A Multimodeling basis for across-trophic-level ecosystem modeling : the Florida Everglades example

MISSING IMAGE

Material Information

Title:
A Multimodeling basis for across-trophic-level ecosystem modeling : the Florida Everglades example
Series Title:
Department of Computer and Information Science and Engineering Technical Reports
Physical Description:
Book
Language:
English
Creator:
Fishwick, Paul A.
Sanderson, James G.
Wolff, Wilfried F.
Affiliation:
University of Florida -- Department of Computer and Information Science and Engineering
University of Florida -- Department of Wildlife Ecology and Conservation
University of Miami -- Department of Biology
Publisher:
Department of Computer and Information Science and Engineering, University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:
Copyright Date:
1997

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
System ID:
UF00095398:00001


This item has the following downloads:


Full Text








A Multimodeling Basis for Across-Trophic-Level Ecosystem Modeling: The
Florida Everglades Example

Paul A. Fi-h', -i.
Department of Computer and Information Science and Engineering
University of Florida

James G. Sanderson
,-11, i, 11. Biological Survey
Department of Wildlife Ecology and Conservation
University of Florida

Wilfried F. Wolff
Department of Biology
University of : .i ini

November 5, 1997


Abstract

We present a new modeling method for use in large-scale physical systems, such as the Ever-
glades ecosystem. T111 current work that has been done in the ATLSS (Across-Trophic-Level
System Simulation) project-which focuses on simulating key Everglades system components
relies on code integration. While this represents a necessary first step in analyzing the dy-
namics of species within the Everglades, it falls short of true model integration. We have
constructed a methodology called object-oriented physical modeling (OOP'. ), which allows
a comprehensive knowledge representation to be constructed for large scale systems. OOP-. I
enforces the idea that an implementation of computer code can be accomplished in an in-
cremental fashion by starting with a conceptual model and progressing to more detailed
models. During this evolutionary procedure, a minimal amount of code is written since the
emphasis is on developing the conceptual model so that it not only represents the intuitive
aspects of the model, but that it is also executable. OOP'. provides a kind of I-lil' 11 i11i"
for ecologists, biologists and hydrologists to communicate and integrate models effectively.

1 Introduction

. ideling is an essential enterprise in ecology when combined with empirical "in the field"
studies. Field work and modeling are being used in conjunction to evaluate proposed changes
to the Florida Everglades as part of a nationally supported restoration effort. One problem
with modeling is that most models are at different abstraction levels. For example, in
studying a landscape such as the Everglades, some species' populations are homogeneous,









and so a continuum approach is warranted to track time-dependent changes in density, mass,
and age structure. For other species, such as wading birds, panthers, deer, and other higher-
trophic organisms, models capable of incorporating individual behavior are required. Three
model types have been proposed: general population, structured population, and individual
based models to model across-trophic-level interactions. Historically, these models are coded,
often using different programming languages, and then executed for their singular purpose.
However, the definition of i-1 I" is one where interactions among species and with the
landscape must be considered. We need to begin with the meaning of the words inI, .,i and
"code." We define model as an abstract-often visual-formal representation of a system.
It is I, ii.,!" in the sense that the representation is provided with a level of clarity that
affords execution of the representation on a computer. Examples of model types are: finite
state machines, System Dynamics models (Forrester 1969; Forrester 1971; Richardson and
Pugh 1981), Petri nets (Peterson 1981), Bond Graphs (Tio. 1.1,1 1975; Cellier 1991; Rosenberg
and Karnopp 1983), and sets of ordinary or partial differential equations. A drawing that is
informal might exhibit the facade of a model, but if it does not conform to strict syntactical
rules, then we label it as an informal, conceptual model. We define code as representing a
linear sequence of source lines written with a programming language such as FORTRAN, C,
C++ or Pascal. Code modules represent packages or libraries of source lines. Both models
and code are translated, but they are at two different abstraction levels. Code is translated
into machine-translatable binaries-or object code. '. ldels are translated into code. While
there is only one level of abstraction separating the concepts of model and code, the distance
traversed from model to code is significant. '. ldels are economic in their representations of
dynamics and represent powerful cognitive tools for reasoning about system behavior.
With reference to .i'i pliii" in the Abstract, an analogy will help to clarify the inte-
gration problem from a modeling perspective. Ti, problems facing model integration are
analogous to the carpenter, plumber, and electrician constructing a house. Each has sep-
arately created a working network: the carpenter has built the frame, the electrician has
soldered wires and boxes for a completely wired house circuit, and the plumber has assem-
bled the plumbing and heating. No blueprint has been used to guide each of the three, and so
arguments arise and fast but inelegant fixes are made just to iiii,!;, it all work." Ti!, three
work in isolation in performing their functions. Ti, problem with this scenario is founded on
a lack of knowledge-sharing. Without a blueprint to serve as a design of house integration,
the house cannot be constructed. To model the Everglades properly, many models must be
written and assembled. We must find a way for the modelers to communicate by specifying
their models using a common framework. Glue and paste methods may initially be nec-
essary, but a knowledge representation and design framework must eventually be created.
Only then, can we focus on models instead of code. This will ease the burden on the ecologist
who wishes to use simulation while investigating species population growth and decay.
.lily of the above concerns are also discussed by Zeigler (Zeigler 1984), whose work
shares a common interest with ours regarding the importance of formal model representation
over code specification. In particular, the System Entity Si, .. !, re, supports a similar object-
oriented knowledge representation with more of a focus on integrating formal representations.
Our work concentrates on model types that represent large classes of models that are used
in practice. We develop a methodology to achieve proper integration of models. First,
we present background on the Everglades in Sec. 2 so that the reader knowledgeable in









simulation will gain some insight into the historical and contextual reasons why modeling is
being done. We follow with a brief discussion of the current state of ATLSS in Sec. 3. We then
proceed to a proposed model integration procedure using OOP. I described in Secs. 4 and
5. -.1 1,,i .i,-. described in these previous two sections are applied to an Everglades example
in Sec. 6. Conclusions are summarized in Sec. 8.

