an organic matter and nitrogen dynamics model for the ecological

7/25/2019 An Organic Matter and Nitrogen Dynamics Model for the Ecological

1/12

Environmental Modelling & Software 17 (2002) 571582

www.elsevier.com/locate/envsoft

An organic matter and nitrogen dynamics model for the ecologicalanalysis of integrated aquaculture/agriculture systems:

I. model development and calibration

D.M. Jamu 1, R.H. Piedrahita

Department of Biological and Agricultural Engineering, University of California, One Shields Avenue, Davis, CA 95616, USA

Received 25 July 2000; received in revised form 20 September 2001; accepted 11 January 2002

Abstract

A dynamic and mechanistic mass balance model (Integrated Aquaculture/Agriculture System model, or IAAS) for predictingnitrogen and organic matter outputs from aquaculture ponds and their subsequent recycling in conventional agriculture practiceshas been developed and calibrated using data from Honduras, Thailand and Malawi. The model, developed using Stella modeling

software (High Performance Systems, Inc. Hanover, New Hampshire), simulates individual fish growth, organic matter, nitrogen(organic, total ammonia and nitrate), dissolved oxygen and phytoplankton, crop growth, soil nitrogen concentrations and soil waterbalance. Processes included in the model are fish growth, crop biomass growth, allochthonous and autochthonous organic matterproduction, organic matter decomposition, nitrogen input, nitrogen mineralization, nitrification, denitrification, diffusion, uptake andleaching. The model has been calibrated using literature and observed parameter values from experiments. The calibration procedureinvolves running the model using inputs from observed data, comparing the model output to observed data, and making appropriateadjustments to parameter values until a general fit between the model and observed values is observed. The structure of the modelallows users to modify parameter values to suit different simulation scenarios via a user interface display that also includes graphs

and tables for model output.

2002 Elsevier Science Ltd. All rights reserved.

Keywords: Nitrogen cycling; Organic matter; STELLA; Aquaculture; Dynamic model

Software availability

Name of software: IAASHardware requirements: Laptop or desktop IBM-com-

patible or Macintosh PC. Microsoft Windows3.1 or Windows/95/98 or NT with minimum2Mb RAM and 1 Mb of hard disk space forIBM compatible PC and 68020 processor or bet-ter or System 7 or later for Macintosh computers(requirements depend on Stella version used)

Software requirements: Stella Research or Stellasave-disabled demo available athttp://www.hps-inc.com

Contact address: Raul H. Piedrahita, Biological and

Corresponding author. Tel.: +1-530-752-2780; fax: +1-530-752-

2640.

E-mail addresses: [email protected] (R.H. Piedrahita);

[email protected] (D.M. Jamu).1 Current address: ICLARMThe World Fish Center, P.O. Box

229, Zomba, Malawi.

1364-8152/02/$ - see front matter 2002 Elsevier Science Ltd. All rights reserved.

PII: S1 3 6 4 - 8 1 5 2 ( 0 2 ) 0 0 0 1 6 - 6

Agricultural Engineering Department, One Shi-elds Avenue, University of California, DavisCA 95616-5294 USA

E-mail: [email protected]: Free, available upon request

1. Introduction

Integrated aquaculture/agriculture systems are increas-ingly being promoted as an environmentally sustainablemethod for producing aquatic and terrestrial crops. Inintegrated systems, wastes from one system componentare recycled as inputs to another system component.Through waste recycling, integratedaquaculture/agriculture systems can be used to treataquaculture effluents, increase farm productivity throughefficient resource utilization, spread financial riskthrough diversification and reduce system nutrient losses(Singh et al., 1996; Williams, 1997).


2/12

572 D.M. Jamu, R.H. Piedrahita / Environmental Modelling & Software 17 (2002) 571582

To take advantage of the attributes of the integrated

aquaculture/agriculture system stated above, adequate

scientific information is required on the important pro-cesses governing nitrogen dynamics and system pro-

ductivity, and how they affect different functions of theintegrated system. Because of the complexity of inte-

grated aquaculture/agriculture systems (Edwards et al.,1988), it is impractical to conduct basic field experi-ments that answer questions concerning the contribution

of different processes to the structure and function ofthe system (Thornley and Verberne, 1989). Previously,

researchers have used conventional tools for agroecosys-

tem analysis like energy budgets and bioresource flowconceptual models to describe and analyze integrated

systems. However, this approach does not capture the

dynamic properties of the system (Conway, 1987; Light-foot et al., 1993).

Simulation models are useful tools in the analysis of

complex systems and biogeochemical cycling of nutri-

ents (Anderson, 1992). Models assist scientists in con-

ceptualizing, summarizing and analyzing complex

phenomena (Hall and Day, 1977). Simulation models

have been used to simulate different terrestrial and

aquatic systems (Hall and Day, 1977; France and

Thornley, 1984). Existing models developed to analyzeintegrated aquaculture/agriculture systems have largely

been empirical (van Dam, 1995; Schaber, 1996, steady

state (Dalsgaard, 1996) or have failed to incorporate

major components of the system (van Dam, 1995;

Dalsgaard, 1996; Schaber, 1996). The empiricism of cur-

rent integrated systems models, their lack of dynamical

properties, and exclusion of important system compo-nents limit their extendability and utility in the ecologi-cal analysis of dynamic ecosystems.

A new model (Integrated Aquaculture/Agriculture

System model or IAAS) has been developed for simulat-

ing organic matter and nitrogen dynamics through aqua-

culture and integrated conventional agriculture. Given

the scope of the material related to the description of

IAAS, it was decided to present it in two parts. The

objective of this first part is to present an overview ofthe model and some details of significant model compo-nents. A brief overview of the model calibration is also

presented in this first part. Application of the model toidentify important processes and priority areas for future

research in integrated aquaculture/agriculture systems is

presented in the second paper (Jamu and Piedrahita,

2002). Further details are available elsewhere (Jamu,1998).

2. Model overview

The model described in this paper is a dynamic, deter-

ministic and mechanistic model that simulates fish andcrop production, organic matter, phytoplankton, dis-

solved oxygen (DO), and nitrogen dynamics in an inte-

grated aquaculture/agriculture system. The model builds

upon existing pond ecosystem models (Nath, 1996; Pied-

rahita, 1990) and a general crop model (SUCROS1)

(Spitters et al., 1989). The model, developed usingStellamodeling software (High Performance Systems,

Inc, Hanover, New Hampshire), has a simulation timestep of 0.125 d and the model differential equations are

solved numerically using Eulers method. The model canbe used to study long-term (i.e. several years) trends bysimulating consecutive production cycles.

