Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Analysis of the economic risks caused by environmental issues in large infrastructure projects De Zoysa, Garumuni Niranjan Sanjaya 2000

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2000-0381.pdf [ 12.16MB ]
Metadata
JSON: 831-1.0063994.json
JSON-LD: 831-1.0063994-ld.json
RDF/XML (Pretty): 831-1.0063994-rdf.xml
RDF/JSON: 831-1.0063994-rdf.json
Turtle: 831-1.0063994-turtle.txt
N-Triples: 831-1.0063994-rdf-ntriples.txt
Original Record: 831-1.0063994-source.json
Full Text
831-1.0063994-fulltext.txt
Citation
831-1.0063994.ris

Full Text

ANALYSIS OF ECONOMIC RISKS CAUSED B Y E N V I R O N M E N T A L ISSUES IN L A R G E INFRASTRUCTURE PROJECTS by G A R U M U N I N I R A N J A N S A N J A Y A DE Z O Y S A B.Sc. (Engineering) Honors, University of Moratuwa, 1996 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF M A S T E R OF APPLIED SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Civil Engineering) We accept this thesis as conforming ifxthe required standard THE UNIVERSITY OF BRITISH COLUMBIA August 2000 © Garumuni Niranjan Sanjaya De Zoysa, 2000 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of C i u i l g n . c p r v e e r ' i r t The University of British Columbia Vancouver, Canada Date | | S e p t e m b e r 2.OOP . DE-6 (2/88) Abstract The objectives of the research described in the thesis are the identification and characterization of environmental issues associated with large infrastructure projects which affect the project's economic performance, and the development of a framework that will enable the analysis of the uncertainty in project economic variables caused by such issues. The focus of the research is on the natural environment. The risk analysis framework is developed with the aim of enabling project proponents and other stakeholders to identify, characterize and quantify environmental risks during the project appraisal stage. The risk identification process is based on the use of an extendable checklist or register of probable risks. This risk register acts as the starting point of the identification process, as well as a knowledge base that can be utilized on projects of a similar nature. The risks are categorized into 4 types based on the effect of environmental issues on project work packages. The categorization allows the selection of a suitable quantification method for the risks. Monte Carlo simulation and influence diagrams are employed in developing the quantification methods for the four risk types. Quantification of the economic variables is carried out by considering them in terms of their decomposed primary variables. Adjustment coefficients that represent the effect of external uncertainties are utilized in quantifying the cost and duration uncertainty of work packages subject to external uncertainty factors. Influence diagrams are used both to represent, and as an aid to quantifying such external uncertainty factors. The Pearson family of frequency i i curves is used to characterize the input distributions to the quantification process while the rank correlations between the primary variables are evaluated by considering their transformations into uniform variables on the unit interval [0,1]. A parameter of the joint distribution between the transformed variables is used to assess the rank correlation coefficient. The information elicitation process developed to obtain the information required for the quantification process makes use of decomposition, the use of directed questions, and appropriate feedback to elicit information of high quality from experts. The uncertainty associated with the economic variables is represented in a form suitable for incorporation into the overall project decision making process. This final output of the framework should help to enable the effective appraisal of the risks posed by environmental issues to the economic performance of the project. iii T A B L E O F C O N T E N T S Abstract ii List of Tables vi List of Figures vii Acknowledgements x CHAPTER 1 Introduction 1 1.1 Background for the Research 1 1.2 The Research Objectives 4 1.3 Structure of the Thesis 5 CHAPTER 2 Large Infrastructure Projects and the Environment 8 2.1 Introduction 8 2.2 The Natural Environment and its Sensitivity 10 2.3 Scope of Engineering Projects 13 2.4 Public Perceptions 14 2.5 Government Regulations 16 2.6 The Interaction between Projects and the Environment 26 2.7 Summary 38 CHAPTER 3 Environmental Risk Analysis Framework 40 3.1 Introduction 40 3.2 Risk Categorization Schemes 41 3.3 Risk Management Frameworks 50 3.4 Environmental Risk Analysis Framework 54 3.5 Summary 72 CHAPTER 4 Quantification of Type A Environmental Risks 73 4.1 Introduction 73 4.2 Economic Risk Quantification Methods 77 4.3 Quantification Methodology for Type A l and Type A2 Risks . 88 4.4 Conclusion 116 iv CHAPTER 5 Quantification of Type B Environmental Risks 118 5.1 Introduction 118 5.2 Influence Diagrams 121 5.3 Requirements for the Analysis of External Risks 132 5.4 External Risk Quantification Methods based on Influence Diagrams 136 5.5 Quantification Method for Type B Risks 143 5.6 Application of the Quantification Methodology 163 5.7 Summary 170 5.8 Conclusion 172 CHAPTER 6 Elicitation of Information from Experts 174 6.1 Introduction 174 6.2 Evocation of Information from Experts 175 6.3 Decomposition of Variables 178 6.4 Biases in Information Elicitation 180 6.5 Information Requirements for Analyzing Environmental Risks 183 6.6 The Information Encoding Process 185 6.7 The use of Information Technology and Multimedia in Information Elicitation 195 6.8 Application of the Environmental Risk Analysis Framework 200 6.9 Summary 221 CHAPTER 7 Conclusions and Recommendations 222 7.1 Conclusions 222 7.2 Suggestions for Future Research 225 Bibliography 227 Appendix A Preliminary Risk Register for Large Hydropower Projects 236 Appendix B Transformation of Random Variables 261 v List of Tables Table Page 3.1 Summary of Risk Categories adopted by Wideman 43 3.2 Risk Categorization Scheme adopted by Perry and Hayes 44 3.3 Risk Factors identified by Kangari and Boyer 45 3.4 Risk Categorization Scheme adopted by Al-Bahar and Crandall.... 47 4.1 Correlated Values of Labor Productivity and Work Quantity 85 4.2 Cumulative Probability Values of the Productivity Variable 113 5.1 Percentile Values of the Primary Variables used in the Example of Excavation Duration Estimation 167 5.2 Adjustment Coefficients used in the Example of Excavation Duration Estimation 167 5.3 Statistical Outputs obtained for the Excavation Duration 170 6.1 Directed Questions and Prompts based upon Argument Types 177 6.2 Use of Information Technology and Multimedia in Information Elicitation 199 6.3 Dam Removal Project on Ecstall River - Initial Estimates for the Statistical Parameters of Duration Variables 211 6.4 Dam Removal Project on Ecstall River - Possible States of the External Factors Affecting Sediment Disposal Duration 215 A . l Risk Register for Large Hydropower Projects 239 vi List of Figures Figure Page 2.1 Factors Affecting the Relationship between a Project and the Natural Environment 10 2.2 Types of Environmental Issues having an Uncertain Impact on the Economic Performance of an Infrastructure Project 30 2.3 Relevant Environmental Issues during the Different Phases of the Project Lifecycle 36 3.1 The Environmental Risk Analysis Process 55 3.2 Description of Work Packages for the Purpose of Environmental Risk Analysis ' 61 3.3 Description of the Natural Environment for the Purpose of Environmental Risk Analysis 62 3.4 Extract from a Risk Register for Large Hydropower Projects 64 3.5 Classification of Risks 66 4.1 Diagonal Band Distribution with Parameter 8 105 4.2 Correlated Values of a Uniform Distributions with a Rank Correlation Coefficient of 0.9 106 4.3 Correlated Values of a Uniform Distribution and a Normally Distributed Variable with a Rank Correlation Coefficient of 0.9 .... 110 5.1 Types of Nodes in Influence Diagrams 121 5.2 Influence Diagram for Evaluating Fuel Spill Clean up Cost 123 5.3 Probability Tree for Evaluating Fuel Spill Clean up Cost 124 5.4 Influence Diagram for Evaluating Fuel Spill Clean up Cost in the Presence of External Factors 127 5.5 Probability Tree for Evaluating Fuel Spill Clean up Cost in the Presence of External Factors 128 vn 5.6 Influence Diagram for Evaluating Fuel Spill Clean up Cost in the Presence of Interdependent External Factors 130 5.7 Influence Diagram for Evaluating the Excavation Cost in an Environmental Restoration Project 137 5.8 Probability Tree for Evaluating the Excavation Cost in an Environmental Restoration Project 137 5.9 Assignment of Multiplicative Factors to Primary Variables 146 5.10 Influence Diagram with a Hierarchy of Chance Nodes 147 5.11 Influence Diagram for the Example of Evaluating Excavation Duration 164 5.12 Probability Tree for the Example of Evaluating Excavation Duration 166 6.1 Representation of the Percentage Reduction in Uncertainty 194 6.2 Flowchart of the Information Encoding Process for Type A Risks . 196 6.3 Flowchart of the Information Encoding Process for Type B Risks . 197 6.4 Dam Removal Project on Ecstall River - Definition of the Project. 203 6.5 Dam Removal Project on Ecstall River - Definition of the Environment 203 6.6 Dam Removal Project on Ecstall River - Use of the Risk Register . 205 6.7 Dam Removal Project on Ecstall River - Updating the Risk Register 207 6.8 Dam Removal Project on Ecstall River - Project Work Package Network 209 6.9 Dam Removal Project on Ecstall River - Basic Influence Diagram for Evaluating Sediment Disposal Duration 213 6.10 Dam Removal Project on Ecstall River - Refined Influence Diagram for Evaluating Sediment Disposal Duration 214 6.11 Dam Removal Project on Ecstall River - Probability Tree for Evaluating Sediment Disposal Duration 216 vni 6.12 Dam Removal Project on Ecstall River - Percentile Estimates of Primary Variables 217 6.13 Dam Removal Project on Ecstall River - Probability Distribution of Sediment Quantity 218 6.14 Dam Removal Project on Ecstall River - Primary and Binary Adjustment Coefficients for the Productivity Variable 220 ix Acknowledgement I gratefully acknowledge the guidance of my supervisor, Dr. Alan Russell during the preparation of this thesis. His words of advice and encouragement are greatly appreciated. I would also like to extend my sincere thanks to Dr. Thomas Froese for reviewing the thesis as well as for his useful comments. Finally, my special thanks to Dr. Malik Ranasinghe, Ahmed Abdel-Aziz, Marina Dimitrijevic and Mohammed Hassanain for their continued support and encouragement. To my parents for their support and encouragement in all my endeavours Chapter 1 Introduction 1.1 Background for the Research Construction can be thought of as the modification of the natural world in order to make it more suitable for a variety of human activities. In addition to creating structures to provide shelter for humans from the natural elements, construction also enables the smooth functioning of modem society by providing it with its infrastructure needs such as roads, bridges, dams, tunnels, water treatment plants and energy generation facilities. However, during the process of creating infrastructure, major impacts may be made upon the existing natural environment. These include impacts such as the physical disruption of ecological systems caused by the construction of large dams across natural waterways and the impacts of highway construction on the stability of fragile hillsides. While in some cases such impacts can be predicted and suitable mitigation measures taken, in other instances the impacts can either be totally unexpected or be of an unexpected magnitude. Such unexpected impacts can lead to severe ecological damage and environmental pollution. From a project proponent's or contractor's point of view, such occurrences mean an increase in project duration and cost due to necessary remediation and clean up operations as well as due to subsequent litigation and compensation. The mitigation measures used to control environmental impacts will also play a major role in determining the way a project is executed. Mitigation measures will influence the selection of the methods of construction and operation as well as of waste disposal. 1 Statutory and regulatory controls brought about by various levels of government in relation to environmental issues will also affect large projects in many ways. Such regulations include those governing environmental assessment procedures, and those governing the control of air, noise and water pollution. As public awareness about the environment increases, these regulations will be refined and developed further, making them wider in scope and generally more stringent. Actions taken by non-governmental organizations and public interest groups with regard to environmental issues can also have an adverse effect on large infrastructure projects. Public agitation over the environmental consequences of projects has contributed to the abandonment of several large projects throughout the world while major modifications have had to be made in a significant number of other projects in order to accommodate public concerns about the environment. Public pressure resulted in the cancellation of the Nagymaros dam in Hungary in 1989 (Ofori 1992) and contributed to the shelving of the Great Whale hydroelectric project in Northern Quebec in 1994. In the UK, the construction cost of the Channel Tunnel is estimated to have been increased by US$ 1.4 billion due to the modifications carried out to accommodate public concerns over environmental issues (Ofori 1992). Therefore, it can be concluded that issues relating to the environment have a significant effect on project cost and duration. This conclusion is further reinforced by figures from the Northumberland strait bridge project in which the environmental costs were estimated to be in excess of 73 million Canadian dollars out of a total capital cost of approximately 800 million (Public Works & Government Services Canada and Strait Crossing Development Inc. 1996). 2 The success of projects is often measured in terms of the three-fold criterion of meeting cost, schedule and performance targets (Williams 1995). The performance criterion relates to the degree of success of the project in meeting its technical and social objectives. An economic evaluation of project success or feasibility is mainly concerned with the first two of these criteria, i.e. the cost and the duration. A model that can be used to predict or evaluate the economic success of a large project of necessity must take into account the cost components and revenue streams of the project as well as the time at which they occur in the project life cycle. Due to the ability of environmental issues to affect the cost and duration of various project elements, such issues will have a direct bearing on the economic success of a project. In some cases, environmental issues might also affect the revenue streams of a project, as described in Chapter 2. In general, the environmental impacts of large infrastructure projects are assessed by carrying out Environmental Impact Studies. Various legislative acts which governs this process, exist in most countries, states, and provinces throughout the world. Environmental Impact Studies contribute to identifying the impacts of a project upon the existing environment as well as in developing suitable mitigation methods that will minimize or eliminate harmful and undesirable impacts. However, due to the ability of environmental issues to significantly affect the economic performance of the project, it is also important to quantify the cost and duration associated with these impacts and the mitigation measures that will be adopted in order to control environmental impacts. Similarly, it is also important to assess the cost and duration associated with other environmental issues relating to large construction projects. 3 In most projects, the information available during the feasibility stages is not sufficient to arrive at deterministic cost and time parameters for various project components, including those related to environmental issues. This brings about a requirement for carrying out the assessment in terms of uncertain cost and time parameters. An early assessment of the uncertainties and risks of the environmental issues would give project management personnel significant leeway to make changes in the scope of the project before large sums of money are irrevocably committed. Clearly, there is a need for a framework that project stakeholders can use in identifying the environmental issues related to the economic viability of a large construction project, and in assessing the cost and time uncertainty surrounding such effects and their effect on the project economic model. 1.2 The Research Objectives The objectives of the research described in this thesis are to: 1. Identify and characterize the natural environmental issues associated with large infrastructure projects that will have an impact on the projects' economic performance. These issues associated with all stages of the project life cycle are considered with the aim of acquiring a better understanding of the interaction between large projects and the environment. 2. Develop a framework that will enable the identification and quantification of the economic uncertainty associated with environmental issues in a large infrastructure project. Time and cost risk associated with the issues are addressed in order to integrate their treatment into the overall project decision making process. 4 It is envisaged that the framework will enable project proponents and contractors to make a conscious effort to treat time and cost risk associated with environmental issues in large infrastructure projects. The ability to identify and address some of the uncertainty associated with environmental issues would improve the quality of management decisions regarding project appraisal. The identification and quantification of environmental risks during the early stages of the project would also provide several other benefits that go beyond the project appraisal stage. They include: • The provision of feedback into the design and planning process in terms of selecting suitable construction, operation and demolition methods; • Provision of a starting point for the allocation of the environmental risks to the project parties who are in the best position to deal with the risk; and, • Assistance in the development of mitigation measures in the form of response selection and contingency planning. 1.3 Structure of the Thesis Discussed in Chapter 2 is the interaction between the natural environment and a large infrastructure project from a project economics perspective. The characteristics of large infrastructure projects and the natural environment are analyzed with the aim of acquiring an understanding of their complex interaction and its ability to influence the project's economics. Factors that further influence and shape this relationship are also addressed in this chapter. The chapter concludes by characterizing the various environmental issues that are important with respect to the project economic model. 5 An environmental risk analysis framework that provides a formal, logical and systematic tool aimed at assisting project stakeholders in identifying and analyzing environmental risks is presented in Chapter 3. The framework is intended to be used during the project appraisal stage when definitive information regarding the project and the natural environment might not be readily available. Chapter 4 introduces a risk classification scheme for the categorization of environmental risks. It then proceeds to outline a practical methodology of quantifying Type A environmental risks. A Type A risk is the cost or duration uncertainty associated with a work package of certain occurrence, which is not influenced by factors external to the project. The methodology utilizes the commonly used Monte Carlo simulation method as its basis and introduces means of overcoming the shortcomings of simulation as used in current practice. The quantification process for Type B risks is described in Chapter 5. A Type B risk is the cost or duration uncertainty associated with a work package, the occurrence and scope of which is affected by external factors. The use of influence diagrams to represent risk events external to the project is central to the quantification methodology described. A novel approach towards representing and quantifying the effect of multiple external uncertainties is also presented as part of the quantification process. An example that shows the application of the quantification method is presented at the end of the chapter. Chapter 6 describes the process used to obtain the information necessary to carry out effective analysis of environmental risks. The information elicitation process described in this chapter makes use of strategies such as decomposition and the use of directed questions based on argument types in order to ensure the overall success of 6 the encoding process. A detailed example that shows the application of the risk analysis framework is also presented. Chapter 7 presents the conclusions of the research as well as the recommendations for future work. Appendix A of the thesis presents a preliminary risk register intended for use in large hydropower projects. It contains a list of probable work packages of a generic hydropower project, that arise or are affected by environmental issues, and are of uncertain cost and duration. The variable transformation process that forms the basis for the method of obtaining rank correlation coefficients between variables is described in Appendix B. 7 Chapter 2 Large Infrastructure Projects and the Environment 2.1 Introduction The relationship between large infrastructure projects and the environment is a complex one. Rarely can it be said that an engineering project does not have a significant impact on the existing natural environment. While there are instances in which large projects result in beneficial impacts on the environment, in most cases the impacts are of an adverse nature. Environmental protection measures are therefore taken in order to prevent or reduce such adverse impacts. These measures may be (Finch 1992, Gifford 1997, Ofori 1992): (a) Initiated voluntarily by the project proponents; (b) Adopted in order to satisfy statutory requirements; or (c) Adopted in response to public pressure. Environmental protection measures adopted due to any one of (a), (b) or (c), or due to any of their combinations may be encountered in a single infrastructure project as a result of the large diversity of impacts associated with such projects. Steps to protect the environment can be taken during each stage of the project life cycle. Such measures include: 1. Environmental Impact Assessment (EIA) studies; 2. Environmental monitoring programs; 3. Changes in design to incorporate environmental friendly measures; 8 4. Use of environmentally friendly construction materials; 5. Mitigatory and safety measures adopted during construction and operation; and 6. Compensatory measures taken to offset the losses caused by the project. These measures and the consequences that arise due to not taking them into consideration as early as possible in the project life cycle will require that time and other resources be expended, thus having an impact on the project's economic performance. Therefore, the relationship between a project and the natural environment in which it is situated can be viewed as one in which the environment and the project are mutually affected. An understanding of the factors that affect this relationship should provide a better understanding of the potential of environmental issues to affect project performance as expressed in an economic model. The relationship between the natural environment and a large irifrastructure project is dependent upon several factors as shown in figure 2.1. The ecological, air, water and land resources that make up the natural environment will be affected in differing proportions and ways during the individual stages of the project. The extent of these effects is a function of the sensitivity of the environment and the scope of the project. This relationship would be further modified by the perceptions and the involvement of the public, and by relevant government regulations. This chapter provides a description of the factors that affect the project-environment relationship with the aim of contributing towards the understanding of 9 Fig. 2.1. Factors Affecting the Relationship between a Project and the Natural Environment the complex interaction between projects and the environment and its relevance to the economic risks of the project. The changing nature of environmental issues that confront the project during the different phases of the project life cycle and their impact on the cost and duration components of the project are discussed in order to achieve this objective. 2.2 The Natural Environment and its Sensitivity The natural environment consists of physical and biological components that are interrelated and affect the growth and development of living organisms (Sadar 1996). For the purpose of this research the natural environment is defined as the ecological, air, water and land resources which are, or have the potential to be affected by the direct or indirect actions of the engineering project under consideration. 10 The sensitivity of the environment, in which the infrastructure project is located or has the potential to affect, plays an important role in the relationship between the project and the natural environment. Highly sensitive areas are most likely to produce adverse reactions to a large infrastructure project. From an ecological viewpoint, sensitive areas can be defined as natural sites of a unique nature or sites that are of particular importance because of their biological productivity and diversity, and which can be expected to suffer serious loss due to the activities of the project (Allison 1981). Similarly, important air, land, and water resources which can be expected to diminish significantly in quality, productivity, or value due to the activities of the project, can also be considered as being sensitive. The importance of such resources can be measured in terms of their current usage and in terms of their potential for future use. Several natural areas that can be considered as sensitive environments are described below (Environment Canada 1999a, 1999b, Hummel 1995). They include: 1. Sites with steep slopes or unstable soils where construction activities would lead to a high rate of soil erosion and increase the possibility of landslides. Removal of the natural vegetation from slopes would increase soil erosion and subsequently cause an increase in the turbidity levels of the streams and rivers, for which the area serves as a watershed. Increased loading due to the construction of new structures or being subjected to increased infiltration due to the alteration of the natural drainage regime can cause instability of slopes leading to landslides. 2. Sites that support concentrations of rare or endangered plant and animal species. Rare floral and faunal species which are at risk of becoming extinct are normally classified under three categories. Species which face the threat of imminent 11 extirpation or extinction, are classified as Endangered Species. Several other species also face the likelihood of becoming endangered if factors such as diminishing habitats and fatalities due to collisions with vehicles are not addressed. Such species are classified as Threatened Species. The third category of Vulnerable Species takes into account species that display high sensitivity to activities such as construction. Areas that serve as a habitat for these types of species or are in close proximity to such habitats are considered as ecologically sensitive areas. The importance of such areas would increase significantly if the area were the sole habitat for the species. 3. Sites within and in close proximity to designated wilderness areas, national parks, national natural landmarks, wild and scenic rivers, wildlife refuges, and sanctuaries. 4. Land units or land components which include unusual, unique, or interesting geological or geomorphological formations or features. These include land areas which are considered to be visually appealing and of significant aesthetic value. 5. Unusual or special areas of animal concentration. These include nesting areas for birds, breeding sites, denning areas, wintering grounds and hibernation areas. 6. Unusual plant communities. These include old growth forests and wetlands. Areas which are classified as wetlands include the freshwater edges of lakes and rivers, inland marshes, swamps, sloughs and peatlands, and the marine waters of estuaries and the tidal ocean shoreline. These are critical habitats for many species of flora and fauna. 7. Migration routes and resting areas for migratory species such as fish, butterflies, birds, and mammals including marine species. 12 8. Watersheds of high importance supporting streams rich in biodiversity or streams which harbor threatened species. 9. Watersheds feeding streams and rivers which play an important role in economic activity such as hydroelectricity generation, commercial fisheries, or act as a source of water for domestic or industrial consumption. 10. Areas, which act as natural buffers for important natural resources such as streams, reservoirs and wetlands. 11. Special sources of water such as sole-source aquifers, wellhead protection areas, and other water sources that are vital in a region. 2.3 Scope of Engineering Projects The scope of a project helps to determine the extent of its interaction with the natural environment. The parameters that can be used to characterize the scope of a project differ from one type of project to another. While one project type might be characterized using storage capacity, another type of project could be characterized using the energy usage rate. Parameters that are used to describe the scope are: • Physical size (length, areal extent, height) • Capacity (production, carrying, storage) • Industrial Classifications (e.g. Standard Industrial Classification as published by the Minister of Supply and Services, Canada) • Energy usage • Resource consumption • Use of hazardous substances or substances harmful to the environment 13 • Visual impact • Waste discharge rates • Use of new or untested technology • Number of employees In addition to determining the extent of its physical interaction with the environment, the scope of a project will also affect the application of environmental regulations to a project. Environmental regulations specify the types of projects to which various regulations and clauses apply. They will also specify limiting or threshold values of its scope attributes, beyond which particular regulations will apply. In the case of new projects, limiting values are specified in terms of attribute values, while in the case of a modification to an existing facility they might be specified in terms of attribute values or as a percentage of the original attribute value. 2.4 Public Perceptions In the context of the relationship between civil engineering projects and the environment, the term "public" refers to concerned individuals, community organizations, first nations people and special interest groups such as environmental and outdoor recreation groups (British Columbia 1995). Public pressure in relation to environmental issues can have many consequences for civil engineering projects. As mentioned in Chapter 1, protests over their effects have led to the modification of many large infrastructure projects, while in some cases, projects at various stages of completion have even been abandoned altogether due to such reasons. 14 The perception of the public towards the environment in which a proposed project is to be located in, can be expressed in terms of Valued Ecosystem Components. A Valued Ecosystem Component can be defined as an environmental attribute or component which can be identified as having legal, scientific, cultural, economic, commercial or aesthetic value (Sadar 1996). The presence of Valued Ecosystem Components within the spatial area that will be affected by the project activities can be considered as a catalyst for public concern regarding the project. Similarly, negative public sentiment can be created by the lack of effort on the part of the project proponent to adequately notify the public of the project and its actions, to provide sufficient distribution of information, enter into meaningful discussions with the public, and to adequately respond to genuine public concerns. Ineffective responses to negative public perceptions can lead to: 1. Law suits; 2. Demonstrations and blockades; 3. Vandalism; and 4. Boycott of the products or services of the project. These actions in turn will have a negative impact on the economics of the project through delays and additional expenditures. Blockades during the operational phase, and the boycott of the products or services of the project are also likely to affect the revenue stream of the project. 15 2.5 Government Regulations Four types of regulations that are relevant to the relationship between large infrastructure projects and the environment can be identified. These are: • Environmental impact assessment regulations; • Regulations that govern the usage of natural resources; • Regulations that govern the disposal of waste; and • Regulations promulgated to protect natural spaces and life forms. 2.5.1 Environmental Impact Assessment The Environmental Impact Assessment (EIA) process, in essence involves (Glasson, Therivel and Chadwick 1994, Ofori 1992, UNEP 1980): (a) Screening of projects in order to determine the applicability of the EIA process to each project. In general, applicability depends upon the scope characteristics of the projects. Projects that are greater in scope than stipulated guideline values are subject to ELAs. The stipulated guidelines are determined by the regulations operating in the jurisdiction at that time. In addition, projects that are likely to generate a significant amount of public concern may be subject to an EIA even in instances where the scope is less than the stipulated guideline value. (b) Preliminary identification of the key, significant issues that need to be addressed. This aim of this step is to determine the issues that need to be addressed more rigorously than others. (c) Consideration of alternatives in order to ensure that all feasible approaches, including alternative project locations, scales, processes, layouts, operating 16 conditions as well as the "No action" option have been considered in the decision making process. (d) Description of the proposed project, its alternatives as well as the existing environment. This includes a clarification of the purpose and rationale of the project and its components. (e) Description of the baseline environment based on an understanding of the characteristics of the natural environment in which the project is located. The present state and the future state of the environment in the absence of the project is considered in this description. (f) Identification of all significant impacts of the proposed development on the existing environment, and estimating the nature and magnitude of such impacts. (g) Identification of mitigatory measures including measures to avoid, reduce, remedy, or compensate any significant adverse impacts. (h) Making recommendations for acceptance or rejection of the project and/or its alternatives and the remedial and mitigatory actions that should be taken. (i) Making recommendations for carrying out monitoring and inspection procedures during the project execution. (j) Public consultation in order to ensure that the views of the public are adequately taken into account during the decision making process, (k) Reviewing of the EIA statement by the project approving authorities. (1) Decision making on the project by the relevant authorities, (m) Post decision monitoring in order to record the outcomes associated with the development. 17 (n) Auditing. This involves comparing the actual out comes with the predicted outcomes in order to assess the quality of the predictions and the effectiveness of the mitigation measures. In Canada, EIA regulations have been enacted by both the federal and provincial governments. The federal environmental assessment and review process establishes an overall framework and implements EIA procedures for federal-level projects (Glasson, Therivel and Chadwick 1994). The provinces have their own EIA processes which differ widely among the provinces (Smith 1991). The Canadian Environment Assessment Act (Canada 1992) requires that an environmental assessment of a project be carried out in instances where: a) A federal authority is the proponent of the project; b) A federal authority makes or authorizes payments, or provides a guarantee for a loan or any other form of financial assistance to the proponent for the purpose of enabling the project to be carried out; c) A federal authority has the administration of federal lands and sells, leases or otherwise disposes for the purpose of enabling the project to be carried out; and d) The Governor in Council or a federal authority issues a permit or license, or grants an approval in order for the project to be carried out. The environmental assessment process in the province of British Columbia is governed by the Environmental Assessment Act (British Columbia 1996c). The Act requires that the proponent of a reviewable project obtain a "Project Approval 18 Certificate" before carrying on construction, operation, modification, or dismantling of the project. The Act also provides for participation in the assessment of reviewable projects by the public, project proponents, first nations people, municipalities and regional districts, the government and its agencies, the federal government, and the governments of neighboring jurisdictions. 2.5.2 Environmental Protection Regulations Environmental Impact Assessment regulations relate to a specific project or specific program. However, the other types of regulations, i.e. regulations that govern the usage of natural resources, regulations that govern the disposal of waste, and regulations promulgated to protect natural spaces and life forms do not in general refer to specific projects but apply to all aspects of human activity. However, the implications of such environmental regulations are also felt by individual projects. Such legislation that can collectively be referred to as environmental protection regulations, exists in the province of British Columbia as well as in Canadian federal law. Environmental protection legislation enacted in British Columbia includes: a) Waste Management Act The Waste Management Act (British Columbia 1996i) governs the introduction of waste to the environment. This legislation empowers officials to issue pollution abatement orders and to request the preparation of contingency plans when an organization is in possession of a substance that has pollution potential. The statute also deals with the confinement and disposal of special waste and governs the use of ozone depleting substances 19 within the province. A permit is required under this legislation in order for an organization to introduce waste into the environment. b) Sewage Disposal Regulations of the Health Act These regulations of the Health Act (British Columbia 1996f) ensure that sewage does not reach the surface of land or discharge into a surface body of fresh water. It sets outs guidelines for the location of sewage disposal systems in environmental control zones. These regulations require that a permit be obtained prior to constructing, installing, altering or repairing a sewage disposal system. c) Water Act The Water Act (British Columbia 1996j) requires that licenses be obtained for the acquisition of water rights for use in irrigation, mines and for use in power development. An organization making a change in a stream is required to ensure that no substance, sediment, debris or material is allowed or permitted to enter or leach or seep into the stream. Other requirements include the maintenance of minimum flow levels, protection of fish and the restoration of the site subsequent to construction activities. d) Forest Practices Code of British Columbia Act This Act comes into effect with respect to large infrastructure projects when a project stakeholder wishes to use or manage a portion of crown land in a provincial forest or wilderness area (British Columbia 1996e). Such uses include the construction of temporary roadways through crown land, or the construction of a temporary airstrip or helipad. A special use permit is required in such instances. Forest road regulations of this Act set out 20 guidelines for road layout, drainage design, road surfacing and deactivation of roads through forest lands. The Act specifies that a permit be obtained to construct works on any part of a forest service road right of way for any purpose other than the passage of public vehicular or pedestrian traffic. e) Petroleum and Natural Gas Act The Act requires that air quality monitoring stations be installed and maintained as specified by the British Columbia Ministry of Environment, Land and Parks in the event of operations being carried out at a well or a production facility (British Columbia 1996h). It also requires that on completion or final abandonment of a well, test hole or facilities, that the area be cleared of all refuse material and the site left in a condition near to the original condition. The submission of an Application for Certificate of Restoration is part of this requirement. f) Forest Act Forest fire prevention regulations of this Act requires that fire fighting equipment be maintained while carrying out industrial operations within 1 km of any provincial forest (British Columbia 1996d). It also prescribes the regulation of activities of the project according to the fire "danger class" applicable to the locality in which the operations are being carried out. This Act also governs the removal of timber from crown lands. The timber felled during the clearing activities of a project are subject to Stumpage fees, including royalty under the provisions of the Forest Act. 21 Ecological Reserves Act This Act empowers the provincial government to designate land areas as ecological reserves (British Columbia 1996a). Lands that may be designated as reserves include areas that are representative examples of the natural ecosystems of British Columbia and areas that serve as habitat to rare or endangered native plants and animals. The Ecological Reserves Act as well as other acts such as the National Parks Act (Canada 1985d) do not have an immediate and direct impact on large infrastructure projects. However, designation of areas as ecological reserves would elevate lands to the status of a Valued Ecosystem Component in the eyes of the public. Thus, projects located in close proximity to such lands are likely to come under heavy public scrutiny during the environmental assessment stage. Environment Management Act This act enables the provincial government to recover any costs undertaken by the government in taking steps for the purpose of preventing, lessening or controlling the hazard presented by an environmental emergency such as a major oil spill from the person or organization whose acts caused the environmental emergency (British Columbia 1996b). The Act also allows the responsible minister to issue environmental protection orders restricting, modifying or prohibiting any work or undertaking that is considered to have a major detrimental impact on the environment. Wildlife Act Land may be acquired by the government for the purpose of conservation or management of wildlife under this Act (British Columbia 1996k). Such lands 22 would be regarded as Valued Ecosystem Components during the appraisal of a project located in the vicinity of such lands, j) Land Act This Act establishes procedures for the disposal of crown lands (British Columbia 1996g). Applicants wishing to develop land for the purpose of an infrastructure project are required to submit a plan detailing the environmental impact minimization steps that will be taken. Federal legislation that relate to environmental protection include: a) Navigable Waters Act This legislation regulates the dumping and deposition of material in navigable waters (Canada 1985e). b) Canadian Environmental Protection Act This Act is relevant to the use, manufacture and storage of toxic materials in projects as well as to air and marine pollution issues (Canada 1999). This Act empowers the minister to specify codes of practice respecting pollution prevention. The minister also has the right to request users of toxic materials to develop and implement a pollution prevention plan. The Canadian Environmental Protection Act also authorizes the federal government to bring forth regulations with respect to the quantity or concentration of a substance that may be released into the environment and the manner and conditions under which the substance may be released into the environment. The Act also provides for greater public participation in implementing the Act by providing for the public the right to bring civil suits in cases of significant 23 damage to the environment if the government fails to enforce the Act. Permits required under this Act include: • Permits for the import, export or convey in transit of a hazardous waste; • Permits to incinerate a substance on board a ship, a platform or another structure in an area of the sea; and • Permits to dispose of a substance in specific areas of the sea. National Parks Act This Act empowers the Governor in Council to establish new national parks and to add land to existing parks (Canada 1985d). Similar to ecological reserves established under provincial legislation, the establishment of national parks in close proximity to the project is likely to affect the perceptions of the public towards the project. The Oceans Act and the Canada Wildlife Act are also described below as having an impact on infrastructure projects based on the same argument. Oceans Act The Governor in Council may make regulations designating marine protected areas under the powers vested by the Oceans Act (Canada 1996). The Act also allows the zoning of protected areas and the prohibition of classes of activities within the protected areas. 