2 Background and Motivation

2.1 Everglades Restoration History

In February 1996, Vice President Albert Gore and Carol '.1 Browner, Environmental Protec-
tion Agency Administrator, announced the Cl 11iii., administration's environmental restora-
tion plan for the south Florida Everglades. Tit, restoration of the greater Everglades land-
scape was declared a top environmental priority. During this century, the Everglades were
successively divided and conquered. By the time the southern Everglades became Everglades
.it1i iI .1i, Park in 1949, the contextual setting of the park made clear that management deci-
sions outside the park would eventually define what happened inside the park. Today, much
of the south Florida landscape supports agriculture and large human populations whose
demands for freshwater define the goals of water management in the region.
As a result of changes in the hydrology brought about by the need to supply human
demands including flood control, many regional features have been changed (Davis and Og-
den 1994). i:'i. 11! ecological processes were altered and organisms responded. What was
once uninterrupted sheet flow became a water flow fragmented by levees and impoundments.
Extensive short hydroperiod marshes were degraded by overdrainage or lost altogether to
development. Attenuated seasonal changes in water depth became more pronounced. '.i-
jor drydowns in the sloughs, once rare, became more frequent. Because of water diversions
from Lake Okeechobee and the Everglades, reduced flows into estuaries enabled seawater
penetration further inland and also contributed to the demise of Florida Bay, a once produc-
tive fishery. Ti, oligotrophic water of the historical Everglades became eutrophic in places
through agricultural runoff and sparse sawgrass marshes changed to chocked cattail stands.
T111 freshwater alterations led not only to a change in the spatial and temporal distri-
bution of living organisms but also to a decline in numbers of native and endemic species.
Animal populations suffered dramatic declines and abnormal dieoffs. Before the turn of the
century, millions of wading birds populated south Florida (Frohring, Voorhees and Kushlan
1988). Now these populations are estimated to be a mere 5% of their previous numbers (Og-
den 1994).
Gunderson et al. (1995) discussed how government policy directed increased control of
the hydrological resources of south Florida. Tit, dramatic decline in wading bird populations
has been blamed on changes in the hydrology, in particular the area and the length of time
land is inundated by water. Ti, U.S. Army Corps of Engineers was given responsibility for
flood control and insuring that agricultural interests were protected (Halloway 1994). Now,
the U.S. Army Corps of Engineers is responsible for the restoration of the same landscape
that is expected to require decades of work and cost hundreds of millions of dollars. Already,
lands north of Lake Okeechobee along the Kissimmee River have been allowed to return to
a more natural flood regime as a result of actions taken to reduce flood control (Dahm et al.









1995).


2.2 Ecosystem Modeling

Dale and Rauscher (1994) reviewed the state of biological modeling. Ti, ~ discussed models
that addressed the impacts of forests at four levels of biological organization: global, regional
or landscape, community, and tree. Tli ;- y-i--.I -fI the development of landscape vegetation
dynamics models of functional groups as a means to integrate the theory of both landscape
ecology and individual tree responses to climate change. Furthermore, they recommend
(1) linking socioeconomic and ecological models; (2) interfacing forest models at different
scales; (3) obtaining data on susceptibility of trees and forests to changes in climate and
disturbance regimes; and (4) relating information from different scales.
Hunsaker et al. (1993) reviewed terrestrial ecosystem models and stressed the use of
geographic information system (GIS) map layers. Ti l.iill GIS is not time dependent the
authors believed a macro language embedded in the GIS can be used to implement time
dependent modeling. Languages such as Arc i., ro Language (A .L) are cumbersome at
best and offer no special data structures beyond arrays. Tli, authors note that "Tli, problem
of integrating models with different temporal and spatial scales is not trivial and requires
special methodologies."
Tli, development of models -hdl. by Dale and Rauscher (1994) and Hunsaker et
al. (1993) depends upon a critical understanding of what is being modeled and the computing
power required to achieve the desired level of understanding. We consider an ecosystem to be
a relatively homogeneous area of vegetation that responds to governing physical processes
fairly uniformly across the area extent (Forman 1995). In contrast, a landscape consists
of two or more ecosystems and responds to the governing physical processes in spatially
and temporally varying ways. We are interested here in developing a methodology that
facilitates modeling heterogeneous landscapes at different spatial and temporal scales and
across trophic levels.

3 The ATLSS Project

Tli, Across-Trophic-Level System Simulation (ATLSS) (ATLSS 1997) is a landscape scale
ecosystem represented by conceptual models, computer programs and code from several
different sources such as the US Geological Survey, University of Tennessee, University of
.Ii.iini and University of '.I~iryland. Ti, plan for ATLSS is for it to be used to assist in
the evaluation of proposed system-wide changes in the hydrology of south Florida. ATLSS
will predict changes in landscape vegetation and multi-level organism responses to proposed
changes in the hydroperiod brought about by restoration efforts. For example, ATLSS will
predict the responses of wading birds to changes in system-wide water levels over a twenty
year period. One potential outcome might be that wading bird populations increase for 10
years and then decline because of some unforeseen change that unfolds only after a decade
of, say, vegetation change. Such unforeseen outcomes can only be studied using complex,
detailed models. Restoration attempts that "do -, iii I,!iiii and then "correct it" based on
short term outcomes might cause irreparable long term damage (Kushlan 1979).









Deer, Panther
., .'.) .'t Alligators Wading Birds Snail kite Cape Sable seaside sparrow
individual-based modules N



Age-size modules Amphibians Crayfish Piscivorous fish Apple snails
reptiles Planktivorous fish



Process modules Zooplankton Macrophytes Vegetation
Periphyton Detritus community
Benthic insects

Abiotic Process Hydrology Fire,freeze,hurricane Nutrients
modules

Figure 1: Planned collection of ATLSS modules (ATLSS 1997)


To simulate abiotic and biotic processes of south Florida, ATLSS was organized in a way
that permitted the use of code modules that represent ecological knowledge at widely varying
levels of detail and across many different ecosystems while at the same time maintaining the
ability to incorporate existing legacy code. Code integration development, toward an inte-
grated ATLSS, has been accomplished by our ATLSS partners at the University of Tennessee
(UT) (ATLSS 1997), with an initial code integration ii lp.11, i-i-.- -ti by one of the au-
thors (Sanderson). ATLSS requires different fidelity simulation codes to be integrated in a
top down fashion. For instance, the hydrology code in ATLSS is cell based, the lower trophic
organisms are coded using stable, time dependent non-linear differential equations, and the
higher organisms are coded using structured population or individual-based codes (DeAnge-
lis and Gross 1992; Wolff 1994). Because appropriate scales are used to simulate organism
responses at each trophic level, reliable qualitative predictions across the south Florida land-
scape are possible and system-wide integrity can be better understood (Fleming, DeAngelis
and Wolff 1995).
Ti, underlying spatial structure of ATLSS is a grid of varying resolution down to '--',
that covers the study area from Lake Okeechobee in south central Florida to Florida Bay
lying south of the peninsula, and east-west from coast to coast. All codes execute on some
part of this grid, though not all codes are grid-based. Fig. 1 illustrates the implementation
plan for ATLSS code modules.
Individual-based codes are behavioral modules that use realistic decision making capa-
bilities of the individual organisms of the species the codes represent. T111 decisions must
be made within the landscape occupied by the individuals. Ti, environmental information,
biotic and abiotic, available to the individual varies as a function of position. '., !vement and
dispersal are functions of the information that is available and perceived by the organism.
Animal movement and habitat selection have long been the subject of behavioral ecologists
hence an organism's rules can be both complex and realistic. ATLSS represents an attempt
to code the behavioral ecology of organisms at the scale of the landscape (Lima and Zollner
1996).









4 Toward Model Integration


Until the present time, the ATLSS project has evolved various code modules that have been
integrated by the University of Tennessee (UT) team. A general lack of resources, and
real deadlines for simulation results, has made it difficult for the UT team, and the overall
ATLSS team as a whole, to research the possibility of integrating models together instead of
large segments of code. At the University of Florida, our purpose is to construct a modeling
language with two distinct objectives: 1) to serve as a visual modeling interface for ecologists,
and 2) to serve as a potential model integration method that ATLSS might adopt in the
future if enough time and resources are provided. ATLSS as it is defined and constructed
today does not use models in the sense that we have defined model in Sec 1. T!iI University
of Florida initiative represents an exploratory project to determine whether ATLSS might
use model integration in the future, rather than to remain an integrated system of code
modules.
T111 first objective is consistent with the current emphasis on code module integration.
Thi, visual model is automatically translated into a code module that can be integrated
directly with the remaining ATLSS code modules. All modules can then be subsequently
directed to "('1pi i,' over specific time and landscape scales. Tli, second objective is more
ambitious, but we feel that the OOP. 1 methodology represents a potentially strong candidate
for a model integration method for ATLSS. We now proceed to discuss the modeling approach
which satisfies the above two objectives.

