1

Introduction

Human activity has been altering—for better and for worse—natural

environments for thousands of years. Beginning several decades ago
societies decided that they wanted a better understanding of human
interactions with other species and the abiotic environment. Out of this
concern have come laws and regulations requiring the evaluation of
potential environmental impacts caused by continuing and new hu-
man activities. While science plays a role in the evaluation, the criteria
by which activities are judged are those that reflect the societal values
of the regulating jurisdiction. These governments, and other environ-
mental impact assessment (EIA) professionals, need practical and ef-
fective tools in order to be effective.
There is a hierarchy which can be used to organize the practical ap-
plication of these tools. At the top of this hierarchy is environmental
risk management. Beneath that are levels of environmental manage-
ment systems, regulatory oversight and approval, compliance assur-
ance, violation enforcement systems and reclamation activities (which
include remediation, enhancement and creation, among others). Envi-
ronmental risk management is a process to guide businesses through
the potential uncertainties of environmental laws and regulations. This
includes bringing all their facilities and operations into compliance
with existing regulations, maintaining full compliance, providing as
much certainty and predictability as is possible under current condi-
tions and planning for change. Governments, too, manage environ-
mental risk through their statutory and regulatory programs. Envi-
2 1 Introduction

ronmental management systems are the detailed implementations of

the environmental risk management policies. Environmental impact
assessments are usually part of the lower hierarchical levels, imple-
mented by national, regional and local governments.
Environmental impact assessment have been statutory and regu-
latory requirements for businesses around the world for more than
three decades. In the United States these assessments are required un-
der the National Environmental Policy Act (NEPA) in the form of En-
vironmental Assessments (EA) and Environmental Impact Statements
(EIS). At the state level assessments may also be required regardless
of whether a major federal action that triggers NEPA compliance is
involved. Washington State has its State Environmental Act (SEPA),
Oregon has its Statewide Planning Goals (in particular, Goal 5: Natural
Resources, Scenic and Historic Areas. and Open Spaces) and Califor-
nia has its California Environmental Quality Act (CEQA). Other states
have their environmental assessment laws and regulations, too. These
regulations apply to industry (e.g., mining, forestry, electrical genera-
tion/transmission), commerce (e.g., transportation) and development
by ports, resort operators, housing builders and others. In other coun-
tries there are equivalent national and subnational requirements for
development and industrial projects.
No matter where the project is located or what it involves, there
are a number of EIA characteristics and procedures with the poten-
tial to frustrate the various interests involved in the process. Too often
the frustration produces contention which increases EIA cost and time
without resolving or reducing the underlying concerns of all parties.
One of the most important characteristics which creates the potential
frustration elements is the inherently subjective nature of environmen-
tal impact assessments. Because societal values of significance and ac-
ceptability are impossible to measure objectively, decisions may seem
arbitrary and capricious to groups who do not like the outcome. Much
too frequently, these decisions end up in the legal system, where costs
and time delays continue to accumulate while the subjectivity is not
removed.
There are also real problems that arise during the environmen-
tal impact assessment process, and these may also lead to delays, in-
creased costs, and dissatisfaction by one or more s in the process. Ex-
amples of real problems are—
 1 Introduction 3

• Quantitatively describing existing conditions (environmental, so-

cial and economic) so they can be compared with alternative out-
comes.
• Establishing the relative importance of various condition compo-
nents so the most important ones are included in the assessment.
• Ensuring the equal and objective analysis of each alternative.
• Having a technically sound and legally defensible basis for making
a decision based on the assessments.
Several of these potential problems—very important ones—will be
thoroughly documented in Part I of this book. Real problems and “frus-
tration elements” may be related in any specific assessment process.
A method of solving the identified problems, a mathematically
rigorous, alternative environmental impact assessment process using
fuzzy logic and approximate reasoning, will be described and justified
in Part II. To illustrate how such a computational intelligence approach
would work in performing an assessment, an example is presented in
Part III. The computer modeling system used for the example in Part III
is FuzzyEI-Assessor™. Other fuzzy system modeling software is avail-
able at no cost or can be licensed from the developers. These could be
used to implement the modern approach described in Part II.
Two factors differentiate the approach using fuzzy system models
from those techniques used in the past:
1. It takes advantage of advanced computational intelligence tech-
niques (such as fuzzy sets and logic) for quantifying and manip-
ulating (in a mathematically rigorous way) subjective, inherently
uncertain or imprecise values and concepts.
2. While the protocol is widely applicable and valid under many dif-
ferent environmental statutes, regulations, and societal values, its
application is site specific but consistent.
In addition, the process works much more effectively with the active
input of s, the public, and everyone else with an opinion. The ap-
proach to conducting EIAs described in this book is technically clas-
sified as multi-objective, multi-criteria decision-making under conditions of
uncertainty. It is part expert system, part decision support system, and
all well supported in set theory, operations research, decision science,
ecosystem ecology, and other disciplines.
4 1 Introduction