24 e) Canada Wildlife Act The Wildlife Act (Canada 1985c) allows the government to assign public lands for wildlife research, conservation and interpretation. It also enables the Governor in Council to establish protected marine areas. f) Canada Water Act The Water Act (Canada 1985b) empowers the government to formulate comprehensive water resource management plans in cooperation with provincial governments for any waters deemed to be of significant national interest. In instances where a water body is subject to a management plan, a permit is required under this legislation to carry out disposal of waste into the waters. The permit sets out the admissible quantities and concentrations of the waste. g) Arctic Waters Pollution Prevention Act This Act governs the disposal of waste in the mainland and waters of the Canadian Arctic (Canada 1985a). The Governor in Council may make regulations prescribing the type and quantity of waste that may be discharged. The Governor may also require persons who carry out exploratory work or construction to provide evidence of financial responsibility in the form of insurance or an indemnity allowing claims based on pollution of the Arctic to be recovered directly from such instruments. 25 2.6 The Interaction between Projects and the Environment The nature of environmental issues that confront an infrastructure project will change throughout the various stages of the project life cycle. The life cycle defines the start and end points of a project. The project life cycle is normally divided into several phases in order to provide better management control (Project Management Institute 1996) and to signify the collection of activities that contribute towards the completion of a particular project deliverable such as the Final Design or the Completed Structure. The execution of these phases can be carried out sequentially as was done traditionally, or several phases may be overlapped in order to reach a desired phase early, as is the case in fast-tracked projects (Rahemtulla 1989). Several diverse approaches have been adopted by various authors towards the exercise of breaking down the project life cycle into phases. Morris (1994) has adopted a four-stage breakdown that considers a project to be made up of a Feasibility stage, a Planning & Design stage, a Production stage, and a Turnover & Startup stage. Chapman and Ward (1997) have proposed an eight stage breakdown which goes some way towards highlighting sources of process risks. This is an extension of a simpler breakdown that divides the project life cycle into the four phases of Concept, Planning, Execution and Termination. Chapman and Ward have considered Conceive, Design, Plan, Allocate, Execute, Deliver, Review and Support as the distinct stages in the life of a project. However, these approaches do not provide a convenient breakdown structure for the purpose of analyzing environmental risks in civil engineering projects. Some approaches do not adequately cover the time length during which environmental issues affect the project while other life cycle breakdowns suggested in the literature take into account phases such as a project conceive phase that are unimportant with 26 respect to the effect of the environment on project economic variables. Therefore, a five-stage breakdown of the project life cycle is proposed herein for the purpose of analyzing environmental risks. The five phases are: Planning & Design; Construction; Operation; Modification; and Closure phases. This particular breakdown has been selected for two main reasons: 1. The breakdown structure effectively covers the time span in which the project economics will be affected by environmental issues. As mentioned previously, phases such as the conceive stage suggested in some life cycle breakdowns are not relevant to the treatment of environmental risks. Other breakdowns do not consider modification of the original project as well as its termination. Environmental issues such as obtaining environmental approvals and environmental restoration of the site cannot be addressed without considering modification or termination stages. Therefore, the proposed project life cycle breakdown can be considered to effectively encompass the environmentally sensitive activities of the project by including all the stages relevant to environmental issues and by neglecting non-relevant stages. 2. It provides a simple, intuitive breakdown that effectively separates the various types of environmental issues affecting the project economic model. The nature of the environmental issues confronting a project changes throughout the project life cycle. It is important to make a distinction between these different types of issues in order to analyze them effectively. The five phases of the proposed life cycle breakdown are described below. 27 Planning and Design Phase The Planning & Design phase includes the development of preliminary designs and plans, carrying out feasibility studies, appraising the environmental consequences of the project, and obtaining project approvals. It also involves the detailed designing of structures and processes of the project in accordance with building codes, safety regulations and environmental regulations, and the development and awarding of contracts. Construction Phase The construction phase includes the construction of the envisaged structures that make up the project, as per the final design. It also includes the erection of temporary structures that are created to facilitate the overall construction process, as well as the implementation of mitigation measures to reduce or eliminate damage to the environment. Operational Phase The operation of the facility in order to fulfil its intended use occurs during this phase. Minor repairs and alterations, as well as regular maintenance also takes place during this phase. Environmental monitoring will be carried out during this phase in many projects to determine the actual impact of the project upon the environment. Modification The modification of the structures or the processes of the project, which can be construed as a major alteration to the initial operating state of the project make up the 28 modification stage. This phase, which includes phased development, would not necessarily exist in all large infrastructure projects. Environmental issues involved during this stage include appraising the environmental consequences of the modification and the implementation of mitigatory measures during the modification process. C l o s u r e Termination involves the decommissioning of the facility and disposal of the project components. It could also involve environmental clean up and restoration activities. However, this phase would not exist in projects such as highways, which can be considered to have an indefinite lifetime. Several types of environmental issues that arise during the different project stages, and have an uncertain impact on the economic performance of the project can be identified. These can be categorized into three types as shown in figure 2.2 based on the nature of their occurrence. The categorization of environmental issues serves as the starting point for the categorization of environmental risks in carrying out risk analysis as set out in Chapter 3. 2.6.1 Pe rm i t s a n d E I A Procedures Requirements for permits and licenses, and the need to carry out EIA studies arise as a consequence of relevant government regulations. Common permit requirements pertaining to environmental issues have been identified previously in section 2.5.2. In addition to procuring environmental permits, environmental 29 EIA Procedures and Requirements for Permits Changes in Environmental Regulations and By-laws Law Suits based on Environmental Issues Projects' Economic Performance Environmental Damage and Pollution -Mitigation Measures / Accidents Fig. 2.2. Types of Environmental Issues having an Uncertain Impact on the Economic Performance of an Infrastructure Project impact assessment studies need to be carried out during the initial stages of the project as described in section 2.5.1. Requirements for permits and EIA regulations are issues that are mostly encountered during the design and planning stage of a project. However, some permits such as those relating to waste disposal are required to be renewed periodically. EIA and other regulations also come into effect during the modification and closure phases. Modifications that exceed limiting values, normally specified in legislation as a percentage of original scope parameter values are required to undergo an EIA assessment. 30 In most cases the need to obtain certain permits and licenses and the need to carry out an EIA study can be identified by consulting government officials and through the review of government statutes. The cost components associated with these processes include: 1. Licensing fees on a one-time or periodic basis. In most cases the fees are fixed amounts stipulated in regulations. However, licenses that have to be renewed periodically are subject to uncertainty due to potential increases in the licensing fee. 2. Payments made to experts in order to carry out appraisal of project components. Expert appraisal such as those made regarding the environmental impacts of the project are necessary for the preparation of reports required as part of the licensing procedure. A certain amount of uncertainty would be prevalent regarding the magnitude of this cost component during the project feasibility stage. The same type of intrinsic uncertainty afflicts the cost components described below as well. 3. Costs of carrying out testing procedures. Such tests include field tests, laboratory tests and computer simulations carried out in order to predict future impacts of the project on the environment as well as tests carried out during the construction, operation, modification and closure phases in order to determine actual impacts upon the environment. Results of such tests are required as part of the EIA process. Test results will also be required for licenses pertaining to waste discharges. 4. Costs of carrying out public consultation as part of the EIA process. Public consultation forms an integral part of the EIA process and includes providing notification and information to the public about the project as well as ensuring 31 ongoing consultations. Costs are incurred by the proponent in placing advertisements in newspapers and other media, and in organizing public meetings such as "Open houses" regarding the environmental issues of the project. 5. Costs of documentation. This cost component includes office overheads and miscellaneous costs incurred during the preparation of permit applications and EIA reports. The duration of a permit approval process and an environmental assessment process is made up of: 1. An identification stage in which the requirement is reviewed by the proponent. The allocation of responsibilities to members of the project team, making budgetary allocations and identifying experts are tasks that are carried out during this stage. 2. The preparation of the application. This stage could simply involve filling out a standard form or be a complex process of assessing environmental impacts, obtaining expert opinions, public consultation and preparing supporting documents. The scope of this stage would depend on the type of permit under consideration. 3. Review of the application. This stage involves the reviewing of the application by authorities of the licensing institution. In the case of the EIA process a public hearing may be held during this stage as well. Even though time frames for EIA procedures and permit approval processes are set out in some government regulations, duration overruns are considered to be 32 common in practice (Wood 1995). Therefore, uncertainty can be considered to be prevalent with regard to the duration of permit approval and EIA processes similar to their cost components. 2.6.2 Changes in Regulations and Law Suits Changes in environmental regulations can bring about a requirement for a change in the design of the project and its processes. The need to accommodate such unanticipated changes can cause unprecedented cost and schedule overruns particularly if they come into effect when construction is partially complete or when operation of the facility has commenced. The occurrence of such a change in environmental regulations may itself be an uncertain event. Similarly, lawsuits brought on by negative public perceptions can also have adverse effects on the economic performance of the project. Lawsuits may be filed by the public seeking redress from perceived irregularities in approval processes, claiming damages for the harm caused to the environment due to the normal operations of the project, as well seeking damages for environmental pollution caused by events such as accidents. The occurrence of such lawsuits themselves would be considered an uncertain event during the project appraisal stage. Regulatory changes and lawsuits may take place during any stage of the project. However, as mentioned previously such issues are likely to have a more significant impact on the project if they come into effect when the project has already entered the construction stage. Regulatory changes occurring during the initial stages of the project can be accommodated by making necessary changes in the design. Changes occurring at latter stages are less easily accommodated when substantial amounts of 33 funds have been expended on creating structures and operational procedures based on the previous set of regulations. Regulatory changes and lawsuits affect project economics due to the cost and duration associated with: • Obtaining additional permits that may be required under new or changed regulations; • Carrying out re-design of project structures and operational procedures to suit the new regulations or in order to comply with a court ruling; • Making physical changes in existing project structures and operations to suit the new regulations or in order to comply with a court ruling; • Implementing environmental protection measures required as part of the changed regulations during the construction, operation, modification or closure phases; • Legal defense of the project organization in response to lawsuits based on environmental issues; • Injunctions issued by courts in response to lawsuits; and • Damages paid as part of the settlement of lawsuits. 2.6.3 Environmental Damage and Pollution In general, an effort is made to eliminate damage to the environment by utilizing appropriate mitigatory measures in most large infrastructure projects. However, in some instances environmental damage and pollution could be unavoidable adverse impacts of the project. Mitigatory measures designed to minimize the harmful effects on the environment are adopted in such cases. A monitoring program that evaluates 34 the efficiency of the mitigatory measures as well as the actual impact of the project on the environment is also required in such instances. Environmental damage and pollution can also result from unanticipated accidents that occur during the lifetime of the project. Clean up and restoration procedures need to be implemented in the event of environmental accidents. Environmental pollution brought on by the normal .operations of the project and environmental accidents will have an impact on the project economics due to the cost and duration associated with: • The implementation of mitigatory measures designed to eliminate, reduce, or control environmental pollution caused by the project; • The implementation of monitoring plans to determine the effectiveness of the mitigatory measures and to determine the actual changes to the environment brought on by the project; • Compensatory measures undertaken to offset the losses caused to the environment; • Measures taken to control and combat environmental accidents; • Clean up and restoration operations subsequent to environmental accidents. The relevance of the individual issues described above would vary during the different phases of the project as shown in figure 2.3. For example, issues such as lawsuits are relevant during all stages of the project, while the implementation of mitigatory measures become an issue only after the commencement of the construction phase. 35 o O 1 < 111 t o c u c CO -£= O >i t o O t o *—' ^—1 _ t o E t o rj i U) CD CO c u CL _l t o c u J t o c o c u c o o c c u E o > c LU c : c u -g u o < i—H CN 36 The significance of the environmental issues described above would also be a function of the project delivery mode. As mentioned earlier in this section, some of the project phases may be overlapped (see figure 2.3) during project execution with aim of achieving a desired phase early. However, a major risk of such a fast-track strategy is that despite the expenditure of greater resources, the reduced duration will not be achieved (Russell and Ranasinghe 1991). Environmental issues can contribute significantly towards the realization of such a risk. Several instances where environmental issues can adversely affect fast-track projects are described below, (a) Requirements for some environmental permits may not be identified prior to the start of construction in fast-track projects that involving an overlap between the design and construction phases. The probability of such an occurrence is heightened by the fact that changes from the preliminary design is considered to be an inevitability in such projects (Rahemtulla 1989). Therefore, a possibility exists that the requirement for some environmental permits will be identified only when the detailed design is being carried out. However, the duration of the permit approval processes as well as the uncertainty associated with the duration could be such that it creates an unanticipated interruption of the construction phase. An example would be an instance where the construction phase of a fast-track project is started with the design of project components being in advance of their construction by 3 weeks. However, should an environmental permit that governs the start of construction of a key project component be identified only when the detailed design is being carried out, it is likely that the fast-track schedule would be delayed should the permit approval process takes more than 3 weeks to complete. 37 (b) Environmental permits that were not obtained prior to the start of construction may place unanticipated restrictions that affect the planned schedule of construction. An example would be a permit that restricts the construction of a key project component to the daylight hours. This would significantly affect the fast-track duration in an instance where a 24-hour construction schedule was originally planned. (c) Environmental protection measures such as impact mitigation measures and waste disposal methods might not be identified prior to the detailed design of project components. Therefore, the additional duration required for the implementation of such measures would also affect the planned fast-track schedule of construction. 2.7 Summary Large infrastructure projects have the potential to cause significant impacts of an adverse nature on the natural environment. Similarly, issues related to the natural environment can also have adverse effects on the project cost and schedule creating a negative impact on the projects' economic performance. Therefore, the relationship between a project and the natural environment in which it is situated can be viewed as one in which the environment and the project are mutually affected. The interaction between the natural environment and a large infrastructure project has been discussed in this chapter from a project economics perspective with the aim of contributing towards understanding this complex interaction and its relevance to the economic risks of a project. The factors that affect the project-environment relationship have been identified as part of this discussion. These are: Scope of the project; Sensitivity of the environment; Government regulations, and Public perceptions. The various types of 38 issues that have an impact on the project's economic performance have also been described in this chapter. Such issues include requirements for environmental permits, changes in regulations, lawsuits, and environmental damage and pollution. It is believed that a greater understanding of the complex relationship between the natural environment and large infrastructure projects, towards which this chapter contributes, would enhance the ability to address the economic risks that arise due to environmental issues. 39 Chapter 3 Environmental Risk Analysis Framework 3.1 Introduction The potential of environmental issues to have an uncertain impact on the project parameters of cost and schedule is a fact that has been acknowledged in many contributions to the literature on project risk management. The recognition given to environmental risks is more evident in recent literature, as public concern over environmental degradation has continued to grow over the years. This chapter presents an analysis of the identity given to environmental risks in various risk categorization schemes adopted in the literature. Five such categorization schemes, which are representative of the many different schemes available are considered in this study. This chapter also reflects upon project risk analysis frameworks suggested in the literature. Such frameworks facilitate the analysis and management of the various economic risks encountered during large projects such as contractual and legal risks, and construction related risks. While such frameworks provide a formalized approach for the analysis of overall project risk, they do not provide'sufficient rigor in aspects such as: (a) the provision of risk quantification methods capable of analyzing different forms of uncertainty, and (b) in providing elicitation methodologies designed to obtain the information necessary to carry out meaningful analysis of environmental risks. The different forms of uncertainty encountered in the analysis of economic risks due to environmental issues includes the uncertainty inherently present in economic variables such as labor productivity, and the uncertainty caused by factors external to 40 the project components being studied. Such factors include public perceptions towards the project, the sensitivity of the environment in which the project is situated, and changes in government regulations. Therefore, a framework devoted solely to the analysis of risks due to environmental issues is presented in the final section of the chapter with the aim of facilitating a more comprehensive treatment of environmental risks as opposed to the general treatment provided by other frameworks. The proposed framework includes methods for effectively quantifying risks due to environmental issues, procedures for obtaining the various types of information required for the analysis process, as well as means of building up a database of probable environmental risks for projects of a similar nature. 3.2 Risk Categorization Schemes The large number of risks associated with engineering projects can be categorized into various groups in accordance with their similarities and dissimilarities. Literature on project and construction management suggest a variety of categorization techniques for this purpose. Most categorization techniques acknowledge the existence of risks relating to environmental issues. However, the definitions provided for Environmental Risks as well as the uncertain events that are included within this category show a large variation when different categorization schemes are compared. This section presents a brief outline of 5 risk categorization schemes found in the literature. The usefulness of the categorization schemes and their treatment of environmental risks are discussed at the end of the section. 41 A classification scheme for project risks, based upon the cause of the risk has been developed by Wideman (1986). Project risks have been divided into the five categories of External Unpredictable Risks, External Predictable Risks, Internal Risks, Technical Risks and Legal Risks (see Table 3.1). Risks relating to environmental issues can be found under several categories in this scheme. Wideman has acknowledged that some environmental risks are not predictable. These include unanticipated government intervention in the project by way of new regulations or changes to existing ones, as well as unexpected side effects upon the environment by the project. Some environmental impacts have also been deemed to be predictable but of an uncertain magnitude. Both these types of project risks are characterized as being due to sources external to the project. The risk of delays in the project schedule due to difficulties in obtaining regulatory approvals is classified as an Internal Risk. Internal Risks are risks associated with actual project components as opposed to risks due to external factors such as changes in regulations. Approvals that would give rise to Internal Risks include project permits required under environmental impact assessment legislation and environmental protection legislation. Project risks associated with lawsuits have been classified as a Legal Risk. Lawsuits associated with environmental issues such as those claiming damages for environmental pollution caused by the project would be included within this category. 42 Table 3.1. Summary of Risk Categories adopted by Wideman External External Internal (non Technical Legal Unpredictable Predictable - technical) Risks Risks Regulatory Market Risk Schedule Changes in Licenses Unanticipated Changes in: Delays due to: Technology government - Availability - Regulatory Patent Rights interventions in: of raw approvals Performance - Supply of raw materials - Labor - Quality Contractual materials - Competition shortages - Productivity Failure - Environmental - Unforeseen - Reliability issues site Outsider Suit - Site location Environmental conditions Risk Specific Impact to Project's Insider Suit Technology Completion Social Impact Cost - In creating Force Failure to Overruns due - In operating Majeure complete: Currency to: - Supporting Changes - Schedule infrastructure by delays others Inflation - Procurement - Due to political strategy unrest Unexpected Side Cash Flow Effects Interruption - Environmental - Social Source: Adapted from Wideman, M.R. 1986. Risk management. Project Management Journal, September, 20-25. Perry and Hayes (1985) have categorized the risks associated with the design, construction and operational phases of a project (see Table 3.2). Environmental Risks have been considered as a separate risk category in this scheme. The risk sources that have been considered under this category include ecological damage, pollution, waste treatment, and public inquiry. Changes to law and legislation have been considered under the separate category of Political Risks in this scheme. 43 Table 3.2. Risk Categorization Scheme adopted by Perry and Hayes Risk Category Uncertain Events Physical Loss or damage by fire, earthquake, flood, accident, landslip Environmental Ecological damage, pollution, water treatment Public inquiry Design New technology, innovative applications, reliability, safety Detail, precision and appropriateness of specifications Design risks arising from surveys, investigations Likelihood of changes Interaction of design with methods of construction Logistics Loss or damage in the transportation of materials and equipment Availability of specialized resources - expertise, designers, Contractors, suppliers, plant, scarce construction skills, materials Access and communications Organizational interfaces Financial Availability of funds, adequacy of insurance Adequate provision of cash flow Losses due to default of contractors, suppliers Exchange rate fluctuations, inflation Taxation Legal Liability for acts of others, direct liabilities Local law, legal differences between home country and home countries of suppliers, contractors and designers Political Political risks in countries of owner and suppliers, contractors -war, revolution, changes in law Construction Feasibility of construction methods, safety Industrial relations Extent of change Climate Quality and availability of management and supervision Operational Fluctuations in market demand for product or service Maintenance needs Fitness for purpose Safety of operation Source: Perry, J.G. and Hayes, R.W. (1985) Risk and its management in construction projects, Proceedings of the Institution of Civil Engineers, U.K., Vol.78. 44 Another risk categorization scheme considers six types of risks encountered during the construction phase of a project (Kangari and Boyer 1987). The six categories that have been identified are Construction related Risks, Physical Aspects, Contractual and Legal Risks, Performance and Management Risks, General Economic Factors, and Political and Public Risks (see Table 3.3). In this scheme, water, air and noise pollution have been considered as risks pertaining to Environmental Protection. The sub category of Environmental Protection falls under the broad classification of Political and Public Risks, along with the subcategories of risks due to Public Disorder, and Government Acts and Regulations. Risks due to changes in regulations and the risks involved in the permit approval process have been included within the subcategory of Government Acts and Regulations. Table 3.3. Risk Factors identified by Kangari and Boyer Risk Category Uncertainty Factors Construction related Risks Uncertainty in labor Equipment related risks Uncertainty in material Late completion risk Defective design work Defective construction Delayed site access Quantity variation Design changes risk Physical Aspects Uncertainty in subsurface Acts of god Contractual and Legal Risks Failure of payment Delayed disputes resolution Coordination failure Uncertainty in change orders Labor disputes risk Insurance coverage risk 45 Table 3.3. - Continued Risk Category Uncertainty Factors Performance and Management Uncertainty in productivity Uncertainty in quality Accidents/safety Mistakes Managerial competence General Economic Factors Uncertainty of inflation National and International Financial uncertainty Political and Public Environmental protection Public disorder Government acts/regulations Source: Adapted from Kangari, R. and Boyer, L.T. (1987) Knowledge based systems and fuzzy sets in risk management, Microcomputers in Civil Engineering. 2, 273-283. Smith and Bohn (1999) have identified eight categories of risk. These broad categories of risk are Natural Risks, Design Risks, Logistics Risks, Financial Risks, Legal and Regulatory Risks, Political Risks, Construction Risks and Environmental Risks. The Environmental Risks identified include ecological damage, pollution and waste treatment. They have been characterized as being risks that are generated outside the project and which are in many cases not controllable by the parties to the project. Environmental Risks have also been considered as having a reasonable certainty of occurrence. Changes in regulation have been categorized under Legal and Regulatory Risks. Risks associated with the permit approval process have also been considered under the same category. Such risks have also been characterized as being risks that are generated outside the project. However, unlike the risks within the category of Environmental Risks, Legal and Regulatory Risks have been characterized as risks of uncertain occurrence. 46 Several purposes of classifying risks have been set out by Al-Bahar and Crandall (1990). The purposes, which have been identified, are: (a) To expand the risk awareness of the project stakeholders; (b) Develop common risk mitigation strategies to counter risks of similar nature. Al-Bahar and Crandall have proposed a risk classification scheme that classifies risks according to their nature and potential consequences. Six risk categories have been identified in this classification as shown in Table 3.4. The category of Political and Environmental Risks take into account changes in laws and regulations as well as pollution and safety rules. Requirements for permits such as project approval certificates granted subsequent to the environmental assessment processes are also classified under the Political and Environmental Risk category. However, this scheme has avoided any direct reference to the uncertain economic consequences associated with pollution such as the schedule delays and costs incurred due to environmental accidents. Table 3.4. Risk Categorization Scheme adopted by Al-Bahar and Crandall Risk Category Typical Risks Acts of God Physical Financial and Economic Flood, earthquake, landslide, fire, wind, lightening Damage to structure, damage to equipment, labor injuries, material and equipment fire or theft Inflation, availability of funds from client, exchange rate fluctuation, financial default of subcontractor, non-convertibility 47 Table 3.4. - Continued Risk Category Typical Risks Political and Environmental Design Construction-related Changes in laws and regulations, war and civil disorder, requirements for permits and their approvals, pollution and safety rules, expropriation, embargoes Incomplete design scope, defective design, errors and omissions, inadequate specifications, different site conditions Weather delays, labor disputes and strikes, labor productivity, different site conditions, defective work, design changes, equipment failures Source: Al-Bahar, J.F. and Crandall K.C. 1990. Systematic risk management approach for construction projects. Journal of Construction Engineering and Management. Vol.116.No.3, September. As identified by Al-Bahar and Crandall (1990) the classification of risks can assist in improving the risk awareness of project stakeholders. A logical classification scheme can assist in the identification of risks by serving as a preliminary checklist of risk sources. Categorization of risks can also simplify the identification process by allowing the analysts to concentrate upon a limited area of risk at any one time, which decreases the possibility of overlooking important risks within that area as opposed to looking at all risk types simultaneously. As identified further by Al-Bahar and Crandall (1990), classification can also assist in the development of some risk mitigation strategies such as the transfer of risks to other project parties. Project components that are subject to a particular risk type may be transferred to a party that is best equipped to deal with such risks. An example would be the transfer of risks associated with labor disputes that occur during construction to the contractor. In this 48 instance, classification would assist in identifying the project components that are subject to risks associated with labor disputes. While the categorization of risks can have several advantages as outlined above, the risk categorization schemes found in literature differ significantly from one another. The lack of consistency extends itself to the identification and categorization of Environmental Risks as well. The review of categorization schemes illustrates that a constant definition for Environmental Risk does not exist in the literature. The risk events, which have been considered as Environmental Risks in the different categorization schemes, include: (a) Changes in environmental regulations and by-laws (b) Ecological damage (c) Water, Air and Noise pollution (d) Waste treatment (e) Public inquiry (f) Requirements for permits and their approval None of the categorization schemes have included all the above risks under the Environmental Risk category. Most categorization schemes have determined Environmental Risks to be closely related to Political Risks, and have considered both as a single category. Such categorization schemes however, have considered Environmental Risks to be only those due to changes in laws and regulations as well as the permit approval process. The schemes that have considered Environmental Risks as separate to Political Risks, mostly take into account the risks due to pollution and waste disposal. 49 Economic risks that arise due to environmental issues have been termed as Environmental Risks in this thesis as opposed to other definitions provided in the literature. Such a definition is introduced in view of the conflicting definitions provided for Environmental Risks, as well the variety of risk events that have been included within such definitions. 3.3 Risk Management Frameworks The process of risk management in large projects can be considered to be made up of several stages or sub-activities. Perry and Hayes (1985) have considered the risk management process to be made up of three-sub activities: (1) Risk identification, (2) Risk analysis, and (3) Risk response. Kangari and Boyer (1987), in developing a knowledge based system for risk management, have considered five processes. They are: (1) Risk identification, (2) Risk policy description, (3) Risk sharing and allocation, (4) Risk evaluation, and (5) Risk minimization and response planning. Risk management frameworks provide a disciplined, formal approach for applying these sub-processes towards the treatment of risks in large projects. The motivation for adopting such formal approaches stems due to several reasons. These include: (a) A formal risk framework allows management decisions regarding risk to be based upon a sound mathematical and scientific foundation as opposed to disorderly human j udgement. (b) A formal plan for the appraisal and management of economic risks is required for projects funded by some government organizations. Such organizations include the U K Ministry of Defense (Chapman and Ward 1997). 50 (c) They provide a useful tool in convincing top level management, government officials and lending institutions to agree to economically viable project alternatives that would be considered "too risky" in the absence of a formal process. (d) A formal process facilitates the systematic collection and appreciation of project data relevant to risks. Such data would be extremely useful in planning and executing projects of a similar nature in the future. Chapman and Ward (1997) outline a formal risk management process for projects in generic terms. The framework consists of 9 distinct phases with each phase being associated with a broadly defined deliverable. The 9 phases identified by Chapman and Ward (1997) are described below. 1. Define phase - Relevant information about the project is consolidated with a view of obtaining a clear, unambiguous, shared understanding of all key aspects of the project. The documentation, verification and reporting of key information about the project is the deliverable of this phase. 2. Focus - Preparation of a strategic plan for the risk management process at an operational level. The documentation, verification and reporting of the key aspects of the risk management plan is the deliverable of this phase. 3. Identify - Identifying the sources of risk and appropriate responses as well as the possible faults that lie in the responses. The classification, characterization, documentation, verification and reporting of key risks and responses is the deliverable of this phase. 51 4. Structure - The Structure phase involves exploration of the interactions between the risks and developing an ordering of risks based upon a qualitative assessment of their impact on the economic viability of the project. Where appropriate this phase includes testing simplifying assumptions and refining the classification adopted in the Identify phase. Obtaining a clear understanding of the implications of any simplifying assumptions about the relationships between risks, responses and base plan activities is the deliverable of this phase. 5. Ownership - Allocation of risks to project parties. Providing clear and legally enforceable ownership of risks is the deliverable of this phase. 6. Estimate - Obtaining a basis for understanding which risks and responses are important, and estimating risk impacts in numeric terms are the deliverables of this phase. This is achieved by providing simple, numeric subjective probability estimates to the risks. Risks that are deemed to be of significant uncertainty based upon the value of the subjective estimate are subjected to a more refined quantification process. 7. Evaluate - Synthesis and evaluation of the results of the estimate phase. Deliverables of this phase include a diagnosis of all important difficulties, and comparative analysis of the implications of responses to these difficulties. Specific deliverables of this phase include a prioritized list of risks, or a comparison of the base plan and contingency plans. 8. Plan - Producing a project plan and associated risk management plans for the project management process. Complete and appropriate plans are the deliverables of this phase. 