Fish Growth Modelling

Universidad
de Las Palmas
de Gran Caaria
Universitat
de les Illes
Balears
Universit
de Lige
(coordinateur)
Universit
de Bretagne
Occidentale Brest
Universit
de Corse
Pascal Paoli

Universit
Pierre et Marie Curie
Paris VI
Universidade
Nova de Lisboa
Universitat Politcnica de
Catalunya
Universidade Tcnica
de Lisboa

DIPLOME DTUDES APPROFONDIES EUROPEEN EN
MODELISATION DE LENVIRONNEMENT MARIN

SOCRATES/ERASMUS

FISH GROWTH MODELLING
Growth of the European anchovy (Engraul i s
encrasi col us) in the Tagus Estuary, Portugal

Rita Susana Oliveira MONTEIRO

Anne 2001/2002

DEA ? Modelling of the Marine Environment ? 01/02

If you can imagine it, you can create it.
If you can dream it, you can become it.
William Arthur Ward

i
INDEX

1 INTRODUCTION.............................................................................................................. 1
1.1 Tagus Estuary............................................................................................................. 3
1.1.1 Morphology........................................................................................................ 3
1.1.2 Physical and Biological Properties..................................................................... 4
1.1.3 Fish in the Tagus Estuary................................................................................... 5
1.2 European Anchovy, Engraulis encrasicolus ............................................................... 6
2 METHODOLOGY............................................................................................................ 9
2.1 State of the Art ........................................................................................................... 9
2.2 Conceptual Model .................................................................................................... 12
2.3 Individual Growth Model (IGM) ............................................................................ 13
2.3.1 Interaction between fish and habitat................................................................. 15
2.4 Population Growth Model........................................................................................ 16
2.4.1 Interaction between fish population and habitat .............................................. 18
2.5 Fish movement ......................................................................................................... 19
3 MODEL IMPLEMENTATION....................................................................................... 22
3.1 Individual Growth Model......................................................................................... 22
4 RESULTS AND DISCUSSION...................................................................................... 26
4.1 Individual Growth Model......................................................................................... 26
4.2.1 Life Cycle in the Fish Population..................................................................... 30
4.2.2 Scenarios .......................................................................................................... 31
5 CONCLUSIONS.............................................................................................................. 38
6 RECOMMENDATIONS FOR FUTURE WORK........................................................... 40
7 ACKNOWLEDGMENTS................................................................................................ 41
8 REFERENCES................................................................................................................. 43
A APPENDIX..................................................................................................................... 49
A.1 MOHID2000 ............................................................................................................ 49
A.1.1 Lagrangean Tracers Module............................................................................. 51
A.1.2 Eulerean Transport Module.............................................................................. 51
A.1.3 Cohesive Sediment Transport Module ............................................................. 52
A.1.4 Water Quality Module...................................................................................... 52

1
1 INTRODUCTION
Coastal zones are the most heavily populated and ecologically vulnerable areas worldwide.
Here many different and often contradictory interests exist which have to be carefully
managed. For this purpose it is absolutely necessary to enhance the forecast capacity
regarding the medium to long-term evolution of the coastal marine system in view of human
activities and natural changes.
Forecast models for coastal waters, complemented by continuous monitoring systems and
supported by efficient databases, can provide this kind of assistance, helping in the
development of management concepts compatible with sustainable growing and with the
preservation of ecosystems integrity. These integrated systems, if used in a regular basis by
local authorities, can help them to achieve important improvements in coastal management
practice.
Regarding all the uses that an estuary is submitted to, it is important to assess the required
environmental quality for the maintenance of the fish stocks. It is important to know the
location of the specific coastal areas chosen by the fish for their spawning and feeding
activities, so that decision makers may have the knowledge of the more relevant areas to
protect and therefore the measures needed to be implemented at an upper level.
To be able to predict where is the fish going one has to gather biological and ecological field
data and the results of experiments that provide an estimate of the fish environmental
preferences. Oceans, lakes and ponds offer their inhabitants environments that are variable in
space and time. Movements in these systems, either vertically or horizontally, impose
gradients in growth and survival through the effects of temperature, food concentration,
sensory capabilities, predator density and detection risk. All large and commercially
important marine fish stocks undertake seasonal horizontal migrations. The extent of these
migrations varies with age, size and environmental conditions. Understanding the forces that
create spatial distributions is a major challenge to ecology, but also has clear economic
benefits to humans, by allowing more precise assessments of managed stocks (Giske et al,
1998). While field studies may reveal patterns at a given time and place and laboratory
investigations may isolate effects of single causes, models may combine several forces in
continuous space and time.
2
An effective management of the fisheries resources requires an understanding and the ability
to predict long-term fluctuations in abundance and biomass production due either to human
perturbations or environmental variations. Therefore it is required a detailed study of the life
stage specific processes of the fish. Recent study has indicated that very often it is insufficient
to study the averaged early-life dynamics of a year class to understand recruitment processes
in fish populations. To understand these dynamics it may be necessary to follow diary growth
and mortality for a period of time (Cowan Jr et al, 1999). Processes that operate during a
specific time of the year or during a specific life stage may be critical in determining annual
recruitment level. However, examinations of life stage specific processes are difficult to make
in the field. Simulation modelling is an attractive alternative that provides a context for
evaluating the effects of many interacting factors potentially important affecting recruitment.
This work was developed in the scope of the european DEA (Diploma Eruditionum Altarum)
in Modelling of the Marine Environment. The goal of this work is the development of a
growth model for a fish population in a coastal ecosystem. In the near future, this growth
model will provide the basis to build a model of the movement of fish in the marine
environment according to their environmental preferences, permitting then to predict the fish
preferred areas. These models will be inserted in the already developed ecohydrodynamic
system, MOHID2000 (for more information see section A.1).
The Tagus Estuary (Portugal) is selected as the framework to study the growth of the
population of european anchovy, Engraulis encrasicolus, since there inhabits a considerable
population of this species.
3
1.1 Tagus Estuary
1.1.1 Morphology
The Tagus estuary is the largest estuary of Portugal and one of the largest european estuaries.
It is located near Lisbon (3844N, 908W), capital of Portugal and the most important city
built on its margins (see Fig.1). The metropolitan area has nowadays around 2 million
inhabitants, an important harbour and big industrial complexes around the estuary. It covers
an area of about 320 km
2
, has a length of 80 km from the upstream limit of the dynamic tidal
effect (near Muge) to the river mouth (near S. Julio da Barra Bugio) (Ferreira et al, 1994;
Costa, 1999). The submerged surface changes from 30000 ha in the neap tide to 34000 ha in
the spring tide (Cabrita & Moita, 1995).

Fig.1 Tagus Estuary.
The estuary is composed of a deep, straight and narrow inlet channel and a broad, shallow
inner bay. The inlet channel, allowing the entrance of seawater in the estuary, is about 15 km
long and 2 km wide (reaching a 40m depth), while the bay is about 25 km long and 15 km
4
wide (with a complex bottom topography that includes channels, tidal flat areas and sand
banks). Upstream, a single narrow channel marks the entrance of the Tagus river. The
maximum depth is 42 m near the mouth (with an average of 10 m), but a considerable
upstream area has 2-3 m. It is a mesotidal estuary with semi-diurnal tides with tidal ranges
varying from 1 to 4 m (Costa, 1999).
The main source of fresh water of the estuary is the Tagus river, with smaller contributions
from other rivers (Tranco, Sorraia). The discharge pattern of the Tagus river reflects the
regularization with Spain, the dry and wet seasons and large inter-annual variations: the
average discharge varies between 100 and 2200 m
3
.s
-1
(Pina, 2001; Costa, 1999; Cabrita &
Moita, 1995). The residence time of fresh water is highly variable and ranges from 65 days at
a discharge of 100 m
3
.s
-1
to 6 days at a discharge of 2200 m
3
.s
-1
; the residence time to the
modal discharge of 300-400 m
3
.s
-1
is of about 23 days (Martins et al, 1984).

1.1.2 Physical and Biological Properties
The estuary is a mixing place of river and oceanic waters. The salinity distribution depends
mostly on the river flow and on the mixing imposed by the tidal regime, which is the main
mechanism controlling the distribution of aquatic organisms and suspended particulate matter
in the estuary. There is a partial two-layer stratification: salinity is higher in the deeper layers
because the mixing of fresh and saltwater is slow (Cabrita & Moita, 1995).
Temperature variations influence the estuarine communities, by influencing the dissolved
oxygen and several pollutants concentrations. In the Tagus estuary, temperature range is
higher upstream, with maximum values at Vila Franca de Xira. There is a considerable
thermic homogeneity during winter compared with a pronounced longitudinal gradient during
summer.
The Tagus estuary is a highly productive ecosystem, and has considerable conservation value,
since it provides an optimum habitat for many crustacean, molluscs, fish and bird species.
This has led to the creation of a Natural Reserve in 1976, covering a large surface of estuarine
water, mud banks, saltpans, salt marshes, islands and agricultural land (Ferreira et al, 2001).
5
The Tagus estuary presents a high productivity and the most abundant primary producers are
the salt marsh vegetation, microphytobenthos, and macrophyte algae; together they provide
food for benthic fauna and habitat for many species. The primary producers are responsible
for the removal of high amounts of nutrients, which improves the estuarine water quality.
Although the estuary is recognised as a highly productive ecosystem, some of its natural
resources have been degraded in the last 30 years due to the increased water and soil
pollution. Directly or indirectly the estuary is the final disposal place of the catchment-
produced effluents. The implementation of the EU Water Directives is progressively reducing
the organic matter and nutrients load in the estuary. The quantification of the maximum
admissible loads discharged from each affluent sub-basins (Tagus, Tranco and Sorraia and
smaller streams) requires a good understanding of the biological processes inside the estuary
and in the coastal zone near the estuary's mouth, which is an important recreational and
leisure zone.
All those factors contribute for the need of efficient managing tools that allow the local
authorities to make decisions based on the better knowledge that they may have available in
each moment. To achieve this knowledge it is necessary to collect information about the
different parameters that play an important role in the ecosystem, integrate this information in
order to make it representative of the whole estuary and make it available in a comprehensive
way to the decision makers.

1.1.3 Fish in the Tagus Estuary
The maintenance of the coastal fisheries stocks is to a large extent dependent on the natural
conditions of the estuary, since several fish species use it as a nursery area, or migrate through
the estuary at different stages in their life cycle. The nursery zones are located in the
shallowest upstream part of the estuary, near Vila Franca de Xira, Montijo and Alcochete.
The estuarine community varies throughout the year and is rare the fish that inhabit
permanently in the estuary. As the strong currents of the estuary can drag eggs, the majority
spawns in the sea. Nevertheless the species whose life cycle develops in the estuarine
environment have to arrange mechanisms that hinder this drag of eggs (Costa, 1999).
6
In 1996 and 1997, Costa (1999) identified 44 fish species in this estuary, including resident
species, species that use the estuary as a nursery, migrant and occasional species. Some of the
most abundant species in this estuary is the european seabass (Dicentrarchus labrax), the
common sole (Solea vulgaris), the gilthead seabream (Sparus aurata) with considerable
commercial interest, the sand and marbled goby (Pomatoschistus minutos and P. microps) and
the european anchovy (Engraulis encrasicolus), this last with commercial interest for
fisheries.
The anchovy and the goby live in this estuary throughout their life cycle, while the european
seabass, the common sole and the gilthead seabream use it as a nursery. These last species
spawn in the sea and the juveniles use the estuary as foreground nursery, where they find
greater amount of food, minor number of predators and good conditions of temperature; they
remain during a variable period depending on the species and return to the sea. Some adults
enter occasionally in the estuary to feed (Costa, 1999; Monvoisin, 1997).
The survival of the estuarine fish is affected by diverse factors as pollution, overfishing and
dams. The overfishing of several species in this estuary causes the decrease of the fish stocks
because juveniles are caught, thus preventing the adult population to increase and
consequently producing a minor number of eggs.