The aquaculture ecosystem model developed here is,

however, markedly different from previous models

because of its explicit treatment of organic matter and

nitrogen dynamics and the inclusion of aquaculture pond

sediments as part of the mass balance calculations.Nitrogen is treated as a major component in this model

because of its role as a major limiting nutrient con-

trolling the diversity, dynamics and productivity of ter-

restrial and aquatic systems (Vitousek et al., 1997). On

the other hand, organic matter provides the main link for

integration of aquaculture and agriculture, and plays a

pivotal role in the recycling of nutrients in ecological

systems (Ruddle and Zhong, 1984; Voght et al., 1986).

Although the agriculture component model incorporatedin the integrated model is not very different from the

original SUCROS1 model, the soil nitrogen and water

balance models have been greatly simplified, and nowinclude simple and reduced statements to describe soil

water and nitrogen processes and their effects on crop

growth. The model strengths over existing integrated

aquaculture/agriculture models (van Dam, 1995;Dalsgaard, 1996; Schaber, 1996) are that the present

model is largely mechanistic and includes the important

components of the integrated aquaculture/agriculture

system, while previous models are largely empirical andexclude the major system components/processes. The

model has been developed using a modular approach

where modules, sub-modules, and sub-models are

developed based on biophysical and functional similarity

of the state variables (Table 1). The model consists of

two modules, i.e. aquaculture and agriculture (Table 1).Each module consists of mass balance calculations for

the state variables which describe the state or condition

of the system (Shoemaker, 1977). Figure 1 shows the

major components (sub-modules) of the model and the

main pathways and linkages between them and the sur-

rounding environment. The model contains 19 state vari-

ables that are expressed in units of concentration: kg

ha1 for nutrient and biomass concentrations and numberof individuals ha1 for population densities (Table 1).

Thirty-nine variables and parameters can be modified byusers. Initial values must be input for the following statevariables: fish weight, chlorophyll a concentration, dis-solved oxygen, terrestrial leaf area index, fresh and

stable organic matter content in terrestrial soil. For the


3/12

573D.M. Jamu, R.H. Piedrahita / Environmental Modelling & Software 17 (2002) 571582

Table1

Modules,sub-modules,statevariablesan

dexternalforcingfunctionsforNdynamic

sinintegratedaquaculture/agriculturesyste

m

Module

Sub-module

Submodel

Statevariables

Forcingfu

nction

Aquaculture

Fishpondwatercolumn

Fishgrowth

Fishbiomass,phytoplanktonandorganicmatter

Watertem

perature,

solar

biomass,Totalammon

ianitrogen(TAN),Dissolved

radiation

oxygen(DO)

Fishpopulation

Fishpopulation

Feedquality

Feed(natural,artificial);nitrogenandcarbon

concentration

Phytoplankton

Phytoplanktonbiomass,DO

,TANandnitrate-N;fish

Watertem

perature;solar

biomass

radiation

Waterquality

Organicmatterbiomass;nitrate-NandTAN

concentration;organic

nitrogen;fishbiomass

Fishpondsedime

nt

Pondsedimentorgan

icmatter

Carbohydratesandcru

deprotein

,cellulose,

ligninandWatertem

perature

stableorganicmatterc

oncentrations;TAN

concentration;nitrate-N

concentration,

organic

nitrogen,

DO

Pondsedimentnitrog

en

TANconcentration;ni

trate-Nconcentration;organic

Watertem

perature

nitrogenconcentration

Agriculture

Terrestrialsoil

Croptranspiration

Leafareaindex;rootbiomass;soilnitrogen

Canopyan

dairtemperature;

concentration

windspeed;Solarradiation

Soilevaporation

Deadleafbiomass;soilwater

Soiltemperature;solar

radiation

Irrigation

Cropbiomass;water

Freshorganicmatter

Rootbiomass;shootandtotalleafbiomass

Soiltemperature,

soilwater

Stableorganicmatter

Microbialbiomass;fre

shorganicmatter;stable

Soiltemperature

organicmatternitrogen

Soilnitrogen

Cropbiomass;soilwa

ter;organicmatterbiomass,

waterquality

Terrestrialcrop

cropgrowth(leaf

,stem,

storage/grain

Rootbiomass;rootlen

gth;storagebiomass;fresh

Airtemperature;soilwater,

androot)

organicmatterbiomass;soilnitrogen;pondsediment

solarradia

tion

organicmatter


4/12


Fig. 1. Conceptual model for the integrated aquaculture/agriculture

system showing pathways and linkages between different components

and the surrounding environment.

other state variables (Table 1), initial values can be set

to zero when their values are not known.The aquaculture and agriculture modules can be run

separately or as a single integrated

aquaculture/agriculture model. Integration of aquacul-

ture and agriculture is established through the creation

of water, organic matter, and nutrient pathways between

the two modules. The major pathways between the two

systems are through the utilization of: (a) agricultural

plant wastes as pond fertilizers, (b) pond water to irrigate

agricultural crops and (c) pond sediments to fertilizeagricultural crops. The transfer of plant waste to fertilize

the aquaculture pond is assumed to occur after harvest

of the agricultural crop and irrigation occurs during the

fish production cycle. In cases where sediment organicmatter is removed from the pond at the end of the fishproduction period, the sediment mineral nitrogen and

organic matter concentrations are used as the initial sedi-

ment values for the next production cycle.

3. Agriculture module components and processes

The agriculture module is used to simulate crop

biomass growth, soil organic matter and nitrogen con-

centrations, and crop soil water loss through crop tran-

spiration and soil surface evaporation. The module con-sists of two main sub-modules: terrestrial crop and soil.

Each of the sub-modules is further broken down into

sub-models as shown in Table 1.

3.1. Terrestrial crop sub-module

The terrestrial crop sub-module simulates the crop

biomass growth rate as a function of soil water avail-

ability, air temperature, solar radiation and soil nitrogen

concentration. The crop biomass growth rate is simulated

using the SUCROS1 sourcesink model (Spitters et al.,1989). In this approach, gross assimilated carbon feeds

a pool of carbohydrates, and the carbohydrates are par-

titioned into roots, shoots, and storage organs as a func-

tion of phenological age of the plant. Information on

gross partitioning coefficients of assimilated carbon todifferent crop components, crop phenological develop-

ment rate, maximum carbon assimilation rate and respir-

ation rates of the different crop organs is available from

the literature (Penning de Vries, 1982). The crop con-

sidered in the model is corn (Zea mays).

3.2. Terrestrial soil sub-module

The second sub-module in the agriculture system

component of the model is the terrestrial soil sub-module

(Table 1, Fig. 1). The terrestrial soil sub-module simu-

lates soil nitrogen and organic matter concentrations, andsoil water balance. The modeling of soil nitrogen avail-

ability and uptake follows the approach of van Keulen

and Seligman (1987) whereby nutrients are available

through decomposition of organic material based on firstorder kinetics. The processes affecting soil nitrogen are

fertilization (inorganic and organic), organic matter

decomposition, leaching, and crop uptake. It is assumed

that nitrogen input through wet deposition in rainfall isinsignificant relative to the other sources, and nitrogenis primarily available to the crop plants through mass

flow with the transpiration stream.The terrestrial soil gains organic matter through root

death, reincorporation of crop waste and pond water irri-

gation. The terrestrial soil loses organic matter primarily

due to decomposition. In the model, a multiple pool

approach, where the organic matter is divided into sev-

eral pools based on the reactivity of each pool (Thornleyand Verberne, 1989), is used to simulate organic matter

decomposition. In the present model, the organic matter

is divided into four organic matter pools (easily decom-

posed, moderately easy to decompose, slowly decom-posable, and stable). The organic matter pools decay at

different rates based on their reactivity expressed in a

unit of d1 (in parenthesis) as shown in Eq. (1) (van

Keulen and Seligman, 1987).