52 9. Manage - Monitoring, controlling and developing plans for immediate implementation. Deliverables of this phase include prioritized lists of risk/response issues, exception reporting after significant events, and associated replanning. The Institution of Civil Engineers, and the Faculty and Institute of Actuaries (1998) have also developed a structured approach to risk management. The process is termed RAMP, an acronym for Risk Analysis and Management for Projects. The objective of RAMP is to provide a comprehensive and systematic process for identifying, evaluating and managing risk during the entire life cycle of capital investment projects. The process consists of four activities: a) Process launch A risk management team is appointed to implement the R A M P process during the Process launch stage. The baseline objectives, scope and plans for the project are defined and the underlying assumptions on which they are based are identified. b) Risk review The risk review stage involves systematically identifying risks and entering them into a risk register. Next, the risks are evaluated to determine their likelihood and magnitude. Relationships that exist between the risk are also identified during this stage. Mitigation measures are then identified to avoid, reduce or transfer risks. The measures are incorporated in a risk mitigation strategy. An investment model is used to identify the viability of the project based on the risks that remain. A risk response plan is prepared in the final step of this stage. The risk review stage is repeated at key stages or decision points throughout the life of the investment. 53 c) Risk management This stage involves implementing the risk mitigating strategy and risk response plan developed during the preceding risk review stage. The project activities are monitored to identify new or changing risks. d) Process close-down The last activity is the closing down of the RAMP process. A review of made of the investment in terms of its success in meeting its objectives. The different risk management approaches suggested in the literature consider the risk management process at varying degrees of detail and in a diverse manner. However, several core elements that are common to all the approaches can be identified. These are: 1. Risk identification; 2. Evaluation of the risks; 3. Risk allocation and development of mitigation strategies; and 4. Controlling of risk throughout the project lifecycle. 3.4 Environmental Risk Analysis Framework This section presents a risk analysis framework for environmental risks in large infrastructure projects, which is reflective of risk analysis frameworks suggested in past literature. The purpose of this framework is to enable the effective analysis of the uncertainty caused by environmental issues in project economic variables. It consists of a structured set of activities (see figure 3.1) that lead to the effective analysis of environmental risks. The proposed framework provides a formal, logical and 54 ysi Anal lisk "ca men! iron] Envi Launch Process -a T> w m m <D O O i— PH CO e < > G W 55 systematic tool that helps project stakeholders in identifying and analyzing environmental risks. The framework is intended to be used by project proponents at the project appraisal stage when definitive information regarding the project and the natural environment might not be readily available. However, it can be readily adopted for use by other project stakeholders such as contractors during any stage of the project. The input that would be available to the framework would tend to be refined during its application at later stages of the project. A larger and more definitive body of information would emerge regarding project components as the project progresses. Similar results can be expected with respect to information regarding the environment as studies such as EIAs yield detailed information about the natural environment. Public perceptions too have a likelihood of changing as the project progresses. Initial perceptions may change due to positive or negative public awareness campaigns that are carried out by the project proponents, environmental lobbyist, and government agencies. The early actions of the project such as site investigations and the construction of temporary structures may also affect public perceptions depending upon the short-term environmental impacts of such activities. 3.4.1 Definitions The terms risk and uncertainty are often used interchangeably in existing literature on risk management (Al-Bahar and Crandall 1990). Risk and uncertainty will be treated as synonyms in this thesis as making a distinction between them is considered to be unnecessary in project risk analysis (Perry and Hayes 1985). 56 Risk In the context of project economic analysis, risk can be characterized by three components (Al-Bahar and Crandall 1990): 1. The risk event - the event that leads to an impact on the project economics; 2. The uncertainty of the event - the likelihood of the event occurring; and, 3. Potential loss or gain - the consequence of the risk event. In this thesis, risk is defined as the exposure to the possibility of economic or financial loss or gain as a consequence of the uncertainty associated with a particular component of the project. Expressed in quantifiable terms it combines a probabilistic measure of the occurrence of an uncertain event with the consequences of such an event, which can also be probabilistic. Environmental Risk As described previously in this chapter, a consistent definition for environmental risks does not exist within project management organizations and the civil engineering profession. For the purpose of this thesis, an environmental risk is defined as the exposure to the possibility of economic or financial loss or gain as a consequence of the uncertainty associated with a particular project component that arises or is affected by environmental issues. 57 3.4.2 Scope and Characteristics of the Framework The emphasis of the framework is on the analysis of environmental risks from the perspective of their effect on the project's economic performance. Focus is therefore given to work package components whose cost and duration are subject to uncertainty as a result of environmental risk. Uncertainty in the revenue streams of the project due to environmental issues can be treated in a manner similar to cost components and thus a distinction is not made in the thesis. The framework allows the user to characterize and quantify the uncertainty associated with the parameters of the work packages subject to environmental risks so that they can be readily incorporated into the overall decision making process of the project. However, it does not facilitate the development of risk transfer strategies or other risk mitigation strategies. It is assumed that the development of risk mitigation strategies and their implementation would remain part of the overall risk management process of the project. The framework has the following characteristics, which enable the effective analysis of environmental risks in large infrastructure projects. a) The analysis methodology provided by the framework is structured and objective. This allows the user to follow a clearly defined path in analyzing risks as shown in figure 3.1, rather than doing so intuitively. This ensures that all work packages subject to environmental risks are identified and analyzed. b) It allows the classification and subsequent quantification of environmental risks. Classification enables risk events having similar characteristics to be treated in unison. This assists the user in selecting a suitable quantification methodology for their analysis. From a global project perspective, risk classification also assists in the development of risk mitigation strategies. The quantification of risks instead of 58 a qualitative analysis allows the user to make a comparative evaluation of the impact of the individual risks. This enables risks to be prioritized based on a common criterion such as the most extreme magnitude of the risk outcome. The quantification of risks also enables the user to express the individual risks in a manner suitable for incorporation into the overall risk management process of the project. c) The framework provides for the systematic accumulation of information regarding risk identification. The framework ensures that risks identified during the implementation of the framework are systematically documented and carried over to future projects. 3.4.3 The Framework The proposed framework for the analysis of environmental risks consists of three stages. Each of these stages consists of several sub-steps. The aggregate of these steps make up the risk analysis process as shown in figure 3.1. Define Stage The Define stage initiates the risk analysis process. The objective of this stage is to obtain a clear, unambiguous, yet general definition of the project and the environment in which it is located. The definition of the project is provided in terms of its scope and its component work packages. Scope parameters that are commonly used to describe projects have been identified previously in section 2.3. Examples of scope parameters include the physical size of the development such as the length of a highway, the capacity of the 59 project such as the production capacity, and the energy usage rate of the project. The project scope is defined using the parameters that are most commonly used to describe the particular type of project as well as in terms of parameters that are of particular relevance to the environment such as the waste discharge rate and the energy usage rate. An example would be an instance where the scope of a water storage reservoir is specified in terms of the dam height, the maximum volume of water that will be impounded, and the area that will be inundated at full capacity. Thereafter, the component work packages of the project are identified. A work package is defined as a distinct project component made up of a collection of activities, which consumes resources and is of a finite duration. In identifying the work packages, the five project stages of Design and Planning, Construction, Operation, Modification and Closure are considered separately. This step results in a taxonomy of work packages that constitute each of the five stages of the project. A general description would accompany each work package. Such a description would include: 1. The magnitude of the output of the work package; 2. The schedule of the work package including the start date, end date and whether the activities are continuous throughout the work package duration or otherwise; and 3. A brief outline of the work method. An example of such a description is shown in figure 3.2. As definitive project information might not be available during the Define stage the work package description is provided using approximate values based on the information available at that stage. 60 Proje ct — R a p i d T ran s it Pro j e ct Location — Vancouver Scope - 19.5 km in length - Dual Track Work Package Details Excavation for column footings - Quantity - Approx. 15,000 cubic meters Locations A42 -A51 Start date -1st week of May 2000 End date . 3rd week of May 2000 Continuous - 2 days at each location Using C A T - 3 0 0 series excavators) L Fig. 3.2. Description of Work Packages for the Purpose of Environmental Risk Analysis The next step in the Define phase involves the definition of the natural environment. The Valued Ecosystem Components (VECs) of the environment are identified as part of this step. The tools available for making such an assessment include public surveys, literature surveys and expert interviews. In the next step of the process, the values of the environmental components are estimated at a general level in line with the accuracy required of the overall process. While several methods of valuation exist with regard to the value of the natural environment (Munasinghe 1993), any direct estimation of the economic value of the natural environment is considered to be difficult and uncertain (Ranasinghe 1999). Therefore, it is suggested that the characterization of environmental components as Highly Important, 61 Moderately Important, and Slightly Important is sufficient in view of the specificity required and in view of the limited amount of information that may be available during the feasibility stage. The characterization of environmental components at a very general level is aimed at highlighting the more important environmental components and serving as an aid to identifying the risks that may be associated with such components. A description of the factors that contribute to the importance of the environmental components should accompany each component that is identified. An example of such a description is provided in figure 3.3. Pro je Ct — Rap i d Tran s it Proj e ct Location — Vancouver Important Environmental Components Environmental Importance Description Component Spotted Owl (Strix High Listed as an Endangered occiden talis) species. Moose Creek Medium Creek supports salmon and trout runs ^ ^ ^ ^ ^ Fig. 3.3. Description of the Natural Environment for the Purpose of Environmental Risk Analysis 62 The output of this stage is a document that lists: 1. The scope of the project; 2. The component work packages accompanied by a description of their schedule and the magnitude of their output; and 3. The important environmental components of the area accompanied by a characterization of their importance and a description of the factors that contribute to their importance. Risk Identification Stage The aim of this phase of the risk analysis process is to: 1. Identify as exhaustively as practicable, all work packages that arise or are affected by environmental issues and are of an uncertain duration and cost; and, 2. Characterize and classify the risk associated with the work packages. Prior to the actual identification step, an analysis of the public perception towards the project is carried out in order to assess possible influences on the relationship between the project and the natural environment. The appraisal should identify the major causes for public concern regarding the environmental effects of the project. The appraisal is carried out using surveys and by interviewing key public figures such as politicians, community leaders, leaders of first nations groups and public officials. Discussions should also be held with representatives of special interest groups such as environmental and outdoor recreation groups with aim of identifying key environmental concerns with respect to the proposed development. 63 A risk register serves as a starting point for the risk identification step. A risk register is a comprehensive listing of the environmental risk events that might be encountered on projects of the type under consideration. Risk registers maybe compiled and updated using information from previous projects and with the aid of literature surveys. An extract of a risk register developed for large hydropower projects is shown in figure 3.4. In the risk identification step the analyst browses through all the risk events listed in the register and determines their applicability to the project under review. In the next step, experts such as project managers with experience in similar projects are consulted in order to further develop the list of risks. Participants in this process are encouraged to list even seemingly unlikely risks and scenarios. It is also useful to review relevant literature describing case studies of similar infrastructure projects in order to develop the risk register even further. Government regulations pertaining to environmental issues are also analyzed during this stage in order to Work Package with Uncertain Cost and Duration Description of Environmental Issues that Create / Contribute to the Work Package Cost and Duration Relevant Project Phases Risk Type Removal of hazardous waste from areas that will be inundated Hazardous wastes include mine tailings and industrial wastes deposited during previous land use. Unremoved wastes will lead to increases in bio-accumulation of heavy metals and toxic organic compounds in fish. Construction, Modification Bl Installation of a sediment controlling system to treat aggregate wash water A sediment controlling system is required to prevent fines from reaching fish habitats. Fine sediment particles smother fish eggs and fry, and reduce the availability of food and oxygen in the water by retarding photosynthesis ^, Construction, Modification Al Fig. 3.4. Extract from a Risk Register for Large Hydropower Projects 64 identify regulations relevant to the project and the environment. This step will result in the identification of the permits and approvals that will be required for the execution of the project as planned. In the Risk Identification stage, all probable risks are listed without making a pre-judgement on the extent of their impact. Initial assessments carried out regarding public perceptions and applicable government regulations are used in identifying risks as well as in determining the applicability of risks to the project under consideration. The outputs of the Define stage, which set out the attributes of the project and the environment, are used in a similar manner to determine the relevance of a particular risk to the project as well as in generating the list of probable risks. The risk events are grouped into 4 types in the next step of the process. The 4 types follow from an initial subdivision of risks into 2 types. Initially the risks are subdivided based on the nature of their occurrence. The cost and duration uncertainty associated with a project component that occurs with certainty during the lifecycle of the project is termed as a Type A risk. The cost and duration uncertainty is classified as Type B should the occurrence of the project component itself be an uncertain event. The risk types are further subdivided into 2 groups based on the role environmental issues play in creating the project component. If an environmental issue results in a need for a work package that would normally not be part of the project, the cost and duration uncertainty associated with such a risk is termed as a Type 1 risk. Therefore, in accordance with the previous classification of risks, the cost and duration uncertainty associated with a work package that occurs with certainty during the project lifecycle and arises due to environmental issues is classified as a Type A l risk. 65 Project component of uncertain cost and duration (e.g. Construction of a Fish Ladder) Yes Project component Type = A New Addition Type = A l Type = A2 Project component Type = B Addition New Type= B1 Type= B2 Fig. 3.5. Classification of Risks Similarly, the cost and duration uncertainty associated with a work package that arises due to environmental issues but is of an uncertain occurrence is classified as a Type B l risk. Should environmental issues contribute to the cost and duration of a work package that is normally part of the project, the cost and duration uncertainty associated with such a work package is termed as a Type 2 risk. In accordance with the previous classification of risks, the cost and duration uncertainty associated with an addition to an existing work package would be classified as a Type A2 risk provided 66 that the addition occurs with certainty during the project lifecycle. Similarly, the cost and duration uncertainty associated with an addition to a work package that arises due to environmental issues but is of an uncertain occurrence is classified as a Type B2 risk. The classification process is shown graphically in figure 3.5. Formal definitions of the risk types are presented in Chapter 4. The output of this stage is a comprehensive and exhaustive list of risk events that have the potential to affect the economic performance of the project. The risk events would be grouped into the 4 types identified above. The risk events identified using sources external to the risk register are added to the original register in order to develop it former for use in future projects. Analysis Stage The objectives of the Analysis stage are: 1. To determine the work packages with uncertain cost and duration that are most relevant to the project's performance in terms of their capability to affect the economic parameters of cost and duration; and 2. To quantify the cost and duration uncertainty of such work packages. The analysis stage consists of several distinct steps. The first step of the process is a screening step in which the identified environmental risks are screened with respect to their possible effect on the economic parameters of cost and duration. The number of risks that arise due to environmental issues might be very large in considering infrastructure projects. The risk register for large hydropower projects that is provided in Appendix A is used as an illustration. This risk register contains 67 approximately 100 components that are of uncertain cost and duration. The number of risks could increase further in an actual hydropower project as the list is refined by identifying site-specific risks. Some of the risks that are identified during the risk identification stage would be capable of having a major impact on the economic success of the project due to the significance of the outcome of the risk events, such as a delay in obtaining an environmental permit that prevents construction from commencing prior to the onset of winter. However, a large number of risks that are incapable of having such major effects would also exist. In an ideal situation, the uncertain cost and duration consequences of all environmental risks would be evaluated using a comprehensive quantification methodology. However, in a real-life scenario, time and other resources that are required for such an exercise might not be available. Alternatively, the benefits derived from such an exercise could be considered as being insufficient in relation to the resources that need to be expended to carry out such a comprehensive analysis. Therefore, a screening process that can be used to identify risks which warrant a full-fledged analysis is required as part of the risk analysis framework. Environmental risks can be prioritized based upon their likelihood of having a significant impact on project cost and duration. Work packages that can be considered to be associated with significant cost and duration risk include the work packages that; (a) Have a high variability associated with their cost or duration; (b) Have uncertain cost or duration values that have a possibility of achieving extreme magnitudes in value even though such an occurrence might be unlikely; (c) Lie on the critical path of the project network, and the timing and duration of which can significantly affect other work packages of the project. 68 The spread of the possible values of a variable represents the degree of uncertainty associated with work package cost or duration. Extreme values of economic parameters represent unlikely, yet possible outcomes that could have a significant impact on the overall economic viability of the project. The cost and duration uncertainty associated with work packages that lie on the critical path of the project network are of particular importance since changes in their duration can affect the timing of successor activities as well as the overall project duration. Cost and duration risks due to environmental issues can be prioritized based on the characteristics described above. Project components that have a high degree of uncertainty associated with their cost or duration, or hav| a possibility of achieving extremely high cost or duration values, or have the ability 'io affect critical activities of ' -1," \ . • •- • \ V| the project can be considered to be of a higher priority than other-project components. Therefore, an appropriate basis for prioritization would consist of a set of conditions based on the characteristics described above. Work packages of uncertain cost and duration that meet such standards would be subject to a rigorous evaluation. An example set of conditions would be work packages with uncertain duration such that: (1) the difference between the maximum possible duration and the minimum possible duration is greater than 10 days; or, (2) the difference between the maximum possible duration and the minimum possible duration is greater than 4 days should the activity be a critical activity. The selection of such a set of values for prioritization would depend on the characteristics of the project such as its total budgeted cost, project deadlines that have to be met and the consequences of non-compliance, the float available for non-critical activities, as well as on the risk perceptions of the project 69 management personnel. An appropriate basis should be decided upon by the analyst at the start of the Analysis stage. However, information relating to parameters such as the expected value and the extreme values of the cost or duration of a work package would not be readily available at this stage of the analysis. Therefore, initial estimates need to be obtained from experts regarding the expected value of the economic variable, and its minimum possible value and maximum possible value in order to carry out risk prioritization. Estimates for work packages subject to Type B risks would need to be provided considering the whole spectrum of outcomes of external factors that affect such risks. The elicitation procedure utilized for obtaining this information is described in detail in Chapter 6. It is noted that the parameters obtained in this manner would not be of high accuracy as the direct estimation of the moments of variables and the estimation of extreme values is considered as being difficult and inaccurate (Keefer and Verdini 1993). However, they have been selected for use in risk screening due to the need to achieve a compromise between a comprehensive analysis and the resources required for such an analysis. These values can then be used in selecting risks that warrant further analysis along with any information relating to the criticality of work packages should such information be available. The expected value of the economic variable which is obtained in this step can be used to characterize the cost and duration of work packages that are considered to be non-significant. The cost and duration of such work packages are thus characterized by a single point estimate. The screening step can therefore be considered to consist of a decision step in which an appropriate basis for the prioritization of risks is selected. The next step consists of an estimation stage in which initial approximations for the expected value, 70 and the end points of the variable are obtained for all identified risks. Risks that warrant further assessment are selected using the basis selected in the primary step. The next step of the Analysis stage involves obtaining the necessary information required to carry out a comprehensive analysis of the cost and duration uncertainty of the work packages that are deemed to be significant, information is obtained using data from previous projects of a similar nature or from experts using an elicitation process. These risks are then quantified using different quantification methodologies in accordance with their risk classification. These two steps are described in detail in Chapters 4, 5 and 6. The representation of the uncertainty associated with work packages subject to environmental risks needs to be carried out in a standardized form in order to incorporate the uncertainty into the overall decision making process of the project. Uncertainty is in general characterized by a probability distribution that describes the relationship between the probable outcomes of an event with their associated probabilities. The probability distribution can be described by its moments or by expressing it in the form of a commonly used statistical distribution. The risk quantification tools employed in this framework enables the uncertainty to be expressed in terms of standard statistical distribution in the event of the distribution being unimodal. It also allows the first four moments of the distribution to be obtained irrespective of the type of the distribution. The uncertainty representation step is the concluding step in the risk analysis process. 71 3.5 Summary Various risk management frameworks have been presented in the literature for the analysis and management of project risks. However, consistent and comprehensive treatment of risks due to environmental issues has not been presented in these frameworks. The environmental risk analysis framework presented in this chapter provides a formal, logical and systematic tool that is aimed at assisting project stakeholders in identifying and analyzing environmental risks. The framework is intended to be used at the project appraisal stage when definitive information regarding the project and the natural environment might not be readily available. The framework presents methodologies that allow the user to take into account the ability of external factors to affect the economic parameters of work packages. The uncertainty of the work package performance measures of time and cost is characterized in a manner suitable for incorporation into the overall decision making process. A n information accumulation process and risk quantification processes that are part of the risk analysis framework are described in the following chapters. 72 Chapter 4 Quantification of Type A Environmental Risks 4.1 Introduction In economic risk analysis of large civil engineering projects, one is mainly concerned with cost, revenue and duration uncertainty associated with the activities of the project. For the purpose of risk analysis, it is useful to consider the economic variables of the project in the form of a hierarchy. During, the feasibility stage of the project, work packages can be considered as the lowest level of such a hierarchy which can be meaningfully analyzed. As described in chapter 2, a variety of environmental issues can cause uncertainty in the component work packages of the project. The risks that arise due to environmental issues can be classified into 4 types in accordance with their effect on the project work packages as identified in the previous chapter. The four types have been termed as Type A l , Type A2, Type B l and Type B2 environmental risks in this thesis. This chapter provides a detailed definition of the four risk types and describes the proposed quantification methodology for risk Types A l and A2. A review of existing risk quantification methodologies and their elements is also presented in this chapter. 4.1.1 Type A l Environmental Risks A Type A l environmental risk is defined as the cost and duration uncertainty associated with a work package that arises due to environmental issues and which will 73 be executed with certainty during the life cycle of the project. In the case of such a work package it will be known with certainty during the project appraisal stage that time and other resources will have to be expended for the work package which might be initiated in response to environmental regulations, in order to prevent environmental damage, to compensate for environmental loss or in response to public pressure. A work package of this type is similar to most of the traditional work packages of a project in that while its occurrence is certain its cost and duration are of an uncertain nature. An example of a Type A l risk is the cost and time uncertainty associated with the construction of a fish ladder during the execution of a hydroelectricity generation project. A fish ladder is a device that enables upstream migrating fish to by-pass man made obstructions such as dams and weirs. Man made barriers can have disastrous consequences for migratory fish. A single dam with a height greater than a few feet will block species such as salmon from migrating to their ancestral spawning grounds, eliminating all traces of the fish from upstream waters. A fish ladder provides a passage with a gentler gradient and a flow regime which the fish will instinctively pursue. In practice a fish ladder is often a 10-percent-graded flume interrupted with vertical, slotted partitions which provides a link between the upstream and downstream sides of the dam (Rainey 1997). Fish ladders are generally constructed using material such as wood or aluminum. In the case of salmon, a maximum one-foot drop in water level is provided at each partition. 74 Uncertainty can arise in the cost and duration of the fish ladder construction work package due to several factors: 1. The quantity of material that will be needed will not be known with certainty during the project appraisal stage. 2. The unit cost of material will be uncertain due to escalation and market variations. 3. The productivity of the labor force that will be employed in the construction of the fish ladder will be of an uncertain nature. The productivity would depend on the experience and motivation of the labor force, on the weather conditions, on learning curve effects, and on managerial competence. 4. The productivity of the equipment used would be uncertain and dependent upon site conditions such as the geology, on the weather conditions, and on operator competence. 4.1.2 Type A2 Environmental Risks In addition to creating a necessity for the execution of additional work packages, environmental issues can also affect the scope of work packages that are normally part of the project. In such instances, environmental issues contribute to the time and cost uncertainty inherently present in the work package. Such work packages are defined as being subject to Type A2 environmental risks. An example of this type of risk is the cost and time uncertainty associated with the site clearing operation during the construction of a large dam. In addition to the land clearing which will be carried out under normal circumstances, an additional effort would have to be made to remove vegetation and organic matter from the area that will be inundated, due to environmental considerations (Sadar 1996). If vegetation 75 and other organic matter are not removed, bacteria feeding on the drowned vegetation convert mercury naturally present in organic matter into methyl mercury. Methyl mercury, which is more toxic than mercury itself, can then be released into the food chain and accumulate in living organisms. Subsequently, it can move up the food chain and poison humans who consume contaminated fish species (Foulke 1994). Therefore, removal of vegetation should be carried out in instances where the fish living in the reservoir or in the downstream reaches of the river are likely to be consumed by humans. 4.1.3 Type B Environmental Risks A Type B environmental risk is fundamentally different to risks of Type A in the sense that the occurrence of a work package, or the occurrence of an addition to an existing work package cannot be predicted with certainty during the project appraisal stage. The need for a new work package to be executed, or for an addition to an existing work package to be implemented, as well as their scope once the need arises would depend on several external factors. A Type B l environmental risk can be defined as the cost and duration uncertainty associated with a work package, the existence and magnitude of which are uncertain, and are subject to the influence of factors external to the project. External factors refer to events external to the project over which project management personnel are unable to exert a large amount of control. An example of a Type B l risk is the cost and duration uncertainty associated with the cleanup of an accidental oil spill. Similarly, a Type B2 risk is defined as the cost and duration uncertainty associated with an uncertain addition to an existing work 76 package. In this case the occurrence and magnitude of the addition are both uncertain and are subject to the influence of factors external to the project. An example of a Type B2 risk is the cost and duration uncertainty associated with a change in environmental regulations that creates a need for extraordinary changes in the quantity of contaminated soil to be excavated from a site. Type B l and B2 risks are discussed in depth in Chapter 5, which also presents detailed examples of such risks. 4.2 Economic Risk Quantification Methods The uncertainty surrounding the cost or duration of a work package can be represented as a probability distribution which outlines the relationship between cost or duration values with the probability of their occurrence. These distributions can be estimated holistically by the analyst or derived from a functional relationship which links them to their primary variables. In assigning a distribution to decision variables such as the cost and duration, decomposing the variable into its primary variables is often regarded as a useful technique for reducing the complexity of the estimation problem. More importantly it also allows the development of a link between the two decision variables of cost and duration. In general terms, the expressions for the work package cost and duration can be represented as Y = g ( X „ X 2 , X 3 X„) (4.1) in which Y is the decision variable and X is the vector of primary variables such as quantity, unit cost and productivity. Once the uncertainty of primary variables are quantified using actuarial or frequency data from past projects or using knowledge 77 elicited from experts, the uncertainty of Y can be assessed using analytical methods (Kottas and Lau 1978,1982, Ranasinghe 1990, Russell and Ranasinghe 1991) or by using a simulation technique such as Monte Carlo simulation (Lorance 1992, Rao and Grobler 1995, Songer, Diekmann and Pecsok 1997, Uher 1996). 4.2.1 Monte Carlo Simulation In Monte Carlo simulation, a value is chosen randomly from the cumulative distribution function of each primary variable. The corresponding value of the decision variable is calculated using the vector of values for the primary variables which is generated randomly and the functional relationship Y = g ( X , , X 2 , X 3 X n ) . The same procedure is repeated over a large number of iterations and each output value is assigned a probability of 1/N, where N is the total number of iterations. The probability distribution for the decision variable is generated from the set of output values. The main advantage of the Monte Carlo method is its simplicity. However, the need to specify a cumulative distribution function for the input variables in a standard statistical form can be considered to be a disadvantage of the Monte Carlo method. Even though empirical distributions can be used to specify the cumulative distribution function of a primary variable when the large amount of frequency data necessary to develop such a distribution is available, the use of such distributions has several disadvantages (Law and Kelton 1991). These include the inability of empirical distributions to portray extreme values outside the range of the original data set and the necessity for the usage of a large amount of computer disk storage in order to store representative values of the empirical distribution. Additionally, in the event of 78 subjective estimates having to be used in the absence of frequency data, it is not feasible to elicit a large enough data set necessary to define an empirical distribution. This necessitates a need for the subjective probabilities elicited by the experts, as well as frequency data to be expressed in a standard statistical form such as a triangular, beta, lognormal or normal distribution. Even though parameters such as the mean and the standard deviation of the variable can be calculated fairly accurately from the subjective estimates provided by experts, the choice of the distribution type to represent primary variables remains an unresolved issue in engineering economic analysis. Representation of variables in standard statistical form The literature dealing with the selection of a standard statistical form to represent the uncertainty of variables in Monte Carlo simulation deals mainly with the cost and duration parameters of a work package. However, the discussions are still valid for primary variables such as resource productivity, where the uncertainty has to be assessed subjectively. Chau (1995a) discusses the validity of the triangular distribution, commonly used to represent cost uncertainty in construction. Three reasons for the popularity of the triangular model have been identified by Chau. These are: (a) The triangular distribution has the ability to model the skewed nature of construction costs; (b) The parameters needed to describe the triangular distribution, which are the minimum (L), most likely (M) and the maximum (H) values are relatively easy to elicit from experts; and 79 (c) The use of a triangular model simplifies the computational process. However, because very high values of cost are often the result of unexpected events with low probability, the right tail of the cost distribution can be expected to be a lot thinner than is implied by a triangular distribution (Chau 1995b). Therefore, even though the triangular distribution retains the ability to portray skewness, its inability to model a long thin tail is likely to lead to an overestimation of the probability of exceeding the most likely value. An empirical study carried out by Chau (1995a) utilizing data from electrical services contracts has shown that the assumption of a triangular distribution leads to an upward bias in the calculation of the probability of exceeding the most likely estimate. Chau (1995b) suggests the use of the log-triangular distribution to overcome the inability of the triangular distribution to represent the long thin tail of the construction cost distribution. The log-triangular distribution is essentially an exponential transformation of the triangular distribution. If X is the variable under consideration that is assumed to be distributed as log-triangular, then ln(X) ~ A (ln(L), ln(M), ln(H)) (4.2) Therefore, the distribution can be fully determined by the same three points estimated for the triangular distribution. However, comprehensive validation of the ability of the log-triangular to model the uncertainty of construction costs does not exist. Touran and Wiser (1992) have also identified that the distributions of cost items are unimodal and have skewed distributions. Unit cost data items from building 80 projects for work such as concrete, masonry and finishes have been used in their analysis. Normal, log-normal, beta and gamma distributions have been fitted to the histograms of the data and evaluated using the chi-squared test of goodness of fit. Touran and Wiser report that in almost all cases the log-normal distribution gave the best fit followed by the beta distribution. The log-normal distribution is defined only for values greater than zero and is positively skewed with a long tail, which theoretically makes it a good candidate for representing cost uncertainty. An added advantage of using the log-normal distribution to represent cost data is the relative ease with which correlated random numbers can be generated for Monte Carlo simulation in the event of the cost items being correlated (Touran and Wiser 1992, Touran 1993). However, Wall (1997) states that the beta distribution remains the most widely used distribution by practitioners in Britain for representing cost uncertainty. The beta distribution is also used to represent duration uncertainty most notably in the Program Evaluation and Review Technique (PERT). In traditional PERT analysis the mean (E(X)) and the standard deviation (CT) of the beta distribution is obtained from the minimum (L), most likely (M) and the maximum (H) values using the equations (Clark 1961), E(X) = H + 4M + L (4.3) V 6 (4.4) 81 Although, these approximations are widely used in the industry (Keefer and Verdini 1993) they have been criticized as having no empirical justification and as being prone to significant errors (Chau 1995, Keefer and Verdini 1993, Lau and Lau 1998). Several alternative estimates have been suggested to equations (4.3) and (4.4) including an estimate of the mean as E(X) = 0.63P50 + 0.185(P5 + P95) (4-5: where P5, P 5 0 and P 9 5 are the subjective estimates for the 5 t h, 50 t h and 95 t h percentile values respectively (Keefer and Verdini 1993) and an estimate for the standard deviation (Lau and Lau 1998) as c7 = 0.34(P 9 5-/> 5 0) + 0.28(P 5 0-/>) ( 4 6 However, Kamburowski (1997) has defended the use of the classical PERT formulas given in equations (4.3) and (4.4). Kamburowski has based his attempt to validate the formulas on the characteristics of the kurtosis of the beta distribution. It is clear from this literature review that consensus does not exist within the industry about the suitability of a particular distribution to represent cost and time uncertainty as well as about the estimation of the parameters of the distributions. 