mp

f

«-«

Northwest and >

Alaska i
Fisheries Center

National Marine '

Fisheries Service j

U.S. DEPARTMENT OF COMMERCE

NWAFC PROCESSED REPORT 83-01

Skeleton Bulk Biomass
Ecosystem Model
(SKEBUB)

January 1983 ^ WHoi

DOCUMENT

COLLECTION

This report does not constitute a publication and is for information
only. All data herein are to be considered provisional.

W H 0 I

DOCUMENT

COLLECTION

m

D

_□

tr

o ■

D

3-

g

D

0

S^^

r^

a

m

CD

CD

SKELETON BULK BIOMASS
ECOSYSTEM MODEL (SKEBUB)
By
Nicholas Bax-

Report for NOAA, NMFS Contract No. 82-001i»5

Resource Ecology and Fisheries Management Division

Northwest and Alaska Fisheries Center

National Marine Fisheries Service

National Oceanic and Atmospheric Administration

2725 Montlake Boulevard East

Seattle, Washington 98112

January I983
* Compass Systems Inc., kG^O Jewell St. #20i», San Diego, CA 92109

CONTENTS

Page

1. Introduction 1

2. Formulae in SKEBUB 3

Monthly changes in biomass 3

Apex predators 3

Fish species and benthos 3

Zooplankton and phytoplankton ^

Components of the biomass equation k

Growth (GS|^ ^) ^

Natural mortality (GMS ,) 5

Consumption by apex predators (SS) 6

Consumption by other species (CCj^ ^) 7

Total food requirement of each species grouping (FOODj^) 7

Amount consumed of each prey by each predator (CPN|^ 1^) 8

Calculation of equilibrium biomasses 9

3. Simulation of food suitability and availability dependent feeding 10

Simulations of feeding in SKEBUB 12

Resul ts and Limi tat ions 13

h. Preliminary results from SKEBUB 13

5 . References 1°

6. Appendix tables 25

I I

LIST OF TABLES AND FIGURES

Table 1 . --Prel iminary output data from SKEBUB.

Figure 1. --Annual mean biomasses of fish species groups and benthos, showing the

change from input to equilibrium values.
Figure 2. --Annual mean biomasses of fish species groups and benthos after

equilibrium, showing the effect of a 75% reduction In the biomass

of the silver hake group at year 10.
Figure 3-~"Annual mean biomasses of fish species groups and benthos after

equilibrium, showing the effect of increasing the fishing pressure

on the silver hake group fourfold at year 10.

I I I

LIST OF APPENDIX TABLES

Table 1. --Sample composition of biota for SKEBUB.

Table 2. --Input biomasses and percent similarity in their diets.

Table 3---Sample inputs: initial biomasses, growth, mortality, and food

coef f icients .
Table '♦.--Monthly mean temperatures, acclimatization temperatures, and plankton

parameters .
Table 5 . --Parameters for simulation of monthly mean apex predatory biomasses, and

their food composition.
Table 6. --Mean food composition of sample fish biota.

INTRODUCTION

Fisheries ecosystems have been modelled since at least the turn of the century.
Nowadays the management of the world's fisheries is based, in part, on several
well established models of varying complexity; for example, Schaeffer's surplus
production model, Beverton and Holt's, or Ricker's, analytical models, and virtual
population analysis. These models are based on a species by species approach to
fisheries ecosystems, a fact which has led to criticism of their continued use, since
the realization of the importance of species interactions.

To answer this criticism, models have been introduced which provide for inter-
action between the species in the management unit or ecosystem. This interaction
has most commonly taken the form of interspecific predation and these models
have undergone considerable development since their introduction to fisheries in
the late 60's and early 70's. Ursin (1982) summarizes the models which have been
applied to the management of marine fisheries.

A consequence of the increased complexity of these models over the single
species approaches has been the requirement for more and more data and, when these
data are not available, their estimation or substitution with unproven formulae.
This has led to a reluctance to incorporate multispecies models into fishery
management decisions and occasionally their derision in the scientific press
(Gulland 1982), A diversity of available models has resulted, from the concise
multispecies virtual population analysis (e.g. Pope 1979) with biologically
identifiable data requirements, to the detailed analytical model of Anderson and
Ursin (1977) with requirements for parameters of ambiguous biological meaning.
Both approaches have their merits and it is only through the continued development
of models and their comparison that a holistic view of fisheries ecosystems will
be obta i ned .

A model lying between the two extremes in complexity is presented here. It
is a simplification of the extensive biomass-based models of Laevastu and Larkins
(1981) without spatial resolution. The formulae used in this model correspond to
those of the larger models (DYNUMES and PROBUB) and are characterized by their
straightforward relationship to the available data. Unknown parameters and
coefficients for which there are little or no data are kept to a minimum. In
general, the formulae are of a linear form, in the absence of any biologic
information to the contrary.