1.1 Making Decisions

There is a subset of operations research called “decision science” that
focuses both on how decisions are made and on how to make better
decisions. Decision science also incorporates elements from psychol-
ogy, business, computer processing of information, and other interdis-
ciplinary aspects of importance to managers. Many of the ideas and
techniques are appropriate to the environmental impact assessment
process. After all, the underlying purpose of conducting an environ-
mental assessment is to create a basis for making informed decisions
on the condition or acceptability of a natural ecosystem (e.g., the status
of Coho salmon populations along the Pacific Ocean coast of Oregon),
whether to allow an industrial or commercial project to be developed,
or to determine the appropriate alternative for such development. This
book focuses on identifying the alternatives among which a decision
will be made and predicting the significance of the future environ-
ments under each alternative.

1.1.1 Decision Support Systems

There are two components involved in making a decision [23]:
1. Judgment. The experience, feelings, and insight that the decision-
maker uses consciously or unconsciously.
2. Knowledge. The project-specific information that has been col-
lected and analyzed for use in making an informed decision.
The environmental impact assessment process is supposed to be based
on knowledge of the specific environment and similar projects under-
taken in the past. Judgment—including the “best professional judg-
ment” of the technical staff—is always a factor. The goal of the process
should be a result that is informed and defensible. Several factors make
it difficult to reach this goal, including having multiple objectives, the
presence of many (often conflicting) constraints to be satisfied, and the
accumulation of a large amount of project-specific information to be
properly analyzed and understood in the context of the project. This
last factor reflects the many scientific, economic, social, and political
considerations that cannot be ignored in the decision-making process.
Within the broad category of decision science, environmental impact
assessments are in the broad category of multi-objective, multi-criteria
decisions.
 1.1 Making Decisions 5

[Link] Multi-Objective

Multi-objective means that there are different goals to be attained. For

example, operation of a dam and reservoir system may need to be op-
timized for flood control, irrigation, barge traffic, hydroelectric power
production, and recreational use. Land development at a port may
need to be optimized for heavy industry, light industry, maritime, rail
and truck shipping, cargo storage, wildlifeprotection, and general eco-
nomic gain. These objectives can sometimes seem to be mutually ex-
clusive, but with the proper techniques and approaches an optimal
balance can almost always be achieved. This balance maximizes all ob-
jectives relative to each other. Linear programming (out of the field
of operations research) is a long-standing example of multi-objective
optimization in decision-making and is sometimes taught in graduate
courses of systems ecology.

[Link] Multi-Criteria

Multi-criteria describes the different values that constrain achievement

of the identified objectives. In operations research multi-criteria deci-
sions result from the use of a “bottom up” approach when the values of
different public groups shape the outcome. Criteria for a project might
include increasing the number of available jobs, providing a high rate
of return on investment, protecting the environment, increasing the
size of a population of fish or wildlife, and being æsthetically pleas-
ing.
In addition to having multiple objectives and being constrained by
multiple criteria, the EIA process occurs under conditions of future un-
certainty and frequently involves more than one decision-maker. This
makes it very difficult to reach the appropriate decision; one that can
be satisfactorily explained and completely justified. Historically, decid-
ing on a decision method emphasized the psycho-social interactions
of the group and how it reached consensus. The powerful computa-
tional intelligence tools now available bring advantages to all levels of
decision-making. Among these advantages are their ability to quantify
subjective values, beliefs, and language and the ability to minimize, or
eliminate, qualitative pressures from the decision.
Several mathematical approaches to handling imprecise concepts
have been developed and used in different situations. Each approach
6 1 Introduction