82 Treatment of correlation among primary variables In order to obtain realistic distributions of the decision variables such as cost and duration it is required that proper treatment of the correlation among primary variables be carried out (Chau 1995b, Moselhi and Dimitrov 1993, Ranasinghe and Russell 1992, Rao and Grobler 1995, Touran and Wiser 1992, Wall 1997). Failure to take account of the interdependence among primary variables will result in the variance of the decision variable being smaller than or greater than the true value (Moselhi and Dimitrov 1993). This can lead to an underestimation or an overestimation, respectively of the risk associated with a decision variable such as cost or duration. The covariance of two variables is a measure of the linear interdependence between the variables. The dependency is generally expressed in terms of the correlation coefficient which is a normalized version of the covariance. The correlation coefficient, also known as the Pearson correlation coefficient is defined as, PX,Y = V ax°rr j (4.7) where cov(X,Y) = the covariance of X and Y CTX = standard deviation of X a Y = standard deviation of Y Extreme values of the Pearson correlation coefficient, i.e. values that approach -1 or 1, imply that the relationship between the variables is approximately linear, whereas 83 values closer to zero imply that the linear trend between the variables are weak (Benjamin and Cornell 1970). However, the use of Pearson correlation coefficients in Monte Carlo simulation is constrained by the practical difficulty in generating correlated random numbers. The joint density function of the correlated variables is not readily obtainable unless the primary variables are distributed normally, or unless they can be described by a transformation of the normal distribution such as the log-normal or sinh"1 normal distribution (Touran and Suphot 1997, Touran and Wiser 1992). However, as the previous discussion on the shape of the primary variable distributions has shown, it is important that flexibility exists in order for variables to be described in terms of other distributions such as the beta or triangular distributions. The use of Spearman rank correlations has been suggested as an alternative method to overcome this difficulty (Touran and Suphot 1997). The Spearman correlation coefficient, R is defined as (4.8) where rj = rank of the i value of variable 1 Sj = rank of the i * value of variable 2 r = mean of the ranks of variable 1 s = mean of the ranks of variable 2 84 In order to calculate R, the values of the two random variables under consideration are ranked in ascending order. In the case of two or more sample values of a variable being equal, the average rank of the tied data is assigned to the values. The value of R is then calculated using equation 4.8. The computation of the two correlation coefficients is illustrated using the set of values given in table 4.1. The values relate to labor productivity in masonry projects (Thomas and Sakarcan 1994) and shows the relationship between labor productivity and the quantity of work. The Pearson correlation coefficient between productivity and quantity can be calculated by directly applying the values given in columns 1 and 3 to equation 4.7. In this case the coefficient is obtained as - 0.112. The values given in columns 1 and 3 need to be ranked in order to obtain the rank correlation coefficient. Table 4.1. Correlated Values of Labor Productivity and Work Quantity Quantity (Q) -m2 Rank ofQ Productivity (P) -Hours /m 2 Rank OfP (1) (2) (3) (4) 4,173 1 0.096 3.5 29,138 3 0.096 3.5 39,251 7 0.105 6 48,111 9 0.1 5 38,984 6 0.089 2 32,160 4 0.067 1 33,952 5 0.141 9 14,068 2 0.123 8 40,147 8 0.119 7 Source: Data from Thomas, H.R. and Sakarcan, A.S. (1994) Forecasting labor productivity using factor model. Journal of Construction Engineering and Management. 120, 228-239. 85 Each set of values is ranked in ascending order with the average rank being assigned in the event of ties. The ranks are shown in columns 2 and 4 of table 4.1. The rank correlation coefficient is obtained by applying the rank values to equation 4.8. In this case R is obtained as 0.124. As can be seen by this example, significant differences in value can exist between the two correlation coefficients. Iman and Conover (1982) have developed a distribution-free approach for invoking correlated random variables based on the assumption that Spearman correlation coefficients are a meaningful way of describing the dependence which exists between the variables. Results of a study carried out by Touran and Suphot (1997) involving construction cost data have indicated that the use of rank correlations was as effective as the use of the Pearson correlation coefficient in modeling interdependence between variables. Most of the commercially available software for carrying out Monte Carlo simulation make use of rank correlations in order to provide flexibility in specifying the shape of the distribution of the primary variable, while allowing the simultaneous modeling of interdependence (Palisade Corp. 1997). However, the main disadvantage of using rank correlations to specify interdependence is the difficulty in assessing a subjective value for the rank correlation coefficient in the absence of frequency data. Providing a physical meaning to the concept of the rank correlation coefficient is considered to be difficult, hindering the direct estimation of its value from experts (Fackler 1991, van Dorp and Duffey 1999). 86 4.2.2 A n a l y t i c a l M e t h o d s of R i s k Quan t i f i ca t i on Analytical methods provide an alternative to the use of Monte Carlo simulation in risk quantification. The simplest analytical approach is to use the mean and variance of the primary variables to calculate the first two moments of the target variable. The expected value of a decision variable Y can be obtained by applying the expectation operator to the truncated, second order Taylor series expansion of Y . The expected value of Y is obtained as •idx 2 ~ xi ^2_j xv PxixjGxi^xj (=1 1=1 (4.9) Similarly, the variance of the Taylor series expansion Y is given by, Var(Y)« J 2E I (=i d2g CQViX„Xj) (4.10) when 2 n d and higher order terms are neglected. Equations (4.9) and (4.10) provide an estimate of the uncertainty of the decision variable Y , in terms of the expected values, variances and covariances of the primary variable. Unlike in Monte Carlo simulation, the shape of the probability distribution of the primary variables need not be specified in this method. However, the truncation of higher order terms in the Taylor series expansion as well as in the expression for the variance of the derived variable can lead to errors in the estimation of uncertainty. 87 Additionally, a standard statistical shape will have to be assumed for the probability density function of the decision variable in order to make probabilistic statements. The choice of standard statistical shapes will be limited as only distributions which can be specified using the two known values of E(Y) and Var(Y) can be used in the analysis. Kottas and Lau (1982) have overcome this disadvantage by using a four moment approach for a class of stochastic management models. Analytical solutions have been developed by Kottas and Lau for the first four moments of a derived variable which is related to its primary variable by a function which involves only the four basic arithmetic operators of addition, subtraction, multiplication and division. A variety of conditions such as statistical independence and normality of variables have been taken into account in developing the class of models. The probability density function of the derived variable has been assumed to belong to the versatile four parameter Pearson family of distributions. Ranasinghe (1990) has extended the use of four moment analysis to the more general case of functional relationships in the form Y = g (X , , X 2 , X 3 X„) by considering the truncated Taylor series expansion of Y . However, the complexity of the functional relationships between variables and the variation in the functional form of the relationships between different scenarios are more conveniently accommodated by Monte Carlo simulation which has made analytical approaches less appealing than simulation in economic risk analysis. 4.3 Quantification Methodology for Type A l and Type A2 Risks As described previously, work packages subject to environmental risks of Type A l arise due to an environmental issue. Their cost and duration can be expressed in 88 terms of their uncertainty causing primary variables as a function of the form Y = g ( X „ X 2 , X 3 X n ) . In considering Type A2 risks, the work packages would comprise of several additive components, some of which are brought on by environmental issues while the remainder is normally part of the project. In the example of site clearing introduced in section 4.1.2, an additional duration of Y 2 may be required for the removal of organic matter as a result of concerns about mercury poisoning, and a further duration of Y 3 may be required to remove topsoil contaminated with agricultural pesticides. This would require that a total duration of (Y, + Y 2 + Y 3 ) be expended for site clearing where Y, is the duration of the work package when environmental issues are not present. While working at the decision variable level allows the total value of the variable to be obtained by the simple addition of the 3 components, a similar approach is not feasible at a primary variable level. The difficulty of such an approach can be illustrated by considering the treatment of a primary variable such as labor productivity. In considering the value of labor productivity (P) for a work package such as land clearing, let the values of the productivity for the work durations Y b Y 2 and Y 3 be P,, P 2 and P 3 respectively. Unlike in the case of decision variables, the productivity for the complete work package cannot be expressed as the sum of the 3 disaggregate values, i.e. the productivity for the complete work package cannot be expressed as being equal to (P, + P 2 + P3). Other primary variables such as resource usage rates (R) and unit costs also display a similar behavior. An exception to this behavior is provided the quantity of work, where the total quantity of work, Q can be expressed as Q = Qi + Q 2 + Q3, adopting a similar notation as above. However since, P 89 * Pi + P2 + P3 and R * R, + R2 + R3 (resource usage rates) it is not possible to express the total duration, Y as, Y = — = — _ = — ( 4 . 1 1 ) (Pi+P2+PM+R2+R3) Three different approaches can be adopted in carrying out the quantification of the cost and duration uncertainty of work packages under such circumstances. These are illustrated using the same example employed above. The 3 approaches are: (a) Carry out the analysis at decision variable level and quantify the contribution of each environmental issue to the total uncertainty. This can be accomplished by estimating the probability distributions for the durations Y b Y 2 and Y 3 by considering them holistically and subsequently adding them together. Therefore, the total duration Y is obtained as Y = Y, + Y 2 + Y 3 . However, in using this method the advantages offered by decomposition, i.e. the ability to analyze the effect of environmental issues on the primary variables leading to a more accurate estimate and the ability to develop a link between the cost and duration, cannot be utilized. (b) Carry out the analysis at primary variable level and assess their uncertainty by considering all relevant environmental issues in their totality. This would be in effect be equal to providing a probability distribution function to each of the primary variables which accounts for the summation of the contributions to the work package by environmental issues together with the extent of the work 90 package in the absence of environmental issues. In the example under consideration probability estimates would be provided to Q, P and R considering the work package in its totality. Thereafter, the total duration Y is obtained using the relationship, Y = Q /PR. This approach allows the analysis to be carried out at primary variable level. However, since estimates for Q, P and R are assigned considering the total scope of work it is not possible to assess the individual magnitudes of Y b Y 2 and Y 3 . (c) Divide the work package into several components according to the contribution of each environmental issue and then carry out decomposition of the individual components. Symbolically this can be expressed as Y = + h C4 i2) PA P2R2 W 1 ' This approach allows decomposition to be used in the analysis. Additionally, Y, , Y 2 and Y 3 can be calculated separately should the need arise. However, this approach involves carrying out the analysis at a level more detailed than the work package level. Approach (c) is the most attractive with regard to the outputs that can be obtained. However, the work package level is considered as the most detailed level at which meaningful information can be obtained during feasibility analysis (Russell and Ranasinghe 1992). Therefore, approach (c) cannot be adopted during the feasibility stages of the project. Of the remaining two approaches, the second approach has been chosen in the analysis of environmental risks in this thesis due to the better accuracy it 91 provides as a result of considering primary level variables and due to the ability to establish a link between cost and duration by carrying out the analysis at such a level. By adopting approach (b), the cost and duration of work packages subject to A2 risks can also be represented by the general equation (4.1). Therefore, Type A2 risks mirror Type A l risks in terms of the manner in which the relationship between the decision variables and their uncertainty causing agents are represented. This allows environmental risks of Type A l and Type A2 to be considered in unison for the purpose of risk quantification. Therefore, the same analysis procedure will be adopted for both these risk types. The risk quantification method proposed for Type A l and Type A2 risks is based on the premise that the set of primary variables which are functionally related to the decision variable can be represented by continuous, unimodal probability distributions. The proposed method uses Monte Carlo simulation as the quantification tool while aiming to overcome its shortcomings by dispensing with the need for the user to pre-specify a standard statistical distribution shape to describe the uncertainty of the input variable. This approach which utilizes the Pearson family of frequency curves, derives its motivation from the work carried out by Ranasinghe (Ranasinghe 1990, Ranasinghe and Russell 1991) in developing an analytical method of risk quantification. A flexible method that does not limit the input distributions to a single user specified distribution will complement the use of rank correlations to describe the dependence between the variables. The steps in the risk quantification method are described below. 92 4.3.1 Identifying the Set of Primary Variables In carrying out analysis of environmental risks it is assumed that the accuracy of the estimation of the uncertainty of a decision variable can be improved by estimating it from a set of functionally related primary variables rather than by a direct estimation of its uncertainty. This assumption is discussed in detail in chapter 6. The specification of the decision variable in terms of a function of its set of primary variable is fundamental to the Monte Carlo approach (Rubinstein 1981) as well as to analytical approaches based on the moments of the variables. At a very general level, the constant dollar cost of a work package subject to risk types A l and A2 can be expressed as, Cost (C) = K x [ Mobilization cost + Labor cost + Material cost + Equipment cost + Subcontractor cost ] (4.13) where K = a factor which accounts for profit margin, overheads and market conditions At a more detailed level, an expression for the work package cost can be written in terms of variables such as resource usage levels, productivity of resources and unit cost of resources. 93 C = K\MnR+C 'Lf + CQQ + C, L (4.14) where, MOB = Mobilization expenses c L = Unit cost of labor Q = Work quantity PL = Labor productivity Q = Unit cost of material c E = Unit cost of equipment PE = Equipment productivity s = Subcontractor cost The work package duration (T) can be expressed as T = — (4-15) PL where, Q = quantity of the work L = resource usage level of a resource, R P = productivity of resource R (Output / Input) 94 Equations (4.13) to (4.15) and the set of primary variables that are included in the equations will be used to illustrate the quantification process. However, in general any function of the form Y = g ( X b X 2 , X 3 X„) can be used in the analysis. 4.3.2 Evaluating the Uncertainty of the Primary Variables The probability distributions that are used to represent the uncertainty of the primary variables are assumed to have the following characteristics. (a) The function, f(x) should be continuous on an interval [L,U], where L is the minimum value and U is the maximum value. (b) The function should be unimodal. Therefore, f(x) < f(m) for all x * m, where m is the modal value of x. (c) The end points should be non-negative, distinct and finite so that the variable is bounded and non-degenerate. These characteristics which have been identified by MacCrimmon and Ryavec (1964) for characterizing the duration of activities are applied to all the variables considered in carrying out environmental risk quantification. Such variables include cost, productivity, quantity and unit rates for resource usage. The uncertainty of the primary variables can be quantified using frequency data from past projects or with the aid of subjective estimates obtained from experts. In the event of frequency data being available, the data points are fitted to a standard statistical distribution using computer software available for curve fitting. The standard distribution obtained in this manner is used to represent the uncertainty of the primary 95 variable. In the absence of frequency data from previous projects, it is assumed that an expert is capable of providing accurate, subjective estimates for the percentiles of the probability distributions of the primary variables. However, the feasible number of subjective estimates that can be elicited from an expert is limited. The elicitation of a large number of percentile is infeasible in terms of the effort required by the experts as well as in terms of accuracy. The accuracy of estimates for percentiles that have little intuitive meaning would be minimal. Therefore, standard methods for distribution selection such as the maximum likelihood estimator method or the method of moments (Bury 1975, Law and Kelton 1991) cannot be used in the selection of a statistical distribution due to the lack of a sufficient number of data points (Mirkhani, Chitra and Lazur 1987). In such cases, the parameters of the distribution need to be estimated using available empirical formulae. A number of formulae that can be used to obtain approximate values for the mean and the standard deviation of a specific distribution type can be found in the literature. These include methods suggested by Farnum and Stanton (1987), Golenko-Ginzburg (1988), Keefer and Verdini (1993) and, Lau and Lau (1998) for the beta distribution. Robust estimators for the expected value, E[X] and the standard deviation, a of a general random variable X , have been derived by Pearson and Tukey (1965), utilizing the 2.5 th, 5 t h, 50 th, 95 t h and 97.5th percentile values. Let P2.5, P 5, P 5 0 , P 9 5 and P 9 7 5 denote the 2.5 th, 5 t h, 50 th, 95 t h and 97.5th percentile points respectively. The estimator suggested by Pearson and Tukey for the expected value is E(X)= P 5 0 + 0.185 A (4.16) 96 where A = (P95 + P 5)-2(P 5 0) (4.17) The estimator for the standard deviation has been suggested as the larger value of rj0.05 or C 0 . 0 2 5 , where °0.05 — P - P max{3.29-0.1(V. )2,3.08} / cr 0.05 (4.18) where '0.05 P - P r95 5 3.08 (4.19) and P - P 1 97.5 1 2.5 0.025 max{3.98-0.138(y . )2,3.66} / CT 0.025 (4.20) where ' 0.025 P -P 3.66 (4.21) 97 The accuracy of these estimators have been verified by Pearson and Tukey for the Pearson family of curves which encompass most of the commonly used standard statistical distributions. Verification has been carried out for the parameter ranges of Vp, = 0.0 - 2.0 and (p2 -p, -1) = 0.0 - 15.0. The parameters Vpb the skewness of the distribution and the kurtosis, p2 are defined as where p„ = n"1 central moment about the mean (n = 2,3,4). Assuming the estimates of the percentile values to be accurate, the error in the estimation in the mean is shown to be less than 0.1%, except in the cases of some Pearson curve J- and U - shaped distributions which are however not considered as being important in representing common physical phenomena. Similarly, in the case of the estimation of the standard deviation, it has been shown that the error is less than 0.5% over a very large part of the field of bell-shaped Pearson curves, which is the type of most interest in economic risk analysis. In the proposed quantification methodology, elicitation would be carried out to obtain the 5 t h, 25 t h, 50 th, 75 t h and 95 t h percentiles of the subjective probability distribution for each uncertain primary input variable. Chapter 6 describes the probability elicitation procedure used for this purpose. Although the use of the 2.5 th and the 97.5th values instead of the 25 t h and the 75 t h percentiles has the additional (4.22) (4.23) 98 advantage of being able to be used to calculate a more accurate value of the standard deviation using equation (4.20) for some types of Pearson curves, the 25 t h and the 75 t h percentiles have been retained for the proposed methodology. This is due to the fact that estimates obtained for extreme fractiles such as the 2.5 t h and the 97.5 th which are typically outside the realm of the experiences of experts are more likely to be erroneous than points further removed from the extremes (Keefer and Verdini 1993). th th In addition the use of the 25 and 75 points, which lie halfway between the median and the extremes provides a better definition of a curve than in the case when the median and four extreme points are used. Ranasinghe (1990) has used 7 percentile values, namely the 2.5th, 5 t h, 25 t h, 50 th, 75 t h, 95 t h and the 97.5th percentiles in carrying out a similar analysis. However, the use of more than 3 estimates has been considered as a shortcoming of an uncertainty quantification process by Lau and Lau (1998). An intermediate number of percentiles are used in the present analysis. The expected value and the standard deviation of the primary variable are obtained using equations (4.16) through to (4.19). The percentiles are then fitted to a distribution of the Pearson family of curves (Ord 1985). The use of Pearson curves in defining the uncertainty of the variables is necessitated due to the limited number of percentile points that are available. As the number of data points is insufficient in order to obtain meaningful results using methods such as the maximum likelihood estimator method, it becomes necessary to compare the data points to tabulated values of standard distributions. Although, tabulated values of distributions such as the normal, gamma, beta and student's t distribution exist they are neither in standard units nor are the parameters immediately comparable. Thus, several tables provided in different measures have to be referred in order to obtain an optimum fit. The Pearson family of 99 curves provides an attractive alternative in which data encompassing several distribution types are tabulated in a standard measure. The Pearson frequency curves are obtained from the solution of the differential equation \ dY = - ( x + c,) ( 4 2 4 ) Y dx c0 + cxx + c2x2 where ft(4A-3fl) 0 10/J 2-12/?,-18 V 5 M T ( A + 3 ) ( 4 2 6 ) 10/?2 -12/?, -18 ( 2 A - 3 A - 6 ) 2 10/J 2 -12/?,-18 when the origin for x is at the mean. The Pearson system encompasses most of the probability distributions used for applications in engineering. These include the normal, beta (Type 1), gamma (Type 3), exponential (Type 10), uniform, lognormal (Type 5), Student's t (Type 7), chi-square (Type 3) and F (Type 6) (Ord 1985, Ranasinghe 1990). Since the distributions commonly used in Monte Carlo simulation to describe input variables such as log-normal, beta and normal (Touran and Wiser 1992, Wall 1997) are included within the 100 Pearson system it is reasonable to assume that the subjective estimates obtained from experts can be represented by a curve of the Pearson system. Tables of percentage points of Pearson curves compiled by Johnson, Nixon and Amos (1963), and Amos and Daniel (1971) provide in standard measure the median and the lower and upper 0.25, 0.5, 1.0, 2.5, 5.0, 10.0 and 25.0 percentage points for different values of Vp\ and p2- The tables also provide in the same standard measure, the terminals of the distribution if they are not infinite. In order to fit the subjective estimates to a curve of the Pearson system, the standard deviate Xp of the estimates should be calculated in order to allow a comparison to be made with the standardized tabulated values. The standard deviate is defined in the following manner (Johnson, Nixon and Amos 1963). If Y = f(x) lx<x<l2 = 0 x < I and x > l2 (4.28) is a distribution of the Pearson system with (4.29) 101 The standardized deviate X p is defined as xP - E(x) (4.30) CT where x p is the value corresponding to the P m percentile of the cumulative distribution of Y. For the 5 t h, 25 t h, 50 th, 75 t h and 95 t h percentile values, the value of P will be 0.05, 0.25, 0.50,0.75 and 0.95 respectively. The standardized values are then compared to the corresponding tabulated values of the Pearson system of distributions. The "best fit" distribution is selected by minimizing the sum of squared deviations (Ranasinghe 1990). A maximum acceptable level of error can be expressed as a fraction of the standard deviation in order to ensure the consistency of the data with the Pearson family of distributions. The error (8) is expressed as where X T P is the tabulated value of X for a given P. In the case of 8 exceeding a maximum acceptable value specified by the analyst, the elicitation process will have to be repeated. Once the Pearson curve that best fits (4.31) the data and the corresponding values of Vpi and p2 are identified, the values of c0, c. 102 and c2 are calculated using equations (4.25) to (4.27). A large number of percentile points which corresponds to the particular curve are then obtained by numerically solving equation (4.24) for different values of x. The computer tool utilized for this purpose has been developed by Ahmed Abdel-Aziz as part of his Ph.D. work at the University of British Columbia (Abdel-Aziz 2000). His contribution in this regard is gratefully acknowledged. The set of values obtained from the numerical integration process as well as the end points are converted back into actuarial values using the relationship xP = XPo- + E(x) (4-32) These points are then fitted to a standard statistical distribution using computer software available for curve fitting in order to express the distribution in a format suitable for standard Monte Carlo software. The distribution which represents the percentile values most closely is chosen to describe the uncertainty of the primary variable. The graph of the probability density function obtained in this manner is used to perform a consistency check with the beliefs of the expert who provides the subjective estimates. 4.3.3 Computation of Rank Correlations Among Primary Variables Most modem software available to carry out Monte Carlo simulation make use of rank correlations in order to specify interdependence between variables. As 103 described earlier in this chapter, the main disadvantage of rank correlations is the difficulty in assessing a subjective value for the rank correlation coefficient in the absence of frequency data. However, an alternative method of modeling statistical dependence proposed by Cooke and Waij (1986) and later refined by van Dorp and Duffey (1999) provides an indirect method of obtaining the rank correlation between two interdependent variables. The method proposed by Cooke and Waij is based on the Copula of two interdependent variables. A Copula is defined as a bivariate distribution whose marginals are uniform on the interval [0,1] (Genest and MacKay 1986). Let X, and X 2 be two random variables with uniform marginal distributions that are limited to the range [0,1]. In the case of Xiand X 2 being independent, the joint distribution can be thought of as a uniform distribution of probability mass over the unit square in a Cartesian coordinate system with the possible values of X, and X 2 set out along the axes. In the event of the two variables being perfectly and positively correlated, the possible values of X, and X 2 will lie along the diagonal X 2 = X,. In the event of strict negative correlation the possible values of X, and X 2 will lie along the diagonal X 2 = 1 - X,. The Diagonal Band distribution has been used by Cooke and Waij (1986), and van Dorp and Duffey (1999) in order to model intermediate correlation between X, and X 2 . An example of such a distribution, when an intermediate degree of positive interdependence exists between the variables is shown in figure 4.1. 104 Figure 4.1. Diagonal Band Distribution with Parameter 9 The Diagonal Band distribution is generally characterized by the parameter, 0 which is equal to (1 - distance from the origin to the edge of the band along an axis). The case 9 = 1 reduces the diagonal band to a straight line and corresponds to perfect correlation. The correlation lessens as the band-width increases, finally enveloping the unit square when the correlation is zero, which corresponds to the case 9 = 0. The diagonal band distribution has been chosen to model the joint distribution of two uniform variables due to its efficiency in providing bivariate samples and due to the relative ease with which the parameter of the Diagonal Band distribution may be elicited from experts (van Dorp and Duffy 1999). Actual data values for correlated uniform distributions are found to closely mirror the shape of the Diagonal Band distribution. An example scatterplot shown in figure 4.2 containing 1000 actual data 105 points for two correlated uniform variables is used as an illustration (Iman and Davenport 1982). The rank correlation coefficient between the variables for this case is equal to 0.9. 4. 000 3.S00 3. 000 2.300 2. 000 1.300 1 . 000 .3000 0. 000 -.3000 -1 . 000 -1.300 -2. 000 -2.500 -3.000 -3.300 - I 1 1 1 1 -• r. :-V,.: ; \". : ' •• • ». : r : • i t . -4.000 -4 . -1 I I I 1 ' • ' -1 I 1_ 00 - 3 . 00 - 2 . 00 - 1 . 00 0. 00 1. 00 2. 00 3.00 Figure 4.2 - Correlated Values of Uniform Distributions with a Rank Correlation Coefficient of 0.9. Reprinted, by permission, from Iman, R.L. and Davenport, J.M. Rank correlations plots for use with correlated input variables. Communications in Statistics. (New York: Marcel Dekker 1982) 11. The probability density p(X,, X 2) of the Diagonal Band distribution shown in figure 4.1 is defined as (van Dorp and Duffy 1999), p (X„X 2 ) =(1-0)"' in areas A and C. = {2x(l-9)}"' in area B. 106 outside the diagonal band. (4.33) The probability density of the end areas A and C is twice that of the central area, B as indicated by equation (4.33). This ensures that the marginal distributions of Xiand X2 remain uniform within the unit interval. This can be illustrated as follows. Consider the instance when X, = 0. The probability density of X, at X, = 0 is equal to (1- 9) x (1 -9)"1 which is equal to unity. When X, is within the central area of the diagonal band, for example when X, = a, the density is obtained as 2 x (1- 9) x {2 x (1 - 9)}"' which is also equal to unity. Similarly a value of 1 is also obtained when intermediate values within the end areas are considered, which implies that the marginal distribution of X, remains uniform throughout the interval [0,1]. Similar results are obtained for values ofX2. A relationship between the rank correlation coefficient of the two variables and the parameter, 9 of the Copula has been developed by van Dorp and Duffy (1999). The rank correlation coefficient R has been shown to be related to 0 as R = 9 + 9 2 - 9 3 (4.34) Therefore, if two random variables are uniformly distributed on the interval [0,1] and if the value of 0 can be elicited from experts, then the rank correlation coefficient can be obtained using equation (4.34). However, in most cases random variables are neither uniformly distributed nor are they restricted to the interval [0,1]. In such an event, the value of the rank correlation coefficient can be obtained by considering the 107 transformation of the two variables into uniformly distributed distributions on the unit interval. Using standard distribution theory it can be shown that any absolute continuous distribution may be derived from the uniform distribution using an appropriate transformation and vice versa. This procedure is illustrated in Appendix B. Unlike the Pearson correlation coefficient, the value of which changes with the transformation of the variables (Iman and Davenport 1982), the rank correlation coefficient remains constant under transformations that produce strictly increasing distributions (van Dorp and Duffey 1999). The value of 9 can be obtained from experts by considering its association with a physical parameter. Consider a random value a of X , in the Diagonal Band distribution (see figure 4.1). In the absence of any correlation between X , and X 2 the range of values of X 2 associated with the value oc should range from 0 to 1. However, due to the correlation between the two variables the value of X 2 is limited to the range [p,q]. Therefore, if the value of X , is known with certainty, it would be possible to reduce the uncertainty associated with X 2 by {100 - (q - p)}%. Extrapolating this result to the total range of X , yields that the total percentage reduction in the uncertainty over the whole range of X , is equal to the area outside the Diagonal Band distribution. This area is equal 0 2 . However, as mentioned previously the distributions of the natural variables such as productivity and quantity of work will not be uniform in most instances nor will they be restricted to the interval [0,1]. In providing a value for the percentage reduction in uncertainty the experts will be thinking in terms of the natural distributions and their range of values. Therefore, the value experts provide as the percentage reduction in 108 uncertainty in such an event will not equal 9 2 in most instances. This is due to the fact that when the transformation of a uniform distribution to another distribution such as a beta distribution or a lognormal distribution is carried out, the ratio of the area to which the values are limited due to correlation, to the area enclosed by all the possible values would change in most cases. The ratio would remain constant in cases where the transformations along both axes are strictly linear. When the transformations are non-linear over the range [0,1] the ratio of the area within and outside the band would change as the height and width of the Diagonal Band distribution do not themselves remain constant over the interval [0,1]. On transformation back to the natural state, the set of possible values of the two variables will lose the uniformity of the Diagonal Band distribution and will be spread out in a complex fashion as shown in the example scatterplot in figure 4.3. This scatterplot portrays the relationship between a uniformly distributed variable with a normally distributed variable in an instance when the rank correlation coefficient is 0.9. Therefore, the values obtained from experts for the percentage reduction in uncertainty cannot be directly used in the analysis. However, this drawback can be overcome by transforming the value obtained for the percentage reduction to a corresponding value on the Diagonal Band distribution. This procedure is described below. 109 4.000 3.300 3. 000 7.300 2. 000 1.300 1 . 000 .3000 0. 000 -.3000 -1.000 -1.300 -2.000 -2.300 -3.000 -3.300 -4.000 -4 • ..'.••-'-'vr. 00 -3. 00 -2. 00 -1 . 00 0. 00 2. 00 3. 00 4. 00 Figure 4.3. Correlated Values of a Uniformly Distributed Variable and a Normally Distributed Variable with a Rank Correlation Coefficient of 0.9. Reprinted, by permission, from Iman, R.L. and Davenport, J.M. Rank correlations plots for use with correlated input variables. Communications in Statistics. (New York: Marcel Dekker 1982) 11. In the elicitation process described in chapter 6, the value for the percentage reduction in uncertainty is obtained by considering the range to which a variable is limited to when the exact value of another variable is known. For example in considering the interdependency between quantity and productivity, an expert would be asked, "to what range of values could you limit your estimate for the productivity if you knew that the quantity was exactly Q,? " Let the answer be "from d, to d2". As described in Appendix B, the values d 2 and d, can be transformed into their 110 corresponding values on the Band Width distribution using the transformation F(d,) and F(d2), where F is the marginal cumulative distribution function of the productivity. F(d,) and F(d2) will correspond to points such as p and q on the Diagonal Band distribution as shown in figure 4.1. In cases where a closed form equation for the marginal cumulative distribution does not exist, the corresponding probability values can be obtained by considering the set of data points generated during the process of obtaining the marginal distributions. As described in section 4.3.2, a large number of data points that correspond to various values of the cumulative probability are generated using numerical integration for each variable. Therefore, probability values that correspond to d, and d 2 can be obtained using values from the available data points that are equal or close to d, and d 2 and by carrying out interpolation where necessary. Therefore, the corresponding percentage reduction in uncertainty in the transformed environment can easily be obtained as [1 - (F(d2) - F(d,)}] / (1 - 0). This value can then be used in order to calculate the rank correlation coefficient between the two variables. In order to calculate the total percentage reduction in uncertainty of a variable, several values for the reduction should be obtained by considering the reduction at several equally spaced percentile values of the other variable. A value for the total percentage reduction, given by the area outside the diagonal band is obtained by multiplying the uncertainty reduction by the interval length and summing up the values. This process is detailed in Chapter 6. If the value obtained from an expert for the total percentage reduction in uncertainty is k then 111 x = e2 (4.35) Substituting this value in equation (4.34) yields R = X m + X - X m (4.36) The value of the rank correlation coefficient can be obtained for the analysis from equation (4.36). An example is used to illustrate the procedure to calculate the rank correlation coefficient described above. Consider the excavation and removal of contaminated soil from a work site and assume that the exact quantity of excavation is unknown. In analyzing the cost and duration of this work package it will be necessary to identify the degree of correlation between the quantity (Q) and excavator productivity (P). Assume that the calculation procedure outlined in section 4.3.2 has been carried out to establish the marginal distributions of P and Q. Let the marginal distribution of P be a lognormal distribution with parameters 27.25 and 4.03, and let the lower and upper bounds of the productivity be 20 and 35 mVhour. Let the quantity be triangularly distributed with parameters 3188,5710 and 7130 m3. In order to calculate the rank correlation coefficient, an expert is asked "To what values could you restrict your estimate of the range of the productivity if you knew that the value of the quantity is exactly 6000?" Let the answer be "Between 25 to 30 nf/hour". As a closed-form expression for the cumulative distribution function of the lognormal distribution does not exist, the probability values corresponding to 25 and 30 are estimated from the set 112 of cumulative distribution values obtained during the selection of the marginal distribution of the productivity as set out in section 4.3.2. A limited number of such values are shown in table 4.2. Using this set of values, the cumulative probability corresponding to productivity values of 25 and 30 are found to be 0.3 and 0.76 respectively. Therefore the percentage reduction in uncertainty is equal to {1 - (0.76 -0.3)}. The value of A, is therefore, equal to 0.54. The rank correlation coefficient can then be obtained using equation (4.36) by substituting the values of A.. In this case the coefficient is equal to 0.88. Table 4.2. Cumulative Probability Values of the Productivity Variable Value of Productivity (d) 22.34 23.82 24.96 25.97 26.96 27.98 29.12 30.51 32.55 34.33 Cumulative Probability 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.95 4.3.4 Obtaining the Probability Distribution of the Decision Variable In the final step of the quantification process a Monte Carlo simulation is carried out using the distributions of the primary variables and the rank correlations obtained in the previous steps as input data. The values generated for the decision variable are used to calculate its moments while the statistical shape of the distribution is obtained by fitting a standard statistical distribution to the histogram of the generated values. The statistical distribution or the moments of the generated data are used to represent the uncertainty of the decision variable. 113 4.3.5 Summary The steps in the quantification method for environmental risks of Types A l and A2 are summarized in this section. The methodology has been implemented utilizing the software MS Excel®, Mathcad®, BestFit® and @Risk® in a prototype computer system. Step 1 - The set of primary variables are identified by selecting a suitable decomposition method. Step 2 - Estimates for the 5 t h, 25 t h, 50 th, 75 t h and 95 t h percentile values are elicited from experts for each primary variable. Step 3 - The mean and the standard deviation are obtained for each primary variable using the Pearson and Tukey estimators given by equations (4.16) to (4.19). Step 4 - The percentile values are converted into standardized values using equation (4.30). The mean and standard deviation calculated in Step 3 are utilized for this purpose. Step 5 - The standardized values are compared to tabulated values of the Pearson curves. The curve which best fits the data is selected by minimizing the sum of squared deviations. If the sum of squared deviations is larger than the minimum acceptable error the analyst should return to step 2. A default value of 10% is used in the prototype system. The end points of the selected curve and the values of and p2 are obtained as outputs of this step. Step 6 - The values of Co, ci and C2 are calculated using equations (4.25) to (4.27) for the values of Vp, and p2 obtained in Step 5 and the value of a obtained in Step 2. A 114 large number of percentile points which corresponds to the particular curve can then be obtained by numerically solving equation (4.24) for different values of x. This operation which has been implemented in Mathcad® yields by default, a set of 44 data points corresponding to different values of the primary variable. Step 7 - The set of data points generated in Step 6 which corresponds to the Pearson curve selected in Step 5 are expressed in a format for existing Monte Carlo simulation software by fitting the data values to a standard distribution using the curve fitting software BestFit® (Palisade Corp. 1995). BestFit® uses maximum likelihood estimators in fitting a distribution to a data set. The use of a large number of data points which correspond to a Pearson curve, as well as the end points, ensures that the selected distribution is representative of the original data set. The expected value and the standard deviation calculated in Step 2 is compared with the parameters of the selected probability distribution in order to provide a consistency check. Step 8 - The selected probability distribution is displayed graphically to the expert who provided the probability estimates. The graphical capabilities of BestFit® are utilized for this purpose. If grave concerns regarding the statistical shape are raised, the analysis procedure is repeated starting from Step 2, with some or all of the estimates being adjusted in accordance with the new estimates made by the expert. Step 9 - Estimates for the percentage reduction in uncertainty due to knowing the exact value of the other variable is elicited from experts for each pair of variables. The rank correlation coefficient is calculated for each pair of variables using equation (4.36). Step 10 - A Monte Carlo simulation is carried out using the distributions for the primary variables obtained in Step 7 and the rank correlations obtained in Step 9. This 115 process is implemented using @Risk (Palisade Corp. 1997) and has default settings of Latin hypercube sampling and 100 iterations. Step 11 - The values for the decision variable generated as a result of Step 10 are imported into BestFit® in order to fit a standard statistical distribution. The standard statistical distribution which represents the uncertainty of the decision variable as well as its moments are obtained as results. 