Conceptual Modeling and Dynamic Simulation of Brownfield
Redevelopment
Todd K. BenDor
Department of Urban and Regional Planning
University of Illinois at Urbana-Champaign
111 Temple Buell Hall
611 Taft Drive
Champaign, IL 61820
Tel: 217-333-5172
Fax: 217-244-1717
Sara S. Metcalf
Department of Geography,
University of Illinois at Urbana-Champaign
Urbana, IL 61801
ssm@uiuc.edu
Abstract
The negligent upkeep of many abandoned industrial sites (“brownfields”) throughout the
twentieth century has had grave impacts on the urban landscape of American and European
cities. In recent years, brownfield redevelopment has come to be viewed as a strategy for
sustainable land use and urban revitalization. This study assesses the feasibility of the
construction of a dynamic simulation model of urban brownfield redevelopment. Literature
surrounding brownfield redevelopment is reviewed and used to construct a dynamic hypothesis
of brownfield redevelopment as it relates to site liability, economic viability, and availability of
redevelopment funding. Finally, an initial system dynamics model of the brownfield
redevelopment process is constructed. This quantitative analysis is performed using the 2003 US
Conference of Mayors brownfield survey, which serves as a dataset on brownfield distribution
and average site size. We conclude with suggestions for the extension of the model to capture
spatial feedback in order to assess redevelopment effects on the surrounding matrix of urban
land-uses.
Keywords: Brownfields, brownfield redevelopment, urban development, urban modeling, urban
planning, urban revitalization
Introduction
The negligent upkeep of many industrial sites throughout the twentieth century has had
grave impacts on the urban landscape of North American and European cities (De Sousa, 2001;
Harrison and Davies, 2002). Only during the last several years has the failure to reuse and
redevelop contaminated urban lands become a major concern for many municipalities. This
heightened awareness of the brownfield problem has occurred as estimates of the number of
brownfields in the United States have grown to between 500,000 and | million (Kretchik, 2002).
The mid-1980s saw a shift in planning and policymaking attention towards measures designed to
improve the quality of life in urban areas (De Sousa, 2001; De Sousa, 2003) with the United
States Environmental Protection Agency (U.S. EPA) creating a mission to “empower States,
communities, and other stakeholders in economic development to work together in a timely
manner to prevent, assess, safely clean up, and sustainably reuse brownfields (U.S. EPA,
2003b).” One aspect of urban revitalization garnering widespread political support has been the
redevelopment of under-utilized brownfield sites that are often located in dilapidated urban core
areas. In recent years, brownfield redevelopment has emerged as a sustainable land use strategy
and one of several ways to address urban sprawl and promote economic development through
new job creation (Thomas, 2002; Kirshenberg et al., 1997 Tam and Byer, 2002).
However, as we will show throughout this study, much of the literature on brownfield
redevelopment has highlighted a critical paradox created by state and federal legislation: past
decades have seen governmental attempts to revitalize contaminated urban lands into areas
beneficial to the surrounding community. While doing this, the same governments attempt to
ensure public health by enacting regulations that impose uncertain liability risks on individuals
interested in redeveloping a contaminated area. These conflicting actions have slowed the
possibility of revitalization in areas whose economic viability is already inherently in question
(otherwise they already would have been redeveloped) and undermined the original societal goal
of urban revitalization. This paradox was evidenced in a survey of 231 American cities in which
the most frequently identified barrier to redevelopment of brownfields was lack of clean-up
funds (82 percent), liability issues (59 percent), and the need for environmental assessments (51
percent) (US Conference of Mayors, 2003).
The project described herein was initiated as part of a set of Illinois state-funded ventures
‘sist in community and land-use planning issues in the St. Louis metropolitan area. These
projects include the East St. Louis Action Research Project (ESLARP) and the Land-use
Evolution and Impact Assessment Model (LEAM), among others. Several of these projects are
specifically focused on addressing East St. Louis, a city that has experienced significant
economic collapse over the last forty years and is the current focus of intense redevelopment
efforts and academic research by the University of Illinois at Urbana-Champaign (Reardon,
1995, 1998; Reardon and Shields, 1997). '
There are two major objectives of this study. Our first aim is to create a framework for
understanding the redevelopment process. We initiate this through an exploration of literature
looking at the barriers to urban brownfield redevelopment, followed by the construction of a
dynamic hypothesis incorporating these elements. Here, the hypothesis of system behavior is
dynamic since it tries to explain the dynamics characterizing the system in terms of the
underlying feedback mechanisms that control system or problem structure.
' East St. Louis lies on the Illinois side of the Mississippi River, directly adjacent to the City of St. Louis, MO.
Our second objective is to abstract this process through the use of the system dynamic
modeling methodology, which has the ability to incorporate feedbacks and delays into dynamic
models. Throughout this construction we utilize a number of probab: c elements to represent
the uncertainty underlying the many facets of the redevelopment process. We envision that the
model created herein will have the further potential to examine specific, location-dependent
policing relating to brownfield redevelopment.
In addition to policy testing possibilities, we recognize the importance of considering
social attitudes relating to urban revitalization and contaminated urban areas. We will explore
what we have identified as the major factors contributing to the likelihood of redevelopment,
while attempting to capture these aspects as part of a dynamic hypothesis of brownfield
redevelopment.
Literature Review and Background
As brownfield redevelopment has become a major part of urban revitalization as a whole,
a growing body of literature has formed around the brownfield redevelopment process. In 1997
the U.S. EPA promulgated a widely accepted definition of brownfields that defined them as,
“abandoned, idled, or under-used industrial and commercial facilities where expansion or
redevelopment is complicated by real or perceived environmental contamination (De Sousa,
2003).” The Comprehensive Environmental Response, Compensation, and Liability Act
(CERCLA) of 1980 (42 U.S.C 9601, sect. 101) defined brownfields as, “...real property, the
expansion, redevelopment, or reuse of which may be complicated by the presence or potential
presence of a hazardous substance, pollutant, or contaminant (EPA, 2003a).”