1.2 European Anchovy, Engraulis encrasicolus
The european anchovy Engraulis encrasicolus (family -
Engraulidae, order Clupeiformes) is a very abundant
species from Norway to England, all western Europe,
Mediterranean, Black and Azov seas, South Africa, Gulf
of Suez, western Indian Ocean, Mauritius, Seychelles and
upwelling area around Somalia, as shown in Fig.2 (Costa,
M.J., 1999; Whitehead, 1990; Whitehead et al, 1988).

Fig.2 Engraulis encrasicolus distribution (FAO in Froese & Pauly, 2000)
7
It is a pelagic and mainly coastal marine species, forming large schools and feeding on
planktonic organisms. It is a eurythermic and euryhaline species that tolerates salinities from
5-41 ppt and in some areas, enters estuaries, lagoons and lakes, especially during spawning
activity. Tends to move further north and into surface waters in summer, retreating and
descending in winter. This species adult mean weight is of 30 g and a maximum length of 20
cm (see Fig.3).

Fig.3 European anchovy, Engraulis encrasicolus (Frose & Pauly, 2000)
The european anchovy presents a short life span (three to four years), seasonal migrations, an
intense growth and an age of first maturity with one year (r selection strategy, i.e. selection
for maximum rate of population reproduction in a fluctuating environment). It spawns from
April to November with peaks usually in the warmest months. In the Tagus estuary the
European anchovy spawns one time in the April/May period (R, 1986). Eggs are ellipsoidal
to oval, floating in the upper 50 m and hatching in 24-65 hrs, presenting a biomass of 0,226
mg wet weight (R, 1986). Since the eggs are pelagic, their shape prevents their transport
through the estuary, because it offers less resistance to the currents. This fish species is a
multiple spawner (often more than ten batches within a season). This mechanism is an
efficient means of ensuring complete use of food and increases the probability of population
survival, as well as being an adaptation for unstable environmental conditions and high
mortality in early stages of ontogeny.
Survival of anchovy larvae is determined by the food availability in a period of omnivorous
feeding, and there is a close synchrony between key larval abundance and the zooplankton
population maximum. The anchovy feeds mainly on copepod nauplii larvae, mollusc larvae,
phytoplankton, copepods and cirripeds (Berdnikov et al, 1999).
The european anchovy is a highly commercial species, marketed fresh, dried, smoked,
canned, frozen, made into fish meal and used as bait (Frimodt, 1995). As a target fisheries
species, the anchovy has received intensive study (Winkler, 2000; Berdnikov, S.V. et al,
1999; Costa, 1999; Santos et al, 1997; Tudela & Palomera, 1997; Magoulas et al, 1996;
8
Myers et al, 1995; Koranteng, 1993; Morales-Nin & Pertierra, 1990; Karacam & Dzgnes,
1990; Palomera & Lleonart, 1989; Spanakis et al, 1989; Palomera et al, 1988; Erkoyuncu &
Ozdamar, 1989; Uriarte & Adtudillo, 1987).
The anchovy fisheries management of the southwestern Europe and of the Mediterranean Sea
is difficulted by the large interannual variability in the populations abundance. The anchovy
life cycle seems to be adapted to a highly variable environment, presenting both high growth
rate and fecundity and also active migrations to areas that gather better conditions (Palomera
et al, 1988).
The model built in the framework of this thesis is applied to this fish species, since (1) it has
an important role maintaining this ecosystems trophic web, (2) it is a resident species in the
Tagus estuary and (3) because it is highly used in fisheries. The anchovy life cycle is quite
similar to the sardines cycle (R, 1986).
9
2 METHODOLOGY
2.1 State of the Art
Fish simulation models have been developed for assessment on fisheries and environmental
quality management, simulating fish population using individual-based models IBMs
(Lindholm et al, 2001; Cowan Jr et al, 1999, 2000; Cowan Jr et al, 1999; Monvoisin et al,
1999; Rose et al, 1999; DeAngelis et al, 1991).
Spatial modelling of fish with individual behaviour was initiated by Balchens group
(Balchen, 1976 and Reed & Balchen, 1982 in Giske et al, 1998). IBMs are created on the
recognition that individuals differ in their characteristics and abilities and that such
differences may be important in ecology and population dynamics. These models simulate the
fish growth in terms of the biomass variations or using an energetic-based approach
(Levebvre et al 2001; Hanson, 1997). One of the main features of IBMs with regard to spatial
distributions has been the recent development of spatially explicit models (Tyler & Rose,
1994), which incorporate spatial heterogeneity, individual variability and individual
movement. One of the problems with IBMs is that they are little more than a way of
accounting for individuals in a population. If IBMs are to become an important tool for
explaining the behaviour of individuals and populations, models that describe why individuals
are motivated for actions are needed. This means that individual actions should be viewed in
the light of what evolution has found favourable.
It has been ongoing the development of simulation models for fish behaviour (Breitburg et al,
1999; Ficksen et al, 1998; Heath et al, 1998; Sekine et al, 1997; Sekine et al, 1996; Eckman
et al, 1993; Sekine et al, 1991a; Sekine et al, 1991b), including fish movement (McDermot &
Rose, 2000; Sutton et al, 2000; Railsback et al, 1999; Giske et al, 1998), to predict the fish
distribution and the influences of environmental conditions on fish survival (Lindholm et al,
2001; McDermot & Rose, 2000; Lamouroux et al, 1999a; Lamouroux et al, 1999b; Jaworska
et al, 1997).
These studies often include a model of the fish biomass growth, dependent on its processes of
feeding, respiration and excretion. In fact, environmental conditions affect these processes
(directly or indirectly) and so these models consider biological constraints (food quantity and
10
quality, predators of the studied species), systems physical-chemical conditions (bottom
material, water temperature, current velocity, depth, dissolved oxygen, salinity, turbidity,
ammoniacal nitrogen) and human-related stress (fishing, pollution). Predicting the fish
biomass is much more inaccurate compared to predicting the behaviour of a fish. There are so
many factors affecting fish growth that modelling 60% of the fish biomass change is in fact a
good estimate (Sekine, M. personal communication).
Direct Density Dependence
Population processes such as survival, growth, reproduction and movement are said to be
density dependent if their rates change as a function of population abundance. Processes that
limit population growth at high abundances (e.g. slower growth, increased emigration, lower
survival) or increase numerical growth at low population abundances (e.g. faster growth,
increased immigration, higher survival) are examples of direct density dependence or
compensation. Direct density dependent processes operate as a negative feedback and tend to
stabilize population abundance. If processes such as survival and growth decrease at low
population abundances or increase at high abundances, they are referred to as inverse density
dependence or depensation. Inverse density dependent processes operate as a positive
feedback and tend to destabilize populations.
While compensation is a population-level phenomenon, it is important to note that it is
individual responses to abundance changes that cause both compensatory and depensatory
density-dependence (Rose & Cowan, 2000; DeAngelis et al, 1991). Individuals in a fish
population compete with one another for a limited pool of environmental resources vital to
their survival and successful reproduction. Examples of these environmental resources are
food, refuge from predators and suitable sites for spawning (Berdnikov, 1999).
Compensations principle is that the greater the number of fishes in relation to this pool of
vital resources, the more intense the competition. Increased intensity of competition for these
resources may then express itself in decreased survival rates, slower growth rates, reduced
reproduction or increased emigration from the area of competition. The result is that very
dense populations that exceed the available resources decline in numbers until the competitive
pressure is reduced. On the other hand, when population abundance is low in relation to the
available environmental resources, the intensity of competition is minimized. The result is an
11
increase in survival, growth and reproduction and a decrease in emigration, which operate to
increase population abundance.
An increment in an existing mortality, or even a new source of mortality, imposed on a fish
population will at first tend to diminish the numbers of fish; this decrease in numbers of
individuals will cause a population response to compensate the increased mortality. The
magnitude of the compensatory response and the life stages in which it operates vary among
species, depending on their history strategy (survival, growth and reproduction rates), and can
vary among populations due to site-specific conditions (e.g. environmental resources
available, food web arrangement). Rose & Cowan (2000) confirm that the existence of
compensation is found in spawner-recruit data, empirical studies and also in fisheries
management. Berdnikov et al (1999) demonstrated that anchovy recruitment numbers in the
Azov Sea usually greatly exceed survivors from the previous generation, presenting a high
sustainability of the population in face of unfavourable environmental conditions and heavy
fishing.
The question of compensation is important in assessing the effects of anthropogenic stressors
on fish populations (Jaworska et al, 1997). Harvest levels or entrainment and impingement
rates that exceed the compensatory reserve of a population would cause population declines.
The consequences of overestimating compensation can be severe: population decline to a
severely low abundance. To be protective of the resource, people usually resort to the no
compensation hypothesis unless one can demonstrate a clear and unequivocal understanding
of the compensation mechanism (Rose & Cowan, 2000). Rarely there is sufficient empirical
evidence to comprehend how the various life stages and processes combine to cause the
compensatory response of a particular population. Assuming no compensation, or
underestimating compensation, is extremely protective of the natural resources but it lowers
economic efficiency and causes potentially excessive and unnecessary financial costs.
Understanding compensation at the process-level permits to bound the magnitude of
compensation appropriate to use in population models.
Individual-based modelling offers a promising approach for modelling population and
community dynamics (DeAngelis et al, 1991) and has features that should help in quantifying
compensatory responses of fish populations. Representing local interactions in space, size-
based interactions, episodic effects, movement and stochasticity, all of which are important to
12
realistically simulate fish population dynamics and compensation, is relatively easy in IBMs.
Anthropogenic stresses in fish populations can be directly inserted in these models, and then
be used to scale the numbers of individuals loss up to the population level. If one can
realistically represent how individuals grow, survive, reproduce and move, then population-
level phenomenon such as compensation should be obtained by summing over all of the
individuals in the model. Compensation becomes an emergent property of the model, rather
than being imposed in relationships used to construct the model (Rose & Cowan, 2000).

2.2 Conceptual Model
The model developed in the framework of this thesis involves two distinct sub-models: the
individual growth model (IGM) and the population growth model (PGM). The main goal of
this thesis is to build the population growth model, but since there are processes significant at
Fig. 4 . Overview of the Growth Model. Tw is
water temperature; I
S
is ingestion of stage s, E
S
is egestion of stage s, excretion due to VM
S
(vital metabolism of stage s) and AM
S
(active
metabolism of stage s). The IGM produces time
series of W
S
(biomass of the individual fish in
stage s) and W
Sc
is the mean biomass of the
individual fish in the class c of stage s; g
S c
is the
scope for growth of the individual fish in the
class c of stage s, calculated in the PGM.

g
S

Individual Growth
Model
Phy Zoo
Tw
Nutrients
Light
Pina, 2001
Is E
s
VM
s
AM
s

Ws
W
S
c
c
Population Growth
Model
13
the individual fish scale, the IGM is also developed. In the near future, with the PGM it will
be possible to build a model of the fish movement according to the populations
environmental preferences. Field observations indicate that fish inhabit different areas
according with their life stage. Therefore, fish prefer different environmental conditions in
each life stage (Berdnikov et al, 1999) and this should be considered in the model of the fish
movement.
Fig.4 provides a simplified view of the model. The model considers that the individual growth
depends on food availability and water temperature. This model creates series of results of the
individual fishs variation of biomass in time, for each of the stages of its life cycle. These
results are then treated to estimate a mean biomass for each life stage. In fact, as the PGM is a
class transition model, in this phase it is estimated mean biomass values for defined classes of
each life stage. With this procedure is then possible to calculate the scope for growth for these
mean biomass values for each class of the life stages and develop the PGM.