4.4 Conclusion Cost and duration uncertainty caused by environmental issues can be classified into 4 types in accordance with their effect on the work packages of the project. Of these, types A l and A2 can be analyzed in unison due to similarities in behavior. The quantification method adopted for risk types A l and A2 is based on Monte Carlo simulation. However, the proposed method overcomes a shortcoming of the Monte Carlo method by dispensing the need for the user to pre-specify a standard statistical distribution shape to describe the uncertainty of the input variable. This disadvantage is overcome by utilizing the Pearson family of frequency curves which encompass most of the commonly used standard statistical distributions. A method which allows the accurate use of rank correlations in Monte Carlo simulation is also presented as part of the risk quantification methodology. Rank correlation coefficients allow the interdependence between two random variables to be modeled effectively irrespective of their distributional form. However, in the absence of frequency data it has been difficult to subjectively estimate the value of the rank correlation coefficient due to the difficulty in physically conceptualizing its value. A 116 method based on the transformation of two interdependent variables into copulas is used to provide a physical meaning to the concept of rank correlation in the analysis. The use of the Pearson family of curves to identify the distribution of the primary variable and the methodology adopted in order to obtain an accurate estimate of the rank correlation coefficient is expected to contribute towards the effectiveness of Monte Carlo simulation in quantifying Type A l and Type A2 environmental risks. These can be considered as the major contributions of this chapter. 117 Chapter 5 Quantification of Type B Environmental Risks 5.1 Introduction As described in the previous chapter, Type B environmental risks are fundamentally different to environmental risks of Type A. Unlike work packages subject to risk Type A, the occurrence of a work package or the occurrence of an addition to an existing work package is itself an uncertain event in the case of Type B risks. An approach similar to the one adopted in dealing with Type A2 risks will be adopted in estimating the cost and duration associated with Type B2 risks. The cost and duration uncertainty of work packages subject to Type B2 risks will be estimated considering the work package in its totality allowing a common quantification method to be adopted for the treatment of the risk types B1 and B2. An example of a Type B l risk is the time and cost uncertainty associated with the clean up operation subsequent to an accidental fuel spill during marine construction. During the project appraisal stage, it would not be possible to predict with certainty that a clean up operation will have to be carried out during the project life cycle, as the occurrence of a fuel spill in itself is an uncertain event. Marine fuel spills could occur in association with marine construction activities as vessels and equipment used in marine construction such as dump scows, dredges and barges will have built-in fuel storage capabilities or will be equipped with self-dyked tanks. Fuel spills could occur as a result of a leak in these storage units, during the refueling process, or as a result of a marine collision (Strait Crossing Inc. 1992). Given that a 118 fuel spill occurs, the cost and duration of the clean up operation will depend on the magnitude of the spill. The magnitude of the spill may be such that it could be contained using on-site facilities, or its magnitude may be such that external assistance such as from the Coastguard will be required. The cost and duration of the clean up operation will also depend upon the ice condition of the area in the case of ice infested waters as the method of clean up is dependent on the ice conditions. For example, in the case of the area having less than 10% ice cover, containment booms could be deployed to recover the spilt fuel whereas in the case of heavy ice cover the fuel may have to be burnt using igniters. High winds and fast currents will also play a role in determining the cost of clean up. High winds and fast currents will hasten the dispersion of the fuel slick making clean up more difficult. Additional resources will also have to be expended by the contractors or the proponents if marine wildlife is affected by the fuel spill. If wildlife such as seabirds are affected by the spill, holding and cleaning facilities will have to be established for the oil-fouled birds. Therefore, it can be seen that the cost and duration of the fuel spill clean up work package is influenced by several external factors such as the ice condition of the waters, the wind and current conditions at that time, and the presence of marine wildlife in the area. An example of a Type B2 risk would be the cost and duration uncertainty associated with an excavation operation carried out on a site in which there is a likelihood of soil contaminated with hazardous waste being encountered. In this case, the excavation operation would be a part of the construction phase of the project even under normal circumstances. However, should contaminated soil be encountered during the excavation process, additional excavation may be necessary to completely remove the contaminated soil, while the excavated soil would also have to be treated 119 prior to disposal (Lens, Simon and Martin 1991). The productivity of the work force would also be affected by such a discovery as additional safety measures would need to be adopted including the use of protective clothing. In this example, the addition to the scope of the excavation operation would be contingent upon the discovery of contaminated soils within the site, thus making the occurrence of the addition itself an uncertain event. The scope of the addition would be further modified by external factors such as changes in regulations that govern the excavation and disposal of contaminated soils. Unlike in the case of a traditional work package, it is not possible for the cost and duration of work packages subject to Type B risks to be expressed as a function of the form Y = g ( X , , X 2 , X 3 X„) in terms of its uncertainty causing elements due to the presence of external factors. In almost all instances, mathematical relationships that predict the impact of external factors such as changes in regulations on project economic variables do not exist. The development of such relationships is hampered by: (i) the lack of physical principles that govern the relationship between economic variables and external factors; and (ii) the difficulty in generating empirical equations due to the lack of data and the uniqueness of each project. The ability to express the decision variable as a function of its primary variables in the form Y = g(X b X 2 Xn) is fundamental to moment based analytical risk quantification methods as well as to Monte Carlo simulation thus preventing their direct application to the analysis of Type B risks. Therefore, alternative methods need to be employed in instances where external factors affect project economic variables. The effect of external factors are best characterized using conditional statements which account for the different levels of impacts made upon the variable by the various states of the 120 external factor. Analytical tools such as influence diagrams can play a useful role in such analyses. 5.2 Influence Diagrams Introduced in the 1970s, the influence diagram is one of the most effective tools available for the representation and evaluation of decision problems (Howard 1988). The main advantage of this tool is that while providing a formal description of a problem for mathematical and computer analysis, it is also simple enough to be easily understood by people with a wide range of technical proficiency. Irifiuence diagrams can be used to specify the existence and the functional characteristics of relationships between deterministic or probabilistic variables (Howard and Matheson 1983). Essentially, an influence diagram is a directed acyclic network consisting of nodes and arrows. Four types of nodes can be used to represent various parameters and decisions in an influence diagram as shown in figure 5.1. Chance node Deterministic node Decision node Value node Fig. 5.1. Types of Nodes in Influence Diagrams 121 Chance nodes, depicted as circles or ovals are used to represent a continuous or discrete random variable, or a set of events. Deterministic nodes, which are a special case of a chance node depict parameters which depend deterministically on other nodes. Decisions are represented by square nodes while the quantity which is to be optimized is represented by a hexagon. An arrow entering a node implies that the probability assignment given to the node in the case of a chance node, or the decision made in the case of decision node is conditional on the node at the other end of the arrow (Howard 1988). Diagrams that contain only chance nodes and arrows are the most common form of influence diagrams encountered during risk analysis. Diagrams, which contain only chance nodes and arrows are alternatively called Knowledge maps (Howard 1989). Influence diagrams are a useful tool for representing and analyzing project risks and have been used in the development of several risk analysis models (Attoh-Okine and Ahmad 1995, Diekmann et. al. 1996, Jeljeli and Russell 1996, Pemg 1988). Pemg (1988) has utilized influence diagrams in developing an intelligent system for construction risk identification. In this case, influence diagrams have been used as an unambiguous language for representing uncertainties and their cause-effect relationships. Jeljeli and Russell (1996) have developed a model based on the influence diagramming method that estimates a contractor's liability in carrying out environmental clean up projects. Influence diagrams have been used to model the effects of a comprehensive set of factors that are likely to have an impact upon the contractor's liability. An influence diagramming approach can be particularly useful in analyzing uncertain events which are in turn affected by other uncertain factors or events, such as 122 is the case of Type B environmental risks. The use of influence diagrams in risk analysis as well as their advantages and disadvantages are illustrated by considering the example of an accidental fuel spill during marine construction. An influence diagram, which can be used to evaluate the cost uncertainty surrounding a marine fuel spill clean up, is shown in figure 5.2. In this case the cost of the marine fuel spill clean up will depend upon whether the spill actually occurs, and in the event of it occurring, its magnitude. / Cleanup cost \ / Fuel spill Fig. 5.2. Influence Diagram for Evaluating Fuel Spill Clean up Cost This influence diagram implies that the analyst assigns a probability distribution to the cost (C) of the fuel spill clean up work package conditioned upon the knowledge of the state or value of the fuel spill (S) node. In mathematical notation this is P(C | S, £), the conditional probability of the cost given the state of the fuel spill. Examples of the state of the spill (S) includes "major oil spill has occurred" and "5,000 gallons of oil has spilled". S represents the total life experience of the analyst. The corresponding probability tree for the problem is shown in figure 5.3, assuming that three states of the cost will be specified for each risk event. Two distinct 123 states of spills, Major and Minor have been considered. Major spills can be construed as an event where more than X gallons of oil is spilled, whereas the spill would be minor if less than X gallons is spilled. The value of X would need to be set by the analyst taking into consideration factors such as (a) the magnitude of a spill that can be handled using project resources alone without the aid of external organizations, (b) regulations that require spills above a certain quantity to be reported to relevant authorities, and (c) quantities of spills that would require other activities of the project to be temporarily halted. An upper bound on X would be placed by the maximum possible amount of oil that is likely to be spilled indicated by the capacity of oil storage tanks or the capacity of the fuel tanks of marine vessels. Major Spill ^ ^ P 4 CA CB 7.P.1 /~i Minor Spi l l ~ ~ P 5 P6 P 7 C c C D C E \A P2 P8 P9 CF No Spi l l \ P3 Q Cost = p Fig. 5.3. Probability Tree for Evaluating Fuel Spill Clean up Cost 124 In assigning values to the probability tree, either the cost values (CA to C F) or the conditional probability values (P4 to P9) could be assigned subjectively while keeping the values of the other parameter at fixed levels preset by the analyst. In the "probability method" the probability values are fixed at predetermined levels and the corresponding cost values are assessed while in the "value method" probabilities are assessed for fixed values of the cost (Merkhofer 1987). An example would be a case of assigning the values $ 10000, $ 50000 and $ 90000 to C A , C B and C c respectively and thereafter assessing the values of P 4, P 5 and P 6 . This example corresponds to the "value method". A study carried out by Hora, Hora and Dodd (1992) has indicated that these two methods can be considered to be equivalent in terms of the results generated. In either case, the probability assignments given to the probability tree should satisfy the axioms of probability. The axioms are summarized below. Axiom 1 The probability of an event is a number greater than or equal to zero but less than or equal to unity: 0<P[A]<1 (5.1) Axiom 2 The probability of the certain event S is unity: P[S] = 1 (5.2) where S is the event associated with all the sample points in the sample space. 125 Axiom 3 The probability of an event which is the union of two mutually exclusive events is the sum of the probabilities of the two events: P [ A u B ] = P[A] + P[B] (5.3) In order for the probability axioms to be satisfied for the case under consideration: 0<pi<l fori =1,2...9 (5.4) and Pi + Pi + Ps = P4 + Ps + Pe = Pv + Ps + P9 = 1 (5.5) Once the prior probabilities for S, and the conditional probabilities for C are assigned, the joint probability P (C,S , S) can be obtained using the multiplication rule of probability. P(C,S ,S) = P ( C | S,S) x P(S | 8) (5.6) Evaluation of the probability tree using the assigned cost and probability values will yield the probability distribution of the oil spill clean up cost. Now consider the more complex case when the analyst believes that ice conditions and wind conditions play a major role in determining the cost of clean up. The influence diagram for this case shown in figure 5.4 asserts that the analyst will assign a probability distribution to the cost conditioned upon the knowledge of the state of the fuel spill, the condition of the wind at that time, and the ice conditions at 126 that time. It also implies by the absence of arrows between the wind, ice and spill nodes, that these events have been considered to be independent of each other. It is noted that the development of an influence diagram such as the one shown in figure 5.4 is a subjective process that would depend upon the knowledge and perceptions of the experts and the analysts. In some instances, different experts and analysts may identify different sets of external factors and relationships which would lead to diagrams that are different from one another for the same problem. In evaluating the influence diagram in this example, prior probabilities are first assigned to the ice, wind and spill nodes. Conditional probabilities for the cost can then be assigned according to the different scenarios. Figure 5.4. Influence Diagram for Evaluating Fuel Spill Clean up Cost in the Presence of External Factors 127 A probability tree in which two states each of wind and ice conditions are considered is shown in figure 5.5. Cost A No Spill Figure 5.5. Probability Tree for Evaluating Fuel Spill Clean up Cost in the Presence of External Factors 128 Using the multiplication rule of probability, P(C,W,I ,S ,S) = P ( C | W,I,S,S) x P ( W | I,S,S) xP(I | S,S) xP(S | S) (5.7) However, I,S and W have been considered to be independent events in this case. By definition, two events A and B are said to be independent if and only if A major advantage of the influence diagramming method is that different experts can be used to assign probabilities to the different nodes in the diagram. In the scenario under consideration, a meteorologist might assign the probability of winds being severe during the time marine construction is taking place while a marine construction expert might assess the probability of a spill occurring. Now consider the case when the analyst decides that some of the conditioning variables are not independent of each other. For example, the analyst might consider high wind conditions cause accidents during marine construction and therefore W and S are dependent on each other. The corresponding influence diagram is shown in figure 5.6. In this case the probability distribution for S should be assigned with P ( A | B ) = P(A) (5.8) It follows that, P(C,W,I ,S ,S) = P ( C | W,I,S,S) x P ( W | 8) xP(I | S) xP(S | S) (5.9) 129 Figure 5.6. Influence Diagram for Evaluating Fuel Spill Clean up Cost in the Presence of Interdependent External Factors knowledge of the state of W. Therefore, the conditional probability of "a spill occurring given the wind condition" should be assessed. However, there might exist some cases in which the expert might be more comfortable assigning a distribution to the wind condition based upon knowledge of the state of the spill. Influence diagrams offer certain flexibility regarding such preference of the assignment order as shown below. From Bayes' Theorem, for two events A and B, P(A | B) x P(B) = P(B | A) x P(A) (5.10) 130 Considering the two events, S and W it can be written P(S | W,S) x P(W | S) = P(W | S,S) x P(S | S) (5.11) This implies that the assessment order can be reversed provided that both W and S are based on the same set of information S. Therefore it follows that the arrow between W and S can be reversed. This illustrates another advantage of the use of influence diagrams in risk analysis in that, with suitable manipulation of the diagram probability assignments can be made in a sequence more comfortable for the experts providing the information. However, the main drawback of the influence diagramming method is that the number of probability assignments that have to be made becomes very large as the number of conditioning variables increase. In the example considered (see figure 5.5) experts will have to provide 24 conditional probability values of the clean up cost, as well as additional assignments for the prior probabilities of the conditioning events. Additionally, when the external factors are not independent of each other, conditional probabilities would have to be assessed in order to assess the probability of various conditioning events, as implied by equation (5.7). In such a case, reduction of the problem to the much simpler state depicted in equation (5.9) is not possible. Assessment of conditional values could also become complicated if the number of conditioning factors increase. An example situation in which the number of factors increase further would be when the analyst decides to take into account the possibility of wildlife such as sea birds being contaminated by the fuel oil and the clean up cost 131 being affected as a result of the need to treat and rehabilitate the birds. In such an event, experts will find it difficult to evaluate the combined effects of the large number of conditioning variables. Therefore, it may not be possible to make realistic assignments for the conditional probabilities of the cost values as the dimensionality of the problem increases. Another drawback of using influence diagrams for risk analysis is that it is difficult to decompose the decision variable into its primary variables with the aim of achieving higher accuracy. Since it will be necessary to obtain several probability values of the primary variables for each combination of events, the task of decomposing the decision variable would be impractical in terms of the number of subjective values that would have to be assigned. 5.3 Requirements for the Analysis of External Risks The above section has illustrated the use of influence diagrams in isolation, or what may be termed as the "straightforward use" of influence diagrams in the analysis of economic risks due to external factors. The approach has several advantages such as the ability to portray the different states of external factors and its ability to facilitate and simplify the probability assessment process. However, several disadvantages were also identified. These include: (i) The approach is extremely data intensive. (ii) The combined effect of several external factors need to be considered simultaneously in making probability assignments to the target variable. This would become difficult as the number of external factors increase. 132 These disadvantages hamper the straightforward use of influence diagrams in analyzing risks due to external factors. Therefore, the need for a quantification methodology that retains the ability to model the effect of external factors efficiently remains as it cannot be satisfied by the use of influence diagrams in isolation. This section identifies the characteristics that such a methodology should possess in order to carry out effective analysis of Type B risks. Several general characteristics that can be considered to be desirable in an ideal quantification methodology dealing with Type B risks are: (a) Have the ability to model the various distinct states that an external factor might attain. The impact brought on by the external factor would depend on the state of the factor. Therefore, it is important to have an ability to model the various states of the factor. An example of the distinct states of an external factor is provided by the presence of wildlife in the example of an oil spill clean up operation. The wildlife that may be present in the area might be limited to birds during certain months of the year, while in other cases marine mammals such as whales may be affected if the spill occurs during the migration season of such species. Therefore, it would be necessary to consider 3 distinct scenarios in analyzing the problem. These would be a scenario in which a significant impact is not made on any wildlife, one in which an impact is made upon a limited number of wildlife species such as birds, and one in which a major impact is made upon wildlife, including impacts upon species of significance such as whales. Each of these states would have varying degrees of impact on the cost and duration uncertainty of the oil spill clean up work package. 133 (b) Allows the analysis of the problem at primary variable level in instances where such an analysis is considered to be useful. As will be discussed subsequently in chapter 6, the decomposition of decision variables into their primary variables is widely regarded as a method of improving the accuracy of the estimation process. Often, more informed assessments of an uncertain variable can be obtained by disaggregating a variable into its component variables (Chapman and Ward 1997). Decomposition also allows the analyst to develop a link between the decision variables of cost and duration. Therefore, it is important that a quantification methodology developed with the aim of quantifying the impact of external factors allows the analysis to be carried out at primary variable level. (c) Have the ability to accurately model the impact of a specific state of an external factor on the variable under consideration. Under circumstances in which external factors are not considered, the value of the variable would be expressed as a probability distribution, or as a deterministic value. The methodology should be able to accurately model the impact the external factor has on the probability distribution or the deterministic value that describes the variable. The impact on a probability distribution in general, can be described in terms of the effect on the location, scale and shape of the probability distribution. Ideally, the modeling process should be able to portray the impact of the external factor on these 3 characteristics of the probability distribution in the case of the variable being stochastic. An example would be the instance in which the impact of extreme wind conditions on the labor cost of an oil spill clean up operation is being evaluated. Consider, that the cost is distributed triangularly with parameters (1000,2000, 3500) in the absence of external factors such as high winds. Ideally, it should be 134 possible to predict whether: (1) The distribution is merely shifted along the value axis due the effect of the external factor, thereby becoming a triangular distribution with parameters (1100,2100, 3600); (2) The distribution is scaled up (or down) thereby becoming a triangular distribution with parameters (1250,2500,4375); (3) changes shape and becomes a lognormal distribution with parameters (2170,620) leaving the mean and standard deviation relatively unchanged; or (4) undergoes a combination of the 3 changes described previously. There also exists a possibility that an external factor may create a need for a variable to be analyzed stochastically, even though it could be treated deterministically under conditions when external factors are inactive, and vice versa. A methodology developed for the treatment of uncertainty caused by external factors should also be capable of handling such a transformation. (d) Have the ability to accurately model the simultaneous impact of several external factors on the variable under consideration. Instances can arise when several external factors need to be considered in unison. In the example of an oil spill, high winds as well as thick ice cover may exist at the time of the spill creating a need to be able to quantify the simultaneous effect of external factors on a variable such as labor cost. As in the case of a single external factor, an analyst should ideally be able to identify the impact of the multiple factors on the location, scale and shape of the probability distribution of a stochastic variable. Additionally, it should also be possible to model any transformations that may occur between the stochastic and deterministic domains as in the case of a single external factor. (e) Have a minimum data requirement. The amount of data that is required as well as the difficulty in obtaining such data should be a minimum. 135 5.4 External Risk Quantification Methods based on Influence Diagrams As identified in section 5.2, the straightforward use of influence diagrams in analyzing risks due to external factors has several disadvantages. This section describes an existing risk quantification methodology introduced by Diekmann et al. (Diekmann and Featherman 1998, Diekmann 1997, Diekmann et al. 1996) that has attempted to overcome the disadvantages of using influence diagrams in isolation. The degree of concordance of the method proposed by Diekmann et al. with the characteristics identified in the previous section, as well as the disadvantages of the method are also discussed in this section. Several risk quantification models have been developed by Diekmann et al. (Diekmann and Featherman 1998, Diekmann 1997, Diekmann et al. 1996) that integrate the use of influence diagrams and Monte Carlo simulation with aim of analyzing external risks. Of these models, the fixed probability factor model (Diekmann and Featherman 1998) which is empirical in nature has been identified as being the model which yields forecast values closest to the actual values, based on the results obtained from three environmental restoration projects. An example used by Diekmann and Featherman (1998) is used to illustrate the use of the fixed probability factor model. Consider the case when the excavation cost related to an environmental remediation site is affected by the external influences of an uncertain waste quantity and design changes. The influence diagram and the associated probability tree considered by Diekmann and Featherman for this case is shown in figures 5.8 and 5.9 respectively. 136 Fig. 5.7. Influence Diagram for Evaluating the Excavation Cost in an Environmental Restoration Project. Source: Adapted from Diekmann, J.E. and Featherman, D.W. 1998. Assessing cost uncertainty: Lessons from environmental restoration projects. Journal of Construction Engineering and Management. 124 (6), 445-451, figure 7. Waste Quantity A s Expected (Ps. P50, P95) Wo Design Chanqes / Quantity More FA* (P«, P50, P95) Than Expected (Cost factor = F2) \ FB " (Ps, PJO, P95) Design Changes (Cost factor = Fi) F 0 » (P5 .Pso .P95) Fig. 5.8. Probability Tree for Evaluating the Excavation Cost in an Environmental Restoration Project. Source: Adapted from Diekmann, J.E. and Featherman, D.W. 1998. Assessing cost uncertainty: Lessons from environmental restoration projects. Journal of Construction Engineering and Management. 124 (6), 445-451, figure 7. 137 In the first step of the risk quantification process, the uncertainty of the excavation cost is evaluated using Monte Carlo simulation by decomposing the decision variable into its primary variables such as excavation quantity, equipment cost and equipment operator wage rate. In carrying out this step of the analysis it is assumed that the two external uncertainties are inactive. This is considered as the base case for the analysis. The 5 t h, 50 th and 95 t h percentile estimates obtained from the simulation are considered as the base percentile values of the excavation cost and are incorporated into the branch of the probability tree which corresponds to the case of both the external factors being inactive. Deterministic multiplicative factors termed as Cost Factors, shown in figure 5.9 as F A , F b and F c have been defined in order to quantify the percentage change that a nominal cost could experience as the result of external uncertainty factors. It has been assumed that the cost in the presence of external uncertainties is given by the multiplication of the original distribution by the cost factor. The values of the cost factors would be elicited from experts in order to carry out the analysis. For elements that are affected by more than one external factor, an adjusted product function has been developed to combine the multiple cost factors. The adjusted product function (F) has been defined as F = [F ,xF 2 xF 3 Fn]{i.o-(o.2x(n-i))} ( 5 1 2 ) where F, = factor between the conditioning variable and the decision variable n = number of conditioning variables 138 The derivation of the adjusted product function has been based on an empirical rule in estimating known as the "Six-tenths rule". The base estimates for the 5 t h, 50 th and 95 t h percentile values are then multiplied by the corresponding cost factors and incorporated into the probability tree. In the final step of the analysis process, the probability tree is evaluated, which yields a distribution for the excavation cost for the example considered. The method developed by Diekmann et al. presents an important advancement in the ability to take into account external uncertainty factors in a simplified or aggregated way in the economic risk analysis of projects. However, it is noted that while the method has appeal due to its simplicity, it lacks a solid intellectual underpinning. The method satisfies some of the general requirements set out in the previous section that can be considered as desirable ingredients in an effective methodology designed to quantify the effects of external factors. The conditions that have been satisfied are: (i) The methodology proposed by Diekmann and Featherman reduces the information requirement needed for the analysis by using cost factors to model the change in probability distributions instead of re-deriving the distribution of cost for each combination of external factors using percentile estimates or by using the distributions of the primary variables. Therefore, a single value (the Cost Factor) is used in place of several probability estimates, thus reducing the information that is required. (ii) The methodology allows various states of an external factor to be modeled. A distinct cost factor can be assigned for each state of the external factor that needs to be considered. 139 (iii) The impact of multiple factors acting in unison can also be modeled with the use of cost factors. The adjusted product function allows the calculation of cost factors in cases where multiple external factors are active. However, several shortcomings of this process can also be identified. These include: (i) The use of the adjusted product function in developing the fixed probability factor model can be considered as being arbitrary. The adjusted product function used by Diekmann et al. is based on the six-tenths rule, a rule of thumb commonly used by cost estimators in the construction industry (Mular 1982). The six-tenths rule is applied in such a manner that if the cost C A of a work package having a quantity Q A is known, then the cost for a similar work package of quantity Q B will cost C B where The mathematical form of the adjusted product function given as F = [F, x F 2 x F 3 F n ] { 1 0 " ( 0 2 x ( n" 1 ) ) } based on the six-tenths rule has several flaws. Although, the adjusted product function is intended to be used in a case when the decision variable is influenced by more than one external uncertainty (Diekmann and Featherman 1998), the functional form used for the development of the fixed probability factor model given in equation (5.12) produces an adjusted value of the factor even in the case of one external uncertainty being active. For 140 example if F, = 1 and F 2 = 2 in the example shown in figure 5.9, it would correspond to a situation in which only the waste quantity has changed. However, the application of the adjusted product function to this situation would yield a value of 1.74 (Diekmann and Featherman 1998), whereas by definition, the cost should increase by a factor of 2 when subject only to waste quantity changes. Since an expert in making an assessment of the percentage change of the cost would already be taking into account rules of thumbs such as the six-tenths rule in making the assessment, application of the formula based on the six-tenths rule in instances where only one factor is active would lead to a duplication of the adjustment of the cost. Additionally, should the number of external factors be equal to 6, the power term in the adjusted product function would be equal to zero. This would create an effect whereby the value obtained by the use of the adjusted product function would always be equal to unity irrespective of the values of the individual factors. This illustrates further undesirable behavior of the adjusted product function. Therefore, even though the use of Cost Factors allows the effect of external factors to be modeled, the accuracy of the modeling process would be gravely affected by the flaws in the adjusted product function, (ii) As stated previously, a probability distribution of a random variable could theoretically undergo changes in location, scale and shape due to the effect of an external factor on the values of the variable. However, the use of a cost factor only enables the shifting of the limits of the distribution along the value axis while the distribution shape is retained. The inability to model any changes in distribution shape that may occur can be considered as a shortcoming of the use of a single 141 value (the Cost Factor) to model the effect of the external factor on a random variable. Additionally, any transformation of the variable from a stochastic domain to a deterministic domain and vice versa, that may occur as a result of the external factors cannot be modeled with the use of cost factors, (iii) In the probability factor model, the base estimates are derived from the primary variables using decomposition. However, in all other cases when the external uncertainties are active, the cost factors are defined taking into account only the change in the derived variable. Therefore, decomposition has only been considered when the distribution of the derived variable is obtained for the base case. Based on the premise that decomposition improves the accuracy of estimates, it can be concluded that the estimates obtained for the instances when the external uncertainties are active, are less accurate than the estimates obtained for the base case. A method of quantifying Type B risks that aims to overcome the shortcomings of the probability factor model and the straightforward influence diagramming approach is presented in the following section. The method that is presented is one of the many approaches that can be considered as a candidate for use in the quantification of Type B risks. Another approach is a methodology in which the environmental impacts of a work package are tied to the cost and duration of the work package. As an example, the cost of an oil spill clean up may be tied to the probable number of seabirds affected by the spill and expressed as "Unit cost per bird x Probable number of birds". Distributions for the unit cost and the number of birds can then be assessed for each scenario. However, all such approaches have individual advantages and 142 disadvantages. The method proposed in the following section has been selected due to its ability to complement the analysis procedure for Type A risks. The method described in the following section allows; (a) decision variables to be decomposed in a manner similar to the risk quantification methodology for Type A risks; (b) correlation among the primary variables to be accommodated; and (c) obtain a large number of sample points for the decision variable using Monte Carlo simulation, that can be used to calculate the first four moments. Therefore, the proposed quantification methodology bears close resemblance to the methodology described in Chapter 4 for Type A risks, particularly the level of detail at which the analysis is carried out. Therefore, the two methodologies are ideally suited for use in the same environmental risk analysis framework as the types of input, output and the level of detail of the analysis are similar. 5.5 Quantification Method for Type B Risks As described earlier in this chapter, influence diagrams provide a means of incorporating the uncertainty due to external factors in the analysis of Type B risks. However, in using influence diagrams in isolation, it is not feasible to take advantage of the improvement in accuracy and the increase in simplicity brought about by decomposition. This is due to the large number of subjective probability assignments that would have to be made in such an event. A straightforward influence diagramming approach would also require the elicitation of probability estimates that are conditioned by multiple factors. This section introduces a risk quantification methodology for Type B risks that avoids the need to make complex conditional estimates, while simultaneously utilizing the advantages offered by decomposition. 143 This methodology derives its motivation from the work carried out by Diekmann et al. in developing risk quantification models that marry influence diagrams with Monte Carlo simulation. The following section presents a brief overview of the quantification methodology. This is followed by a detailed description of each of the individual steps and an example that shows the application of the quantification methodology. 5.5.1 Outline of the Quantification Methodology The methodology utilizes influence diagrams for the representation of the relationship between a decision variable and the external factors that affect its outcome. Therefore, the first task in the analysis is the development of an influence diagram that shows external factors that are likely to have a major impact on the cost or the duration of the work package under consideration. This is followed by the derivation of the probability tree that corresponds to the influence diagram. Thereafter, prior probabilities in the event of the external uncertainties being independent, or conditional probabilities in the event of the externalities being dependent are obtained from experts and assigned to the branches of the probability tree except to the final level of nodes which depict the decision variable. In the next stage of the analysis, decomposition of the decision variable into its primary variables is carried out by selecting a suitable decomposition method. The base case of the problem, i.e. when the external factors are inactive, is considered initially in assigning probability distributions, similar to the probability factor model introduced by Diekmann and Featherman. Probability distributions are assigned to each of the primary variables using frequency data or subjective estimates as set out in section 4.3.2 in this thesis. 144 The impact of an external factor on a variable is modeled by a multiplicative factor as in the probability factor model. However, multiplicative factors will be obtained for the primary variables instead of the decision variable in carrying out the analysis. The multiplicative factors will first be assessed considering external uncertainty factors individually. Thereafter, external uncertainty factors will be considered in a pair-wise manner. An example of a pair-wise combination of external uncertainties would be "when both high winds and thick ice cover is present". Multiplicative factors are obtained for each pair of external uncertainties. Multiplicative factors for events that involve a combination of 3 or more external uncertainties are calculated employing a procedure that is described in detail in the following section of the thesis. The assignment of multiplicative factors would yield several sets of primary variables each with distinct multiplicative factors, that correspond to each end node of the probability tree. Part of such a probability tree is shown as an example in figure 5.9. In the next stage of the analysis, Monte Carlo simulation is carried out at each branch of the probability tree. This yields a set of values obtained as the simulation output for the decision variable at each end node of the probability tree. This set of values and the event probabilities for each branch of the probability tree are used to calculate the first four moments of the decision variable. 