Development Of Fuzzy Multi-Criteria Decision Analysis Approach
For Contaminated Site Management

Mohammad Habibur Rahman
B.Sc, Independent University, Bangladesh, 2004

Thesis Submitted In Partial Fulfillment Of
The Requirements For The Degree Of
Master Of Science
in
Natural Resources and Environmental Studies
(Environmental Science)

The University of Northern British Columbia
March, 2008
© Mohammad Habibur Rahman, 2008

1*1

Library and
Archives Canada

Bibliotheque et
Archives Canada

Published Heritage
Branch

Direction du
Patrimoine de I'edition

395 Wellington Street
Ottawa ON K1A0N4
Canada

395, rue Wellington
Ottawa ON K1A0N4
Canada

Your file Votre reference
ISBN: 978-0-494-48803-4
Our file Notre reference
ISBN: 978-0-494-48803-4

NOTICE:
The author has granted a nonexclusive license allowing Library
and Archives Canada to reproduce,
publish, archive, preserve, conserve,
communicate to the public by
telecommunication or on the Internet,
loan, distribute and sell theses
worldwide, for commercial or noncommercial purposes, in microform,
paper, electronic and/or any other
formats.

AVIS:
L'auteur a accorde une licence non exclusive
permettant a la Bibliotheque et Archives
Canada de reproduire, publier, archiver,
sauvegarder, conserver, transmettre au public
par telecommunication ou par Plntemet, prefer,
distribuer et vendre des theses partout dans
le monde, a des fins commerciales ou autres,
sur support microforme, papier, electronique
et/ou autres formats.

The author retains copyright
ownership and moral rights in
this thesis. Neither the thesis
nor substantial extracts from it
may be printed or otherwise
reproduced without the author's
permission.

L'auteur conserve la propriete du droit d'auteur
et des droits moraux qui protege cette these.
Ni la these ni des extraits substantiels de
celle-ci ne doivent etre imprimes ou autrement
reproduits sans son autorisation.

In compliance with the Canadian
Privacy Act some supporting
forms may have been removed
from this thesis.

Conformement a la loi canadienne
sur la protection de la vie privee,
quelques formulaires secondaires
ont ete enleves de cette these.

While these forms may be included
in the document page count,
their removal does not represent
any loss of content from the
thesis.

Bien que ces formulaires
aient inclus dans la pagination,
il n'y aura aucun contenu manquant.

Canada

Abstract
Selection of remediation alternative is an important task in the decision making process of
contaminated site management. The number of available remediation alternatives is
increasing over the years as a result of progress in scientific research. Decision makers face a
confounded situation to select the best acceptable alternative by satisfying various
preferences of different stakeholders. In this research, a fuzzy multi-criteria decision analysis
(FMCDA) approach was developed. Since most information available in the decision making
process is not deterministic, fuzzy-set theory was used to deal with such uncertainty. The
developed FMCDA approach ranks the alternatives according to the utility values. Different
stakeholders' opinions were effectively incorporated in the developed approach, allowing for
a robust decision making for contaminated site management. The developed method was
then applied to the management of a site in northern British Columbia to examine its
applicability.

TABLE OF CONTENTS
ABSTRACT

i

LIST OF TABLES

v

LIST OF FIGURES

vi

ACKNOWLEDGEMENT

viii

Chapter 1 Introduction

1

Chapter 2 Literature Review
2.1 General background of contaminated sites

9
9

2.2 Remediation technologies for contaminated site management
2.2.1 In-situ methods
2.2.2 Ex-situ methods

9
10
12

2.3 Multi-criteria decision analysis (MCDA) for environmental management
2.3.1 Process of MCDA

13
17

2.4 Description of three MCDA methods
2.4.1 Simple additive weighting method (SAW)
2.4.1.1 Procedure
2.4.2 Technique for order preference by similarity to ideal solution (TOPSIS)
2.4.2.1 Procedure
2.4.3 Weighted product method (WPM)
2.4.3.1 Procedure

20
20
20
21
22
23
23

2.5 Fuzzy-set theory
2.5.1 Handling uncertainty through fuzzy-set theory
2.5.2 Conversion of linguistic criteria preferences
2.5.3 Conversion of fuzzy-sets into crisp weights

24
26
28
31

2.6 Application of fuzzy-set approach in environmental problems
2.7 Summary of literature review

33
35

Chapter 3 Methodology

37

3.1 Overview of methodology

37

3.2 Step 1 Stakeholder selection

38

3.3 Step 2 Criteria selection
3.3.1 Remediation alternative evaluation criteria
3.3.2 Description of criteria
A. Cleanup time (in-situ and ex-situ)

39
40
41
41

u

B. Overall cleanup cost (in-situ and ex-situ)
C. Minimum achievable concentration
D. Community acceptability
E. Availability
F. Regulatory permitting acceptability
G. Development status of a technology
H. Technology maintenance requirement

41
41
41
42
42
42
43

3.4 Step 3 Establishing membership functions of fuzzy criteria
3.4.1 Triangular fuzzy numbers (TFNs)
3.4.2 Selection of criteria value range
3.4.3 Membership functions of fuzzy criteria
A. Time required for cleanup
B. Cleanup Cost
C. Minimum achievable concentration
D. Community acceptability
E. Technology availability and other criteria

43
43
45
47
47
53
59
62
65

3.5 Step 4 Developing fuzzy multi-criteria evaluation matrix
3.5.1 Conversion of linguistic variables
3.5.2 Fuzzy evaluation matrix

70
70
76

3.6 Step 5 Defuzzification

77

3.7 Step 6 Development of a decision support system

78

Chapter 4 Case Study, Results and Discussions

81

4.1 Description of study site

81

4.2 Description of remedial alternatives in the system
4.2.1 Information about remedial alternatives

82
82

4.3 Fuzzy processing of criteria information

83

4.4 Processing of input data
4.4.1 Aggregation of membership values and criteria weights

85
94

4.5 Defuzzification and ranking of alternatives

95

4.6 Ranking of alternatives

97

4.7 Comparison of results of MCDA techniques for the same case study site
4.7.1 Simple additive weighting (SAW) method
4.7.2 Technique for order performance by similarity to ideal solution (TOPSIS)
4.7.3 Weighted product method (WPM)
4.8 Results from MCDA methods

98
100
102
104
106

4.9 Comparison of results

107

4.10 Sensitivity analysis
4.10.1 Single input value

108
108
in

4.10.2 Change in criteria importance weight
4.10.3 Change in remediation alternative performance values

111
115

Chapter 5 Conclusions and Recommendations
5.1 Summary
5.2 Future extensions

117
117
120

References

122

APPENDIX I Questionnaire

141

APPENDIX II Introduction of Remediation Technologies

151

1. Enhanced bioremediation (in-situ biological treatment)

151

2. Bioventing (in-situ biological treatment)

152

3. Phytoremediation (in-situ biological treatment)

153

4. Soil vapor extraction (in-situ physical/chemical treatment)

153

5. Biopiles / static pile (ex-situ biological treatment)

154

5. Landfarming (ex-situ biological treatment)

155

6. Slurry phase treatment (ex-situ biological treatment)

156

7. Low temperature thermal desorption-LTTD (ex-situ thermal treatment)

156

8. Soil washing (ex-situ physical/chemical treatment)

157

9. Soil flushing (in-situ physical chemical treatment)

158

IV

LIST OF TABLES
Table 2.1 Comparative assessment of in-situ soil remedial technologies
11
Table 2.2 Comparative assessment of ex-situ soil remedial technologies
12
Table 2.3 Examples of linguistic universes
31
Table 3.1 Examples of typical stakeholder interests
38
Table 3.2 Criteria value range
46
Table 3.3 Survey on the fuzzy-sets of in-situ cleanup time
48
Table 3.4 Survey on the fuzzy-sets of ex-situ cleanup time
51
Table 3.5 Survey on the fuzzy-sets of in-situ cleanup cost
54
Table 3.6 Survey on the fuzzy-sets of ex-situ cleanup cost
57
Table 3.7 Survey on the fuzzy-sets of minimum achievable concentration
60
Table 3.8 Survey on the fuzzy-sets of community acceptability levels
63
Table 3.9 Triangular fuzzy numbers (TFNs) defined for development of fuzzy membership
functions of remediation alternative evaluation criteria
69
Table 3.10 Determination of criteria value using scale three
71
Table 3.11 Criteria value determination using scale six
72
Table 3.12 Conversion of linguistic terms into crisp values
73
Table 4.1 Information on remedial alternatives
82
Table 4.2 Input values of remedial alternatives
84
Table 4.3 Fuzzification of remediation alternative criteria
92
Table 4.4 Membership values of each alternative after aggregation
95
Table 4.5 Criteria input value for MCDA methods
99
Table 4.6 Calculations using SAW method
101
Table 4.7 Normalized criterion ratings
103
Table 4.8 Weighted normalized values of criteria
103
Table 4.9 Positive ideal and negative ideal solutions
103
Table 4.10 Separation measurefrompositive and negative ideal solutions and calculated
value function of each alternative
104
Table 4.11 Calculations of weighted product method
105
Table 4.12 Single input value used in the fuzzy multi-criteria approach
109

v

LIST OF FIGURES
Fig. 2.1. Common soil remediation technologies
Fig. 2.2. Taxonomy of classical MCDA methods
Fig. 2.3. Overall procedure of typical MCDA application
Fig. 2.4. A hierarchy of criteria for remediation alternative selection
Fig. 2.5. Classification of fuzzy theory
Fig. 2.6. Linguistic term conversion into fuzzy-set
Fig. 2.7. Conversion of linguistic term into crisp values
Fig. 2.8. Linguistic terms conversion scales
Fig. 2.9. Illustration of determining crisp value
Fig. 3.1. Overall methodology of the developed fuzzy-multi criteria decision analysis
approach
Fig. 3.2. Example of a triangular fuzzy-set
Fig. 3.3. Membership functions of fuzzy-sets "cleanup time (in-situ)" criterion
Fig. 3.4. Membership functions of fuzzy-sets of "cleanup time (ex-situ)"
Fig. 3.5. Membership functions of fuzzy-sets of "cleanup cost (in-situ)"
Fig. 3.6. Membership functions of fuzzy-sets of "cleanup cost (ex-situ)"
Fig. 3.7. Membership functions of fuzzy-sets of "minimum achievable concentration"
Fig. 3.8. Membership functions of fuzzy-sets of "community acceptability"
Fig. 3.9.. Membership functions of fuzzy-sets of "technology availability" criterion
Fig. 3.10. Membership functions of fuzzy-sets of "regulatory acceptability" criterion
Fig. 3.11. Membership functions of fuzzy-sets of "development status" criterion
Fig. 3.12. Membership functions of fuzzy-sets of "maintenance requirement" criterion
Fig. 3.13. Conversion of linguistic variables by scale three
Fig. 3.14. Conversion of linguistic variables by scale six
Fig. 3.15. Average criteria importance weights
Fig. 3.16. Selection of remedial alternatives
Fig. 3.17. Criteria data input in the system
Fig. 3.18. Ranking order of remedial alternatives
Fig. 3.19. Remediation decision process by the system
Fig. 4.1. Data input for in-situ cleanup time
Fig. 4.2. Data input for in-situ cleanup cost
Fig. 4.3. Data input for ability to reduce contaminant concentration
Fig. 4.4. Data input for community acceptance
Fig. 4.5. Data input process of technology availability criterion
Fig. 4.6. Data input for regulatory acceptance criterion
Fig. 4.7. Data input for development status
Fig. 4.8. Fuzzification of input data for required maintenance criterion
Fig. 4.9. Final membership functions of remediation alternatives
Fig. 4.10. Ranking of remediation alternatives by fuzzy multi-criteria method
Fig. 4.11. Utility values of alternatives obtained from MCDA methods
Fig. 4.12 Ranking order of remedial alternatives by MCDA and fuzzy multi-criteria and
Fig. 4.13. Comparison of results when uncertainty is considered in the evaluation
VI

10
14
17
19
26
27
28
30
32
37
44
50
53
56
59
62
65
66
66
67
67
71
72
75
78
79
79
80
85
86
87
88
89
90
90
91
96
97
106
107
111

Fig. 4.14. Sensitivity analysis of remediation alternatives by changing "overall cost"
criterion
113
Fig. 4.15. Sensitivity analysis of remediation alternatives by changing the importance
weights "overall cost", "cleanup time" and "community acceptability" criteria 114
Fig. 4.16. Sensitivity analysis for SVE alternative by changing performance values of
"community acceptability, "cleanup time" and "cleanup cost" criteria
115

vu

ACKNOWLEDGEMENT
I would like to thank my supervisor Dr. Jianbing Li, for his patience and providing me all sort
of supports throughout my graduate studies at the University of Northern British Columbia
(UNBC). He made many efforts discussing and developing my ideas, working with me through
difficulties, and reviewing my papers. His guidance and willingness to expand myself
intellectually and the opportunities he afforded me are truly appreciated.
I would also like to thank my co-supervisor Dr. Jueyi Sui and committee member Dr. Liang
Chen, for their guidance throughout this research. Despite busy schedules, they were more than
willing to meet with me when I needed advice.
I am in debt to Dr. Michael Rutherford, Associate Professor, Environmental Science,
UNBC for his valuable feedback on questionnaire survey. He also helped me to network with
environmental professionals for this research.
I would like to thank Doug McMillan, Scientific Research Officer of SNC-Lavalin, Prince
George for his valuable suggestions and input in my research work.
I am grateful to Bharath Reddy, Computer Science, UNBC and MA Razzaq, Senior
Application Architect, Beendo Corporation, USA for their help to develop the user-friendly
decision support system.
Together with, I would like to thank my brother, Muhammad Rahman and his family for
providing me accommodation and all sorts of supports throughout my sojourn.
Finally, I want to acknowledge my parents for their sacrifice to let me study in Canada.
Their perpetual emotional supports, and never letting me forget the bigger picture helped me to
reach this far.
Consequently, I can honestly say that this research has been a group effort and I thank
everyone for an amazing experience that I will carry with me throughout my life.

vm

Chapter 1 Introduction
Numerous operations in petroleum exploration, production and transportation have been
causing various environmental problems in Canada and worldwide (Amro, 2004). It is
estimated that there are an excess of 10,000 contaminated sites in Canada and these sites pose
significant threats to the ecosystem and human health (Siciliano and Germida, 1998; Sousa,
2001). The effective management of such contaminated sites is of critical importance and is
often a liability of the federal and provincial governments as well as the industries. Thus, a
solid decision making process for site management is necessary. However, the decision making
of contaminated site management is complex due to the presence of many uncertainties in
evaluation criteria and different stakeholder interests. Each stakeholder may define the risks
and potential benefits in a remedial alternative by different (even unique) criteria, and
implementation of the alternative may be prevented by raising objections that seem
unnecessary to other stakeholders (Seager et al., 2007). Over the past 30 years, the concept of
contaminated site management has been changed markedly (Pollard et al., 2004). In the mid1970s the focus was on cost-centered approaches, in the mid-1980s technological feasibility
was emphasized, in the mid-1990s risk based approaches were taken and in this new
millennium, environmental decisions must be socially-robust (Urban Task Force, 1999; ESRC
Global Environmental Change Programme, 2000). The selection of remedial alternatives
requires formation of partnerships between technology developers, manufacturers, regulators,
end-users and public (Seager et al, 2007). The existing decision making approaches have many
limitations in a number of components, such as (a) evaluation of remedial alternatives, (b)

1

evaluation criteria of remedial alternative, (c) stakeholder involvement, and (d) uncertainties in
remedial alternative evaluation process.

Development and implementation of remedial alternatives for contaminated sites have
received much attention during the past decades (Riser-Roberts, 1998; Li et al., 2001). The
number of in-situ and ex-situ remedial technologies for cleaning up contaminated sites has
been growing over the years due to advancements in science and technological research. Thus,
decision makers face complex problems in identifying the best alternative from a wide range of
remedial alternatives where none of them are dominating (Khadam and Kaluarachchi, 2003).
Since most technologies are site-specific, the selection of appropriate technologies is often
difficult. Being able to make these difficult decisions, with consideration to all stakeholders, is
an extremely important step in successful management of contaminated sites (Khan et al.,
2004). Most of the developed decision support systems provide a list of applicable remedial
alternatives, but the problem remains in selecting the best alternative.
In addition, selection of a remedial alternative involves a multi-criteria evaluation process,
requiring a multi-criteria analysis approach. Multi-criteria analysis refers to screening,
prioritizing, ranking and/or selecting a set of remedial alternatives under independent or
conflicting criteria. A wide range of criteria (e.g., flexibility, compatibility, time, cost,
environmental impact, public acceptance) need to be considered in an evaluation process.
These criteria may conflict with each other in terms of their trade off values. For example,
some remedial alternatives might be economically feasible but require a lengthy treatment
process, while other alternatives might be expensive but require shorter clean up periods.
2

Moreover, the management of contaminated sites is not limited to choosing the right
solution solely by a decision maker or by an environmental engineer. According to Akter et al.
(2005), decision on contaminated site management is no longer limited to the selection of the
most preferred alternative among the non-dominated solutions; the analysis needs to be
extended to account for diverse opinions of multiple decision makers. The importance of
stakeholders (e.g., government, industry, regulatory agency) involvement in the decision
making process can be found in many literatures (Gregory et al., 1994; Kamnikar, 2001;
Balasubramaniam et al., 2007). Discussions with stakeholders are needed not only for the
related model building process, but also to provide stakeholders a voice which facilitates the
development of stakeholder trust in the policy-implementation process (Lind and Tyler, 1988;
Lind, 1995). Borsuk et al. (2001) identified that stakeholders do not only value a particular
environmental problem, they also care about how they are involved in the decision making
process. In the past, the manager of a contaminated site remediation project needed to be
skilled only in excavating, but now a manager must combine the skills and talents of engineer,
lawyer, scientist, and negotiator (Cole, 1994). One of the recommendations Kamnikar (2001)
made for contaminated site management is that one has to understand the importance of early
community involvement and this will increase the acceptance of specific remediation projects,
acceptance of new or alternative remediation techniques, and will establish trust and good
working relationships. As a result, the decision for environmental contaminated site
management needs to take into account the inputs from different stakeholders with different
priorities and objectives (Linkove et al, 2006). Existing decision analysis approaches fail to
address different stakeholder opinions in the process of contaminated site management and
3

limited efforts have previously been made to address this issue (Zahedi, 1986; Juang and Lee,
1991; Cheng, 1996).
There are many uncertainties involved in the evaluation process of remediation alternatives.
For example, the criteria of "cleanup cost" may vary significantly even for a particular
technology ranging from $30 per m3 to $300 per m3 of soil. Similarly, the criteria (e.g.,
"impact on the environment", "community acceptability" etc.) without units are measured on a
numeric scale (e.g., 1 to 10, 0 to 1). Generally, in a numeric scale the lower values are given to
represent less preference and higher values are given to express high preference. Existing
multi-criteria analysis methods assume that ratings of alternatives and the weighting factors of
criteria are deterministic values. For example, rating on "clean up cost" criteria is rated as "10
= when cost is less than $100/m3; 5= when cost is $100-300/m3; and 1= when cost is more than
$300/m3". By applying the above rating method, when a remediation alternative costs $99/m3,
it will receive 10 points. However, if the cost is $101/m3 it will receive 5 points, and this
significantly underestimates the remedial alternative with only slightly higher cost (e.g., when
cleanup cost for 2 alternatives are $101/m3 and $99/m3 respectively, the difference in cost is
only $2/m3). In existing rating practices, the rating value of criteria is either in or not in the
crisp set. However, a rating value can partially belong to a crisp set or belong to more than one
set. Therefore, it can be stated that the ratings of criteria are not best represented by only
deterministic or single crisp value. These criteria ratings could be best represented by a range
of values, or by linguistic terms.

4

Selection of criteria importance weight for remedial alternative evaluation is associated
with another source of uncertainty. Different stakeholder has different preferences on each
criterion. Thus the values of criteria importance can vary significantly. Most of the existing
multi-criteria analysis models apply the analytical hierarchy process (AHP) (Saaty, 1994) to
calculate criteria importance weight. In AHP, a decision problem is presented hierarchically
and this method synthesizes various assessments for ranking alternatives in a systematic way
(Yeh et al., 2000). In AHP method, stakeholders are asked to compare two criteria at a time
and provide a crisp rating on the comparison. However, such approach is often criticized for
inappropriateness of the crisp ratio representation and for cumbersome procedure (Zahedi,
1986; Juang and Lee, 1991; Cheng, 1996). According to Petrovic and Petrovic (2002)
stakeholders with different technical and non-technical backgrounds feel more comfortable
with linguistic expressions to express their opinions (e.g., high, medium, low).

There are many multi-criteria decision analysis (MCDA) methods available to support
environmental decision making, such as the simple additive weighting (SAW) method,
weighted product method (WPM), preference ranking organization method for enrichment
evaluations (PROMETHEE), and technique for order performance by similarity to ideal
solution (TOPSIS). Most of the widely used multi-criteria analysis methods are effective in
dealing problems with quantitative data (Hwang and Yoon, 1981; Bana e Costa, 1990; Yeh et
al, 2000). However, the applicability of existing MCDA methods are seriously reduced when
dealing with situations where imprecision and subjectiveness of the decision making process
are present (Chen and Hwang, 1992; Hellendoorn, 1997; Bender et al., 2000; Petrovic and

5

Petrovic, 2002). By neglecting these qualitative inputs, the robust capability of decision making
is lost. According to Linkove et al. (2006) current decision analysis practices do not offer a
comprehensive approach for incorporating the varied types of information and opinions of
multiple stakeholders. The effective incorporation of multiple stakeholder perspectives within
the decision making process of contaminated site management are thus of critical importance
and should be a principal task of environmental professionals and regulatory agencies (Testa
and Winegardner, 1991; Li et al., 2000, 2001; USEPA, 2001).

The uncertainties involved in qualitative data, qualitative criteria weights or the subjective
and imprecise assessments of the decision problem can be better expressed by fuzzy logic and
fuzzy set theory (Klir and Folger, 1988; Yeh et al, 2000). The application of fuzzy-set theory
(Zadeh, 1965) to multi-criteria problems provide an effective way for solving decision
problems in a fuzzy environment where little information is known (i.e., imprecise knowledge
from descriptions of human language) and the information is subjective (Bellman and Zadeh,
1970; Carlsson, 1982; Dubois and Prade, 1994; Zimmermann, 1996; Herrera and Verdegay,
1997; Sadiq et al, 2004b; Chang et al., 2007; Li et al. 2007). It is also an effective tool to
incorporate linguistic preferences from different stakeholders, and can address decision making
problems under uncertainty in number of environmental management areas (Li et al., 2007).
There has been much theoretical work done on the use of fuzzy-set theory in multi-criteria
decision analysis (MCDA) during the last two decades, however little attention has been given
to integrating these ideas and developing a fuzzy multi-criteria decision support (FMCDS)
system (Cheng, 2000). Particularly, few efforts have been made to apply this approach to

6

address the hydrocarbon impacted site management issues. In this thesis a hybrid method,
fuzzy multi-criteria decision analysis (FMCDA) will be developed and applied to evaluate and
rank applicable remediation alternatives for oil contaminated site by considering various
stakeholder preferences and uncertainties.

Consequently, there are 3 main objectives in this research including (a) to identify the
criteria for remedial alternative selection and determine the criteria importance weight
according to stakeholder preference, (b) to integrate and address uncertainty issues in
remediation alternative evaluation process (e.g., cost, time, and stakeholder preferences) into a
general decision analysis framework, and (c) to develop an effective fuzzy multi-criteria
approach for evaluating and ranking remediation alternatives by comprehensively considering
various independent and/or seemingly conflicting criteria.

This thesis is organized as follows. In Chapter 2, a review on different component of
contaminated site management issues are discussed. As well, existing practices of decision
making and their limitations are discussed in this chapter. Also, literature review on the
proposed fuzzy multi-criteria approach is presented. In Chapter 3, the overall methodology of
developing a fuzzy multi-criteria approach is described. Moreover, acquisition of criteria
importance weight, development of fuzzy membership functions, and development of a userfriendly decision support system is described in this chapter. In Chapter 4, the results of a case
study site using the developed method is presented. A comparison of results from fuzzy multicriteria and the existing multi-criteria methods are also presented. Besides, results of various

7

sensitivity analyses are discussed. In Chapter 5, conclusion of the research work is drawn,
while future direction of the research is discussed.

8

Chapter 2 Literature Review
2.1 General background of contaminated sites
Contaminated land is primarily a post-1800s problem worldwide in terms of cause but a
post 1970s phenomenon in terms of risk management (Petts et al., 1997; Sousa, 2001).
Significant industrialization has occurred over the past years in many developed nations (e.g.,
Canada, United States, and United Kingdom). Particularly, in Canada the oil and gas industries
play an important role in its economy. It is expected that this industry will continue to expand
in the future. However, a number of environmental concerns (e.g., soil and groundwater
contamination) are associated with such development and expansion (Dowd, 1985; Newton,
1991). It is estimated that there are 200,000 underground storage tanks (USTs) in Canada. The
leakage from these USTs causes contamination to the surrounding environment and significant
economic losses to the petroleum industries (CCME, 1993). In general, the number of
suspected/known contaminated site in USA is 384,400 and 20,000-30,000 in Canada (NRTEE,
1997; Simons, 1998).

2.2 Remediation technologies for contaminated site management
A great number of remediation technologies have been developed and implemented for
contaminated site management during the past years, and more remediation alternatives will be
available in the future due to continued competition among environmental service companies,
technology developers and development in technology researches (Vranes et al., 2000). The
existing remediation technologies can be divided into two categories of in-situ and ex-situ. For
9

in-situ remediation, no excavation is required, while ex-situ technologies require the removal,
usually by excavation, of contaminated soils. Some of the common remediation alternatives
can be further divided into biological, physical/chemical and thermal treatment methods (Fig.
2.1).

Soil remediation

In-situ methods

Jl

*

Physical/chemical methods

Thermal methods

• •

Soil vapor extraction
Soil flushing
Electrokinetics
Solidification/ stabilization

Soil heating
Vitrification

Ex-situ methods

*

*

Biological methods

i

+

Physical/chemical methods Thermal methods Biological methods

r

r

Natural attenuation
Soil washing
Enhanced bioremediation Solvent extraction
Bioventing
Chemical dechlorination
Air stripping
Electrokinetics
Solidification/stabilization

r

!r

Thermal desorption
Incineration
Vitrification

Bioremediation
Biopiles
Slurry-phase
Composting
Land farming

Fig. 2.1. Common soil remediation technologies (from Reddy et al., 1999)
2.2.1 In-situ methods
In-situ remediation methods treat the contaminated soil and/or water without being
excavated or transported. In-situ methods are advantageous because they are often cost
effective, make little site disruption, and possess increased safety due to a lessened risk of
accidental contamination exposure to both on-site workers and the general public (Reddy,
1999). However, in-situ methods generally require a longer time period, and there is less
certainty about the uniformity of treatment because of the variability in soil and aquifer
characteristics. A comparison of some in-situ soil remediation approaches is listed in Table 2.1.
10

Different in-situ technologies have some strengths and limitations in their applications. For
example, soil vapor extraction is a well developed technology for remediation of hydrocarbon
contaminated sites. However, this technology becomes less effective when the soil
characteristic is heterogeneous and there is low hydraulic conductivity. Similarly, there are
differences in technology availability and cleanup cost. The contaminated site manager needs
to have clear idea about such characteristics of each alternative under evaluation.