2.3 Individual Growth Model (IGM)
This section explains the basic equations used in the Individual Growth Model (see Fig.5).
The IGM model is based on a biomass balance. This model simulates the biomass growth of
an individual fish, for each of its life stages, considering the fishs metabolism.
Fig.5 Individual Growth Model. For the life stage s, the individual fishs biomass (W
S
) is variated by the
difference between daily production (ingestion, I
S
) and destruction (egestion, E
S
, excretion due to the vital
metabolism, VM
S
and to the active metabolism, AM
S
) and this difference is the individual fishs biomass
variation (? W
S
). Ingestion is a function of water temperature (Tw) and food availability (phytoplankton, Phy
and zooplankton, Zoo) and the excretion (VM and AM) is a funcion of the fishs biomass, food availability and
water temperature.
IGM
W
s

E
s
VM
s
AM
s

? W
s

Fish
I
s

Phy
Zoo
Tw
14
This model uses modified equations of the Shallow-Sea Ecological Model (Sekine et al,
1991a). The biomass growth is described as follows:
? ? AM VM E I
t
W
? ? ? ?
?
?
(1)
where W is the individual fishs biomass (mgC.individual
-1
), I is the amount eaten or ingestion
(mgC.d
-1
), E is the egestion, i.e. the amount of ingested food thats not assimilated (mgC.d
-1
)
and VM+AM is the excretion component, i.e. the losses of biomass due to the fishs
metabolism, where VM is the vital metabolism (mgC.d
-1
) and AM is the active metabolism
(mgC.d
-1
). McDermot & Rose (2000) developed an individual-based model to piscivore
stocking where the growth function also uses these two metabolic losses of fish biomass. The
ingestion is given by:

Ks Wfood
Wfood
. W . ax Im I
Ka Ki
?
? (2)
where Imax is the maximum ingestion rate or maximum growth rate (d
-1
), Ki is a stage-
specific ingestion coefficient, Ka is a stage-specific coefficient that translates the scope of
biomass growth in the specific stage, Wfood is the number of available preys or food amount
(individuals.L
-1
) and Ks is the half-saturation constant (individuals.L
-1
). The maximum
ingestion rate is a function of water temperature:
) Tw . Kc exp( . Kb ax Im ? (3)
where Kb is the maximum ingestion rate at 0 C (d
-1
), Kc is a temperature constant (C
-1
) and
Tw is the water temperature (C).
The amount of available food Wfood, is calculated as follows by:
wp
Z P
wp
Z
Wfood
. ? ?
? ? (4)
where Z is the prey concentration, wp is the weight of one individual prey (mgC.prey
-1
), P is
the concentration of phytoplankton in the environment (mgC.L
-1
), Z is the concentration of
15
zooplankton in the environment (mgC.L
-1
) and ? is a coefficient of zooplankton intake by the
individual fish, dependent on the fish size.
Destruction processes presented in equation (1) are given by:
E = I . (1 - K
E
) (5)
VM = W . K
VM
(6)
AM = I . K
AM
(7)
where K
E
is the net assimilation efficiency, K
VM
is the vital metabolism biomass consumption
rate (d
-1
) and K
AM
is the active metabolism excretion rate.
All the equations are developed for the larvae, juvenile and adult stages. In the egg stage,
there is no ingestion and egestion, vital metabolism and active metabolism are considered
null, resulting in no growth. Each of the life cycles stages (egg, larvae, juvenile and adult)
presents different characteristics in the physiological processes and so this model
contemplates those differences by considering stage-specific coefficients values for each of
the four stages of the fishs life cycle.

2.3.1 Interaction between fish and habitat
The fish interacts with habitat by feeding and excreting. In this model, phytoplankton and
zooplankton constitute de fish food (or prey), and it is considered that there is always enough
food in the system for this single fish. This model assumes that there is a dependence between
the prey size and the fish size, considering that the fish can eat phytoplankton since the
beginning of the larval stage, but the amount of zooplankton intake increases with the fish
size and hence the fishs age, if one considers that the fish has a growth where higher biomass
is equivalent to more age.
In the IGM model, the destruction processes are egestion (E) and excretion due to vital and
active metabolisms (VM an AM). Both egestion and active metabolism are simulated as partial
losses of the ingested prey biomass. The ammonia production may be predicted from the vital
metabolism consumption of fish biomass. Organic matter inputs to the environment may be
16
calculated from the non-assimilated food (egestion) and the active metabolism consumption
of biomass.

2.4 Population Growth Model
The population dynamics is simulated through a modification of a class transition model
(Ferreira et al, 1998 in Simas et al, 2001; Berdnikov et al, 1999). This model considers the
fishs four stages of its life cycle: eggs, larvae, juveniles and adults, taking into account the
individual stage-specific biological processes. In each stage there are several classes and so
the model considers:

The transition of the number of individuals of the fish population between classes in each of
the stages of the life cycle is given by:

? ? ? ?
C C
C C C
S S
S S S
n . m
s
g . t n
t
n
?
?
?
?
?
?
(8)
where n is the number of individuals of stage s and class c, g is the scope for growth of stage s
and class c (d
-1
), m is the mortality rate of stage s and class c (d
-1
) and t is time (d). Therefore,
Fig. 6. Population Growth Models classes for each stage (one class for the eggs, four classes for larvae stage,
L, three classes for the juvenile stage, J and four classes in the adult stage, A, being YOY the Young-of-the-Year
individuals), with characteristic mortality for each class, m and transition of stages rates (EtoL, LtoJ and
JtoYOY, transition of eggs to larvae, larvae to juveniles and juveniles to adults young-of-the-year,
respectively).

Eggs
L1 ? L2 ? L3 ? L4
J3 ? J2 ? J1 A3 ? A2 ? A1 ? YOY
meggs mL1 mL2 mL3 mL4
mA3 mA2 mA1 mAyoy mJ3 mJ2 mJ1
EtoL
L
t
o
J

JtoYOY
S
p
a
w
n
i
n
g

PGM
17
the number of individuals in each class depends on the individual scope for growth and on a
natural mortality rate. Fig.7 illustrates how the number of individuals varies along the life
stage according to Equation (8). This model considers only the natural mortality rate, but in
the future it will develop to consider human pressure on the population (e.g. fishing,
pollution).

The scope for growth, g is described as:

? ?
C C
S S S S
S
ws ws
AM VM E I
g
C C C
C
?
? ? ?
?
?1
(9)
where I
S
, is ingestion (mgC.d
-1
), E
S
is egestion (mgC.d
-1
), VM
S
is vital metabolism (mgC.d
-1
)
and AM
S
is active metabolism (mgC.d
-1
) in the class c of stage s; ws
c
and ws
c+1
are the mean
weights of the individual fish (mgC) in the class c and c+1, respectively. Mean weights of
each class are estimated from the time series of the biomass results simulated by IGM and
ingestion, vital metabolism and active metabolism are calculated as in IGM but using mean
biomass values in each class.
The transition between stages is simulated considering that it is required a period of time for
the development of the fish in each stage. Fish need a period of time in the stage of larvae
before being juveniles and so on. Nevertheless, as this is a population growth model, fish
n
s
c

eggs
larvae

juveniles

adults

n
eggs
n
larvae
n
juvenile
n
adults
Fig.7 Variation of the number of individuals per stage of the fish life cycle. In each stage there is a mean
number of individuals. One of the reasons for considering several classes per stage is that it greatly diminishes
the error in the estimate of these numbers of individuals.
18
arent all in the same stage in a given moment in time, existing eggs, larvae, juveniles and
adults at the same moment and in each stage there are individuals more developed than others.
Therefore it is reasonable to consider a transition rate (stos+1), as follows:

s
f
t
s
1 s to s ? ? (10)
with s
f
being the last class of stage s (eggs, larvae or juveniles; never the stage adults, since
this is the last stage of the life cycle) and t
s
the characteristic duration of the specific stage
(t
eggs
=2 d, t
larvae
=31 d, t
juveniles
=330 d). As seen in Fig.6 there is a transition of eggs to larvae,
EtoL, larvae to juveniles, LtoJ and juveniles to adults, JtoYOY.
Spawning is a function of the number of individuals in each class of the adult stage, n
Ac
:
young-of-the-year, A1year, A2years and A3years (number of adults) and of each classs
effective spawners (not all fish are able to spawn), provided there is a specific batch size for
each fish species. Effective spawners of class c is the percentage of fish in the class c that are
actually able to spawn (the ability to spawn depends on the fishs sex and reproductive
maturity):

?
?
c A c E
n S BS Spawning . . (11)
where BS is the batch size (eggs.individual
-1
), S
Ec
is the number of effective spawners in the
class c and n
Ac
is the number of individuals in class c of the adult stage. In the PGM spawning
period occurs in April/May and hence the spawning is of 1 time per year.
Mortality fluctuates along the life cycle. In fact, eggs and larvae present the highest natural
mortalities (Lindholm et al, 2001; Jaworska et al, 1997; DeAngelis et al, 1991). In this model
each class of each stage presents a characteristic mortality, so that in the overall population
there is a mortality rate curve as described by Fig. 10.
2.4.1 Interaction between fish population and habitat
The fish population interacts with habitat by feeding and excreting. Concerning the dynamics
of the fish population predation over the phytoplankton and zooplankton in the ecosystem this
19
model considers variations of the prey concentration throughout the year and how it affects
the populations growth. In fact, this influence is included in the individual growth model,
that produces the results from which it is then calculated each stages mean biomass and also
in the population growth model, in the ingestion and destruction processes calculus. This
model assumes that the fish population is not able to reduce extensively the phytoplankton
and zooplanktons concentrations in the system and therefore the starvation-induced mortality
is not measured.
The organic matter and ammonia production by the fish population is easily estimated from
the PGM results for the egestion and active metabolism (organic matter) and vital metabolism
(ammonia) partial consumption of fishs biomass, multiplied by the number of individuals in
the population. As this model will be integrated in the MOHID system (for more information
see section A.1) in the near future, it is important to include the fish populations input of
ammonia and organic matter in the environment.

2.5 Fish movement
Movement rules are a critical component of spatially explicit IBMs because movement is an
essential process used by fish to adapt to changing environment and competitive conditions.
In fact, fish move to improve food intake and growth, to reduce vulnerability to predation
risks that vary with fish size and environmental conditions, to seek shelter during high flows
and avoid stranding as flows decrease, and to avoid competition with dominant fish. Models
must be able to simulate these movements correctly if they are to be useful for predicting
population responses to changes in habitat, mortality rates or competition (Railsback et al,
1999). Furthermore, verifying that movement rules produce realistic movement decisions is
one of the model evaluation steps needed for IBMs to be accepted for decision-making and
research.
Railsback et al (1999) studied fitness measures and movement rules, in stream fish IBMs.
Spatially explicit IBMs use movement rules to determine when an animal departs its current
location and also to determine its movement destination. These rules generally compare
locations using some measures of a fishs expected fitness and assume that fish make
20
movement decisions at least in part to increase their fitness. Movement rules are critical to
realistic simulation of how individuals, and therefore populations, respond to changes in
habitat and population density. Recent research shows that many fish move quickly in
response to changes in physical and biological conditions, so movement rules should allow
fish to rapidly select the best location that is accessible. Regarding movement rules, Railsback
et al (1999) concluded that (1) IBMs should assume that the fish are familiar with their
surroundings over an area greater than their current location (models should let fish move any
time that better habitat is available within the area they are familiar with) and (2) models
should let fish select the best habitat available instead of moving randomly. If routine non-
exploratory movement is simulated as partially random, it should occur over a considerably
smaller time scale than habitat variation.
Fitness measures are used in movement rules to rate alternative destinations. The theory that
minimizing the ratio of mortality risk to food intake maximizes a fishs fitness is not
applicable to typical IBM movement decisions and can cause serious errors in common
situations. Instead, Railsback et al (1999) developed fitness measures from unified foraging
theory that are theoretically compatible with IBMs. This fitness causes fish to select habitat
that maximizes its expected probability of survival over a specified time horizon, considering
both starvation and other risks. This fitness measure is dependent on the fishs current state,
making fish with low energy reserves more willing to accept risks in exchange of higher food
intake. Another measure represents the expectation of reaching reproductive maturity by
multiplying expected survival by a factor indicating how close to the size of first reproduction
the fish grows within the time horizon.
Sekine et al (1991a, 1991b, 1997) studied fish movements according to their environmental
factors. Using an object-oriented programming language, they developed the Shallow-Sea
Ecological Model (SSEM), which simulated fish movement according to their preference for
environmental conditions as well as competition among species. The movement component
considered advection and diffusion and also the movement by swimming. The swimming of
living organisms was caused by their preferences and the criteria to decide the movement
were food amount, depth, water temperature and the nature of the sea bottom.
Later, Sekine et al (1997) developed a model of fish distribution according to the fish
preference for environmental factors such as water temperature, depth, cover, shade, current
21
velocity, turbidity, food amount, biomass, stem of aquatic plants and composite conditions. In
this model the parameter values for the environmental preference and the weight values of the
environmental factors could be introduced without calibrating the preference parameters
again. This models calibration depended on experimental simulations. In these experiments,
a determined number of fish was placed in tanks and several environmental conditions were
tested to measure fish preferences (in terms of their distribution in the tank).
Lindholm et al (2001) modelled fish movement according to habitat complexity, using a
simple population model with a modified Beverton-Holt equation to simulate the effects of
habitat-mediated processes on population responses at the landscape scale. It considers a
population subdivided into n regions, where each region is represented by a particular level of
habitat complexity. Habitat complexity is interpreted as a measure of the vertical relief and
cover from predation provided by a particular substratum (e.g. pebble-coble, sand, boulder)
and any associated emergent epifauna (e.g. sponges, amphipod tubes, cerianthid anemones,
shell deposits, biogenic depressions). It also considers that habitat complexity is uniform
within any region and that each region is either a marine protected area in which no fishing
disturbance occurs or that i t is outside a marine protected area and altered by fishing activity.
In Lindholm et al model net per-month movement between regions is a function of the
percentages of individuals moving between regions and it is specified that movement rates are
the inverse of a habitat complexity score. Lindholm et al (2001) concluded, among other
things, that movement rates were critical for predicting the effects of marine protected-area
size on fish survival.