Abbreviated ecosystem models are an aid in evaluating the relative importance
of component processes in the ecosystem. They have the advantage of being readily
assimilable by both the modellers and their audience, but they are a simplification,
Lack of spatial heterogeneity and a lack of recruit variability in particular,
limit the results to average solutions. Their purpose is not to attempt to define
the ecosystem and its processes, but rather to foster an understanding of the
general interactive processes.

This paper presents the formulae used in the model, followed by a more detailed
examination of the interactive processes in the feeding routines. The reader is
referred to Laevastu and Larkins (I98I) for a comprehensive discussion of the
attributes of biomass-based models in general. It is stressed that the formulae
presented herein are not invariable, but should be changed to reflect available
information on the species under consideration or the beliefs or hypotheses of
the researcher. Sample input data and parameters are presented in the following
sections, together with results from this preliminary modelling. The species
composition has a rough correspondence with that of Georges Bank, but the data
do not necessarily correspond to published data.

FORMULAE IN SKEBUB

Monthly Changes in Biomass
APEX PREDATORS

The apex predators are divided into three g roups- -mamma 1 s , birds, and sharks.
Because of the high mobility of these predators, their biomasses are not computed
in this model. The mean annual biomass of each predator group is an input value.
Monthly deviations from this mean biomass are described by a simple harmonic
formula, (cosine), where the frequency and magnitude of the oscillations are input
values, and can be used to describe migrations in and out of the area.

FISH SPECIES AND BENTHOS

The following set of equations applies to benthos and all groups of fish
species with the exception of the squids. The data on the squids are insufficient
to enable the dynamic computation of their biomass, consequently, their biomass
is maintained in the same manner as the apex predators. Other species do prey on
the squids and the squids do prey on other species.

The central formula for the monthly updating of biomass of a species grouping
(N) in month k is:

^^N.k - ^^^N,k - ^''N,k - S,k

^^N,k= ^^N,k-1 " ^

where GS,, , is the growth, CMS,, , the natural mortality excluding predation,

N ,k N ,k

FP.. , the fishing mortality, and C , the consumption of N by other species.

IN J K iN J K

An iterative procedure is used to update this equation:

a) an initial, and approximate, computation uses the previous month's (k-1)
values of growth, natural mortality, fishing mortality, and predation. This

-i»-

computation is used to determine the current month's feeding computations.

b) the computation is repeated with the component parameters adjusted to
reflect the current month's feeding and biomasses, as indicated below.

c) step b is repeated until the biomasses values of successive iterations
converge. In practice it has been found that the values converge rapidly and
step b is computed once only.

ZOOPLANKTON AND PHYTOPLANKTON

The monthly biomasses of zooplanl<ton and phytoplankton are computed in the
same manner as those of the apex predators, namely, an input mean annual biomass
is adjusted by a cosine function to simulate seasonal variation. The period and
magnitude of this seasonal variation are controlled by input parameters.

Components of the Biomass Equation
All components of the biomass equation are adjusted monthly by any or all of:
temperature, starvation and, following equilibrium, density dependent factors.

GROWTH (GS,, , ^
N,k;

Species specific input values for growth {G^.) are adjusted in the following
equations. The last of these equations (density dependence) is only computed once
equilibrium has been reached.

(J_ .L.)

^ re - r ■■- o TA TM'

Growth is adjusted by the difference between the current month's temperature,
TM, and the optimum, or acclimatization, temperature for the species, TA. Both
values are input parameters. The form of the equation follows Krueger (196'*)-

^) ^^N,k = ^^N,k - ^''h -' 0-°' ' 'h,k^

-5-

The growth rate of any species is reduced by the fraction of its food
requirement that it could not obtain in the current month, (SCj^ - 0.01, where
SC is a percentage). This adjustment is thus dependent on the biomass of the
species itself (i.e., its food requirement), and the biomass of the other species
which provide, or remove, potential prey items. This effect is assumed to be
linear, following Jones and Hislop (1978).

To provide for density dependent growth once the equilibrium position has
been reached, the growth for each species is adjusted by the ratio of the
equilibrium biomass of that species (V^) to its mean biomass in the previous
year, (BB ). This ratio is divided by 12 to correspond to the monthly growth

IN , ,

coefficient. The natural logarithm of this ratio is taken to correspond with
growth, which is expressed as an instantaneous rate.

Density dependent growth has not been identified as a universal characteristic
of fish species and potential researchers will have to consider the suitability
of this equation in their own situations. Implicit in the use of this equation
in this instance is the assumption that decreases in the biomass are linked with
reduction in the percentage of adults in the population (e.g., resulting from
fishing mortality). Individual growth rates decrease with increasing age
(Paloheimo and Dickie 1965) thus the decrease in mean age of the population will
produce an increase in mean biomass growth rate.

NATURAL MORTALITY {GHS^ ^)

Natural mortality excludes that due to predation, which is computed separately,
It includes spawning stress and senescent mortality, together with residual

mortalities, for example, mortality due to disease (Laevastu and Larkins I98I).