Table 2.1 Comparative assessment of in-situ soil remedial technologies (from Reddy et al.,
1999)
Technology
Strengths
Factors affect the
Cost range Commercial
availability
treatment
Soil vapor
extraction

It is a proven
technology

Soil flushing

It is effective for
residual contaminant
reduction
Useful for low hydraulic
conductivity soils and
mixed contaminants
This technology
converts contaminants
into non hazardous
substance; it requires
low cleanup cost
Hydrocarbon can be
easily recovered by this
technology
Effective for treatment
of mixed contaminants

Electrokinetics
Bioremediation

Soil heating
Vitrification

Solidification
/stabilization
Phytoremediation

It is a proven
technology
It produces less
secondary waste and it
has capability of treating
broad range of
contaminants

Not effective for
heterogeneous and low
hydraulic conductivity soils
Flushing solution may trap
in soil; not effective for low
conductivity soils
Not effective for metallic
contaminants

<$100/ton

Widespread

$80-$165/ton

Very limited

$90-$130/ton

Very limited

It requires lengthy treatment
time; not effective for low
hydraulic conductivity soils

$27-$310/ton

Widespread

Not effective for metallic
contaminants and low
hydraulic conductivity soils
It converts contaminated soil
into glassy structured soil;
not effective for metallic
compounds
Not effective for low
hydraulic conductivity soils
Applicable to limited to
shallow depths and low
concentration levels; it
requires lengthy treatment
time and there is risk of food
chain contamination

$50-$100/ton

Limited

$350-$900/ton

Limited

$100-$150/ton

Widespread

<$100/ton

Very limited

11

2.2.2 Ex-situ methods
Ex-situ methods treat contaminated soils and/or groundwater after excavation. They often
require shorter cleanup time period as compared to in-situ methods. Another advantage of exsitu methods is that they provide more certainty about the uniformity of treatment because of
the ability to homogenize and continuously mix the soil. Excavated soil can be treated on site
or off-site depending on the site-specific conditions. Ex-situ treatments require excavation of
soils, resulting in increased costs. Moreover, the urban settings characteristics, including
neighboring buildings and narrow streets, often limit the use of onsite treatment facilities.
Therefore, off-site treatment, requiring transport of contaminated materials to a treatment
facility, is often necessary at urban contaminated sites. Ex-situ treatment methods are attractive
because consideration does not need to be given to subsurface conditions. Ex-situ treatments
also offer greater control and monitoring during remedial activity implementation (Reddy et
al., 1999). A comparison among ex-situ remediation technologies is shown in Table 2.2.

Table 2.2 Comparative assessment of ex-situ soil remedial technologies (from Reddy et al.,
1999)
Technology
Strengths
Factors affect the
Cost range
Commercial
treatment
availability
Soil washing
Solvent extraction

Chemical
dechlorination

Electrokinetics

Volume of
contaminated soil is
reduced significantly
It has capability of
treating broad range of
contaminants
It reduces toxicity of
contaminants; it can be
used with other
technologies
Applicable for low
hydraulic conductivity
soils and mixed
contaminants

It is not effective in fine
textured soil

$100-$300/ton

Widespread

Not effective in clays

$100-$500/ton

Limited

Not applicable in the sites
with inorganic pollutants

$300-$500/ton

Limited

Not effective for remediation
of metal contaminated soils

$90-$130/ton

Very limited

12

Thermal
desorption
Incineration

Vitrification
Bioremediation

Solidification

It requires lower cost
than incineration
It has capability of
treating broad range of
contaminants
Effective for treatment
of mixed contaminants
It is a simple technology
to apply; it is cost
effective
It is a proven
technology; it has
capability of treating
broad range of
contaminants

Not effective in clays

$74-$184/ton

Widespread

The cleanup cost is high

$500-$1500/ton

Widespread

The cleanup cost is high

$90-$700/ton

Very limited

Different environmental
factors affect the
effectiveness of this
technology
Not applicable for organic
soils

$27-$310/ton

Widespread

$50-$250/ton

Widespread

2.3 Multi-criteria decision analysis (MCDA) for environmental management
Multi-criteria decision analysis (MCDA) methods have been applied to aid environmental
managers to ensure better decision by selecting the best alternative. MCDA provides a
systematic way to clarify the problems in a decision making process, and helps to evaluate the
alternatives based on the decision maker's values and preferences. Conventional decision
methods, including cost-benefit analysis, fixed target approach, and single objective linear
programming, dominated to solve multi-criteria problems until the end of the 1960s (Nijkamp
et al., 1990). These methods mainly considered cost criterion, and failed to address the issue of
multiple criteria and trade-offs among criteria. Since the early 1970s, MCDA was introduced in
order to cope with this problem. The MCDA method is a simple and intuitive approach that
helps to address potential areas of conflicts among stakeholders (Cheng 2000; Linkov et al.
2006). According to Hwang and Yoon (1981), MCDA tools fall into a group of 17 methods
based on the type and salient features of information received from decision makers (Fig. 2.2).

13

Salient feature of
information

Type of information from
the decision maker
JNO information

Dominance
Maximin
Maximax

\-

| Standard level

Multiple criteria I
decision making [

Major classes of methods

Conjunctive method
|— (Satisfying method)
Disjunctive method

Ordinal

Lexicographic method
Elimination by aspects
Permutation method

Cardinal

Linear assignment method
Simple additive weighting (SAW)
ELECTRE
TOPSIS

1

Marginal rate of
substitution

Hierarchical trade-offs

Pairwise preference —I

LINMAP
Interactive SAW method

Information on criteria [

1

information on
alternative

Order of pairwise
Nproximity

MDS with ideal point

Fig. 2.2. Taxonomy of classical MCDA methods (from Cheng, 2000)
However, these MCDA methods were modified by Hwang (1987). Three new methods
were added and six methods were removed from the list. The new added methods are
lexicographic semi order method; weighted product method (WPM) and distance from target
method. The methods permutation, hierarchical trade-offs, analytical hierarchical process
(AHP), the linear programming techniques for multidimensional analysis of preference
(LINMAP), the interactive simple additive weighting method and the multidimensional scaling
(MDS) were removed from the previous taxonomy.

14

Depending on criteria information characteristics, the information is divided into three
categories, a) Standard data: decision maker provides the minimum acceptable value for each
criterion. Either conjunctive method or disjunctive method is used at this level; b) Ordinal
data: decision maker provides ordinal data (position of the data in a series) on the criteria
weights. At this level the methods that can be applied are lexicographic, elimination by aspects
and Lexicographic Semiorder methods, and c) Cardinal data: cardinal data (opinion on relation
between criteria) on the criteria weights is provided by the decision makers. In such cases
many methods are applicable including linear assignment (LA), simple weighted addition
(SWA), elimination and choice expressing the reality (ELECTRE), technique for order
performance by similarity to ideal solution (TOPSIS), weighted product method, and distancefrom-target (DT) methods.

There is a plethora of references regarding application of various MCDA tools in
environmental management projects and related areas. For example, the allocation of Jordan
River Basin water among bordering nations was determined by ELECTRE (Bella et al., 1996).
Joubert et al. (1997), Ning and Chang (2002), Gregory and Failing (2002), and Gregory and
Wellman (2001) applied the multi-attribute utility theory (MAUT) for water and coastal
resources management projects. Al-Rashdan et al. (1999) used outranking method preference
ranking organization method for enrichment evaluations (PROMETHEE) for prioritization of
wastewater projects in Jordan. Rogers et al. (2004) used the same method to select novel
technological alternatives for sediment management. There have been numerous applications
of MAUT in environmental management projects. For instance, MAUT was applied by Arvai

15

and Gregory (2003) for identifying radioactive waste cleanup priorities at the Department of
Environment (DOE) sites. Prato (2003) also applied MAUT for selection of management
alternatives for Missouri River. In addition to MAUT, Ganoulis (2003) applied outranking
method (e.g., ELECTRE) for evaluating alternative strategies for wastewater recycling and
reuse in the Mediterranean. Mardle et al. (2002) applied AHP for analyzing priorities in fishery
management. The application of AHP can also be found in the works of Fernandes et al.
(1999), Soma (2003), Yurdakul (2004), Al-Ahmari (2007), Moran et al., (2007) and Wong et
al. (2008).

The MCDA approaches have also been combined with many other decision support
techniques to develop a number of decision support systems (DSS). For example, Hong et al.
(1991) designed a spreadsheet-based DSS integrated with SAW to perform loan approval
judgments. French (1996) applied MCDA methods to build a DSS for emergency responses on
nuclear accidents. Norbis et al. (1996) applied multi-objective integer programming to the DSS
in order to solve resource constrained scheduling problems. For resource planning, AlShemmeri et al. (1997) developed an effective monitoring system using an outranking method
(PROMETHEE) to deal with the use of water resources. Qin et al. (2006) developed a DSS for
the management of petroleum contaminated sites. Again, Hajkowicz and Higgins (2008)
applied different MCDA methods (i.e., weighted summation, range of value, PROMETHEE II,
evamix and compromise programming methods) for water management decision making
problems.

16

2.3.1 Process of MCDA
MCDA is a process of making decisions in the presence of multiple, usually conflicting
criteria (Chen et al., 1992, Figueira et al., 2005). A typical MCDA problem can be solved by
the following steps (Fig. 2.3):

Define Problems
Identify alternatives
Select criteria by which to judge
alternatives
Value judgment on relative importance
of criteria
Evaluation matrix or Application of
MCDA

^
Rank/select finaT*--^,,^
^"•""-•.^alternativesi^.
Fig. 2.3. Overall procedure of typical MCDA application (from Cheng, 2000)
In MCDA process the decision maker needs to define the problem at the beginning. Then
he/she has to identify a number of alternatives applicable to solve the problem. The next step is
to define the criteria to evaluate these alternatives. Many references indicate that defining
criteria for a problem is a difficult and time-consuming task (Hwang et al., 1981; Chen et al.,
1992 and Yoon et al., 1995). To identify the criteria representing a desired purpose, Chen et al.
(1992) suggests that the analyst should use either a deductive or an inductive approach to build

17

a hierarchy tree of criteria. A number of criteria are listed at the top level for a decision making
problem. In the next stage the criteria are divided into sub-criteria. The process is continued at
the bottom of a branch where information about the criteria is known or it is measurable.
According to Cheng (2000) such criteria hierarchy tree has several advantages to deal with
decision making problems, including (a) clarification of the intended meaning of the criteria at
higher levels; (b) enabling to consider the criteria as independent entities among which
appropriate trade-offs can be made later on, and (c) preventing undesirable double-counting of
the same criterion. For example, Bonano et al. (2000) applied a hierarchy tree of criteria for
evaluating remediation technologies by considering six major criteria (Fig.2.4). The criteria are
programmatic issues, cost, socioeconomic issues, cultural resources, environment and human
health. The programmatic issues can be divided into four-sub-criteria including time, type of
waste generated, availability of technology and reliability of technology. Similarly, other
criteria can be divided into various sub-criteria. The criteria are divided into sub-criteria until a
criterion is measurable.

After defining the criteria for the problem, the next step is data acquisition on criteria
importance. To solve a multi-criteria problem with uniform parameters or uniform data (e.g.,
only crisp data or only linguistic data) the classical multi-attribute decision making (MADM)
methods can be used. The MADM methods should be modified when mixed input parameters
are present (Chen et al., 1992).

18

Socioeconomic Issues

19

Fig. 2.4. A hierarchy of criteria for remediation alternative selection (from Bonano et al, 2000)

Transportation of waste

jProtection of ground water res.

Public acceptance

Reliability of technology

Short term risk to public &
worker health

•(Long term risk to public health

Human health & safety

Impact on low income & minority populations ^Reduction of contaminant concentration

H
Impact on
wildlife

Environment

Availability of technology

Protect cultural resources

Cultural resources

Impact on local economy

Completion cost

X

Life cycle cost

Type & quality of waste generated

Cleanup time

Programmatic issues

Selection of remediation technology

2.4 Description of three MCDA methods
The last step of MCDA process is to rank/select final alternatives through establishing the
evaluation matrix by various methods. Three existing MCDA tools will be introduced here.
2.4.1 Simple additive weighting method (SAW)
The SAW is the simplest and widely used MCDA method. In this method the overall score
of an alternative is computed as the weighted sum of the criteria values (Yoon and Hwang,
1995).
2.4.1.1 Procedure
(1) For each alternative, a score is computed by multiplying the scale rating of each criteria by
its importance weight and summing these products over all criteria;
(2) The alternative with the highest score is selected. Mathematically, the value of an
alternative can be selected as:
V(Ai) = Vi=fjwJvj(xij),

i = l,

,m; j = \,

,n

(2.1)

Where, V{At) is the value function of alternative At, and wy and v ; () are weight and value
functions of criteria ., respectively. xtj is the outcome of the ith alternative about the
j t h criterion. However, Yoon and Hwang (1995) suggest that the value of alternative A. can be
rewritten as:
n

V

i=Yuwjrir

/ = 1

'

m

>

J=l>

>n

.7=1

20

(2-2)

Where ri} is the normalized rating ofxtj. This procedure (calculation of rtj) divides the rating of
each criterion by its norm, so that each normalized rating of xtJ can be calculated as below
(Yoon and Hwang, 1995).
rv for benefit criteria (the greater the criteria value the more its preference):

r•=-!

x , - x fJ n

—

Xj

(2.3)

Xj

rtJ for cost criteria (the greater the criteria value the less its preference):
x* - x,.

Where x* is the maximum value of j ' h criteria and x™n is the minimum value of j ' h criteria.
By applying the above equations, the scale of measurement varies precisely from 0 to 1 for
each criterion. The worst outcome of a certain criteria implies rtj -0, and the best outcome
implies rtj = 1.

2.4.2 Technique for order preference by similarity to ideal solution (TOPSIS)
According to Hwang and Yoon (1981) the chosen alternative should have the shortest
distance from the ideal solution and the farthest distance from the negative-ideal solution. The
following steps are followed to apply this method (Yoon and Hwang, 1995).

21

2.4.2.1 Procedure
(1) In TOPSIS vector normalization is used for calculation ofrtj:
rti = ,

,

i = l,

,m; j = 1,

,n

(2.5)

Where rtj is normalized rating of criterion for each alternative, i is the index related to the
alternatives, j is the index related to the criteria, andx(> is rating of each criterion for each
alternative;

(2) Then the weighted normalized decision matrix is calculated. The weighted value (v,y) is
calculated as:
v =w r

v

jv

i = 1

>

,™; j = l,

'w

(2-6)

Where Wj is the weight of the j ' h criterion;
(3) Then the positive-ideal (A") and negative-ideal solutions (A~) are calculated. The
positive-ideal solution is the composite of all best criteria ratings attainable, and is calculated
as:
A* =\y\,v*2,...,v),...,v*n}

(2.7)

Where v* is the best value for the j ' h criterion among all alternatives.
The negative-ideal solution is the composite of all worst criteria ratings attainable, and is
calculated as:
A

~ =K,v2,-,vT,...,v;j

(2.8)

Where vj is the worst value for the j ' h criterion among all alternatives;
22

(4) Then the separation or distance of each alternative from the positive-ideal solution (A*) is
calculated as:

^=J2>.- V *-) 2 '

' = 1.

,«

(2-9)

Similarly, the separation from the negative-ideal solution is calculated as:

5

r=J5>#- v J) 2 '

'=i.

>m

( 2 - 10 )

(5) Then the relative closeness (C*) to the positive-ideal solution is calculated. For example,
the relative closeness of an alternative At with respect to A * is calculated as:
C* = S* /(S* +S7), With,0 < C* < 1 and i = 1,

,m

(2.11)

(6) Finally the alternatives are ranked according to C" in descending order.

2.4.3 Weighted product method (WPM)
In weighted product method a product instead of a sum of the values is made across the
criteria to penalize alternatives with poor criteria values (Easton, 1973).
2.4.3.1 Procedure
In this method, the scale rating of each criterion of each alternative is raised to a power
equal to the importance weight of the criteria. Then the resulting values are multiplied over all
criteria. The alternative with the highest product is selected. Mathematically, the most preferred
alternative A, can be calculated as:
V(A) = Vi=1[lx;< ,

i = \,...,m

(2.12)
23

Where xtj is the outcome of the i'h alternative about the j t h criterion, with a numerically
comparable scale, and Wj is the normalized importance weight of the j ' h criterion.

Alternative values obtained by this method do not have a numerical upper bound (Yoon et
al. 1995). Therefore, it is convenient to compare each alternative value with the value of ideal
alternative. The value ratio between an alternative and the ideal alternative can be shown as
(Yoon etal. 1995):
n

V(A)
R-L¥LL

V(A)

fK'
i = l,...,m

= -J2

(2.13)

ft(l.r

Where x* is the most favorable value for the j ' h criterion. And the preference of At increases
when i?. approaches 1.

2.5 Fuzzy-set theory
The conventional MCDA approaches are challenged by the uncertainties existing in various
environmental management systems. Many data on criteria might not only be in numeric form.
For example, in remediation technology selection problem, it is difficult to determine the
public acceptance level of a technology. The information about these criteria could be numeric
or linguistic (e.g., high, medium, low). The fuzzy-set theory is a powerful mathematical tool
used for modeling and controlling uncertain systems. Fuzzy MCDA methods act as facilitator
for approximate reasoning in decision making in the absence of complete and precise

24

information (Gutierrez et al., 1995). Lotfi Zadeh introduced a simple and intuitive concept of a
fuzzy-set in his seminar paper 'Fuzzy-Sets' in 1965; according to him fuzzy-set theory was
developed and extensively applied in previous decade (Zadeh, 1965). A fuzzy-set is an
extension of the traditional set theory (in which x is either a member of set A or not) and it is
defined by membership function. An example of fuzzy sets is adapted here from Kucheva et al.
(2000). If U is an ordinary set with elements ui, U2, u%

, um, a fuzzy set A on U is defined

by assigning a degree of membership between 0 and 1 to each w, e U, usually with regard to a
linguistic term. For example, let t/be the set of integers from 1 to 100 denoting the age of a
person and let ,4 be 'middle aged'. We can define a (subjective) function that assigns to each w,
a degree of membership//^(M;)G [0,l]. Degree 0 denotes non-membership and degree 1
denotes full membership. A plausible model of "middle aged" will be obtained by using a
function (membership function) that yields high values between, say 40 and 55 and gradually
decreases towards the two edges of the scale. Thus, the degree of membership of 37, jUA(37),
can be 0.75, and of 82, jilA(S2) =0.1. In Fuzzy method, vagueness and imprecision associated
with qualitative data can be represented more logically. Wang (1997) classified fuzzy theory
into five major branches (Fig. 2.5), including (1) fuzzy mathematics: classical mathematics
concepts are extended by replacing classical sets with fuzzy-sets; (2) fuzzy logic and artificial
intelligence: approximations to classical logic are introduced and expert systems are developed
based on fuzzy information and fuzzy reasoning; (3) fuzzy systems: include fuzzy control and
fuzzy approaches in signal processing and communications; (4) uncertainty and information:
different kinds of uncertainties are analyzed; and (5) fuzzy decision-making: considers
optimization or satisfaction problems with soft constraints.
25

Fig. 2.5. Classification of fuzzy theory (from Wang, 1997)

These branches of fuzzy-set theory are dependent on each other and there are strong
interconnections among them. For example, multi-criteria decision support system (MCDSS) is
in the class of fuzzy decision-making that deals with satisfaction problems and it uses the
concept from fuzzy mathematics (i.e., fuzzy-sets).

2.5.1 Handling uncertainty through fuzzy-set theory
The fuzzy-set theory can be applied to solve decision making problems with uncertainties
described by linguistic variables (Chen et al, 1992). The linguistic terms are mostly
encountered in the data acquisition of MCDA methods. Studies have shown that among the
weightings techniques the decision makers are most comfortable with ordinal (linguistic)
rankings of criteria importance (Hajkowicz et al., 2000). In the remediation technology

26

selection problem the criteria (e.g., technology availability, public acceptance, etc.) can be
adequately expressed in linguistic terms (i.e., "high", "medium", "low"). When such linguistic
terms need to be counted in alternative evaluation process, the analyst needs to consider a
fuzzy MCDA method. Nijkamp et al. (1990) suggests that input parameters containing fuzzy
information can be converted into crisp values before applying any MADM methods.

Application of fuzzy-set theory for fuzzy input transformation includes two steps. First, the
linguistic-term conversion is performed to convert the verbal terms into a fuzzy-set (Fig.2.6).

Linguistic term

Linguistic conversion
(assigned by analyst)

Fuzzy-set
Fig. 2.6. Linguistic term conversion into fuzzy-set (adapted from Cheng, 2000)

A fuzzy-set is a class of objects with a continuation of membership grades (Zadeh, 1965).
A membership function is assigned for the input value. Usually, the membership grades are
[0,1]. When the grade of membership for a value in a set is one, this object is absolutely in that
set; when the grade of membership is zero, the value does not convey any absolute significance
(Hwang etal. 1992).

27

2.5.2 Conversion of linguistic criteria preferences
A numerical approximation system was proposed by Chen and Hwang (1992) to
systematically transform linguistic terms to their corresponding fuzzy-sets. A schematic
diagram of the transformation process is shown in Fig.2.7.

Linguistic
Fuzzy-sets
terms
^ Linguistic-term
conversion

Defuzzifier

Crisp
values

t-

Fig. 2.7. Conversion of linguistic term into crisp values

According to them, the transformation requires eight conversion scales (Fig.2.8). These
conversion scales are proposed by synthesizing and modifying previous works (Baas et al.,
1977; Bonissone, 1982; Efstathiou et al., 1979; Efstathiou et al, 1982; Wenstop, 1976). It is
assumed that the given figures can cover the universe of expressing the given terms "high" vs
"low". One of the figures was applied when certain linguistic terms were provided. The
determination of the number of conversion scales is intuitive. Miller (1965) suggested that
"seven plus or minus two" of linguistic variables represent the greatest amount of information
that an observer can give about the objects based on their preference and judgment. Miller's
theory was also adopted by Chen and Hwang (1992) to develop their linguistic conversion
scales (Fig. 2.8).

28

(b)

(a)

J

J

,3 A $

,6 .? J

J

f

.? ,« J

1

Parameter value
(c)
m

,

lw

tow

4

4

,3

.4 3

.6

l

1\

A A

A

A /

X A XX *T

/ \ / \ / \ / \ / \
k

,
.1

.2

3

.4

5

J

.7

.8 J

! Y i Y i Y i Y i '-

,1 2

I

29

J

.4

J

J

,? J

J

1

(g)

.1 2

3

A S

(h)

S

.1 M 3

,t

I

2

3

A

S

£

,7

M $

I

Fig. 2.8. Linguistic terms conversion scales

In the linguistic term conversion procedure, a scale figure is selected that contains all the
verbal terms given by the decision-maker. Then the membership functions of those verbal
terms are calculated to represent the meaning of those verbal terms. If the provided verbal
terms exist in more than one figure, the simplest one should be considered. The verbal terms
used in the above eight scales are in the universe: U = { "excellent", "very high", "high to very
high", "high", "more or less high", "medium", "more or less low", "low", "low to very low",
"very low" and "none" }. This universe of verbal terms is suitable to describe technology
selection criteria like cost, maintenance, availability and community acceptability. But this
universe of verbal terms is not applicable for clean up time criteria. Because to describe the
clean up time the possible universe of linguistic terms will be U= {"extremely long", "very
long",

, "extremely short"}. This universe and the proposed universe are different. Chen

and Hwang (1992) suggest that the latter universe can be adjusted according to the nature of
the criteria used in the decision problem. Therefore, "very long" cleanup time can be treated as
"very high", and "very short" as "very low", respectively. A pair of words that represents

30

extreme meanings can always be found for evaluation of any type of criteria. Eight pairs of
opposite words are listed in Table 2.3. More examples on pairs of words can be found in
Osgood (1975).

Price

Table 2.3 Examples of linguistic universes
Size
Distance Weight Hazard Technique Experience

high

expensive

large

far

heavy

danger

advanced

good

low

cheap

small

local

light

safe

basic

poor

General

2.5.3 Conversion of fuzzy-sets into crisp weights
In order to determine a crisp score for a fuz2y-set M, it is necessary to compare the fuzzysets with a maximizing fuzzy-set (fuzzy max, Umax) and a minimizing fuzzy-set (fuzzy min,
P-min) (Chen and Hwang ,1992). The membership values of these two fuzzy-sets are calculated
by equation (2.14) and (2.15).
f x,
[0,
fl-x,
A-.(*)= n
[0,

0 <x< 1
otherwise
0<x<l
•
tu
otherwise

(2 15)

-

As well, the following Fig.2.9 shows the procedure of determining crisp values. This figure
is similar to Chen and Hwang's (1992) conversion scale, Fig.2.8 (b).

31

Medium

0.1

0.2

0.3

0.5

0.4

0.6

0.7

0.8

0.9

Fig. 2.9. Illustration of determining crisp value

In Fig.2.9 the maximum membership function (umaX) and minimum membership function
(Umin) are computed and presented as two-diagonal dashed lines respectively. The right score
[juR (M) ] refers to the intersections of the fuzzy-set M with the fuzzy max. The right score of
M can be determined using:
M

MR ( )

= SUV[MM (*) A >"max (*)]

(2-16)

Likewise, the left score [juL (M) ] of M can be determined using:
(2.17)

jUL (M) = sup[// M (x) A //min (*)]

Given the left and right scores of M, the total score [juT(M)] of M, can be calculated using:
MM)

= [juR(M) +

(2.18)

l-jUL(M)]/2

32

The membership functions of Mi, M2, M3 are (Fig.2.9):

/"M,W:

X

0<x<0.2

0.4 - x
{ 0.2 '

0.2<x<0.4

x-0.2
0.3 '
/VO) = '- 0.8-x
{ 0.3 '

/V(*)

:

x-0.6
0.2 '
1,

(2.19)

0.2<x<0.5
(2.20)

0.5<x<0.8

0.6<x<0.8

(2.21)

0.8<x<l

The set of total score [jUT(M)] can substitute the original linguistic terms.