22
3 MODEL IMPLEMENTATION
3.1 Individual Growth Model
The conceptual model of the Individual Growth (IGM) uses the Tagus estuary as the case
study and Engraulis encrasicolus as the studied fish species. The model was implemented
using the Powersim
TM
software. A first sensitivity analysis is done, considering the
environmental variables (phytoplankton, zooplankton and temperature) constants in space and
time.
This model uses temporal series of phytoplankton and zooplankton calculated by Pina (2001)
for a central area in the Tagus estuary and these results are in good agreement with field
measurements. Pinas results indicated that the daily mean concentrations of phytoplankton
and zooplankton were described throughout the year as shown in the curves of Fig.8.
Fig. 8. Daily mean concentration (mgC.L
-1
) of phytoplankton and zooplankton for the year 1999 in a central box
of the Tagus estuary (Pina, 2001).

The IGM simulates the growth of the individual fish in each of its life stages. The parameters
used in the individual growth models are presented in Table.1, resulting from the calibration
of the model after introducing some typical values presented by Sekine (1991b) and Gypens
et al (2001).
0,0
0,2
0,4
0,6
0,8
1,0
1,2
0 30 60 90 120 150 180 210 240 270 300 330 360
time (days)
m
g
C
.
L
-
1
phyto
zoo
23
Table .1 Parameter values used in the IGM.
Name Units Meaning Larvae Juvenile Adult
Ki - Ingestion coefficient 0,3 -0,25 -0,25
Ka - scope of biomass growth in the specific stage 0,7 0,68 0,64
Kb d
-1
maximum ingestion rate at 0 C 0,1210 0,0339 0,0146
Kc C
-1
temperature constant 0,0693 0,0693 0,0693
Tw C water temperature 18 18 18
?
-
coefficient of zooplankton intake by the
individual fish
0,5 1 1,5
Wp mgC. prey
-1
weight of one individual prey 0,01 0,01 0,01
Ks individuals.L
-1
half-saturation constant 4 4 3
K
E
- net assimilation efficiency 0,7 0,8 0,8
K
AM
- active metabolism excretion rate 0,2 0,1 0,2
K
VM
d
-1
vital metabolism biomass consumption rate 0,1780 0,0670 0,037

The conceptual model of the Population Growth (PGM) uses the Tagus estuary as the case
study and Engraulis encrasicolus as the studied fish species. The model was implemented
using the Powersim
TM
software. First, the PGM was tested with several simulations with
different initial conditions for the number of individuals in each class of the fish population
and no irregularities were found in the results. For that reason it is chosen that the population
initial number of individuals is of 150.000 eggs (the same number of eggs as in an individual
batch; this value is confirmed by Melo, 1994).
In the population growth model, individual mean weights necessary for the scope for growth
estimate in each class of each life stage, are given by the IGM results (see Fig.9b for
individual mean weights the mean values are calculated upon the biomass curve throughout
the life cycle illustrated in Fig.9a). Ingestion and destruction processes are estimated as in the
IGM (using the values presented in Table.1), but using the mean biomass values.
The PGM uses a characteristic mortality rate for each class. The PGM simulations consider
the curve of Fig.10 and plus two different situations of higher mortality rates in the adult stage
(see Table.2, scenarios 3 and 4). In the first few days of life, fish face an intense mortality,
either because they are driven by the currents and therefore may end up in places too adverse
(in the stage of egg and the L
1
and L
2
classes of the larvae stage), either because they are too
young and dont know how to find food or how to escape from predators (larvae stage). In the
24
egg stage only 10% of the eggs survive and in the larvae stage there is a total mortality of
59% in one-month period. According to R (1986) in the first period of life fish may present
60-100% mortality rates. At the end of their life cycle, fish present a higher mortality rate
because they present a higher risk of dying of age and/or diseases (in the class A
3
).

Fig. 9 a) IGM results for a simulation of the anchovys biomass growth in the Tagus estuary for 4 years,
providing the individual fish biomass (mgC). From these results it is possible to calculate the b) mean individual
weight of an individual fish in each class of the stages of larvae, juvenile and adult used in the PGM. Comparing
curves a) and b) it is concluded that b) represents the biomass mean values throughout all the fishs life stages.

a)
0
200
400
600
800
1000
1200
0 360 720 1080 1440
time (days)
m
g
C
b)
0
200
400
600
800
1000
1200
L1 L2 L3 L4 J1 J2 J3 Ayoy A1 A2 A3
Class
m
g
C
25
Four different scenarios are developed to test the PGM model (see Table.2). Scenario 1 is an
ideal situation without environmental or anthropogenic stresses; scenario 2 is a situation with
a less reproductive population (due to environmental stress), where the number of spawners in
the adults stage is less significant; scenario 3 is a situation considering anthropogenic stresses
(e.g. fishing) over all the classes of the adult stage, represented by a mortality 10 times higher
than in scenario 1; scenario 4 is a modification of scenario 3, considering fishing stress over
the adult classes except for the young-of-the-year individuals.

Table.2 PGM parameters values and state variables initial values. These values result from the calibration of the
model after introducing some typical values presented by Melo (1994), R (1986) and Henrique Cabral
(pers.com).
Scenario
Name Units Meaning
1 2 3 4
BS eggs.individual
-1
batch size (per individual fish) 150000 150000 150000 150000
S
Eyoy
- percentage of spawners in class A
YOY
0,2 0,1 0,2 0,2
S
E A1
1
0,4 0,2 0,4 0,4
S
E A2
2
0,3 0,2 0,3 0,3
S
E A3
3
0,3 0,2 0,3 0,3
m
E
d
-1
characteristic mortality in class Eggs 0,9 0,9 0,9 0,9
m
L1
d
-1
characteristic mortality in class L
1
0,4 0,4 0,4 0,4
m
L2
d
-1
2
0,1 0,1 0,1 0,1
m
L3
d
-1
characteris tic mortality in class L
3
0,08 0,08 0,08 0,08
m
L4
d
-1
4
0,01 0,01 0,01 0,01
m
J1
d
-1
characteristic mortality in class J
1
0,008 0,008 0,008 0,008
m
J2
d
-1
2
0,008 0,008 0,008 0,008
m
J3
d
-1
3
0,008 0,008 0,008 0,008
m
Ayoy
d
-1
characteristic mortality in class A
yoy
0,001 0,001 0,01 0,001
m
A1
d
-1
1
0,001 0,001 0,01 0,01
m
A2
d
-1
2
0,001 0,001 0,01 0,01
m
A3
d
-1
3
0,005 0,005 0,05 0,05
Fig. 10. Mortality rates used in the
PGM for each class.

0,0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1,0
Eggs L1 L2 L3 L4 J1 J2 J3 Ayoy A1 A2 A3
Class
d
a
y

-
1
26
4 RESULTS AND DISCUSSION
4.1 Individual Growth Model
The IGM was calibrated and simulations starting in April with the parameters values
presented in Table.1 provide the results discussed in this section.
This model produces the individual fishs biomass throughout its life cycle (Fig.9a) in
accordance with the expected. Before the adult stage (first 360 days) fish present an
exponential growth - a considerable part of the ingested food allows the increase of the
individuals biomass and not only the individuals survival. In addition, this exponential
growth presents scope variations due to the seasonal variations in food availability. In the
adult stage (after 360 days) the ingested food tends to be completely consumed by the fishs
vital and active metabolisms and therefore the oscillations in the biomass result from seasonal
variations in food availability. There is the exception of the young-of-the-year adults (the first
semester of the second year (i.e. days 360 to 540) that present an extensive growth, which is
in accordance to what it is usually observed in the field.
Considering that carbon biomass (mgC) represents 40% of the fishs dry weight (mg dw) and
that fishs dry weight is 10% of its wet weight (mg ww), multiplying the IGM individuals
biomass results by a factor of 25 provides the fishs wet weight. Therefore, according to the
IGM results (see Fig.9a), an adult fish weights 27,4 g ww (1095 mgC), which is in agreement
with field observations (25 to 30 g ww - Dr. Henrique Cabral pers. com.).
Concerning the larval growth the model considers the larvae metabolism as shown in Fig.11.
The area between the IngestionL and DestructionL curves gives the larvae biomass growth.
Destruction processes are discriminated with rectangular unfilled marks (egestion, active
metabolism and vital metabolism). The vital metabolism consume of biomass is the highest,
after half of the larval stages period. However, in the first 17 days, egestion plays the main
role of the destruction processes.

27
0
0,4
0,8
1,2
1,6
2
2,4
2,8
0 5 10 15 20 25 30 35
IngestionL
DestructionL
EgestionL
ActMetabolismL
VitMetabolismL

Fig.11 Larvae Metabolism.
In the juvenile stage there is an important biomass growth - at the end of this stage the fishs
biomass is about 80 times higher than at the end of the larval stage (see Fig.9a,b). As this
stage presents a characteristic time higher than that of the larvae, seasonal variations in
phytoplankton and zooplankton are relevant as observed in Fig.12 - there is a period during
winter when ingested food is not enough to supply the fishs destruction processes and
therefore forcing a consumption of the fishs biomass.

Fig.12 Juvenile metabolism.
As explained earlier, the adult metabolism is somewhat different of that observed in the
previous stages (see Fig.13). In the first six months of the adult stage (the Young-Of-the-Year
class in the population growth model) there is a significant biomass growth when the fish
attains its first maturity age. At the same time there is an important increase in the
0
10
20
30
40
50
60
30 60 90 120 150 180 210 240 270 300 330 360
time (days)
m
g
C
.
i
n
d
i
v
i
d
u
a
l
-
1
IngestionJ
DestructionJ
EgestionJ
ActMetabolismJ
VitMetabolismJ
28
individuals vital metabolisms consumption of biomass, thus preventing further growth with
the same scope.
The adult stage lasts for about three years and consequently the winter collapses of food
availability is very evident (compare Fig.13 with Fig.14). In these periods, ingestion decreases
to a point where destruction processes are higher, which causes the individuals biomass to
decrease in those periods (as observed in Fig.9a).

Fig.13 Adult metabolism.

Fig.14 Available food for larvae (FoodL), juvenile (FoodJ) and adult (FoodA).

