
3KLINC
·..'\
·
\
"
/
\. • I
/,,/
.1/.)
'.
.... Ir) ..
'.'
•
..
:.'.
.·Z,'·
r-- :...-----
-;--- ~.
~'.:.
·
..
V·~ / '.......
I'-...
'.
I
~
· _.~ ~
'
'.
'.
'
/ .
~
'.'
':' ~
-~
'.
I'-
:
'V
,
l/.
I
1/
3LLINC
·, ...... V,
.....
V
· ....
.'
11:.
\
'
.
.,11
.'
.
·
. 1/·: . . .: .\
·. · .
·.
·.
.... ) . : .' .·· . /f\ .
'( :.
..
.. /
3Kwrnc
V - ......... ' V '.,
~
.
....,
..· - t::" ... . .' ~
.
k
'.
:'
·
."
.'. ~.'~
.. ...
.'
.'
· ~
1'- .
"-------·v .
.'-.
/. \ .
3Lwrnc V
V
". 1/' ...'\
.
I
.
.
1/.
·
·.
· ..
..
I\0 ::'.1.J . . I\: .... )
· · .·· : - ';
. . · . ··· · · ·..
J .. \
/
.\ 1/ V
I : \
. \ 2JCOND
·
,
.
..
'.
,
..
·
.
.
.;.: / . · . · · · .. J. ·
.. . : , .' . . ....
....
, / .'
., . J
.. /
;/ 1\ '.'
/
.. / ,.
/ ..
· ... \
11 .'
.. .'.
·... 3KCOND ".\
/~ ··V
\
·
·
.
.
I~.),
. .. .
.' .'
~'.'
· . '. '.,: /. · .
. ... .. ·.:: J '\ ··.1 : / ·... /: I\"~'
3LCOND
..
'\ ':~v.
.\
·
V'.
~
Ir
.'.
· . , . . .. . . .
·
.
..
-. ..'.
'
'
(J)'
','
"
I
'.'
.y
~~ k~
v.
v:
.
('.....
i'-
I'-
'
',-
\
~
'
~
() \
'
'
/
'
r.:.
0.0 -0.8
I'
y'
I
•••
{.
-0.5 0.5 UD.5 -0.1 0.3 -1.0
2JWrnC
I
u'10 .&
/.)
.
/
,
~:)1(:·':
:~"l
1/·'
·
'
. . . . 1'---....
0.00.&0.4 0.0 0.4 -1.0
·
:J
1\\
0.0 -0.5
./
j
/
t
•
.
: / ~:
0.5 -0.4 0.00.30.85 1.00
\
~
"'"
'
·
0.90 1.00
0.90 1.00
Fig. 15. Scatter plot matrix of average annuaI water column temperature (AVET) and bottom
temperature anomaly (BOTAN) from Station 27 measurements against growth and condition data
(same abbreviations as in Fig. 13).
•
J
28
Scatter Plot Matrix
45~ HAMCIL
35
25
~+----:::~~~~~--:;;,f-=----::;;i-~=:-i---:;;=:::-k=---+:-;---+-;;::.;:--b:=:::::---f;:::=::::--i:=::::::::--l-o::;:::::==:i
I
40
:n
2:l
0.0
-0.5
·1.0
0;::~t=-~~::::"-h--.--~:::::::~~+-::~i-:;:::'~~+;:~~:::::+;::::::==+==~---11
0.2
-0.2
-O.8+---r-~+...".....-=--+-r-"""""~-,-~:"'-H--:-"""",""+---r::-""-+r-~~=-I--:"'--:::;::>--t--:T.'\--t~""",,---:-1b--;-,,,;+-"?"::---+-::7.""-""<j
0.4
I
0.0
-0.4
0.0
-0.5
·1.0
h-~-';'4-~--':+..-~-4~~..:J1~""":"-;,1-.l,~+-r--~~~4-.L;""-r~-=----'l:-t"'--~""';"'R"':-oC.-t--=--"'.-l1
1.0J-...L.....:....j.~---..::.l.I-----':........:+-....:L-...:l.+-..:...L-+-L-:.L.-f-=--...L..--+---~~=!-=---'--+----:....4-.f.--.....;~-p---.,;..--i1
0.5
0.0
-0.5
0.4J-~-+----,~--.J---.,....-+------,,--+.~-:-I---:-~..J..-.....--;'-+----:---Ir-="=~f---,~-h-~...,...--:-"----I1
0.2
0.0
-0.2
-O.4:1-~....:.,.:~~....,.....:L+--.;;~~W~,..:L./>.--~~J--.-\,..:..,/-+;,..---~:..",..-~....1.:.;...t.r--1---.:::---e~~+-=~~+--..,;:--L.j1
1.05
0.95
0.85J--+-~,.+---'-:~-l..~~';":'-h.L:';"'4~":"'-4f---JL.:.:M+-7-~L.,f--=7...--:4-LT!"\+--;~....L+-:T-:-tr~~:;:,;",t-~",~1
1.05
1.00
0.95
0.90
1.00
0.95
0.90
W-:""':"'~:"'-"'-A-~~H--,-<~~4-.J,.....!H"'7'""-'-.L......;+--~4-++~:::""-H~-+-""","",-+::~~1
15 25 35 452:1 3J 40
ro
25 3J 35 2:1 3J 40
-1.0 0.0 ·0.8 0.0 0.5-0.4 0.0 0.4 ·1.0
0.0 -0.5 0.5 -0.4 0.0 0.3 0.85 1.00
0.90 1.00
0.90 1.00
Fig. 16. Scatter plot matrix of average annual area of the cold intermediate layer on the Hamilton
Bank (HAMCIL), Bonavista (BONCIL) and Flemish Cap (FLEMCIL) as weIl as the annual
average of all three (AVECIL) against growth and condition data (same abbreviations as in Fig.
13).
•
•
~
.
29
I
40
~
35
c -E
~
~~
30
G)
~]
20
<>
::
- - 0 - 2JSLlNC
I
- - 0 - 3KSLINC
I
G)
25
E E
.:.::
1
- - + - - 3LSLINC
1
•
RUNCIL
11---+---'-+--+1-+-1--+-1---'11--+--+1-+-1--+1-jl-+-I--+1-+-1--+-I-I(---j
~ ~ CD r--. CX) m 0
T""
C\I c") ~ LO CD I"- CX) Cl 0
C")
I"-
I"-
r--. r--. r--.
I"-
I"-
CX)
CX)
CX)
CX)
CX)
CX)
CX)
CX)
CX)
CX)
Cl
Cohor1
-Cl
cG)
1600
'-
u
c 1000
c
ca
-E
.:.::
tr
11l
800
600
400
..J
U
1
I
I
,~
1400 I
1200
E
G)
'C
40
200
0
+ /.
"".
+/X
-~ + /+"",
"'"+
+-+
I
t
35
30
25
- - 0 - 3KSWINC
20
3LSWINC
15
I
- - D - - 2JSWINC
•
RUNCIL
10
I
I
5
0
I
LO
r--.
CD
I"-
r--.
r--.
CX)
l"-
Cl
r--.
o
CX)
T""
CX)
C\I
CX)
c")
CX)
~
CX)
LO
CD
r--.
CX)
CX)
CX)
CX)
CX)
Cl
CX)
Cohor1
Fig. 17. The total increment in length and weight in each cohort from the beginning of age 2 to the
end of age 5 plotted together with the average area of the eIL over that period (all three sections
combined).