has a type of application where it works very well. Environmental

impact assessments are characterized by the importance of linguistic
variables that are not directly measurable (that is, they are terms eas-
ily expressed in language that are not represented by discrete numbers
or thresholds). Fuzzy sets are used to convert the subjective linguistic
variable into a number and fuzzy logic manipulates these numbers in
a mathematically rigorous way.
The above overview describes what is required by a decision sup-
port system (DSS). Environmental impact assessment are a form of
DSS, so the approach used to conduct the EIA should accommodate
multiple objectives and multiple criteria under conditions of uncer-
tainty. Over the past few decades the power of computers has been ap-
plied to the complexity of decision-making. Computers are far more ca-
pable than people at making sense out of large amounts of data. There
are many different decision support system applications available to
the business community, and it is certainly reasonable to demand that
such an application facilitate decisions when assessing environmental
impacts of a proposed project.
However, people are far superior to computers when it comes to
judgment and the extrapolation of past experience to new situations.
Human expertise is another facet of environmental impact assessments
that contributes to informed and legally defensible decisions.

1.1.2 Expert Systems

The capture of an expert’s knowledge within computer software was

one of the first artificial intelligence (AI) efforts in the 1960s and 1970s.
These knowledge-based expert systems, or expert systems for short,
capture what an expert would do in a particular situation as a series
of IF-THEN rules. While in many ways the promise of expert systems
was over-hyped, they are commonly used today in fields as diverse as
medical diagnosis, psychiatric screening, new product pricing, insur-
ance fraud detection, financial portfolio evaluation and other business
processes [8].
During the EIA process technical experts are employed to analyze,
interpret and explain the large amount of technical data that describe
natural ecosystems. More importantly, IF-THEN “rules” are precisely
how alternatives are analyzed, but not in a structured, formal way. For
 1.1 Making Decisions 7

example, the EIA report might include statements such as, “if the wet-
land is filled, then migratory waterfowl habitat will be decreased” or
“if the ’No Action’ alternative is selected, then 80 high-paying, perma-
nent jobs will not be created.” Translating the project-specific IF-THEN
rules into a form that can be used by a computer permits these rules to
be uniformly applied to all alternatives. When human reasoning is em-
ulated by computer software, the result is called an approximate rea-
soning model. Because this software applies the collective knowledge
of experts to solve the problems of environmental impact assessments
the application is also an expert system.
What the appropriate expert system model provides to the decision-
maker is an objective characterization of the baseline environmental
conditions along with projected environmental characterizations of the
future conditions when each project alternative is applied. These alter-
native characterizations are calculated with input of all stakeholders on
values such as “acceptability,” “significance,” “sustainability,” “biodi-
versity,” and other such concepts that cannot be objectively measured.
This makes the process a true decision support system rather than a
decision-making system.

1.1.3 Decision-Making in the EIA Process

Environmental decision-making involves four broad steps:

1. Problem identification.
2. Options (alternatives) identification.
3. Predictions of the future environments associated with each alter-
native.
4. Selecting one of the options.
Problem identification is almost always defined under a law at the na-
tional, regional, or local level and will not be further discussed. The
next two steps (alternative identification and predictions of potential
futures) are the focus and purpose of this book. The final component—
selecting one alternative action—almost always is based on social val-
ues, economic priorities, and political considerations and will not be
further considered. However, if the identification and characterization
of alternative options is done with sufficient scientific objectivity and
8 1 Introduction

mathematical rigor then the selection of a particular option should be

simplified—even in the most sensitive and contentious situations.1
The decision-making process can be further organized into the
categories of information gathering and analyses. These categories
include—
• Identifying values.
• Characterizing (not only describing) the existing environment.
• Characterizing the social, economic and political setting.
• Characterizing the legal and regulatory setting.
• Integrating information.
• Forecasting effects of project alternatives.
• Assessing options.
• Post-decision assessment and justification.
Various tools, methods, and techniques have been applied to these cat-
egories. This book will address the needs of decision-makers in most
of environmental impact assessment components.

1
This is not to assume naively that everyone will be pleased with the out-
come, but the steps leading to the decision will be visible and shown not to
be arbitrary and capricious.