0
0,2
0,4
0,6
0,8
1
1,2
1,4
1,6
0 180 360 540 720 900 1080 1260 1440
time (days)
m
g
C
.
L
-
1
FoodL
FoodJ
FoodA
0
10
20
30
40
50
60
70
80
360 540 720 900 1080 1260 1440
time (days)
m
g
C
.
i
n
d
i
v
i
d
u
a
l
-
1
IngestionA
DestructionA
EgestionA
ActMetabolismA
VitMetabolismA
29
Running the IGM model with the parameters values presented in Table.1 but using a series of
food amount in the environment of five times smaller of that shown by Fig.8 (daily mean
concentrations of phytoplankton and zooplankton) causes a smaller individual biomass
growth, as illustrated in Fig.15, which is in agreement with what it would be expected. Here it
is clear the model constrain of the fishs growth with the available food amount: less food
availability causes lower growth.

Fig. 15 Individual fish biomass
(mgC) - IGM results for a
simulation of the anchovys
biomass growth in the Tagus
estuary for 4 years, using series
of food amount five times
smaller than that shown by
Fig.8.

Another environmental factor considered by the IGM model is the water temperature (Tw),
also of great importance to the fishs growth. The european anchovy withstands temperatures
ranging from 6 to 29C. Comparing the IGM results of the individual biomass growth
considering a Tw of 10C (Fig.16), with the results obtained with a Tw of 18C (Fig.9a), and
all the other parameters remaining the same (Table.1), it is evident that when the water
temperature gets closer to the limit, the fishs growth is lower.

Fig. 16 Individual fish biomass
(mgC) - IGM results for a
simulation of the anchovys
biomass growth in the Tagus
estuary for 4 years, with water
temperature of 10C.
0
100
200
300
400
500
600
700
800
900
0 360 720 1080 1440
time (days)
m
g
C
0
200
400
600
800
1000
1200
0 360 720 1080 1440
time (days)
m
g
C
30
The PGM was simulated for five years to develop different scenarios, starting with a
population of 150.000 eggs (the equivalent to one individual batch Melo, 1994 -, not
representative of real life systems, but it facilitates the interpretation of the results). In the
PGM, ingestion and destruction processes are considered as in the IGM, but using the mean
biomass values calculated from the individual biomass IGM results (see Fig.9b). The defined
characteristic mortality rates (see Fig.10) are a key element in the model, as it would be
expected.

4.2.1 Life Cycle in the Fish Population
There is an input of 150.000 eggs and the number of eggs rapidly decreases to zero because
mortality in this stage is very high (m
E
=0,9) and because the characteristic time of this stage is
of 2 days the survivors shift to the larvae stage (Fig.17a, Fig.18a, Fig.19a, Fig.20a). These
results may be very useful when modelling fish movements according to their environmental
preferences in time scales smaller than 1 year. When modelling fish movement it is possible
to observe the importance of currents variations throughout the year, since fish in the egg
stage move with the water currents. Month-to-month currents variations cause different larvae
settlement and consequently lead to different levels of survival.
In the larvae stage the characteristic mortality rates decrease from L
1
to L
4
(from 0,4 to 0,01
see Fig.10) and therefore the number of individuals in class L
1
is considerably higher than that
of classes L
2
, L
3
and L
4
, as illustrated in Fig.17b, Fig.18b, Fig.19b and Fig.20b.
When moving up to the juveniles and adults stage, the food availability constraint plays an
important role (see Fig.17c,d, Fig.18c,d, Fig.19c,d, Fig.20c,d). In fact, the seasonal variations
in the amount of available food, determined by the phytoplankton and zooplankton
concentrations in the system, force variations in the scope for growth, and thus the number of
fish in each class. From stages J
1
to J
3
the number of individuals decreases due to each classs
characteristic mortality and the shifts to the next class. In the adult stage however, depending
on the characteristic mortality (reference scenario or anthropogenic stress scenario, both
31
presented in the next section 4.2.2) the number of individuals may increase or decrease
(Fig.17d, Fig.18d, Fig.19d, Fig.20d).
4.2.2 Scenarios
In this section are presented the results obtained in four different scenarios (see Table.2), thus
permitting an efficient analysis of the PGM parameters.
Scenario 1 is an ideal situation without environmental or anthropogenic stresses. When
comparing the different scenarios, scenario 1 is the reference scenario.
Scenario 2 is a situation with a less fertile population, considering a less significant number of
spawners per class in the adult stage. This second scenario represents a situation where the
fish population experience environmental stress (other than food availability and water
temperature, already included in the model), consequently declining the reproduction activity,
therefore being less fit (Railsback et al, 1999).
Scenario 3 is a situation considering anthropogenic stresses (e.g. fishing) over all the classes
of the adult stage, represented by an increase in the characteristic mortality 10 times higher
than in the reference scenario.
Scenario 4 is a modification of scenario 3, considering fishing stress over the adult classes
except for the young-of-the-year individuals (to simulate the result of fisheries law restriction
in fishing mesh size). In this fourth scenario it is assumed that the present mesh size limit
imposed by the law is so that it forbids catches of fish in the young-of-the-year adults class.
Although the PGM differentiates individuals in each class using biomass values, there is a
relationship between the fishs biomass and fishs length, hence permitting to compare this
value with the limit imposed by the law. Froese & Pauly (2000) developed a database with a
and b values of the following linear regression:
b
L a W . ? (12)
where W is the fishs biomass (g ww) and L is the fishs length (cm total length). Using
Pereda and Villamor (1991) a and b values (0,001 and 3,448 respectively) for Engraulis
32
encrasicolus, for an adults biomass of 1095 mgC (27,4 g ww), the correspondent total length
is of 19,4 cm. This length value is consistent with field observations (R, 1986; Froese &
Pauly, 2000; Dr. Henrique Cabral, pers. com.).
In the reference scenario (Fig.17), although the population starts with a single batch, its
evolution is quite productive in five years the population increases from zero to 1100 adult
fish (Fig.21a). In this scenario the number of individuals in the fish population varies
according with seasonal variations in food (Fig.8), provided that there is no other stress to the
population.

Fig.17 PGM results for the simulation of Scenario 1 or reference scenario for 5 years (day 0 corresponds to the
beginning of April). a) Number of eggs in the fish population; b) number of fish in each class of the larvae stage;
c) number of fish in the classes of the juvenile stage; d) number of individuals in each class of the adult stage
(see Section 2.4 for more information about these classes).
a)
0
5.000.000
10.000.000
15.000.000
20.000.000
25.000.000
0 360 720 1080 1440 1800
time (days)
b)
0
200.000
400.000
600.000
800.000
1.000.000
1.200.000
1.400.000
0 360 720 1080 1440 1800
time (days)
n
L1 L2 L3 L4
c)
0
1000
2000
3000
4000
5000
6000
0 360 720 1080 1440 1800
time (days)
J1 J2 J3
0
100
200
300
400
500
600
0 360 720 1080 1440 1800
time (days)
YOY A1 A2 A3
d)
33
When considering environmental stress (other than food availability and water temperature),
fish react by decreasing their reproduction activity, due to being less fit (Railsback et al,
1999). This is the situation simulated in scenario 2. The comparison of scenario 2 (Fig.18)
with the reference scenario (Fig.17) puts in evidence the consequences of a lower number of
spawners to the total number of fish in the population. In fact, after five years, the population
presents an increase of 350 adult fish (Fig.21b). Therefore, diminishing the percentage of
spawners in scenario 2 to half or higher than that of scenario 1 (as shown in Table.2) causes
an evolution in the adults population 64 % weaker (almost 2/3 less).
Fig.18 PGM results for the simulation of Scenario 2 or less fit scenario for 5 years (day 0 corresponds to the
beginning of April). a) Number of eggs in the fis h population; b) number of fish in each class of the larvae stage;
c) number of fish in the classes of the juvenile stage; d) number of individuals in each class of the adult stage
(see Section 2.4 for more information about these classes).

Regarding scenarios 1 and 2 (Fig.21a,b), it is clearly identifiable the consequences of a less fit
population (scenario 2) with the same food availability and same water temperature: the
population grows exponentially in both scenarios, but more rapidly if the population is more
a)
0
300.000
600.000
900.000
1.200.000
1.500.000
1.800.000
2.100.000
0 360 720 1080 1440 1800
time (days)
b)
0
20.000
40.000
60.000
80.000
100.000
120.000
0 360 720 1080 1440 1800
time (days)
L1 L2 L3 L4
c)
0
100
200
300
400
500
600
0 360 720 1080 1440 1800
time (days)
J1 J2 J3
0
10
20
30
40
50
60
0 360 720 1080 1440 1800
time (days)
YOY A1 A2 A3 d)
34
fit. Therefore the percentage of spawners parameter S
Ec
is an important parameter in the
PGM model.
If one considers anthropogenic stress on the fish population in the PGM model, the results are
quite different from those presented above, as it would be expected. However this is an
important scenario to assess the consistency of the PGM results, since anthropogenic stresses
are an ever-growing reality in nearly all real aquatic systems.
In scenario 3, anthropogenic stress is incorporated in the PGM model, affecting the four
classes of the adult stage, as explained above. This is a scenario that represents the situation of
an aquatic system where fishing activity takes place with no rules (no law restriction). The
consequence of a higher mortality in the adult stage results in the diminishing of the number
of individuals in each class and after few years, in the disappearance of the fish population
(Fig.19).

Fig.19 PGM results for the simulation of Scenario 3 or anthropogenic stress scenario for 5 years (day 0
corresponds to the beginning of April). a) Number of eggs in the fish population; b) number of fish in each class
of the larvae stage; c) number of fish in the classes of the juvenile stage; d) number of individuals in each class
of the adult stage (see Section 2.4 for more information about these classes).
a)
0
20.000
40.000
60.000
80.000
100.000
120.000
140.000
160.000
0 360 720 1080 1440 1800
time (days)
b)
0
1.000
2.000
3.000
4.000
5.000
6.000
7.000
8.000
0 360 720 1080 1440 1800
time (days)
L1 L2 L3 L4
c)
0
5
10
15
20
25
30
35
40
0 360 720 1080 1440 1800
time (days)
J1 J2 J3
0,0
0,3
0,6
0,9
1,2
1,5
1,8
2,1
2,4
0 360 720 1080 1440 1800
time (days)
YOY A1 A2 A3
d)
35
When changing this third scenario to consider fisheries law restriction in fishing mesh size
resulting in the fourth scenario - only by keeping the young-of-the-year fish mortality as in
the reference scenario, the results do not lead to the extinction of the fish population, but to a
steady population with short growth (Fig.20).

Fig.20 PGM results for the simulation of Scenario 4 or anthropogenic stress with law restriction scenario, for 5
years (day 0 corresponds to the beginning of April). a) Number of eggs in the fish population; b) number of fish
in each class of the larvae stage; c) number of fish in the classes of the juvenile stage; d) number of individuals
in each class of the adult stage (see Section 2.4 for more information about these classes).

Comparing the results of scenarios 3 and 4 permits to evaluate the effect of the fisheries law
restriction. By catching only the fish with 19 cm or bigger - which by Froese & Paulys
(2000) length-weight relationship is a fish with 25 g ww (1000 mgC) the fish population is
saved from extinction (see Fig.21c,d).
a)
0
20.000
40.000
60.000
80.000
100.000
120.000
140.000
160.000
180.000
200.000
0 360 720 1080 1440 1800
time (days)
b)
0
1.000
2.000
3.000
4.000
5.000
6.000
7.000
8.000
9.000
10.000
0 360 720 1080 1440 1800
time (days)
L1 L2 L3 L4
c)
0
5
10
15
20
25
30
35
40
45
0 360 720 1080 1440 1800
time (days)
J1 J2 J3 d)
0,0
0,5
1,0
1,5
2,0
2,5
3,0
3,5
0 360 720 1080 1440 1800
time (days)
YOY A1 A2 A3
36

Fig.21 PGM results for the total number of individuals in the adult stage, in the fish population: a) scenario 1, or
reference scenario, b) scenario 2 or less fit scenario, c) scenario 3 or anthropogenic stress scenario and d)
scenario 4 or anthropogenic stress with law restriction scenario.