This definition characterizes an enormous number of properties as brownfields since the
severity of contamination is not specific. Characterizing brownfields using this definition also
permits areas containing perceived contamination to be classified as brownfields with the same
ease as areas containing documented contamination. Brownfields are often located in urban core
areas and industrial suburbs whose history has included intense periods of traditional
manufacturing (McCarthy, 2002). These sites can also include small commercial and residential
lots such as gas stations and dry-cleaners that are suspected of contamination.
U.S. law creates a distinction between extremely contaminated sites and sites possibly
contaminated with low levels of ordinary, non-hazardous waste. Hazardous sites are generally
governed federally by the Comprehensive Environmental Response, Compensation, and Liability
Act of 1980 (CERCLA). This law placed a tax on the chemical and petroleum industries and
provided broad Federal authority to respond directly to releases or threatened releases of
hazardous substances that posed a threat to public or environmental health. Over five years, $1.6
billion was collected for a trust fund targeted at cleaning up abandoned or uncontrolled
hazardous waste sites (commonly called the Superfund). One of CERCLA’s goals is to initiate
“long term remedial response actions, that permanently and significantly reduce the dangers
associated with releases or threats of releases of hazardous substances (EPA, 2003c),” on
contaminated sites that are placed on the National Priorities List (NPL). In 2002, there were
approximately 1300 Superfund sites on the NPL containing toxic waste or dangerous heavy
metals such as lead or mercury (McCarthy, 2002).
Most brownfield sites have relatively low levels of contamination when contrasted with
Superfund sites. As such, most sites are not governed directly by CERCLA, but rather fall under
the jurisdiction of state superfund laws often modeled after CERCLA that contain similarly strict
liability provisions (Kirschenberg, 1997). Each state has different types of cleanup standards,
procedures for identifying sites and provisions for apportioning liability.
Brownfield sites containing even small amounts of contaminants may still be extremely
challenging to remediate. In light of this, the benefits of redeveloping brownfields often do not
immediately outweigh the costs. McCarthy (2002) argues that brownfield redevelopment
presents a “dual policy challenge.” Barriers to private-sector redevelopment of brownfields must
be reduced while encouraging the connection of brownfield reuse to the broader goals of the
community. Uncertainties surrounding the first policy challenge in reducing the barriers to
redevelopment include uncertain liability provisions, cleanup standards, funding opportunities,
and legal regulations.
Barriers to Redevelopment: Liability Issues
Entities interested in redevelopment activities must be aware of the potential liability
problems associated with a brownfield site. In the past, U.S. law (both federal and state) has
held that anyone working with contaminated sites can be held liable for all cleanup costs, thus
prompting business owners and potential developers to avoid abandoned industrial sites, even if
contamination problems are relatively minor. Liability concerns are commonly believed to be
responsible for a diversion of capital away from brownfield redevelopment, thus limiting the
possibility of redevelopment (Wright, 1997, Tam and Byer, 2002).
Liability as a disincentive for redevelopment has been well documented by the banking industry,
where loan officers often require costly environmental assessments in order to assure that by
providing a mortgage the bank cannot be held liable under the same strict liability logic placed
on other developers (Rafson and Rafson, 1999). Under CERCLA (and many similarly worded
state laws), liability for property contamination may be imposed on the owner of property based
solely on his or her status as property owner. U.S. common law imposes strict liability where a
person undertakes an “ultra-hazardous” or “inherently dangerous” activity (Nurad, Inc. v. W.E.
Hooper & Sons, Inc., 966 F.2d 837, 1191, 10" Cir. 1997). Courts have concluded that storage of
hazardous substances can constitute an inherently dangerous activity. Therefore, under
CERCLA, “a property owner may be held responsible for remediation of property even if the
environmental condition was in existence prior to the current owner’s purchase of the property
(Rafson and Rafson, 1999, pg. 10). Concurrently, CERCLA imposes joint and several liability
on individuals or entities identified as partial brownfield owners. This type of liability holds that
where the conduct of two or more persons combines to create an indivisible harm, either
defendant can be held responsible for the entire harm (Wisconsin Natural Gas Co. v. Ford, Bacon
& Davis Construction Corp., 291 N.W.2d 825, Wis. 1980). Thus, the fear of liability
surrounding brownfield redevelopment is compounded. An individual who may not have been
responsible for contamination may be forced to assume the entire burden of the cleanup cost with
no outside assistance.
A 1990 American Bankers Association survey found that 62.5% of U.S. lending
institutions had rejected loan applications based purely on the possibility of environmental
liability (Byrne and Greco, 1994). Liability concerns have also forced many industrial site
owners to stop placing old sites on the market in order to avoid discovery of contamination that
might force them to initiative an expensive remediation program. Tam and Byer (2002) develop
a flexible decision methodology looking at the preferred remedial action and future use of
contaminated sites from the perspective of site owners. This methodology focuses on the
creation of a cost/benefit analysis of various future land-uses and remediation methods which
attempts to maximize site value, while minimizing liability and remediation cost. Although this
decision methodology delves more into more detail in terms of the quantification of liability and
the resolution (site scale) that is much more fine-grained and agent-based than we explore in this
study, it holds the promise of being an extremely important component of more sophisticated
future models of urban brownfield redevelopment.