2.6 Application of fuzzy-set approach in environmental problems
Fuzzy-set theories have been applied to a number of areas including the environmental
sciences e.g., soil, forest air pollution, meteorology and water resources (Kuncheva et al.,
2000). Kuncheva et al. (2000) proposed a fuzzy model of heavy metal loadings in Liverpool
Bay. Bender et al. (2000) applied a fuzzy compromise approach in water resource systems
planning under uncertainty. Vranes et al. (2000) developed a DSS to evaluate remedial
alternatives considering technical, financial, environmental and social criteria. The system
requires input of the numeric weight for criteria and technology performance information, thus
their system fails to incorporate uncertainty in the evaluation process, van Moeffaert (2003)
used fuzzy outranking method for choosing a sustainable wastewater treatment system in
33

Surahammar, Sweden. Li et al. (2003b) developed an integrated fuzzy-stochastic model to
examine the fate of petroleum contaminants in groundwater and risk assessment. Hu et al.
(2003) developed another rule based fuzzy expert system for hydrocarbon contaminated site
characterization. Sadiq et al. (2004a) applied an analytical hierarchy process with a technique
called fuzzy synthetic evaluation to determine the best management alternatives in a case of
petroleum drilling waste discharge scenario. Sadiq et al. (2004b) developed a fuzzy based
method to evaluate soil corrosiveness. To obtain diversified opinions of a large number of
stakeholders and to deal with uncertainties in flood management problems, Akter el al. (2005)
proposed fuzzy-set theory and fuzzy logic approaches. A risk based prioritization of air
pollution monitoring using a fuzzy synthetic evaluation technique was proposed by Khan et al.
(2005) to incorporate exposure frequency and potential hazard. Fuzzy classification combined
with spatial prediction was used to assess the state of soil pollution by Amini et al. (2005).
Najjaaran et al. (2006) developed a fuzzy expert system to assess corrosion of cast/ductile iron
pipes from backfill properties. Li et al. (2006) developed a fuzzy-set approach for addressing
uncertainties in health risk assessment of hydrocarbon-contaminated sites. The fuzzy-set
approach was also used to evaluate different alternatives against different criteria due to a lack
of crisp data. Fuzzy concepts in ranking were introduced by Baas and Kwakernak (1977), they
assumed that criteria values and the relative importance of criteria were fuzzy numbers and the
final evaluation of alternatives was computed by membership functions. Application of fuzzy
multi-criteria method (FMCM) involves the evaluation of alternative actions with respect to
multiple criteria given either in numeric or linguistic form (Altrock and Krause, 1994). The
applications of FMCM found in literature have demonstrated a number of advantages in

34

handling qualitative, unquantifiable criteria (Baas and Kwakernak, 1977; Altrock and Krause,
1994; Teng and Tzeng, 1996; Mclntyre and Parfitt, 1998; Tang et al., 1999). For example,
FMCM was applied by Chang et al. (2007) for ranking different landfill sites of Harlingen city
of Texas, USA.

2.7 Summary of literature review
There are a number of contaminated sites in Canada. As well, a number of remediation
alternatives are available in the market for remediation of contaminated sites. Meanwhile,
perpetual progress in research and competition among technology developers has created a
trend of more remediation alternatives in the future. As a result, it is difficult for a decision
maker to select the most appropriate remediation option for a site. Besides, various
stakeholders with different interests require to be involved in the decision making process of
contaminated site management. Thus, there are many uncertainties (e.g., differences in
stakeholder opinion, conflict among evaluation criteria) to be dealt with in a contaminated site
decision making process. Though there are many decisions making tools available to aid the
decision managers in their decision making process, most of these tools fail to deal with the
uncertainty issues and to consider stakeholder opinions. Incorporation of fuzzy-set theory
concept with existing multi-criteria decision analysis approach has been found to be the best
way to deal with such problems. Application of fuzzy multi-criteria approach in various
environmental decision making process has proven to be successful. However, there are limited
applications of fuzzy multi-criteria approach particularly in contaminated site management.
Though there is a plethora of literature and applications of fuzzy multi-criteria approach into
35

different decision making problems, practical application of this idea into contaminated site
management is still lacking. Considering this, in this study a fuzzy multi-criteria decision
analysis (FMCDA) approach was developed for contaminated site management. The
developed method was then applied to a case study site in northern British Columbia.

36

Chapter 3 Methodology
3.1 Overview of methodology
The following flow chart (Fig. 3.1) shows the steps that were followed to develop the fiizzy
multi-criteria decision analysis approach in this research.

Step 1 Selection of stakeholders

Step 2 Selection of evaluation criteria

~l
Selection of criteria value
range

T

Stakeholder involvment

Step 3 Membership function of
fuzzy criteria
Selection of criteria importance
weight

I

Fuzzy multi-criteria decision analysis

Step 4 Fuzzy evaluation matrix of alternatives
Aggregation
Step 5 Defuzzification
Ranking of alternatives

l_
Fig. 3.1. Overall methodology of the developed fuzzy-multi criteria decision analysis
approach
37

3.2 Step 1 Stakeholder selection
The management of a contaminated site may involve a number of stakeholders, and
different stakeholders have different perspectives, priorities and concerns. Table 3.1 presents
the common concerns of stakeholders regarding contaminated sites.

Table 3.1 Examples of typical stakeholder interests (adapted from USEPA, 2005)
Stakeholder
Interests
a) achieve regulatory compliance
Facility owner
b) utilize risk-based techniques
c) minimize/eliminate disruption of operations
d) minimize costs
e) reduce long-term treatment and liabilities
a) protect human health and the environment
Regulatory agencies
b) protect groundwater resources
c) achieve regulatory compliance
d) eliminate off-site impacts to receptors
e) involve stakeholders
f) maintain reasonable schedule
g) obtain reimbursement for oversight costs
Other stakeholders
a) optimize zoning
(Local/county agencies,
b) maximize tax revenues
property owners, special c) accelerate schedule
interest group, etc.)
d) protect human health and the environment
e) maximize quality of life
f) protect groundwater resources
From the above table, it can be seen that a wide range of interests are existing among
different stakeholders for contaminated site management. The interest of the facility owner lies
in minimizing costs; however, regulatory agencies would emphasize stakeholder involvement
or environmental resources protection. Again, people within the community may have
concerns about quality of life.
38

Therefore, the owner of a site is not the only stakeholder to be considered in the decision
making process of cleaning up a contaminated site. The principal stakeholders in the
remediation decision making process include those who have interests in the land, its
redevelopment, and impact on the environment. Bardos et al. (1999, 2000) identified the
following potential stakeholders to be considered in a contaminated site management decision
making process, including (a) site owners; (b) regulatory and planning authorities; (c) site
users, workers, visitors; (d) financial community (banks, insurers); (e) site neighbors (tenants,
dwellers, visitors); (f) advocacy organizations and local pressure groups; (g) consultants,
contractors and technology vendors; and (h) researchers.

In this research, a number of written sources related to environmental projects were
consulted to select potential stakeholders in the contaminated site management process. The
following stakeholders were contacted for their input in developing fuzzy multi-criteria
decision analysis (FMCDA) approach for contaminated site management, including (a)
institution (i.e., universities); (b) site owner (i.e., oil and gas industries); (c) federal
government; (d) provincial government; (e) research organizations; (f) non-governmental /
non-profit organizations; (g) environmental consulting firms; (h) first nation community; (i)
general public, and (j) other.

3.3 Step 2 Criteria selection
Selection of evaluation criteria is an important task in environmental decision making
problems. This is also true in case of remedial alternative evaluation for contaminated sites.
39

According to Susan (2003), the most appropriate remediation option must meet the general
criteria including practical design, feasible implementation, low impact on the landscape and
reasonable cost. The author also provided a complete list of criteria for technology selection.
Vranes et al. (2000) developed a decision analysis tool to select appropriate remediation
technologies for different contaminants. In their developed system a wide range of criteria was
applied, including overall cost, minimum achievable concentration, clean up time, public
acceptability, reliability and maintenance. Again, Bonano et al. (2000) considered six criteria
for evaluation, ranking and selection of remediation alternatives for Chromium (Cr) and
Trichloroethylene (TCE) contaminated soil. These criteria included programmatic assumptions,
cost, socio-economics, cultural, human health and archeological/historic resources. Janikowski
et al. (2000) considered time, cost, effectiveness, social acceptance and feasibility criteria to
select an alternative for pollutant management in agricultural lands.

3.3.1 Remediation alternative evaluation criteria
The following criteria were selected to develop the remediation alternative evaluation
system. These criteria were selected after surveying pertinent references, literatures, previous
case studies, professional experts and technology selection matrix of US Environmental
Protection Agency (EPA). The criteria applied in the developed approach are: (a) clean up time
(in-situ/ ex-situ); (b) overall cost (in-situ/ ex-situ); (c) minimum achievable concentration/
ability to reduce contaminant concentration; (d) community acceptability/public acceptability;
(e) availability of the technology; (f) regulatory permitting acceptability; (g) development
status; and (h) maintenance requirement.

40

3.3.2 Description of criteria
A. Cleanup time (in-situ and ex-situ)
The criterion of "time to complete cleanup" refers to the period of time required for
completion of remedial activities, including site closure time. Required clean up period is an
important criterion for technology selection. A lengthy cleanup process will increase the cost of
the operation. It will also pose a negative impact on the public's opinion around a contaminated
site. In general, the remedial alternative that requires less time for cleanup are more likely to be
selected compared to other alternatives that require a longer period of time.

B. Overall cleanup cost (in-situ and ex-situ)
Cleanup cost is the most important deciding factor in the selection of a remedial alternative
(Kamnikar, 2001). In this research, the criterion of "overall cleanup cost" includes design,
construction, operations and maintenance (O and M) costs of a remedial alternative.
Conversely, it excludes the cost of mobilization, demobilization, and pre- and post- treatment.

C. Minimum achievable concentration
"Minimum achievable concentration" criterion refers to the degree to which the technology
is able to meet remediation objectives. It can be measured by capability of contaminant
concentration reduction or efficiency of a remedial alternative.

D. Community acceptability
"Community acceptability" criterion refers to the level of technology acceptability by
41

members of the general public that live or work near the contaminated site. Acceptability of a
remedial technology in a community is an important concern in contaminated site management
problem. Therefore, it is important to consult the affected community about the type of
remedial strategy that is going to be applied to remediate a site.

E. Availability
The criterion of "availability of a remediation alternative" refers to the numbers of vendors
which can design, construct, and maintain the technology. Availability of a remedial alternative
depends on location of the contaminated site. In general, commercially available technologies
are preferred over other emerging technologies.

F. Regulatory permitting acceptability
"Regulatory permitting acceptability" for a technology affects the acceptance of a remedial
technology as part of a remediation plan. The level of acceptance can be expressed as a degree
of existing regulatory and permitting acceptability of the technology. If regulators are reluctant
to support a particular technology, it may be less attractive than those technologies that have
more support.
G. Development status of a technology
"Development status" criterion refers to the current status of the technology, (i.e.,
laboratory scale, pilot scale, full scale). Remedial alternatives which are in a lower state (i.e.
laboratory state), have a lower chance of being selected. However, these technologies should

42

not be discounted. These alternatives can attain commercial status, if applied appropriately
with alternatives in demonstrated status.

H. Technology maintenance requirement
The criterion of "technology maintenance requirement" refers to the level of complexity of
the technology and how easy it is to maintain. If high maintenance is required for a technology,
this will indicate that the technology has low reliability.

3.4 Step 3 Establishing membership functions of fuzzy criteria
Development of membership function is an important step in the application of fuzzy
theory (Turksen, 1991). There are many approaches for generating fuzzy membership
functions. However, triangular membership function is applied by many researchers to describe
vagueness, ambiguity and uncertainties in the real-world system (Civanlar and Trussel, 1986;
Lee 1996; Dou et al., 1997; Freissinet et al., 1999; Mohamed and Cote, 1999; Cheng and Lin,
2002; Li et al., 2003a; Mikhailov and Tsvetinov, 2004; Sadiq et al., 2004b; Chang et al., 2007,
Li et al, 2007).

3.4.1 Triangular fuzzy numbers (TFNs)
A triangular fuzzy number (TFN) can be defined by specifying three numbers: (1) the most
credible value, (2) the lowest possible value, and (3) the highest possible value. Fig.3.2
presents an example of a triangular fuzzy-set.
43

The most credible value is assigned a

membership value of 1, and any number that falls short of the lowest possible value or exceeds
the highest possible value will get a membership grade of 0. The intermediate membership
grades can be obtained by linear interpolation.
Most credible value

*\ Parameter value X

Fig. 3.2. Example of a triangular fuzzy-set

For example, a triangular fuzzy number N can be defined by a triplet a<b<c,

where b, a

and c are the most credible value, the lowest possible value and the highest possible value,
respectively. Then, the membership function (jU^(x)) can be defined as (Kaufmann et al.,
1985; Mikhailov and Tsvetinov, 2004; Fenton and Wang, 2006):
(x — a)/(b-a),
jU$(x) = Uc-x)/(c-b),
0,

a<x<b,
b<x<c,
Otherwise.

(3-1)

According to Khan et al. (2002) the values of alternative performance rating can be crisp,
fuzzy and/or linguistic. In the developed fuzzy multi-criteria evaluation system, TFNs were
used to convert the remediation alternative performance information into lowest possible value,
most credible value and highest possible value.
44

3.4.2 Selection of criteria value range

The evaluation criteria of remediation alternative are usually associated with uncertainty.
For example, the overall cleanup cost for enhanced bioremediation for a particular
contaminated site could vary from $50/m3 to $300/m3. In this research, such uncertainty will be
addressed through fuzzy-set approach. Each criterion was divided into five fuzzy-sets of "low",
"low to medium", "medium", "medium to high", and "high". For example, the criterion of
"overall cleanup cost" was further divided into five sets ranging from "low" cost to "high"
cost. Each fuzzy set was characterized by a triangular fuzzy number (TFN), and the highest
possible values and the lowest possible values were selected as a value range for each criterion
to develop the TFN fuzzy membership functions. The criteria value ranges were selected from
pertinent references, previous case studies, and the U.S. environmental protection agency's
remediation technologies screening matrix (USEPA, 1993). Table 3.2 lists the value ranges of
criteria.

45

C

Table 3.2 Criteria value range
^ .. .
Criteria value range
Criteria
„. .
——;—^
Cleanup time In situ
6
60
Minimum
Maximum
Ex situ (excavation) 4
12
Overall cost
In situ
50
275
Ex situ (excavation) 100
300
Minimum achievable concentration
10
90

D

Community acceptability

10

90

E

Availability

1

10

,. ,
Indicator
month
month
$/m3
$/m3
% reduced
concentration
%of
community
population
Dimensionless

F
G
H

Regulatory permitting acceptability
Development status
Technology maintenance
requirement

1
1
1

10
10
10

Dimensionless
Dimensionless
Dimensionless

A
B

¥

From the above table, it can be seen that overall cleanup cost of in-situ alternatives vary
from $50/m3 to $275/m3. And for ex-situ alternatives the costs vary from $100/m3 to $300/m3.
Even, cleanup time for in-situ treatment vary from 6 to 60 months and for ex-situ the cleanup
period vary from 4 to 12 months. However, other criteria values, including "development
status", "availability", "maintenance requirement" and "regulatory permitting acceptability"
have no unit to measure. Hence, a dimensionless and numerical scale consisting of values from
1 to 10 was applied to measure these criteria, where, 1 presents the lowest possible value and
10 presents the highest possible value. For the criterion "availability" of technology a rating of
1 means low availability, 5 means average availability and 10 means high availability,
respectively. As well, the intermediate values are used to rate other possibilities of technology
availability. Again, for the criterion "development status" a rating of 1 means the technology is
at a laboratory stage and a rating of 10 means that the technology is widely applied and it is a

46

proven technology. In the same way, for the criterion "technology maintenance requirement" a
rating of 1 means that the technology requires least maintenance and a rating of 10 means that
the technology requires higher maintenance during operation.

3.4.3 Membership functions of fuzzy criteria
The existing approach of probabilistic distributions (e.g., Bayesian approach) can not
quantify the uncertainties in evaluation criteria adequately. However, these uncertainties can be
better explained by linguistic variables and then quantified by fuzzy-set theory (Petrovic and
Petrovic, 2002; Li, 2003a). In this research, the construction of fuzzy-sets of criteria depends
on experiences of the experts and stakeholders who have in-depth knowledge on site
management (Fayek and Sun, 2001). The stakeholder opinions were collected through a
questionnaire survey, and fuzzy membership functions of evaluation criteria were then
developed based on stakeholder opinions (Appendix I). The answer with highest survey
response rate for each fuzzy-set was selected as full membership (i.e. 1). Other response rates
of that fuzzy-set were covered as lowest possible and highest possible value. Such process was
applied to address uncertainties in a fuzzy-set. However, the representative responses were
counted only to construct the corresponding fuzzy membership function.

A. Time required for cleanup
In-situ treatment
From the survey it was found that 50% of respondents indicated that the option in-situ
cleanup time required by an alternative should be approximately 1 year or less to be considered
47

as "short" cleanup time, therefore 1 year was selected as the most credible value with a
membership function of 1 in "short" cleanup time fuzzy-set. Next, 26.32% of the respondents
selected that the option in-situ cleanup time required should be approximately 18 months and 2
years to be accepted as "short to medium" cleanup time. As the percentage of respondents were
similar, an average value of 18 months and 2 years (i.e. 1.7 years) was selected as full
membership in the "short to medium" cleanup fuzzy-set. Then, 47.37% of the respondents
selected that "in-situ cleanup time required by an alternative should be approximately 3 years"
to be accepted as a "medium" cleanup period. Subsequently, 26.32% of the respondents also
selected that "in-situ cleanup time should be approximately 3 years" to be considered as
"medium to long" time period. But this response rate was similar with the response rate of
"medium" cleanup time. However, similar percentage of response (18.42%) was found for 42
months and 4 years. For these reasons, an average time period of 3 years, 42 months and 4
years (i.e.3.5 years) was selected as most credible value or full membership in the "medium to
long" cleanup time fuzzy-set. Lastly, fifty percent of the respondents selected that "in-situ
cleanup time required should be approximately 5 years or greater" to be considered as a "long"
cleanup period. For this reason, 5 years or greater cleanup period was selected as "long"
cleanup time. Table 3.3 lists the values of stakeholder opinion on in-situ cleanup time.

Table 3.3 Survey on the fuzzy-sets of in-situ cleanup time
(1) Survey on short cleanup time
Cleanup time should be approximately:
Response percentage (%)
6 months or less
23.68%
1 year or less
50.00%
18 months or less
5.26%
2 years or less
13.16%
30 months or less
0.00%
3 years or less
0.00%
48

42 months or less
4 years or less
54 months or less
5 years or less
No opinion
Total

0.00%
0.00%
0.00%
5.26%
2.635
100.00%

(2) Survey on short to medium cleanup time
Response percentage (%)
Cleanup time should be approximately:
6 months
2.63%
1 year
18.42%
18 months
26.32%
2 years
26.32%
30 months
5.26%
3 years
10.53%
42 months
0.00%
4 years
0.00%
54 months
0.00%
5 years
5.26%
No opinion
5.26%
Total
100.00%
(3) Survey on medium cleanup time
Cleanup time should be approximately:
Response percentage (%)
6 months
0.00%
1 year
2.63%
18 months
5.26%
2 years
23.68%
30 months
7.89%
3 years
47.37%
42 months
2.63%
4 years
2.63%
54 months
0.00%
5 years
5.26%
No opinion
2.63%
Total
100.00%
(4) Survey on medium to long cleanup time
Cleanup time should be approximately:
Responsi e percentage (%)
6 months
0.00%
1 year
2.63%
0.00%
18 months
2 years
2.63%
30 months
10.53%
3 years
26.32%
49

42 months
4 years
54 months
5 years
No opinion
Total

18.42%
18.42%
2.63%
7.89%
10.53%
100.00%

(5) Survey on long cleanup time
Cleanup time should be approximately:
Response percentage (%)
6 months or greater
0.00%
1 year or greater
2.63%
18 months or greater
0.00%
2 years or greater
0.00%
30 months or greater
0.00%
3 years or greater
10.53%
42 months or greater
2.63%
4 years or greater
18.42%
54 months or greater
10.53%
5 years or greater
50.00%
No opinion
5.26%
Total
100.00%
In the following Fig. 3.3 the developed membership functions of the fuzzy-sets of in-situ
cleanup time are shown. The membership functions of these fuzzy-sets (i.e., short, short to
medium, medium, medium to long and long) are developed based on the above collected data.
Short to Med.

Short

O

0.5

1

1.5

2

Med.

2.5

3

Med. to long

3.5

4

Long

4.5

5

5.5

6

Cleanup time (in year) in-situ

Fig. 3.3. Membership functions of fixzzy-sets "cleanup time (in-situ)" criterion

50

Ex-situ treatment
From the survey it was also found that 60.53% of the respondents indicated that "ex-situ
cleanup time should be approximately 4 months or less" to be considered as "short" cleanup
time; 52.63% of the respondents selected that "ex-situ cleanup time should be approximately 6
months" to be considered as "short to medium" cleanup time; 47.37% of the respondents
selected that " ex-situ cleanup time should be approximately 8 months" to be accepted as
"medium" cleanup time; 34.21% of the respondents selected that "ex-situ cleanup time should
be approximately 10 months" to be considered as "medium to long"; and 60.53% of the
respondents selected that "ex-situ cleanup time should be approximately 12 months or greater"
to be considered as long cleanup time. Table 3.4 shows the values collected from the
stakeholders on ex-situ cleanup time.
Table 3.4 Survey on the fuzzy-sets of ex-situ cleanup time
(1) Survey on short cleanup time (ex-situ)
Cleanup time should be
Response percentage
approximately:
(%)
4 months or less
60.53%
5 months or less
10.53%
6 months or less
21.05%
7 months or less
0.00%
8 months or less
0.00%
9 months or less
0.00%
10 months or less
0.00%
11 months or less
0.00%
12 months or less
5.26%
No opinion
2.63%
Total
100.00%

51

(2) Survey on short to medium cleanup time (ex-situ)
Response percentage

Cleanup time should be
approximately:
4 months
5 months
6 months
7 months
8 months
9 months
10 months
11 months
12 months
No opinion
Total

(%)

5.26%
7.89%
52.63%
10.53%
7.89%
5.26%
0.00%
0.00%
0.00%
10.53%
100.00%

(3) Survey on medium cleanup time (ex-situ)
Response percentage
Cleanup time should be
approximately:
(%)
2.63%
4 months
5 months
0.00%
6 months
10.53%
7 months
2.63%
47.37%
8 months
9 months
10.53%
10 months
5.26%
11 months
0.00%
12 months
7.89%
No opinion
13.16%
100.005
Total
(4) Survey on medium to long cleanup'time (ex-situ)
Response percentage
Cleanup time should be
approximately:
(%)
4 months
2.63%
5 months
0.00%
6 months
0.00%
7 months
7.89%
8 months
7.89%
9 months
18.42%
10 months
34.21%
11 months
2.63%
12 months
13.16%
No opinion
13.16%
100.00%
Total

52

(5) Survey on long cleanup time (ex-situ)
Cleanup time should be
Response percentage
approximately:
(%)
4 months or greater
2.63%
5 months or greater
0.00%
6 months or greater
0.00%
7 months or greater
0.00%
8 months or greater
2.63%
9 months or greater
2.63%
10 months or greater
10.53%
11 months or greater
10.53%
12 months or greater
60.53%
No opinion or greater
10.53%
Total
100.00%
Fig. 3.4 shows the developed membership functions for the fuzzy-sets of ex-situ cleanup time.
Short

Short to med.

4

5

6

Med.

Med. to long

Long

7

Cleanup time (in month) ex-situ

Fig. 3.4. Membership functions of fuzzy-sets of "cleanup time (ex-situ)"
B. Cleanup Cost
In-situ treatment
From the survey it was found that 57.89% of respondents selected that "in-situ cleanup cost
should be approximately $50 per m3 or less" to be considered as "low" cleanup cost; 26.32% of
the respondents indicated that "in-situ cleanup cost should be approximately $100 per m3 to be
considered as "low to medium" cleanup cost; 36.84% of the respondents indicated that "in-situ

53

cleanup cost should be approximately $150 per m3" to be considered as "medium" cleanup
cost; 28.95% of the respondents selected that "in-situ cleanup cost should be approximately
$200 per m3" to be accepted as "medium to high" cleanup cost; 28.95% of the respondents
selected that " the in-situ cleanup cost should be approximately $275 per m3" to be considered
as "high" cleanup cost. Table 3.5 shows the values for establishing membership functions of
in-situ cleanup cost collected through a questionnaire survey.

Table 3.5 Survey on the fuzzy-sets of in-situ cleanup cost
(1) Survey on low cleanup cost (in-situ)
Cleanup cost should be
Responise percentage
approximately:
(%)
$50 per m3 or less
57.89%
3
$75 per m or less
15.79%
$100 perm3 or less
10.53%
3
$125 perm or less
0.00%
$150 perm 3 or less
0.00%
$175 perm 3 or less
0.00%
$200 per m3 or less
0.00%
3
$225 per m or less
0.00%
$275 per m3 or less
0.00%
No opinion
15.79%
Total
100.00%
(2) Survey on low to medium cleanup cost (in-situ)
Cleanup cost should be
Response percentage
approximately:
(%)
$50 per mj
7.89%
$75 per m3
23.68%
$100 per m3
26.32%
$125 perm3
18.42%
$150 perm3
2.63%
3
$175 perm
0.00%
$200 per m3
0.00%
$225 per m3
0.00%
$275 per m3
0.00%
No opinion
21.05%
Total
100.00%
54

(3) Survey on medium cleanup cost (in-situ)
Response percentage
Cleanup cost should be
approximately:
(%)
$50 per m3
2.63%
3
$75 per m
10.53%
$100 perm3
13.16%
$125 perm3
10.53%
3
$150 per m
36.84%
$175 perm3
7.89%
$200 per m3
0.00%
$225 per m3
0.00%
$275 per m3
0.00%
No opinion
18.42%
Total
100.00%
(4) Survey on medium to high cleanup cost (in-situ)
Response percentage
Cleanup cost should be
approximately:
(%)
$50 per m3
0.00%
3
$75 per m
0.00%
$100 perm3
10.53%
$125 perm3
13.16%
3
$150 perm
2.63%
$175 perm3
18.42%
$200 per m3
28.95%
3
$225 per m
0.00%
$275 per m3
5.26%
No opinion
21.05%
Total
100.00%
(5) Survey on high cleanup cost (in-situ)
Response percentage
Clean up cost should be
approximately:
(%)
$50 per m3 or greater
2.63%
3
$75 per m or greater
0.00%
$100 perm3 or greater
0.00%
$125 per m3 or greater
7.89%
3
$150 perm or greater
13.16%
$175 per m3 or greater
5.26%
$200 per m3 or greater
5.26%
3
$225 per m or greater
18.42%
$275 per m3 or greater
28.95%
No opinion
18.42%
Total
100.00%

55

Fig. 3.5 shows the developed membership functions of the fuzzy-sets of in-situ cleanup cost.
Low

Lowtomed.

Med.

Med.tohigh

High

0.90.80.7a o.6& 0.5-

1

0.4-

0.20.1 00

25

50

75

100 125

150

175 200 225 250 275 300
3

In-situ cost ($/ m )
Fig. 3.5. Membership functions of fuzzy-sets of "cleanup cost (in-situ)"

Ex-situ treatment
55.26% of the surveyed respondents selected that "ex-situ cleanup cost should be
approximately $100 per m3 or less" to be considered as "low" cleanup cost; 28.95% of the
respondents selected that "ex-situ cleanup cost should be approximately $150 per m3" to be
considered as "low to medium" cleanup cost; 31.58% of the respondents selected that "ex-situ
cleanup cost should be approximately $200 per m3" to be considered as "medium" cleanup
cost; 21.05% of the respondents selected that "ex-situ cleanup cost should be approximately
$225 per m3" to be considered as "medium to high" cleanup cost; 36.84% of the respondents
selected that "ex-situ cleanup cost should be approximately $300 per m3" to be considered as
"high" cleanup cost. Table 3.6 shows the values collected through the questionnaire survey on
ex-situ cleanup cost.