Input, species specific, natural mortality (SM ) is assumed to increase as a

linear function of starvation (SC.,) :

N

FISHING MORTALITY (FP„ , )

N ,k

The fishing mortality has two components--a constant rate of mortality, FK ,
and a density dependent component, FM . Before equilibrium, total fishing mortality
is simply set at twice the constant component (i.e. 2 x FK ) . This value corresponds
to the available data on fishing mortality. After equilibrium, total fishing
mortality is equal to the constant rate plus the density dependent component.
The form of this density dependence is currently represented by the following
equat i on :

^N
FM., = FK. ■' ■'

N N ' V„

FP, = FM^ . FK^

The division of the fishing mortality into two components allows the
consideration of the relatively constant components (e.g., bycatch and
artisinal fisheries) separately from the density dependent component (e.g.,
management adjustments). This formula may be modified in future simulations to
incorporate species specific factors.

CONSUMPTION BY APEX PREDATORS (SS)

The food requirements of the apex predators are a function of their biomasses,
Monthly variations in the biomass, and therefore the food requirements, do occur,
but from year to year they are invariant. Consequently, their food requirements
are computed once at the beginning of the model. Together with the percentage

food composition table for each predator, this gives the monthly consumption of
each prey group by the apex predators. This is removed each month irrespective
of the current biomass of any of the prey groups. In areas with high apex
predator biomasses this would need adjusting to allow prey switching in response
to prey availability.

CONSUMPTION BY OTHER SPECIES (CC^^ ^)

Food requirements of these species groupings are dependent on their growth in
the current month. Initial estimates of growth in any month are computed using
the previous month's values for the parameters in the biomass equation. Once
consumption is computed, the growth is recomputed. It is possible to recompute
consumption at this stage to reflect the updated growth, but in practice this has
not been found to affect the results. Thus the consumption is computed only once
in each month. The steps are as follows:

Total Food Requirement of Each Species Grouping (FOOD )

The food requirement consists of a food requirement for growth and a food
requirement for maintenance. The food requirement for growth is computed from
the actual growth in biomass, (GRO) , that is, growth uncorrected by fishing
mortality, natural mortality, or consumption:

GS
GRO = BBI 'Me '^''^ - 1)

The maintenance requirement is computed as the maintenance requirement for

the mean of the current and the previous month's biomass, BBI. An arithmetic

mean is used following Ricker (1975, p. 239). Thus the total food requirement

can be expressed as:

FOOD,, =(BBI -'■ FRM -- 30) + (gRO " FRG)
N

-8-

where FRM is the coefficient for daily food requirement for maintenance, and FRG
is the coefficient of food requirement for growth. Both coefficients are expressed
as a fraction of the biomass or growth in biomass, and are input parameters.

Amount Consumed of Each Prey by Each Predator (CPN )

There are two sets of inputs that determine the amount of consumption of a
prey group by a predator. Firstly, a table of the percentage food composition of
each predator (derived from stomach analyses modified to include the juvenile
feeding) is input, CF , , and together with the food requirement for each predator,
(FOOD ) , the actual food required by each predator group from each prey group,
(CPN ), is determined. These values are summed over each prey group to give

IN J K

the total amount of food required from each group, (CC , ) .

Total food requirement (CC.. ,) is compared with the allowable consumption of

IN I K

each prey group's biomass, (AC,). The allowable consumption is the second set
of input values, where the allowable consumption is input as a percentage of
that prey group's biomass, (AP, ) , and the allowable consumption, (AC,), is
computed from the current biomass of that group.

At this stage, then, the food required by each predator group of each prey
group and the allowable consumption for each prey group are known. These values
are now used to modify the input food composition table to reflect the influence
of prey abundance on feeding. The ratio of allowable to required food consumption
of each prey group is formed, (FCOC, ) , and the adjusted percentage food composition
of the predator's diet is made a function of this ratio and the input percentage
food composition:

1 + Ae"^
FCN., , = CF. •'- ' ^^

N.k N,k , ^ ^^-B^VFCOC^

-9-

where A and B are input parameters which determine the magnitude and rate of any
change in food composition, respectively. The effect of this equation is
described in more detail in a later section. The mod i f ied food composition
table, (FCN) , is now used to recompute the actual food requirement of each
predator from each prey group, which is again compared to the amount allowable
from that prey group, and if overconsumpt ion of any prey group is still indicated,
the excess is removed from the predator's food for that month. The difference
between the required and the actual food consumption for any group is then
allotted to starvation, (50^,).