The scope for growth function determining the variation of the number of individuals in the
population (see Equation 8) depends on the fishs individual metabolic characteristics (see
Equation 9). Although several simulations were made with different ingestion, egestion, vital
metabolism and active metabolism to test the PGM model, the one presented here is the one
that characterizes the anchovys growth. But it is unmistakable that if one changes the
metabolic rates then the individuals mean biomass in each class would be different and so
would be the scope for growth and consequently the populations growth.
In all the scenarios developed above, scope for growth, g, fluctuates as illustrated in Fig.22 in
each class of each stage.
a)
0
200
400
600
800
1.000
1.200
0 360 720 1080 1440 1800
time (days)
n
b)
0
20
40
60
80
100
120
0 360 720 1080 1440 1800
time (days)
n
c)
0
1
2
3
4
0 360 720 1080 1440 1800
time (days)
n
d)
0
1
2
3
4
5
6
7
0 360 720 1080 1440 1800
time (days)
n
37

The scenarios were developed without changing the scope for growth, since this function
translates the species metabolic characteristics and environmental conditions for food and
water temperature, hence permitting to analyse the parameters of the PGM and not those of
the IGM. The curve of g
function has a mean value for
each class as illustrated by
Fig.23. The mean value of g
increases in the YOY and A
2

classes of the adult stage are
due to the small variation in
the mean individual biomass,
characteristic of this life stage.

Fig.23 Mean g (scope for growth) in each class of each stage of the life cycle.

a)
0,00
0,02
0,04
0,06
0,08
0,10
0,12
0 360 720 1080 1440 1800
time (days)
gL1 gL2 gL3 b)
0,000
0,005
0,010
0,015
0,020
0,025
0 360 720 1080 1440 1800
time (days)
gJ1 gJ2
c)
0,000
0,005
0,010
0,015
0,020
0,025
0,030
0,035
0,040
0 360 720 1080 1440 1800
time (days)
gYOY gA1 gA2
Fig.22 Scope for Growth, g (d
-1
), for each class
of the a) larvae stage, b) juvenile stage and c)
adult stage, considered in the simulations of the
scenarios presented in this section.
0
0,02
0,04
0,06
0,08
0,1
L1 L2 L3 L4 J1 J2 J3 YOY A1 A2
Sc
m
e
a
n

g

(
d
a
y
-
1
)
38
5 CONCLUSIONS
The results obtained by the model developed in this thesis are in conformity with field
measurements to what it concerns biomass values of the individual fish and growth responses
to changes in the environment (the IGM model evaluated water temperature and food
availability).
Results of the individual growth model indicate that the growth of the european anchovy
throughout the four years of the life cycle culminates in an adult biomass of 27,4 g ww (1095
mgC), which is in agreement with field observations. This biomass value corresponds to a
19,4 cm length fish (also consistent with field measurements).
Concerning the IGM parameters, if one considers a five times smaller food amount (Wfood)
in the environment (with all the rest of the IGM parameters remaining the same), the adult
biomass is smaller (17,5 g ww). Comparing this result with the first simulation, it is
concluded from the IGM model that less food availability causes lower growth.
A test was made to analyse the influence of the water temperature (Tw). Decreasing this
parameter from 18 to 10 C, with all the other parameters remaining as in the first simulation
(Table.1) the model result for the adult biomass is of 23 g ww. Comparing this result with the
first simulation it is concluded that when the water temperature gets closer to the limit
supported by the european anchovy, the fishs growth is lower.
Also the results of the population growth model are in agreement with what is expected in
each of the four scenarios presented in section 4.2.2. In all the scenarios the population starts
with a single batch (150.000 eggs) and the eggs mortality was considered constant and equal
to 0,9 (i.e. 10% survival). The mortality rates play a crucial role in the evolution of the
number of fish in the population.
Scenario 1 is an ideal situation without environmental or anthropogenic stresses. In the first
scenario there is a reasonably productive evolution and in five years the population increases
from zero to 1100 adult fish. Scenario 2 considers a less significant number of spawners per
class in the adult stage (S
Ec
), representing a situation where the fish population experience
environmental stress that declines the reproduction activity. When considering environmental
stress fish react by decreasing their reproduction activity, being less fit; in a five years
39
simulation the population presents an increase of 350 adult fish. Comparing scenarios 1 and 2,
it is concluded that a decrease in the percentage of spawners in the population to half or
higher, causes an evolution in the adults population 64 % weaker (almost 2/3 less).
Therefore, the consequences of a less fit population are very clear: the population grows
exponentially in both scenarios, but more rapidly if the population is more fit. Therefore the
S
Ec
parameter is very important in the PGM model.
Scenario 3 considers the existence of anthropogenic stresses (e.g. fishing) over all the classes
of the adult stage, with an increase in the characteristic mortality 10 times higher than in the
first scenario. In this scenario, the results are quite different from those in scenarios 1 and 2. A
higher mortality in the adult stage results in the declining of the number of individuals in each
class and after few years, in the disappearance of the fish population Scenario 4 is a variation
of the third scenario, considering fishing stress over the adult classes except for the young-of-
the-year individuals (to simulate the result of fisheries law restriction in fishing mesh size).
In this fourth scenario the results do not lead to the extinction of the fish population, but to a
steady population with a small growth. The comparison of the results of scenarios 3 and 4
evaluates the effect of the fisheries law restriction. By catching only the fish with 19 cm or
bigger (with 25 g ww or more) the fish population is saved from extinction.
The scope for growth function translates the species metabolic characteristics and responses
to environmental conditions for food and water temperature. If one changes the metabolic
rates then the individuals mean biomass in each class is different and so it is the scope for
growth and consequently the populations growth.

40
6 RECOMMENDATIONS FOR FUTURE WORK
Despite the growth of the individual fish is modelled here as a function of its metabolic
characteristics and environmental conditions water temperature and food availability, the
evolution of the population growth model to include movement rules according to the fishs
environmental preferences is crucial to get closer to the real world. Therefore, the model
developed in this thesis will gain by introducing (1) an individual growth model dependent of
a higher number of relevant environmental variables and (2) a space concept in the population
growth model, not only with advection-diffusion variables, but also with a spatial grid of the
fishs environmental preferences. In this case, the individual growth will be dependent of the
location of the fish in the estuary, since its fitness is not the same throughout the system. One
good example is that of the eggs. Eggs are driven by the water currents, not being able to
chose where to swim. Therefore, depending on the water currents in an estuary, eggs may end
up in the sea (with zero recruitment to the population in the estuary), or in more protected
areas hence permitting them to survive and develop to the larvae stage.
After gathering the knowledge of what as been done so far in the modelling of fish
populations area it is in theory somewhat easy to predict the pathway that this model should
follow in the near future. However, this pathway is dependent on empirical knowledge gained
in field data processing. For instances, to model the fishs environmental preferences its
essential to study the fishs responses to changing environmental variables (as studied by
Breitburg et al, 1999; Cowan Jr et al, 1999; Jaworska et al, 1997; Lindholm et al, 2001;
McDermot & Rose, 2000; Rose & Cowan Jr, 2000; Rose et al, 1999a, 1999b; Sekine et al,
1991a, 1991b, 1996, 1997; Tyler & Rose, 1994).
In the Tagus estuary it is necessary to gather the complete information about each species
location according to the life stage, since fish inhabit different areas of the estuary depending
on the stage of the life cycle. In fact, to be capable to accomplish an efficient management of
this living resources in Portugal, it is necessary an intensive study in all the portuguese
estuaries and coastal zones, since fishing activity is of such a great importance to the
Portuguese economy. Therefore, to a sustainable practice of the fishing activity, the
processing of field data will permit the development of a more close to the reality model that
includes each site characteristics and their influence on the evolution of the fish populations.
41
7 ACKNOWLEDGMENTS
Im very grateful to Prof. Kenneth A. Rose (Coastal Fisheries Institute and Department of
Oceanography and Coastal Sciences from the Louisiana State University, USA) and Prof.
Masahiko Sekine (Department of Civil Engineering from the Yamaguchi University, Japan)
for their expertise and knowledge of the individual-based modelling and modelling of fishs
growth according to their preference for environmental conditions, respectively.
I would like to thank Prof. Ramiro Neves for the idea of this thesis initial theme, for his good
reception in the research team (MARETEC) and also for the reviewing of the final
manuscript. Im grateful to MSc. Pedro Pina (from MARETEC) for the results of his model
and his support for this thesis.
Also of great importance was the knowledge of Prof. Paul Nival (Laboratoire dOcanologie
Biologique et dcologie du Plancton Marin, Station Zoologique de Villefranche-Sur-Mer,
France), taught during the DEA training period in Villefranche-Sur-Mer.
From the New University of Lisbon I would like to thank Prof. Joo Gomes Ferreira, MSc.
Joo Pedro Nunes and MSc. Maria Teresa Simas for their support and ideas in the conception
of the model presented in this thesis.
Im very thankful for Aida Alvera Azcarates support and companionship throughout this
thesis, also for the reviewing and improving of the final thesis presented here.
I also want to thank Dr. Henrique Cabral (Oceanographic Institute of the University of
Lisbon, Portugal) for offering his time to find field data concerning the european anchovys
ecology.
My sincere thanks are also to Dr Gavin Begg (CRC Reef Research Centre of the James Cook
University, Australia), Ph.D. James Lindholm (National Undersea Research Center for the
North Atlantic & Great Lakes from the University of Connecticut, USA), Steve Railsback
(Lang, Railsback & Assoc., California, USA), Prof. Ragnar Arnason (University of Iceland),
Trevor Hutton (CEFAS, UK), Victor Selyutin and Serg Berdnikov (Rostov State University,
Russia) and Anthony R. Wilbur (Massachusetts Coastal Zone Management, Boston, USA).
42
Finally, but not least, I thank my family and friends for their continuous support during all
this DEA at the University of Lige, at the Zoological Station of Villefranche-Sur-Mer and
also during all the process of developing this thesis in Lisbon. A very special thank you for
my parents for making it all possible!
Rita Monteiro
Lisbon, 18
th
of December, 2001