Easily decomposable (0.8)

Moderately decomposable (0.05) (1)

Slowly decomposable (0.0095)

Stable (8.3 105)

The easily decomposable organic matter is made up


5/12


of readily metabolizable subgroups of organic matter,

composed primarily of carbohydrates and crude proteins.

The moderately decomposable organic matter is made

up of cellulose a, and the slowly decomposable organic

matter is composed of lignin compounds. Stable organicmatter is defined as the organic matter that has under-

gone decomposition at least once (van Keulen and Selig-man, 1987).

Soil water balance is simulated using simplified calcu-lations for a homogeneous soil layer as affected by eva-potranspiration rates, soil type and soil water content at

field capacity (Addiscott and Whitmore, 1987). For sim-plicity, clay and loam are considered to be the two main

soil types. Any intermediate soil types between clay and

loam are treated as either loam or clay depending on

the dominant soil mineral component, e.g. loamy clay isclassified as clay. This classification removes the needfor fine delineation of soil types.

The leaching of water from the agriculture component

through downward transport of water outside the corn

root zone is calculated as the difference between actual

soil water content and water content at field capacity forthe soil type. Water loss due to surface runoff is not

taken into account in the calculation of soil water. Irri-

gation is calculated based on a fixed frequency schedulewhereby the total amount of water for irrigation is calcu-

lated as the cumulative crop evapotranspiration between

irrigations. Irrigation is equal to zero when rainfall is

greater than crop evapotranspiration.

4. Aquaculture module components and processes

The aquaculture module consists of the fishpond watercolumn and fishpond sediment sub-modules. The fish-pond water column sub-module consists of sub-modelsthat are used to simulate fish growth and population, feedquality, phytoplankton biomass, and water quality (Table1). The fishpond sediment sub-module is used to simu-late pond sediment organic matter and nitrogen dynam-

ics (Table 1).

4.1. Fish growth

The fish growth model simulates the growth of anindividual fish using a fish bioenergetic model adaptedfrom Bolte et al. (1994) that is parameterized for Niletilapia (Oreochromis niloticus). The model (Bolte et al.,

1994) has been modified to include the effects of feedquality, feed preference and digestibility of different feed

resources onfish growth. The model has also been modi-fied to include a multiplier that accounts for genetic dif-ferences between Nile tilapia and other tilapia species.

Details of these modifications are described elsewhere(Jamu and Piedrahita, 1996; Jamu, 1998).

4.2. Fish population

The total number of fish in the aquaculture pond isdependent on the initial number stocked, mortality and

reproduction:

d(population)dt

stocking birthdeathharvest (2)

Fish harvest in Eq. (2) occurs at the end of the pro-

duction period. It is assumed that stocking, birth and har-

vest are equal to zero during the production cycle. The

assumption of zero birth rates may be a simplificationin so far as mixed sex tilapia production is concerned.

However, inclusion of births would require the consider-ation of an age-structured population model. The simul-

ation of age structure is not important in the present

model since the effects of recruits on the pond carrying

capacity is implicitly included in the fish bioenergetic

growth model. Fish population in the pond is assumedto be affected by mortality rates only. Fish mortality inaquaculture ponds is a result of a variety of factors like

high-unionized ammonia concentrations, low DO con-

ditions, predation, poor handling during sampling, and

disease. While it is desirable to predict fish mortalityfrom simulated water quality conditions, the problems

of identifying the actual causes of fish mortality makesthis approach impractical (Nath, 1996). The fish mor-tality rate (death in Eq. (2)) is expressed using a logistic

equation (Pearl and Reed, 1920):

dP

dt

g(PaP)

Pa P (3)

where: g=mortality rate (d1); Pa=management allow-

able fish population; P=fish population at time, t (fish.ha1). The management allowable population is cali-

brated based on survival data recorded at fish harvest.

4.3. Feed intake and preference

Tilapias show a preference for phytoplankton, which

disappears under phytoplankton-limited conditions(Schroeder, 1978). In nature, tilapias have a predomi-

nantly herbivorous diet consisting of phytoplankton,

vegetable debris, zooplankton, and benthic organisms

(Philippart and Ruwet, 1982). Under waste-fed con-

ditions common in integrated aquaculture/agriculture

systems, supplementary feed is applied to ponds to aug-ment the natural food supply. Therefore, both natural and

supplemental feed may be available to the fish. In themodel, it has been assumed that tilapias prefer natural

feed (phytoplankton and detritus) to artificial feed in thefollowing order: phytoplanktondetritussuppementalfeed (Schroeder, 1978; Spataru, 1978; Philippart and

Ruwet, 1982). While this preference series may be true

for waste-fed tilapia ponds, it may not hold for systems


6/12


where fish are fed high quality feeds. Apart from feedquality,fish physiology and learning behavior may alsoaffect fish preference for certain feed types. Therefore,the feed preference series adopted here should be viewed

as a simplification of the complex behavioral and physio-logical factors that influence feed preference by fish in

aquaculture ponds. The feed intake for feed resource iis dependent on the total feed intake rate, a preference

factor (analogous to a half saturation constant), and the

feed concentration. Feed intake rate for the ith feedresource can be generally presented as:

ri Rci

Ksi ci(4)

whereri = intake rate for feed i (g d1); andR is the total

feed intake rate (g d1) (Nath, 1996),Ksi=half saturation

constant for uptake for the ith feed resource (mg l1),

and ci=concentration of the ith feed resource (mg l1).

TheKs value for feed uptake captures the effects of feedpreference by fish and feed concentrations on the feedintake rate. Based on Eq. (4), fish in the model areallowed to initially feed on phytoplankton by setting theKs for phytoplankton uptake by fish to a lower valuerelative to theKs for detritus. When the fish cannot meettheir daily requirements, they supplement their total feed

requirements by consuming detritus and artificial feed.This approach requires the determination of feed uptake

coefficients and feed uptake factors that are used to cal-culate feed intake rates for each of the three feed

resources. The feed uptake coefficient is defined as theMonod coefficient for the feed resource expressing feeduptake as a function of its concentration and correspond-

ing half saturation constant. For example, the phyto-

plankton uptake coefficient (PUC) is expressed as:

PUC Chla

Ksp Chla (5)

where:Chla=measure of phytoplankton biomass density

(chlorophylla, mg l1);Ksp=half saturation constant for

phytoplankton uptake (mg l1). A similar equation can

be developed for the detritus uptake coefficient (DUC)as follows:

DUC DwcKsd Dwc

(6)

where: DWC=the concentration of detritus in the water

column (mg l1) and Ksd=half saturation constant fordetritus uptake (mg l1).