Calculation of Equilibrium Biomasses
Equilibrium is attained when the biomasses of the species groups are constant
from one year to the next. Monthly fluctuations do occur due to seasonal
variation in temperatures, apex predator consumption, and phytoplankton and
zooplankton biomasses. At equilibrium the growth in biomass must equal the losses
to each biomass from fishing, predation, and natural mortality. Thus, to attain
equilibrium, either the growth of the species can be varied, or the sources of
loss can be varied, but in an opposite direction. Growth is determined from
empirical data, and the mortality coefficients for fishing and natural causes
are assumed to remain constant from year to year (before equilibrium). Predation
is the logical variable to adjust to reach equilibrium and, rather than adjust
the percentage food composition (derived from stomach samples), the input
biomasses of the species groups are modified. An iterative procedure is used
which adjusts the biomasses at the end of each year's computations:

BB,, ,0 . = BB., , + BB,, ,„ - BB,, ,
N,12,b N,l,a N, 12 ,a N,l ,a

AGA

■10-

where BB,, ,„ , is the new December biomass, BB , is the previous year's January
N , I / , b N , I ,a

biomass. and BB. ,^ is the current December biomass. The iteration constant,
' N, 12 ,a

AGA, controls the degree of convergence towards the equilibrium and varies from
3.0 at the beginning of the iterative procedure to ^.8 at the end of the iterative

procedure. Thirty iterations, or "years", have been found necessary to reach a

. • '/
stable position.—

SIMULATION OF FOOD SUITABILITY AND
AVAILABILITY DEPENDENT FEEDING

Predator-prey relationships provide a readily comprehensible mechanism for
species interactions and are used for this purpose in many ecosystem models
(e.g., Andersen and Ursin 1977, Pope 1979, Helgason and Gislason 1979, Laevastu
and Larkins 1981). Ursin (1981) has discussed the constraints of the models of
Pope(l979) and Helgason and Gislason (1979), in particular their assumption
that growth and food consumption rates are independent of food concentration.
The remaining two models allow underconsumption by a predator species when prey
sources are limiting. In both instances the prey consumption by a predator is
made dependent on an index of suitability and the prey's biomass.

Andersen and Ursin (1977) compute the suitability, or vulnerability, index
as a composite of suitability with respect to size, the fractional overlap of
predator and prey in time and space, and the chance of encounter due to the
behavior of predator and prey. This index is frequently compared with available
stomach contents data to check the realism of the arrays. Laevastu and Larkins
(1981) have taken a more direct approach and base their suitability index on

1/ A "stable position" is defined as one where further iterations do not affect
~ year and biomasses. in practice this stabilization occurs when the percent

annual change in individual biomass varies from 0 to about 5%. Global adjustment
to the allowable percent consumption parameter is used to adjust the equilibrium
position so that the mean of the percent annual changes in individual biomasses
(which can be + or -) is close to zero.

-11

empirical data from stomach content analyses.

The proportion of a prey actually "occurring" in a predator's diet in each
time step is then made a function of this suitability index and the prey's current
biomass. In the Andersen and Ursin (1977) model, prey switching by a predator is
implicit and is in direct proportion to the relative proportion of each prey biomass.
Laevastu and Larkins (1981) explicitly allow prey switching; the suitability of
any prey type at any time step is a function of the input suitability index and
the ratio of required to available biomass for each species. This latter approach
requires the input of a parameter to determine the proportion of each biomass
available for consumption in each time step; otherwise overgrazing would occur.
This parameter is a fixed proportion of the growth for all fish species. In the
model of Andersen and Ursin (1977), actual food consumption by any predator is
then determined from the available prey and a half saturation constant, or
coefficient of rate of search. Once the adjusted suitability index has been
computed in the Laevastu and Larkins (1981) model, no further adjustment to
feeding occurs; if more food is required from a prey biomass than is deemed
allowable, the predators requiring that food suffer starvation. Both models,
then, require the estimation of a parameter unverifiable given current data. The
method of Andersen and Ursin (1977) computes the half saturation constant for
each species by trial in the model. Laevastu and Larkins (I98I) designate allowable
consumption as a fixed proportion of growth for all species. In SKEBUB this
proportion is assumed identical for all fish species, reducing the required parameter
estimates to one. This global allowable consumption parameter is adjusted so that
the estimates of the percent of required food that is unobtained (starvation) appear
consistent with available data. The method of Laevastu and Larkins (I98I) as used

■12-

in SKEBUB is presented in more detail below. The function relating the composition
of the diet to the biomass of the prey is modified to allow the modeller to
control the dependence of the final food composition of each predator on either
the input table of select ivit ies or on the allowable composition of each species.

SIMULATIONS OF FEEDING IN SKEBUB

At the beginning of each monthly time step the growth of each biomass and the
food required for the maintenance and growth of this biomass are computed. The
food required by each predator of each prey is then determined through the input
food selectivity table (CF^ , ) . The total requirement for all predators from
each prey type is computed and compared with the amount of each prey biomass
designated available. The input food selectivity table is now adjusted by a
function of the ratio of available to required food for each prey biomass,
(FCOCj^) :

f^^M ^ = CF, , [(1 + Ae"^ / (1 + Ae"^^''"^k))]
N ,k N,k

where FC , is the adjusted percentage food composition, and A and B are constants.