The 1990s saw an effort by many state and federal agencies to lower the role that liability
plays in slowing brownfield redevelopment. Since most brownfield sites are not contaminated to
a point that would warrant federal involvement, many states have responsibility over brownfield
programs. Many states have created State Voluntary Action Programs (VAPs) intended to allow
private parties to voluntarily investigate and remediate a property while receiving some level of
protection from future state enforcement action (Wright, 1997; McCarthy, 2002). On the federal
level, the EPA has attempted to calm fears of federal action by entering into Superfund
Memoranda of Agreement (MOAs) with individual states. MOAs create agreements to refrain
from direct federal involvement at sites utilizing VAPs. The critical problem from a federal and
state governmental perspective is achieving a liability balance which encourages redevelopment
while ensuring that sites are adequately remediated and do not pose a danger to public and
environmental health.
Barriers to Redevelopment: Uncertain Cleanup Standards
Uncertainty surrounding site assessment considerations can also represent a major barrier
to redevelopment based on the lack of straightforward site remediation standards. State and
federal regulators can require remediation at many different levels depending on the anticipated
future use of the site. Sites whose future use will be restricted to industrial functions may require
less remediation than sites slated for residential or open space use by young children or other
individuals that may be endangered by very low levels of contamination. Contaminated
groundwater often requires remediation to drinking water standards before it leaves the property
and enters wells used for drinking, showering, and cooking (McCarthy, 2002). This failure of
state and federal agencies to determine widespread cleanup standards and the complex and
interconnected legal nature of federal and state agency involvement in redevelopment
coordination can form a major barrier to brownfield redevelopment.
Barriers to Redevelopment: Availability of Funding
Another major barrier to brownfield redevelopment is derived from the uncertainty
surrounding the possible cost of environmental assessments and remediation. Significant
financial investment is often required in order to remediate areas to the level required by law.
Thus, many private investors require stance from lending institutions, insurance firms and
government agencies (McCarthy, 2002). In fact, of the 12 major federal brownfield
revitalization programs reviewed by the Interdisciplinary Environmental Clinic (2003) at
Washington University in St. Louis, every single one involved the creation of grant or loan
programs to assist with brownfield redevelopment. Many individual state and local governments
also attempt to provide grants, loans, loan guarantees and tax credits to stimulate redevelopment.
With the increase in redevelopment funding made available to investors during the mid-1990’s
came an increase in the number of funding organizations — increasing the complexity in securing
loans and grants for redevelopment. Heightened complications in securing funds have also
confronted investors in recent years through the regulatory system that governs the
redevelopment of brownfields in different places throughout the country, thereby making fund
acquisition complicated from both a local and regional perspective.
Barriers to Redevelopment: Complicated Network of Regulation
Compliance with federal, state and local remediation regulations can involve substantial
time and financial costs, thus creating another major barrier to redevelopment. This process has
also been complicated by a lack of information or database integration among the different levels
of regulatory oversight agencies. Most local and state GIS programs have not integrated
brownfield maps into their databases or websites. In a survey of 23 cities, 57% indicated that
they held brownfield redevelopment partnerships with their county or state governments (US
Conference of Mayors, 2000). These partnerships are not necessarily archived in any widely
accessible database that would facilitate regional brownfield redevelopment. One suggested
solution to the complex regulatory framework that has emerged is the establishment of local
brownfield redevelopment authorities that could act as major points of contact for information
and financial assistance (Borak and Meek, 1999; Papper, 1997). However, disagreement exists
as to whether or not this entity would function in the private or public sector (McCarthy, 2002).
Brownfield Redevelopment and the Community
Although brownfields can be found almost anywhere, they commonly occur in the urban
core areas of major cities. Observations of the patterns of growth across the nation have shown
an exodus of capital and population away from major downtown areas during the last several
decades (Interdisciplinary Environmental Clinic, 2003; Simons, 1998; Wright, 1997). Much of
the downtown areas of major industrial cities such as St. Louis have suffered significantly as
relocating entities settle in suburban “greenfields” which are easily cleared for new development
and contain no actual or perceived contamination. Viewed in this frame, brownfield
redevelopment can be perceived as one of several methods currently being sought by many
historically industrial cities in revitalizing the economic and environmental health and viability
of the urban core (Wright, 1997). McCarthy (2002) argues that the pattern of redevelopment of
these contaminated lands therefore must be connected to broader community goals in order to
revitalize inner cities. Any attempts to remove the aforementioned barriers to redevelopment
must avoid conflict with the ever-present need to protect the environmental and economic health
of local residents. McCarthy (2002) is arguing here that governmental responsibility necessitates
studies of the future marketability of brownfield sites as well as a social cost-benefit analysis of
brownfield redevelopment coinciding with the ongoing development of participatory dialogues
with the community.
Another issue relating to community goals is the allocation of brownfield redevelopment
funds. Should redevelopment funds be focused on areas that suffer from higher contamination or
should they be channeled specifically to brownfields in good locations that are more likely to be
economically viable? McCarthy (2002) discuss common strategy in which redevelopment
funds are channeled towards high profile brownfields with strong prospects for successful reuse
in an attempt to trigger a “domino effect” of revitalization. However, this strategy neglects the
large number of non-economically viable sites in dilapidated communities where many
brownfields have little or no market value and negative images of crime drive a vicious cycle of
economic hardship.” The practice of avoiding the lowest market value parcels commonly
excludes disadvantaged neighborhoods from redevelopment programs. This behavior may
widen inequalities between wealthier and poorer neighborhoods thus undermining the basis of
inner-city revitalization (Leigh, 2000). Given this situation, there are currently few strategies for
remediating and redeveloping economically non-viable sites even if site redevelopment is
socially desirable (Leigh, 2000).