145 Ice Cover > 10% / P = 0.3 O Labor cost = 0.8 x lognormal (600,25) Material cost = 0.5 x Triangular(2000,3000,4200) Equipment cost = 0.75 x lognormal(4000. 650) High Winds / p = 0.2 Ice Cover < 10% p = 0.7 o Labor cost = 1.3 x lognormal (600,25) Material cost = 1.4 x Triangular(2000,3000,4200) Equipment cost = 1.3 x lognormal (4000,650) Fig. 5.9. Assignment of Multiplicative Factors to Primary Variables Each of the stages identified above are described in detail in the following section. 5.5.2 Derivation of Influence Diagram for the Problem The first step of the analysis of Type B risks is the preparation of an influence diagram that shows the relationship between the external factors and the target variable. In some cases there might be uncertain events which influence the external factors themselves leading to an influence diagram having three or more levels of nodes as shown in figure 5.10. The process of deriving the influence diagram for the problem using information obtained from experts is described in chapter 6. Thereafter, the corresponding probability tree is drawn up and probabilities assigned to each branch except to the final level of nodes which depict the decision variable. In the case of independent factors, only the prior probabilities corresponding to each state of the individual factors need to be assessed. In the event of dependency between factors, conditional assessments will be required. 146 Fig. 5.10. Influence Diagram with a Hierarchy of Chance Nodes The branch which corresponds to all external factors being inactive is considered as the base case. However, any scenario can theoretically be considered as the base case. Therefore, in the absence of a scenario in which all the external factors are inactive, another scenario may be chosen as the base case. In the example of a fuel spill clean up, the scenario in which a minor spill occurs when thick ice and high winds are not present may be considered as the base case. 147 5.5.3 Assignment of Probability Values to the Primary Variables A suitable decomposition method is chosen to model the decision variable as in Step 1 of the analysis of Type A risks. This leads to an identification of the primary variables that need to be considered in the analysis. Probability estimates for the 5 t h, 25 t h, 50 t h, 75 t h and 95 t h percentiles are elicited from experts for each of the primary variables considering only the base case scenario of the problem. The uncertainty of each primary variable is then quantified adopting the method described for risk type A. The correlation coefficients are assessed as described in section 4.3.3 of the previous chapter. 5.5.4 Assignment of Adjustment Coefficients In order to utilize the benefits offered by decomposition, individual factors which account for the percentage change in the value of the primary variable relative to the base case due to the effect of externalities will be defined for each primary variable. These will be termed as adjustment coefficients instead of Cost Factors as used by Diekmann and Featherman (1998), in order to avoid confusion with the term, external factor. It will be assumed that the change in the distribution of a variable due to an external factor can be modeled by multiplying it with a suitable coefficient. As in the probability factor model introduced by Diekmann and Featherman (1998) the multiplication of the distribution by a coefficient will result in the expected value and the standard deviation of the base case distribution being increased by a factor equal to the coefficient. However, the shape of the distribution would remain unchanged. The inability to model any changes in the statistical shape of the distribution may be considered as a disadvantage of using a single point estimate to model the effect of 148 external factors. A more detailed analysis would require the adjustment coefficients to be expressed as a distribution function instead of a single point estimate. This would be impractical in terms of the detailed information that would be required. Additionally, this approach also cannot model the transformation of a deterministic variable into a probabilistic variable due to the effect of external factors, or vice versa. However, the correlation matrix between the primary variables will remain unchanged subsequent to the application of coefficients since the value of the rank correlation coefficient is invariant under transformations that produce strictly increasing distributions. The probability distributions that are obtained by multiplying the base distribution by a coefficient is simply a linear transformation of the original distribution. Therefore, the same correlation matrix can be used in all scenarios pertaining to the problem. This can be considered as an advantage of the approach. The example of the fuel oil spill clean up can be used to illustrate the advantages of using primary variables instead of derived variables in the analysis. If the decomposition of the clean up cost is carried out on a general level as given in equation (4.13), the clean up cost can be expressed as Clean up cost (C) = [ Mobilization cost + Labor cost + Material cost + Equipment cost + Subcontractor cost ] (5.14) In this case overheads, profit margins and market conditions have been neglected for simplicity. The subcontractor cost in the event of a fuel spill would be the cost of reimbursing the Coastguard and other agencies for their efforts in controlling the spill. 149 If the probability factor model was used, the distribution of the clean up cost would first be evaluated using Monte Carlo simulation for the base case corresponding to a minor spill, when the effects of wind and ice conditions are non-existent. In obtaining a distribution for the case of a minor spill with ice cover being more than 10%, the cost distribution for the base case would merely be adjusted by a factor, Fj. The benefit of improved accuracy brought on by decomposition is not utilized in making such an assessment. The expert providing the factor Fj would have to collectively assess the impact of the ice cover being more than 10% on the individual primary cost components and provide an estimate of a factor which accounts for this collective impact. An approach whereby the expert could assess the individual impact of the ice condition on the individual cost component is likely to be more accurate than a collective estimate as the complexity of the mental computation the expert would be required to carry out is reduced. Therefore, the use of individual coefficients that take into account the change in the primary variables due to the external uncertainties will be obtained and used in the proposed analysis method. However, the need for a function that takes into account the combination of several external uncertainties remains. A method of incorporating the effect of several external uncertainties acting in combination on a variable is necessary in order to assess the percentage change in the values of the variable due to two or more external uncertainties which act upon it at the same time. For the case when the incremental changes caused by two or more external uncertainties are mutually exclusive, the combined effect is given simply by the addition of the percentage changes. This can be illustrated as follows. 150 If (FA - 1) is the percentage change expected in the number of personnel needed for an activity S due to an external uncertainty A, and if (FB — 1) is the percentage change expected in the number of personnel needed for S due to an external uncertainty B, and if s was the original number of personnel when A and B were inactive then, The change in personnel due to A only = s x (FA - 1) (5.15) The change in personnel due to B only = s x (FB - 1) (5.16) If the increases in usage of the personnel due to A and B are mutually exclusive, then Personnel needed when both A and B occur = s x (FA - 1) + s x (FB - 1) (5.17) Therefore, Percentage change in S due to both A and B = (s x (FA -1) + s x (FB -1)) / s = (F A -1) + (F B -1) (5.18) However, in general the incremental effects might not be mutually exclusive. A new modeling paradigm is introduced in this thesis which enables the analysis of the combined effects of multiple external uncertainties. This modeling paradigm enables 151 the combined effect of the external uncertainties to be quantified irrespective of their degree of mutual exclusiveness. Modeling paradigm for the analysis of the combined effects of multiple external uncertainties Consider a case when thick ice conditions are present during the clean up of an accidental oil spill. Assume that the base case scenario constitutes of low wind conditions and an ice cover of less than 10 percent. Due to the thick ice conditions, the value of a variable such as labor productivity would decrease relative to the base case as clean up crews based on boats would find it difficult to maneuver around the ice. If high wind conditions were present instead of thick ice, the productivity would also decrease relative to the base case due to difficulties in navigation that arise due to high winds. Therefore, it can be seen that both high winds and thick ice conditions decrease productivity when considered individually. However, in a situation when both winds and ice are present the decrease in productivity might not be equal to the sum of the individual decrease amounts. This is due to the fact that in some cases the ice would have a moderating effect on the waves induced by high winds thus making navigation easier than in a case when high winds are present in isolation. Therefore, while the difficulty of navigating through ice infested waters remains, the additional difficulty caused by high winds could be of a reduced magnitude. This physical phenomenon can be thought to mirror the behavior of sets in mathematics. Consider two sets, A and B. The union of A and B is not simply (A + B) but {A + B - (A n B)}. This is analogous to the behavior of the effect on productivity caused by ice and wind. If the 152 effect of ice on the productivity is considered as a set of minute effects instead of a single effect and if the effect due to the wind is considered in a similar manner, their union or their combined effect can be thought as the union of the two sets of minute effects. The difference between the sum of their individual effects and actual combined effect can be thought of as the intersection between the two sets of minute effects. This mathematical analogy is used in this thesis to model the combined effect of several external factors upon a variable. The analogy of sets is applied to two additional scenarios in order to test the robustness of the analogy. The first scenario consists of an example in which two external factors bring about opposite effects in a variable. Consider the case when the increase or reduction in the value of a variable when two external factors are active, is not equal to the difference between the individual effects of the two factors. The behavior of the variable Material Cost, in the example of an oil spill clean up is used as an illustration. Should high winds (W) be present in isolation the material cost would be increased as more sorbents would need to be employed to collect the spilt oil. Should thick ice (I) be present in isolation, the material cost would be reduced significantly as clean up consists of simply burning the spilt fuel. However, should winds and ice both be present in combination, the fuel would still be set on fire thus reducing the material cost sizably in this case as well. In this case the reduction might be greater than the value obtained as the difference between the two individual effects. Applying the analogy of sets to the effects of the external factors yields ( W u I ) = W + I - ( W n I ) (5.19) 153 If however, the value (W u I) is in reality not equal to W +1, then (W n I) should have a value other than zero. However, such an explanation seems unintuitive in this instance since wind and ice cause effects in the opposite direction and thus should have no common effects. Positive effects increase the value of the variable from its base case value by adding onto its value, while negative effects decrease the value of the variable. Therefore, negative and positive effects should be distinctly different from each other, thus leading to the statement that they share no common effects. However, this scenario is easily explained by looking at the effect of the external factors upon the material cost in detail. In the event of thick ice being present, the need for material such as sorbents and chemical dispersants in combating the oil spill is nullified. However, a need for igniters to bum the fuel now arises. Therefore, the material cost has actually decreased and then increased with respect to its base case value. The overall effect is a decrease of the material cost. Recognition of the fact that the overall effect is considered of a positive as well as a negative component provides an explanation of the interaction between the two external factors in this case. Let the negative component of the effect of ice on the material cost be IN and the positive component be IP. Then the overall effect of ice denoted by I would be equal to (IP + IN). In this example L would consist of the percentage increase in cost caused by the need for igniters while IN would be equal to the percentage decrease in material cost caused by foregoing the use of sorbents and chemicals. Similarly the total effect of wind (W) can be written as W = W P + W N . The combined effect of wind and ice can now be written as 154 (W u I) = (WP + WN) u (IP + IN) (5.20) As the intersection of a positive effect and a negative effect is zero, equation (5.20) effectively reduces to Equation (5.22) explains the scenario under consideration in which the combined effect of two externalities that create opposite effects is not equal to the difference between the two effects. (W u I) in this case need not be equal to (W +1) as {(WP n IP) + (WN n IN)} can have a value that is not equal to zero. The second scenario considered reflects an instance in which the change in the value of a variable due to a combination of two factors having an effect in the same direction is more than the sum of the individual changes caused by the two factors. An example of such a scenario would be the impact of design changes and regulatory changes upon the productivity of an excavation operation at a hazardous waste remediation site. Assume that design changes require that a large part of the excavation be carried out in an area with a steep gradient as opposed to the fairly level area associated with the base case. In this case the productivity of the equipment WuI = (WP + Ip + WN + IN)-{(WPnIp) + (WNnIN)} (5.21) which can be written as (W u I) = (W +1) - {(WP n IP) + (WN n IN)} (5.22) 155 operators would decrease due to the increased difficulty of the excavation. Let the regulatory changes that may be forthcoming require that equipment operators wear cumbersome safety clothing. This would also decrease the productivity of the operators by a certain amount. In considering the case when design changes and regulatory changes occur together, the operators would have to wear cumbersome safety clothing while carrying out a complex excavation. A possibility exists that the productivity loss in such a case could actually be greater than the sum of the two individual reductions. Such a scenario could also be explained by considering the set analogy at a level in which the positive and negative components of the effect of an external factor are considered separately. Let D denote the set of effects caused by design changes and R denote the set of effects caused by regulatory changes. Adopting notation similar to the previous scenario, it can be written (D u R) = (D + R) - {(Dp n R P) + (DN n RN)} (5.23) The term (D u R) which is the combined effect of the external factors would be greater than (D + R) when {(DP n R P) + (DN n RN)} is negative. Therefore, the scenario can also be successfully modeled using the analogy of sets. The model in this case suggests that the intersection between the negative components is greater than the intersection between the positive components of the effects of the two external factors. It is noted that the adjusted product function developed by Diekmann and Featherman (1998) does not accommodate the analysis of scenarios of this type. 156 Therefore, it is suggested that the use of sets can be employed in drawing a robust analogy to the interaction between the effects of various external factors in a variety of different situations as illustrated by the scenarios considered above. A calculation procedure based on the use of sets is therefore used in obtaining values for the adjustment coefficients in the case of multiple external factors acting in combination upon a variable. The following premises are introduced in order to provide a robust framework for the development of the calculation procedure. Premise 1 The effect of an external uncertainty on a variable can be expressed as a set of a finite number of mini effects. Two mutually exclusive types of these mini effects are assumed to exist, mini effects that cause positive changes in the variable and mini effects that cause negative changes in the variable. The two types are defined to account for increases and reductions in the value of the variable. Each of these mini effects is assumed to cause the variable to increase or decrease its value by a single percentage point from its base case value. For example, i f ice conditions decrease the material cost by 25% it is assumed that this is the overall effect of several separate effects, the net number of which is 25 negative mini effects. The number of effects has been tied to the percentage change for simplicity. Alternatively, it could have been assumed that the 25% change is caused by the sum of 250 different effects each of which causes a 0.1% change in the variable. Since each of these mini effects is assumed to increase or decrease the value by a single percentage point, the individual mini effects are similar in magnitude. In the above example, the value of a 25% decrease could be the net value of 45 negative mini effects and 20 positive mini 157 effects. The concept of a mini effect has been introduced to tie the physical phenomenon of a change in a variable, with the concept of sets that exists in mathematics. The mini effects are thought of as the elements of the set. Premise 2 The impact of each mini effect is considered to be independent of the value of the variable. This premise allows the percentage change in the variable for a particular scenario to stay the same whatever the value of the variable may be. As an example, application of this premise to the effect of high winds on the material cost would imply that the material cost would increase by 80% due to high winds irrespective of whether the material cost for the base case condition was $ 25,000 or $ 75,000. This premise plays an important in role in dealing with probability distributions of variables. Since a probability distribution encompasses a range of values of a variable, all the values can be multiplied by the same adjustment coefficient based on this premise. However, several disadvantages of using a single point coefficient exist as has been discussed previously in section 5.4.4. The calculation procedure for adjustment coefficients that is based on the premises introduced above consists of several steps. In step (I), the percentage changes in a variable due to the external factors are evaluated by considering the scenarios in which the external factors act individually. An example would be the percentage change in productivity due to high winds, when all the other external factors are at their base case values. The value of this percentage change is obtained from experts through a suitable elicitation process as described in chapter 6. The percentage change 158 values obtained in this step will be termed as individual adjustment coefficients to reflect the fact that they are due to individual external factors. Let A be the effect brought on by an external factor A. A will consist of a set of mini effects that cause the change in the variable under consideration. An example would be an instance where winds cause a 25% decrease in productivity. In this case A would consist of a set of mini effects whose net impact is a 25% reduction in the value of the productivity. In step II of the process, the factors are considered in a pair-wise manner. The percentage change due to each pair of the external factors will be elicited from experts for all pairs of external factors that can feasibly exist together. Since the expert has to consider the combined effect of only two uncertainties at the same time it is assumed that the assignment of an accurate coefficient is within the capability of an expert. An example would be the percentage change in productivity in conditions of high winds and thick ice, when all the other external factors are at the base case values. The percentage change values obtained in this step will be termed as binary adjustment coefficients to reflect the fact that they are due to pairs of external factors. Let A and B, be two effects brought on by two external factors A and B. Invoking the analogy of sets, the combined effect of A and B is given by (A u B), Therefore, it can be written, {AvB) = (A) + (B)-{Ac\B) (5.24) The percentage value obtained from an expert in this step is therefore equal to (A u B). In step III, adjustment coefficients are obtained for scenarios in which 3 or more external factors are acting in unison. While theoretically such values can also be 159 elicited from experts, the accuracy of the estimates would decrease as the number of external factors increase. This is due to the fact that the mental computation an expert would have to carry out in order to provide a value for the adjustment coefficient while accounting for the interrelationship between factors would become very complex as the number of factor increase. The likelihood of an estimate provided by an expert being erroneous could be considered as being high in such a scenario. Therefore, the values of adjustment coefficients that relate to scenarios involving 3 or more external factors are calculated using the information obtained in step I and II. If A , B and C are 3 external factors then their combined effect given by (AKJBUQ can be expressed as, (A u B u Q = (A) + (B) + (O - (A n B) - (A n Q - (B n Q + (A n B n Q (5.25) by invoking the analogy of sets. However, the term (Ar\Br\C) contains only the positive and negative mini effects that are common to A, B and C. Therefore, this term will tend to have only a very small number of mini effects. Therefore, it is assumed that the term (Ar\Br\C) can be neglected without a significant loss in accuracy. Equation (5.25) can now be written as 04ufluQ~ (A) + (B) + (Q-(AnB)-(AnQ-(BnQ (5.26) 160 The assumption made above is used to simplify the calculation process and the data requirements in obtaining the value of the adjustment coefficients for scenarios that involve 3 or more external factors. The assumption can be expressed in general terms as (Ai n A2c\ A3 Ak.inAk)~0 (5.27) where k> 3. The value of (A) and other individual coefficients in equation (5.26) can be obtained from step I, while values for (A n B), (B n Q and (Ar\C) can be obtained using the binary adjustment coefficients obtained in step II. For example (A n B) can be obtained as (AC\B) = (A) + (B)-(AKJB) (5.28) Substituting for values such as (A n B) with values obtained from equation (5.28) in equation (5.26), it can be written 0 4 u 5 u Q ~ (AVB) + (AKJQ + (BKJC)-(A)-(B)-(C) (5.29) A value for (A u B u Q is obtained in this manner using equation (5.29). 161 Similarly, in the case of four externalities A, B, C and D the combined effect is given by (A{JB(JC{JD) = A+B + C + D-(Af]B)~(Bf]C)-(Af]Q-(Af]D) -(Bf]D)-(Cf]D) + (AnBf]Q + (Bf]Cf]D) + (CnDf]A) + (Df]AnB)-(Af]BC\Cf]D) (5.30) Neglecting terms that involve the intersection of the effect of 3 or more variables (A U B U C U D) can be expressed as (A\JB[JC\JD) r A + B + C + D-(Af]B)-(Bf]Q-(Af]Q -(Af]D)-(Bf]D)-(Cr\D) (5.31) Substituting values obtained from step II, for (A f] B), (B f) Q, (A f] Q, (A f] D), (B D D) and ( C f l - D ) , it can be written that (A[jB[jC[jD) ~ (A[jB) + (B[jQ + (A[jC) + (A[jD) + (B[jD) + ( C U D) - 2A -IB - 2C - 2D (5.32) Similar results can be obtained for even larger numbers of external factors. The adjustment coefficients obtained in this manner model the percentage change the value of a variable undergoes as a result of the external factors. Therefore, in order to obtain the value of a variable under the changed circumstances the original value is multiplied by' 1 + adjustment coefficient'. 162 5 . 5 . 5 The Probability Distribution of the Decision Variable The adjusted distributions of the primary variables will serve as input to a Monte Carlo simulation carried out for each combination specified by the probability tree of the problem. An equal number of iterations are carried out for each scenario represented by the branches of the probability tree. Let the number of iterations be N . Then a sample value generated in carrying out a Monte Carlo simulation for a particular branch of the probability tree, with a joint probability of P A will have a probability of occurrence of P A x (1/N) with regard to the final distribution of the decision variable. The moments of the distribution of the decision variable are calculated using the values generated from each branch along with their associated probabilities. Since this distribution is an aggregate of several individual distributions with considerable differences in spread, the result would be a multi-modal probability distribution in most instances. Thus, it would not be possible to characterize the decision variable with the aid of a standard uni-modal statistical distribution. Therefore, the uncertainty surrounding the decision variable is represented by the moments of its probability distribution. 5.6 Application of the Quantification Methodology The use of the methodology described above is illustrated by considering an example of assessing the duration of excavation in an environmental remediation project. The primary step in the analysis methodology is the derivation of the influence diagram for the problem utilizing expert opinion. Let figure 5.11 be the outcome of this step, considering a hypothetical project in which the excavation duration is affected by the external factors of (1) regulatory changes; (2) design changes requested 163 by stakeholders; and (3) the outcome of a lawsuit filed by a member of the public. Regulatory changes and design changes would bring about a change in the volume of excavation thus affecting the duration. The change in the difficulty and complexity of the excavation caused by these changes would also affect the duration through its primary variable of equipment productivity. In this example it has been assumed that both the quantity as well as the difficulty of excavation would increase as a result of design and regulatory changes. It is also assumed that the lawsuit filed in this case is aimed at limiting the amount of excavation that is to be carried out. Therefore, an unsuccessful defense against the lawsuit would result in a reduction in the amount of excavation that can be carried out as part of the environmental restoration project. The probability tree corresponding to the influence diagram is prepared in the next step of Fig. 5.11. Influence Diagram for the Example of Evaluating Excavation Duration 164 the quantification methodology. The probability tree for this example is shown in figure 5.12. Two states each of regulatory changes and design changes, i.e. their existence and non-existence, as well as two states of the outcome of the lawsuit, i.e. successful and unsuccessful, have been considered in the analysis. As is implied by the influence diagram, the three external factors have been assumed to be independent of each other. The prior probabilities of the various states are elicited from experts. The branch probabilities of the probability tree are calculated based on the prior probability values obtained from experts. The prior probability values used in the analysis as well as the calculated values of the branch probabilities are shown in figure 5.12. The base case scenario of the problem is considered in detail in the next step of the analysis. The base case corresponds to a scenario in which regulatory changes and design changes do not occur while the outcome of the lawsuit is successful. The duration (T) of the excavation work package which is the decision variable in this case is decomposed into its primary variable as, T = — (5.33) PL where, Q = quantity of the work L = resource usage level of a resource, R P = productivity of resource R (Output / Input) 165 Fig. 5.12. Probability Tree for the Example of Evaluating Excavation Duration Probability distributions are obtained for the primary variables considering the base case scenario. In obtaining the probability distributions for the variables, the 5 t h, 25 t h, 50 t h, 75 t h and 95 t h percentile estimates are elicited from experts. Suitable standard probability distributions are selected using the distribution selection procedure described in the previous section. The values used in the analysis and the outputs obtained are shown in Table 5.1. The primary variables have been assumed to be uncorrelated for simplicity. 166 Individual and Binary adjustment coefficients are elicited from experts considering the three external factors and their pair-wise combinations, in the next step of the analysis. Adjustment coefficients are obtained for all three primary variables. The hypothetical values used in the analysis are shown in Table 5.2. Table 5.1. Percentile Values of the Primary Variables used in the Example of Excavation Duration Estimation 5th percentile 25m percentile 50th percentile 75th percentile 95m percentile Probability distribution Q(m 3) 14,000 14,600 15,200 15,900 16,800 Triangular (13,200, 15,200, 17,400) L (Excavator-hours / hour) 3 3 3 3 3 3 P (m3 / Excavator -hour) 82 86 90 94 98 Beta(1.43, 1.14)* 24.97 + 75.15 Table 5.2. Adjustment Coefficients* used in the Example of Excavation Duration Estimation Design changes (Occur) Regulatory changes (Occur) Lawsuit (Lose) Des. change + Reg. change Des. Change + Lawsuit Reg. Change + lawsuit Q 100 60 -20 110 70 50 L 0 0 0 0 0 0 P -20 -10 5 -20 -18 -8 Values are provided as percentages of the base-case values. The adjustment coefficients for the combination of the 3 external factors are calculated using the values given in Table 5.2. An example calculation is shown considering the 167 quantity of work (Q). From equation (5.29), the combined effect, (D u R u L) of the design changes (D), regulatory changes (R) and the lawsuit (L), is given as (DvRvL) = (DvR) + (DvL) + (RvL)-D-R-L (5.34) Therefore, ( D u i ? u Z ) = 1 1 0 + 70 + 50-100-60-(-20) ( D u / ? u l ) = 90% Similarly, the adjustment coefficient for the productivity, when all external factors are active is obtained as -21%. In the next step of the analysis, the probability distribution for each scenario is obtained by multiplying the base case distribution by the value ' 1 + adjustment coefficient'. The scenario when all three factors are active is used as an illustration. In this case, the adjustment factors obtained for the quantity, resource usage and the productivity are 0.9, 0 and -0.21 respectively as shown in the previous calculation. The base case probability distributions for the variables are multiplied by "1 + adjustment factor" to obtain the distribution for the scenario when all external factors are active. Therefore, the probability distributions for the scenario when all external factors are active would be 1.9 x {Triangular (13,200,15,200,17,400)}, 3 (deterministic) and 0.79 x {Beta(1.43,1.14) * 24.97 + 75.15}. 168 A Monte Carlo simulation is then carried out using the probability distributions given above and the functional relationship T = Q /PL. Let the number of iterations be 100. Therefore, each sample value generated for the duration would have a probability of 1 /100 within the scenario considered. However, the probability of the scenario occurring is 0.09. Therefore, from a global perspective each sample value of the duration would have a probability of occurrence of 0.09 x 0.01. Monte Carlo simulation is carried out for all scenarios of the problem. Simulation software such as @Risk® allow the simultaneous simulation of several scenarios which reduces the input effort. The output of the simulation process is a set of values for the duration, each with an associated probability of occurrence as described above. The set of values and their associated probabilities can be used to calculate the first four moments of the duration using the standard statistical definitions for the first four moments. An example would be the expected value, E(T) which is defined as where t is the sample value and p is its probability. The results obtained for the statistical parameters of the duration using the procedure described above are shown in table 5.3. The results correspond to a case where 100 iterations were used in carrying out Monte Carlo simulation. (5.35) 169 Table 5.3. Statistical Outputs obtained for the Excavation Duration Stat is t ica l P a r a m e t e r V a l u e Expected value (mean) 97.134 2 n d moment about the mean 1297.152 3 r d moment about the mean 3830.895 4 t h moment about the mean 3614231.912 Standard deviation 36.016 Skewness 0.082 Kurtosis 2.148 5.7 S u m m a r y The steps in the quantification method for Type B environmental risks is summarized in this section. Step 1 - The relationship between the decision variable and external uncertainty factors which are likely to have an impact are identified by preparing an influence diagram which depicts the relationship of the decision variable to external factors. Step 2 - The probability tree corresponding to the influence diagram drawn up in Step 1 is prepared considering the possible states of the external factors. Prior probabilities in the event of the external uncertainties being independent, or conditional probabilities in the event of the externalities being dependent are assigned to the branches of the probability tree except to the final level of nodes which depict the target variable. 170 Step 3 - The scenario which corresponds to all the external factors being inactive is considered as the base case for the analysis. Steps 1 to 9 in the analysis of risk type A are performed considering the base case. Step 4 - Adjustment coefficients are obtained from experts for each primary variable considering each individual factor separately. Step 5 - Adjustment coefficients are obtained for each primary variable considering the effect of two externalities being active at the same time. A l l the external uncertainties are considered in a pairwise fashion in this step of the analysis. Step 6 - Adjustment coefficients corresponding to situations when multiple factors are active are obtained by applying the values obtained in Steps 5 and 6 to equations (5.29), (5.30), or similar equations corresponding to an even larger aggregate of externalities. Step 7 - The probability distributions of the primary variables obtained in Step 3 are multiplied by the value ' 1 + adjustment coefficient' to obtain the probability distribution for the scenario under consideration. This procedure is repeated for each branch of the tree, which correspond to all the possible scenarios. Step 8 - Monte Carlo simulation is carried out for each collection of the adjusted distributions as well as for the base case scenario. An equal number of iterations is carried out at each branch in order to simplify the analysis. Step 9 - The joint probability of each value generated by the simulation is calculated by dividing the joint probability of each branch by the number of iterations. This yields the probability of each value generated as part of the simulation process. Step 10 - The first four moments of the decision variable are calculated using the values generated from the Monte Carlo simulation and their associated probabilities. 171 The uncertainty of the decision variable is represented by the first four moments obtained in this manner. 5.8 Conclusion Environmental risks of Type B can create a need for additional work packages to be executed during the life cycle of the project or cause extraordinary changes in the scope of the existing work packages of a project in other instances. Due to their strong relationship to external events and uncertainties, the scope of such work packages are difficult to express in a form suitable for analysis using Monte Carlo simulation or moment based analytical methods. Influence diagrams, which are capable of representing and analyzing the uncertainties of such work packages, have disadvantages of their own. These include the large number of probability assignments that need to be made in carrying out the analysis and difficulties in assessing conditional values required for the analysis. The risk quantification method described in this chapter makes use of multiplicative factors in order to model the effect of external uncertainties. The use of such factors is not novel and has been used previously in analyzing economic risks. However, the quantification methodology described herein enables consideration of the effects of external uncertainties at a primary variable level which is unique to this approach. This allows the benefits offered by decomposition to be utilized in carrying out the analysis. In addition, a modeling paradigm which enables the representation and quantification of the effect of multiple external uncertainties is also introduced. The disadvantages of methods previously suggested in literature to model multiple uncertainties are overcome by this approach. 172 The treatment of external uncertainties at the primary variable level as well as the introduction of a novel approach towards representing and quantifying the effect of multiple external uncertainties are considered as the major contributions of this chapter. 173 Chapter 6 Elicitation of Information from Experts 6.1 Introduction An expert is identified as an individual who has superior knowledge in a particular field brought on by training, skill, and experience, and who is capable of imparting that knowledge in a coherent and accurate manner. The use of experts in obtaining subjective information is particularly useful in economic risk analysis of projects when actuarial or relative frequency based data are unavailable. The unavailability of such data within a project proponent's or contractor's organization can be due to several reasons: (1) Data from past projects of a similar nature might be unavailable or unreliable due to the absence of proper record keeping; (2) Difficulties in getting suitable and reliable data from other organizations due to the reluctance of organizations to share information (Russell and Froese 1997); and, (3) Due to the unique nature of civil engineering projects, the attributes of a project such as location, methods of construction and characteristics of the end product differ widely between projects. Therefore, it might not be possible to use data obtained from one project in evaluating the feasibility of another project. In addition to providing estimates for parameters which cannot be quantified using frequency data, expert judgement can also be used in the development of 174 scenarios and in making value judgements (Hora 1992). In analyzing environmental risks as set out in the previous chapters, expert assessment of various parameters, and events will be required. The term "information elicitation" will be used throughout this thesis as a wide range of information, not limited to probability estimates, needs to be obtained from experts in order to carry out environmental risk quantification. The use of strategies such as decomposition of decision variables into their primary variables can contribute to the overall success of the information elicitation process. This chapter presents a discussion of such strategies as well as of the factors which impedes the elicitation process. An information elicitation process for the analysis of environmental risks that incorporates decomposition, the use of influence diagrams, and the use of directed questions based on argument types is also presented in this chapter. The final section of the chapter illustrates the use of the environmental risk analysis framework by considering the hypothetical example of a dam removal project. 6.2 Evocation of Information from Experts The information generation process within an expert's mind consists of two distinct phases (Curley and Benson 1994). In the belief assessment phase, relevant information is evoked by the expert and applied towards the formation of a belief. In the response assessment phase, a numerical qualifier is attached to the belief, which serves as an output of the elicitation process. Humans make use of several argument types in the processes of belief formation and numerically qualifying such a belief (Brockriede and Ehninger 1960, Curley and Benson 1994). However, the limitations of the information processing capability of the human mind acts as an impediment to 175 the evocation of knowledge from experts. External aids such as directed questions and prompts based upon the argument types used by humans in reasoning can play an important role in overcoming impediments and assisting the evocation of relevant information (Browne, Curly and Benson 1997). A study carried out by Browne, Curly and Benson has shown that prompts and directed questions based upon argument types can aid an expert in evoking a much larger body of information than in a case when such assistance is not given. A significant loss in the quality of the information has not been observed during the process of generating additional information. A taxonomy of such argument types identified by Brockriede and Ehninger (1960) as well as examples of directed questions based upon such arguments, which are relevant to civil engineering projects is shown in table 6.1. Influence diagrams have also been found useful in evoking knowledge and capturing disparate knowledge in a coherent form (Howard 1989). In addition to serving as a direct quantification tool, influence diagrams can also serve as an evocative tool which evokes proper considerations in making numerical assignments of probability. An influence diagram, which shows several levels of dependency as in the example shown in figure 5.10, is useful in ensuring that all the necessary information is considered prior to making numerical assessments. The numerical assessment itself can be simplified by assessing only the nodes that are immediate predecessors of the decision variable. The other nodes, whose primary function would be to evoke knowledge and provide consistency to the problem in the case of multiple experts being utilized, can be considered as being evocative (Howard 1989). Evocative nodes are linked to other nodes using dashed line arrows in a diagram. 