The feed (phytoplankton, detritus, artificial) uptakefactor is defined as the fraction of feed consumed by thefish for each of the feed resources after accounting forthe effects of feed concentration and preference. Sinceit has been assumed that tilapia prefer phytoplankton to

other feed resources, the phytoplankton uptake coef-

ficient is always equal to the phytoplankton uptake factor

(PUF=PUC). Therefore, the phytoplankton uptake rate

(rp, g d1) is equal to the feed intake rate multiplied

by PUF:

rp R.PUF (7)

The PUF is always less than one, although it can

approach one if the phytoplankton concentration is muchgreater than the half saturation constant for uptake.

Therefore the phytoplankton uptake by fish (rp) willalways be less than the total feed intake rate (R). The

balance; R(1-PUF), will have to be met through fishfeeding from the detritus and artificial feed pools. In thecase of the present model the next preferred feed is

assumed to be detritus and the detritus uptake factor

(DUF) is calculated as:

DUF (1PUF)DUC (8)

The detritus uptake factor is then used to calculate the

detritus feed intake rate (rd, g d1

) as:rd R.DUF (9)

If artificial feed is available, it will constitue the bal-ance of the total diet, R(1PUFDUF). Therefore, the

artificial feed uptake factor (AFF) is:

AFF 1(PUF DUF) (10)

and the artificial feed factor by the feed uptake rate (ra,g d1):

ra R.AFF (11)

4.4. Feed quality

Fish growth in integrated aquaculture/agriculture sys-

tems can be affected by low feed quality. The effect of

feed quality on growth is incorporated in the model

based on the work of Urabe and Watanabe (1992) and

the definition of a balanced diet as presented by Sternerand Hessen (1994). The model for the effect of feed

quality on fish growth is expressed as:

q N:CF

QCEifQCE N:CF

or q ifQ

CEN:CF

(12)

where:QCE Kc(N:Cz) is the critical food nitrogen to

carbon ratio below which fish production would be lim-ited by nitrogen, N:CF=nitrogen to carbon mass ratio infood (g N/ g C), Kc=carbon gross growth efficiency (gfeed C incorporated infish biomass/g feed C consumed),N:Cz=nitrogen to carbon mass ratio in fish (g N/g C).The parameter Kc is determined by calculating the ratio

of fish biomass growth rate to food intake rate, allexpressed in terms of carbon (Lucas, 1996). Feed N:C

ratios in excess of the critical feed nutrient requirements

results in decreased fish growth rates due to the meta-


7/12


bolic costs associated with the breakdown of excess

nitrogenous products from the fish. However, low feedN/C ratios due to low protein feed and inadequate N

fertilization are more common in waste-fed aquaculture

where herbivorous and detritivorous fish such as tilapiaare reared (Bowen, 1987) and the effects of high feed

N:C ratios on fish growth have been ignored.

4.5. Phytoplankton biomass

The phytoplankton mass balance calculations incor-

porated in the IAAS model account for phytoplankton

growth, grazing byfish, respiratory losses, settling, deathand sporadic effluent losses through irrigation water. Thephytoplankton sub-model is linked to: (i) the fish growthsub-model through fish grazing, (ii) organic matterdynamics through dead phytoplankton and, (iii) nitrogen

dynamics through nitrogen uptake and mineralization of

dead phytoplankton. Phytoplankton growth rate is a

function of phytoplankton biomass, solar radiation, tem-

perature, and nutrient concentrations. Nitrogen is con-

sidered to be the only limiting nutrient for phytoplankton

growth based on the fact that nitrogen is the primary

nutrient limiting productivity in tropical aquatic and

aquaculture systems (Moss, 1969; Melack et al., 1982;Knud-Hansen et al., 1991). Phytoplankton settling and

death are expressed as first order processes with respectto phytoplankton biomass. The temperature limitation

coefficient is expressed as a skewed normal function(Svirezhev et al., 1984; Nath, 1996).

4.6. Water quality

The major parameters affecting water quality in the

aquaculture module are organic matter, dissolved oxygen

(DO), and total ammonia nitrogen (TAN). Organic mat-

ter decomposition results in the depletion of DO due to

microbial respiration and nitrification. Decomposition oforganic matter also releases nitrogen which undergoes

ammonification to form TAN. TAN levels above the tol-erance threshold for thefish and DO below the tolerancethreshold for fish affect fish growth rates (Boyd, 1995).

The major sources of organic matter in a waste-fed

aquaculture pond are applied feed and organic fertilizers.

The organic matter mass balance calculations for the

water column account for losses that occur during

organic matter decomposition, pond effluent dischargeduring harvest and/or irrigation, and fish consumption.Organic matter is produced within the pond through phy-

toplankton death, fish death, and fecal production. Inaddition, the pond gains organic matter through organic

fertilization and pond influent. The simulation of organicmatter decomposition in the pond water column andsediment is based on the multi-G model of Westrich and

Berner (1984). The multi-G approach is similar to the

approach adopted in the modeling of terrestrial crop

organic matter described earlier. The decomposition of

organic matter can be expressed as:

rGi kiTcGi (13)

Gtm

m

i 1

Gi (14)

where: rGi=rate of decomposition for the ith group

organic matter group (kg ha1 d1); Gi =concentration

of organic matter in the ith group (kg ha1); Gtm=the

total concentration of organic matter (kg ha1);ki=decay

rate constant of theith organic matter group as a function

of the substrate C:N ratio (d1); Tc=temperature multi-

plier. The concentrations of different groups or organic

matter fractions are determined from proximate analyses

for different feeds or fertilizers being applied to the

pond. Literature values (Gohl, 1981) can be used if the

proximate analyses are not available. These data areinput to the IAAS model by the user.

Dissolved oxygen concentrations are simulated using

the DO sub-model that follows the procedures described

by Piedrahita (1990). In this sub-model, the water col-

umn DO mass balance calculations include photosyn-thetic DO production, DO exchange with the atmos-

phere, consumption through biological (fish,phytoplankton and microbial populations) respiration in

the water column and sediment oxygen demand. The

demand for oxygen in the sediment is primarily due to

the oxidation of TAN through nitrification, and of meth-

ane and hydrogen sulfide by a wide range of bacteria.Only nitrification oxygen demand in the sediment ismodeled explicitly. The remaining oxygen sinks(methane and hydrogen sulfide oxidation, macrobenthosrespiration) are combined into a single sediment respir-ation term that is calibrated from measured sediment res-

piration rates (Boyd and Pippopinyo, 1994). Sediment

respiration rate is expressed as a function of temperature

and water column oxygen concentrations (Walker and

Snodgrass, 1986):

rOs ksO2O2

KO2 O2(15)

where rOs=sediment respiration rate (mg O2 l1 d1);

ks=specific sediment respiration rate (d1); O2=watercolumn oxygen concentration (mg O2l

1 ), andKO2=half

saturation constant for oxygen (mg O2 l1).