The Redevelopment Process
The EPA has constructed an outline of the brownfield redevelopment process which the
progression into Five major phases: Site Assessment and Due Diligence (Phase I), Site
Investigation (Phase II), Development of Remedy Plan, Remedy Implementation, and
Redevelopment Activities (Construction) (EPA, 2002). For our purposes, we will be assuming
that the development of a plan of remediation will take place during the Phase II site
investigation. The brownfield remediation process is typically performed by highly skilled
environmental consulting agencies with experience in dealing with dangerous contaminants.
Typically, the brownfields redevelopment process begins with a Phase I site sment.
The site assessment process provides an initial screening to explore owner records and site
history, extent of contamination, and possible legal and financial risks. If no apparent
contamination and no significant health or environmental risks are revealed, redevelopment
activities may begin immediately. If the site appears to contain unacceptably high levels of
contamination, a reassessment of the project’s viability may be necessary. Phase I site
assessment also includes the performance of due diligence to look at “preliminary cost estimates
for property purchase, engineering, taxation and risk management (EPA, 2002, pg 8). Due
diligence also involves a study of the market viability of the redevelopment project.
A phase II site investigation involves chemical sampling of the site in order to provide a
comprehensive understanding of the contamination. If this investigation reveals no significant
sources of contamination, redevelopment activities may commence immediately. If the phase IT
site investigation reveals a manageable level of contamination it is next possible to evaluate
possible remedial alternatives. If no feasible remedial alternatives are found, the project viability
must again be reassessed. Otherwise, the next step is to select an appropriate remedy and
develop a remedy implementation plan. The discovery of additional contamination following
remedy implementation would necessitate the reenactment of the remediation process.
Methodology
Scope and assumptions
Based on the literature outlined above, we began the process of constructing a model
framework. At the core of the model, we wanted to represent the process of redevelopment as it
? This type of positive feedback has become prevalent in many cities such as East St. Louis, IL where negative
perceptions surrounding crime and economic viability have caused an exodus of social and financial capital which,
over time, has amplified negative perceptions of crime and economic hardship, leading to a further drain of capital
(Reardon, 1998; Interdisciplinary Environmental Clinic, 2003).
occurs through different phases and delays. In addition to this core structure, which would
reduce the stock of existing brownfields over time, it is important to capture some of the key
barriers to redevelopment. We have identified funding for redevelopment, liability of
redevelopment, and economic viability of the site as key factors influencing the probability of
redevelopment for a given site. Moreover, redevelopment of brownfields at a given locale would
in turn influence those factors for surrounding brownfields in the form of spatial feedback.
We chose to represent the brownfield redevelopment process at the aggregate level,
utilizing data from a recent survey of over 200 cities across the nation (US Conference of
Mayors, 2003). The city-specific estimates were treated as samples indicative of parameter
distributions for average site size in acres, the fraction of redeveloping acres at a given point in
time, and estimated tax and job benefits from redevelopment. While considering the aggregate
representation of the process to be most appropriate, utilizing the distribution of data enables
representation of site-specific processes.
We identified three primary research questions: What are the most important factors
influencing brownfield redevelopment? What are the fundamental behavioral modes of
brownfield redevelopment? How should we represent delays in the system? While the model
inherently abstracts the redevelopment process as it takes place in reality, we believe that any
dynamic model of brownfield redevelopment should attempt to answer these questions.
Modeling Techniques
Modeling brownfield redevelopment is complicated partly because the process
simultaneously involves multiple system components such as site remediation, permitting,
liability, funding, and economic viability. However, real complexity in this system emerges as
system components dynamically interact and occur over time. In many systems where long-
term studies or experimental manipulations are not possible, which is often the case in complex
urban and economic systems, representative models have been shown to be helpful in filling
knowledge gaps and assisting in decision-making and policy formation activities (Sterman,
2000). The extensive literature supporting system dynamics as an aid to cognitive proc and
comprehension presents this methodology as an ideal technique for enhancing our conceptual
understanding of the brownfield redevelopment process.
Dynamic Model Structure
Before beginning the model-building process, we developed a representation of our
dynamic hypothesis as a causal map of the major feedback loops, or circular chains of causation
(Figure 1). Such feedback loops may be either reinforcing or balancing in nature, depending on
whether a variable, if increased initially, will be further reinforced or balanced after the ripple
effect of the other variables in the loop reaches it. This causal map hypothesis is a tool for
conceptualizing the system. Only the key variable relationships are highlighted in this
representation. Specifically, reinforcing dynamics of economic viability (or more precisely, the
lack of viability) in areas proximate to brownfields are highlighted, as are the reinforcing effects
of perceived contamination. Perceived contamination increases fear of liability, while funding
alleviates it. These effects are combined in the presence of laws that may positively or
negatively impact liability.
Figure 1. Dynamic Hypothesis
,, Peroeived and Potential
77” Contamination Level aan
/ \ Laws
Fear of Liabllity
Economie Viability _
of Adjacent Areas
romaine | ol an
ne
Brownfields Probability
[' Delays te Redevelop
“pect Redevelopment wa gece - Senitue
werse Relationship
Redeveloped
Land
This dynamic hypothesis (Figure 1) illustrates state variables such as funding,
brownfields, and redeveloped brownfields. For simplicity of illustration, details of the many
system delays and the impact of brownfields on tax base are not shown in Figure 1. These are
exposed in the sections that follow.