5 Object-Oriented Physical Modeling

5.1 Overview

Tli, basis for physical modeling in a large-scale system such as ATLSS begins with ;I-_-_ re-
gation and object-oriented design concepts. We will briefly explore the background on both
of these topics. Tli, ;,-.-,ii 1-.,iii .o problem (Zeigler 1976; Zeigler 1979; Zeigler 1985) has long
been a concern to simulationists and ecologists. Zeigler's DEVS formalism and the related
problem of multiple ;,-.-_i .-i i, .i levels have been applied to ecological problems whether be-
havioral or at the level of the landscape ('. I1,iwell and Costanza 1995; Vasconcelos and Zeigler
1993; Vasconcelos, Zeigler and Pereira 1995; Zeigler, Vasconcelos and Pereira 1994; Zeigler
and M.i on 1996). 'ii:i iiindeling, as originally developed by Fi-l, i, 1 (Fi-l, i, 1 1991) in a
paper on heterogeneity in high-level model integration, is a method for integrating specific
high-level model types most often used by ecologists and others in science and engineering
contexts (Fi-li"i 1: 1995). Ti, taxonomy of models in (Fi-li"i 1: 1995) reflects the taxonomy
of programming language styles used in computer science (Fi-li' i, 1: 1996a). li iiniideling
is grounded in a more mathematical systems formalism such as DEVS (Fi-l, i, 1. and Zei-
gler 1992). In this sense, multimodeling (Fi-,li i 1 1995) and multiformalism (Zeigler 1976;
Zeigler 1979) are co-related and at different abstraction levels. Furthermore, multiformal-
ism and multimodeling techniques are founded on system theoretic concepts and modeling
methods.
With regard to object oriented methodology, our approach has been to employ a new
methodology called Object-Oriented Physical .ildeling (Fi-l- i, 1: 1996b), which is an ex-










Class

Attributes C3
Cl C2 C3 Cl C2 C3
Methods
Methods Generalization Aggregation Dual Relationship
Hierarchy Hierarchy Hierarchy
Note:
Ci = cardinality constraint such as "=4" or "<2"

Figure 2: Structure of a class with three relations.


tension to object-oriented (00) design as expressed by the Software Engineering 00 com-
munity (Booch 1990; Rumbaugh et al. 1991). Tli, extension is made specifically to define
how physical models should be incorporated into the kind of object-oriented designs found
in software engineering (Rumbaugh et al. 1991; Booch 1991; Fowler and Scott 1997). Our
work is not the first to combine object-oriented design with simulation. Related work in
simulation has focused on objected oriented techniques (Rothenberg 1989; Hill 1996; Zeigler
1990). Our specific contribution, relating to this work, is that we build a detailed description
of how physical object dynamic and static models are mapped into an object-oriented design
framework using the model taxonomy presented in (Fi-lr-i, 1: 1995). This taxonomy divides
models into categories reflective of programming language styles: declarative, functional and
constraint (Fi-lr' i 1 1997). Primitive models are created from these three types, and then
multimodels (Fi-r, -i, 1: 1991; Fi-l'-i, 1: 1993; Fi-l',-i, 1: et al. 1994; Fi-h, -i. 1: and Zeigler 1992)
are constructed to create the larger models. To create each model within a potential next-
generation ATLSS, a class graph with relations among classes was constructed. Furthermore,
attributes and methods within classes were identified. An object is an instance of a class. A
method is an action within a class and hence within an object. For instance, there exists a
base class Bird, and a wading_bird object is an instance of class Bird. '. 1 II, i y() is in class
Bird, and so fy() is also a method within object wading_bird. When we instantiate, say,
1000 wading bird objects, we are creating 1000 individual wading birds that can be modeled
separately using a common set of attributes such as age, sex, and location, and methods.
T1li act of coding in an object-oriented language, is not ultimately a substitute for model
design. As an example, C++ provides many object-oriented capabilities, but does not enforce
object-oriented design. Norman (1988) -il-f-, the need for visual, conceptual models
in general design for improved user-interfaces to physical instruments and devices. T111
importance of design extends to all scientific modeling endeavors. '.idels must provide a
map between the physical world and the virtual world produced by computers.

5.2 Conceptual Model

A key part of conceptual modeling is identifying classes. This procedure is ill-defined but
some rules and approaches do exist (Fi-l, i, 1: 1995; Graham 1991) to aid the model engi-
neering process. Fig. 2 illustrates generalization (0) and .-.- i t-, -.i i (D) relations. We also
permit an analyst to specify any given relation as both ;.-.-ii .-.i ii .11 and generalization. This
is delineated with a circle inside a square. It is not necessary to group all relations into one














Landscape Tidepool Swamp Organic Non-Organic


Land Water Artificial
Patch Mangrove Cypress

Figure 3: Conceptual :'.idel, Part I.

Organic




UpperTrophic LowerTrophic



Panther Deer Colony Fish Periphyton




Bird



Nestling Fledgling Adult

Figure 4: Conceptual :.idel, Part II.


graph or hierarchy-multiple graphs or hierarchies are possible. Figures 3 and 4 provide the
scenario for samples classes necessary for integrated Everglades modeling. In Fig. 3, we
show two trees of the conceptual model for class Habitat. In the left tree, Landscape, Tide-
pool and Swamp are defined as types of ecosystems or Habitat. :'. i ever, all Habitat-type
objects are ;.---ii, .-,i, of organic and inorganic objects. In Fig. 4, there are two types of
Organic class: LowerTrophic and Upper_Trophic.
Our extension of object-oriented design specifies that an attribute is one of two types:
variable or static model. Likewise, a method is of one of two types: code or dynamic model. A
method can be of a functional (representing a function) or constraint (representing a relation)
nature. Once the conceptual model has been constructed, we identify the attributes and
methods for each class. An attribute is a variable, whose value is one of the common data
types, or a static model. A method can be code, whose form depends on the programming
language, or a dynamic model. Tli, structure of a class is seen in Fig. 5. Variables and code
are described in 00 languages such as C++ (Stroustrup 1991). We define a static model as
a graph of objects and a dynamic model as a graph of attributes and methods. Ti1, model


Habitat


Habitat










Class Name OOP
extension

Variables
Attributes Variables
SStatic Models


Methods Code
SDynamic Models


Figure 5: Structure of a Class.


types of interest here are dynamic. However, the concept of static model complements the
concept of dynamic model: methods operate on attributes to effect change in an object.
Dynamic models operate on static models and variable attributes to effect change.

5.3 Representing Dynamic Models

How are dynamic model components represented in the physical modeling methodology? We
will illustrate two model types (functional and declarative), each with two model sub-types.
For the functional model types, we use a block model and a System Dynamics model. For
the declarative model type, we will use an FSA and a Petri net. Ti, following notation,
consistent with the previous notations, will be used throughout this discussion:

Objects: An object is represented as obj. obji represents a sub-object of obj (in its
;r .-.-, .-t iii, hierarchy), and obj' represents a super-object that is composed, in part,
of obj. When indices i and j are used, it is possible that i = j or i j. This relation
rests with the particular application.