176 Table 6.1. Directed Questions and Prompts based upon Argument Types Argument Type Description Example Cause Non-intentional causal link between data and claim. What do you think were the causes of the cost overruns in the last highway project the company undertook? Motivation Intentions of human agent used as a cause of some result. Have you thought about whether the labor force has sufficient incentive to follow these environmental protection guidelines drawn up to reduce environmental accidents? Sign Covariational, but not directly causal. So you think the cool summer we've had is a sign that the timing of the salmon run would change and throw the construction schedule off target? Generalization Induction from a sample to its population. Do you think the equipment operators the company hires on a contract basis will act in the same environmentally responsible manner as the ones on the permanent work force? Classification Concluding from a general population to a specific instance. In general, is there a high turnover rate in the casual labor force during the summer months, which could cause a loss of productivity on this project during that period? Parallel case Concluding based on a similarity between instances. Can you think of an instance of a hazardous material spill occurring on a similar project that you've been involved in? Analogy Concluding based on a similarity of relationship with something outside of the current domain. Are you saying that productivity is like a train, starting off slowly and then building up speed? Would this mean that larger quantities result in higher average productivity values? Authority Appeal to an external source as having relevant information. Have you read or heard of any specialists endorsing the new fuel spill containment booms the company intends to use on this project? 177 6.3 Decomposition of Variables The use of decomposition in the elicitation process is based on the principle of divide and conquer. The assumption that experts find it easier to analyze individual events rather than a complex combination of such events provides the platform for the decomposition concept. In economic risk analysis of engineering projects, decomposing a decision variable such as work package cost into its primary variables is often regarded as a useful technique for reducing the complexity of the estimation problem. The necessity of developing a link between time and cost in engineering economic analysis is considered to be the main reason for decomposing decision variables by Ranasinghe and Russell (1993). In addition, decomposition has also been found to be useful in reducing biases which affect the accuracy of judgmental forecasts (Merkhofer 1987). Andradottir and Bier (1998) have considered a form of decomposition in which a number of mutually exclusive, collectively exhaustive conditioning events are identified, unknown quantities of interest are estimated conditional upon such events, and finally the resulting estimates are aggregated. This type of decomposition mirrors the quantifying methodology adopted for Type B environmental risks. The study carried out by Andradottir and Bier has shown that decomposition is useful when the disaggregate estimates are likely to be accurate but prone to errors when disaggregate estimates have large errors. However, decomposition has been found to be useful even in cases where the errors in the disaggregate estimates are larger than those in the direct estimation process, provided that the estimated marginal probabilities are sufficiently precise. Ravinder, Kleinmuntz and Dyer (1988) have carried out analysis of the same type of decomposition. They concluded that decomposition could 178 contribute towards the reduction of the random errors associated with the probability encoding process. However, the marginal increase in accuracy has been shown to decrease as the number of conditioning events increase. However, Ravinder, Kleinmuntz and Dyer have also noted that dependencies among variables could adversely effect the accuracy brought on by decomposition. A more common method of using decomposition in risk analysis is to express a decision variable as a known function of several component variables (Hora, Dodd and Hora 1993, Ranasinghe and Russell 1993). The estimate for the decision variable is obtained by the recombination of the estimates for the primary variables. This type of decomposition is common to the analysis methodology adopted for all the environmental risk types. An empirical study carried out by Hora, Dodd and Hora has shown that probability distributions obtained for variables using decomposition were much better calibrated than those obtained holistically. A disadvantage of carrying out decomposition which has been highlighted by Ranasinghe and Russell (1993) is the danger of the experts losing the mental awareness of interdependencies between the primary variables while providing judgmental forecasts. Several authors have also questioned the validity of the belief that decomposition leads to more accurate estimates. Henrion, Fischer and Mullin (1993) have carried out an empirical study, the results of which show that decomposition does not significantly affect either the accuracy of the assessed median or the calibration of the subjective probability distributions. The results of the study have indicated that decomposition changes the underestimation bias in the direct estimation method to an overestimation bias. However, the majority of literature suggests that decomposition is beneficial in carrying out information elicitation. 179 6.4 Biases in Information Elicitation A bias in the knowledge encoding process can be defined as a conscious or subconscious discrepancy between the elicited information from the expert and an accurate description of the expert's underlying knowledge (Spetzler and Stael Von Holstein 1975). Biases may be introduced into the output knowledge due to several reasons. Adjustments in the responses caused by a perceived set of rewards are termed as motivational biases, while biases introduced by the mental information processing technique of the expert are termed as cognitive biases (Spetzler and Stael Von Holstein 1975). Motivational bias may be introduced when an expert views a variable as a goal rather than an uncertain quantity for which an estimate should be made. Bias due to an inaccurate perception of the role of an expert can encourage the expert to provide estimates which underestimate the uncertainty of the target variable. A belief that experts are supposed to be certain in their estimates would give rise to this type of a motivational bias (Merkhofer 1987). Cognitive bias can arise in experts due to several reasons. These include giving too much weight to events in recent memory, aligning estimates around a convenient value, and due to stereotyping. Different modes of judgement employed by experts and their contribution towards producing biases have been identified by Spetzler and Stael von Holstein (1975). These basic modes of judgement are described below. (a) Availability - Due to the comparative ease with which recent events can be recalled compared to ones more distant in the past, more recent events are given too much weight by experts. Therefore, outcomes which resemble recent events are assigned an inappropriately high probability. 180 (b) Adjustment and Anchoring - In some cases, the estimates are anchored around a convenient value instead of being assessed independently. Central bias, the most common form of bias caused by anchoring results when estimates are anchored around the median. (c) Representativeness - The estimation of the probability of an event based on its representativeness to the process from which it originates leads to a mode of judgement in which probability judgements are reduced to judgements of similarity. (d) Unstated Assumptions - As a result of unstated assumptions being made, the probability distribution will not reflect the total uncertainty of the variable. (e) Coherence - The assignment of probability to an event based on the ease with which a plausible scenario that would lead to its occurrence can be fabricated is termed as coherence. Several steps can be taken in order to prevent bias being active in expert judgment. These steps have been constructed taking into account the reasons that are likely to trigger biases in expert judgement. 1. Analysis of the relationship between the expert and the information which is being elicited can lead to the identification of motivational bias. For example, a manager who is responsible for the sales volume of a department might provide a motivationally biased estimate when asked for a forecast of future sales in order to ensure that actual sales have a higher probability of exceeding the predicted value (Spetzler and Stael Von Holstein 1975). In such cases the estimate can be 181 restructured, for example by using decomposition, in such a manner that the expert would not be aware of the link between his or her interests and the estimate. 2. Appraising the expert of the purpose of the elicitation process would also contribute towards reducing motivational bias. The expert should be made to understand that the purpose of the elicitation process is to obtain the best judgement of the expert, including an accurate assessment of the variability involved with the estimate. This would discourage the expert from attempting to make a prediction of the value of the target variable which encompasses minimum uncertainty (Merkhofer 1987). 3. Experts can be prompted to recollect relevant events which occurred a significant period into the past in order to overcome bias caused by the comparative availability of recent events. This will prevent experts from giving more weight than necessary to recent events. 4. In most instances, the median serves as a convenient value around which other values of the estimate can be aligned. Central bias can be overcome by encouraging experts to consider scenarios in which extreme outcomes can occur (Spetzler and Stael Von Holstein 1975). In carrying out the elicitation of several percentile values of a variable, eliciting the extreme values initially is also considered as being useful in reducing central bias (Ranasinghe and Russell 1993). 5. Discussion of the possibility of the scenarios which correspond to the elicited estimates can be carried out with the expert. Experts can be made aware that the ease or difficulty with which scenarios can be constructed does not necessarily correspond to the possibility of the actual event occurring. 182 In addition to eliciting the estimates, the analyst should also strive to obtain the explicit and implicit assumptions the expert may have made during the estimation process. This would ensure that all experts and analysts share a common set of assumptions. It would also perform as a consistency check as to whether the expert has been provided with all the available information necessary for the estimation process. 6.5 Information Requirements for Analyzing Environmental Risks Several types of information need to be elicited from experts in order to analyze environmental risks as described in the previous chapters. The assessments which would be required from experts are described below. Initial estimates for the statistical parameters of the decision variables As described in section 3.4.3, the identified risks need to be screened in order to identify the most important risks. Estimates for the expected value, and the minimum and maximum possible values of the cost and duration of the work packages needs to be obtained from experts in order to prioritize risks. Identification of the external uncertainties which affect primary variables In analyzing Type B risks, it is necessary to identify the external uncertainty factors which are capable of making a significant impact on the decision variables of cost and duration. In some instances it would be necessary to consider other uncertain events which in turn are likely to affect the outcome of the external uncertainties which directly affect the target variable. The identification of external uncertainty factors which would be carried out as part of the information elicitation process would lead to 183 the development of an influence diagram which would serve as the foundation for the analysis of a Type B risk. Estimation of the probability distributions which characterize uncertain variables and events. Subjective estimates for the 5 t h, 25 t h, 50 t h, 75 t h and 95 t h percentile values would be required for each primary variable. In the case of Type B risks, the percentile values are obtained for the base case. In addition, event probabilities would have to be assessed in order to assign probability values to the probability tree derived for the analysis of Type B risks. Estimation of adjustment coefficients The adjustment coefficients which characterize the change in the primary variable due to external uncertainties should be obtained from experts. Two types of assessments need to be made in this regard. 1. Individual adjustment coefficients should be obtained for each state of all the externalities. 2. Binary adjustment coefficients should be obtained considering the external factors in a pairwise manner. Estimation of the interdependency between primary variables The percentage reduction in uncertainty of a primary variable due to the knowledge of the exact value of a second primary variable is required in the analysis in order to calculate the value of the rank correlation coefficient between the two variables. A l l 184 primary variables are considered in a pair-wise manner in order to obtain the rank correlation matrix. 6.6 The Information Encoding Process The encoding process involves the generation of accurate information regarding the parameters and events described in section 6.5. This process also involves the adoption of measures to reduce biases, carrying out consistency checks, and adopting steps to generate a complete body of information using information evocation methods. The information encoding process consists of 2 phases. Phase I involves eliciting information required to carry out risk screening, while Phase II involves obtaining information required to carry out in-depth analysis of selected risks. The encoding process is described in detail in the following sections. 6.6.1 Phase I - Eliciting Information for Risk Screening The elicitation of information for risk screening involves obtaining initial estimates for the statistical parameters of all the environmental risks that were identified as being relevant to the project under consideration during the risk identification process. The statistical parameters that need to be elicited are the expected value, and the minimum and maximum possible values of the economic variable. Information elicitation for risk screening should be carried out using a limited number of experts such as project managers and estimators who preferably have experience in projects of a similar nature. Information regarding the project should be presented to these experts in advance of the actual encoding process. 185 The values for the statistical parameters maybe elicited from experts by asking direct questions related to the parameters. The minimum and the maximum values can be obtained by asking questions of the form "What is the minimum / maximum possible value you would expect the parameter P of work package Wto be, assuming that everything associated with this work package goes as planned / goes wrong?". An example of such a query would be "What is the maximum possible value you would expect the duration of the re-vegetation program to go up to assuming that everything associated with this program that could go wrong, goes wrong?". An estimate for the expected value of the variable can be obtained by asking questions of the form "What is the value of the parameter P of the work package W, you would expect on average given the circumstances of the project?". 6.6.2 Phase II - Eliciting Information for In-depth Analysis of Selected Risks Phase II of the information encoding process consists of 11 distinct steps, some of which are however applicable only to Type B environmental risks. Step 1 - Selection of experts The primary step in the elicitation process involves the identification of suitable experts who will be able to provide the required information. The comprehensive analysis of environmental risks would involve dealing with issues in a wide range of specializations such as ecology, wildlife conservation, meteorology and hydrology. Therefore, experts external to the project proponent's organization would have to be approached in most cases to obtain the required information when carrying out in-depth analysis of important risks. In instances where the environmental issues under 186 consideration are under heavy public scrutiny it might be necessary to ensure that the experts who are used in the elicitation process have a reputation for being impartial and are well qualified in their field. Experts external to the organization can be selected using registries of professional organizations and consulting firms, listings of universities and government agencies as well as through literature searches (Hora 1992). Experts within the organization can be selected by considering previous experience in similar projects, overall work experience and their involvement with the project under scrutiny. Experts who are directly involved with the project and whose personal performance could possibly be tied to the performance of the project should not be used unless an alternative expert is not available, in order to reduce the possibility of motivational biases. Step 2 - Definition of the problem Information regarding the project in general and specific information pertaining to the experts' field of specialization should be presented to the expert well in advance to the actual encoding process. The information provided to experts should include a description of the decomposition method which will be used, the parameters for which the expert is required to make estimates as well the assumptions which are required to be made. Step 3 - Identification of external uncertainty factors (Type B risks only) This step involves the identification of the external factors that need to be taken into account while making estimates for the decision variable or its decomposed primary variables, initially, the analyst should prepare a basic influence diagram which shows 187 the relationship between an event occurring, and the time or cost uncertainty associated with it. Figure 5,2 is an example of such a diagram. The analyst and the expert should then proceed to add additional nodes which represent chance events that need to be taken into account in order for an accurate estimate of the variable to be made. The analyst should make use of directed questions and prompts based upon argument types to evoke information from the expert. Subsequent to identifying the external factors, the analysts should repeat Step 1 of the encoding process and identify suitable experts who could estimate the uncertainty of the external factors. However, in some cases the analyst might be able to make use of existing frequency data in order to obtain such estimates. In the example of a marine oil spill the analyst might be able to use frequency data to calculate the probability that the water will be ice infested during the period marine construction is taking place. Steps 2 and 3 should be repeated with any new experts who are selected. This will lead to further development of the influence diagram. For Type B risks, the iteration among the first three steps should be carried out until the influence diagram is developed to a level sufficient enough to calculate the probability distribution of the variable under consideration. The number of distinct states of the external factors that need to be considered should be obtained from the expert. The analyst should use his or her judgement in selecting the nodes for which numerical assessments are required. Limiting the quantitative assessment to nodes that are immediate predecessors of the decision variable node is suggested. A l l the other nodes should be considered as evocative. In some cases even nodes that are immediate predecessor to the decision variable need 188 not be assessed numerically. As an example consider the case when an expert affirms that the presence of marine wildlife would affect the cost of a marine fuel spill cleanup as indicated in the influence diagram shown in figure 5.10. However, i f further discussion indicates that the change in cost due to the presence or absence of wildlife at the time of the spill is minimal, the chance node pertaining to the presence of marine wildlife can be considered as being evocative in order to simplify the analysis. However, it is important to portray such nodes in the influence diagrams in order to ensure all relevant knowledge is considered in the analysis as well as to encourage further generation of information. Step 4 - Receive feedback from the expert regarding the problem definition The analysts should discuss and receive feedback from the expert regarding gaps in the information which are necessary to be filled in order to make an educated estimate of the parameters, the validity of the assumptions that have been made, additional assumption which need to be made, and the level of comfort of the expert regarding the decomposition method that has been adopted. Step 5 - Redefinition of the problem Redefinition of the problem is only necessary should any expert request a major change regarding the assumptions that have been made or the type of decomposition which has been carried out. In such an event the problem will have to be restructured and all the experts who are affected by the changes will have to be consulted. This necessitates for steps 4 and 5 to be repeated until all the experts are comfortable with the problem definition. 189 Step 6 - Training in probability judgement The training process is designed to assist the experts in expressing their knowledge as probability estimates and to make them aware of the biases that might affect their judgement. The training session should include a brief introduction to the basics of probability theory. Topics that should be explained are the probability axioms, the concept of correlation and the application of Bayes' rule. These topics should be explained in a very general manner. Training in the basics of probability is necessary as experts in substantive fields such as engineering may not be effective in expressing their beliefs in the form of probability distributions (Hora 1992). Research carried out by Gebotys and Claxton-Oldfield (1989) has indicated that subjects who received training in probability theory provided much better probability assignments than subjects who did not receive such training. The effect of such training has been shown to be evident even several weeks after the training. Subsequent to training in the basics of probability, the experts should be given a brief introduction to the biases that could affect their judgement as well as to the modes of judgement that leads to such biases. The training session should also include a practice session designed to allow experts to gain confidence in the process. The use of questions from almanacs is suggested as parameters to be estimated during the practice session. Step 7 - Encoding of probability estimates for primary variables The 5 t h, 25 t h, 50 t h, 75 t h and 95 t h percentile values of primary variables need to be obtained as subjective estimates from experts. Questions of the form "What is the value of variable V, which you feel has only a P % chance of being / not being 190 exceeded?". An example query would be "What is the value of the excavation quantity that you feel has only a 5% chance of being exceeded?". In order to reduce central bias the extreme values should be elicited initially (Ranasinghe and Russell 1993, Spetzler and Stael Von Holstein 1975). The analyst should ask directed questions based upon argument types in order to encourage the expert to take a wider range of issues into account in order to reduce biases such as availability bias. Step 8 - Consistency checking Once the required probability values of the primary variables are available, a probability density function should be generated for each primary variable using computer graphics and displayed to the experts. A probability density function is a more preferred mode of presentation than cumulative distribution plots as characteristics such as skewness are more clearly presented by a probability distribution plot. Percentile values, or extreme values obtained from the probability distribution can be read out to the expert in order to check the validity of the results. The expert should be asked to comment on the final result. Should grave concerns be voiced regarding the values, the analyst should consider repeating steps 7 and 8 of the elicitation process (AbouRizk and Sawhney 1993). Step 9 - Encoding of event probabilities (Type B risks only) In executing this step the analyst should derive the probability tree corresponding to the influence diagram obtained in step 3. Prior probabilities for events which are not dependent upon any other event in the influence diagram should be elicited initially 191 from the relevant experts. Direct questions such as "What is the probability of the seas around the project location being ice infested during the construction period from April to December?" can be used to elicit the probability values associated with each state of the event. Once prior probabilities are assigned to independent events, conditional probabilities are elicited for dependent events. Direct questions such as "What is the probability that a large number of seabirds will be present in the area, given that the ice cover is more than 10 percent?" can be used in such instances. Step 10 - Encoding of adjustment coefficients (Type B risks only) Adjustment coefficients, defined in Chapter 4, need to be elicited from experts in the case of Type B environmental risks. Individual adjustment coefficients can be obtained by asking questions of the form "By what percentage would the average value of parameter P change if state A of external uncertainty Z is present?" An example of such a query related to the marine fuel spill work package would be "By what percentage would the labor force need to be increased to combat the spill, on average, if high winds are present during the time of the fuel spill?" Adjustment coefficients for two simultaneous uncertainties can be assessed by asking questions of the form "By what percentage would the average value of parameter P change if state A of external uncertainty Z and state L of external uncertainty Y are present at the same time?" An example of such a query would be "By what percentage would the labor force need to be increased to combat the spill, on average, if high winds are present during the time of the fuel spill and the waters are ice infested at the same time?" 192 Step 11 - Elicitation of interdependency parameters The process of eliciting interdependency parameters between variables involves identification of the existence of interdependence, and in the event the variables are correlated, elicitation of the rank correlation coefficient from an expert. The percentage reduction in the uncertainty of one variable due to the knowledge of the exact value of the other variable is utilized in order to identify the existence of interdependence as well as to obtain an indirect value for the rank correlation coefficient (van Dorp and Duffey 1999). Subsequent to selecting a pair of variables X , and X 2 , the expert is asked if knowing the exact value of X , would help in reducing the uncertainty associated with X 2 . The choice between X , and X 2 should be made taking into account the relative ease with which an expert can visualize the conditional values of one variable given another. For example, an expert would find it intuitively easier to think of the values of productivity conditional upon the knowledge of the exact value of the size of the work force, rather than about the values of the size of the work force based upon the knowledge of the value of the productivity. In the event of the expert answering 'No' to the above question, the two variables would be deemed to be independent. In the event of an 'Yes' answer the two variables are interdependent, necessitating the elicitation of a value for the rank correlation coefficient. In order to obtain the rank correlation coefficient, the expert is asked for the percentage reduction in the uncertainty of one variable due to the knowledge of the exact value of the other variable. Questions of the form "To what values or range could you restrict the value of X 2 if you knew thatthe value of X ! is exactly A ? " should be asked from the expert. If the expert answers that the value of X 2 will now 193 fall within M and N , the corresponding cumulative probabilities can be obtained from the marginal distribution of X 2 as described in section 4.3.3. If the corresponding cumulative probability values are P M and P N then the corresponding percentage reduction in uncertainty would be equal to {1 - (PN - PM)} x 100 %. The percentage reduction value should be obtained for several values of X , that divide the axis of the diagonal band distribution corresponding to X , into equal parts. The uncertainty reduction of the mid-point of the sub-range is used to represent the sub-range. For example in figure 6.1, "(1 - b)" is used to represent the average percentage reduction in uncertainty in the 1st quartile. The reduction in uncertainty is obtained by multiplying "(1 - b)" by the interval width. Summation of the values obtained for the four intervals would give the overall percentage reduction in uncertainty. Division into four equal parts as shown in figure 6.1 is recommended. This would involve obtaining the conditional range of X 2 at points corresponding to the 12.5th, 37.5th, 62.5 th and 87.5th percentile values of X , . The value of the rank correlation coefficient is obtained using equation (4.36) utilizing the overall percentage reduction in uncertainty as input. 0,0 0.125 0.25 Fig. 6.1. Representation of the Percentage Reduction in Uncertainty 194 Once the interdependence between X , and X 2 has been characterized, another primary variable, X 3 is introduced into the analysis by considering the pairs X ! and X 3 and X 2 and X 3 . The correlation coefficient between these pairs are calculated using the procedure set out above. Subsequently, the correlation matrix between the three variables X , , X 2 and X 3 is evaluated in order to satisfy the theoretical requirement of the correlation matrix being positive definite (Ranasinghe 1990). In the event of the matrix not being positive definite the correlation coefficient between X 2 and X 3 is re-evaluated. In the event of the correlation matrix not being positive definite subsequent to the re-evaluation, the correlation between the other pairs of variables are considered for re-evaluation until this theoretical requirement is met. The information encoding process for the analysis of the two risk types is shown graphically in figures 6.2 and 6.3. 6.7 The Use of Information Technology and Multimedia in Information Elicitation This section briefly identifies the usefulness of Information Technology (IT) and multimedia in information encoding. Advocacy for the use of computers and related tools in knowledge elicitation is limited in the literature due to the established belief that a better quality elicitation results from a direct interaction between an analyst and the expert. This reflects a philosophy that an attempt at self-elicitation without the aid of an analyst would result in inaccurate values (Merkhofer 1987). Spetzler and Stael Von Holstein (1975) recommend against the use of interactive computer programs as the lack of the balancing effect of the analyst can adversely affect the results. However, recent research has started to challenge these beliefs (Walther 1996). 195 Identification of experts Definition of the problem to experts . . — < r Feedback from experts Yes No Training in probability judgement Encoding of percentile estimates U Generate PDF for primary variables Fig. 6.2. Flowchart of the Information Encoding Process for Type A Risks 196 Identification of experts Definition of the problem to experts Identification of external factors Yes No r Training in probability judgement Encoding of percentile estimates Generate PDF for primary variables No Elicit event probabilities, dependency parameters and adjustment coefficients j -W End . 6.3. Flowchart of the Information Encoding Process for Type B Risks 197 The use of computer software and graphics in calculating and generating visual representations of probability distribution functions for the values elicited from experts in order to carry out consistency checks at the end of an encoding process is recommended by Ranasinghe (1990). A tool developed by AbouRizk and Sawheny (1993) for the purpose of fitting beta distributions to subjective data utilizes visual representation of the fitted distributions. The ability of such a representation to immediately reveal the quality of the fit is cited as a significant advantage. It is suggested that further use of new developments in information technology and multimedia can contribute towards the successful elicitation of information while reducing the time and costs involved with the process. Multimedia can be classified as an application that uses two or more media to convey a message or achieve a given purpose (Vanegas and Smith 1994). Information technologies that can be used in conjunction with multimedia in information encoding include Group Decision Support Systems (GDSS), Desktop videoconferencing, Electronic mail (E-mail) and the Internet. Multimedia can assist in the description and definition of the project and the natural environment, while IT tools can provide versatility in helping maintain personal interaction between the analyst and the experts while reducing associated costs simultaneously. A set of instances in which IT tools and multimedia can be used in the information elicitation process, and the tools that can be used in order to facilitate the process are presented in table 6.2. 198 Table 6.2. Use of Information Technology and Multimedia in Information Elicitation Steps in The Elicitation Process Available Tools Definition of the problem - Providing description of the work site, the natural environment of the project area, and proposed construction and operation methods to experts Video clips, Interactive CD-ROMs Definition of the project - Visualization of the finished product and its processes Graphics based computer applications Definition of the project - Distribution of project information among experts at the start of the elicitation process. E-mail, Internet or Intranet, Bulletin boards (GDSS) Receiving feedback from experts - Clarification of information and provision of additional information which may be required by individual experts E-mail, Videoconferencing, Internet chat Training in probability judgement Videoconferencing, Multimedia educational environments (Interactive CD-ROMs, Interactive H T M L documents on Intranet), video clips Information encoding Videoconferencing Consistency checking - Providing feedback to experts and carrying out consistency checks on the information obtained Interactive distributed computer applications The advantages of the use of IT tools and multimedia in information elicitation can be summarized as follows: 1. Multimedia based descriptions of the project site, the surrounding natural environment, the construction, operation and demolition methods, and the finished project components can assist experts in gaining a better understanding of the 199 project characteristics thus enabling them to apply their knowledge to suit such conditions. 2. The use of IT and multimedia-based tools for the training of experts in probability encoding allows self-paced learning. This is an important factor with regard to experts with tight time schedules. The use of hypertext linkages in CD-ROM or web-based documents allow experts to easily navigate though the project documents and effectively focus attention on topics of relevance. 3. The use of videoconferencing during the actual elicitation process eliminates the need for the experts and the analyst to be in the same location. Thus, videoconferencing can provide the personal interaction required for a successful elicitation while reducing associated costs. 4. The use of distributed computer applications to display outputs of the elicitation process in a graphical form allows consistency checks to be carried out in real-time while the participants are at several different locations. 6.8 Application of the Environmental Risk Analysis Framework This section illustrates the use of the environmental risk analysis framework. Screen snapshots of the prototype computer system developed to implement the framework are used to aid the illustration. The prototype system uses a combination of commercially available software, using Microsoft Visual Basic® as the main programming environment. The example deals with the removal of a dam which is part of a large hydropower project. The application of the risk analysis framework towards studying the duration uncertainty associated with this operation is considered in this hypothetical example. 200 Background Smaller dams may be constructed in conjunction with large hydropower dams for reasons such as to dampen the effects of water fluctuations in the downstream reaches. However, instances can arise when such dams need to be removed. The reasons for such an action could be the prohibitive cost of rehabilitation which makes removal the only viable option in the case of damaged or degraded dams, and due to concerns over negative environmental effects that might have arisen unexpectedly. The hypothetical example introduced in this section concerns the removal of a rock and earthfill dam located on the Ecstall river in British Columbia. The dam is assumed to be 9.3 m in height and 78 m in length. The degraded dam is to be removed as its further use could pose a danger to the public. The planned schedule calls for the removal of the dam during the summer of 2001. The dam is located downstream of a Uranium mine. There exists a possibility that the sediment trapped behind the dam may be highly contaminated with toxic waste discharged from the mine several years ago. While the removal of dams with a height less than 10 m is generally not required to undergo an EIA assessment (British Columbia 1995) the possible presence of contaminants has prompted the Environmental Assessment Office to reconsider the exemption. At the time of this analysis, assumed to be early December 2000, tests are being conducted to determine the concentrations of toxic contaminants within the trapped sediment. The results are expected by the end of the month. The requirement for an EIA would depend largely upon the test results. Therefore, the EIA process regarding this project can be considered as a project component that may or may not arise. 201 The uncertainty associated with the EIA process as well as the uncertainty associated with other project components that may be affected by environmental issues contribute to the overall uncertainty of the dam removal project duration. The environmental risk analysis framework is applied to this problem in order to assist in the determination of the likelihood of being able to complete the dam removal process by the end of September 2001. The risk analysis process is described below. Stage 1 - The Define Stage The define stage involves obtaining a definition of the project and the environment. Definition of the project The project is defined in terms of its scope and component work packages. Generally, the scope of small dams is provided in terms of dam height measured from the lowest point of the foundation to the crest of the dam. The component work packages considered include only the work packages that are required to achieve the basic objectives of the project. Additional work packages that may arise due to environmental issues need not be considered at this stage. The project definition document is shown in figure 6.4. Definition of the natural environment The natural environment is described in terms of the valued ecosystem components that can be found v^thin the project area (see figure 6.5). 202 Pr oj ect - Rjemo val of dam No.3 Location - Ecstall river (Northern BC) Scope- dam height = 9.3 m W o r k Package Details Breach dam Quantity of rock / earth to be removed-Approx. 10,000 m 3 Work duration-May 2001 to September 2001 U sing B ackho e 1 o ader(s) Dispose dam fill material Quantity-Approx. 10,000 m 3 Work duration-May 2001 to September 2001 Using Dump truck(s) Fig. 6.4. Dam Removal Project on Ecstall River - Definition of the Project Project — Removal of dam No.3 Location —Ecstall river (Northern BC) Important Environmental Components E nviro nm ent al C omp onent Importance Description Chinook Salmon High Has high economic value. River harbors juveniles. Extensive spawning migration occurs from May to mid June. Moose Low Infrequently encountered in the project area. Listed as Threatened. Grizzly Bear Low Infrequently encountered in the project area. Listed as Threatened. Ecstall river waters Medium Downstream reaches are used for recreational purposes Fig. 6.5. Dam Removal Project on Ecstall River - Definition of the Environment 203 Stage 2 - The Risk Identification Stage The next stage in the analysis framework is the risk identification stage. A l l work packages that arise or are affected by environmental issues are identified during this stage. Public perceptions are analyzed prior to the actual identification of risks. Analysis of public perceptions Analysis of public perceptions on this project would involve carrying out discussions with first nations people, outdoor recreational groups including anglers and hunters, and representatives of nearby communities with the aim of identifying concerns regarding the project. In this hypothetical project, it is assumed that the only major concern that has been raised is by a downstream community regarding the degradation of water quality of the river during the dam removal operation. Concerns have been heightened due to the possible presence of toxic material trapped behind the dam. The river is used by the community for recreational purposes. Identification of risks A risk register serves as the starting point of the risk identification process. The risk register provided in Appendix A - l that has been incorporated into the prototype system as a Microsoft Access® database is utilized for this purpose (see figure 6.6). The entries in the register are considered individually to determine their applicability to the project under consideration. An example would be the determination of the applicability of the work package "Construction of a sediment trap downstream of the dam to trap sediments that will be eroded". This work package is relevant to the project in view of the presence of juvenile Chinook salmon in the river waters. High 204 ^•Environmental Risk Analysis - [Select Risks] EJ» File,. Window Construction of a sediment trap downstream of , Risk Type A1 the dam ', H i iFi:k • !