56

Table 3.6 Survey on the fuzzy-sets of ex-situ cleanup cost
(1) Survey on low cleanup cost (ex-situ)
Ex-situ cleanup cost should be
Response percentage
approximately:
(%)
$100 per m3 or less
55.26%
$125 perm 3 or less
26.32%
3
$150 perm or less
5.26%
$175 perm 3 or less
0.00%
$200 per m3 or less
0.00%
$225 per m or less
0.00%
$250 per m3 or less
0.00%
$275 per m3 or less
0.00%
3
$300 per m or less
0.00%
No opinion
13.16%
100.00%
Total
(2) Survey on low to medium cleanup cost (ex-situ)
Cleanup cost should be
Response percentage
approximately:
(%)
$100 perm3
13.16%
3
$125 perm
26.32%
$150 per m3
28.95%
$175 perm3
13.16%
$200 per m3
2.63%
$225 per m3
0.00%
3
$250 per m
0.00%
$275 per m3
0.00%
$300 per m3
0.00%
No opinion
15.79%
Total
100.00%
(3) Survey on medium cleanup cost (ex-situ)
Cleanup cost should be
Response percentage
approximately:
(%)
$100 perm3
2.63%
3
$125 perm
7.89%
$150 perm3
18.42%
$175 perm3
13.16%
$200 per m3
31.58%
$225 per m3
0.00%
3
$250 per m
10.53%
$275 per m3
0.00%
$300 per m3
0.00%
No opinion
15.79%
Total
100.00%
57

(4) Survey on medium to high cleanup cost (ex-situ)
Cleanup cost should be
Response percentage
approximately:
(%)
$100 perm3
0.00%
$125 perm
2.63%
$150 perm3
7.89%
$175 perm3
18.42%
$200 per m3
5.26%
$225 per m3
21.05%
$250 per m3
13.16%
$275 per m
10.53%
$300 per m3
0.00%
No opinion
21.05%
Total
100.00%
(5) Survey on high cleanup cost (ex-situ)
Cleanup cost should be
Response percentage
approximately:
(%)
$100 per m3 or greater
2.63%
$125 per m3 or greater
0.00%
$150 perm3 or greater
5.26%
$175 per m3 or greater
5.26%
$200 per m3 or greater
13.16%
$225 per m3 or greater
5.26%
3
$250 perm or greater
2.63%
$275 per m3 or greater
15.79%
$300 per m3 or greater
36.84%
No opinion
13.16%
Total
100.00%

By using the above data fuzzy membership function was developed for ex-situ cleanup
cost. Fig.3.6 shows the developed membership functions of the fuzzy-sets of ex-situ cleanup
cost criterion.

58

Low to med.

0

25

50

75

Med. Med. to high

100 125 150 175 200 225 250 275 300 325
3

Ex-situ cleanup cost ($/ m )

Fig. 3.6. Membership functions of fuzzy-sets of "cleanup cost (ex-situ)"
C. Minimum achievable concentration
From the survey it was found that 26.32% of respondents indicated that the option of
"contaminant concentration should be reduced by approximately 10% or less" to be considered
as "low" minimum achievable concentration; 31.58% of the respondents selected that
"contaminant concentration should be reduced by approximately 40%" to be considered as
"low to medium" category; 36.84% of the respondents selected that the option of "contaminant
concentration should be reduced by approximately 50% to be considered as "medium"
category; 42.11% of the respondents selected that "contaminant concentration should be
reduced by approximately 70%" to be considered as "medium to high" category; and 76.32%
of the respondents selected that "contaminant concentration should be reduced by
approximately 90% or greater" to be accepted as "high" category. Table 3.7 shows the values
on minimum achievable concentration collected from the stakeholders.
59

Table 3.7 Survey on the fuzzy-sets of minimum achievable concentration
(1) Survey on low minimum achievable concentration
Contaminant concentration should be
Response percentage
reduced by approximately;
(%)
10% or less
26.32%
20% or less
18.42%
30% or less
21.05%
40% or less
2.63%
50% or less
21.05%
60% or less
0.00%
70% or less
2.63%
80% or less
0.00%
90% or less
0.00%
No opinion
7.89%
Total
100.00%
(2) Survey on low to medium minimum achievable concentration
Contaminant concentration should be
Response percentage
reduced by approximately:
(%)
10%
5.26%
20%
0.00%
30%
18.42%
40%
31.58%
50%
13.16%
60%
21.05%
70%
0.00%
80%
0.00%
90%
0.00%
No opinion
10.53%
Total
100.00%
(3) Survey on medium minimum achievable concentration
Contaminant concentration should be
Response percentage
reduced by approximately:
(%)
10%
2.63%
20%
0.00%
30%
0.00%
40%
0.00%
50%
36.84%
60%
23.68%
70%
26.32%
80%>
90%
No opinion
Total

2.63%
0.00%
7.89%
100.00%

60

(4) Survey on medium to high minimum achievable concentration
Contaminant concentration should be
Response percentage
reduced by approximately:
(%)
0.00%
10%
20%
0.00%
30%
0.00%
40%
0.00%
50%
0.00%
60%
10.53%
70%
42.11%
80%
31.58%
90%
5.26%
No opinion
10.53%
Total
100.00%
(5) Survey on high minimum achievable concentration
Contaminant concentration should be
Response percentage
reduced by approximately:
(%)
10%> or greater
0.00%
20% or greater
0.00%
30%) or greater
0.00%
40%) or greater
0.00%
50%) or greater
0.00%
60%o or greater
0.00%
70%o or greater
0.00%
80%) or greater
15.79%
90% or greater
76.32%
No opinion
7.89%
Total
100.00%

The membership functions of these fuzzy sets were developed based on the collected data.
Fig. 3.7 represents the developed membership functions for the fuzzy-sets of minimum
contaminant concentration reduction criterion.

61

Low

Low to Med. Med

Med. To high

High

0.90.8a 0.7| 0.6% 0.5| 0.4S 0.30.20.1 00

10

20

30

40

50

60

70

80

90

100

Ability to reduce contaminant concentration (percentage)
Fig. 3.7. Membership functions of fuzzy-sets of "minimum achievable concentration"
criterion
D. Community acceptability
From the survey it was found that 34.21% of respondents indicated that the option of
"community acceptability should be approximately 40% or less" to be accepted as "low"
community acceptability; 42.11% of the respondents selected "community acceptability should
be approximately 50%" to be accepted as "low to medium" community acceptability; 36.84%
of the respondents selected " community acceptability should be approximately 60%" to be
accepted as "medium" community acceptability; 42.11% of the respondents indicated that the
option of "community acceptability should be approximately 70%" to be accepted as "medium
to high" community acceptability and 42.11% of the respondents selected that "community
acceptability should be approximately 90% or greater" to be accepted as "high" community
acceptability. Table 3.8 shows all the values expressed by stakeholders on community
acceptability criteria.
62

Table 3.8 Survey on the fuzzy-sets of community acceptability levels
(1) Survey on low community acceptability
Acceptance of the remedial alternative Response percentage
by community should be
(%)
approximately:
10% or less
21.05%
20% or less
7.89%
30% or less
21.05%
40% or less
34.21%
50% or less
10.53%
60% or less
0.00%
70% or less
0.00%
80%) or less
0.00%
90% or less
0.00%
No opinion
5.26%
Total
100.00%
(2) Survey on low to medium community acceptability
Acceptance of the remedial alternative Response percentage
by community should be
(%)
approximately:
10%
0.00%
20%
7.89%
30%
15.79%
40%
10.53%
50%
42.11%
60%
15.79%
70%
0.00%
80%
0.00%
90%
0.00%
No opinion
7.89%
Total
100.00%
(3) Survey on medium community acceptability level
Acceptance of the remedial alternative Response percentage
by community should be
(%)
approximately:
10%
0.00%
20%
0.00%
30%
5.26%
40%
7.89%
50%
28.95%
60%
36.84%
70%
13.16%
63

80%
90%
no opinion
Total No. of Respondents

0.00%
0.00%
7.89%
100.00%)

(4) Survey on medium to high community acceptability level
Acceptance of the remedial alternative Response percentage
by community should be
(%)
approximately:
10%
0.00%
20%
0.00%
30%
0.00%
40%
0.00%
50%
10.53%
60%
15.79%
70%
42.11%
80%
23.68%
90%
0.00%
No opinion
7.89%
Total
100.00%
(5) Survey on high community acceptability
Acceptance of the remedial alternative Response percentage
by community should be
(%)
approximately:
10% or greater
0.00%
20% or greater
0.00%
30%> or greater
0.00%
40%) or greater
0.00%
50% or greater
5.26%
60% or greater
5.26%
70% or greater
5.26%
80% or greater
36.84%
90% or greater
42.11%
No opinion
5.26%
Total
100.00%

According to Chen and Hwang (1992), the membership functions of these five fuzzy sets
can be constructed based on the collected data. Fig. 3.8 shows the developed membership
functions of the fuzzy-sets of community acceptability criterion. For example, if community
64

acceptance is 80% then it could be categorized as partly (0.5) "medium to high" and partly
(0.5) "high".
Lo.tomed. Med. Med.tohigh

Low

0

10

20

30

40

50

60

70

80

High

90

100

Community acceptance
Fig. 3.8. Membership functions of fuzzy-sets of "community acceptability"
Other criteria such as technology availability, regulatory acceptability, development status
and technology maintainability are mostly technical criteria. For this reason, these criteria were
not included in the survey. Membership functions of these criteria were selected from
references and with consultation of experts (USEPA, 1993; Soesilo, 1997).

E. Technology availability and other criteria
As rating of technology availability criteria has no unit for measurement, a dimensionless
scale of 1 to 10 was adopted. The values between one and ten were divided into 5-tuple fuzzy
sets. The range of "medium" category was considered from three to seven. Again, values
falling below three were categorized as "low". And values above seven were categorized as
"high". Fig.3.9 shows the developed membership functions for technology availability
65

criterion. Similarly, regulatory acceptance, development status and technology maintenance
criteria were fuzzified into five-tuple ("low", "low to medium", "medium", "medium to high"
and "high") fuzzy-sets. Fig. 3.10, 3.11 and 3.12 shows the developed membership functions of
these criteria.
Low

0

Low to med.

1

2

Med.

High

Med. to high

3
4
5
6
7
Level of technology availability

8

9

10

Fig. 3.9.. Membership functions of fuzzy-sets of "technology availability" criterion

Med.tohigh

Low to med.

.9*
U

!

7

10

Level of regulatory acceptance
Fig. 3.10. Membership functions of fuzzy-sets of "regulatory acceptability" criterion
66

Lowtomed.

Med. to high

I
10
Development status
Fig. 3.11. Membership functions of fuzzy-sets of "development status" criterion

Med. to high

Low to med.

i-

1
10
Required maintenance
Fig. 3.12. Membership functions of fuzzy-sets of "maintenance requirement" criterion

Though different criteria uses different linguistic terms in the fuzzy performance scale
figure (e.g., "low", "low to medium", "medium", "medium to high", "high", "short", "short to
medium", "medium", "medium to long" and "long") all fuzzy-sets are categorized using

67

common linguistic variables (i.e., excellent, good, fair, poor and bad). Such conversion process
was developed and applied by Chen and Hwang (1992). Table 3.9 lists the values of developed
membership functions for all criteria. From this table it can be observed that the fuzzfied value
range of "excellent" and "bad" fuzzy-sets follow two sequences. For example, "excellent"
fuzzy-set is defined by the highest values for certain criteria (e.g., technology availability,
community acceptability, minimum achievable concentration). And the sequence of
fuzzification for these criteria is highest values as "excellent" fuzzy-set and lowest values as
"bad" fuzzy-set. On the other hand, for certain criteria "excellent" fuzzy-set is defined by the
lowest values (e.g., cleanup cost, cleanup time). And the sequence of fuzzification for these
criteria are opposite of the previous fuzzification sequence.

68

3

Development status
(0 to 10)
Technology Maintenance
requirement
(0 to 10)
1,1,3

69

1, 3, 5

9,7,5

9,7,5

9,9,7
9,9,7

9,7,5

9,9,7

Availability
(0 to 10)
Regulatory permitting
acceptability (0 to 10)

(%)

90,90,70

Community acceptability

90,70,60

3, 5, 7

7,5,3

7,5,3

7,5,3

70,60,50

5,7, 9

5,3,1

5,3,1

5,3,1

60,50,40

7,9, 9

3,1,1

3,1,1

3,1,1

50,40,40

Table 3.9 Triangular fuzzy numbers (TFNs) defined for development of fuzzy membership functions of remediation
alternative evaluation criteria
Criteria
TFNExceiient
TFNGood
TFNFair
TFNPoor
TFNBad
Cleanup time
1,1,2
1,1.7,2.5
1.5,3,3.5
2.5,5,5
3.5,5,5
(in-situ, in year)
4,6,8
8,10,12
10,12,12
Cleanup time
4,4,6
6,8,10
(ex-situ, in month)
Overall cost
50,50,100
50,100,150
100,150,200 150,200,275 200,275,275
(in-situ, $/m3)
Overall cost
100,100,125 100,125,200 125,200,225 200,225,300 225,300,300
3
(ex-situ, $/m )
90,90,70
90,70,50
50,40,10
Minimum achievable
70,50,40
40,10,10
concentration (%)

3.5 Step 4 Developing fuzzy multi-criteria evaluation matrix
A fuzzy multi-criteria evaluation matrix was developed for remedial alternative evaluation.
In a general setting, the process of fuzzy multi-criteria analysis (FMA) can be conveniently
described by pointing out relationships between a collection of pattern features and their class
membership vectors. A fuzzy multi-criteria decision problem with m alternatives
m) and n criteria

Aiiy=\,...,

Cy (7 =1,...,n) can be concisely expressed as:D = [S^.J and W = (w-J,

where D is the fuzzy decision matrix, x represents the fuzzy rating of alternative At with
respect to criterion Cj, W is the weight vector, w is the fuzzy weight of criterion Cs.

3.5.1 Conversion of linguistic variables
In this research, stakeholder opinion on criteria was collected through a questionnaire
survey. Each stakeholder expressed their opinion through linguistic variables. Then, the
linguistic variables were converted into crisp values by using Chen and Hwang's (1992)
conversion method. The conversion method of linguistic variables is discussed in details in
chapter two. A scale was chosen from the eight scales (Chen and Hwang, 1992) that contained
all the linguistic variables given by a respondent. For example, if a respondent preferred the
variables "very low", "low medium", "high" and "very high", to rate all the criteria, scale three
was selected for conversion of those linguistic variables (Fig. 3.13).

70

Very low

Medium

Very high
]i#s)

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Fig. 3.13. Conversion of linguistic variables by scale three (Chen and Hwang, 1992)

In this Fig.3.13 five linguistic variables ("very low", "low", "medium", "high", and "very
high") are presented as five fuzzy-sets. The membership function of each fuzzy-set is shown as
a continuous line. The y axis presents the membership function values and the x axis presents
the values for each fuzzy set. The left score [jUL (M) ] and right score [juR (M) ] for each fuzzyset are shown as dashed lines. Consequently, Table 3.10 represents the converted values of
linguistic variables. From the table it can be seen that "very low" received numeric value of
0.0945 and "very high" linguistic variable received 0.9055.

Table 3.10 Determination of criteria value using scale three
i
Linguistic variables n«(Mi) UL(MJ)
UT(MJ)
1
Very low
0.189
1
0.095
2
Low
0.350
0.775
0.288
3
Medium
0.588
0.588
0.500
4
High
0.775
0.350
0.713
5
Very high
1
0.189
0.906

71

Similarly, scale six was applied when a respondent used the linguistic variables "Very
low", "Low", "Low to medium, "Medium", "Medium to high", "High" and "Very high" to rate
all the criteria. Fig.3.14 shows the conversion process.

Very low

low

more or
less low Medium

0

0.2

0.3

0.1

0.4

0.5

more or
less high

high

Very high

0.6

0.8

0.9

0.7

1

Fig. 3.14. Conversion of linguistic variables by scale six (Chen and Hwang, 1992)
The converted criteria importance weights are shown in Table 3.11. These weights were
calculated by using scale six. In this scale, linguistic variable "very low" and "high" has a value
of 0.0875 and 0.9125 respectively.
Table 3.11 Criteria value determination using scale six
i
1
2
3
4
5
6
7

Linguistic variables
Very low
Low
Low to medium
Medium
Medium to high
High
Very high

HR(M0

0.175
0.275
0.463
0.538
0.725
0.813
1

72

HiXMi)
1
0.813
0.725
0.538
0.45
0.275
0.175

HT(M0

0.0875
0.231
0.369
0.5
0.6375
0.769
0.9125

The crisp values for all the 8 conversion scales are shown in Table 3.12. According to
Chen and Hwang (1992) the principle of this conversion procedure is simply to pick a figure
that contains all the verbal terms given by the decision maker and then use the fuzzy
membership function from the figure to represent the meaning of those verbal terms.

Table 3.12 Conversion of linguistic terms into crisp values
Scale
Linguistic
UL
UT
m
variables
1
Medium
0.66
0.5
0.58
High
0.32
0.75
0.82
1
0.16
2
Low
0.32
0.62
0.5
Med
0.62
0.32
0.84
High
1
4
Low
0.2
1
0.1
0.3
Med. low
0.4
0.8
Med. low
0.6
0.5
0.6
Med. high
0.4
0.7
0.8
High
0.2
0.9
1
5
Very low
1
0.09
0.18
Low
0.82
0.23
0.28
More or less low
0.72
0.35
0.42
Medium
0.52
0.52
0.5
More or less high
0.72
0.42
0.65
High
0.82
0.28
0.77
Very high
1
0.18
0.91
1
0.09
7
V. low
0.18
Low to very low
1
0.125
0.25
Low
0.82
0.265
0.35
Med low
0.42
0.7
0.36
0.58
0.5
Med
0.58
0.64
Med. high
0.42
0.7
High
0.82
0.35
0.735
High to very high
0.875
1
0.25
V. high
1
0.18
0.91
8
None
0.09
1
0.045
V. low
0.14
0.18
0.9
Low to very low
0.42
0.9
0.26
0.335
Low to very low
0.42
0.75
Med. Low
0.62
0.415
0.45
73

Med. Low
Med. High
High
High to very high
Very, high
Excellent

0.58
0.62
0.75
0.9
0.9
1

0.58
0.45
0.42
0.42
0.18
0.09

0.5
0.585
0.665
0.74
0.86
0.955

From Table 3.12 it can be observed that similar linguistic variables have different values in
different scale. For example, values for "medium" and "high" are 0.58 and 0.75 respectively in
scale one. Again, in scale two the values for "medium" and "high" are 0.5 and 0.84
respectively.

A questionnaire survey (Apendix I) was conducted to obtain the criteria importance
weights. Criteria importance weights were first collected through linguistic variables. Then,
these linguistic variables were converted into crisp weights. Subsequently, an average
importance weight was calculated for each of the criterion. For example, the average
importance weight of criterion "overall cost" was calculated by adding all the importance
weights given by each stakeholder and then the total importance weight was divided by the
number of respondents. The sum of importance weights of "overall cost" given by all the
stakeholders was 26.622 and total number of respondents was 38, therefore the average
criterion importance weight of "overall cost" is (26.622 / 38) = 0.701. Similar procedure was
applied to calculate the average criteria importance weight of other criteria. Fig. 3.15 shows the
average criteria importance weights of all criteria. From this figure it can be observed that,
most of the stakeholders provided high importance weight on regulatory acceptability criterion.

74

Such emphasis on regulatory acceptability criterion reflects that, the stakeholders are
concerned that a remediation technology is implemented with in compliance with existing rules
and regulations. As well, minimum achievable concentration and community acceptability
criteria were rated as the next most important factors in the remediation alternative evaluation
process. However, the criterion development status was rated as lowest important factor among
all the criteria.

3

1.000
0.900
0.800

0.116

If 0.700 -I Q*638
*
8
|

0.600
0.500
0.400

I

0.300

0-701

& 0.200
* 0.100
0.000

•&S?

<r

>cr

^°

0:694

0:694
H

-—
0.622

H

—
Wm

0:605

o;669
=

x•&'
f

&
r&

J$<y

Ji
d*

«£
<s?

<*

iiiln
°°

#

,*°

G

^°

^

^

<?'
^

Criteria
Fig. 3.15. Average criteria importance weights
To use the average criteria importance weights in the developed fuzzy multi-criteria
approach the criteria importance weights were normalized. Otherwise the membership
functions in the fuzzy-sets may have a membership of more than 1 in the fuzzy evaluation
matrix results. The normalization procedure is conducted by dividing each criterion importance
weight by the summation of all criteria importance weight. For example, the sum of all criteria
importance weight is 5.399, so the normalized criterion weight of "technology availability" is

75

0.622 / 5.399 = 0.115. Similarly, the normalized criteria importance weights of other criteria
are obtained. The normalized criteria importance weights are 0.118, 0.130, 0.129, 0.129, 0.144,
0.112 and 0.124 for "cleanup time", "overall cost", "minimum achievable concentration",
"community acceptability", "regulatory permitting acceptability" and "technology maintenance
requirement", respectively.

3.5.2 Fuzzy evaluation matrix
In this research fuzzy evaluation matrix was developed for each alternative. The procedure
of fuzzy evaluation matrix is shown below:
Fuzzy-set (Bi) for an alternative =
MFO,\

Mp0,\ MBO,\

MFO,2

MEI,2

MGO,2

Mpo,2

MBU,2

MEI,3

MGO,T, M-Fa,3 Mp0,3

MBO,3

[ wl w2 w3 wA w5 w6 w1 w% ] • MEIA

MGOA

MFUA

M-POA M-BaA

ME1,5

MGO,S

MFO,5

Mp0,5

MBU,5

MEI,6

MGO,6

Mpa,6

Mp0,6

MBO,6

MEI.1

MGOJ

MFU,7

MPOJ

MBOJ

MEI.S

t*Go$ M-Fa,% Mpo,S MBO,8

Where i = 1,2,

,8; Wi represents normalized criterion importance weight; piEli is the

membership function of the fuzzy-set "excellent" of criteria i; juGoi is the membership function
of the fuzzy-set "good" of criteria i; juFai is the membership function of the fuzzy-set "fair" of
criteria i; juPo t is the membership function of the fuzzy-set "poor" of criteria i; and juBal is the
membership function of the fuzzy-set "bad" of criteria i.
76

Five membership values were generated for each alternative through the above matrix
multiplication.
Final fuzzy-set (Bi) for an alternative = [W • A]

(3.3)

These five membership (MEiMcoMFaMpoMBa) values are the aggregated values of an
alternative.

3.6 Step 5 Defuzzification
A two step process was applied for defuzzication of final fuzzy-sets and to get a utility
value of an alternative. In the first step, fuzzy-sets were normalized such that the cardinality of
the fuzzy-sets was in unity. In second step, the normalized five tuple fuzzy-sets were converted
into a utility value using Equation (3.4).

Ua = max(jUEljuGo/lPa/iPo[iBa)

(3.4)

Where Ua is the predominant level that is the highest value of membership. This value decides
the classification of the fuzzy-set. Again, Ua with the highest value represents the best
management alternative among all.

77

3.7 Step 6 Development of a decision support system
A decision support system was built by using Java script computer program to make the
developed fuzzy multi-criteria decision analysis approach user-friendly. There are three frames
in the developed system. These three frames are used for data input, data evaluation and data
presentation purposes. On the first frame the user has to select remedial technologies to be
evaluated for a contaminated site. Fig. 3.16 shows a screen shot of the first frame in the
developed system.
I<fc«5Ki»*»«II»*f«»K» tcHrftximotoggiieas l i s t ?

Iwi-- s i t u
f'l

Bioventing
Entranced bioretnediation
Pbytansmediatiofi

f n - -Stlftll f*i*;3'»lL«r»l/«;I*«?irttic:fls.l

DD0D1Q

l ^ t c s e r t r o l c i r u s t i c ; s&epjpsursatiork
S o i l flutsHirigg
S k » i l v a p o r oscta-sMstiont
S t o l i d i f* o j a * i ojnt/sstmfc*i iisK^ttrs o n

I . i -slt-u I h e r m a l
I

t**«ta*t«»€;**Jt

t r e a t m e n t

! Tnh«ssrm«I

tteatment

i : « - ~S»itKft i i » t e j j " " » f « l

t r e a t m e n t

S o i l vtxpar
tsx t r a c t i o n *
A i r sparging
L :, I ' n e u i T V B t i o f t a t o t u r t n g ;

orti\&rnctematGmt:

l i 2 x ~ » I * M . i > t o l o t g £ i « : » l ***<e«**ii**«swfc* <«jKcrs»-vj»iJ«*»i>

con
E x -

»it*» p.h>'!iit-a»l/«:h*sim»c»I

trvMMhnaMHt*

k. .1 C2M«szrki«£s*l « w « t r * % e H o r *
(TH C h e m i c x l r e < J k » o t i o r * / a x * < d » t i a t t
T ! SUsfMurattioR
l_J S o i l
w u M n g
E x H r i M

«l»*?t~**iii«*l c*-«?*fc*»riNewt < * ; * * * * * % ' » * * « * . * * >
JHU»t i&saes <cl€5r4^c>rmt£Mmii3nt»tito£it
0|»oar* tyonri

i" i

m-x- ~«Ift»ji «so*»*:«ijBi*iaM£i**

r: :• L j u i d f f l l

osip

I *<M»X« |

Fig. 3.16. Selection of remedial alternatives
78

Then the user needs to input information on all criteria for each technology. Fig. 3.17
shows the process of criteria value input.

Remidial
Alternatives

Development
Required Technology Community
status of
maintenance availability acceptance
technology

8,9,10
Bioventing
Enhanced
8,9,10
bioremediation
Soil vapor
8,9,10
extraction
Landfarming 85,10
Slurry phase
8,9,10
biological
treatment
Thermal
M.10
desorption
tttext]

1A3

6,7,10

80,90,100

Ability to
Regulatory
reduce
acceptance contaminant
concentration
5,6,7

80,90,100

Time

Cost
26,70,80

12,3

•

8,75,10

80,90,100

1,2.5,3

50,65,70

50,1250,150

05,2,3

J 1,2,3

6,7.5,10

80,90,100

8,9,10

50.60,70

50,125,150

0.5,0.9,1

15,3

6,7.6,10

50,60,70

5,6,7

50,60,70

75,100,125

6,9,12

5,6,7

6,7,10

50,60,70

8,9,10

50,60,70

200,250,3(10

6,9,12

5,6,7

6,7,10

;$9,60,70

5,6,7

80,90,100

200.250,300

6,9,12

5,6.5,7
,

„

Fig. 3.17. Criteria data input in the system
On the third screen the results for all alternatives are shown in a ranking order (Fig. 3.18).
Final Results
I - Utility

Remidiation Technology
Thermal desorption
Slurry phase biological treatment
Enhanced bioremediation

[0.299

Landfarming

'0J45

iSoil vapor extraction

!o.sos

Bioventing

10.473

Rank

j

6

!

"£
"T

!
!