Attributes: obj.a represents a variable attribute and obj.A represents a static model
attribute, a is short for any string beginning with a lower case letter; A is short for
any string beginning with an upper case letter. Attribute references (i.e. names) and
values are relevant: a name is just obj.a whereas the value of attribute a is denoted
as v(obj.a). Thi following special attributes are defined for all objects: obj.input,
obj.output and obj.state and represent the input, output and state of an object at the
current time.

Methods: obj.m() represents a code method and obj..\f() a dynamic model method.
m is short for any string beginning with a lower case letter; [f is short for any string
beginning with an upper case letter. Tih following special methods are defined for
all objects: obj.input() and obj.output() and represent the input and output time
trajectories.

Ti, block model contains a collection of blocks coupled together in a network. Each block
calculates its input, performs the method representing that block, and produces output. A
block is not an object since it represents a method within an object. Without any refinement
within a block, a function takes no time to simulate. Time is ultimately associated with
state change. All obji represent sub-objects of obj. Fig. 6 displays a sample functional









obj.DynModel0


Figure 6: Functional Block model.


obj.DynModel()

obj.input(i _\ obj 2statel=ml() ob 4state2=m2() o\/ bj output()
obj state 1=M() A bj 4state2=M2()
obj al obj al Obj al obj 5al



Figure 7: System Dynamics model.


model. Fig. 7 shows a System Dynamics model that is similar to the block model except
that instead of methods represented by the network nodes, a node represents an object's
state (a variable attribute). Ti, same kind of multimodeling represented in Fig. 6 (refining
a block into a dynamic model) can be done for the model in Fig. 7; a block is refined
into a System Dynamics model. This multimodel refinement is particularly useful for our
two declarative model types shown in Figs 8 and 9 where input() and output() are not as
obvious unless we capture the model inside of a functional block. '.1 i1i ini, dealing is denoted
by drawing a dashed functional block around the model, denoting model refinement. Tli,
methods input() and output() are essential to perform c.( 'ipl../ of models together. An FSA
will have an input() method and a state variable attribute obj.state, but we require coupling
information somewhere if we are to decide where the input is coming from and where the
output is going to. This coupling information for a method of obj is present in some obj'.
For example if the FSA belongs in a bird, defining the dynamics of state change the input
must come from a physical object outside of the bird but within a more encapsulating object
such as the colony.
Ti, predicates pi and p_ in the FSA model in Fig. 8 require further explanation. A
predicate is defined as a logical proposition, whose value is boolean, containing internal and
external events. External events are of the form obj.input() and internal events are of the
form obj.state or obji.state. Rules are another convenient dynamic model (of the declarative
type) that express changes of state and event. A rule-based model is composed of a collection
of rules within an object obj, each rule i of the form: IF pi(event, state) THEN obj.mi() or










obj.DynModel()




obj.input()



vl(obj.state)=n
or
vl(obj.state)=M


obj.outputQ



=m2()
=M2()
---


pi) = boolean valued predicate containing arguments
external event of the form obj.input()
or internal event of the form obji state

Figure 8: A finite state automaton.


obj.DynModel()


obj.inputQ


Figure 9: Petri net.


obj..fi(). T!i, phrase pi(event, state) defines the same logical proposition discussed for the
FSA.
Regarding coupling and closure under coupling of model components, the rule is that
coupling on a single level (i.e., for components within a model at the same level of abstraction)
is performed by coupling the output of one function fi into the input of another f2 to obtain
f2 (f()). This operation represents functional composition. Inter-level coupling is achieved
by implementing any function as one of the \1 () or mi() that are specified in the above
figures. Tli, outside of every model type is a function with an input and output, and the
inside has multimodel entry points in the form of functions 11 and mn. For example, Fig. 8
is a function with an input and output as designated by the functional block surrounding
it. However, new functions may be substituted for the mr and \1 defined as part of the
model. This makes it so that the FSA can be embedded within a larger model as one if
its components, and that the FSA contain refined models of different types that are refined
through the mr and \1 functions. Currently, no provision is made for type checking at the
couplings, although this would be a valuable addition in the implementation.









.li! i i models are executed in the following manner. For any given scenario, execution
begins by invoking the method Execute () within the most general composing object. Hybrid
behavior involving both event scheduling and numerical integration is handled uniformly by
forcing all events to be routed through a single future event list. Continuous behavior is
interpreted to involve discrete events with regular time spacing.

6 Everglades Modeling

To apply the OOP .1 methodology to the Everglades scenario, we first construct a conceptual
model using classes and the ;.-i .-i .-, ili, and generalization relations. Figs. 3 and 4 display
hierarchies for two classes: Habitat and Organic. This provides a logical structure to the
physical objects found in the Everglades. Ti, following are sample relations drawn from the
conceptual models in Figs. 3 and 4:

Landscape, that is composed of multiple Patches is a type of Habitat. From this, we
can create an object called everglades of type Landscape.

Colony is an ;l.-.-regate of Bird of which AdultBird is a type (i.e., class).

Ti-lt represents a type of LowerTrophic form of life.

This conceptual model serves as a base to define attributes and methods. We proceed
to focus mainly on dynamic model methods of importance to the Everglades. Ti, I are
three prominent model types employed: 1) general population, 2) structured population,
and 3) individual-based models. In the first type, one constructs a model that represents
the dynamics of an entire population without regard to transitions within the population
or species structure. General population models have wide coverage in the literature since
they are by far the most utilized model type, mainly due to the ease of closed-form solution
and simple form ('.i i nly 1990). Unfortunately, this simplicity may be artificial rather
than reflecting true behavior. For more accurate models, one can structure a population
by physiological attributes such as age or size (Hallam et al. 1992; Ulanowicz and Platt
1985). Tlii third model type encompasses models where each individual in the population is
modeled separate from the others. T!i, modeling of periphyton is a general population model,
without structure, whereas fish are modeled using structure. Upper trophic animals such as
panther, deer, alligators, and wading birds, such as wood storks, are modeled individually.
DeAngelis et al. (1992; 1992) create a structured hierarchy for these model types. A
p-state model corresponds to a general population model. An i-state model has two sub-
types: distribution and configuration. An i-state distribution model corresponds to the
structured population model and an i-state configuration model corresponds to modeling
each individual. Two sample papers deal with specific tradeoffs with the two i-state ap-
proaches (DeAngelis and Rose 1992; DeAngelis et al. 1993). Wolff (Wolff 1994) refines the
individual approach into two sub-categories: individually-based models and individually-
oriented models. An individually-based model (ii ,-.1) is one where individuals are modeled,
but generally with equation structures that are common among all individuals. An example
of this is provided by Hallam et al. (Hallam et al. 1992) for the change in biomass for an









individual as a function of time and age structure:

a a
a +- m+ = g=(mn, m 2,..., m) (1)
at Oa
where mi represents the mass of individual i and a is age. Note the homogeneity in this
formulation for individuals; individuals are treated as being fundamentally the same and
with simple dynamics. Individually-oriented models (IO.1), instead, model individuals in a
way not possible with the continuous equations in II ,-. models, by adding discrete events,
rules and more elaborate model structures. Tli IO.1 therefore, involves more complex
dynamics for the individual. '. i, .reover, each individual in an IO. 1 model can conceivably be
modeled with different dynamics. In cases where the individuals are simple, as is generally
the case towards the lower trophic end of the food web, an II i. or even a general population
model, may more reasonable, but for higher trophic level animals, such as birds, panthers
and deer, an 10 may be more appropriate.
A key aspect of the OOP' .1 methodology is that we can integrate each of these three mod-
eling paradigms within the same logical structure. Lower trophic levels are generally modeled
with a population model since there are too many individuals to make an individually-based
model practical at this level. ., 1,_reover, the individuals have a sufficiently coarse type of
; _--regate behavior for which a population model is sufficient. Ulanowicz (1994) discusses
this modeling level in some depth.