This equation is a modification of the logistic equation, forcing the point
of inflection to be at x = y = 1 (i.e. when the food required from a prey type
is equal to the amount available, (FCOC, = 1 ) , no adjustment is made
(FC|^ , = CF ,)). The maximum upward adjustment of the percentage food requirement
from any prey item is equal to:

MaxUp = 1 + Ae ,
and the maximum downward adjustment:

MaxDown = MaxUp/1 + A
The rate of change is specified by B, although a value of 1.5 has been found
reasonable in practice. The effect of increasing A is a parallel increase in
the maximum upward and downward adjustments.

•13-

Results and Limitations

Complete utilization of all available prey can be obtained with a high value
for A or by iterating the procedure several times. Five iterations with A = 1.5
and B - 1.5 produced complete consumption of all prey biomasses when prey was
limiting. Although, superficially, this appears reasonable, it is apparent
that the feeding routine becomes highly dependent on the amount of each biomass
designated available for consumption which is difficult, if not impossible,
to quantify. In practice, one iteration with B - 1.5 and A - 1.5 was found
satisfactory, producing a stable equilibrium position with the equilibrium food
composition reflecting the input distribution.

The approach is versatile enough to enable emphasis to be put on either the
input food composition table, or on the estimated prey availability. The level
of adjustment can be made specific to each predator biomass and used to investigate
the effects of a high or low prey selectivity by a predator.

PRELIMINARY RESULTS FROM SKEBUB

These results illustrate the output from a simulation on the sample data
given in Appendix Tables 1 to 6. They are presented here to provide the reader
with an illustration of the simulations in SKEBUB and are not intended to test
biological hypotheses.

Many data result from a single run of SKEBUB. Laevastu and Bax (1982)
detail possible data outputs. Here we present mainly annual mean biomasses of
the various species groupings. There are two stages in the simulation. The
first stage produces an equilibrium situation given the input data. The
equilibrium biomass of a species group is that biomass which can be sustained
such that the growth of the biomass equals its mortalities. The equilibrium
situation may be considered an unnatural one, but it is a necessary starting

■]h-

point and base for the comparison of the effects of various factors on the
ecosystem (Laevastu and Larl<ins 198l)- Figure 1 illustrates the changes in
the species groups biomasses from their input until equilibrium is reached 30
years after the start of the run. The biomasses change again immediately
following equilibrium when the equ i 1 i br iat ion procedure is removed and the input
food selectivity table is replaced by the one adjusted at equilibrium. This
replacement of the input food composition table is not necessary to the running
of the model and is inadvisable when the input food composition table is
estimated with confidence.

After equilibrium the density dependent influences on growth come into play
and maintain the biomasses at a fairly constant level (years 1 to 9, Fig. 2).
In the tenth year following equilibrium two severe and arbitrary adjustments
were made to the biomasses to study the maintaining effect of the density
dependent processes. In Figure 2 the biomass of the silver hake species group
was reduced by 75 percent at the start of year ten. In the subsequent ten years
all biomasses fluctuate until approximately the original equilibrium distribution
of biomasses is regained. In Figure 3 the fishing pressure on the silver hake
species group was raised fourfold at the start of the tenth year following
equilibrium. Again equilibrium was regained in the subsequent ten years, but

in this instance a new equilibrium distribution of biomasses was reached,

2/
reflecting the continuing effect of the increased fishing pressure.—

Sample output data are given in Table 1. These data are the annual mean
values at equilibrium.

2/ The ordinate scale varies for each species group on these graphs. The
~ biomass of benthos was reduced by a factor of 10 for plotting.

-15-

SKEBUB can thus be run in two modes. In the first instance, input variables
are modified and the output data at equilibrium compared. This provides
information on the sensitivity of the simulation to changes (or error) in the
input data and demonstrates the interactive effects of the model. This mode
is used to investigate the relative importance of parameters or species groups
to the overall biomass and to the biomasses of other species groups. The
second mode starts when the equilibrium forcing constraints are replaced with
density dependence and is primarily of use in studying the interactions of the
species groups following a perturbation (e.g. increased fishing effort on one
species). The time scale of these interactions is strongly dependent on the
form of the density dependent formulae.

As is the case with all simplified ecosystem models it is the repeated
running and hypothesis testing that provides the most information, rather than
the actual output data from any one simulation. For this reason the code of
the model has been kept simple to facilitate users in developing a clear
understanding of the component processes. Processing time for a 70 year
simulation with 13 species groups is under 30 sees, on a Burroughs B7800
mainframe computer, and has a core requirement of less than 8000 words. The
FORTRAN code for SKEBUB is available as printed output or on 5 W^ inch
soft-sectored disl<ettes written on an Osborne I micro-computer,

■16-

REFERENCES
Andersen, K.P., and E. Ursin.

1977. A multispecies extension to the Beverton and Holt theory of fishing,
with accounts of phosphorous circulation and primary production. Medd . Dan.
Fisi<. Havunders. 7:319-'*35.