In contrast to the core dynamics described in Figure 1, an overview of the model structure
is provided in Figure 2 to illustrate relationships among the major sub-sectors of the model. The
three sectors of economic viability, liability, and funding all influence the probability of
redevelopment, and funding in turn influences liability of redevelopment. The subsequent
spatial relationship effects are illustrated with dotted lines in this diagram. The site ssment
and redevelopment sector follows from the probability of redevelopment and in turn influences
the economic viability and funding for redevelopment. Site assessment and redevelopment also
influences the resultant tax base in terms of the jobs created from the redevelopment process.
Figure 2. Model Sector Overview
Economic Liability of Funding for
Viability Redevelopment Redevelopment
snes
Probability of
Redevelopment
cr--l-r st
| Site Assessment |
Hs and a
| Redevelopment |
lananh Gime retreat =
———"_ TaxBase_!
The detailed structure of the site assessment and redevelopment sector is illustrated in
Figure 3 as composed in the STELLA modeling software. This sector provides the core of the
model in the form of an aging chain structure. The process of brownfield redevelopment is
depicted as flows between stocks from left to right, with the undeveloped stock of brownfields
at the left-most side. This stock is increased in one of two ways. The first inflow is brownfield
generation (in our base case this is zero, as we do not concentrate here on the dynamics of
brownfield generation or classification). Another inflow is the recycling back of rejected
brownfields from the phase I process, in the case that contamination is too high for the
stakeholders to manage. Demand for brownfield redevelopment provides the outflow from the
brownfields stock and is influenced by the stochastic parameter of initiating redevelopment,
which can be zero or one. If this parameter is one, an amount will be extracted based on the
existing brownfield stock in acres, multiplied by the redeveloping acre fraction.
Figure 3. Site Assessment and Redevelopment Sector
Fraction with Addtion
Contamination
Redeveloping Aere Fraction &
‘Additional Contamination Found
Phase 1 frase? Redeveloped Ae
smplatoe Completion Construction Redevel res
ott comeletion rrr Se
2Phaze hase
- ire fein
BF Generation Demand for BF Remedy implementation
Redevelopment sum of Rajacted
Fractions
Fraction sites suffering Clean Site Construction
weg 4
povcved outa bey re
Redevelopment
Phase 1
Rejection Rate
Sum of Rejected
Fractions High Contamination
Phase 1 Rejected | /“ Construction on Clean Sites
Fraction Sites with
High Contamination
Developing from Clean Sites
The redeveloping acre fraction is a probabilistic exponential distribution based on the
data from the 2003 US Conference of Mayors Survey. The form of this distribution was verified
using the @RISK software (Figure 4). Including the zero fractions, beta (analogous to the
mean) was close to 11% of total brownfields in redevelopment. We further accounted for the
distribution of acres in redevelopment among the different phases, according to the expected
delay times in each phase. We used a time unit of months for the model. As the average time to
redevelopment is around 3.5 years, we divided the exponential distribution of redeveloping acre
fraction by 42 months to convert it to units of acres per month (EPA, 2002).
> @RISK is a risk analysis and simulation plug-in used in the Microsoft Excel spreadsheet software that specializes
in modeling uncertainty through the use of sophisticated probability distribution functions (Palisade, 2004)
10
Figure 4. Distribution of Redeveloping Acre Fraction by City
Expon(0.11261) Shift=-0.00087291
12 a 7 y
Frequency (Cities)
Redeveloping Acre Fraction
C= TT se
0.0049 0.3365
To keep the order of redeveloping sites consistent, we utilized the conveyor form of stock
representation for each of the phases of redevelopment. Conveyor stocks allow for discrete
accumulating and output, unlike normal reservoir-type stocks that assume continuous
accumulating and output (which would imply that some small fraction of a site is being assessed
at any given time). Here, the conveyor allows for a first in, first out (FIFO) behavior, with
allowance for leakage and for probabilistic delay times. The duration in Phase I (site
assessment) was assumed to be an exponential distribution of a delay time of 2 months. The
initial acres in Phase I was also determined using the relative delay and the total redeveloping
acres at the time of the 2003 US Conference of Mayors Survey. As the site assessment process
is the initial introduction of uncertainty about potential contamination, we enabled a rejection
rate for sites that were either too highly contaminated to address with the given resources, or for
sites that were found to be uncontaminated and ready to redevelop. The bulk of the brownfield
acres (90%) would continue on to phase II (site investigation), with an exponentially distributed
delay time of 12 months, and then to the implementation of remedy (US EPA, 2002). This third
phase is inclusive of the development of a remedy plan, remedy implementation, and
redevelopment activities. We have assumed an exponentially distributed delay time of 30
months for the third phase, after is a stock of redeveloped acres (US EPA, 2002). Because the
third phase includes so many activities, we also allow for 5% recycling back to phase one if
additional contamination is found.
We have started with the overall process of redevelopment that is at the heart of this
construction. The input to this sector that ties it in with other sectors is initiating redevelopment,
the probabilistic parameter that could either be zero or one. The drivers of this parameter are
outlined in Figure 5. At the top center of Figure 5 is the probability of redevelopment sector.
11
The initiating redevelopment parameter is influenced by the probability to redevelop and a “roll
the dice” random parameter returning a value between zero and one. If the probability to
redevelop is greater than this random number generator, then initiating redevelopment will be
one. This construction is one way of explicitly representing stochasticity in the redevelopment
process and has been applied to land-use models such as the Land-use Evolution and impact
Assessment Model (LEAM) (Deal, 2001, Deal and Schunk, 2004).
Figure 5. Drivers of Redevelopment
Initiating Redevelopment
Roll the Dice
Probability to Redevelop
Ezononmic Mability
of Brownfield iability.