"i2

"1

Fig. 3.18. Ranking order of remedial alternatives
In summary, the overall computation process by the developed system can be viewed in the
following flow chart (Fig.3.19). The first task in the system is to gather site information. The
site information is stored in the system as a database and for future references. Then, a list of
79

potential remediation alternatives is prepared for the site. As well, information on each
alternative is collected from expert opinion, pertinent references and case studies for evaluation
and ranking of these alternatives. Since, there are 8 evaluation criteria in the developed system
the user has to provide information on those criteria for evaluation and ranking.

R e m e d i a t i o n techniques
investigation

Site investigation

R e mediation
alternatives

Remediation alternative evaluation

r
i

<L>
E-i

>

<_> ca

HI

<L>

7}
o>

e

73

O"

o

2

1\

s

&

ity

a

I

*
J =

uirem

JP

aila

chn

o

•S"

epta

63.S-

I
lat

1

•a

a, &
.3
Pi

C3

Ranking of alternatives
Utility v a l u e s and
r a n k i n g of alternatives

Fig. 3.19. Remediation decision process by the system

80

Chapter 4 Case Study, Results and Discussions

A contaminated site was selected to apply the developed method. The obtained results were
compared with three existing MCDA methods, including SAW, TOPSIS and WPM.

4.1 Description of study site
The study site is located about 150 km north of Fort St. John in northern British Columbia.
An oil leak was discovered in May, 2002. The leaked volume of crude oil was recorded about
200 m3. Crude oil spilled over an area of approximately six hectares. The site is surrounded by
agricultural land. This spill area is within the traditional use area of First Nations and
approximately 80 kilometer from the Doig Reserve. The volume of contaminated soil was
estimated to be 10,000 m3 (13,080 yd3). Contaminant concentration was measured higher than
the provincial guideline value of Total Petroleum Hydrocarbon (TPFf) level. Identified
contaminants were Benzene, Ethylbenzene, Toluene and Xylene (BTEX). After primary site
investigations and site characterization a list of potential remedial alternatives were selected to
recover the contaminated soil. After screening out the list of available remedial alternatives, the
following alternatives were selected for the above case, including (a) enhanced bioremediation
(in-situ); (b) bioventing (in-situ); (c) soil vapor extraction (SVE; in-situ); (d) landfarming (exsitu); (e) slurry phase treatment (ex-situ), and (f) low temperature thermal desorption (LTTD,
ex-situ). There are three in-situ and three ex-situ alternatives in the potential alternative list.
The site manager has to evaluate these alternatives to select the most appropriate remediation
alternative from this list by considering stakeholder involvement and other uncertainties.

81

4.2 Description of remedial alternatives in the system
A brief description on remedial technologies was stored in a database as a source of
reference and to assist the users of the developed system (Appendix II). The remedial
alternatives were divided into two categories, ex-situ and in-situ. The user can refer to this
database for background information on technology and technology evaluation criteria. This
reference will help the users to measure the performance of each alternative for each criterion
and then decide the input values.
4.2.1 Information about remedial alternatives
Information on these alternatives was collected based on previous application results, case
studies and from remedial alternative evaluation matrix. Table 4.1 lists information on remedial
alternative used for the case study site.

Criteria
Technology availability
(1-10)

Table 4.1 Information on remedial alternatives
In-situ
A
B
C
D
More
than 4
vendors
Better

More
than 4
vendors
Better

More
than 4
vendors
Better

More than
4 vendors

Ex-situ
E

F

More
More than
than 4
4 vendors
vendors
Average Average

Community
Average
acceptability (10-100%)
Min.ach.concentration
Average Better
Average Average
Average Better
(10-100%)
Development status
Full
Full
Full
Full
Full
Full
(1-10)
Maintenance req. (1-10) Average Low
Low
Low
Average Average
Time to clean up
0.5-3
1-3
0.5-1
0.5-1 year 0.5-1
0.5-1 year
(month or year)
years
years
year
year
Overall cost ($/ton)
Average Low
Average Low
High
High
Reg, acceptability (1-10) Worse
Average Better
Average
Better
Average
A= in-situ enhanced bioremediation, B= in-situ bioventing, C= in-situ soil vapor extraction (SVE), D= ex-situ
landfarming, E= ex-situ slurry phase, F= ex-situ low temperature thermal desorption (LTTD)
82

Most of the remediation alternative information was in linguistic form. Therefore, this
information is required to be processed before applying in the developed system.

4.3 Fuzzy processing of criteria information
As the developed fuzzy multi-criteria approach is capable of dealing with a range of values,
it was not mandatory to input a single crisp value for each criterion. Triangular fuzzy numbers
(TFNs) are an effective way to represent uncertain values. Table 4.2 represents the converted
criteria values.

83

H

G

F

E

D

C

B

A

5

8

1

6

80%

50%

$50

0.5yr

6.5

9

2.5

7.5

90%

65%

$125

2yr

7

10

3

10

100%

70%

$150

3.5yr

1

8

5

6

80%

80%

$25

lyr

2

9

6

7

90%

90%

$70

2yr

3

10

7

10

100%

100%

$80

3yr

Max

1

8

8

6

80%

50%

$50

0.5yr

Min

2

9

9

7.5

90%

60%

$125

0.9yr

MLV

84

Min=Minimum, MLV=Maximum Likely Value, Max=Maximum, yr=year, m=month.

Regulatory
acceptability
(1-10)
Dev. status
(1-10)
Maintenance
requirement
(1-10)

Cleanup time
(month/year)
Overall cost
($/m3)
Min.achievable
concentration
(10-100%)
Community
acceptability
(10-100%)
Availability
(1-10)

MLV

Min

Max

Min

MLV

Bioventing

Enhanced bio
remediation

3

10

10

10

100%

70%

$150

lyr

Max

Soil vapor extraction

1

8

5

6

50%

50%

$75

6m

Min

2

9

6

7.5

60%

60%

$100

9m

MLV

Landfarming

Table 4.2 Input values of remedial alternatives
In-situ

3

10

7

10

70%

70%

$125

12m

Max

5

8

8

6

50%

50%

$200

6m

Min

6

9

9

7

60%

60%

$250

9m

MLV

Slurry phase

Ex-situ

7

10

10

10

70%

70%

$300

12m

Max

5

8

5

6

50%

80%

$200

6m

Min

6

9

6

7

60%

90%

$250

9m

MLV

7

10

7

10

70%

100%

$300

12m

Max

Low temperature
thermal desorption

4.4 Processing of input data
The input data processing of in-situ enhanced bioremediation is described below.
A. Cleanup time
The possible values of cleanup time for in-situ enhanced bioremediation are 0.5, 2 and 3.5
years. This array refers to a Triangular Fuzzy Number (TFN) which has most credible value 2
and lowest and highest possible values are 0.5 and 3.5. When the TFN is plotted on Fig. 4.1,
the memberships of cleanup time to the five fuzzy-sets (0.5, 0.89, 0.69, 0.4, 0.0) are obtained
(Fig. 4.1), where these five fuzzy-sets include "short", "short to medium", "medium", "medium
to long" and "long" cleanup time, respectively.
1
0.9
0.8 0.7

f

0.6

|

0.5

Short to Med. Med. Med. to long

Short

\

Long

A A

/

\
<fy0 6 \
A..C../4'

1 0.4
"
I 0.3

£0.69
Y
f
\ / \

/ 7

W

/
/

/
/

/
/

—t—l— —r

1

/

0

2

2.5

0.5

1

\

/
V

\

/

J VV /

7 '••/

0.2
0.1
0

A.

A

1.5

\

V

\ \ /
\.A /

\
\

M/

1

3

¥

3.5

1

4

r

4.5

\
l

i

5.5

Cleanup time (in year) in-situ

Fig. 4.1. Data input for in-situ cleanup time
Fig. 4.1 shows that the TFN intersected the fuzzy-sets of "short", "short to medium",
"medium" and "medium to long" with a certain membership, i.e. short (0.5), short to medium
(0.6 & 0.89), medium (0.69), medium to long (0.4) and long (0.0). The TFN intersected, the
85

fuzzy-set "short to medium" at 0.6 and 0.89 (Fig. 4.1), therefore, the maximum operator is used
to determine the membership to fuzzy-set "short to medium", which is 0.89 (Yager and Filev,
1994). Likewise, the other criteria input data is processed similarly. Fig. 4.2 to 4.8 presents the
processing of input data for enhanced bioremediation.
B. Overall cost
The estimated cleanup cost by in-situ enhanced bioremediation is $50 per m3 to $150 per
m3 . The maximum possible value for this criterion is $125 per m3. And, the lowest possible and
highest possible values are $50 and $150 per m3 respectively. Therefore, the vertex of TFN of
this criterion is leaned towards the "medium cost" fuzzy-set (Fig. 4.2). After interpolation of
the TFN the memberships obtained for this criterion regarding in-situ enhanced bioremediation
alternative are 0.4, 0.8, 0.68, 0.0 and 0.0 to "low", "low to medium", "medium", "medium to
high", and "high" cleanup costfozzy-sets,respectively .
Med.tohigh

Lowtomed.

0

25

50

75

100

125

150

175 200 225 250 275 300

In-situ cost ($/ m )

Fig. 4.2. Data input for in-situ cleanup cost

86

C. Minimum achievable concentration
The value of ability to reduce contaminant concentration of in-situ enhanced
bioremediation criterion varies from 50% to 70%. The most credible likely value was 65% for
this criterion. The vertex of the TFN of the criterion was leaned towards "medium to high"
category (Fig. 4.3).
LowtoMed. Med.

0

10

20

30

40

50

Med. To high

60

70

80

90

100

Ability to reduce contaminant concentration (percentage)
Fig. 4.3. Data input for ability to reduce contaminant concentration

D. Community acceptability
Again, community acceptability of the in-situ enhanced bioremediation alternative is within
80% to 100%. The TFN for this alternative is 80, 90 and 100. Five membership values (1.0,
0.32, 0.0, 0.0, 0.0) are obtained after plotting the TFN in the fuzzy performance scale (Fig.
4.4).

87

Lo. to med. Med. Med. to high

Low

0

10

20

30

40

50

60

70

80

High

90

100

Community acceptance
Fig. 4.4. Data input for community acceptance
Since the most likely value is 90, the vertex of the TFN is leaned toward "high" fuzzy-set.

£. Availability
Availability of in-situ enhanced bioremediation alternative varies from 6 to 10. The TFN
for this criterion is 6, 7.5, andlO. Since, the most credible value is 7.5 the vertex of TFN is
leaned towards "good" fuzzy-set (Fig.4.5).

88

Membership

Low
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

Low to med. Medium

Medium to high

High

\

V / X
/ X / / /x\ \ \ /
Xj
X. /
X I / i \ /, \
\

/

\

J
\

/

\ /

—(-—,—V

\ /

, Y /

i V1
i

\

Y \
10

Level of technology availability
Fig. 4.5. Data input process of technology availability criterion
The memberships obtained for this criterion are 0.67,0.85, 0.28, 0.0, and 0.0. That means
the TFN has a membership of 0.67 to "high", 0.85 to "medium to high", 0.28 to "medium"
and no membership to "low to medium" and "low" fuzzy-sets.

F. Regulatory permitting acceptability
The regulatory permitting acceptability value of in-situ enhanced bioremediation varies
from 1 to 3, and the most credible value is 2.5. Therefore, the vertex of the TFN is leaned
towards the "low to medium" fuzzy-set (Fig. 4.6).

89

Low to med.

Med.

Med. to high

3

10

Level of regulatory acceptance
Fig. 4.6. Data input for regulatory acceptance criterion

G. Development status
The numeric scale value of development status of in-situ enhanced bioremediation
technology varies from 8 to 10. The most credible value of this criterion is 9. Therefore, the
vertex of TFN of the criterion intersected the vertex of the "high" fuzzy-set (Fig. 4.7).
Low to med.

Med

Med. to high

10
Development status
Fig. 4.7. Data input for development status

90

H. Maintenance requirement
The possible values of maintenance requirement criterion of in-situ enhanced
bioremediation alternative are 5, 6.5 and 7. This array refers to a triangular fuzzy number
(TFN) which has most credible value 6.5. The lowest and highest possible values are 5 and 7
respectively. Five membership values (0.0, 0.0, 0.6, 0.8 and 0.0) are obtained for "low", "low
to medium", "medium", "medium to high", and "high" maintenance requirement, respectively
(Fig. 4.8).
Low
0.9 0.8 - £ 0.7g 0.6•§ 0.5£ 0.4 - 2
0.30.2 0.1 0—
0

Low to med.

1

2

3

Med.

4

5

6

Med. to high

7

8

High

9

10

Required maintenance
Fig. 4.8. Fuzzification of input data for required maintenance criterion

In the same way, the input data of other remediation alternatives are handled, and the
membership values to the fuzzy-sets of each criterion are then obtained. Table 4.3 lists the
results after converting different linguistic terms to common linguistic variables of "excellent",
"good", "fair", "poor", and "bad", respectively.

91

Soil vapor extraction
(in-situ)

Bioventing (in-situ)

Remediation
alternative
Enhanced
bioremediation
(in-situ)

92

Cleanup time
Overall cost
Minimum achievable concentration
Community acceptability
Technology availability
Regulatory permitting acceptability
Development status
Technology maintenance requirement
Cleanup time
Overall cost
Minimum achievable concentration
Community acceptability
Technology availability
Regulatory permitting acceptability
Development status
Technology maintenance requirement
Technology availability
Cleanup time
Overall cost
Minimum achievable concentration
Community acceptability
Technology availability
Regulatory permitting acceptability
Development status
Technology maintenance requirement
0.5
0.4
0.0
1.0
0.67
0.0
1.0
0.0
0.32
0.88
1.0
1.0
0.6
0.0
1.0
0.5
0.6
1.0
0.4
0.0
1.0
0.67
1.0
1.0
0.5

0.89
0.8
0.8
0.32
0.85
0.0
0.32
0.0
0.85
0.35
0.35
0.32
1.0
0.5
0.32
0.5
1.0
0.0
0.8
0.69
0.32
0.85
0.32
0.32
0.5

0.69
0.68
0.59
0.0
0.28
0.0
0.0
0.6
0.6
0.0
0.0
0.0
0.32
0.5
0.0
0.0
0.32
0.0
0.68
0.69
0.0
0.28
0.0
0.0
0.0

Table 4.3 Fuzzification of remediation alternative criteria
Criteria
MEI
MGO
Mfa
0.4
0.0
0.0
0.0
0.0
0.82
0.0
0.80
0.25
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

MPO

0.0
0.0
0.0
0.0
0.0
0.58
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0

MBa

Low temperature
thermal
desorption
(ex-situ)

Slurry phase
(ex-situ)

Remediation
alternative
Landfarming
(ex-situ)

93

Cleanup time
Overall cost
Minimum achievable concentration
Community acceptability
Technology availability
Regulatory permitting acceptability
Development status
Technology maintenance requirement
Cleanup time
Overall cost
Minimum achievable concentration
Community acceptability
Technology availability
Regulatory permitting acceptability
Development status
Technology maintenance requirement
Cleanup time
Overall cost
Minimum achievable concentration
Community acceptability
Technology availability
Regulatory permitting acceptability
Development status
Technology maintenance requirement

0.0
1.0
0.0
0.0
0.67
0.0
1.0
0.5
0.0
0.0
0.0
0.0
0.6
1.0
1.0
0.0
0.0
0.0
1.0
0.0
0.6
0.0
1.0
0.0

0.4
0.5
0.69
0.5
0.85
0.5
0.32
0.5
0.4
0.0
0.69
0.5
1.0
0.32
0.32
0.0
0.4
0.0
0.35
0.5
1.0
0.5
0.32
0.0

0.8
0.0
0.69
1.0
0.28
0.5
0.0
0.0
0.8
0.32
0.69
1.0
0.32
0.0
0.0
0.5
0.8
0.32
0.0
1.0
0.32
0.5
0.0
0.5

Table 4.3 Fuzzification of remediation alternative criteria (continued)
Criteria
ME,
MFa
Mao
0.8
0.0
0.0
0.5
0.0
0.0
0.0
0.0
0.8
0.81
0.0
0.5
0.0
0.0
0.0
0.5
0.8
0.81
0.0
0.5
0.0
0.0
0.0
0.5

MPO

0.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.4
0.6
0.0
0.0
0.0
0.0
0.0
0.0
0.4
0.6
0.0
0.0
0.0
0.0
0.0
0.0

MBU

From Table 4.3 it can be observed that each technology has different membership Junctions
in fuzzy-sets. For example, membership functions of overall cleanup cost of in-situ enhanced
bioremediation are 0.4, 0.8, 0.68, 0.0, and 0.0 for the fuzzy-sets of "excellent", "good", "fair",
"poor", and "bad", respectively. Again, membership functions of overall cleanup cost of in-situ
bioventing are 0.88, 0.35, 0.0, 0.0 and 0.0 for the fuzzy-sets of "excellent", "good", "fair",
"poor", and "bad", respectively.

4.4.1 Aggregation of membership values and criteria weights
In the next step, the membership values of all criteria are multiplied by the criteria weights
for aggregation. Since, there are 8 criteria weights and 5 fuzzy-sets for each criterion the
multiplication provides the membership values to the five fuzzy-sets for each alternative. For
example, equation (3.2) is applied to aggregate the criteria importance weights and in-situ
remediation alternative performance values. The calculation process is shown below:

0.5, 0.89, 0.69, 0.4, 0.0
0.4, 0.8, 0.68, 0.0, 0.0
0.0, 0.8, 0.59, 0.0, 0.0
1.0, 0.32, 0.0, 0.0, 0.0
[ 0.118,0.130,0.129,0.129,0.115,0.144,0.112,0.124 ]«
0.67,0.85, 0.28, 0.0, 0.0
1.0, 0.32, 0.0, 0.0, 0.0
0.6, 0.94, 0.68,0.5, 0.0
0.0, 0.0, 0.0, 0.82, 0.58

[0.429 0.487 0.353 0.264 0.084]

94

Similar procedure is followed to aggregate the criteria importance weight and remediation
alternative performance values of all remediation alternatives. Table 4.4 shows the aggregated
results of each alternative. In this table, it can be seen that, in-situ enhanced bioremediation has
an aggregated membership value of 0.429 in "excellent", 0.487 in "good", 0.353 in "fair",
0.264 in "poor" and 0.084 in "bad" fuzzy-sets, respectively.

Table 4.4 Membership values of each alternative after aggregation
Remediation
Cardinality
MPO
MFa
ME,
MGO
MBa
alternatives
Enhanced
0.429
0.264
0.084
1.617
0.487
0.353
bioremediation
0.265
0.052
0.301
0.218
0.164
(in-situ)
Bioventing
0.653
0.517
0.180
0.030
0.000
1.379
(in-situ)
0.474
0.375
0.130
0.021
0.000
SVE
0.694
0.476
0.210
0.000
0.000
1.380
(in-situ)
0.503
0.345
0.152
0.000
0.000
Landfarming
0.381
0.533
0.417
0.159
0.047
1.537
(Ex-situ)
0.248
0.347
0.271
0.103
0.031
Slurry phase
0.325
0.398
0.453
0.326
0.125
1.627
(Ex-situ)
0.200
0.244
0.278
0.201
0.077
Low temperature
0.310
0.380
0.436
0.326
0.125
1.577
thermal desorption
0.197
0.079
0.241
0.276
0.207
(ex-situ)
* Normalized values are in bold

4.5 Defuzzification and ranking of alternatives
In this step, the membership values are normalized. The normalization process is done by
dividing each membership value with cardinality of an alternative. According to Sadiq et al.
(2004) the term "cardinality" is referred as the sum of all the membership values of all of its
fuzzy sets (i-e;/Um,jUGo,jUFa,jUPo,jUBa). For example, the cardinality for five fuzzy-sets of insitu enhanced bioremediation is 1.617 and the normalized membership functions are 0.265,
0.301, 0.218, 0.164, and 0.052 in "excellent", "good", "fair", "poor" and "bad" fuzzy-sets,
95

respectively. Similar procedure is followed to obtain normalized membership functions of
other alternatives. Table 4.4 shows the normalized values of each alternative in bold.
Afterwards, the alternatives are ranked by using the maximum normalized value (within five
fuzzy-sets) of an alternative.
Final membership functions of alternatives are shown in Fig. 4.9 in form of a possibility
mass function. The height of the bars in the figure represents normalized memberships to 5
fuzzy-sets. These are "excellent", "good", "fair", "poor" and "bad". These memberships
represent the overall fuzzy multi-criteria score for each alternative after aggregating all the
criteria. From Fig. 4.9 it can be seen that in-situ soil vapor extraction (SVE) and in-situ
bioventing technology has higher membership in "excellent" fuzzy-set. In-situ enhanced
bioremediation also has a higher membership in "excellent", "good" and "fair" fuzzy-sets.

1.000
0.900
0.800
0.700
0.600
Meniership 0.500
0.400
0.300
0.200
0.100
0.000

Fig. 4.9. Final membership functions of remediation alternatives

96

4.6 Ranking of alternatives
Each of the fuzzy-sets was divided by the cardinality for all fuzzy sets. Then the maximum
utility value within 5-tuple fuzzy-sets was selected for ranking of the alternatives. Fig. 4.10
shows the ranking order of the remedial alternatives.
1.000
0.800
, 0.600
0.400
0.200
0.000

0.474

0303

• I 111
• A
0.276

Enhanced
Bioventing (inbiorem. (in-situ)
situ)

SVE(in-situ)

Landferm (exsitu)

Slurry phase
(ex-situ)

LTTD (ex-situ)

Remediation alternative

Fig. 4.10. Ranking of remediation alternatives by fuzzy multi-criteria method
It was observed that in-situ soil vapor extraction alternative scored 0.503 and became the
most preferred alternative according to the developed fuzzy multi-criteria approach. Again, insitu bioventing became the second most preferred option with a score of 0.474. In-situ
enhanced bioremediation and ex-situ landfarming alternatives were third and fourth preferable
options, respectively. From the remedial alternative input data in Table 4.2, it was observed
that criteria values of "technology availability", "community acceptability", "overall cost" and
"regulatory acceptability" of in-situ SVE alternative contributed mostly for its top preference.
For example, the cost range of in-situ SVE is $50-$150/m3 and the cost range of low
temperature thermal desorption is $200-$300/m3. Similar observations were made for other
criteria values of other alternatives.
97

4.7 Comparison of results of MCDA techniques for the same case study site
Different multi-criteria decision analysis (MCDA) techniques (i.e., SAW, TOPSIS, and
WPM) techniques were applied for the same case study site for analysis of results. A
description of these MCDA tools is given in Chapter 2. Since existing MCDA methods can
deal with only crisp rating values, the collected criteria information needed to be converted into
single crisp values before using in these MCDA methods. The conversion process of selecting
crisp values was done in reference with existing rating system and USEPA (1993), remediation
alternative screening matrix. According to USEPA the criteria of "overall cost", "minimum
achievable concentration", "cleanup time", "required maintenance" and "community
acceptability" are rated by means of linguistic terms of "better", "average" and "worse". These
linguistic ratings are then divided into representative numerical values such asl= better, 2=
average, and 3= worse (ICS-UNIDO, 2000; UNECE, 1997). Assigning numerical values to the
linguistic terms is dependent on the decision maker's preference. Table 4.5 lists the
remediation alternative evaluation criteria information in linguistic and crisp form. In this table
the values in bold represent the single crisp values for each criterion. It should be mentioned
here that, the applied numerical rating scale has a value range of 1 to 10 and 10 to 100 for
certain criteria (e.g., technology availability and minimum achievable concentration).

98

l-3yrs
24m
Low
$50/m3
Better
90%
Better
90%
More than 4
vendors
8
Average
6
Full
9
Low
2

l-3yrs
24m
Average
$100m3
Average
60%
Better
90%
More than 4
vendors
8
Worse
2
Full
9
Average
6

Full
9
Low
2

Better
90%
More than 4
vendors
8
Better
9

0.5-lyr
9m
Average
$100/m3
Average
60%

Full
9
Low
2

Average
60%
More than 4
vendors
8
Average
6

0.5-lyr
9m
Low
$100/m3
Average
60%

99

*values in bold are single crisp values
a. En.bio= Enhanced bioremediation, b. SVE= Soil vapor extraction, c. LTTD= Low temperature thermal desorption.

Technology maintenance
requirement
(1-10)

Regulatory permitting
acceptability
(1-10)
Development status (1-10)

Cleanup time
(months)
Overall cost
($/m3)
Minimum achievable
concentration
(10-100%)
Community acceptability
(10-100%)
Availability
(1-10)

Criteria

Table 4.5 Criteria input value for MCDA methods
En.bio8 rem. Bioventing
Landfarming
SVE"
(in-situ)
(in-situ)
(in-situ)
(ex-situ)

Full
9
Average
6

Average
60%
More than 4
vendors
8
Better
9

0.5-lyr
9m
High
$250/m3
Average
60%

Full
9
Average
6

Average
60%
More than 4
vendors
8
Average
6

0.5-lyr
9m
High
$250/m3
Better
90%

Slurry phase LTTDC
(ex-situ)
(ex-situ)

4.7.1 Simple additive weighting (SAW) method

Simple additive weighting method was applied for the case study site. The calculations are
shown in Table 4.6. Performance weights of each alternative were calculated by using equation
(2.1). Then these performance weights were converted into utility values.

100

0.164
0.103
0.164
0.082
0.138
0.185
0.021

0.115

0.129

0.129

0.112

0.124

0.118

0.130

0.144

Technology
availability
(1-10)
Community
accep.
(10-100%)
Min. achxoncen
(10-100%)
Dev.status
(1-10)
Maintenance
req. (1-10)
Time to cleanup
(months)
Overall cost
($/yd3)
Reg.acceptabilit
y(i-io)
Total
Normalized
utility

A
Perf.
0.119

0.136

0.115

0.136

61.938
0.131

1.621

0.085

12.277 0.139

8.800

5.431

10.279 0.136

7.113

11.714 0.136

U
9.052

Perf.
0.145

0.070

82.468
0.203

8.100

0.104

11.190 0.140

8.800

10.860 0.166

10.280 0.166

11.380 0.104

11.710 0.104

U
9.050

0.083

0.132

Perf.
0.116

69.408
0.188

8.103

9.208

4.400

0.132

0.123

0.149

10.863 0.132

10.279 0.132

7.113

7.321

U
9.052

Perf.
0.149

0.106

83.268
0.190

12.965 0.170

10.095 0.1432

101

0.158

0.0986

Perf.
0.138

5.431

0.204

59.631
0.140

12.965 0.0986

9.208

5.431

10.279

11.380

7.321

U
9.052

59.037
0.148

8.103

14.119

0.0665 4.400

0.0789

10.279 0.158

7.113

7.321

U
9.052

11.733 0.0716 4.400

10.863 0.0849

10.279 0.170

7.113

11.714 0.106

U
9.052

E

CW= criteria weight, Perf.= performance, U= utility, A= in-situ enhanced bioremediation, B= in-situ bioventing, C= in-situ soil vapor extraction,
D= ex-situ landfarming, E= ex-situ slurry phase, F= ex-situ low temperature thermal desorption

Perf.
0.144

CW

Criteria

Table 4.6 Calculations using SAW method
B
C
D

4.7.2 Technique for order performance by similarity to ideal solution (TOPSIS)
Another MCDA technique TOPSIS was applied for ranking of these remediation
alternatives. The procedure of TOPSIS method is discussed in Chapter 2. Since, each
evaluation criterion has a different measurement value; a criterion normalization process was
applied (equation 2.5). The normalized criteria rating values are shown in Table 4.7 and Table
4.8. Next, these normalized criteria rating values were multiplied with the criteria importance
weights (equation 2.6) to calculate the weighted normalization values (Table 4.8).Then, the
positive ideal and negative ideal solutions were calculated by using equation (2.7) and equation
(2.8). Table 4.9 shows the calculated values of positive ideal and negative ideal solutions.
Subsequently, the values of separation measures from positive ideal and negative ideal
alternatives were calculated using equation (2.9) and equation (2.10). Finally, the values of
similarities to ideal solutions (Q) were calculated by using equation (2.11) and the calculated
values are shown in Table 4.10.