43
8 REFERENCES
1 - Berdnikov, S.V., V.V. Selyutin V.V. Vasilchenko and J.F. Caddy , 1999. A Trophodynamic
Model of the Black and Azov Seas Pelagic Ecosystem: Consequences of the Comb Jelly,
Mnemiopsis leydei, Invasion. Report on the INTAS Project 96-2015. Fisheries Research 42:261-
289.
2 - Breitburg, D.L., K.A. Rose and J.H. Cowan Jr, 1999. Linking water quality to larval survival:
predation mortality of fish larvae in an oxigen-stratified water column. Mar. Ecol. Prog. Ser.
178:39-54.
3 - Cabrita, M.T. and M. T. Moita , 1995. Spatial and Temporal variation of physicochemical
conditions and phutoplankton during a dry year in the Tagus Estuary. Netherlands Journal of
Aquatic Ecology 29:323-332.
4 - Costa, M.J., 1999. O Esturio do Tejo. Ed. Cotovia (Lisboa). 196 p.
5 - Cowan Jr, J.H., K.A. Rose, E.D. Houde, S. Wang and J.Young, 1999. Modelling effects of
increased larval mortality on bay anchovy population dynamics in the mesohaline Chesapeake
Bay: evidence for compensatory reserve. Mar. Ecol. Prog. Ser. 185:133-146.
6 - Cowan Jr, J.H., K.A. Rose and D.R. DeVries, 2000. Is density-dependent growth in young-of-the-
year fishes a question of critical weight? Reviews in Fish Biology and Fisheries 10:61-89.
7 - DeAngelis, D.L., L. Godbout, B.J. Shuter, 1991. An Individual-based approach to predicting
density-dependent dynamics in Smallmouth Bass Populations. Ecological Modelling 57:91-115.
8 - Eckman, J. E., F.E. Werner and T.F. Gross, 1993. Modelling some effects of behavior on larval
settlement in a turbulent boundary layer. Deep-Sea Research II 41(1):185-208.
9 - Erkoyuncu, I. and E. Ozdamar, 1989. Estimation of the age, size and sex composition and growth
parameters of anchovy, Engraulis encrasicolus in the Black Sea. Fisheries Resources. 7:241-247
in Froese, R. and D. Pauly. (eds). 2001. FishBase 2000: concepts, design and data sources.
ICLARM www.fishbase.org, 07 August 2001
10 - Ferreira, J.G., P. Duarte, B. Ball, 1998. Trophic capacity of Calingford Lough for oyster culture
analysis by ecological modelling. Aquatic Ecology 31:361-378.
44
11 - Ferreira , J.G., R. Neves, A.S. Cmara, P. Duarte, P. Vicente, L. Portela, and L. Cancino, 1994.
Modelo Ecolgico do Esturio do Tejo. JNICT. Project PMCT 687-91/MAR.
12 - Ferreira, J.G., T. Simas, and J.P. Nunes, 2001. Organic Matter Cycling Models in Estuaries. in
EUROSSAM Final Report. DGXII Environment Programme. Project ENV4-CT97-0436. 63 p.
13 - Ficksen, , A. C. W. Utne, K. E. Aksnes, J. V. Helvik and S. Sundby, 1998. Modelling the
influence of light, turbulence and ontogeny on ingestion rates in larval cod and herring. Fisheries
Oceanography 7(3/4):355-363.
14 - Frimodt, C., 1995. Multilingual illustrated guide to the world's commercial warmwater fish.
Fishing News Books, Osney Mead, Oxford, England. 215 p.
15 - Froese, R. and D. Pauly. (eds). 2000. FishBase 2000: concepts, design and data sources.
ICLARM www.fishbase.org, 07 August 2001
16 - Giske, J., G. Huse and . Fiksen, 1998. Modelling spatial dynamics of Fish. Reviews in Fish
Biology and Fisheries 8:57-91.
17 - Gypens, N, R. Monteiro, B. Pasquer, S. Casas and M. Mathieu, 2001. Stage de Modlisation des
Rseaux Trophiques Marins. DEA Europen en Modlisation de lEnvironnement Marin.
Observatoire Ocanologique de Villefranche-Sur-Mer. Universit Pierre et Marie Curie, Paris VI.
18 - Hanson, P., T.B. Johnson, D.E. Schindler and J.F. Kitchell, 1997. Fish Bioenergetics 3.0 User
Manual. Center of Limnology, University of Wisconsin Sea Grant Institute, Project A/AS-2.
19 - Heath, M., H. Zenitani, Y. Watanabe, R. Kimura and M. Ishida, 1998. Modelling the dispersal of
larval Japanese sardine, Sardinops melanostictus, by the Kuroshio Current in 1993 and 1994.
Fisheries Oceanography. 7: 3/4, pp 335-346.
20 - Jaworska, J.S., K.A. Rose and L.W. Barnthouse, 1997. General response patterns of fish
populations to stress: an evaluation using an individual-based simulation model. Journal of
Aquatic Ecosystem Stress and Recovery 6:15-31.
21 - Karacam, H. and E. Dzgnes, 1990. Age, growth and meat yield of the European anchovy
(Engraulis encrasicolus) in the Black Sea. Fisheries Resources 9:181-186 in Froese, R. and D.
Pauly. (eds). 2001. FishBase 2000: concepts, design and data sources. ICLARM
www.fishbase.org, 07 August 2001
45
22 - Koranteng, K.A., 1993. Size at first maturity of the anchovy (Engraulis encrasicolus) in
Ghanaian waters and suggestions for appropriate mesh size in its fishery. NAGA, The ICLARM
Q. 16(1):29-30.
23 - Lamouroux, N., H. Capra, M. Poully and Y. Souchon, 1999a. Fish habitat preferences in large
streams of southern France. Freshwater Biology 42:673-687.
24 - Lamouroux, N., J.M. Olivier, H. Persat, M. Poully, Y. Souchon and B. Statzner, 1999b.
Predicting community characteristics from habitat conditions: fluvial fish and hydraulics.
Freshwater Biology 42:275-299.
25 - Lefebvre, S., C. Bacher, A. Meuret and J. Hussenot, 2001. Modelling Nitrogen Cycling in a
Mariculture Ecosystem as a Tool to Evaluate its Outflow. Estuarine, Coastal and Shelf Science
52: 305-325.
26 - Lindholm, J.B., P.J. Auster, M. Ruth and L. Kaufman, 2001. Modelling the effects of Fishing and
Implications for the design of Marine Protected Areas: Juvenile Fish Responses to Variations in
Seafloor Habitat. Conservation Biology 15(2):424-437.
27 - Magoulas, A., N. Tsimenides and E. Zouros., 1996. Mitochondrial DNA phylogeny and the
reconstruction of the population history of a species: the case of the European anchovy
(Engraulis encrasicolus). Mol. Biol. Evol. 13(1):178-190.
28 - Martins, M.C., J.G. Ferreira, T. Calvo and H. Figueiredo, 1984. Nutrientes no esturio do Tejo.
Comparao da situao em caudais mdios e em cheia, com destaque para alteraes na
qualidade da gua. In I Simpsio Luso-Brasileiro de Engenharia Sanitria e Ambiental (Lisboa).
29 - McDermot, D. and K.A. Rose, 2000. An individual-based model of lake fish communities:
application to piscivore stocking in Lake Mendota. Ecological Modelling 125:67-102.
30 - Melo, Y. C., 1994. Spawning frequency of the anchovy Engraulis capensis. S. Afr. J. Mar. Sci.
14:321-331.
31 - Monvoisin, G., 1997. A trophodynamic model of the anchovy (Engraulis encrasicolus) in the
Tagus estuary (Portugal). MSc Thesis. Faculdade de Cincias e Tecnologia, Universidade Nova
de Lisboa. 97 p.
32 - Monvoisin, G., C. Bolito and J.G. Ferreira, 1999. Fish dynamics in coastal food chain:
Simulation and Analysis. Bol. Inst. Esp. Oceanogr. 15:431-440.
46
33 - Morales-Nin, B. and J.P. Pertierra, 1990. Growth rates of the anchovy Engraulis encrasicolus and
the sardine Sardina pilchardus in the Northwestern Mediterranean Sea. Marine Biology 107:
349-356 in Froese, R. and D. Pauly. (eds). 2001. FishBase 2000: concepts, design and data
sources. ICLARM www.fishbase.org, 07 August 2001
34 - Myers, R.A., J. Bridson and N.J. Barrowman, 1995. Summary of worldwide stock and recruitment
data. Can. Tech. Rep. Fish. Aquat. Sci. 2024. 274 p.
35 - Neves, R.J.J. and A.J.P. Santos, 1991. Radiative artificial boundaries in ocean barotropic
models. Proceedings of the 2
nd
International Conference on Computer Modelling in Ocean
Engineering (Barcelona).
36 - Nihoul, J.C.J., 1984. A three-dimensional general marine circulation model in a remote sensing
perspective. Annales Geophysicae 2(4):433-442.
37 - Palomera, I., B. Morales and J. Lleonart, 1988. Larval growth of anchovy, Engraulis encrasicolus
in the Western Mediterranean Sea. Mar. Biol. 99:283-291 in Froese, R. and D. Pauly. (eds). 2001.
FishBase 2000: concepts, design and data sources. ICLARM www.fishbase.org, 07 August 2001
38 - Palomera, I. and J. Lleonart, 1989. Field mortality estimates of anchovy larvae, Engraulis
encrasicolus, in the western Mediterranean. Journal of Fish Biology 35:133-138 in Froese, R.
and D. Pauly. (eds), 2000. FishBase 2000: concepts, design and data sources. ICLARM
www.fishbase.org, 07 August 2001
39 - Pereda, P. and B. Villamor, 1991.Relaciones biometricas en peces de la plataforma Cantabrica.
Inf. Tc. Inst. Esp. Oceanogr. 92. 39 p.
40 - Pina, P., 2001. An Integrated Approach to Study the Tagus Estuary Water Quality. M.Sc Thesis.
Instituto Superior Tcnico, Universidade Tcnica de Lisboa. 69 p.
41 - Railsback, S.F., R.H. Lamberson, B.J. Harvey and W.E. Duffy, 1999. Movement rules for
individual-based models of stream fish. Ecological Modelling 123:73-89.
42 - R, P., 1986. Ecologia da postura e da fase placntnica de Engraulis encrasicolus no esturio do
Tejo. Publ. Inst. Zool. Dr.Augusto Nobre 196:1-45.
43 - Rose, K.A and J.H. Cowan Jr, 2000. Predicting fish population dynamics: compensation and the
importance of site-specific considerations. Environmental Science & Policy 3:S433-S443.
47
44 - Rose, K.A., J.H. Cowan Jr, M. Clark, E.D. Houde and S. Wang, 1999a. An Individual-based
Model of Bay Anvhovy population dynamics in the mesohaline region of Chesapeake Bay.. Mar.
Ecol. Prog. Ser. 185:113-132.
45 - Rose, K.A., E.S. Rutherford, D.S. McDermot, J.L. Forney and E.L. Mills, 1999b. Individual-
based model of Yellow Perch and Walleye Populations in Oneida Lake. Ecological Monographs
69(2):127-154.
46 - Santos, R.S., F.M. Porteiro and J.P. Barreiros, 1997. Marine fishes of the Azores: annotated
checklist and bibliography. Bulletin of the University of Azores. Sup. 1. 244 p.
47 - Sekine, M., H. Nakaniski and M. Ukita, 1991a. A Shallow-Sea Ecological Model using an
Object-Oriented Programming Language. Ecological Modelling 57:221-236.
48 - Sekine, M., H. Nakaniski and M. Ukita, 1996. Study of fish mortality caused by the combined
effects of pesticides and changes in environmental conditions. Ecological Modelling 86:259-264.
49 - Sekine, M., H. Nakaniski, M. Ukita and E. Kanehiro, 1991b. A Shallow-Sea Ecological Model to
Assess the Impact of Coastal Development. 4
th
International Conference on Computing in Civil
and Building Engineering Extended Abstracts, 189 (Tokyo, Japan).
50 - Sekine, M., T. Imai and M. Ukita, 1997. A model of fish distribution in rivers according to their
preference for environmental factors. Ecological Modelling 104:215-230.
51 - Simas, T, J.P. Nunes, J.G. Ferreira, 2001. Effects of Global Change on Coastal Salt Marshes.
Ecological Modelling 139:1-15.
52 - Spanakis, E., N. Tsimenides and E. Zouros., 1989. Genetic differences between populations of
sardine, Sardina pilchardus, and anchovy, Engraulis encrasicolus, in the Aegean and Ionian
seas. Journal of Fish Biology 35:417-437 in Froese, R. and D. Pauly. (eds). 2001. FishBase 2000:
concepts, design and data sources. ICLARM www.fishbase.org, 07 August 2001
53 - Sutton, T.M., K.A. Rose and J.J. Ney, 2000. A Model Analysis of Strategies for Enhancing
Stocking Success of Landlocked Striped Bass Populations. North American Journal of Fisheries
Management 20:841-859.
54 - Tudela, S. and I. Palomera, 1997. Tropic ecology of the European anchovy Engraulis
encrasicolus in the Catalan Sea (northwest Mediterranean). Marine Ecological Progress Series.
160: 121-134.
48
55 - Tyler, J.A. and K.A. Rose, 1994. Individual variability and spatial heterogeneity in fish
population models. Reviews in Fish Biology and Fisheries 4:91-123.
56 - Uriarte, A. and A. Adtudillo, 1987. The anchovy in the Bay of Biscay : new data and analysis of
the fishery 1974-1987. ICES Doc. C.M. 1987/H:20 in Froese, R. and D. Pauly. (eds). 2001.
FishBase 2000: concepts, design and data sources. ICLARM www.fishbase.org, 07 August 2001
57 - Vale, C. and B. Sundby, 1987. Suspended sediment fluctuations in the Tagus estuary on
semidiurnal and fortnightly time scales. Estuarine, Coastal and Shelf Science 27:495-508.
58 - Whitehead, P.J.P., 1990. in J. C. Quero, J. C. Hureau, C. Karrer, A. Post, and L. Saldanha (eds.)
Checklist of the fishes of the eastern tropical Atlantic (CLOFETA). JNICT (Lisbon), SEI (Paris)
and UNESCO (Paris). Vol 1: 228-229 in Froese, R. and D. Pauly. (eds). 2001. FishBase 2000:
concepts, design and data sources. ICLARM www.fishbase.org, 07 August 2001
59 - Whitehead, P.J.P., G.J. Nelson and T. Wongratana, 1988. FAO species catalogue. Vol. 7.
Clupeoid fishes of the world (Suborder Clupeoidei). An annotated and illustrated catalogue of the
herrings, sardines, pilchards, sprats, shads, anchovies and wolf -herrings. Part 2 - Engraulididae.
FAO Fish. Synop. 7(125). 579 p.
60 - Winkler, H.M., K. Skora, R. Repecka, M. Ploks, A. Neelov, L. Urho, A. Gushin and H.
Jespersen, 2000. Checklist and status of fish species in the Baltic Sea. ICES CM. 15 p.