Water column and sediment TAN is produced from

the ammonification of organic nitrogen which occursduring the decomposition of organic matter. The nitro-

gen transformations are simulated using first order mod-els adopted in other fish pond ecosystem models(Piedrahita, 1990; Kochba et al., 1994). The rate of nitro-

gen mineralization is calculated as a first order reactionfrom the organic matter decomposition rate (Eq. (13)).


8/12


In addition to organic matter mineralization, the pond

nitrogen mass balance also includes nitrogen fertilization

influent and effluent discharge, diffusion between sedi-ment and water column, denitrification and fishexcretion. Fish nitrogen excretion is modeled explicitlyusing mass balance calculations by expressing the fish

bioenergetic model in terms of nitrogen. Any excessnitrogen not utilized by the fish is assumed to beexcreted as TAN. The simplified equation for calculatingfish TAN excretion is:

rTAN rF rFC rET (16)

whererTAN=TAN production rate,rF=TAN excreted dur-

ing fasting catabolism,rFC=TAN excreted during feeding

catabolism and rET=TAN excreted in excess of fishrequirements for growth and metabolism. All units in

Eq. (16) are in kg ha1 d1. It is assumed that all the

nitrogen in excess of the fish requirements for growthand respiration is excreted as TAN. In addition, the

nitrogen model includes diffusion terms for nitrate nitro-

gen (NO3 N) and TAN. The diffusion of NO

3 N

and TAN is obtained from (Blackburn and Blackburn,

1992):

Diffusion rate (Porosity)(Diffusion coefficient) (17)

(Concentration difference)

Depth

where: Diffusion rate=diffusion rate for nitrogen species

(kg m2 d1), Porosity=void volume fraction(dimensionless), Diffusion coefficient (m2 d1), Concen-

tration difference=difference in concentration betweenwater and sediments (kg m3), and Depth=sediment

depth (m). Porosity values are calculated from published

bulk density and organic matter particle density valuesas follows:

Porosity 1 Bulk density

Particle density (18)

where: Bulk density and Particle density have units of

kg m3. Mean values for sediment bulk and particle den-

sity from a pond sediment characterization study by

Munsiri and Boyd (1995) are used in the model asdefault values. It is assumed that all the pore spaces in

the sediment are occupied by water, therefore the vol-

ume of pore space in the sediment is equal to sediment

pore water volume.

4.7. Pond sediment organic matter and nitrogen

dynamics

The mass balance calculations for pond sediment

include settling, resuspension, and decomposition.Organic matter settling and resuspension are represented

as first order processes, while the multi-G model (Eqs.(13) and (14)) is also used to simulate decomposition of

organic matter in the sediment. The sediment is divided

into two zones: the upper 1 mm layer whose oxygen

status is dynamic and dependent on the oxygen concen-

tration in the water column, and an anaerobic layer at

sediment depths greater than 1 mm. The first order decayconstants for the different organic matter pools in the

aerobic layer are similar to those adopted in the watercolumn. However, for the anaerobic layer, lower (about

1020% of the aerobic values, Prats and Llavador, 1994;Reddy and Patrick, 1984) decay rate constants are usedto reflect the overall lower efficiency of other electronacceptors in the oxidation of organic matter.

Sediment nitrogen simulations are generally similar to

those of the water column. However, in addition to dif-

fusion, leaching, and nitrification, sediment pore waterTAN is adsorbed by the sediment exchange complex.In this model, unidirectional movement of TAN to the

exchange complex is assumed to occur during TAN

adsorption. It is further assumed that equilibrium

between porewater and adsorbed TAN is reached instan-

taneously. The maximum amount of TAN adsorbed by

sediments is determined by a potential upper limit for

TAN adsorption calculated from sediment cation

exchange capacity (CEC) values (Mehrani and Tanji,

1974). The potential upper limit for TAN adsorption cal-culated from CEC values is only approximate, since not

all exchange sites on the sediment are occupied by TAN

as other cations compete for the exchange sites. TAN

adsorption only occurs when the potential upper limit

for TAN adsorption is greater than the TAN adsorbed

by the sediment, i.e. when the exchange sites are not

completely occupied by TAN. The TAN adsorption rateis defined as the rate at which pore water TAN isattracted to the sediment through physical and chemical

processes (Tchobanoglous and Schroeder, 1987). TAN

adsorption rate can be obtained from the following equ-

ation (Deizman and Mostaghimi, 1991):

TANadsorption (PorewaterTAN) (19)

(SpecificTANadsorption)(AdsorbedTAN)

where: TANadsorption=rate of TAN adsorption by the

sediments (kg ha1 d1), PorewaterTAN=TAN in the

pore water (kg ha1), SpecificTANadsorption=specific

rate of TAN adsorption (d1) and AdsorbedTAN=Frac-tion of TAN adsorbed (Dimensionless). The fraction of

TAN adsorbed by the sediment is expressed as (van Raa-

phorst and Malschaert, 1996):

AdsorbedTAN KD

(KD 1) (20)

where:KD=TAN distribution coefficient (dimensionless),defined as the ratio of sediment to pore water TAN.

In addition to the processes occurring in the sediment

accumulated during the growing season, the model also

incorporates processes occurring in the mineral compo-


9/12


nent of the soil. There is a lack of information on the

extent to which the mineral soil participates in pond

sediment and water column processes. However, in the

current pond sediment sub-module, the model bound-

aries are determined by considering the sediment sam-pling depth for the parameters of interest. In the data

used here for model testing, a 5-cm core sampling depthwas used to determine the nutrient and organic matter

of pond sediments (Berkman, 1995). The volume of min-

eral soil involved in sediment nitrogen and organic mat-ter processes is, therefore, defined as the differencebetween the sampling depth and the depth of accumu-

lated organic matter multiplied by the pond area. Using

this definition, the organic matter layer is assumed to bea distinct and separate layer from the mineral layer, an

assumption validated by Munsiri and Boyd (1995). Thenitrogen processes occurring in the mineral sediment are

similar to those occurring in the organic matter layer.

These nitrogen processes include Ficksian diffusion,

leaching through water infiltration losses, denitrificationof nitrate, and total ammonia adsorption by the soil

exchange complex.