Marketabilty :% Demand for BF
is there Adequate Funding) Bedavenemert
Brownfield -‘EPonomio Mabilty
Redevelopment of SurTounding Area
Delay in Increasing
Surounding Eoonomic abilty \ aca rr
Decrease in Brownfields
* Funding
Funding Inflow L__
Funding Needed (4 ‘nding Expenditures
per site
COFunding Needed Per Acre
Po Average Site Size
The probability to redevelop is not a true probability in the mathematical sense, but it is a
measure of the likelihood of redevelopment. As mentioned before and illustrated in Figure 5, it
is positively influenced by the availability of funding and the economic viability of the
brownfield site, but is negatively influenced by liability. The funding sector is represented by a
stock with inflow at $200 million over 12 months, as per the 2001 Brownfields Revitalization
and Environmental Restoration Act. Funding expenditures are drawn down by calculated
demand for brownfield redevelopment (acres/month) multiplied by the funding needed per acre.
The average site size in acres is used in conjunction with this funding requirement to determine
whether there is adequate funding to begin with. Average site size is an exponential distribution
of 16.7 acres/site, also based on the US Conference of Mayors data.
The economic viability sector is also represented in Figure 5. As brownfields are
redeveloped, a decrease in brownfield acreage is recognized after a delay. This proportionate
decrease positively affects the economic viability of the surrounding area and combines with
marketability of the site to affect the economic viability of further brownfield redevelopment.
12
Figure 6. Tax Base Impact of Redevelopment
Jobs per Acre
Tax base increase Taxhase loss
Tax Base
from redevelopment fram outmigration
Job Creation from Mlgraten
Taxes P
me tersion Redevelopment Rate
Annual Taxes Redeveloped Acres Brownfields
per Job =
EXPRND($5470)
Jobs per Acre =
EXPRND(10.39)
$+ $S.on0 $1,000 415.000 $20,000 $25,000 430,000 438.000 $4n0n 45.000 $50,000
With this understanding of the construction of redevelopment drivers, we now turn to the
impact of redevelopment on the tax base. Figure 6 illustrates the tax base sector structure and
the relevant parameters. Redeveloped acres translate to job creation through the jobs per acre
parameter, derived from the US Conference of Mayors (2003) as an exponential distribution
around 10.4 jobs per acre. The number of jobs created is translated into a tax base inflow
through the parameter of annual taxes per job, exponentially distributed around $5470/job per
12 months. This inflow is integrated into the overall tax base, while accounting for the
possibility of outflow if residents migrate due to the presence of brownfields in the area. Again,
as brownfields decrease, the migration rate would decrease as well, further bolstering the tax
base.
Results and Discussion
Because several of the parameters in our model construction were represented by
probabilistic exponential distributions, the first step in our analysis of results was to repeat
several model runs to gain a sense of the inherent uncertainty in the model. The results from this
sensitivity analysis are illustrated in Figures 7 and 8 for redeveloped and remaining brownfields,
respectively. For the redeveloped brownfields, we see at the outset that a very wide range of
possible outcomes emerges. The average, maximum, and minimum redeveloped acreage for 25
stochastic scenarios at each point in time are measured along the left axis. The standard
deviation of these scenarios is shown with a dotted line and measured on the right axis. The
quantity of redeveloped brownfields begins at a value just above 10,000 acres, as initialized from
the US Conference of Mayors (2003) data, and rises to a still increasing average in 2030 of
around 20,000 acres, with a range from 16,000 acres to 22,000 acres. The standard deviation
peaks shortly after the 5" year for this sample of runs, and then plateaus with a slight overall
decline as the base of redeveloped brownfields increases. In contrast to the redeveloped
13
brownfields, the same tests for remaining brownfields indicate an increase in the standard
deviation over time as the base of remaining brownfields decreases (Figure 8). The quantity of
remaining brownfields declines from the initial value of 425,000 to an average of 413,000 acres,
with a range from 411,000 to 416,000 acres.
Figure 7. Uncertainty in Redeveloped Brownfields
— 20,000 4o00
7
FF
& g
< 7 5
2 Wax a
& 15,000 average sooo &
e
§ 2
8 <
ao =
3 10,000 2a, nat RO ES 2000 §
2 . — te 3
2 . Standard Deviation a
$ : 3
3 ‘ ty
3 : ee
ir 5,000 i 1000
a 0
2000 2005 2010 2015 2020 2025 2030
Figure 8. Uncertainty in Remaining Brownfields
425,000 4oo0
420,000
3000
415,000 2000
Remaining Brownfields (Acres)
(seuoy) uonelneg puepueys
410,000 1000
405,000 ao
2000 2005 2010 2015 2020 2025 2030
While the range of uncertainty is substantial, the direction of the progress made is
consistent. Moreover, given that the model scope encompasses the nationwide stock of
brownfields, we had no reason to suppress the compounding stochasticity of this system. With
this grounding in the probabilistic nature of our model as it represents what we hold to be a
14
probabilistic real-world system, we proceeded to test other aspects of the model by assigning
seed values to each random element. The seed values enable reproducibility of simulation runs
while retaining the random variable representation. Figures 9 through 12 represent relevant
variable results for a single run using the same set of seed values for each figure. The variables
charted in Figure 9 represent the acres in each phase of redevelopment. The stock of brownfields
decreases with increasing redeveloped acres in this single run similar to the multiple runs
described above. The acres in the different phases of redevelopment show less directionality,
with phase I acres not accumulating much with its short delay time. The subsequent phases
reveal the effect of uncertainty as it compounds through the system, with increased variability of
acreage levels at later stages in the run.