Gul land, J. A.

1982. Long-term potential effects from management of the fish resources of
the North Atlantic. J. Cons. int. Explor. Mer. '40(1)8-16.
Helgason, T., and H. Gislason.

1979. VPA - analysis with species interaction due to predation. ICES CM.
1979/G:52. 10 p. (Mimeo) .
Jones, R., and J.R.G. Hislop.

1978. Further observations on the relation between food intake and growth of
gadoids in captivity. J. Cons. int. Explor. Mer, 38(2) :2'+'4-251 .

Kreuger, F.

196'*. Neuere matematische Formul ierung der biologischen Temperatur-f unkt ion
und des Wachstums. Helolander Wiss. Meer., 9:'08-124.
Laevastu, T., and H.A. Larkins.

1981. Marine Fisheries Ecosystem: Its Quantitative evaluation and management.
Fishing News Books Ltd., Surrey, Farnham, England. 162 p.

Laevastu, T., and N. Bax.

1982. An abbreviated prognostic bulk biomass ecosystem model (SKEBUB) . Program
Doc. No. ]h. NWAFC, 2725 Montlake Blvd. E., Seattle, WA 98112.

Paloheimo, J.E., and L.M. Dickie.

1965. Food and growth of fishes. I. A growth curve derived from experimental
data. J. Fish. Res. Bd . Canada. 22:521-5'»2.

-17-

Pope, J.G.

1979. A modified cohort analysis in which constant natural mortality is
replaced by estimates of predation levels. ICES CM1979/H:l6. 8 p. (Mimeo)
Sissenwine, M.P., E.B. Cohen, and M.D. Grosslein.

1982. Structure of the Georges Banl< Ecosystem. ICES Symp. on Biolog. Prod.
Continental Shelves in the Temperate Zone of the N. Atlantic. No. 31.
(Mimeo) .
Ursin, E .

1982. Multispecies fish stocl< and yield assessment in ICES. p. 39-'*7. J_n
M.C. Mercer (ed.). Multispecies approaches to fisheries management advice.
Can. Spec. Publ . Fish. Aquat. Sci. 59-

-18-

Z

IaJ

3

O

o

.J

«

QC

c
111

o

o

o

«
X

ii-Hi

X

^ S

V M

s

R
S

?!

s

id
t

Hi

s

X!

s §

ni

•I

- I-

> QC

«t

in ^

J"

n 1^

CD

c

0)

en

c

3
o

i/i

o

x:
+-t
c

0)
XI

XI

c

03

l/l
CL

D

o

I-
en

ui
dJ

O
0)

CL

l/l

E
O

c
ft)

0)

E

C
C
<

dJ

1_

cn

>

E

3
cr

D
C

E
O

3W>i/o>i 0001 ssvwoia

•19-

Z
(A

<A

X

o

i

<A

bl

c
o
o

1-
C71

T~k T~it y n T~i

W M M

n

3W>I/D>I 0001 ssvwoia

-20-

I-
^ t

c

3
O

E

D
cr

If)
O

>-

Q.

o

>

4~t t/1

c

»

■D

C

14-

s

n3

O

S

in
Q.

in
in

M

3

(D

t

m

O

\-
cn

E
O

m4

M

-D

ni

•J

in

s

M

(U
O

v4

o

0)

lU

Cl

c

s

in

*"~

*4

QC

-C

c

111

in

O

H

•—

'—

s

Ilk

!t

14-
<4-

4-J

o

13

O

■D

t

i

a)

(D

1-

n

U

1/1

<K°

«4

>

i/i

LA

t!

r^

t

E
O

OJ

««

■—

««

JD

4-

o

s

C

03

■M

e»

0)

o

E

(U

s

tl

<u

r-

D

C

Q)

s

C
<

Ift

t

(NI

n

•

0)

cn

,W>I/3>I 0001 SSVW0I9

M-

c
o
o

O)

u

jWN/DM 0001 SSVW0I8

-22-

X
Ul

o
o
o

X

«

i:

HI

o

u
0)

O

o

tT^T^l

M
li.

I-
«

•J

b.

iT

s c

a 2

• m

a "^

C; ID

Ilk ^

c

3

O

E

3
0)

o

■l-J
C
(U
X)

■a
c

o

I.
en

in
(U

O
(1)
Q.
1/1

O

a.

3

o

CT)

ID

4)
>

0)

c
o

(U
u
3
1/1
l/l
0)

C

1/1

u-

cn

O

c

n

t/1

u

(T)

in

(1)

l/l

i_

0)

O

(-

c

o

—

XI

V4-

o

c

(D

4-»

0)

o

1-

(U

u-

r^

u-

1)

<u

3

C

0)

C

jr

<

■4-J

fv-i

o;

u

3

CTl

u>

ran MM V « .«

7W>I/D>I 0001 SSVW0I8

■23-

3

<A

V)

o

111

r

liJ

• «• inni <H« «n

2W>i/9>i 0001 ssvwoia

S
»

si:

c
o
o

-2k-

m

CO
UJ

:>^

-D

■3

a.