6.1 General Population Models

Tli, lower trophic level within the Everglades has five key components:

1. Periphyton: micro-algae attached to a plant and bottom surfaces.

2. '.1, rophytes: submerged and emergent aquatic plants.

3. Detritus: dead organic matter.

4. : ..1 ..invertebrates: examples include water fleas and nematodes.

5. .1, roinvertebrates: insect larvae.

Since Periphyton exists as a class in our conceptual model, we create a population object
periphyton from class Periphyton. If we let the state variable (attribute) of periphyton
be pop, and define two parameters for periphyton growth ac and death 3 and neglecting
predation, then we arrive at the object periphy-ton with the following attributes and methods:

Attributes:

Variables: pop, a, 3
Static '.1idels: not defined.

Code: not defined.
Code: not defined.

















-state


Figure 10: Lotka-Volterra population dynamics.


Dynamic '.i dels: (constraint-type) 'pop = ac pop *3 pop2

Tli1 dynamic model of periphyton specifies a simple population model, in the special case
above, a logistic growth model. For many population models with competition for resources,
the equations are only slightly more involved. For example, a Lotka-Volterra model would
be constructed for two population objects preypop and predatorpop, each composed of par-
ticular sub-populations by constructing the necessary classes and objects. Such a model
-i i .-. -.-. that we have a physical scenario composed of an environment (weather), landscape
and populations of organisms. Ti, are two types of populations: predator and prey. For
the sake of the biological metaphor, we choose Panther as the class of predator and Bird and
Deer as sample prey classes. Fig. 10 illustrates the locations of other models that we will
now discuss (i-state and IO.1. model types).
Let's note the rules for generalization and ;i--.i -..ii..ii



1. A specific ;--i i -.ii rule for attribute count:

Population.count = PredatorPop.count + PreyPop.count









A more general ;,-.- ii .-.i .. i,, rule for count, keeping in mind updates and additions
to this conceptual model, is:

C.count = l Ci.count

where C matches any class containing count and Ci matches the sub-classes of C.
2. An ;,---_-i i, ,i rule for dynamic model method Model( in Population:

PredatorPop.rate = PreyPop.birth() PredatorPop.death()
PreyPop.rate = PreyPop.birth() interaction()

3. An -;i-- 1. i. ii, rule for code method interaction() in Population:

interaction() = PreyPop.count x PredatorPop.count


Generalization: We let count be inherited (passed down) but not any of the methods
defined in Population, PredatorPop or PreyPop.

In reviewing our model, we realize several important benefits from the use of OOP'-. in
creating this model. Ti, main benefit is one of knowledge representation that focuses on
class creation and the lexical naming of equational terms such as birth( and interactionO. By
making names explicit, we make the model more comprehensible. Ti, benefits of structure
passing are inheritance and ;-.-ii, .-.iti, Ti, definition of Population.Model() is invariant
to additional PredatorPop or PreyPop sub-classes we may choose to add in the future. For
example, we may later add AlligatorPop under PredatorPop. Since AlligatorPop would pass
count upward via the first ;-.-i, 1-. i.ii i rule, the population model need not be redefined.

6.2 Structured Population Models
An important food source for wading birds, fish populations in the estuaries and freshwater
marshes of the Everglades are modeled (Davis and Ogden 1994). Abiotic factors such as water
level and salinity affect total fish biomass, as does the availability of resources. '., Irtality,
such as caused by periodic or sudden drydowns, and predation reduce fish populations.
Various fish species or groups of species are divided into age- and size-class classes. Fi-lI
survival, growth, aging, and reproduction are modeled using continuous functions of the age-
and size-class of species. '.1 ,vement is also modeled as a function of age- and size-class and
is species specific. For instance, movement of fish between mangrove swamps and interior
marshes during periods of inundation is age- and size-specific for each species of fish. T11i
hydrologic cycle is a principal ecological forcing variable and the total biomass for each class
is updated.
Hallam et al. (1992) define a generic continuous structured model as being of the '. I, I.'I ii 1-
von Foerster type:
Op Op a
p+ p+ (pg) -t (t, a, m, p)p (2)
Ot Oa Om









DeAngelis and Rose (1992) use this type of structure for fish cohorts (groups). This model
is slightly more complex than the general population model since time is not the only inde-
pendent variable; a and m specify the age and biomass of the population. To "stl 1 ii[, a
population by a physiological variable such as age or mass, it is necessary to partition that
dimension. For example, if we are modeling the age a of fish, then we can structure along
this dimension by discretizing age into integral ages (0,1,2, ...). A fish, for example, is in a
larval stage or a adult stage. This provides a structural division of the age dimension into
nominal values "larval" and i.dii;ll "
Ti,, method of incorporating structured models into the OOP'. methodology is similar
to the general population approach. '.Ii,--. m, age a and density p are attributes of the
population object. As in our Lotka-Volterra example, we can add new sub-classes of fish to
a conceptual model without touching the dynamics. This sort of plii'. and 1pl,1 operation
provides a significant level of flexibility in modeling and coding.

6.3 Individual-Based Models

Upper tropic level organisms (i.e., deer, panther, wading birds, and alligators) are modeled
as individuals. Because deer are a primary food resource for panthers, these two species are
combined and modeled in one sub-module. Tli, it behavioral activities, such as movement
across the landscape, are modeled in daily time-steps. Tli, movement of deer is closely
related to the spatial vegetation pattern and, apart from hydrological conditions, mainly
determined by the occurrence and quality of forage the vegetation can provide. On the other
hand wading birds movements, i.e. flights from their colony to their feeding sites and back,
are almost completely determined by the hydrological patterns as they primarily forage in
shallow water areas.
In contrast to deer and panther for which daily time steps are sufficient, wading birds
operate on much smaller time scales. Tl, -, time scales are determined by the various
activities of the birds and can take several hours, e.g. for flying to a particular feeding site,
or as little as several minutes for a bird that arrives at a feeding site with little or no prey
and leaves almost immediately if feeding is poor. Wading birds are therefore simulated using
an event queue.
Tlii 'i ill- in the wading bird model are specified by the various activities of the
individual birds and are determined by behavioral rules. T111 rules can be divided into
two sets: rules that determine what the bird is going to do after its current activity has
terminated and rules that determine how a specific activity is performed. For example, after
an adult bird has returned to its nest, a behavioral rule determines whether it will incubate
I .-.-- or, if the, .-.*-- have hatched, whether the bird will feed its nestlings. Similarly, a rule in
the latter set may be used to determine whether a bird must use flapping flight, thus using
more energy, or may soar and glide which is more energy efficient. Til, behavioral rules are
therefore either of the type:

IF (PI:I1)ICATE) THEN (NEXT_ACTIVITY_OF_BIRD)
IF (Pl:1.)ICATE) THEN (HOW_TO_DO_SOMETHING)

In general, the predicate depends on the current-sometimes even some past-activity as
well as the state or phase of the particular bird, e.g. its location. Some rules also take into









account activities of other birds of the same or of different species. For example, wading
birds often forage in mixed-species flocks that then attract other wading birds to forage
at the same location (Kushlan 1978). Rules that determine where a bird will forage must
therefore take into account the foraging activities of other wading birds, i.e. their location.
Tli, wading birds models execute using an event queue where the events are determined
by the duration of the activities of the birds. A simple example is a flight from the colony to
a feeding site where the duration of the flight is determined by the distance the bird between
the colony and the feeding site and the bird's flight speed which in turn depends on the kind
of flight the bird uses. Using a flight from a feeding site back to the colony as an example
we can assign attributes and methods to an object ., ,I'.l -ird in much the same way as in
Sec. 6.1. If we let one of the state variables (attributes) be the amount of energy the bird has
stored in its body (e.g. as body fat) fat and define two parameters, the flight speed of the
bird speed and the specific energy expenditure (i.e. energy expenditure per distance flown)
specenex, then we arrive at the following attributes and methods for bird flight covering a
distance d:

Attributes:

Variables: fat, speed,specenex
Static '.ldels: not defined.

:'. I !i I, l-

-Code: time = d speed
Dynamic .i, del: Decrease fat by d specenex

Ti, dynamic model for a flight changes the state variable fat of the bird after the flight
has been completed. Ti11 code method specifies the flight time time and therefore determines
when the next event, its arrival at colony, is going to occur at which the bird makes a decision
about its next activity such as relieving its mate from incubating the I ---- or feeding their
nestlings.
Various wading bird species can be incorporated into the OOP. .1 methodology in the same
as way as for general population and structured population models. Wood storks and white
ibis, for example, have different attributes such as weight and size, exploit different food
resources, and some of their behavioral rules are different. T111, plugi and pl operation is,
however, slightly more involved than for the previous model types as some behavioral rules
depend on the presence or absence of other wading species. .'., i. rtheless, a new sub-class
of wading birds, i.e. a new species, can be added to or deleted from the model without
changing the conceptual framework.

7 Model Abstraction: Problems and Issues

T!, OOP'. 1 methodology and its implementation represent a beginning in tackling the model
abstraction problem. Object Orientation, in general, even without the extension of model,
provides for a convenient vehicle for talking about abstraction and dealing with its man-
ifest issues. However, many problems remain to be solved if true model abstraction is to









be practical. Let's first consider what is possible in the implementation of OOP'.I T!11
crux of multimodeling is that it supports the hierarchical refinement of heterogeneous mod-
els through functional coupling. Each dynamic model is represented by the method (i.e.,
function) of a class. Within each model, there are the multimodel entry points defined in
Sec. 5.3. Tli, entry points differ for each model type. Tli, entry points permit coupling
in the traditional sense, thereby achieving a multi-level representation. Currently, '.iOOSE
does not support type checking to ensure that a function's output type exactly matches the
input type for another function's parameter; however, this problem is not an acute one.
Consider two dynamic models: A and B. Our methodology supports the basic case of B
be defined as part of A, or vice versa. It also supports the idea of identifying a behavioral
(but not necessarily structural) equivalent of B by using a neural network (Lee and Fi- li~ i 1:
1996) as the black box mechanism that captures the input-output relation to some arbitrary
degree of accuracy. However, '.iOOSE does not support deriving A from B if A and B are
independent. T111 derivability of models, computationally, is a much harder problem, and is
discussed in more depth in (Zeigler 1984)(ch. 13) and (Fi- '-i, 1: 1995)(ch. 8). Ultimately, we
would like to build a family of models, but it is not clear whether these models can be practi-
cally related since each model may have associated with it different underlying assumptions,
state spaces and parameters. State and event spaces, for instance, should not be arbitrarily
defined since they represent concepts in the minds of the modelers, and in the cognitive
representations of the physical system. Our approach is not necessarily to -, 1 the hard
problems under the uli." but to provide a path to the greater goal of true model simplifi-
cation. Heterogeneity in model creation and behavioral function simplication are stepping
stones toward this goal. For a comprehensive survey of the state of the art in the model
abstraction area, the reader is referred to a recent edited journal on the subject (Fi--l-I i 1:
1996c).

8 Conclusions

We have discussed a proposed method of multimodeling that has the potential to affect fu-
ture ATLSS research. Tli, attempt is to create a framework that helps with model design
and model integration. T! OOP'-. methodology has lead to a multimodeling implemen-
tation program called '.iOOSE ('. i !i iin dealing Object-Oriented Simulation Environment).
.iOOSE-created models (Cubert, Goktekin and Fi-hli~ i 1: 1997) can be used by ecologists to
visually design models which are automatically translated into computer code. This code
can then be integrated into the existing ATLSS legacy code framework. However, we see
OOP'.I and '.iOOSE as having the far more important effect of promoting model design
and integration of models. A key problem, which we are beginning to address, is how to
gradually phase-in visually structured models and phase out legacy code. This is a difficult
problem. It is not enough to assume that legacy code will be abandoned, since that is im-
practical. Instead, we are looking into new approaches that foster a more gradual transition
from legacy code to visual, object-oriented model.
T1i current state of '.iOOSE is that we have a working graphical user interface (GUI)
for the conceptual model and a subset of the dynamic model types (FSA,functional block
model). However, a considerable amount of time is being spent ,,iig up" the code to
make it more robust and bug-free. While it is premature to speculate whether OOP '. and









.iOOSE will solve all problems with legacy code integration, we feel that it is a step in
the right direction. Concerning new code development, the use of OOP'.i will be strongly
encouraged from the beginning so that the code develops from a strong model specification.
Code integration is always a difficult and time consuming task. Creating models within an
object-oriented framework provides the modeler with much greater control and representa-
tion. OOP'.I goes beyond the software engineering visual designs by explicitly organizing
knowledge about physical models. Even in the case where legacy code must be used without
any code rewriting, we have determined that 'O. IOOSE can still be used to organize the legacy
code even though no visual dynamic models are constructed initially.

Acknowledgments

We are grateful to two major groups: the '.iOOSE team and our ATLSS team partners. We
would like to acknowledge the graduate students of the '.iOOSE team for their individual
efforts in making lOOSE a reality: Robert Cubert, Tolga Goktekin, Gyooseok Kim, Jin Joo
Lee, Kangsun Lee, and Brian T1 iilyke. We also thank our colleagues on the ATLSS code
development team Jane Comiskey, Don DeAngelis, D. '.i lrtin Fleming, Louis Gross, .1, 11, I
Huston, Wolff '.,l oij, Phil Nott and Scott Sylvester. It is through this collaboration, that
we have learned of the importance of integrating multilevel models and made the proposal
that OOP'-.I serve as a potential integration vehicle for the Everglades.
We would like to thank the following funding sources that have contributed towards our
study of modeling and implementation of a multimodeling simulation environment for anal-
ysis and planning: (1) Rome Laboratory, Griffiss Air Force Base,' York under contract
F30602-95-C-0267 and grant F30602-95-1-0031; (2) Department of the Interior under grant
14-45-0009-1544-154 and the (3) :'-i-,1 i,11 Science Foundation Engineering Research Cen-
ter (ERC) in Particle Science and Technology at the University of Florida (with Industrial
Partners of the ERC) under grant EEC-94-02989.