Figure 9. Phases of Brownfield Redevelopment
—— Redeveloped Acres
soom Remedy Implementation
ome Phase 2
hase 1
Acres
2000 2005 2010 2016 2020 2026 2030
Working upstream through the system, Figure 10 displays the probability of
redevelopment and the subsequent demand for redevelopment. Again, the actual demand for
redevelopment is mediated by the intervening variable of initiating redevelopment as the roll of
the virtual dice. Moreover, its magnitude also varies with the exponential distribution of the
redeveloping acre fraction. We see that the interaction of these two variables is such that the
probability remains non-zero for several dice rolls until demand for redevelopment is initiated, at
which point the available funding for further redevelopment is likely to be insufficient in the
near-term and so the probability drops to zero until funding is restored. These interactions
produce highly variable or “jagged” results when considered on the scale of a month-to-month
basis, but represent the real constraint of funding availability and the uneven nature of project
starts.
15
Figure 10. Probability Driver and Subsequent Demand for Redevelopment
1 ono
ae --- Probability to Redevelop 3600
-—— Demand for BF Redevelopment
8000
2600
#4 2000
05
1600
Probability to Redevelop
yuauidojenepey 10) pueweg
1000
600
+ L oO
2000 ©2005-2010 «2015 «= 2020-2028 «= (2030
The erratic nature of demand for redevelopment can be seen rippling through the phases
of the system in terms of completion rates. Unlike the stocks, which accumulate net flows, the
completion rates are flows themselves and are represented with probabilistic exponential
distributions of delay times. Figure 11 illustrates the completion rates for phase I and phase II,
as well as the construction rate corresponding with the final phase of redevelopment. The phase
I completion rate is the input to the subsequent phases. While there is some evidence of the
ripple of a spike in acreage, it is difficult to pinpoint due to leakage, the longer delays of
subsequent phases, and the inherent randomness of the system.
Figure 11. Completion Rates for the Redevelopment Phases
4000
3200
----Phase 1 completion
$68 Phase 2 Completion
—— Construction i
200
20nn
+500
Redevelopment Rate (acresimonth)
+00
500
0 |
2000 2005 2010 2015 2020 2025 2030
Finally, we examine the impact of the redevelopment process on the tax base sector and
funding level (Figure 12). The funding periodically runs out as recognized above after
16
expenditures on redevelopment. The tax base increases over the long run, with substantial
variability due to its occurrence at the end of compounded uncertainties, as well as the
uncertainty in the two key parameters of jobs per acre and tax base per job as described earlier.
Figure 12. Impact of Redevelopment on Tax Base and Funding
an 0
os 70
~—Tax Base
Funding
Tax Base (SMM)
(ws) Bupung
In summary, the results have demonstrated that the uncertainty we aimed to represent is
adequately represented. There is an intuitive nature to this model. The brownfield
redevelopment process decreases the stock of brownfields, increasing redeveloped area, which in
turn increases the tax base in the model by both bringing in jobs to the site, and slowing the
exodus of people away from the area that surrounded the former brownfield.
In this model, we are able to observe the interaction of uncertain elements over time in
order to see the ripple effect for downstream phases in the process. Regarding the relevant
factors in redevelopment, we are able to model the fact that funding is a key factor, as it
periodically constrains development activities. Here, our objective was to create a framework
into which greater exploration of specific issues such as liability and land-valuation can be
placed.
There are opportunities for further scenario testing with this model, as there are many
more aspects of the system to digest. Moreover, we want to be sure that the random element of
redevelopment probability represents the real discrete project-based development activity and is
not outweighing other factors.
Conclusions and Implications for Further Research
Many of the conclusions we can draw from the literature involve the need for shifting
redevelopment incentives away from greenfields and towards urban core brownfields. However,
the connection of brownfield redevelopment to broader community goals that is needed in order
for this to be realized is difficult because it involves non-economic factors that can be difficult to
quantify for the type of cost-benefit analysis often utilized for planning purposes. The social
costs and benefits of brownfield redevelopment relate to issues of environmental justice,
environmental quality and regional land use, issues that local governments cannot completely
control. Overall, brownfield redevelopment is a regional endeavor to ameliorate historical
17
problems that have plagued inner cities for decades. By classifying brownfield redevelopment as
part of smart growth initiatives and other planning issues current in the spotlight, funding for
such revitalization purposes may become more readily available.
The implications of further research on this model are wide ranging. Using this type of
model as a framework for policy analysis may hold the key to many promising avenues of
research and policy ideas.’ Furthermore, explicit quantification of various aspects of economic
viability, marketability and liability (especially as discussed in Tam and Byer, 2002) as they
implicitly occur in public and private cost-benefit analyses of redevelopment projects is an
important step in more accurately theorizing a general process by which redevelopment occurs.
This model looks at brownfield redevelopment without spatial context. Further work on
this topic must look at the spatial extension of this model to real urban areas. This aspect would
allow for the full implementation of the feedback loop in Figure | involving the relationship
between economic viability and surrounding brownfields. The dashed line in Figure 2 also
represents the important role that spatial feedback plays in this system. Furthermore, additional
research could determine the efficacy in extending this type of model using agent-based
simulation techniques in which individual brownfield sites could be tracked through the
redevelopment process. By studying the spatial configuration and effects of brownfields and
redevelopment activities alongside the view of brownfields as agents in the larger urban system,
the ability to add functional knowledge to the debate on policies for brownfield redevelopment
may be increased, thus bringing us closer to an overarching goal of improving our urban spaces.
‘ Examples of these include the implementation of a municipally controlled land bank that would initiate brownfield
redevelopment soon after a property underwent tax foreclosure (see Betancur, et. al. (1995) and Leigh (2000) for
more information).
18
Bibliography
Betancur, J., M. Leachman, A. Miller, D. Walker, and P. Wright. 1995. Development without
Displacement Task Force Background Paper. The Chicago Rehab Network. [Online].