5. Model Calibration

Calibration values for parameters in the agriculture

module were available from published reports because

the model has been validated for a wide range of

environmental conditions (Spitters et al., 1989). There-

fore, only the aquaculture model was calibrated. The

calibration procedure involved running the model usinginputs from observed data, comparing model outputagainst observations and making appropriate adjustments

to parameter values until a general fit between modeloutput and measured values was achieved. Given the

scope and structure of the model, the calibration process

was carried out sequentially for each sub-module in the

aquaculture component model (Table 1).

Data inputs from the Pond Dynamics/Collaborative

Research Support Program (PD/A CRSP) database

(Berkman, 1995) were used to provide initial modelvalues for state variables, site characteristics (elevation

and latitude), and environmental variables (wind speed,

solar radiation, air and water temperature). The PD/A

CRSP database contains a variety offish production andenvironmental data obtained during the conduct of dif-

ferent fish culture experiments conducted at five sitesaround the world. Data from the PD/A CRSP experi-

ments conducted at the Rwasave Fish Culture Station,

Butare, Rwanda (2.4 S and 29.45 E) were chosen for

calibrating the model (Table 2). The site is located at an

elevation of 1625 m and has a humid tropical climatewith average annual temperatures between 14 and 28C

and a mean monthly humidity range of 5983%(Bowman and Claire, 1996). The water supply to the

station had a pH range of 6.57.0, and alkalinity of17 mg l1 as CaCO3. The soils at the site are acidic

(pH=4.54.8), and organic matter and cation exchangecapacity range between 0.7 and 5.1% and between 4.5

and 17.6%, respectively. The dataset chosen for calibrat-ing the model consisted of observations collected during

the conduct of an experiment on fertilization of tilapiapond using plant wastes and chicken manure at 1000 kg

ha1 wk1 and 500 kg ha1 wk1 dry weight, respect-

ively. This dataset was chosen because it is the only dat-aset in the PD/A CRSP database where plant wastes

were used as a pond fertilizer. Considering that plant

wastes from crop systems provide a crucial link for inte-

gration, calibration of the model using this dataset pro-

vided realistic parameter estimates for the range of

inputs normally used in waste-fed aquaculture systems.

6. Conclusions

The integration of aquaculture and agriculture activi-

ties is of primary interest for aquaculturists and agricul-

tural ecologists as a means of sustaining system pro-

ductivity and mitigating the negative environmental

impacts of aquaculture effluents. The role of integrationin sustaining system productivity and reducing the nega-

tive environmental impacts of aquaculture can only be

enhanced if more information about the integrated sys-

tem, module interactions and processes and mechanisms

controlling the functioning of the integrated system is

known (Fig. 1).

IAAS represents an important first step in formalizinga model that includes the most important componentsof an integrated aquaculture/agriculture system, and that

quantifies the complex interactions between the differentcomponents of the system. For example, the inclusion

of pond sediments as part of the mass balance calcu-

lations in the model is particularly important because of

the role that sediments can play in pond water quality

as well as in the recycling of nutrients to the agriculture

component. Because the model has been developed in a

modular format, individual sub-modules and sub-modelscan be modified independently without affecting theintegrity of the model. The inclusion of processes and

interactions between the main system components and

the adoption of the modular format to the model pro-

vides a useful tool for exploring the effects of different

management scenarios on the ecological performance,productivity, and environmental effects of different types

of integrated aquaculture/agriculture systems.

Acknowledgements

The research described here was conducted with fund-

ing from the Pond Dynamics/Aquaculture Collaborative


10/12


Table 2

Parameter calibration values for the aquauclture module obtained using fish production and pond data Rwanda (PD/A CRSP Workplan 4, Experiment

3) as model input

Parameter/constant Calibrated Literature range System Reference

O2 consumed per oxidized 3 g g1 1.083.00 g g1 marine sediments; fishponds Jrgensen, 1976; Schroeder,

organic matter 1978Ks for phytoplankton C 1 mg C l1 16 mg C l1 fish ponds; laboratory Nath, 1996; Perschbacher

uptake rate and Lorio, 1993; Northcott

et al., 1991

Denitrification rate k 10 d1 0.0510 d1 fish tanks/ponds; marine Neori, 1996; Acosta-Nassar

sediments et al., 1994; Vanderborght et

al., 1977

TAN diffusion rate k 5.0E10 m2 d1 fish ponds Hargreaves, 1995; Schro-

eder, 1978; Acosta-Nassar et

al., 1994

NO3 diffusion rate k 0.00015 m2 d2 0.000150.00062 m2 d1 marine sediments;fish ponds Acosta-Nassar et al., 1994

Ks for nitrogen uptake 1 mg N l1 0.11 mg N l1 laboratory systems Prats and Llavador, 1994

TAN partition coefficient 10 g g1 540 g g1 fish ponds; agricultural soils Hargreaves, 1995; Mehrani

and Tanji, 1974

Water column respiration 1 d1 fish ponds PD/A CRSP data

rate kSediment respiration rate k 0.5 d1 0.032.57 g O2 m

1 d1 fish ponds Teichert-Coddington and

Carlos, 1993

Upper sediment N mineraliz- 0.05 d1 0.00010.05 d1 lake sediments; fish ponds Kochba et al., 1994

ation rate k

Lower sediment N min- 0.003 d1 0.000010.003 d1 lake sediments; wetlands Reddy and Patrick, 1984

eralization rate k

Phytoplankton respiration 0.005 d1 0.0050.05 d1 laboratory Lee et al., 1991; Prats and

rate k Llavador, 1994

TAN adsorption capacity 0.00012 g g1 pond sediments PD/A CRSP data

Ks for detritus uptake 40 g C l1 40 80 gC l1 fish ponds Nath, 1996

Non-algal light extinction 8.42 m1 fish ponds PD/A CRSP data

coefficient

Chlorophyll a per phyto- 0.02 g g1 0.0050.025 g g1 fish ponds Lee et al., 1991

plankton biomass ratio

Phytoplankton death rate k 0.09 d

1 0.010.09 d

1 Hagiwara and Mitsch, 1994Anaerobic layer mineraliz- 0.001 d1 0.00010.0008 d1 freshwater sediments Jrgensen, 1979

ation rate k

Sediment oxygen demand k 1.4 d1 0.51.4 d1 freshwater sediments Walker and Snodgrass, 1986

Phytoplankton digestibility 0.72 d1 0.380.80 d1 Popma, 1982; Boyd, 1995

Phytoplankton respiration 0.05 d1 0.0050.05 d1 freshwater Lee et al., 1991; DiToro et

rate k al., 1971

Nitrification 0.05 d1 0.050.1 d1 Prats and Llavador, 1991

rate k

Research Support Program (PD/A CRSP), supported byUSAID Grant No. LAG-G-00-96-90015-00 and by con-

tributions from the participating institutions, and by the

Rockefeller Foundation through a fellowship awarded to

the first author. Special appreciation is due to the PD/ACRSP colleagues who collected the data used in this

study. The authors wish to thank Dr Randall Brummett

for permission to use the ICLARM Malawi aquaculture

database. Thanks are also due to the Malawi Meteoro-

logical Department for permission to use the Malawi

weather database. The PD/A CRSP accession number

for this paper is 1229.