49
A APPENDIX
A.1 MOHID2000
The model developed in the scope of this thesis is to be inserted in the System MOHID in the
near future. The development of this system was initiated in the decade of 80, but its
application to different scientific and technological projects has permitted its successive
perfectionings.
MOHID2000 is a full 3D-baroclinic model and has been developed using an object-oriented
programming philosophy, using the ANSI Fortran 95. The object-oriented programming has
proven to be a very useful methodology in the development of complex programs, especially
for those that simulate real world problems, of which the modelling of marine or estuarine
processes are good examples. The use of the ANSI Fortran 95 guarantees the independence of
system MOHID face to the operative system in which the model is executed (Windows,
Linux, Unix, etc.) and permits an easy implementation of the code in any environment. The
option of running several modules (hydrodynamics, turbulence, drift, etc.) with different time
steps allows an optimisation of the required calculation time for the execution of the
simulations. The system uses the dynamic allocation of the memory, optimising the use of the
computer features and allowing the use of a single executable file.
The model is based on a finite volume concept. Finite volumes technique consists in
applying the physical, chemical and biological laws that conduct the processes intended to
simulate, directly to a control volume as a flux divergence. As a consequence, this method
automatically guarantees the conservation of transported properties.
The system has two main modules: one manages the hydrodynamic properties (ex: velocity,
elevation, water fluxes, turbulent viscosity, etc.) and the other manages the water properties
(e.g. salinity, temperature, density, SPM, nutrients, phytoplankton, coliforms). Both modules
solve the equations in the real domain without any space transformation. The geometry
information is carried in the areas and volumes needed to calculate the fluxes. The cells may
have any initial shape and be subjected to any time deformation, allowing several vertical
discretizations, which can be used simultaneously in different regions of the domain. This
architecture is equivalent to a generic vertical coordinate.
50
The hydrodynamic properties evolution is computed solving the three-dimensional primitive
equations in rectangular coordinates for incompressible flows. Hydrostatic equilibrium is
assumed as well as Boussinesq approximation. The turbulent viscosity can be computed using
several models. In the horizontal plan the options are: constant value or Smagorinsky models.
In the vertical plan the models that can be used are: a constant value, a mixing length model
(Nihoul, 1984), a one-equation (K) model or a two-equations (K-L) model. The water
properties evolution is computed solving the advection-diffusion equation explicitly in the
horizontal and implicitly in the vertical. The sinks are computed forward in time and the
sources are computed backwards to avoid negative values of mass.
The system uses five different types of boundary conditions: free surface, bottom, lateral
closed boundary, moving boundary and lateral opened boundary (Neves & Santos, 1991).
Moving boundaries are closed boundaries with a position that varies in time. For the lateral
and moving boundaries the conditions are always null flux. Any exchange between land and
the sea is computed as a discharge (for example a river or an outfall). The discharge module
can compute the discharges of hydrodynamic properties (e.g. momentum) and also of water
properties (e.g. SPM, salinity). For each of the other three boundary conditions, bottom,
surface and open boundary, there are specific modules. The bottom module is responsible for
the fluxes between the bottom and the water column (e.g. shear friction, erosion/deposition of
sediments). The surface module is responsible for the fluxes between the atmosphere and the
water column (e.g. wind forcing, gas emission, solar radiation).
To avoid chaotic growth of the MOHID2000 it is necessary to implement program techniques
together with a set of rules that ensure reliability and maintainability. The object-oriented
programming is the more powerful technique to achieve these goals. This issue is especially
important for large software systems like the MOHID2000. These systems are usually
developed by several collaborators separated in time and in space and for this reason a model
must be able to incorporate new contributions everyday without any difficulties.
Another very important issue is the input data (pre-processing) of complex models. It is
essential to develop graphical interfaces to help users to give the model basic data in a
systematic way. If the user is not a member of the development team, the probability of
introducing input data errors is somehow high if the input data methodology is for example
Ascii files. Output data (post-processing) analysis is also an important issue especially when
51
ones talking of a 4D numerical tool (3 spatial dimension plus time) that is able to compute
the evolution of almost 30 properties. If the users do not have access to tools that allow them
to see the output data in an intuitive way this task can lead to a non-profitable time
consuming. A graphical interface has been developed to allow the user to explore the data
using 2D images (3D slices) animated in time.
A.1.1 Lagrangean Tracers Module
MOHID2000 Lagrangean Tracers Module can be coupled upstream to a finite-differences
(sigma, double-sigma, Cartesian) 3D hydrodynamic model and downstream to a zero-
dimensional water quality model, an oil dispersion model and a sediment model. This model
has been used in a wide number of study cases: impact of organic effluents in estuarine
environments, hydrodynamic model calibration comparing tracer trajectories with real drifter
trajectories, exchanges between different areas in an estuary, exchanges between the deep
ocean and the continental shelf, primary productivity in the ocean, sediment transport, oil
dispersion, etc.
The variety of applications enumerated shows that a Lagrangean dispersion model is a
privileged tool for coupling physics, biology, chemistry and geology in the marine
environment.
A.1.2 Eulerean Transport Module
This module allows calculation of a property characterised by a soft gradient. It solves
explicitly the advective and diffusive horizontal terms and implicitly the advective and
diffusive vertical terms. The advective term can be solved using one of the following
discretization schemes: centred differences, upwind and QUICK.
This module can be coupled both to two and three-dimensional hydrodynamic modules. In the
three-dimensional baroclinic module it is used to calculate the spatial and temporal evolution
of salinity and temperature since these values are used in determining the density value in
each model point.
52
A.1.3 Cohesive Sediment Transport Module
The simulation of cohesive sediment transport processes is performed solving the 3D-
conservative advection-diffusion equation. For the representation of the processes of
flocculation and erosion and deposition of bottom sediments refined empirical formulations
based on data field was adopted. The models have been calibrated and tested by simulating
tidal flows and suspended sediment transport in different estuaries. Results show a good
agreement between the numerical predictions and corresponding field measurements.
A.1.4 Water Quality Module
This is a non-dimensional model, which allows easy coupling to a transport module, both in
the Lagrangean and Eulerean formulations. This model simulates the main biochemical
processes of phytoplankton, zooplankton, phosphorus, nitrogen and oxygen in the pelagic
zone. The nitrogen cycle species are modelled separately and include its three main inorganic
forms - ammonia, nitrate and nitrite - along with three organic forms of nitrogen - refractory
and non-refractory fractions of dissolved organic nitrogen and particulate nitrogen. The
phosphorus cycle species simulated by this model include inorganic dissolved forms of which
the main group is the orthophosphates (PO43-, HPO42-, and HPO4-), as well as organic
dissolved and particulate phosphorus. In relation to oxygen, calculations are made for
dissolved oxygen and CBO. For each property the model calculates the sink and source term,
due to the main biological and chemical processes occurring in the water column.

ABSTRACT

An effective management of the fisheries resources requires an understanding and the ability to predict
fluctuations in abundance and biomass production. In this work it is developed a growth model for individual
fish and correspondent fish population in a coastal ecosystem, using the european anchovy (Engraulis
encrasicolus) in the Tagus estuary (Lisbon, Portugal). The model involves two distinct sub-models: the
individual growth model (IGM) and the population growth model (PGM). Individual fish biomass varies in each
life stage as a function of its metabolic characteristics, food availability in the environment and water
temperature. In the PGM the number of fish in each class of the four life-stages is calculated taking into account
the individual stage-specific biological processes, percentage of spawners in the population, class-specific
characteristic mortalities and class-specific scope for growth. The IGM simulations give an adult individual
biomass of 27,4 g ww with normal food concentration in the Tagus estuary and a mean water temperature of
18C. Additionally, the biomass value decreases when either the food amount either the water temperature are
decreased. The PGM is simulated in four different scenarios. After a five-year simulation, the total number of
adults in the population is higher in the reference scenario (> less fit/environmental stress > fishing stress
with fishing law restrictions > anthropogenic stress). From the comparison of the scenarios results it is
concluded that: (1) when a population is less fit (due to environmental stress), its growth is exponential as in an
ideal situation, but with a wicker growth; (2) the application of fishing law restrictions that save juveniles and
young-of-the-year adults (<19cm) builds a fish population that tends to stabilization. The evolution of the
population growth model to include movement rules according to the fishs environmental preferences is crucial
to get closer to the real world. To manage the fishing activity in a more sustainable way, it is recommended
gathering and the processing of field data that permits the development of a more close to the reality model that
includes each site characteristics and their influence on the evolution of the fish populations.

RESUM

Une gestion soutenable des ressources de la pche exige la comprhension et la capacit de prvoir des
fluctuations de labondance et production de biomasse. Dans le cadre de cet travail c'est dvelopp un modle de
croissance individuelle dun poisson et aussi de la croissance de la population des poissons correspondants dans
un cosystme ctier, en utilisant l'anchois europen (Engraulis encrasicolus) dans l'estuaire de Tagus (Lisbonne,
Portugal). Le modle contient deux sous-modles distincts: le modle de croissance du poisson individuel (IGM)
et le modle de croissance de la population des poissons (PGM). La biomasse individuelle des poissons change
dans chaque tape de la vie en fonction de ses caractristiques mtaboliques, de la disponibilit de nourriture dans
l'environnement et de la temprature de l'eau. Dans le PGM, le nombre de poissons dans chaque classe des quatre
tapes de vie est calcul en tenant compte des diffrents processus biologiques spcifiques de chaque tape, du
pourcentage des spawners dans la population, des mortalits caractristiques spcifiques de chaque classe et de la
scope for growth spcifique de chaque classe. Les simulations du modle IGM rsultent en une biomasse
individuelle de ladulte de 27,4 g ww avec une concentration normale de nourriture dans l'estuaire de Tagus et
avec une temprature moyenne de l'eau de 18C. En outre, la valeur de biomasse diminue quand la quantit de
nourriture ou la temprature de l'eau diminue. Le modle PGM est simul sur quatre scnarios diffrents. Aprs
cinq ans, le nombre d'adultes totales dans la population est plus lev dans le scnario de rfrence (> moins
adapte /pression environnementale > pression de la pche avec des restrictions de la loi de pche > pression
anthropogne). De la comparaison des rsultats des quatre scnarios on conclut le suivant: (1) quand une
population est moins adapte (en raison de la pression environnementale), sa croissance est exponentielle comme
dans une situation idale, mais plus faible; (2) l'application des restrictions de la loi de pche qu'conomise des
juvniles et des adultes young-of-the-year (< 19cm) construit une population des poissons qui tend la
stabilisation. L'volution du modle de croissance de la population pour inclure des rgles de mouvement selon les
prfrences environnementales des poissons est cruciale pour devenir plus prs du monde rel. Pour faire la
gestion de l'activit de pche d'une faon plus soutenable, il est recommand dassembler et de faire le traitement
des donnes du terrain qui permettent le dveloppement dun modle plus prs de la ralit, qui inclut des
caractristiques de chaque endroit et leur influence sur l'volution des populations de poissons.

Contacts:
Rita Monteiro
Rua Amrico Duro, 16, 4B, 1900 Lisboa, Portugal.
ritmonteiro@yahoo.com

Fish Growth Modelling

Uploaded by

Copyright:

Available Formats

Fish Growth Modelling

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fish Growth Modelling

Uploaded by

Copyright:

Available Formats

Universidad

You might also like