4-1

3

o

l_
ID

C

Q.
I

Si
OJ

>■ 4->

■— c

^ (U

4J _C .-
C 4-1 u

2 —

O 14-

\- <4-

cn <U
O

u

o

E

c

<U

-a
>*- o x>
O O 0) r^

«- C

c

■o

TO

<u

OJ

4-^

u

l-

J3

u

•*

0

0)

3

1.

cr

4-1

0)

o

u

c

T3
O
O

U1

TO i/l
I/)

B
O

>- c

— 0)

— E
(TJ 0) -Q

T3 1-

•— y-

c 3 O

(U 0) (5^

z: 1-

ON?

4-1 ;/i

Cl l/l

E rt)

3 E

lA O

C —

O XI

o

c

•— nj

nj 0)

3 E

c
c

< o

a

3
O

I-

o

OJ

u

o.
to

1^

■— -3-

oo o

o —

o

en

LTi

o

en

oo

-a-

— o —

-a-

-3-

cn

ra

1^

CM

<Ti

vD

^

O

vD

oo

, —

-a-

CO

00

00

o

00

o

O

O

O —

CNI —

LTV r~-

^

in

o

in

4-1

<D

rt)

F

o

O

o

ro

r—

J3

<D

3

C

C

<0

C

(U

<

E

00

o

O

LA

O

LA

CM

-3-

crs

CNl

vT)

o

. —

r-~

r-

r~~

1^

1 —

o^

LTV

rA

^

in

y-

■D

OJ

0)

c

.

+

1-

in

3
4-1

O

(U

+

>—

■z

TO

^

14-

c

+

^

^

^-

JH

o

•—

^—

in

in

-i!

(D

o

y-

O)

(U

■—

in

•—

O

in

i-

c

1-

in

14-

O

14-

O

u

(1)

^

L.

.—

0)

■D

w—

x:

4-1

T3

(U

>

o

<u

I-

J^

•—

< —

4-1

(D

-o

E

x>

-—

Q.

x:

1-

o

3

0)

c

(TJ

OJ

o

.—

4->

(U

fO

CT

JI

(U

tZ

I

o

o

00

o

I

z:

00

oo

CQ

-25-

APPENDIX TABLES

-26-

Table 1. --Sample composition of biota for SKEBUB.

Group No.

Name

1.

Species 1

2.

Flatfish

3.

Haddoci<

i*.

Demersal nc.

5.

Cod

6.

Semidemersal

7.

Otiier finfish

8.

Herring

9.

Pelag ic

10.

Squids

11.

Shellfish

12.

Benthos

13.

Zooplankton

]k.

Phytoplankton

Species Composition

Any species under special consideration

Yellowtail, winter flounder, etc.

Haddock

Dogfish, skates, red hake, etc.

Cod

Pollock, redfish, silver hake

Argentines, sand lance, etc.

Herr ing

Mackerel , tuna

1 1 lex, lol igo

Commercial species, including lobsters

APEX PREDATORS

Birds (Fulmars, kittiwakes, storm petrels)

Mammals (Odontocete whales (fish feeders), fin and right whales, pilot whales,
bottlenose dolphins)

Sharks

Man (Fishing)

-27-
TABLE 2. --Input biomasses and percent similarity in their diets.—

Input biomass
Group No. Species composition Percent simi lar ity (kg/km )

Yel lowta i 1 i rno,

Winter flounder)

3. Haddoci<

k. Dogfish / -,n, (

Skates ( '°(cr.o, I 9°^

Red hai<e, etc

iGO%

Cod

{ 69^o

6.

Pollock

Redf ish

S i 1 ver hake

7.

Other finfish

8.

Herr i ng

9.

Mackerel

Tuna

10.

lllex

Lol igo

[37%

li2%

600

300

900

'♦.soo

(500)
(300)
1,100

1,900

1 ,900

300
lop

10,100

10,500

1,700

19,300

1,300
(200)

1,500

280
i^OO

im

II. Shellfish 6,300 12. Benthos 120,000 13- Zooplankton ^45,000
]k. Phytoplankton 60,000

1/ Data from Grosslein et al.

-28-

o O 3

q: o o

Ll. Ll. U

XJ c

z: o —

Qi O (D

Ll_ (J- E

c

0)

w-

c r

<4-

(0 ^

1)

•M O

o

l/l 4-J

o

i^ c tu

la- O O

■D

O

o

~

o

y-

-a

c

(t)

o

E

o
en

in

(D

E
O

X -M o
1/5 U O

o

S u-
O 0)

>- o

3

Q.