References

Acosta-Nassar, M.V., Morell, J.M., Corredor, J.E., 1994. The nitrogen

budget of a tropical semi-intensive freshwater fish culture pond.

Journal of the World Aquaculture Society 25, 261269.

Addiscott, T.M., Whitmore, A.P., 1987. Computer simulation of

changes in soil mineral nitrogen and crop nitrogen during autumn,

winter and spring. Journal of Agricultural Science, Cambridge 109,

141157.

Anderson, T.R., 1992. Modeling the influence of food C/N ratio, and

respiration on growth and nitrogen excretion in marine zooplankton

and bacteria. Journal of Plankton Research 14, 16451671.

Blackburn, T.H., Blackburn, N.D., 1992. Model of nitrification and

denitrification in marine sediments. FEMS Microbiology Letters

100, 517522.


11/12


Berkman H. Central Database. In: Goetze B, Burke D, NcNamara M,

Clair D, editors. Thirteenth Annual Technical Report, PD/A CRSP,

Corvallis, Oregon, 1995, 106 pp.

Bolte JP, Nath SS, Ernst DE. POND: A decision support system for

pond aquaculture, pp. 4867. In: Egna H, Bowman J, Goetze B,

Weidner N. editors. Twelfth Annual Report. Pond

Dynamics/Aquaculture CRSP Management Office, Office of Inter-

national Research and Development, Oregon State University,Corvallis, Oregon. 1994, 205 pp.

Bowen SH. Composition and nutritive value of detritus. In: Moriarty

DJW, Pullin RSV, editors. Detritus and Microbial Ecology in

Aquaculture. ICLARM Conference Proceedings 14. International

Center for Living Aquatic Resources Management, Manila, 1987,

420 pp.

Bowman J, Claire DZ. Pond Dynamics Aquaculture Collaborative

Reports Vol. 1. PD/A CRSP site descriptions. Pond

Dynamics/Aquaculture Collaborative Research Support Program.

PD/A CRSP Management Office, Corvallis, Oregon, 1996.

Boyd, C.E., Pippopinyo, S., 1994. Factors affecting respiration in dry

pond bottom soils. Aquaculture 120, 283293.

Boyd, C.E., 1995. Bottom Soils, Sediment, and Pond Aquaculture.

Chapman and Hall, New York.

Conway, G.R., 1987. Properties of agroecosystems. Agricultural Sys-tems 24, 95117.

Dalsgaard JPT. An ecological modeling approach towards the determi-

nation of sustainability in farming systems. Ph.D. Thesis, Royal

Agricultural and Veterinary University, Denmark, 1996.

Deizman, M.M., Mostaghimi, S., 1991. A model for evaluating the

impacts of land application of organic waste on runoff water qual-

ity. Research Journal WPFC 63, 1727.

DiToro, D.M., OConnor, D.J., Thoman, R.V., 1971. A dynamic model

of the phytoplankton populations in the SacramentoSan Joaquin

Delta. Adv Chem. Ser 106, 131150.

Edwards P, Pullin RSV, Gartner JA. Research and development of

integrated croplivestockfish farming systems in the tropics.

ICLARM Studies and Reviews 16, 53 pp. International Center for

Living Aquatic Resources Management, Manila, 1988.

France, J., Thornley, J.H.M., 1984. Mathematical Models in Agricul-ture. Butterworths, London.

Gohl B.Tropical feeds. FAO animal production and health series; No.

12. Food and Agriculture Organization, Rome, 1981.

Hagiwara, H., Mitsch, W.J., 1994. Ecosystem modeling of a multi-

species integrated aquaculture pond in south China. Ecological

Modelling 72, 4173.

Hall, C.A.S., Day, J.W., 1977. Ecosystem Modeling in Theory and

Practice. An Introduction with Case Histories. University of Colo-

rado Press, Boulder, CO.

Hargreaves JA. Nitrogen biogeochemistry of aquaculture pond sedi-

ments. Ph.D. Dissertation, Louisiana State University, Baton

Rouge, 1995, 241 pp.

Jamu DM. Modeling Organic Matter and Nitrogen Dynamics in Inte-

grated Aquaculture/Agriculture Systems: Effects of Cycling Path-

ways on Nitrogen retention and System Productivity. Ph.D. Disser-tation. University of California, Davis, 1996, 297 pp.

Jamu DM, Piedrahita RH. Aquaculture pond modeling for the analysis

of integrated aquaculture/agriculture systems: Fishpond organic

matter and nitrogen dynamics. In: Burke D, Goetze B, Clair D,

Egna H, editors. Fourteenth Annual Technical Report, PD/A CRSP,

Corvallis, Oregon, 1996, 191 pp.

Piedrahita, R.H., Jamu, D.M., 2002. An organic matter and nitrogen

dynamics model for the ecological analysis of integrated

aquaculture/agriculture systems. II. Model evaluation and appli-

cation. Environmental Modelling and Software 17, 583592.

Jrgensen, S.E., 1976. A model offish growth. Ecological Modelling

2, 303313.

Jrgensen SE. Handbook of Environmental Data and Ecological Para-

meters. Pergamon Press, Oxford, UK, 1979.

Knud-Hansen, C.F., Batterson, T.R., McNabb, C.D., Harahat, I.S.,

Sumantadinata, K., Eidman, H.M., 1991. Nitrogen input, primary

productivity andfish yield in fertilized freshwater ponds in Indone-

sia. Aquaculture 94, 4963.

Kochba, M., Diab, S., Avnimelech, Y., 1994. Modeling nitrogen trans-

formations in intensively aerated fish ponds. Aquaculture 120,

95104.

Lee, J.H.W., Wu, R.S.S., Cheung, Y.K., 1991. Forecasting of dissolvedoxygen in marine fish culture zone. Journal of Environmental

Engineering 117, 816833.

Lightfoot, C.L., Bimbao, M.P., Dalsgaard, J.P.T., Pullin, R.S.V., 1993.

Aquaculture and sustainability through integrated resource manage-

ment. Outlook on Agriculture 22, 143150.

Lucas, A., 1996. Bioenergetics of Aquatic Animals. Taylor and

Francis, London.

Mehrani, M., Tanji, K.K., 1974. Computer modeling of nitrogen trans-

formations in oils. Journal of Environmental Quality 3, 391396.

Melack, J.M., Kilham, P., Fisher, T.R., 1982. Responses of phyto-

plankton to experimental fertilization with ammonium and phos-

phate in an African soda lake. Oecologia 52, 321326.

Moss, B., 1969. Limitation of algal growth in some Central African

waters. Limnology and Oceanography 14, 591601.

Munsiri, P., Boyd, C.E., 1995. Physical and chemical characteristicsof bottom soil profiles in ponds at Auburn, Alabama, USA and a

proposed system for describing pond soil horizons. World Aquacul-

ture Society 26, 346377.