REFERENCES

ATLSS 1997. Across-Trophic System Simulation (ATLSS): An Approach to Analysis of
South Florida Ecosystems. Technical report, Biological Resources Division, US Geolog-
ical Survey, South Florida/Caribbean Ecosystem Research Group, i.ii,1ini Florida.
Booch, G. 1990. On the Concepts of Object-Oriented Design. In .. P. A. and Yeh,
R. T., editors, Modern Software Engineering, chapter 6, pages 165 -204. Van Nostrand
Reinhold.
Booch, G. 1991. Object Oriented Design. Benjamin Cummings.
Caswell, H. and John, A. 1992. From the individual to the population in demographic mod-
els. In DeAngelis, D. J.and Gross, L. J., editor, Individual-based models and approaches
in ... ..I/,/ pages 36-61. C ,1111iniii and Hall,' York, N.Y.
Cellier, F. E. 1991. Continuous System Modeling. Springer Verlag.









Cubert, R. .1, Goktekin, T., and Fi-,-l -i,:, P. A. 1997. '.,!)ose: architecture of an object-
oriented multimodeling simulation system. In Enabling T. ,. l..I, for Simulation Sci-
ence, Orlando, Florida. Part of SPIE AeroSense 1997 Conference.
Dahm, C. N., Cumming, K. W., Valett, H. ':.1 and Coleman, R. L. 1995. An Ecosystem
View of the Restoration of the Kissimmee River. Restoration E ..'.I.', 3(3):229-238.
Dale, V. and Rauscher, H. 1994. Assessing impacts of climate change on forests: the state
of biological modeling. Cl'i,.,'' Cl,.j, '2', 1 5-90.
Davis, S. '.1 and Ogden, J. C. 1994. Everglades: T7, Ecosystem and its Restoration. St.
Lucie Press, Delray Beach, FL.
DeAngelis, D. and Gross, L. J., editors 1992. Individual-based models and approaches in
*.. '.i',., populations, communities, and ecosystems. C11i11iiiiin and Hall, '- ,, York.
DeAngelis, D. L. and Rose, K. A. 1992. Which Individual-Based Approach is '. i .- Appropri-
ate For a Given Problem? In DeAngelis, D. L. and Gross, L. J., editors, Individual-Based
Models and Approaches in E ..',',i, pages 67-87. C,111 ljii:,i and Hall, .: ,, York.
DeAngelis, D. L., Rose, K. A., B., C. L., '-.1i-, Iii!! E. A., and Lika, D. 1993. Fi-!i Co-
hort Dynamics: Application of Complementary '.1, dealing Approaches. TI, American
Naturalist, 1 1'(4):601 1. '
Fi-l, i, 1 :, P. A. 1991. Heterogeneous Decomposition and Coupling for Combined '.ideling.
In 1991 Winter Simulation Conference, pages 1199 -1208, Phoenix, AZ.
Fi- li- 1 i, P. A. 1993. A Simulation Environment for .1ill! i iideling. Discrete Event Dy-
namic Systems: TI .. and Applications, 3:151-171.
Fi-l,-,i, 1:, P. A. 1995. Simulation Model Design and Execution: Building Digital Worlds.
Prentice Hall.
Fi-lli- i1 :, P. A. 1996a. A Taxonomy for Simulation '.i dealingg Based on Programming Lan-
guage Principles. IIE Transactions on IE Research. Accepted for Publication.
Fi-l, -i, 1:, P. A. 1996b. Extending object oriented design for physical modeling. ACM Trans-
actions on Modeling and Computer Simulation. Submitted for review.
Fi-li- i, 1:, P. A. 1996c. Guest Editor for Special Issue on '.i, del Abstraction. Transactions
of the Society for Computer Simulation International, 13(4).
Fi-l- i, 1 :, P. A. 1997. A Taxonomy for Simulation '.i, dealing Based on Programming Lan-
guage Principles. IIE Transactions on IE Research. To be published in the latter part
of 1997.
Fi-l-, i, !:, P. A., :.- i,; .i.H11i, H., Sticklen, J., and Bonarini, A. 1994. '-.ii! iiii,,del approaches
for system reasoning and simulation. IEEE Transactions on Systems, Man and Cyber-
netics. to be published.
Fi-l,-i 1:, P. A. and Zeigler, B. P. 1992. A '.i 1 lil ii ,del '.1 l1,i ll. .1..gy for Qualitative '.i1bdel
Engineering. ACM Transactions on Modeling and Computer Simulation, 2(1):52-81.
Fleming, D., DeAngelis, D., and Wolff, W. 1995. Th1 importance of landscape in ecosystem
integrity: the example of Everglades restoration efforts. In VWestra, L. and Lemons, J.,









editors, Perspectives on Ecological I.!' .. '*, pages 202-217. Kluwer Academic Publish-
ers, 111, i! 'il li -
Forman, R. T. T. 1995. Land mosaics: the ... '.',.. of landscapes and regions. Cambridge
University Press, Cambridge, .i ii--, I!i- tts.
Forrester, J. W. 1969. Urban Dynamics. .ITT Press, Cambridge, '.iA.
Forrester, J. W. 1971. World Dynamics. Wright-Allen Press.
Fowler, '. and Scott, K. 1997. UML Distilled. Addison-XWesley.
Frohring, P. C., Voorhees, S. A., and Kushlan, J. A. 1988. History of wading bird populations
in the Florida Everglades: A lesson in the use of historical information. Waterbirds,
11(2):328-335.
Graham, I. 1991. Object Oriented Methods. Addison XWesley.
Gunderson, L, H., Light, S., and Holling, C. 1995. Lessons from the Everglades. Bioscience.
Supplement: Science and Biodiversity Policy S66-S73.
Hallam, T. G., Lassiter, R. R., Li, J., and KcKinney, W. 1992. '.i dealingg Populations with
Continuous Structured '. idels. In DeAngelis, D. L. and Gross, L. J., editors, Individual-
Based Models and Approaches in E .. '. .' pages 312-337. C1I1.iii i and Hall, :' i- -York.
Halloway, ':. 1994. 'i; llili nature. Scientific American, 270(4):98-108.
Hill, D. R. C. 1996. Object-Oriented Analysis and Simulation. Addison-XWesley.
Hunsaker, C., Ni-Il., I R., Lam, D., Browder, J., Baker, W., Turner, '.1 and Botkin, D. 1993.
Spatial models of ecological systems and processes: the role of GIS. In Goodchild,
.1 Parks, B., and Steyaert, L., editors, Environmental modeling with GIS. Oxford
University Press, Oxford, UK.
Kushlan, J. A. 1978. Feeding ecology of wading birds. In Sprunt, A., Ogden, J. C., and
Winckler, S., editors, Wading Birds, pages 249-298. N.i,',;.ini! Audubon Society, :'
York, N.Y. Research Report No. 7.
Kushlan, J. A. 1979. Design and management of continental wildlife preserves: Lessons from
the everglades. Biological Conservation, 15(4):281-290.
Lee, K. and Fi-lr- i, 1, P. A. 1996. A i-. !,i, ,i,!.gy. for Dynamic ':.Idel Abstraction. Trans-
actions of the Society for Computer Simulation International, 13(4):217-229.
Lima, S. and Zollner, P. 1996. Towards a behavioral ecology of ecological landscapes. Trends
in E .,.,.,/i and Evolution, 11(3):131-135.
.i\xxwell, T. and Costanza, R. 1995. Distributed '.Midular Spatial Ecosystem :,I dealing.
International Journal in Computer Simulation, 5:247-262.
.lii ly, J. D. 1990. Mathematical B'.*l.,.i, Springer Verlag.
Norman, D. A. 1988. T7, Design of Everyday Tl, .,:- Currency Doubleday, N' ,York.
Ogden, J. C. 1994. A comparison of wading bird nesting colony dynamics (1931-1946 and
1974-1989) as an indication of ecosystem conditions in the Southern Everglades land-
scape. In Davis, S. :'. and Ogden, J. C., editors, Everglades: the ecosystem and its
restoration, pages 533-570. St. Lucie Press, Delray Beach, FL.