Available: http://www.uic.edu/~pwright/dwd.html. Accessed: 11-7-03
Borak, D., and C. Meek. 1999. Putting the Pieces Together. International City/County
Management Association: Washington, DC.
Byrne, John. J and Greco, Thomas J. 1994. “Superfund Reform Needed to Keep Credit
Flowing.” American Banker, March 22, 10-11.
Deal, Brian. 2001. Ecological Urban Dynamics: The Convergence of Spatial Modelling and
Sustainability. Building Research and Information, 29: 381-393.
Deal, Brian and Daniel Schunk. 2004. Spatial Dynamic Modeling and Urban Land Use
Transformation: A Simulation Approach to Assessing the Costs of Urban Sprawl.
Ecological Economics, 51:79-95
Interdisciplinary Environmental Clinic. 2003. Brownfield Programs: Federal, Illinois and
Missouri Tools for Revitalization. Washington University in St. Louis School of Law
Interdisciplinary Environmental Clinic. Prepared for the East-West Gateway
Coordinating Council. Submitted at “From Brown to Green Again!” workshop, February
24, 2003. [Online]. Available:
http://www.ewgateway.org/pdffiles/library/ddrc/MajorBrownfieldsRevitalizationProgram
s.pdf Accessed: 10-12-04.
De Sousa, Christopher. 2001. Contaminated Sites: The Canadian Situation in an international
Context. Journal of Environmental Management, 62: 131-154.
De Sousa, Christopher. 2003. Turning brownfields into green space in the City of Toronto.
Landscape and Urban Planning, 62:181-198.
Ford, Andrew. 1999. Modeling the Environment: An Introduction to System Dynamics
Modeling of Environmental Systems. Washington D.C: Island Press.
Harrison, Carolyn, and Gail Davies. 2002. Conserving Biodiversity that Matters: Practitioners’
Perspectives on Brownfield Development and Urban Nature Conservation in London.
Journal of Environmental Management, 65: 95-108.
Kirshenberg, Seth, William Fischer, Charlie Bartsch, and Elizabeth Collaton. 1997.
Brownfields Redevelopment: A guidebook for Local Governments & Communities.
International City/County Management Association: Washington, DC.
Kretchik, Joe T. 2002. Brownfields. Chemical Health and Safety 9,2: 34
19
Leigh, Nancy G. 2000. Promoting More Equitable Brownfield Redevelopment: Promising
Approaches for Land Banks and Other Community Land Development Entities.
Working Paper. Lincoln Institute of Land Policy: Washington, DC.
McCarthy, Linda. 2002. The brownfield dual land-use policy challenge: reducing barriers to
private redevelopment while connecting reuse to broader community goals. Land Use
Policy, 19:287-296.
Palisade. 2004. @RISK Fact Sheet. Palisade Corporation. [Online]. Available:
http://www. palisade.com/htmlI/risk/facts.asp. Accessed 10-18-04.
Pepper, E. 1997. Lessons from the Field: Unlocking Economic Potential with an Environmental
Key. Northeast-Midwest Institute: Washington, DC.
Rafson, Harold J. and Robert N Rafson. 1999. Brownfields: Redeveloping Environmental
Distressed Properties. McGraw Hill: New York, NY.
Reardon, Kenneth M. 1995. Creating a community/university partnership that works: The case
of the East St. Louis Action Research Project. Metropolitan Universities: An
International Forum, 6 (2): 47-59.
Reardon, Kenneth M. 1998. Enhancing the Capacity of Community-Based Organizations in
East St. Louis. Journal of Planning Education and Research, 17: 323-333.
Reardon, Kenneth M. and Thomas P. Shields. 1997. Promoting sustainable
community/university partnerships through participatory action research. National
Society for Experiential Education Quarterly, 23 (1): 22-25.
Simons, Robert A. 1998. Turning Brownfields into Greenbacks: Developing and Financing
Environmentally Contaminated Urban Real Estate. Urban Land Institute: Washington,
DC.
Sterman, John. 2000. Business Dynamics. New York: Irwin/McGraw-Hill.
Tam, Edwin K. L. and Philip H. Byer. 2002. Remediation of Contaminated Lands: A Decision
Methodology for Site Owners. Journal of Environmental Management, 64: 387-400.
Thomas, Michael R. 2002. GIS-based decision support system for brownfield redevelopment.
Landscape and Urban Planning 58,1: 7-23.
US Conference of Mayors. 2003. Recycling America’s Land. [Online]. Available:
http://www.usmayors.org/uscm/brownfields/RecycleAmerica2003.pdf. Accessed: 11-6-
03.
20
US EPA. 2002. Technical Approaches to Characterizing and Redeveloping Brownfields Sites:
Municipal Landfills and Illegal Dumps. National Risk Management and Research
Laboratory: Technology Transfer and Support Division. EPA/625/R-02/002.
US EPA. 2003a. Brownfields Definition. US EPA Brownfields Homepage. [Online].
Available: http://www.epa.gov/swerosps/bf/glossary.htm. Accessed: 11-4-03.
US EPA. 2003b. Brownfields Mission Statement. US EPA Brownfields Homepage. [Online].
Available: http://www.epa.gov/swerosps/bf/mission.htm. Accessed: 1 1-4-03.
US EPA. 2003c. CERCLA Overview. US EPA Superfund Homepage. [Online]. Available:
http://www.epa.gov/superfund/action/law/cercla.htm. Accessed: 11-4-03.
Wright, James G. 1997. Risks and Rewards of Brownfield Redevelopment. Lincoln Institute of
Land Policy: Cambridge, MA.
21