Nath S. Development of a Decision Support System for Pond Aquacul-

ture. Ph.D. Dissertation, Oregon State University, Corvallis, 1996,

273 pp.

Neori, A., 1996. The type of N-supply (ammonia or nitrate) determines

the performance of seaweed biofilters integrated with intensive fish

culture. The Israel Journal of Aquaculture-Bamidgeh 48, 1927.

Northcott, M.E., Beveridge, M.C.M., Ross, L.G., 1991. A laboratory

investigation of the filtration and ingestion rates of the tilapia Ore-

ochromis niloticus feeding on two species of blue green algae.

Environmental Biology of Fishes 31, 7585.

Pearl, R., Reed, L.J., 1920. On the rate of growth of the population

of the United States since 1790 and its mathematical representation.Proc. Natl. Acad. Sci 6, 275288.

Penning de Vries, F.W.T., 1982. Crop production in relation to avail-

ability of nitrogen. In: Penning de Vries, F.W.T., van Laar, H.H.

(Eds.), Simulation of plant growth and crop production. Pudoc,

Wageningen.

Perschbacher, P.W., Lorio, W.J., 1993. Filtration rates of catfish pond

phytoplankton by nile tilapia Oreochromis niloticus. Journal of the

World Aquaculture Society 24, 434437.

Philippart J-Cl, Ruwet J-Cl. Ecology and distribution of tilapias. In:

Pullin RSV, Lowe-McConnell RH, editors. The biology and culture

of tilapias. ICLARM Conference Proceedings 7, ICLARM, Manila,

1982, 432 pp.

Piedrahita, R.H., 1990. Calibration and validation of TAP, an aquacul-

ture pond water quality model. Aquacultural Engineering 9, 7596.

Popma TJ. Digestibility of selected feedstuffs and naturally occurringalgae by tilapia. Ph.D. Dissertation, Auburn University, Auburn,

1982.

Prats, D., Llavador, F., 1994. Stability of kinetic models from waste

stabilization ponds. Water Research 28, 21252132.

Reddy, K.R., Patrick, W.H., 1984. Nitrogen transformations and loss

in flooded soils and sediments. CRC Critical Reviews in Environ-

mental Control 13, 273309.

Ruddle, K., Zhong, G., 1984. Integrated AgricultureAquaculture in

South China: the DikePond System of the Zhujiang Delta. Cam-

bridge University Press, Cambridge.

Schaber J. FARMSIM: a dynamic model for the simulation of yields,

nutrient cycling and resourceflows on Philippine small-scale farm-

ing systems. M.Sc. Thesis, Fachbereich Mathematik/Informatik,

University of Osnabrueck, Germany, 1996, 114 pp.


12/12


Schroeder, G.L., 1978. Autotrophic and heterotrophic production of

micro-organisms in intensily-manured fish ponds and related fish

yields. Aquaculture 14, 303325.

Shoemaker, C.A., 1977. Mathematical construction of ecological mod-

els. In: Hall, C.S., Day, J.W. (Eds.), Ecosystem modeling in theory

and practice: an introduction with case studies. University of Color-

ado Press, Boulder, CO.

Singh, S., Marsh, L.S., Vaughan, D.H., Libey, G.S., 1996. A computersimulation model to optimize greenhouse size for an integrated

(fish production, hydroponics) system. Transactions of the ASAE

39, 22412248.

Spataru, P., 1978. Food and feeding habits of Tilapia zilli (Gervais)

(Cichlidae) in Lake Kinneret (Israel). Aquaculture 9, 4759.

Spitters, C.J.T., van Keulen, H., van Kraalingen, D.W.G., 1989. A sim-

ple and universal crop growth model: SUCROS1. In: Rabbinge, S.,

Ward, S.A., van Laar, H.H. (Eds.), Simulation and systems man-

agement in crop protection. Simulation Monographs. Pudoc, Wag-

eningen.

Sterner, R.W., Hessen, D.O., 1994. Algal nutrient limitation and the

nutrition of aquatic herbivores. Annual Review in Ecology and Sys-

tematics 25, 129.

Svirezhev, Yu.M., Krysanova, V.P., Voinov, A.A., 1984. Mathemat-

ical modelling of afish pond ecosystem. Ecological Modelling 21,315337.

Tchobanoglous, G., Schroeder, E.D., 1987. Water Quality. Addison

and Wesley, Reading, MA.

Teichert-Coddington DR, Carlos N. Soil respiration: Effects of chicken

litter and urea. In: Egna H, Bowman J, McNamara M, edditors.

Eleventh Annual Administrative Reprt, PD/A CRSP, Corvallis,

Oregon, 1993,178 pp.

Thornley, J.H.M., Verberne, E.L.J., 1989. A model of nitrogen flows

in grassland. Plant, Cell and Environment 12, 863886.

Urabe, J., Watanabe, Y., 1992. Possibility of N or P limitation for

planktonic cladocerans: an experimental test. Limnology and

Oceanography 37, 244251.

van Dam AA. Modeling studies offish production in integrated aquac-

ultureagriculture systems. Ph.D. Thesis. Wageningen Agricultural

University, Wageningen, The Netherlands, 1995.

van Keulen, H., Seligman, N.G., 1987. Simulation of Water Use, Nitro-

gen Nutrition and Growth of Spring Wheat. Pudoc, Wageningen.van Raaphorst, W., Malschaert, J.F.P., 1996. Ammonium adsorption

in superficial North Sea sediments. Continental Shelf Research 16,

14151435.

Vanderborght, J., Wollast, R., Billes, G., 1977. Kinetic models of diag-

enesis in disturbed sediments Part 2: Nitrogen diagenesis. Lim-

nology and Oceanography 22, 794802.

Vitousek PM, Aber J, Howarth RW, Likens GE, Matson PA, Schindler

DW, Schleisenger WH, Tilman GD. Human alterations of the glo-

bal nitrogen cycle: Causes and consequences. Issues in Ecology 1,

Ecological Society of America, Washington, DC, 1997, 15 pp.

Voght, K.A., Grier, C.C., Voght, D.J., 1986. Production, turnover, and

nutrient dynamics of above and belowground detritus of world for-

ests. Advances in Ecological Research 15, 303377.

Walker, R.R., Snodgrass, W.J., 1986. Model for sediment oxygen

demand in lakes. Journal of Environmental Engineering 20, 2543.Westrich, J.T., Berner, R.A., 1984. The role of sedimentary organic

matter in bacterial sulfate reduction. The G-model tested. Lim-

nology and Oceanography 29, 236249.

Williams, M.J., 1997. Aquaculture and sustainable food security in the

developing world. In: Bardach, J.E. (Ed.), Sustainable Aquaculture.

Wiley, New York, pp. 1551.

an organic matter and nitrogen dynamics model for the ecological

Documents