Ql

E
(D

I
I

(U

J3
»0

— (/ItNJ

fO in E

— (D J^
4-" E ^

— O en
c — ^

— XI

tn :

(U

— Q.

o D

<u o

Q. 1-

1/1 CD

CNJ

CM

CM

LA

un

o

o

o

r^

r^

r^

vD

vO

r~^

vO

en

LA
CSI

LA

LA

CM

-a-

LA

t~^

CM

OA

LA

LA

LA

LA

LA

LA

LA

LA

LA

LA

-3"

CA

O

O

O

O

O

O

O

O

O

O

O

LA

LA

CM

LA

LA

CS

^—

O

CM

CM

O

O

O

O

O

CSI CSI Csl

O ^ —

O O O

LA CM (SI

o — o

O O O

LA

LA

CO

LA

LA

CXD

LA

LA

o

o

o

O

O

o

O

O

O

O

o

CSI

1—

^—

O

O

o

O

o

O

O

o

o

o

o

LA

00

LA

LA

O

O

LA

LA

O

o

LA

1^

LA

SD

00

CTv

CO

OO

OO

LA

sD

PA

o

o

O

O

O

O

O

O

' —

O

O 1

o o o o o o

o o o o o o

CTv LA Cri 0~\ LA [^

J- — — o ^

o

O

O

O

O

o

o

o

O

OO

O

o

o

o

CA

LA

vJD

rA

o

o

o

(T\

f—

vD

o

LA

o

sO

vO

CTi

o —

-29-

Table '4. --Monthly mean temperatures, acclimatization temperatures, and plankton
parameters,

Month
1
2

3

k
5
6

7

8

9

10
11
12

Water temperature

Species

Accl

imat

ization temperature

TM

Group number

TA

7

1

5.5

2

7.5

k.8

3

7.5

5.0

4

7.0

5.5

5

8.5

7.5

6

9.5

10.0

7

9.5

13.0

8

9.5

13.0

9

11 .0

13.0

10

12.0

11.5

11

7.0

9.5

12
13
14

7.0

Phytoplankton: V(l4) - 60 t/km ZF-30 ZFKA-170
Zooplankton: V(13) - ^5 t/km^ ZZ-20 ZZKA-240

-30-

Table 5 . --Parameters for simulation of monthly mean apex predator biomasses, and
their food composition

2

Area: 53,000 km

2

Birds: Mean 35 kg/km half magnitude of annual change 20 k 220 18^ BWD

2
Mammals: Mean 200 kg/km half magnitude of annual change 120 k 180 ^.S% BWD

2
Sharks: Mean 35 kg/km half magnitude of annual change 15 k 180 2.2^ BWD

2
Squids: Mean 680 kg/km half magnitude of annual change 210 k 180 (0.52 +

Las'* X G)

Food item

(Species group
number)

Birds

Food composition

Mamma 1 s

Sharks

°/

Squ ids

2
3

5
6

7
8

9
10
11
12

13

14

2
2

5
5
7

13
10

10

k

2

36

2

3
8
8
8

12
4
5

18
5
6

21

3
2

15

12

12

15

10

10

12

2

2

5

2

6

3
8
8
12
6

7

2

1

41

Q.

E

o
o

in

<u

u

•

(U

ns

Q.

w

(/)

O

14-

JD

o

j:

(^p

ON

00

c

ro

o

4->

1/1

O

Q.

E

CM

O

o

-a

O

o

y-

—

c

(D

(U

I

s:

fl)

1

in

XI

1

lU

F

•

—

3

vD

O

C

0)

<U

CL

Q,

^

1/5

S

J3

o

flj

L.

LTV
LA

-31

vO 00

v£) v£) LPi r- CO — —

Csl (N] • — — CSI LA 00

LA -3"

LA
LA

o
o

o

o

CVJ

CSI

O O O

— — -T

o

o

o

o

o

o

o

1 —

_

Csl

—

c^

^

o

rA

o o o o o

-3-

-a-

o
oo

o

CM

LA LA O

O O -3-

o

O

o

o

o

O

O

O

—

—

1^

LA

LA

OA

—

t-A

o

o

o

O

O

O

o

O

rA

OA

LA

t^

O

-3-

CM

LA

LA LA O

O O CSI

r--

PA

o

LA

o —

o

CM

PA

o

O

O

O

O

o

O

o

LA

-3-

CM

CNI

r^j

CM

_

,—

o o o
r~~ — —

o

O

O

O

MD

vO

—

—

O

O

o

o

OO

CSI

CA

rA

O

O

O

O

O

O

O

o

O

O

O

-3-

-T

CA

-3-

vD

CM

CSI

^D

oo

CN

CSI

CM

csl

CSJ

CM

LA

O

vD

LA

O

O

O

rA

CA

CA

vD

sD

vO

CA

sD

00

—

_

LAOOOOOOO
CA-3-PA' — CSIrA — •—

o o o

-3- — —

O O O

vD — ^

LALAOOLAOOO
-3-0 — — — CSI — ^

O

r-.

•— o

rv-\ CA