102

c,

0.4083
0.4083
0.4083
0.4083

Soil vapor extraction (in-situ)
Landfarming (ex-situ)
Slurry phase (ex-situ)
LTTD (ex-situ)

C
D
E
F

0.4804
0.3203
0.3203
0.3203

c2

0.4804
0.4804
0.3430
0.3430
0.3430
0.5145

c3

0.3430
0.5145
0.4083
0.4083
0.4083
0.4083

c4

0.4083
0.4083
0.1826
0.1826
0.5472
0.5477

c5

0.5472
0.1826

c6

0.3059
0.2294
0.2294
0.2294

0.6118
0.6118
0.2520
0.2520
0.6299
0.6299

c7

0.2520
0.1260

0.5437
0.3625
0.5437
0.3625

c8

0.1208
0.3625

0.0634
0.0423

0.0474
0.0474

c2

Soil vapor extraction (in-situ)
Landfarming (ex-situ)
Slurry phase (ex-situ)
LTTD (ex-situ)

C
D
E
F

Ci

Remediation alternative
Enhanced biorem. (In-situ)
Bioventing (in-situ)

A
B
0.0634
0.0423
0.0423
0.0423
0.0439
0.0439
0.0439
0.0659

0.0474
0.0474
0.0474
0.0474

0.0223
0.0223
0.0668
0.0668

C4

C5

103

0.0474
0.0223
Negative ideal solution
0.0439
0.0474
0.0668

0.0659

C3

0.0716

c6

0.0268

Table 4.9 Positive ideal and negative ideal solutions
Positive Ideal solution

0.0474
0.0474
0.0474
0.0474

c7

0.0775

0.0310

0.0358
0.0268
0.0268
0.0268

Table 4.8 Weighted normalized values of criteria
V( weighted normalized values)
C2
c4
c3
c6
0.0474
0.0439
0.0668
0.0716
0.0474
0.0634
0.0659
0.0474
0.0223
0.0716
0.0474
0.0634

0.0176

c8

0.0794

0.0310
0.0310
0.0775
0.0775

0.0310
0.0155

0.0794
0.0529
0.0794
0.0529

c8

0.0176
0.0529

CI: Technology availability; C2: Community acceptability; C3: Minimum achievable concentration; C4: Development status; C5: Maintenance requirement;
C6: Time to clean up; C7: Overall cost; C8: Regulatory acceptability.

0.4083
0.4083

Remediation alternative
Enhanced biorem. (In-situ)
Bioventing (in-situ)

A
B

Table 4.7 Normalized criterion ratings
R ( normalized criteria values ) =

Table 4.10 Separation measure from positive and negative ideal solutions and calculated value
function of each alternative
Alternatives
S+
SNormalized
A
B
C
D
E
F

Enhanced bioremediation (in-situ)
Bio venting (in-situ)
Soil vapor extraction (in-situ)
Landfarming (ex-situ)
Slurry phase (ex-situ)
Low temperature thermal desorption
(ex-situ)

c,

0.0910
0.0542
0.0237
0.0404
0.0712

0.0511
0.0894
0.0984
0.0860
0.0763

0.104
0.181
0.234
0.198
0.150

0.0728

0.0611

0.133

S+: Separation measurefrompositive ideal solution; S-: Separation measurefromnegative ideal solution;
Q: Value function (similarities to positive ideal solutions).

4.7.3 Weighted product method (WPM)

In WPM method, the normalized performance index of each alternative was calculated
initially by using available criteria information. Then, utility value of an alternative was
calculated by using equation (2.12). Subsequently, the value ratio (Rj) for each alternative was
calculated by applying equation (2.13). The calculated value ratio is then used for ranking of
the alternatives. Table 4.11 shows the calculated normalized R; values.

104

1.289

0.112
0.124
0.118
0.130
0.144

Maintenance requirement (1-10)

Time to cleanup (month)

Overall cost ($/yd3)

Regulatory acceptability (1-10)
1.741
0.127

1.106

0.568

0.689

2.679
0.196

1.298

0.619

0.689

0.919

1.289

1.774

1.809

2.689
0.181

1.37756

0.568

0.747

0.919

1.289

1.684

1.809

2.485
0.196

1.298

0.568

0.773

0.919

1.289

1.684

1.715

2.060
0.150

1.378

0.508

0.773

0.803

1.289

1.684

1.715

1.274

E

2.045
0.149

1.298

0.508

0.773

0.803

1.289

1.774

1.715

1.274

F

3.083

1.378

0.619

0.747

0.919

1.289

1.774

1.809

1.274

V(A)

105

A= Enhanced bioremediation (in-situ); B= Bioventing (in-situ); C= Landfarming (in-situ); D= Soil vapor extraction (SVE; ex-situ); E= Slurry phase;
F= Low temperature thermal desorption; V(A*)= value of an ideal alternative; V(A;)= value of an alternative; Rj= the normalized value ratio between an alternative
and the ideal alternative.

V(Aj)
Normalized Ri

1.684

0.129

0.803

1.809

0.129

Community acceptability
(10-100 %)
Minimum achieveable
concentration (10-100%)
Development status (1-10)

Technology availability (1-10)

Criteria

Table 4.11 Calculations of weighted product method
Criteria
A
B
C
D
weight
0.115
1.274
1.274
1.274
1.274

4.8 Results from MCDA methods
It was observed that in-situ soil vapor extraction (SVE) alternative was the first preference
both by TOPSIS and WPM methods. In these methods the utility values of SVE is 0.234 and
0.196 (Table 4.10 and Table 4.11). As well, in-situ bioventing alternative became first
preference by SAW method. In this method, the utility value of in-situ bioventing is 0.203
(Table 4.6). As, SVE scored highest ranking order by two methods, this alternative remained
the top preference to remediate the site in hand. Subsequently, enhanced bioremediation
became the least preferable by all the three MCDA methods. The results from existing MCDA
tools of remediation alternatives are shown in Fig. 4.11. In this figure, the X-axis represents the
name of the alternatives, and the Y-axis represents the utility values for each alternative. From
the remediation alternative input data (Table 4.5) it can be observed that criteria value of
cleanup time, and regulatory acceptability of SVE alternative mostly influenced the ranking
order comparing with the criteria value of in-situ enhanced bioremediation.

1.000
0.800
£. 0.600
° 0.400 -••
0.200
1 |§jB
0.000

1 Illii^B

Enhanced
biorcm

• ^ 9 •
Bioventing

SVE
Landferming Slurry phase
Remediation alternatives

•
LTTD

• SAW! TOPSIS • WPM
Fig. 4.11. Utility values of alternatives obtained from MCDA methods

106

4.9 Comparison of results
The average ranking order of each alternative was calculated by dividing each utility value
with the total of three utility values obtained from three MCDA methods. Then, this ranking
order was compared with the ranking order of fuzzy multi-criteria approach. Fig. 4.12 shows
the comparison of results between fuzzy multi-criteria and MCDA methods. In this figure, the
Y-axis represents the ranking order of remedial alternatives and the X-axis represents the list of
remedial alternatives.

Enhanced
biorem

Bioventing

SVE

Landfarm

Remediation alternatives

Slurry phase
HMCDA

LTTD

lFu2zy multi-criteria

Fig. 4.12. Ranking order of remedial alternatives by MCDA and fuzzy multi-criteria and
methods.
From Table 4.2 and Table 4.5 it was observed that the maximum likely values (MLVs) of
the criteria in fuzzy multi-criteria method and single values of the criteria in MCDA methods
were mostly similar. Such similarities in input values were also reflected between the ranking
orders of both methods. In-situ soil vapor extraction (SVE) and in-situ bioventing alternatives
became first and second preference, respectively by both MCDA and fuzzy multi-criteria
methods. Again, the ranking order for ex-situ land farming became fifth in MCDA method and
sixth in fuzzy multi-criteria method.

107

In fuzzy multi-criteria method the ranking order of in-situ enhanced bioremediation was
fourth. However, in MCDA method it became sixth. This difference is observed due to the
variation of criteria values including "minimum achievable concentration", "technology
availability" and "regulatory permitting acceptability". For example, in fuzzy multi-criteria
method the criterion value of "minimum achievable concentration" was 50% to 70% and
maximum likely value was 65%. But in MCDA method the value was 60%.

4.10 Sensitivity analysis
Several sensitivity analyses were conducted to find the effect of parameter value changes
on the remedial alternative ranking order.
4.10.1 Single input value
The remedial alternative performance values were applied as a single crisp value in the
developed fuzzy multi-criteria approach. Table 4.12 lists the input values used in the fuzzy
multi-criteria approach. These input values are similar with respect to the values used in
MCDA methods.

108

90
60
9
2

90
90
9
2
24m
50
6

90
60
9
6
24m
100
2

9

100

9m

8

8

8

6

9

6

250

250

100

109

9m

6

9

90

60

8

9m

6

9

60

60

8

LTTDC
(ex-situ)

9m

2

9

60

60

8

*values in bold are single crisp values
a. En.bio= Enhanced bioremediation, b. SVE= Soil vapor extraction, c. LTTD= Low temperature thermal desorption.

Technology availability
(1-10)
Community acceptability
(10-100 %)
Minimum achieveable
concentration (10-100 %)
Development status
(1-10)
Maintenance requirement
d-10)
Time to cleanup
(month or year)
Overall cost
($/m3)
Regulatory acceptability
(1-10)

Criteria

Table 4.12 Single input value used in the fuzzy multi-criteria approach
Bioventing
En.bioa
SVEb
Landfarming Slurry phase
(ex-situ)
(in-situ)
(in-situ)
(in-situ)
(ex-situ)

A comparison was conducted between the results of uncertainty consideration and the
results of no uncertainty consideration (single input value) in the developed fuzzy multi-criteria
method. It was found that there was no influence of uncertainty consideration in the ranking
order of in-situ bioventing, in-situ soil vapor extraction, ex-situ landfarming and ex-situ slurry
phase remediation alternatives. In both cases (i.e., uncertainties were considered, uncertainties
were not considered) the ranking order of these alternatives remained the same. On the other
hand, the ranking order of in-situ enhanced bioremediation became fourth when uncertainties
were considered and it became sixth when uncertainties were not considered. However, the
ranking order of ex-situ low temperature thermal desorption (LTTD) became sixth when
uncertainties were considered and this alternative became fourth when uncertainties were not
considered. The differences between the ranking order of in-situ enhanced bioremediation and
ex-situ LTTD without uncertainty consideration were due to the differences in remediation
alternative performance values of "minimum achievable concentration", "cleanup time", and
"regulatory acceptability". For example, "Minimum achievable concentration" by in-situ
enhanced bioremediation was 60% whereas, "minimum achievable concentration" by ex-situ
LTTD was 90%; "cleanup time" by in-situ enhance bioremediation was 24m and by ex-situ
LTTD was 9m; "regulatory acceptability" of in-situ enhanced bioremediation was 2, and
"regulator acceptability" of ex-situ LTTD was 6 (Table 4.12).

The differences between ranking order of in-situ enhanced bioremediation and ex-situ
LTTD with uncertainty consideration were due to the differences in remediation alternative
performance values of "overall cost" and "community acceptability". For example, the
uncertainties involved in the "overall cost" criterion of in-situ enhanced bioremediation varied

110

from $50m3 to $150/m3 and the uncertainties of "overall cost" criterion of ex-situ LTTD varied
from $200/m3 to $300/m3 and the uncertainties involved in the "community acceptability"
criterion of in-situ enhanced bioremediation varied from 80% to 100% and the uncertainties in
"community acceptability" criterion of ex-situ LTTD varied from 50% to 70% (Table 4.2). The
overall comparison results between uncertainty consideration and without uncertainty
consideration are shown in Fig. 4.13.

Enhanced
Bioven.(in-sku)
biorera(in-situ)

SVE (in-situ)

Landfarm
(ex-situ)

Slurry phase
(ex-situ)

LTTD (ex-situ)

Remediation alternatives
H Uncertainty considered B Uncetainty not considered (single input value)

Fig. 4.13. Comparison of results when uncertainty is considered in the evaluation

4.10.2 Change in criteria importance weight
The criteria importance weights were changed to determine the sensitivity of remedial
alternative ranking order. In trial 1 the importance weight of "overall cost" criterion was set to
0.26. As the total of all criteria importance weight needed to be 1 (normalized), the weights of
other criteria (i.e., "technology availability", "community acceptability", "minimum achievable
concentration", "development status", "maintenance requirement", "cleanup time" and
"regulatory acceptance") were set to 0.106. From the results of trial 1 it was observed that in-

Ill

situ bioventing became the first preferred alternative and in-situ soil vapor extraction (SVE)
became the second preferred alternative. The ranking order of other alternatives remained
similar to the original ranking order of alternatives for the case study site. In trial 2 the
importance weight of "overall cost" was set to 0.52 and the importance weights of other criteria
were set to 0.069. Likewise in trial 1, in-situ bioventing alternative remained the first preferred
alternative in this trial. Though, the rank order of ex-situ landfarming was third in the original
rank order of case study site and it became the second preferred alternative in trial 2. This
difference was due to the increment of importance weight of "overall cost" criterion from 0.130
to 0.52. In trial 3 the importance weight of "overall cost" criterion was set to 0.85 and the
importance weight of al other criteria were set to 0.022. Though ex-situ landfarming remained
second preferred alternative in trial 3 in-situ SVE became the fifth preferred alternative. The
rank order of in-situ SVE was third in trial 2. However, such difference in the rank order was
due to the difference between the performance values of "overall cost" criterion of these two
alternatives. From Table 4.2 it was observed that the maximum likely value (MLV) of in-situ
and ex-situ landfarming were $125/m3 and $100/m3, respectively. Fig. 4.14 shows the utility
values and the ranking order of remediation alternatives obtained from trial 1, trial 2 and trial 3.

112

1.000
0.900
0.800
0.700
£> 0.600 € 0.500
S> 0.400
0.300
0.200
0.100
0.000

Remediation alternatives
D Trial 1 •Trial 2 • Trial 3 • Original rank order for case study site
Fig. 4.14. Sensitivity analysis of remediation alternatives by changing "overall cost" criterion
importance weight

In trial 4, criteria importance weights of both "overall cost" and "cleanup time" criteria
were set to 0.45 and criteria importance weight of all other criteria (i.e., "technology
availability", "community acceptability", "minimum achievable concentration", "development
status", "maintenance requirement", and "regulatory acceptance") were set to 0.017. In this
trial, in-situ SVE became the first preferred alternative whereas, in-situ SVE was fifth
preferred alternative in trial 3. It was observed that the lower importance weight of "overall
cost" criterion played an important role for higher ranking order of in-situ SVE in trial 4
because the importance weight of this criterion was 0.85 in trial 3. In addition, trial 5 was
conducted by setting the criteria importance weights of "overall cost", "cleanup time" and
"community acceptability" as 0.250. The importance weights of all other criteria were set to
0.05. In-situ SVE remained the first preferred alternative in this trial and overall there were no

113

significant differences in the ranking order of trial 5 compared to the ranking order of trial 4.
The results of trial 4 and trial 5 of sensitivity analysis are shown in Fig. 4.15.
1.000 -I
0.900 0.800 0.700 0.600 g3-S 0.500 » 0.400 0.300 0.200 0.100 0.000 Enhanced
bbremediation

Bbventing

SVE

Landiaiming

Slurryphase

LTTD

Remediation alternatives
D Trial4

I Trial 5

• Original rank orderforcase study site

Fig. 4.15. Sensitivity analysis of remediation alternatives by changing the importance weights
"overall cost", "cleanup time" and "community acceptability" criteria

In summary it can be stated that the criteria importance weights can influence the overall
ranking order of the remediation alternatives. If all the criteria importance weights are close to
each other then all of these criteria will influence the ranking order of the alternatives.
However, if a criterion has significantly higher importance weight compared to other criteria
then this criterion will influence the evaluation results of the alternatives. It was observed that
the criterion "overall cost" played an important role in the ranking order of alternatives when
the importance weight of this criterion was very high.

114

4.10.3 Change in remediation alternative performance values
Since the remediation alternative in-situ soil vapor extraction (SVE) became the best
preference among other alternatives by fuzzy multi-criteria method. A sensitivity test was
conducted by changing the performance values of criteria, "community acceptability",
"cleanup time" and "overall cost" of in-situ SVE to observe whether performance values affect
the ranking order of an alternative. The existing performance value range of criterion
"community acceptability" was 80% to 100%; "cleanup time" was 0.5yr to lyr and "overall
cost" was 50$/m3 to 150$/m3. In the sensitivity analysis these performance value ranges were
set to 30% to 50%, 3yr to 5yr, and 250$/m3 to 300$/m3, respectively. From the sensitivity
analysis, it was observed that the rank order of in-situ soil SVE alternative became fourth from
first ranking place in the original ranking order of remediation alternatives for the case study
site (Fig. 4.16). However, the in-situ bioventing alternative became the first preference in this
ranking order. The ranking order of in-situ bioremediation also improved from previous
ranking order of fourth to third. Fig. 4.16 shows the ranking order and utility values of all the
alternatives obtained from the sensitivity analysis.
1.000

Enhanced
bbremediation

Bioventing

SVE

Landferming

Slurryphase

LTTD

Remediation alternatives
I Rank order of SVE after performance value change

I Original rank order for case study site

Fig. 4.16. Sensitivity analysis for SVE alternative by changing performance values of
"community acceptability, "cleanup time" and "cleanup cost" criteria

115

In summary it can be stated that selection of remediation alternative through an evaluation
process mostly depends on independent performance value (i.e., "cleanup cost", "cleanup
time", "regulatory acceptance" etc.) of each alternative. From the above analysis it was
observed that the in-situ SVE became less preferred option when the performance value of
"community acceptability" criterion was significantly reduced and the performance value
"cleanup time" and "overall cost" criteria were significantly increased.

116

Chapter 5 Conclusions and Recommendations

5.1 Summary
Management of contaminated sites is a complex task. Such decision making process is not
limited only in selection of the best remediation alternative, but also several factors need to be
considered. The factors which influence the decision making process include, involvement of
various stakeholders, consideration of uncertainties, and evaluation and selection of the right
remediation alternative.

Selection of remediation alternative is considered as a multi-criteria problem because a
number of criteria including cleanup time, overall cost, community acceptability and regulatory
acceptability need to be considered for the evaluation of remediation alternatives. There are
many multi-criteria decision analysis (MCDA) tools (e.g., AHP, TOPSIS, PROMETHEE,
ELECTRE, WPM, SAW) available to aid the decision making process of a contaminated site.
Most of these MCDA tools require crisp values as inputs. However, the performance data of
remediation alternatives are not always available in crisp form. For example, performance data
on community acceptability and regulatory acceptability are expressed in linguistic terms (e.g.,
high, medium, low). These linguistic terms need to be converted into numerical scale value or
into crisp values to use in the existing MCDA tools.

Public participation and stakeholder involvement is another important component in the
decision making process of contaminated site management. Due to public concern on
environment, public interest in contaminated sites and regulatory requirement, it is essential to

117

incorporate public opinions in such decision making process. However, the existing decision
making approaches have limitations to involve public and different stakeholders' opinions
adequately.

Moreover, presence of a number of uncertainities in the decision making problems limit the
application of existing MCDA tools. The source of uncertainties includes lack of information
about a contaminated site, different opinions of different stakeholders, and a range of possible
values of remediation alternative evaluation criteria.

Considering these problems, there was a need for development of a holistic approach for
contaminated site management. There are many applications of fuzzy set theory available in
literature, especially in the environmental decision making problems. It was found that the
concept of fuzzy-set theory is appropriate to solve these problems. Fuzzy-set theory provides
an intuitive and effective way for the decision makers for dealing with such linguistic
preferences and uncertainties.

The fuzzy multi-criteria technique involves identification of remediation alternative
selection criteria, estimating criteria weight, fuzzification, aggregation, defuzzification, and
ranking remedial alternatives. In this research, a fuzzy multi-criteria decision analysis approach
was developed.

First, the remediation alternative evaluation criteria were identified. Then, different
stakeholders were contacted to express their opinion on the importance of the selected criteria.

118

Second, the criteria measurement values (e.g., overall cost/m3, cleanup time/year) were divided
into five fiizzy-sets by the stakeholders. Then, performance of each alternative was evaluated
by means of these fuzzy-sets. Subsequently, the criteria importance weight and the membership
functions of the fuzzy-sets were aggregated by applying fuzzy matrix multiplication. This
multiplication produced the final fuzzy-sets for each alternative. Then utility values were
calculated for each alternative. Finally, the alternatives were ranked according to their utility.

One of the objectives of this research was to identify the remediation alternative evaluation
criteria and determine the importance weights of the criteria according to stakeholder
preference. This objective was achieved by selecting the most important evaluation criteria
from literature survey and by involving stakeholder in determining the criteria importance
weights. Another objective of this research was to integrate the uncertainty issues in the
remediation alternative evaluation process. This objective was also achieved by using fuzzy-set
theory. Finally, the last objective regarding development of an effective fuzzy multi-criteria
approach for evaluating and ranking remediation alternatives was met in this research. The
developed fuzzy multi-criteria approach was applied to a contaminated site to solve
remediation alternative selection problem and reasonable results were achieved. In the
developed method, it was possible to use a range of possible values to address the uncertainties
in the alternative evaluation performance data. The involved stakeholders were also satisfied
with the easy process of their participation in the selection of most appropriate remediation
alternative.

119

In addition, existing MCDA methods (i.e., SAW, TOPSIS, and WPM) were applied to
solve the same problem. To apply the existing MCDA methods, remediation alternative
evaluation data was required to convert into single crisp value. Much information was not
counted in the overall evaluation process due to selection of single value. Though, the results
by the both approach were mostly consistent, the fuzzy multi-criteria method proved to be
more efficient in dealing with uncertainties. The results from the developed method are more
acceptable than the results of existing MCDA methods because various stakeholder opinions
were incorporated in the system.

5.2 Future extensions

Selection of remediation alternatives is one of the many challenges in contaminated site
management. The developed fuzzy multi-criteria approach could be applied in the earlier stage
of a decision making process of contaminated sites. For example, site characterization and risk
assessment are two important steps of successful contaminated site management. However, in
the developed method it is assumed that both of these two steps have already been conducted
and the user faces the problem of selecting a remediation alternative. Site characterization tools
and risk assessment tools should be incorporated within the developed method for more
efficiency in the contaminated site management.

Once site characterization and risk

assessment are incorporated the method can provide a decision on the most appropriate
remediation option for a contaminated site.

120

The list of the remediation alternative is fixed in the system. Thereby, a user can apply this
system when the potential remediation alternatives match with the list. However, situation may
arise that integrated remediation alternatives (e.g., bioremediation and thermal treatment) need
to be considered for evaluation and ranking. The future extension of the system should
incorporate the integrated remediation alternatives in the list.

In the developed method, it was assumed that different stakeholders have agreed on criteria
importance weight. The average criteria importance weights were used for remediation
alternative evaluation. This may not satisfy all the stakeholders involved in the decision
making process. An approach should be developed to achieve consensus on a balanced criteria
importance weight among the stakeholders. This can be achieved through group meetings of
stakeholders and by identifying their key interests. Related researches can be found in Bose et
al. (1997) and Zapatero et al. (1997).

Again, in the developed method only eight evaluation criteria were used for remediation
alternative evaluation. However, a number of other criteria (e.g. environmental impact,
produced waste) need to be identified and incorporated within the method. The selection of
criteria should depend on contaminated site characteristics and stakeholder's objectives.

121

References
Akter, T., and Simonovic, S. P. 2005. Aggregation of fiizzy views of a large number of
stakeholders for multi-objective flood management decision-making, Journal of Environmental
Management, 11, 133-143.

Al-Ahmari, A. M. A. 2007. Evaluation of CIM technologies in Saudi industries using AHP.
Int.JAdv.Manuf.Technol. 34, 736-747.

Al-Rashdan, D., Al-Kloub, B., Dean, A., Al-Shemmeri, T. 1999. Environmental impact
assessment and ranking the environmental projects in Jordan, European Journal of
Operational Research, 118(1), 30-45.

Al-Shemmeri, T., Al-Kloub, B., Pearman, A. 1997. Computer aided decision support system
for water strategic planning in Jordan, European Journal of Operational Research, 102(3),
455-472.

Altrock, C. V., and Krause, B. 1994. Multi-criteria decision making in German automotive
industry using fuzzy logic, Fuzzy Sets and Systems, 63, 375-380.

Amini, M., Afyuni, M., Fathianpour, N., Khademi, H., Fluhler, H. 2005. Continuous soil
pollution mapping using fuzzy logic and spatial interpolation, Geoderma, 124, 223-233.

122

Amro, M. M. 2004. Factors affecting chemical remediation of oil contaminated water-wetted
soil, Chemical Engineering Technology, 27, 890-894.

Arvai, J., and Gregory, R. 2003. Testing alternative decision approaches for identifying
cleanup priorities at contaminated sites, Environmental Science & Technology, 37(8), 14691476.

Baas, S. M., and Kwakernaak, H. 1977. Rating and ranking of multiple aspect alternative using
fuzzy sets, Automatica, 13(1), 47-58.

Balasubramaniam, A., Boyle, A. R., Voulvoulis, N. 2007. The appropriateness of multicriteria
analysis in environmental decision-making problems, Chemosphere, 66, 791-798.

Bana e Costa, C , ed. 1990. Reading in Multiple Criteria Decision Aid. Springer-Verlag.

Bardos, R. P., Martin, I. D., Kearney, T. 1999. Framework for evaluating remediation
technologies, Proc, IBC's 10th Conference- Contaminated Land, IBC technical services Ltd.

Bardos, R. P., Mariotti, C , Marot, F., Sullivan, T. 2000. Framework for Decision Support used
in Contaminated Land Management

in Europe

Management Support Inc., Maryland, USA.

123

and North America,

Environmental

Bella, A., Duckstein, L., Szidarovszky, F. 1996. A multicriterion analysis of the water
allocation conflict in the upper Rio Grande basin, Applied Mathematics and Computaion, 77(23), 245-265.

Bellman, R. E., and Zadeh, L. A. 1970. Decision making in a fuzzy environment. Management
Science 17, 141-164.

Bender, M. J., and Simonovic, S. P. 2000. A fuzzy compromise approach to water resource
systems planning under uncertainty, Fuzzy Sets and Systems, 115, 35-44.

Bonano, E. J., Apostolakis, G. E., Salter, P. F., Ghassemi, A. 2000. Application of risk
assessment and decision analysis to the evaluation, ranking and selection of environmental
remediation alternatives, Journal of Hazardous Materials, 71(1-3), 35-57.