Peterson, J. L. 1981. Petri Net TI, .d y and the Modeling of Systems. Prentice-Hall, Inc.,
Englewood Cliff- N.J.
Richardson, G. P. and Pugh, A. L. 1981. Introduction to System Dynamics Modeling with
DYNAMO. ilT Press, Cambridge, '.iA.
Rosenberg, R. C. and Karnopp, D. C. 1983. Introduction to Physical System Dynamics.
.1, Graw-Hill.
Rothenberg, J. 1989. Object-Oriented Simulation: Where do we go from here? Technical
report, RAND Corporation.
Rumbaugh, J., Blaha, ':. Premerlani, W., Frederick, E., and Lorenson, W. 1991. Object-
Oriented Modeling and Design. Prentice Hall.
Stroustrup, B. 1991. T7, C++ Programming Language. Addison Wesley, 2 edition.
T 111o.1i, J. 1975. Bond Graphs: Introduction and Application. Pergamon Press.
Ulanowicz, R. E. 1994. ':.1 dealing Lower Trophic Components of the Everglades Ecosystem.
Technical Report U. iCEES CBL 94-129, University of .liaryland System, C1i -,11, ,Lke
Biological Laboratory, Solomons, .11i) '11 s'S--0038.
Ulanowicz, R. E. and Platt, T. 1985. Ecosystem T '.... for Biological 0 .,.,.,i. ,,/Cana-
dian Bulletin of Fi-l11 ii, -. and Aquatic Sciences 213, Ottawa.
Vasconcelos, '.1 and Zeigler, B. P. 1993. Discrete Event Simulation of Florest Landscape
Response to Fire Disturbances. Ecological Modeling, 65:177-198.
Vasconcelos, '.1 Zeigler, B. P., and Pereira, J. 1995. Simulation of Fire Growth in GIS
Using Discrete Event Hierarchical .i dular '.i, dels. Advances in Remote Sensing, 4(3).
Wolff, W. F. 1994. An Individual-Oriented .,Iodel of a Wading Bird '.i -i-g Colony. Eco-
logical Modelling, 72:75-114.
Zeigler, B. P. 1976. Thi, A.--.--,, -.1i1,ii Problem. In Patten, B., editor, Systems Analysis and
Simulation in E ..'/',., Academic Press.
Zeigler, B. P. 1979. Mi uli-Level `-uli,.il iiiii-iii '.;1delling: An Ecoosystem Example. In
Halfon, E., editor, TI ...:. .i. Systems E .'/ ',., Academic Press.
Zeigler, B. P. 1984. Multifacated Modelling and Discrete Event Simulation. Academic Press.
Zeigler, B. P. 1' .,. lil i li1, ,ted Systems '., Ideling: Structure and Behavior at a i! ipli iily
of Levels. In Individual Development and Social CIL.,. : Explanatory Analysis, pages
265 -293. Academic Press.
Zeigler, B. P. 1990. Object Oriented Simulation with Hierarchical, Modular Models: Intelli-
gent Ai., l! and Endomorphic Systems. Academic Press.
Zeigler, B. P. and .oIon, Y. 1996. Distributed Systems: Speed versus Error Tradeoffs.
Transactions of the Society for Computer Simulation, 13(5):179-190.
Zeigler, B. P., Vasconcelos, '. and Pereira, J. 1994. Simulation of Fire Growth in '. I untain
Environments. In Price, '-. and Heywood, O., editors, C'... ~',.//'. Information Systems
in Mountain Environments. Taylor and Francis Publishing Company.









Biographies


Paul A. Fishwick is an Associate Professor in the Department of Computer and Informa-
tion Science and Engineering at the University of Florida. He received the PhD in Computer
and Information Science from the University of Pennsylvania in 1986. He also has six years of
industrial/government production and research experience working at i'); 'ort ', "- Ship-
building and Dry Dock Co. (doing CAD/CA '.1 parts definition research) and at NASA
Langley Research Center (studying engineering data base models for structural engineer-
ing). His research interests are in computer simulation modeling and analysis methods for
complex systems. He is a senior member of the IEEE and the Society for Computer Simu-
lation. He is also a member of the IEEE Society for Systems, l. iii and Cybernetics, A': i.
and AAAI. Dr. Fi-lr- i 1 founded the comp.simulation Internet news group (Simulation
Digest) in 1987. He has chaired workshops and conferences in the area of computer simu-
lation, and will serve as General Chl.iir of the 2000 Winter Simulation Conference. He was
chairman of the IEEE Computer Society technical committee on simulation (TCSIM) for
two years (1988-1990) and he is on the editorial boards of several journals including the
ACM Transactions on Modeling and Computer Simulation, IEEE Transactions on Systems,
Man and Cybernetics, T7, Transactions of the Society for Computer Simulation, Interna-
tional Journal of Computer Simulation, and the Journal of Systems Engineering. Dr. Fi-l-
wick's WWW home page is http://www.cise.ufl.edu/~fishwick and his E-mail address
is fishwick@cise.ufl.edu.
James G. Sanderson is a Visiting Scientist in the Department of Wildlife Ecology and
Conservation at the University of Florida. He received the PhD in '-1. 111!, I ii, from the
University of '. i1 \i; i, in 1976. He also has 19 years of industrial/government research
experience working on modeling and simulation at Los Alamos :'.--.ii1inil. Laboratory. His
research interests are in ecological modeling and simulation, population modeling, and avian
distributions. He is a member of the Ecological Society of America, and the Society for Con-
servation Biology. Dr. Sanderson's WWW home page is http://www.cise.ufl .edu/-jgs
and his E-mail address is jgs@cise.uf 1.edu.
Wilfried F. Wolff is a Visiting Scientist and Adjunct Professor in the Department of Bi-
ology at the University of '.i,iiini He received the PhD in Physics from the University of
Cologne, Germany, in 1981. He has held various visiting appointments at Stanford Univer-
sity, Xerox Palo Alto Research Center, University of Tennessee at Knoxville, and Oak Ridge
.;, in, 1f11 Laboratory. He is currently on leave of absence from the Juelich Research Center
(KFA), Germany, where he is senior scientist in the Department of Biotechnology 3. His
current research interests are in ecological modeling, in particular individual-based modeling,
and simulation. He is a member of the Ecological Society of America and his E-mail address
is wilfried@fig. cox.miami. edu.