Bonissone, P. P. 1982. A fuzzy set based linguistic approach: theory and applications,
Approximate Reasoning in Decision Analysis, North-Holland, 329-339.

Borsuk, M., Clemen, R., Maguire, L., Reckhow, K. 2001. Stakeholder values and scientific
modeling in the Neuse River Watershed, Group Decision and Negotiation, 10, 355-373.

Bose, U., Davey, A. M., and Olson, D. L. 1997. Multi-attribute utility methods in group
decision making: Past applications and potential for inclusion in GDSS. Omega 25, 691-706.

124

Carlsson, C. 1982. Tackling a MCDM-problem with the help of some results from fuzzy set
theory. European Journal of Operational Research 10, 270-281.

CCME (Canadian Council of Ministers of the Environment). 1993. Environmental code of
practice for underground storage tank systems containing petroleum products and allied
petroleum products. Beauregard Printers Limited.

Chang, N., Parvathinathan, G., Breeden, J. B. 2007. Combining GIS with fuzzy multicriteria
decision-Making for Landfill Siting in a Fast-Growing Urban Region, Journal of
Environmental Management.

Chen, S. J., and Hwang, C. L. 1992. Fuzzy Multiple Attribute Decision-Making, SpringerVerlag, New York.

Cheng, C. H. 1996. Evaluating naval tactical missile systems by fuzzy AHP based on the grade
of membership function, European Journal ofOperational Research, 96, 343-350.

Cheng, C. H., and Lin,Y. 2002. Evaluating the best main battle tank using fuzzy decision
theory with linguistic criteria evaluation. European Journal of Operational Research 142, 174186.

Cheng, S. K. Y. 2000. Development of a fuzzy multi-criteria decision support system for
municipal solid waste management, Masters Thesis, University of Ragina.

125

Civanlar, M. R., and Trussel, H.J. 1986. Constructing membership functions using statistical
data. Fuzzy Sets and Systems 18, (1) 1-13.

Cole, G.,M. 1994. Assessment and Remediation of Petroleum Contaminated Sites, Lewis
Publishers, USA.
Dou, C, Woldt, W., Bogardi, I. and Dahab, M. 1997. "Numerical solute transport simulation
using fuzzy sets approach ". Journal of Contaminant Hydrology 21, no. 1:107-126.

Dowd, R. M. 1985. Leaking underground storage tanks. Environ.Sci. & Tech. 18, 10-15.

Dubois, J. S., and Prade, H. 1994. Fuzzy sets- a convenient fiction for modelling vagueness and
possibility. IEEE transaction on Systems, Man, Cybernetics 2, no. 1:16-21.

Easton, A. 1973. One-of-a-Kind Decisions Involving Weighted Multiple Objectives and
Disparate Alternatives, University of South Carolina Press, Columbia, South Carolina.

Efstathiou, J. 1982. Ranking Fuzzy Sets: A decision theoretic approach, IEEE Transaction on
Systems, Man, Cybernetics, SMC-12, 655-659.

Efstathiou, J., and Rajkovic, V. 1979. Multiattribute decision making using a fuzzy Heurisitic
approach, IEEE Transaction on Systems, Man, Cybernetics, SMC-9, 326-333.

126

ESRC Global Environmental Change Programmed. 2000. Risky choices, soft disasters:
environmental decision-making under uncertainty, University of Sussex, Brighton, 25.

Fayek, A. R., and Sun, Z. 2001. A fuzzy expert system for design performance prediction and
evaluation, Canadian Journal of Civil Engineering, 28(1), 1-25.

Fenton, N., and Wang. W. 2006. Risk and confidence analysis for fuzzy multicriteria decision
making. Knowledge based system 19 (6),430-437.

Fernandes, L., Ridgley, M. A., van't Hof, T. 1999. Multiple criteria analysis integrates
economic, ecological and social objectives for coral reef managers, Coral Reefs, 18(4), 393402.

Figueira, J., Greco, S., Ehrgott, M. 2005. Multiple Criteria Decision Analysis: State of the Art
Surveys, Springer Science, United States of America.

French, S. 1996. Multi-attribute decision support in the event of a nuclear accident, Journal of
Multi-Criteria Decision Analysis, 5, 39-57.

Freissinet, C, Vauclin, M. and Erlich, M. 1999. Comparison of first-order analysis and fuzzy
set approach for the evaluation of imprecision in a pesticide groundwater pollution screening
model. J. Contain. Hydrol 27, 21-43.

127

Ganoulis, J. 2003. Evaluating alternative strategies for wastewater recycling and reuse in the
Mediterranean area, Water Science and Technology, 3(4), 11-19.

Gregory, R., and Failing, L. 2002. Using decision analysis to encourage sound deliberation:
water use planning in British Columbia, Canada, Journal of Policy Analysis and Management,
21(3), 492-499.

Gregory, R., and Wellman, K. 2001. Bringing stakeholder values into environmental policy
choices: a community-based estuary case study, Ecological Economics, 39(1), 37-52.

Gregory, R., and Keeney, R. L. 1994. Creating policy alternatives using stakeholder values,
Management Science, 40(8), 1035-1048.

Gutierrez, I., and Carmona, S. 1995. Ambiguity in multicriteria quality decision, International
Journal of Production Economics, 38(2-3), 215-224.

Hajkowicz, S. A., McDonald, G. T. and Smith, P. N. 2000. An evaluation of multiple objective
decision support weighting techniques in natural resource management. Journal

of

Environmental Planning and Management A2i, no. 4:505-518.

Hajkowicz, S.A., and Higgins, A. 2008. A comparison of multiple criteria analysis techniques
for water resource management. European Journal of Operational Research 184, 255-265.

128

Hellendoorn, H. 1997. After the fuzzy wave reached Europe, European Journal of Operational
Research, 99,58-71.

Herrera, F., and Verdegay, J. L. 1997. Fuzzy sets and operations research: perspectives. Fuzzy
Sets and Systems 90, 207-218.

Hong, I. B., and Vogel, D. R. 1991. Data and model management in a generalized MCDMDSS, Decision Sciences, 22(1), 25.

Hu, Z., Chan, C. W., Huang, G., H. 2003. A fuzzy expert system for site characterization,
Expert Systems with Applications, 24, 123-131.

Hwang, F. 1987. An expert decision making support system for multiple attribute decision
making, Ph.D. dissertation, Kansas state university.

Hwang, C. L., and Yoon, K. 1981. Multiple Attribute Decision Making-Methods

and

Applications, A State-of-the-Art Survey, Springer-Verlag, New York.

ICS-UNIDO and UNECE. 2000. Compendium of soil clean-up technologies and soil
remediation companies, second edition. New York and Geneva.

129

Janikowski, R., Kucharski, R., Sas-Nowosielska, A. 2000. Multi-criteria and multi-perspective
analysis of contaminated land management methods, Environmental

Monitoring

and

Assessment, 60, 89-102.

Joubert, A. R., Leiman, A., de Klerk, H. M., Katua, S., Aggenbach, J. C. 1997. Fybos (fine
bush) vegetation and the supply of water: a comparison of multi-criteria decision analysis and
cost-benefit analysis, Ecological Economics, 22(2), 123-140.

Juang, C. H., and Lee, D. H. 1991. A fuzzy scale for measuring weights of criteria in
hierarchical structures, IFES, 415-421.

Kamnikar, B. 2001. Managing petroleum contaminated soil: department of transportation
perspective, Journal of Environmental Engineering, 127(12), 1080-1088.

Kaufmann, A., and Gupta, M. M. 1985. Introduction to Fuzzy Arithmetic: Theory and
Applications. New York: Van Nostrand Reinhold.

Khadam, I. M., and Kalauarachchi, J. J. 2003. Multi-criteria decision analysis with
probabilistic

risk assessment

for the management

of contaminated

ground

water.

Environmental Impact Assessment Review 23, 683-721.

Khan, F. I., Hussain, T., Hejazi, R. 2004. An overview and analysis of site remediation
technologies, Journal of Environmental Management, 71, 95-122.

130

Khan, F. I., Sadiq, R., Husain, T. 2002. GreenPro-I: A risk-based life cycle assessment and
decision-making methodology for process plant design, Environmental Modelling and
Software, 17(8), 669-692.

Khan, F. I., and Sadiq, R. 2005. Risk based prioritization of air pollution monitoring using
fuzzy synthetic evaluation technique, Environmental Monitoring and Assessment, 105, 261283.

Klir, G. J., and Folger, T. 1988. Fuzzy Sets, Uncertainty and Information. Englewood Cliffs,
NJ: Prentice Hall.

Kuncheva, L. I., Wrench, J., Jain, L. C, Al-Zaidan, A. S. 2000. A fuzzy model of heavy metal
loadings in Liverpool Bay," Environmental Modelling & Software, 15(2), 161-167.

Lee, H. 1996. Applying fuzzy set theory to evaluate the rate of aggregative risk in software
development, Fuzzy Sets and Systems, 79(3), 323-336.

Lehr, J. H. 2004. Wiley's Remediation Technologies Handbook, 1st Ed., John Wiley & Sons,
Inc., Hoboken, New Jersey.

Li, J. B., Huang, G. H., Chakma, A., Chen, Z., Liu, L., Zeng, G. M. 2000. URREM: A decision
support system for subsurface modeling, risk assessment and site remediation, Proceedings of

131

Critical Technologies to the World in 21s Century: Pollution Control and Reclamation in
Process Industries, IAWQ Specialty Conference, Beijing, China, September 2000, 882-889.

Li, J.B. 2003. Development of an inexact environmental modeling system for the management
of petroleum contaminated Sites, University of Regina, Regina, Saskatchewan.

Li, J.B., Huang, G., H., Chakma, A., Zeng, G. M., Liu, L. 2003. Integrated fuzzy-stochastic
modeling of petroleum contamination in subsurface, Energy Sources, 25, 547-563.

Li, J.B., Huang, G., H., Zeng, G. M. 2001. An integrated decision support system for the
management of petroleum contaminated sites, Journal of Environmental Science and Health,
36(7), 1163-1186.

Li, J.B., Huang, G., H., Zeng, G., Maqsood, I., Huang, Y. 2007. An integrated fuzzy-stochastic
modeling approach for risk assessment of groundwater

contamination, Journal

of

Environmental Management, 82, 173-188.

Li, J. B., Liu, L., Huang, G., Zeng, G. M. 2006. A fuzzy-set approach for addressing
uncertainties in risk assessment of hydrocarbon-contaminated Site, Water, Air and Soil
Pollution, 171(1-4), 5-18.

Lind, E. A. 1995. Social Conflict and Social Justice: Some Lessons from the Social Psychology
of Justice, University Press, Leiden, The Netherlands.

132

Lind, E. A., and Tyler, T. R. 1988. Social Psychology of Procedural Justice, Plenum,
NewYork.

Linkov, I., Satterstrom, F. K., Kiker, G., Seager, T. P., Bridges, T., Gardner, K. H., Rogers, S.
H., Belluck, D. A., Meyer, A. 2006. Multicriteria decision analysis: A comprehensive decision
approach for management of contaminated sediments, Risk Analysis, 26(1), 61-78.

Mardle, S., and Pascoe, S. 2002. Modelling the effects of trade-offs between long and shortterm objectives in fisheries management, Journal ofEnvironmental Management, 65(1), 49-62.

Mclntyre, C, and Parfitt, M. K. 1998. Decision support system for residential land
development site selection process, Journal of Architect Engineering ASCE, 4, 125-131.

Mikhailov, L., and Tsvetinov, P. 2004. Evaluation of services using a fuzzy analytic hierarchy
process. Applied Soft Computing 5, 23-33.

Miller, G. A. 1965. The Magic Number Seven, Plus or Minus Seven, Psychological Review,
63, 81-97.

Mohamed, A. M. O., and Cote, K. 1999. Decision analysis of polluted sites-a fuzzy set
approach. Waste management 19, 519-533.

133

Moran, D., A. McVittie, D. J. Allcroft, and D. A. Elston. 2007. Quantifying public preferences
for agri-environmental policy in Scotland: A comparison of methods. Ecological Economics
63, 42-53.

Najjaran, H., Sadiq, R., Rajani, B. 2006. Fuzzy expert system to assess corrosion of cast/ductile
iron pipes from backfill properties, Computer-Aided Civil and Infrastructure Engineering, 21,
67-77.

Newton, J. 1991. Investigating leaking underground storage tanks. Pollution Engineering. 23,
80-83.

Nijkamp, P., Rietveld, P., Voogd, H. 1990. Multicriteria evaluation in physical planning,
Elsevier Science Publishers, Amsterdam, The Netherlands.

Ning, S. K., and Chang, N. B. 2002. Multi-objective, decision-based assessment of a water
quality monitoring network in a river system, Journal of Environmental Monitoring, 4, 121126.

Norbis, M., and Smith, J. M. 1996. An interactive decision support system for the resource
Constrained Scheduling Problem, European Journal of Operational Research, 94(1), 54-65.

134

NRTEE. 1997. Removing barriers: redeveloping national round table on the environment and
the economy, Canadian mortgage and housing corporation. Canada: Golder Associates ltd.
and McCarthy Tetrault.

Osgood, C. E., Suci, G. J., Tannenbaum, P. H. 1975. The measurement of meaning, 9th Ed.,
University of Illinois Press, Urbana.

Petrovic, S., and Petrovic, R. 2002. A new fuzzy multi-criteria methodology for ranking of
alternatives. Intl. Trans, in Op. Res. 9, 73-84.

Petts, J., Cairney, T., and Smith, M. 1997. Risk-based contaminated land investigation and
assessment. Chichester: John Wiley & Sons.

Pollard, S. J. T., Brookes, A., Earl, N., Lowe, J., Kearney, T., Nathanail, C. P. 2004. Integrating
decision tools for the sustainable management of land contamination, Science of the Total
Environment, 325, 15-28.

Prato, T. 2003. Multiple-attribute evaluation of ecosystem management for the Missouri river
system, Ecological Economics, 45(2), 297-309.

Qin, X. S., Huang, G. H., Huang, Y. F., Zeng, G. M., Li, J. B. 2006. NRSRM: A decision
support system and visualization software for the management of petroleum-contaminated
sites, Energy Sources, 28, 199-220.

135

Reddy, K. R., Adams, J. A., Richardson, C. 1999. Potential technologies for remediation of
brownfields, Practice Periodical of Hazardous, Toxic, and Radioactive Waste Management,
3(2), 61-68.

Riser-Roberts, E. 1998. Remediation of petroleum contaminated soil: Biologica, Physical, and
Chemical Processes, Lewis Publishers, Boca Raton, FL.

Rogers, S. H., Seager, T. P., Gardner, K. H. 2004. Combining expert judgement and
stakeholder values with Promethee: A case study in contaminated sediments management,
Comparative Risk Assessment and Environmental Decision Making, Kluwer Academic
Publishers, Boston, 305-322.

Saaty, T. L. 1994. How to make a decision: The analytic hierarchy process. Interfaces, 24, 1943.

Sadiq, R , Hussain, T., Veitch, B., Bose, N. 2004a. Risk-based decision making for drilling
waste discharges using a fuzzy synthetic evaluation technique, Ocean Engineering, 31, 19291953.

Sadiq, R., Rajani, B., Kleiner, Y. 2004b. Fuzzy- based method to evaluate soil corrosivity for
Prediction of Water Main Deterioration, Journal of Infrastructure Systems, 10(4), 149-156.

136

Seager, T. P., Lambert, J. H. and Gardner, K. H. 2007. Fostering innovation in contaminated
sediments management through multicriteria technology assessment and public participation.
Risk Analysis 27, no. 4:1043-1052.

Simons, R. 1998. Turning brownfields into greenback. Washington DC: Urban Land Istitute.

Siciliano, S. D., and Germida, J. J. 1998. Mechanisms of phytoremediation: biochemical and
ecological interactions between plants and bacteria, Environ. Rev., 6, 65-79.

Soesilo, J. A., and Wilson, S. R. 1997. Site remediation planning and managemen.lst ed. USA:
Lewis Publishers.

Soma, K. 2003. How to involve stakeholders in fisheries management-a country case study in
Trinidad and Tobago, Marine Policy, 27(1), 47-58.

Sousa, D., C. 2001. Contaminated Sites: The Canadian situation in an international context,
Journal ofEnvironmental Management, 62(131), 154.

Susan, R. 2003. A decision tree for selecting bioremediation in cold climates, Proc,
Assessment and Remediation of Contaminated Sites in Arctic and Cold Climates, ARCSACC,
Edmonton, 183-197.

137

Tang, M. T., Tzeng, G. H., Wang, S. W. 1999. A hierarchy fuzzy MCDM method for studying
electronic marketing strategies in the information service industry, Journal of International
Information Management, 8(1), 1-22.

Teng, J. Y., and Tzeng, G. H. 1996. Fuzzy multicriteria ranking of urban transportation
investment alternatives, Transportation Plannig Technology, 20(1), 15-31.

Testa, S. M., and Winegardner, D. L. 1991. Restoration of petroleum-contaminated aquifers,
Lewis Publishers, Chelsea, MI.

Turksen, I. B. 1991. Measurement of membership functions and their acquisition. Fuzzy sets
and Systems 40, no. 1:5-34.

UNECE, 1997. Compendium of soil clean-up technologies and soil remediation companies,
first edition, New York and Geneva.

Urban Task Force. 1999. Towards an urban renaissance, E&FN Spon, London.

USEPA (U.S. Environmental Protection Agency). 1993. Remediation technologies screening
matrix and reference guide. Washington, DC.

USEPA (U.S. Environmental Protection Agency). 2001. Abstracts of remediation case studies,
Washington, DC.

138

USEPA (U.S. Environmental Protection Agency). 2005. A decision-making framework for
cleanup of sites impacted with light non-aqueous phase liquids (LNAPL). EPA 542-R-04-011,
Washington, DC.

USEPA (U.S. Environmental Protection Agency). 2008. Remediation technologies screening
matrix and reference guide, version 4.0. [cited January/01 2008]. Available

from

http://www.frtr.gov/matrix2/top_page.html.

van Moeffaert, D. 2002. Multi-criteria decision aid in sustainable urban water management,
Masters Thesis, Royal Institute of Technology, Stockholm, Sweden.

Vranes, E., Gonzalez, V., Lodolo, A., Miertus, S. 2000. Decision support tool: applications in
remediation technology evaluation and selection, Environmental management support, Inc.,
Maryland.

Wang, L. X. 1997. A course in fuzzy systems and control, Prentice-Halljnc, USA.

Wenstop, F. 1976. Fuzzy set simulation models in a systems dynamic perspective, Kybernetes,
5,209-218.

Wong, J. K. W., and Li, H. 2008. Application of the analytic hierarchy process (AHP) in
multi-criteria analysis of the selection of intelligent building systems. Building
Environment 43, 108-125.

139

and

Yager, R. R.; Filev, D. P. 1994. Essentials of fuzzy modelling and control; John Wiley & Sons
Inc.: New York: 1994.

Yeh, C. H., Deng, H. and Chang, Y.H. 2000. Fuzzy multicriteria analysis for performance
evaluation of bus companies. European Journal of Operational Research 126, 459-473.

Yoon, K. P., and Hwang, C. L. 1995. Multiple attribute decision making an introduction, Sage
Publications, Thousand Oaks, California.

Yurdakul, M. 2004. AHP as a strategic decision-making tool to justify machine tool selection.
Journal of Materials Processing Technology 146, 365-376.

Zadeh, L. A. 1965. Fuzzy Sets, Information and Control, 8, 338-353.

Zahedi, F. 1986. The Analytical Hierarchy Process:The Method and its application. Interfaces,
16, 96-108.

Zapatero, E., Smith, C. H., and Weistroffer, H. R. 1997. Evaluating multiple attribute decision
support software. Journal of Multi-criteria Decision Analysis 6, 201-214.

Zimmermann, H. -J. 1996. Fuzzy set theory and its applications. Boston: Kluwer Academic
Publishers.

140

APPENDIX I Questionnaire

Parti:
1. Which best describes your age group? (Please select one)
a. Under 25

[ ]

b. 25-30

[]

c. 31-40

[ ]

d. 41-55

[]

e. 56-65

[ ]

f. Over 65

[]

2. Highest level of education completed? (Please select one)
a. 12th grade, no diploma
[]
b. High school graduate
[]
c. Some college credit, but less than 1 year [ ]
d. 1 or more years of college, no degree
[]
e. Associate degree (i.e., AA, AS)
[ ]
f. Bachelor's degree (i.e.,BA,AB,BS)
[]
g. Master's degree (i.e.,MA, MS, MEng) [ ]
h. Professional degree (i.e., MD,LLB,JD) [ ]
i. Doctoral degree (i.e.,PhD,EdD)
[]
j . Other
[]
3. Which one of the following best describes your professional affiliation? (Please select one)
a. Institution
[ ]
b. Industry
[]
(i.e., universities)
(i.e.,oil and gas industries)
c. Federal government
[ ]
d. Provincial government
e. Research organization
[]
f. Non-governmental / non-profit organization [ ]
g. Environmental consulting firms [ ]
h. First nation community
[]
i. General public
[ ]
j . Other
[]
4. How many years of experience do you have in your profession?
a. less than 2 years
[]
b. 2 -5 years
c. More than 5 years
[ ]
d. Not applicable

[]
[]

Please read these important definitions, these are helpful references for the next sections:
1. Community Acceptability criterion
Refers to the level of technology acceptability by members of the general public that live or work near
the contaminated site.
2. Minimum Achievable Concentration Criterion
Addresses the degree to which the technology is able to meet remediation objectives.
3. Time to Complete Cleanup (in-situ and ex- situ) Criterion
Refers to the time required to complete remediation, including site closure time by a technology.
4. Overall Cost (in-situ and ex- situ) Criterion
Includes design, construction, operations and maintenance (O&M) costs of a remediation technology, exclusive of
mobilization, demobilization, and pre- and post- treatment.
5. Development Status of a Technology
Refers to the current status of the technology, i.e., experimental (laboratory scale), pilot scale, full scale, or the
technology is already being applied as commercially and industrially.

141

6. Availability of the Technology Criterion
Number of vendors that can design, construct, and maintain the technology.
7. Technology Maintenance Criterion
Refers to the level of complexity of the technology and how easy it is to maintain. If high maintenance is required
for a technology this will indicate low reliability of the technology.
8. Regulatory (permitting) Acceptability Criterion
Degree of regulatory and permitting acceptability of the technology.

142

Your opinion on criteria importance

Overall cost

Data requirements

Availability

Maintenance

Regulatory
permitting
acceptability

4

5

6

7

8

3

Minimum
achievable
concentration
Time to complete
clean up

Criteria
Community
acceptability

2

1

Very low

Very low

Very low

Very low

Low

Low

Low

Low

Low

Low

Very low
Very low

Low

Low

Very low

Very low

Low
medium

Low
medium

Low
medium

Low
medium

Low
medium

Low
medium

Low
medium

Medium

Medium

Medium

Medium

Medium

Medium

Medium

143

to

to

to

to

to

to

to

Medium
High

Medium
High

Medium
High

Medium
High

Medium
High

Medium
High

Medium
High

to High

to High

to High

to High

to High

to High

to High

Linguistic scale to rate the importance of criteria
Low
to Medium
Medium
to High
medium
High

Very high

Very high

Very high

Very high

Very high

Very high

Very high

Very high

No
opinion

No
opinion

No
opinion

No
opinion

No
opinion

No
opinion

No
opinion

No
opinion

5. Please rate the following 10 criteria in terms of their importance to you, by selecting the linguistic scale below. For explanation on criteria you may refer
to the previous page.

Different criteria carry different weights among decision makers to select a remediation technology.

Part II:

Part III: Opinion on criteria value range
A. Community acceptability
Refers to the level of technology acceptability by members of the general public that live or work near the
contaminated site.
Note: Provided that the technology is effective for Petroleum Hydrocarbon contaminants
6. When would you consider that acceptance of a technology is low in the community? (You may select more
than one response)
a. If the technology is accepted by 10% or less of the community involved
b. If the technology is accepted by 20% or less of the community involved
c. If the technology is accepted by 30% or less of the community involved
d. If the technology is accepted by 40% or less of the community involved
e. Ifthe technology is accepted by 50% or less of the community involved
f. Ifthe technology is accepted by 60% or less of the community involved
g. Ifthe technology is accepted by 70% or less of the community involved
h. Ifthe technology is accepted by 80% or less of the community involved
i. Ifthe technology is accepted by 90% or less of the community involved
j . No opinion
7. When would you consider that acceptance of a technology is low to medium in the community? (You may
select more than one response)
a. If the technology is accepted by aprox. 10% of the community involved
b. Ifthe technology is accepted by aprox. 20% of the community involved
c. Ifthe technology is accepted by aprox. 30% of the community involved
d. Ifthe technology is accepted by aprox. 40% of the community involved
e. Ifthe technology is accepted by aprox. 50% of the community involved
f. Ifthe technology is accepted by aprox. 60% of the community involved
g. Ifthe technology is accepted by aprox. 70% of the community involved
h. Ifthe technology is accepted by aprox. 80% of the community involved
i. Ifthe technology is accepted by aprox. 90% of the community involved
j . No opinion
8. When would you consider that acceptance of a technology is medium in the community? (You may select
more than one response)
a. Ifthe technology is accepted by aprox. 10% of the community involved
b. Ifthe technology is accepted by aprox. 20% of the community involved
c. Ifthe technology is accepted by aprox. 30% of the community involved
d. Ifthe technology is accepted by aprox. 40% of the community involved
e. Ifthe technology is accepted by aprox. 50% of the community involved
f. Ifthe technology is accepted by aprox. 60% of the community involved
g. Ifthe technology is accepted by aprox. 70% of the community involved
h. Ifthe technology is accepted by aprox. 80% of the community involved
i. Ifthe technology is accepted by aprox. 90% of the community involved
j . No opinion
9. When would you consider that acceptance of a technology is medium to high in the community? (You
may select more than one response)
a. Ifthe technology is accepted by aprox. 10% of the community involved b. Ifthe technology is accepted by
aprox. 20% of the community involved
c. Ifthe technology is accepted by aprox. 30% of the community involved
d. Ifthe technology is accepted by aprox. 40% of the community involved
e. Ifthe technology is accepted by aprox. 50% of the community involved
f. Ifthe technology is accepted by aprox. 60% of the community involved
g. Ifthe technology is accepted by aprox. 70% of the community involved
h. Ifthe technology is accepted by aprox. 80% of the community involved

144

i. If the technology is accepted by aprox. 90% of the community involved
j . No opinion
10. When would you consider that acceptance of a technology is high in the community? (You may select
more than one response)
a. If the technology is accepted by 10% or greater of the community involved
b. If the technology is accepted by 20% or greater of the community involved
c. If the technology is accepted by 30% or greater of the community involved
d. If the technology is accepted by 40%) or greater of the community involved
e. If the technology is accepted by 50% or greater of the community involved
f. If the technology is accepted by 60% or greater of the community involved
g. If the technology is accepted by 70% or greater of the community involved
h. If the technology is accepted by 80% or greater of the community involved
i. If the technology is accepted by 90% or greater of the community involved
j . No opinion

B. Minimum achievable concentration
Different remediation technologies have various level of capability to achieve expected contaminant concentration.
The following questions ask about your opinion to rate a technology depending on its achieved concentration. The
linguistic scale to evaluate this criteria value range includes bad, poor, average, good, and excellent.
Note: Provided that the technology is effective for Petroleum Hydrocarbon contaminants
11. When would you rate a technology as low? (You may select more than one response)
a. If contaminant concentration of interest is reduced 10% or less
b. If contaminant concentration of interest is reduced 20% or less
c. If contaminant concentration of interest is reduced 30% or less
d. If contaminant concentration of interest is reduced 40% or less
e. If contaminant concentration of interest is reduced 50% or less
f. If contaminant concentration of interest is reduced 60% or less
g. If contaminant concentration of interest is reduced 70% or less
h. If contaminant concentration of interest is reduced 80% or less
i. If contaminant concentration of interest is reduced 90% or less
j . No opinion
12. When would you rate a technology as low to average? (You may select more than one response)
a. If contaminant concentration of interest is reduced approx. 10%
b. If contaminant concentration of interest is reduced approx. 20%
c. If contaminant concentration of interest is reduced approx. 30%
d. If contaminant concentration of interest is reduced approx. 40%
e. If contaminant concentration of interest is reduced approx. 50%
f. If contaminant concentration of interest is reduced approx. 60%
g. If contaminant concentration of interest is reduced approx. 70%
h. If contaminant concentration of interest is reduced approx. 80%
i. If contaminant concentration of interest is reduced approx. 90%
j . No opinion
13. When would you rate a technology as average? (You may select more than one response)
a. If contaminant concentration of interest is reduced approx. 10%
b. If contaminant concentration of interest is reduced approx. 20%
c. If contaminant concentration of interest is reduced approx. 30%
d. If contaminant concentration of interest is reduced approx. 40%
e. If contaminant concentration of interest is reduced approx. 50%
f. If contaminant concentration of interest is reduced approx. 60%

145

g. If contaminant concentration of interest is reduced approx. 70%
h. If contaminant concentration of interest is reduced approx. 80%
i. If contaminant concentration of interest is reduced approx. 90%
j. No opinion
14. When would you rate a technology as average to high? (You may select more than one response)
a. If contaminant of interest concentration is reduced approx. 10%
b. If contaminant of interest concentration is reduced approx. 20%
c. If contaminant of interest concentration is reduced approx. 30%
d. If contaminant of interest concentration is reduced approx. 40%
e. If contaminant of interest concentration is reduced approx. 50%
f. If contaminant of interest concentration is reduced approx. 60%
g. If contaminant of interest concentration is reduced approx. 70%
h. If contaminant of interest concentration is reduced approx. 80%
i. If contaminant of interest concentration is reduced approx. 90%o
j. No opinion
15. When would you rate a technology as high? (You may select more than one response)
a. If contaminant concentration of interest is reduced 10% or greater
b. If contaminant concentration of interest is reduced 20% or greater
c. If contaminant concentration of interest is reduced 30% or greater
d. If contaminant concentration of interest is reduced 40%) or greater
e. If contaminant concentration of interest is reduced 50% or greater
f. If contaminant concentration of interest is reduced 60% or greater
g. If contaminant concentration of interest is reduced 70%o or greater
h. If contaminant concentration of interest is reduced 80% or greater
i. If contaminant concentration of interest is reduced 90% or greater
j. No opinion

C. Clean-up time (in-situ and ex-situ):
Refers to the time required to complete remediation, including site closure time, hi general the technologies with
high efficiency might have higher preference over technologies with low efficiency. The following questions are to
survey your opinion on clean up time required for both in-situ and ex-situ remediation treatments.
In-situ treatment: contaminated soil is treated without excavation. In a general setting in-situ treatment is a slow
process compared to ex-situ treatment.
Note: Provided that the technology is effective for Petroleum Hydrocarbon (crude oil) contaminants, the
mentioned time for in-situ treatment is based on a standard site and under favorable conditions. It is
assumed that contaminated site area is 40m X 40m (1600m2); the average depth of soil contamination is
from the surface to a depth of 5m. The contaminated soil mass is approx. 20,000 tons. Only contaminated
soil is considered for treatment

16. When using in-situ treatment, what time period would you consider as a short clean-up time for the
above case? (You may select more than one response)
a. Approximately 6 months or less
b. Approximately 1 year or less
c. Approximately 18 months or less
d. Approximately 2 years or less
e. Approximately 30 months or less
f. Approximately 3 years or less
g. Approximately 42 months or less
h. Approximately 4 years or less
i. Approximately 54 months or less
j . Approximately 5 years or less
k. No opinion
17. When using in-situ treatment, what time period would you consider as a short to medium clean-up time
for the above case? (You may select more than one response)
a. Approximately 6 months
b. Approximately 1 year

146

c. Approximately 18 months
e. Approximately 30 months
g. Approximately 42 months
i. Approximately 54 months
k. No opinion

d. Approximately 2 years
f. Approximately 3 years
h. Approximately 4 years
j . Approximately 5 years

18. When using in-situ treatment, what time period would you consider as a medium clean-up time for the
above case? (You may select more than one response)
a. Approximately 6 months
b. Approximately 1 year
c. Approximately 18 months
d. Approximately 2 years
e. Approximately 30 months
f. Approximately 3 years
g. Approximately 42 months
h. Approximately 4 years
i. Approximately 54 months
j . Approximately 5 years
k. no opinion
19. When using in-situ treatment, what time period would you consider as a medium to long clean-up time
for the above case? (You may select more than one response)
a. Approximately 6 months
b. Approximately 1 year
c. Approximately 18 months
d. Approximately 2 years
e. Approximately 30 months
f. Approximately 3 years
g. Approximately 42 months
h. Approximately 4 years
i. Approximately 54 months
j . Approximately 5 years
k. No opinion
20. When using in-situ treatment, what time period would you consider as a long clean-up time for the
above case? (You may select more than one response)
a. Approximately 6 months or greater
b. Approximately 1 year or greater
c. Approximately 18 months or greater
d. Approximately 2 years or greater
e. Approximately 30 months or greater
f. Approximately 3 years or greater
g. Approximately 42 months or greater
h. Approximately 4 years or greater
i. Approximately 54 months or greater
j . Approximately 5 years or greater
k. No opinion

Ex-situ treatment: contaminated soil is treated after excavation. In a general setting ex-situ treatment is a fast
process compared to in-situ treatment.
Note: Provided that the technology is effective for Petroleum Hydrocarbon (crude oil) contaminants, the
mentioned time for ex-situ treatment is based on a standard site and under favorable conditions. It is
assumed that contaminated site area is 40m X 40m (1600m2); the average depth of soil contamination is
from the surface to a depth of 5m. The contaminated soil mass is approx. 20,000 tons. It is also assumed that
the soil is excavated and treated.
21. When using ex-situ treatment, what time period would you consider as a short clean-up time for the
above case? (You may select more than one response)
a. Approximately 4 months or less
b. Approximately 5 months or less
c. Approximately 6 months or less
d. Approximately 7 months or less
e. Approximately 8 months or less
f. Approximately 9 months or less
g. Approximately 10 months or less
h. Approximately 11 months or le
i. Approximately 12 months or less
j . No opinion
22. When using ex-situ treatment, what time period would you consider as a short to medium clean-up time
for the above case? (You may select more than one response)
a. Approximately 4 months
b. Approximately 5 months
c. Approximately 6 months
d. Approximately 7 months

147

e. Approximately 8 months
g. Approximately 10 months
i. Approximately 12 months

f. Approximately 9 months
h. Approximately 11 months
j . No opinion

23. When using ex-situ treatment, what time period would you consider as a medium clean-up time for the
above case? (You may select more than one response)
a. Approximately 4 months
b. Approximately 5 months
c. Approximately 6 months
d. Approximately 7 months
e. Approximately 8 months
f. Approximately 9 months
g. Approximately 10 months
h. approximately 11 months
i. Approximately 12 months
j . no opinion
24. When using ex-situ treatment, what time period would you consider as a medium to long clean-up time
for the above case? (You may select more than one response)
a. Approximately 4 months
b. Approximately 5 months
c. Approximately 6 months
d. Approximately 7 months
e. Approximately 8 months
f. Approximately 9 months
g. Approximately 10 months
h. Approximately 11 months
i. Approximately 12 months
j . No opinion
25. When using ex-situ treatment, what time period would you consider as a long clean-up time for the
above case? (You may select more than one response)
a. Approximately 4 month or greater
b. Approximately 5 month or greater
c. Approximately 6 month or greater
d. Approximately 7 month or greater
e. Approximately 8 month or greater
f. Approximately 9 month or greater
g. Approximately 10 month or greater
h. Approximately 11 month or greater
i. Approximately 12 month or greater
j . No opinion

D. Overall Cost
Includes design, construction, operations and maintenance (O&M) costs of the core process that defines each
technology, exclusive of mobilization, demobilization, and pre- and post- treatment.

In- situ treatment: contaminated soil is treated without excavation.
Note: Provided that the technology is effective for Petroleum Hydrocarbon (crude oil) contaminants, the
mentioned cost for in-situ treatment is based on a standard site and under favorable conditions. It is
assumed that contaminated site area is 40m X 40m (1600m2); the average depth of soil contamination is
from the surface to a depth of 5m. The contaminated soil mass is approx. 20,000 tons. Only contaminated
soil is considered for treatment.
26. Which value would you consider as low cost (per cubic meter) when in-situ treatment is considered for
the above case? (You may select multiple or single options)
a. Approximately $50 per m3 or less
b. Approximately $75 per m3 or less
3
c. Approximately $100 per m or less
d. Approximately $125 per m3 or less
e. Approximately $150 per m3 or less
f. Approximately $175 per m3 or less
g. Approximately $200 per m 3 or less

h. Approximately $225 per m 3 or less

i. Approximately $275 per m or less

j . No opinion

3

27. Which value would you consider as low to medium cost (per cubic meter) when in-situ treatment is
considered for the above case? (You may select more than one response)
a. Approximately $50 per m3
b. Approximately $75 per m3
3
c. Approximately $100 per m
d. Approximately $125 per m3
e. Approximately $150 per m3
f. Approximately $175 per m3

148

h. Approximately $225 per m3
j . No opinion

g. Approximately $200 per m
i. Approximately $275 per m3

28. Which value would you consider as medium cost (per cubic meter) when in-situ treatment is considered
for the above case? (You may select more than one response)
a. Approximately $50 per m3
b. Approximately $75 per m3
3
c. Approximately $100 per m
d. Approximately $125 per m3
3
e. Approximately $150perm
f. Approximately $175 perm3
3
g. Approximately $200 per m
h. Approximately $225 per m3
3
i. Approximately $275 per m
j . No opinion
29. Which value would you consider as medium to high cost (per cubic meter) when in-situ treatment is
considered for the above case? (You may select more than one response)
a. Approximately $50 per m3
b. Approximately $75 per m3
3
c. Approximately $100 per m
d. Approximately $125 per m3
3
e. Approximately $150 per m
f. Approximately $175 per m3
g. Approximately $200 per m3
h. Approximately $225 per m3
i. Approximately $275 per m3
j . No opinion
30. Which value would you consider as high cost (per cubic meter) when in-situ treatment is considered for
the above case? (You may select more than one response)
a. Approximately $50 per m3 or greater
b. Approximately $75 per m3 or greater
3
c. Approximately $100 per m or greater
d. Approximately $125 per m3 or greater
3
e. Approximately $150 per m or greater
f. Approximately $175 per m3 or greater
3
g. Approximately $200 per m or greater
h. Approximately $225 per m3 or greater
i. Approximately $275 per m3 or greater
j . No opinion
Ex situ treatment: contaminated soil is treated after excavation and excavation cost is assumed $50/ton.
Note: Provided that the technology is effective for Petroleum Hydrocarbon (crude oil) contaminants, the
mentioned cost for ex-situ treatment is based on a standard site and under favorable conditions. It is
assumed that contaminated site area is 40m X 40m (1600m2); the average depth of soil contamination is
from the surface to a depth of 5m. The contaminated soil mass is approx. 20,000 tons. The excavation cost is
$50 per ton. Excavation cost is included in the following values.

31. Which value you consider as low cost (per cubic meter) when ex-situ treatment is considered for above
case? (You may select more than one response)
a. Approximately $ 100 per yd3 or less
b. Approximately $ 125 per yd3 or less
3
c. Approximately $150 per m or less
d. Approximately $175 per m3 or less
3
e. Approximately $200 per m or less
f. Approximately $225 per m3 or less
3
g. Approximately $275 per m or less
h. Approximately $300 per m3 or less
i. Approximately $325 per m3 or less
j . No opinion
32. Which value you consider as low to medium cost (per cubic meter) when ex-situ treatment is considered
for above case? (You may select more than one response)
a. Approximately $ 100 per m3
b. Approximately $ 12 5 per m3
3
c. Approximately $150 per m
d. Approximately $175 per m3
3
e. Approximately $200 per m
f. Approximately $225 per m3
3
g. Approximately $275 per m
h. Approximately $300 per m3
3
i. Approximately $325 per m
j . No opinion
33. Which value you consider as medium cost (per cubic meter) when ex-situ treatment is considered for
above case? (You may select more than one response)
a. Approximately $100 per m3
b. Approximately $125 per m3
3
c. Approximately $150 per m
d. Approximately $175 per m3
e. Approximately $200 per m3
f. Approximately $225 per m3

149

g. Approximately $275 per m3
i. Approximately $325 per m3

h. Approximately $300 per m3
j . No opinion

34. Which value you consider as medium to high cost (per cubic meter) when ex-situ treatment is considered
for above case? (You may select more than one response)
a. Approximately $100 per m3
b. Approximately $125 per m3
3
c. Approximately $150 per m
d. Approximately $175 per m3
3
e. Approximately $200 per m
f. Approximately $225 per m3
g. Approximately $275 per m3
h. Approximately $300 per m3
i. Approximately $325 per m3
j . No opinion
35. Which value you consider as high cost (per cubic meter) when ex-situ treatment is considered for above
case? (You may select more than one response)
a. Approximately $100 per m3 or greater
b. Approximately $125 per m3 or greater
3
c. Approximately $ 150 per m or greater
d. Approximately $ 175 per m3 or greater
3
e. Approximately $200 per m or greater
f. Approximately $225 per m3 or greater
3
g. Approximately $275 per m or greater
h. Approximately $300 per m3 or greater
i. Approximately $325 per m3 or greater
j . No opinion

150

APPENDIX II Introduction of Remediation Technologies
A brief description of all the remediation alternatives used in the system is given below as a reference to the users
of the system. The description of the following remediation alternatives and criteria information is synthesized
from Lehr (2004) and USEPA (2008).

1. Enhanced bioremediation (in-situ biological treatment)
Technology description
Bioremediation is a general term used for the destruction of contaminants in soil, including microorganisms
(e.g., yeast, fungi, or bacteria), by biological mechanisms. In enhanced bioremediation, the activity of naturally
occurring microbes is stimulated by circulating water-based solutions through contaminated soils to enhance in
situ biological degradation of organic contaminants or immobilization of inorganic contaminants. Nutrients,
oxygen, or other amendments may be used to enhance bioremediation and contaminant desorption from subsurface
materials. Bioremediation may rely on either indigenous microorganisms (i.e., those that are native to the site)
or exogenous microorganisms (i.e., those that are imported from other locations. It can take place under aerobic
or anaerobic conditions. Under aerobic conditions, in the presence of sufficient oxygen and other nutrient
elements, microorganisms will ultimately convert many organic contaminants to carbon dioxide, water and
microbial cell mass. Under anaerobic conditions (i.e., in the absence of oxygen the organic contaminants will be
ultimately metabolized to methane, limited amounts of carbon dioxide and trace amounts of hydrogen gas. The
main advantage of the in situ process is that it allows soil to be treated without being excavated and transported,
resulting in less disturbance of site activities. When the clean up goal can be attained in an acceptable time
frame, it can save costs to the projects. This kind of process mostly requires longer time periods and there is an
uncertainty about the quality of the treatment.

a) Community acceptability: Above average;
b) Minimum achievable concentration: Above average;
c) Clean up time: The length of time required for treatment can range from 6 months to 5 years and is dependent
on many site-specific factors;

151

d) Clean up cost: Bioremediation is cost competitive. Typical costs for enhanced bioremediation range from $30
to $100 per cubic meter ($20 to $80 per cubic yard) of soil;
e) Development status: Above average (Full scale);
f) Technology availability: More than 4 vendors and commercially available;
g) Maintenance requirement: Average maintenance required. Access to the site for unexpected repairs,
adjustments, and regular maintenance is likely to be limited;
h) Regulatory acceptability: Average.

2. Bioventing (in-situ biological treatment)
Technology description
Bioventing stimulates the naturally occurring soil microorganisms to degrade compounds in soil by providing
oxygen. Oxygen is most commonly supplied through direct air injection into soil. Passive bioventing systems
use natural air exchange to deliver oxygen to the subsurface via bioventing wells. The rate of natural
degradation is generally limited by the lack of oxygen and other electron acceptors (i.e., a compound that gains
electrons during Biodegradation rather than by the lack of nutrients (i.e., electron donors).
a) Community acceptability: Above average;
b) Minimum achievable concentration: All aerobically biodegradable constituents can be treated by
bioventing. Bioventing techniques have been successfully used to remediate soils contaminated by petroleum
hydrocarbons, nonchlorinated solvents, some pesticides, wood preservatives, and other organic chemicals.
These techniques, while still largely experimental, show considerable promise of stabilizing or removing
inorganics from soil;
c) Clean up time: Average, 1-3 years;
d) Clean up cost: For small site $709 to $742 per cubic yard and for large site $60 to $84 per cubic yard ($928
to $970 per cubic meter or $79 to $109 cubic meter respectively);
e) Development status: Above average (Full).
f) Technology availability: More than 4 vendors. Bioventing is becoming more common and most of the
hardware components are readily available;
g) Maintenance requirement: Above average (low maintenance required);

152

h) Regulatory acceptability: Average;

3. Phytoremediation (in-situ biological treatment)
Technology description:
Phytoremediation is a process that uses plants to remove, transfer, stabilize, and destroy contaminants in soil and
sediment. Contaminants may be either organic or inorganic.
a) Community acceptability: Below average;
b) Ability to achieve minimum concentration: Data not available;
c) Clean up time: Below average, more than 3 years. The time required to remediate a site using bioventing is
highly dependent upon the specific soil and chemical properties of the contaminated media;
d) Clean up cost: For small site $479 to $1775, for large site $479 to $1775 ($626 to $2,322 and$147 to $483
per cubic meter);
e) Development status of the technology: Above average (full scale);
f) Technology availability: 2-4 vendors;
g) Maintenance requirement: Below average (high maintenance required);
h) Regulatory acceptability: Below average.

4. Soil vapor extraction (in-situ physical/chemical treatment)
Technology description
In soil vapor extraction technique a vacuum is applied to the soil to induce the controlled flow of air and remove
volatile and some semi volatile contaminants from the soil. The gas leaving the soil may be treated to recover or
destroy the contaminants, depending on local and state air discharge regulations. Vertical extraction vents are
typically used at depths of 1.5 meters or greater and have been successfully applied as deep as 91 meters.
Horizontal extraction vents (installed in trenches or horizontal borings) can be used as warranted by
contaminant zone geometry, drill rig access, or other site-specific factors.

a) Community acceptability: Average;

153

b) Minimum achievable concentration: Average;
c) Clean up time: The duration of operation and maintenance for in situ SVE is typically medium- to long-term.
In situ SVE projects are typically completed in 1 to 3 years;
d) Overall cost: Cost for SVE of contaminated soil varies from $944-$ 1,100 /cubic yard for a small site and for
large site $300-$722/cubic yard;
e) Development status: Full;
f) Availability of the technology: Above average, more than 4 vendors;
g) Maintenance: Low maintenance and high reliability;
h) Regulatory permitting acceptability: Average;

5. Biopiles / static pile (ex- situ biological treatment)
Technology description:
Excavated soils are mixed with soil amendments and placed in aboveground enclosures. It is an aerated static pile
composting process in which compost is formed into piles and aerated with blowers or vacuum pumps. Biopiles
are engineered systems in which excavated soils are combined with soil amendments, formed into compost piles
and enclosed for treatment. They are commonly provided with an air distribution system by blowers or vacuum
pumps. The leachate must be collected and treated. Several properties of the process such as nutrients and
oxygen can be controlled in order to enhance the remediation procedure. This technology is used to reduce
concentrations of petroleum constituents in excavated soils. The treatment area will generally be covered or
contained with an impermeable liner to minimize the risk of contaminants leaching into uncontaminated soil.
a) Community acceptability: Data not available and it is site specific;
b) Minimum achievable concentration: Data not available; Biopile treatment has been applied to treatment of
nonhalogenated VOCs and fuel hydrocarbons. Halogenated VOCs, SVOCs and pesticides can also be treated;
c) Clean up time: approximate 0.5 to 1 year;
d) Clean up cost: Typical costs with a prepared bed and liner are $130 to $260 per cubic meter ($30 to $60 per
cubic yard);
e) Development status of the technology: Above average (full scale). More than 4 vendors;

154

f) Technology availability: Above average (full scale).The technology is commercially available for treating
fuel contamination;
g) Maintenance requirement: Above average (low maintenance required);
h) Regulatory acceptability: Data not available.
5. Landfarming (ex-situ biological treatment)
Technology description:
Landfarming, also known as land treatment or land application, is an above-ground remediation technology for
soils. It reduces concentrations of petroleum constituents through biodegradation. Contaminated soils are mixed
with soil amendments such as soil bulking agents and nutrients and then tilled into the earth. The soil is spread
over an area and periodically turned to improve aeration. Turning the soil also avoids the disadvantages of
having heterogeneous degradation. Soil conditions are controlled to optimize the rate of contaminant
degradation. The enhanced microbial activity results in degradation of adsorbed petroleum product constituents
through microbial respiration. The petroleum industry has used landfarming for many years. Contaminated soil,
sediment, or sludge is excavated, applied into lined beds, and periodically turned over or tilled to aerate the waste.
Landfarming is extremely simple and inexpensive. Requires no process controls. Relatively unskilled personnel
can perform the technique. Certain pollutants can be completely removed from the soil.
a) Community acceptability: Average;
b) Minimum achievable concentration/ technology applicability: Average;
Land farming has been proven most successful in treating petroleum hydrocarbons and other less volatile
biodegradable contaminants. Because lighter, more volatile hydrocarbons such as gasoline are treated very
successfully by processes that use their volatility (i.e., soil vapor extraction), the use of aboveground
bioremediation is usually limited to heavier hydrocarbons. As a rule of thumb, the higher the molecular weight
(and the more rings with a PAH), the slower the degradation rate;
c) Clean up time: approximate 0.5 to 1 year;
d) Clean up cost: Costs prior to treatment (assumed to be independent of volume to be treated): $25,000 to
$50,000 for laboratory studies; and less than $100,000 for pilot tests or field demonstrations. Cost of prepared

155

bed (ex situ treatment and placement of soil on a prepared liner): Under $100 per cubic meter (under $75 per
cubic yard);
e) Development status of the technology: Above average (full scale);
1) Technology availability: Above average, more than 4 vendors;
g) Maintenance requirement: Low maintenance required;
h) Regulatory acceptability: Average;

6. Slurry phase treatment (ex-situ biological treatment)
Technology description:
An aqueous slurry is created by combining soil, sediment, or sludge with water and other additives. The slurry is
mixed to keep solids suspended and microorganisms in contact with the soil contaminants. Upon completion of the
process, the slurry is dewatered and the treated soil is disposed of.
a) Community acceptability: Average;
b) Minimum achievable concentration: Average ;
c) Clean up time: approximate 0.5 to 1 year;
d) Clean up cost: Treatment costs using slurry reactors range from $130 to $200 per cubic meter ($100 to $150
per cubic yard). Costs ranging from $160 to $210 per cubic meter ($125 to $160 per cubic yard) are incurred when
the slurry-bioreactor off-gas has to be further treated because of the presence of volatile compounds;
e) Development status of the technology: Above average (full scale);
f) Technology availability: Above average, More than 4 vendors;
g) Maintenance requirement: Average maintenance required;
h) Regulatory acceptability: Above average;

7. Low temperature thermal desorption-LTTD (ex-situ thermal treatment)
Technology description:
In LTTD, wastes are heated to between 90 and 320 °C (200 to 600 CF). Decontaminated soil retains its physical
properties. Unless being heated to the higher end of the LTTD temperature range, organic components in the soil
are not damaged, which enables treated soil to retain the ability to support future biological activity.

156

a) Community acceptability: Average;
b) Minimum achievable concentration/ technology applicability: Above average. The technology is targeted
to semi volatile halogenated and non-halogenated organic compounds, as well as other organics. LTTD is a
full-scale technology that has been proven successful for remediating petroleum hydrocarbon contamination in all
types of soil. Contaminant destruction efficiencies in the afterburners of these units are greater than 95%.
c) Clean up time: Less than 0.5 year;
d) Clean up cost: Cleanup cost for LTTD is approximate $75-$232/cubic yard for a small site. For a large site
the cost may vary from $44-$ 110;
e) Development status of the technology: Above average, full scale;
f) Technology availability: Above average, more than 4 vendors. Thermal desorption is a well established
technology;
g) Maintenance requirement: Average;
h) Regulatory acceptability: Average.

8. Soil washing (ex-situ physical/chemical treatment)
Technology description:
Soil washing uses water to remove contaminants from soils. The process works by either dissolving or
suspending contaminants in the wash solution. Contaminants which are absorbed onto soil particles are
separated from soil in an aqueous based system. High pressure water can be used to aid the removal from the
surface.
a) Community acceptability: Above average;
b) Minimum achievable concentration/ technology applicability: The target contaminant groups for this
technology are SVOCs, fuels and heavy metals, including radionuclides. The technology can be used on
selected VOCs and pesticides;
c) Clean up time: Less than 0.5 year;
d) Clean up cost: Cost for soil washing is approximately $142/cubic yard for small site and for large site
approximately $53/cubic yard;
e) Development status of the technology: Above average, full scale;

157

f) Technology availability: The technology of soil washing is used extensively in Europe. Commercialization
in the United States is not yet extensive; More than four vendors, above average
g) Maintenance requirement: Low maintenance required;
h) Regulatory permitting acceptability: Average.

9. Soilflushing(in-situ physical chemical treatment)
Technology description
Water, or water containing an additive to enhance contaminant solubility, is applied to the soil or injected into the
ground water to raise the water table into the contaminated soil zone. Contaminants are leached into the ground
water, which is then extracted and treated.
a) Community acceptability: Average;
b) Minimum achievable concentration/ technology applicability: Below average. The target contaminant
group for soilflushingis inorganics including radioactive contaminants. The technology can be used to treat
VOCs, SVOCs, fuels, and pesticides, but it may be less cost-effective than alternative technologies for these
contaminant groups;
c) Clean up time: 1 to 3 years;
d) Clean up cost: The cost of soil flushing depends greatly on the type and concentration of surfactants used, if
they are used at all. Rough estimates ranging from $25 to $250 per cubic yard have been reported.
e) Development status of the technology: Above average, full scale;
f) Technology availability: Above average, more than 4 vendors;
g) Maintenance requirement: Average;
h) Regulatory acceptability: Below average (worse).

158