•! i Details In carrying out dam removal; sediment tiaps would need to reconstructed to prevent sediment that , were ortgirtally trapped beninci ihejeiam being washed downstream and causing sedimentation in the OK Analyse Fig. 6.6. Dam Removal Project on Ecstall River - Use of the Risk Register loads of sediments introduced into the waters can smother the juveniles. The work package is also relevant in view of the concerns raised by the downstream community. Review of the register as well as review of government regulations leads to the identification of the following project components that contribute to the project duration and its uncertainty: (i) Carrying out Environmental Impact Assessment Studies and obtaining a Project Approval Certificate; (ii) Removal of 9, 000 m 3 (approximate estimate) of possibly toxic sediment from behind the dam prior to breaching; (iii) Disposal of 9,000 m 3 (approximate estimate) of possibly toxic sediment that is removed from behind the dam; 205 (iv) Construction of a sediment trap downstream of the dam to trap sediments that will be eroded; Additionally, the work packages "Breach dam", "Dispose dam fill material", "Remove sediment" and "Dispose sediment" match the characteristics of generic work packages ' A ' , and 'D ' (see Appendix A) as they involve the use of heavy machinery and may have schedules that overlap with the migration period offish. Therefore, these work packages would have to be halted during the migration period of salmon. As the migration occurs during the period late spring - early summer, this effectively means that these work packages cannot commence until the migration period ends. The uncertainty regarding the culmination of the migration would affect the duration uncertainty of the work packages that are scheduled to commence immediately after the migration ends. While these work packages would also need to be halted during wet weather as is characteristic of a generic work package ' A ' , the installation of a sediment trap would nullify the need for such stoppages. The "installation of sediment trap" itself matches the characteristics of generic work package "D". Therefore, the installation cannot be carried out during the salmon migrating season. An attempt is also made to identify additional environmental risks that may not be included in the risk register using expert interviews. In this project it has been identified that the dam material may also be contaminated with toxic mine waste. Therefore, additional precautions will need to be taken during dam breaching. The dam fill material will also need to be treated prior to disposal. These measures will contribute to the duration uncertainty of the "Breach dam" and "Dispose dam fill material" work packages. However, this addition to the scope of work is contingent 206 upon the results of the toxicity determination tests. As the original risk register does not include these risks, it is updated using the register update form of the prototype system (see figure 6.7). ^ Environmental Risk Analysis > [Update Risk Registei] 5 File V'r.dcu" Kelp Update Risk Register i , n,.-., .1,1,, .,i Disposal of dam fill material Please input the work package sut>|ect to environmental* I;>M Please mpjl a de:criphon lor fu>ue rolcrcme Is the existence of thework , component an uncertain'^ event? •. The dam fill material may be _> contaminated with toxic substances ,r; that are introduced in the upstream i|i reaches and are trapped by the dam. f~ Is the work component a new work « ^ package or an addition to an existing , f? Yes r NO • ... work package? r NewWP (• Addit'on Update Register Quit Fig. 6.7. Dam Removal Project on Ecstall River - Updating the Risk Register Risk categorization The risks are categorized in accordance with the definitions set out in Chapter 4. The classifications given to the duration uncertainty caused by the project components are described below. 207 (a) The uncertainty surrounding the "EIA duration" is classified as a Type B l risk in view of the fact that its occurrence is not certain at the time of the analysis, while its magnitude is also of an uncertain nature; (b) The need to remove sediment from behind the dam is known with certainty. However, as the results of the toxicity determination tests are unavailable, it is not known with certainty whether special removal procedures need to be adopted. As the addition to the scope due to special removal procedures is uncertain, the duration uncertainty of the "sediment removal" process is therefore a Type B2 risk. (c) The duration uncertainty of the disposal of sediment can be considered in the same manner as (b) above, and be classified as a Type B21 risk. (d) The provision of a sediment trap is work package that will need to be carried out with certainty. Therefore, the duration risk associated with this work package is classified as a Type A l risk. (e) The scope of the "breach dam" work package would be affected by the discovery of toxic material behind the dam in a manner similar to (b) and (c) above. While the breaching of the dam is a certain event, the addition to its scope is uncertain. Therefore, its duration uncertainty is a Type B2 risk. (f) The duration uncertainty of the "dispose dam fill material" work package can also be classified as a Type B2 risk in a manner similar to (b), (c) and (e) above. The dam removal process can therefore be characterized by the network of work packages shown in figure 6.8. 208 o CO <L> -a o o CN a. (ZD L Z 1 1 on , , 5 S J3 O CQ - o U 22 J3 o g e4 £ CO 8J a .s 5 o 5 CN < C C/3 Q U "5 <=> cS o PJ CN 209 Stage 3 - The Analysis Stage The risks that are most significant to the project duration are identified and quantified during the analysis stage. Risk screening Risk screening involves selecting a basis for risk prioritization and carrying out screening according to the selected basis. In this example, the total time window for the project is 9 months in order to achieve targeted completion by the end of September 2001. In the event of dam removal not being completed prior to the end of the summer, the remainder of the dam removal would have to be carried out in the summer of2002, as harsh weather conditions prevents work from being carried out during the rest of the year. This would involve additional demobilization and mobilization costs as well as increase the risk of the dam or part of it being washed away during the winter. Additionally, ground operations can only commence after the 2 n d week of June as project activities cannot take place during the salmon migration season from May to mid June. In view of these limited time windows available for successful completion of the project, it has been decided that work package durations that have a total uncertainty (maximum duration - minimum duration) of more than 5 days will be investigated comprehensively. Therefore, the maximum and minimum possible values for the duration of each work package is obtained from experts in order to carry out the screening process. The expected value is also obtained in order to characterize the duration of work packages with insignificant duration uncertainty. A suitable expert(s) needs to be chosen to obtain the required information. In this example it is assumed that the services of a Project Manager who has had experience 210 on a similar project is available to provide information with respect to all work packages except the "EIA process". Preliminary statistical parameters for these work packages are obtained by asking questions such as "What would you expect the duration of the sediment removal operation to be on average given the circumstances of the project?". However, an expert is not available'with respect to the "EIA process". In such an event, the analyst would have to refer documents that set out theoretical time frames for EIA studies in British Columbia, and review any data that set out the actual assessment durations of previous dam projects or other types of projects. The analyst would assign estimates to the statistical parameters of the EIA process using such information. Hypothetical values used in this example are given in table 6.3. Table 6.3. Dam Removal Project on Ecstall River - Initial Estimates for the Statistical Parameters of Duration Variables Work Package Classification of Risk Expected value of the duration (days*) Minimum duration (days*) Maximum duration (days*) EIA process B l 90 0 125 Remove sediment B2 27 23 36 Dispose sediment B2 32 24 42 Construct sediment trap A l 3 2 4 Breach dam B2 31 27 42 Dispose dam fill material B2 33 27 43 * Durations have been estimated in work days 211 A l l work packages except the "construct sediment trap" work package meet the criteria of being subject to significant duration risk. Therefore, an in-depth analysis of the other work packages needs to be carried out while the duration of the "construction of sediment trap" is represented by the estimate of its expected value. In-depth analysis of significant risks The "dispose sediment" work package is used to describe the in-depth analysis procedure. The duration uncertainty of sediment disposal is a Type B2 risk. The duration, T is decomposed as T = Q / P M where Q is the quantity of sediment, P is the productivity of the equipment used for hauling and disposal, and M is the usage rate of the equipment. The experts who can provide probability estimates are identified in the first step. As the work package will be affected by the toxicity determination tests, it will be necessary to identify an expert capable of estimating the likelihood of the various outcomes of the tests. Similarly, it is also necessary to identify an expert(s) who will provide percentile estimates for the primary variables, i.e. Q, P and M . In this example, it is assumed that an expert knowledgeable about the sedimentation characteristics of mine waste and who has access to the records of previous waste disposal practices of the Uranium mine will provide information on the possible outcomes of the test results. Similarly, a project manager knowledgeable about the sediment disposal method is assumed to be selected to provide the percentile estimates. Relevant project information is presented to these experts. In the next step, the basic influence diagram for the problem is prepared by the analyst (see figure 6.9). 212 Fig. 6.9. Dam Removal Project on Ecstall River - Basic Influence Diagram for evaluating Sediment Disposal Duration In the next step, other external factors that need to be taken into account in assessing the duration uncertainty are identified. In this example it is assumed that further discussion with the experts has identified the following factors; (a) It is likely that the Environmental Assessment Office will set out certain guidelines and conditions for the transportation and disposal of waste, as part of the EIA approval procedure. The duration would depend on these guidelines; (b) Two alternative locations within the two different regional districts of Stikine and Skeena-Queen Charlotte, have been identified as being suitable to be used as dumping grounds of the sediment. It is planned that treatment of the sediment (if necessary) will be carried out at the dump sites. However, neither of the regional districts has given a complete assurance that their site can be used for dumping. As these sites are at different distances from the dam site and within two different jurisdictions, the duration will depend upon the choice of the dump site. There also exists a possibility that neither of the two sites will allow the dumping to be carried out, especially in cases when the sediment is found to be toxic; and (c) delays in the salmon run would affect the start of the work package effectively extending its duration. However, the analyst has reason to believe that the occurrence of such a delay is unlikely based on past records of 213 salmon runs in the Ecstall river. Therefore, the effect of such a delay will not be considered in the risk quantification. The influence diagram is further developed based upon the information given above. The influence diagramming procedure is implemented in the prototype system using Precision Tree®. The detailed influence diagram prepared using the prototype system is shown in figure 6.10, with the effect of the salmon run being considered as evocative. >J Microsoft ExcelXEcstallRiverDamProiect • © I ' m ren m Hi rsi 1 •/v IQJ 'rflJ g d. V M lilt 11 File Edit View Inseit Format lools Data Window PrecisionTree Help Ei v * & % m & Z— ct? o/ +.0 .00 J :p= g|3 tp A> j .oo +.o j 1 Fig. 6.10. Dam Removal Project on Ecstall River - Refined Influence Diagram for evaluating Sediment Disposal Duration 214 The tests results would influence the outcomes of the specifications and the location as shown in figure 6.10. Additional experts capable of providing probability estimates regarding the location and the specifications are identified in the next step. The analyst then receives feedback from the experts regarding the problem definition expressed by figure 6.10. Should no serious reservation be made, the analyst would continue onto the next step in the analysis, which is the training of experts in probability judgement. In this step the expert is made aware of the fundamentals of probability theory as well the various biases that would affect his or her judgement. The probability tree for the problem is generated taking into account a suitable number of states for each of the influencing nodes. The number of states are determined in consultation with the experts. The states for each node used in this example is given in table 6.4. Table 6.4. Dam Removal Project on Ecstall River - Possible States of the External Factors Affecting Sediment Disposal Duration External Factor States Considered Test result 1. Tests indicate presence of high concentrations (> 20 g / m3) of toxic matter. 2. Tests indicate low concentrations (< 20 g / m3) of toxic matter. Specifications for Treatment and Disposal 1. Stringent specifications will be provided requiring comprehensive treatment of the sediment. 2. Specifications will be lax. Location 1. Dumpsite will be located in the Stikine regional district. 2. Dumpsite will be located in the Skeena regional district. 3. Other location. 215 X Microsoft Excel - EcstallRiveiDamProject tt^) File Edit View Insert Format Tools Data-Window PrecisionTiee Help Fig. 6.11. Dam Removal Project on Ecstall River - Probability Tree for evaluating Sediment Disposal Duration The probability tree for the problem generated using Precision Tree® is shown in figure 6.11. Probability estimates are obtained from the experts in the next step of the analysis. Conditional probabilities would have to be assessed for the branches corresponding to the Specifications and Location as they are dependent upon the Test results. They are obtained from experts by asking direct questions such as "What is the likelihood that the Environmental Assessment Office will specify stringent disposal specifications given that the test results indicate high concentration of toxic matter?". Prior probability of the test results are obtained by asking questions of the form "What 216 is the likelihood that the test results will indicate that the sediment contains high concentrations of toxic matter?". Percentile values that are required for the definition of the probability distributions of the primary variables are obtained in the next step. The percentiles are obtained only for the base case. The scenario that corresponds to tests indicating low concentrations of toxic matter, with the sediment being dumped in Skeena under lax specifications is considered as the base case in this example. Percentile values for the primary variables are obtained from the expert by asking question of the form "What is the possible value of the equipment productivity for the base scenario, which you feel has only a 25% chance of being exceeded?". Biases are counteracted by asking questions based upon argument types such as "Have you heard of an instance in which such equipment have been used and have shown excellent productivity levels?". The percentile values used in this example are shown in figure 6.12 as an input screen of ii Environmental Risk Analysis - [Analysis Input] i I Q.« File Window Help ' :• • - '"^^f'' . •: .... Please input the percentile values which have been elicited from: experts into the appropriate fields. For Type B risks, consider the base case whenassigning these values. • ., '• 5th • . 25th 50th 75th . 95th _ ' ANALYZE ; • ' percentile, percentile percehtile'.iperce'ntile percentile , ' '.x;\.-'s ' Selected' "° ^  sEM. St. D,ey. • Skew Kurtosis Distribution -8100 p 851. 8800 [;|9200 9 7 5 0 Ouamiiv lit Hi 884G.25! 505.244 |; 0.6 ||3.6 \ |Lognorm(8.85e+3,5. ji 1 r 1 1 Equipment Usage , — I I I J27 •  34 J38 •'•2 Productivity I I Fig. 6.12. Dam Removal Project on Ecstall River - Percentile Estimates of Primary Variables 217 the prototype computer system. The mean and the standard deviation shown for the quantity are values obtained using the Pearson approximations (see section 4.3.2). The skewness and the kurtosis are the values that correspond to the selected member of the Pearson family of curves. The probability density function corresponding to the data is selected using the procedure described in section 4.3.2. A graphical image of the probability distribution is generated for each variable and displayed to the expert in order to carry out consistency checking. The graph obtained for the sediment quantity is shown in figure 6.13. The distribution selection process has been implemented in the prototype system using Microsoft Excel® and BestFit®. As can be seen from the mean and the BestFit - [Lognorm ] Erie Edit input Execute' Eraph^Statistes.,, WJndqw. ,Melp'. D > % > % New 1 Open %tSave • Print'. AutoFiti Wizard Resuitsl iHelf Format 0.0009 T 0.0004 + 0.0000 7.0 Probability Distribution for Sediment Quantity -LogNormal(8850,504) 7.8 8.6 B.A Quan t i ty - "000 cubic meters 10.2 11.0 Fig. 6.13. Dam Removal Project on Ecstall River - Probability Distribution of Sediment Quantity 218 standard deviation, the selected lognormal distribution shows excellent agreement with the values obtained from the Pearson approximations using the five percentile values. Once the uncertainty of the primary variables are quantified for the base case scenario, adjustment coefficients are obtained from experts in order to characterize the primary variable distributions for other scenarios. Primary and Binary adjustment coefficients are obtained as input to the system. The coefficient values used for the productivity variable in this example are shown in figure 6.14. The input for the values is arranged in a matrix form in the prototype system. For example, the cell that corresponds to the row "Tests - Highly toxic" and the column "Specifications stringent" is the binary adjustment coefficient for that scenario. The location would be still at its base value 'Skeena" in this case. The symbol '*" is used to denote impossible scenarios in the input to the prototype system (see figure 6.14). Adjustment factors for scenarios in which all three external factors are active are calculated using these inputs. The base distributions are then multiplied by the adjustment coefficients to obtain the distributions for the different scenarios. The reader is referred to section 5.6 for a detailed example that deals with adjustment coefficients. The interdependency between the primary variables is investigated in the next step of the analysis. Primary variables are considered in a pair-wise manner in order to identify whether they are interdependent and if so to evaluate of the rank correlation coefficient between the two variables. In this example, the expert who provided the percentile estimates for the productivity would be asked "Would knowing the exact value of the quantity of the sediment help you in reducing the uncertainty of your estimate for the productivity?". An affirmative answer indicates that the equipment 219 X Microsoft Excel - iisktype3 [Read-Unlyl $ s : N j © £ i s a 13] HMO : & 3a <7 « * I f £ A £i s i a f ?4& Anal - 10 - B I U • i s s , : = E53 i eft «V *s8 .00 | ; r = + == =-•=:--= » /o J .00 +.0 i|(=— 7— D33 ff\ = A ' B c * D I ^ I I I I I S ,1, Risk Factor List ^ T e s t s - H i g h T c x i c L o c . -S t i k i n e Loc. - O t h e r S p e c s ; S t r i n g e n t Tests - High Toxic -0 15 -0.2 -0.25 -0 3 3 - Loc. - Stikine -0 2 0.1 * -0 25 4 Loc. - Other -0.25 -0.15 -0 3 5 Specs. Strinqent -0.3 -0.25 • . -0.3 -0 2 7 G . 27 ii!fiilililill5 Equipment Productivity Fig. 6.14. Dam Removal Project on Ecstall River - Primary and Binary Adjustment Coefficients for the Productivity Variable productivity is correlated with the quantity. In order to obtain the rank correlation coefficient, the expert is asked for the percentage reduction in uncertainty in productivity due to the knowledge of the exact value of the quantity. The reader is referred to section 4.3.3 for a detailed example on the calculation process for the rank correlation coefficient. The elicitation of correlation coefficients concludes the information elicitation process for this example. At this stage of the analysis, the analyst is in possession of primary variable distributions for each scenario, as well as the rank correlation matrix for the primary variables. As described in section 5.5.4, the same correlation matrix can be used for all scenarios. Monte Carlo simulation is then carried out for all scenarios using an equal number of iterations. This step has been implemented in the 220 prototype system using @Risk® which allows the simulation of all the scenarios to be carried out simultaneously. The simulation output values obtained for the duration is used to calculate the first four moments of the duration of the sediment disposal work package. A detailed example of such a calculation process is provided in section 5.6. The process described above is applied to all work packages that were judged as being subject to significant duration uncertainty during the risk screening process. The outputs obtained, i.e. the first four moments of each work package duration, serve as input to the overall project network shown in figure 6.8. Evaluation of the network would yield the total duration uncertainty of the project. 6.9 Summary A variety of information, not limited to probability assessments are required for the analysis of environmental risks. Due to the absence of data from previous projects, such information needs to be elicited from experts in most instances. An information elicitation process that makes use of strategies such as decomposition and the use of directed questions based on argument types, in order to ensure the overall success of the elicitation process was described in this chapter. A detailed example that shows the implementation of the risk analysis framework was also presented. 221 Chapter 7 Conclusions and Recommendations 7.1 Conclusions This thesis is devoted to the analysis of economic risks caused by environmental issues in large infrastructure projects. Many authors have dealt with the topic of economic risks in large projects (Al-Bahar and Crandall 1990, Chapman and Ward 1997, Cooper and Chapman 1987, Perry and Hayes 1985, Wideman 1986). While the presence of environmental risks has been acknowledged by almost all of the authors, economic risks due to environmental issues have even been considered as a separate risk category in some of these works (Perry and Hayes 1985). However, such works do not address the difficulties that arise in applying existing analysis methods such as simulation to environmental risks as well as the difficulties in identifying such risks. The contribution of this thesis to existing knowledge on environmental risk analysis is two-fold. Primarily, it contributes to risk management knowledge by introducing a structured risk analysis approach devoted solely to environmental risks and by presenting suitable risk quantification methods to address different types of environmental risks. The structured framework allows the user to identify, characterize and quantify the impacts of environmental issues on various project components. The quantification methods are an attempt to overcome shortcomings of existing risk quantification methods. They provide a means of overcoming disadvantages of Monte Carlo simulation in the treatment of correlations and in input distribution selection, and also provide a methodology for the treatment of external risk factors. The thesis also 222 contributes in a lesser degree to our understanding of the relationship between large infrastructure projects and the natural environment by identifying the factors that affect the project-environment relationship. The various factors that affect the interaction between project and the environment are discussed from a project economics viewpoint. The highlights of the thesis and main conclusions that follow from the research are as follows: 1. The relationship between a project and the natural environment in which it is situated, is one in which the project and the environment are mutually affected. The extent of the mutual interaction between the project and the environment is dependent upon the scope characteristics of the infrastructure project as well as the sensitivity of the natural environment. 2. The project-environment relationship is further modified by public perceptions and participation as well as by relevant government regulations. 3. Several types of environmental issues that affect the economics of the project were identified. Some environmental issues such as pollution relate to the direct interface between the project and the environment. Another type of issue arises as a consequence of existing government regulations. The third type of issue such as lawsuits arise unexpectedly, mainly due to negative public reaction. 4. The relevance of various environmental issues changes during the various stages of the project life cycle. Project life cycle breakdowns found in the existing literature were found to be unsatisfactory for providing a logical structure to accommodate the changing nature of environmental issues. A five-stage breakdown of the project life cycle consisting of Planning & Design, Construction, 223 Operational, Modification and Closure phases is presented in this thesis as a viable alternative to existing breakdown structures. 5. The environmental risk analysis framework provides for the identification and estimation of environmental risks. It is a three-stage process consisting of a Define stage, a Risk Identification stage and an Analysis stage. 6. The usefulness of maintaining a register of probable risks to aid risk identification was recognized. A preliminary risk register for large hydropower generation projects was presented. 7. The uncertain impact of environmental issues on the economic variables of the project can be quantified by analyzing their effect on individual work packages of the project. The risks were classified into four types in accordance with their effect on the project work packages in carrying out the analysis. 8. Monte Carlo simulation was used to develop the quantification method for two types of environmental risks. A novel approach to the selection of input probability distributions is presented. This approach utilizes the Pearson family of distributions. Similarly a practical method of evaluating the rank correlation coefficient between input variables is also presented in the thesis. The disadvantages of Monte Carlo simulation as used in current practice were overcome in this manner. 9. Current analytical tools are incapable of effectively incorporating the effect of external uncertainty factors in quantifying risks. The method proposed in this thesis for the analysis of the other two types of risks is based on influence diagrams and offers a viable alternative for incorporating external uncertainties. 224 10. Information elicitation from experts plays a vital role in risk analysis of large projects due to the limitations on the availability of data from previous projects. The richness and the quantity of information obtained from experts can be expanded with the use of training procedures, directed questions based on argument types and the use of appropriate feedback. 7.2 Suggestions for Future Research The environmental risk analysis framework introduced in this thesis is an important contribution to the process of treating environmental risks in large infrastructure projects. This section presents a brief introduction to future research that can be carried out to further consolidate this framework as well as the developments that can be based upon the research described in this thesis. The steps that can be taken to consolidate the work described in this thesis are: a) The risk identification process described in this thesis uses a risk register as its starting point. A preliminary risk register intended for use in large hydropower projects is presented as part of this thesis. Similar risk registers need to be developed for other major types of projects such as highways, bridges, tunnels, airports, nuclear power plants, thermal power plants, water treatment plants, offshore oil platforms, and harbours. b) The modeling paradigm for the analysis of the combined effects of multiple external uncertainties presented in Chapter 5 is an important advancement in the ability to treat the uncertainty causing agents external to the project variable. However, validation of this approach needs to be carried out with the aim of improving and modifying this paradigm. 225 c) A significant step in the future use of this framework would be its incorporation into an overall decision support system. Such an incorporation should be accompanied by the development of a comprehensive set of risk mitigation strategies for environmental risks that can be considered at the overall project decision making level. The developments that can be based upon this research include: a) This thesis presents an analysis of environmental risks, which is one of the many risk categories inherent in large infrastructure projects. Several other types of risks such as financial risks due to exchange rate fluctuations and inflation, legal risks, and political risks are also subject to uncertainty caused by external factors similar to Type B environmental risks. Risk analysis frameworks that can be used for their identification and quantification can be developed based on the research work under discussion. b) This research can serve as the platform for the development of an intelligent system for environmental risk analysis. The framework presented herein relies on the user to identify the risks that are applicable to each individual project and its location, and to select the risks that are deemed to be significant. An intelligent system that assists the selection process given the characteristics of the project and the natural environment in which the project is located would reduce the subjective human input that is required. Such a system would also drastically reduce the time required to carry out the analysis as well as the associated costs. 226 Bibliography Abdel-Aziz, A . M . 2000. Generalized economic model, risk analysis framework and decision support system for the analysis and evaluation of capital investment projects. Ph.D. thesis. The University of British Columbia, Vancouver. AbouRizk, S.M. and Sawhney, A. 1993. Subjective and interactive duration estimation. Canadian Journal of Civil Engineering. 20 (3), 457-470. Al-Bahar, J.F. and Crandall K.C. 1990. Systematic risk management approach for construction projects. Journal of Construction Engineering and Management. 116(3), September. 533-546. Allison, L . M . 1981. Sensitive areas : literature review. Mackenzie River Basin Study Report. Supplement 1. Amos, D.E. and Daniel, S.L. 1971. Tables of percentage points of standardized Pearson distributions. Research Report # SC-RR - 710348. NTIS, Virginia. Andradottir, S. and Bier, V . M . 1998. An analysis of decomposition for subjective Estimation in decision analysis. IEEE Transactions on Systems, Management and Cybernetics - Part A: Systems and Humans. 28(4), 443-453. B.C. Hydro. 1976. Revelstoke Project Environmental Impact Statement. B.C. Hydro. Environmental Resources. 1990. Alouette Spillway and Low Level Outlet Rehabilitation. Environmental Management Plan. Report No. ER-90-11. B.C. Hydro. Environmental Resources. 1990. Williston Substation to Kelly Lake Substation 500 kV Transmission Line. Summary of Environmental Impacts and Environmental Protection Plan. Report No. ER-90-04. B.C. Hydro. Environmental Resources. 1991a. Fraser Valley System reinforcement Project. Environmental Planning and Assessment Report. Report No. ER-91-11. B.C. Hydro. Environmental Resources. 1991b. Seven Mile Reservoir Raising Environmental Impact Assessment and Compensation Proposals. B.C. Hydro. Environmental Resources. 1994. Evaluation of the Installation of Unit Five at Revelstoke Generating Station. Benjamin, J.R. and Cornell, C A . 1970. Probability, statistics and decision for civil Engineers. New York: McGraw-Hill. 227 British Columbia. 1996a. Ecological Reserves Act. Statutes of B.C. R.S.B.C.1996. c.103. British Columbia. 1996b. Environment Management Act. Statutes of B.C. R.S.B.C.1996. c.118. British Columbia.- 1996c. Environmental Assessment Act. Statutes of B.C. R.S.B.C.1996. c.119. British Columbia. 1996d. Forest Act. Statutes of B.C. R.S.B.C.1996. c.157. British Columbia. 1996e. Forest Practices Code of British Columbia. Statutes of B.C. R.S.B.C.1996. British Columbia. 1996f. Health Act. Statutes of B.C. R.S.B.C.1996. c.179. British Columbia. 1996g. Land Act. Statutes of B.C. R.S.B.C.1996. c.245. British Columbia. 1996h. Petroleum and Natural Gas Act. Statutes of B.C. R.S.B.C.1996. British Columbia. 19961. Waste Management Act. Statutes of B.C. R.S.B.C.1996. c.482. British Columbia. 1996j. Water Act. Statutes of B.C. R.S.B.C.1996. c.483. British Columbia. 1996k. Wildlife Act. Statutes of B.C. R.S.B.C.1996. c.488. British Columbia. Environmental Assessment Office. 1995. Guide to the British Columbia Environmental Assessment Process. Brockriede, W. and Ehninger, D. 1960. Toulmin on argument: An interpretation and application. Quarterly Journal of Speech. 46, 44-53. Browne, G.J., Curley, S.P. and Benson, P.G. 1997. Evoking information in probability Assessment: Knowledge maps and reasoning-based directed questions. Management Science. 43(1), 1-14. Bury, K . V . 1975. Statistical Models in Applied Science. New York: John Wiley & Sons. Canada. 1985a. Arctic Waters Pollution Prevention Act. Statutes of Canada. R.S.C. 1985. c.A-12. Canada. 1985b. Canada Water Act. Statutes of Canada. R.S.C.1985. c.C-11. Canada. 1985c. Canada Wildlife Act. Statutes of Canada. R.S.C.1985. c.W-9. Canada. 1985d. National Parks Act. Statutes of Canada. R.S.C.1985. c.N-14. 228 Canada. 1985e. Navigable Waters Protection Act. Statutes of Canada. R.S.C.1985. c.N-22. Canada. 1992. Canadian Environmental Assessment Act. Statutes of Canada, c.37. Canada. 1996. Oceans Act. Statutes of Canada, c.31. Canada. 1999. Canadian Environmental Protection Act. Statutes of Canada, c.32. Chapman, C.B. and Ward, S. 1997. Project Risk Management. Chichester: John Wiley & Sons. Chau, K.W. 1995a. The validity of the triangular distribution assumption in Monte Carlo Simulation of construction costs: empirical evidence from Hong Kong. Construction Management and Economics. 13 (1), 15-21. Chau, K.W. 1995b. Monte Carlo simulation of construction costs using subjective data. Construction Management and Economics. 13 (1), 369-383. Clarke, C E . 1962. The PERT model for the distribution of an activity time. Operations Research. 10 (3), 405-406. Cooke, R . M . and Waij, R. 1986. Monte Carlo sampling for generalized knowledge dependence with application to human reliability. Risk Analysis. 6(3), 336-343. Cooper, D.F. and Chapman, C.B. 1987. Risk analysis for large projects: models, methods, and cases. Chichester. John Wiley & Sons. Curley, S.P. and Benson, P.G. 1994. Applying a cognitive perspective to probability construction. In Wright, G. and Ayton, P. (Eds.). Subjective Probability. Chichester: John Wiley & Sons. Diekmann, J., Featherman, D., Moody, R., Molenaar, K. and Rodriguez-Guy, M . 1996. Project cost risk analysis using influence diagrams. Project Management Journal. 27(4), 23-30. Diekmann, J.E. 1997. Risk analysis for environmental remediation projects: A comprehensive approach. Cost Engineering. 39(8), 38-44. Diekmann, J.E. and Featherman, D.W. 1998. Assessing cost uncertainty: Lessons from environmental restoration projects. Journal of Construction Engineering and Management. 124 (6), 445-451. Environment Canada 1999a. Categories of the committee on the status of endangered Wildlife in Canada. Available from http://www.ec.gc.ca/hww-fap/COSEWIC. html. Internet. Accessed 12 November 1999. 229 Environment Canada 1999b. Preserving Canada's wetlands. Science and the Environment Bulletin. Issue No. 12, May/June. Bulletin on-line. Available from http://www.ee. gc.ca/ science/sandemay99/article3_e.html. Internet. Fackler, P.L. 1991. Modeling interdependence: An approach to simulation and elicitation. American Journal of Agricultural Economics. 73(4), 1091-1097. Farnum, N.R. and Stanton, L.W. 1987. Some results concerning the estimation of beta distribution parameters in PERT. Journal of the Operational Research Society. 38, 287-290. Finch, E. 1992. Environmental assessment of construction projects. Construction Management and Economics. 10, 5-18. Foulke, J.E. 1994. Mercury In Fish: Cause For Concern? FDA Consumer. 28 (7). Fujian Province Electric Power Bureau, and Shanghai Investigation, Design and Research Institute 1995. Summary Environmental Impact Assessment of the Proposed Fujian Mianhuatan Hydropower Project in the People's Republic of China. Gebotys, R.J. and Claxton-Oldfield, S.P. 1989. Errors in the quantification of uncertainty: a product of heuristics or a minimal probability knowledge base. Applied Cognitive Psychology. 3, 157-170. Genest, C. and MacKay, J. 1986. The joy of Copulas: Bivariate distributions with uniform marginals. The American Statistician. 40(4), 280-283. Gifford, D. 1997. The value of going green. Harvard Business Review. Vol.75 .No.5, September/October. 11-12. Glasson, J., Therivel, R., and Chadwick, A . 1994. Introduction to Environmental Impact Assessment. London: U C L press. Gleick, P.H. 1992. Environmental consequences of hydroelectric development: The role of facility size and type. Energy. 17 (8), 735-747. Golenko-Ginzburg, D. 1988. On the distribution of activity time in PERT. Journal of the Operational Research Society. 39 (), 767-771. Henrion, M . , Fischer, G.W. and Mullin, T. 1993. Divide and conquer? Effects of Decomposition on the accuracy and calibration of subjective probability Distributions. Organizational Behavior and Human Decision Processes. 55, 207-227. Hora, S.C. 1992. Acquisition of expert judgement: Examples from risk assessment. Journal of Energy Engineering. 118(2), 136-148. 230 Hora, S .C, Dodd, N .G. and Hora, J.A. 1993. The use of decomposition in probability Assessment of continuous variables. Journal of Behavioral Decision Making. 6, 133-147. Hora, S .C, Hora, J.A. and Dodd, N.G. 1992. Assessment of probability distributions for continuos random variables: A comparison of the bisection and fixed value methods. Organizational Behavior and Human Decision Processes. 51, 133-155. Howard, R.A. 1988. From influence to relevance to knowledge. In Influence Diagrams, Belief Nets and Decision Analysis, ed. R . M . Oliver and J.Q. Smith. 3-23. Chichester: John Wiley & Sons. Howard, R.A. 1989. Knowledge maps. Management Science. 35(8), 903-922. Howard, R.A. and Matheson, J.E. 1983. Influence diagrams. In Readings on the Principles and Applications of Decision Analysis, ed. R.A. Howard and J.E.Matheson. 2, 719-762. Menlo Park: Strategic Decisions Group. Hummel, M . 1995. Protecting Canada's endangered spaces. Toronto: Key Porter Books. Hydro-Quebec. 1993. Grande - Baleine Complex. Feasibility Study. Iman, R.L. and Conover, W.J. 1982. A distribution-free approach to inducing rank correlations among input variables. Communications in Statistics. 11(3), 311-334. Iman, R.L. and Davenport, J .M. 1982. Rank correlations plots for use with correlated input variables. Communications in Statistics. 11(3), 335-360. Johnson, N.L. , Nixon, E. and Amos, D.E. 1963. Table of percentage points of Pearson curves, for given V p i and P2, expressed in standard measure. Biometrika. 50, 459-498. Kamburowski, J. 1997. New validations of PERT times. Omega. 25 (3), 323-328. Kangari, R., and Boyer, L.T. 1987. Knowledge based systems and fuzzy sets in risk management. Microcomputers in Civil Engineering. 2, 273-283. Keefer, D.L. and Verdini, W.A. 1993. Better estimation of PERT activity time parameters. Management Science. 39 (9), 1086-1091. Kottas, J.F. and Lau, H. 1978. Stochastic Breakeven Analysis. Journal of the Operational Research Society. 29 (3), 251-257. Kottas, J.F. and Lau, H. 1982. A four-moments alternative to simulation for a class of stochastic management models. Management Science. 28 (7), 749-758. 231 Lau, H. and Lau A . H . 1998. An improved PERT-type formula for standard deviation. HE Transactions. 30 (3), 273-275. Law, A . M . and Kelton, W.D. 1991. Simulation modeling & analysis. New York: McGraw-Hill. Lens, J.E., Simon, R .M. , and Martin, P.J. 1991. Wrapping up environmental issues in an embankment. Preparing for construction in the 21st century : proceedings of Construction Congress '91. 289-294. Leung, H .M. , Chuah, K .B . and Rao Tummala, V . M . (1998). A knowledge-based system for identifying potential project risks. Omega, Vol.26, No.5, 623-638. Lorance, R.B. 1992. Contingency draw-down using risk analysis. Transactions of the American Association of Cost Engineers. 1. MacCrimmon, K.R. and Ryavec, C A . 1964. An analytical study of PERT assumptions. Operational Research. 12 (1), 16-37. Merkhofer, M.W. 1987. Quantifying judgmental uncertainty: Methodology, experiences and insights. IEEE Transactions on Systems, Management and Cybernetics. SMC-17 (5), 741-752. Mirkhani, K. , Chitra, S. and Lazur, J. 1987. Mixed moments estimation for nonnormal analysis. Transactions of the American Society for Quality Control Quality Congress. 41, 539-545. Morris, P.W.G. 1994. The Management of Projects. London: Thomas Telford. Moselhi, O. and Dimitrov, B. 1993. Monte Carlo techniques with correlated random variables - discussion. Journal of Construction Engineering and Management. 119(3), 658-661. Mular, A . L . 1982. Mining and Mineral Processing Equipment Costs and Preliminary Capital Cost Estimations. Montreal: The Canadian Institute of Mining and Metallurgy. Ofori, G. 1992. The environment: the fourth construction objective? Construction Management and Economics. 10, 369-395. Ord, J.K. 1985. Pearson system of distributions. Encyclopedia of Statistical Sciences. 6, 655-659. Palisade Corp. 1995. BestFit User's Guide. Newfield, N Y . Palisade Corp. 1997. Guide to Using @Risk. Newfield, N Y . 232 Pearson, E.S. and Tukey. J.W. 1965. Approximate means and standard deviations based on distances between percentage points of frequency curves. Biometrika. 52, 533-546. Perry, J.G., and Hayes, R.W. 1985. Risk and its management in construction projects. Proceedings of the Institution of Civil Engineers, U.K. 78, 499-521. Project Management Institute. 1996. A guide to the project management body of Knowledge. Public Works & Government Services Canada and Strait Crossing Development Inc. 1996. Environment - The Northumberland Strait Bridge Project. Rainey, W.S. 1997. Fish ladders. Scientific American. October. Journal on-line. Available from http://www.sciam.com/1097issue/1097working.html. Internet. Ranasinghe, K . A . M . K . (1990) Analytical method for quantification of economic risks during feasibility analysis for large engineering projects. Ph.D. thesis. The University of British Columbia, Vancouver. Ranasinghe, M . and Russell, A .D. 1991. Economic risk analysis of large engineering projects. Canadian Journal of Civil Engineering. 18, 416-427. Ranasinghe, M . and Russell, A .D. 1992. Treatment of correlation for risk analysis of engineering projects. Civil Engineering Systems. 9, 17-39. Ranasinghe, M . and Russell, A.D. 1993. Elicitation of subjective probabilities for economic risk analysis: An investigation. Construction management and economics. 11,326-340. Rao, G.N and Grobler, F. 1995. Integrated analysis of cost and schedule risk. Computing in Civil Engineering, Proceedings of the Second Congress held in conjunction with A/E/C Systems '95. 2, 1404-1411. Ravinder, H.V., Kleinmuntz, D.W. and Dyers, J.S. 1988. The reliability of subjective probabilities obtained through decomposition. Management Science. 34(2), 306-312. Rubinstein, R.Y. 1981. Simulation and the Monte Carlo Method. New York: John Wiley & Sons. Russell, A . and Froese, T. 1997. Challenges and a vision for computer-integrated management systems for medium-sized contractors. Canadian Journal of Civil Engineering. 24, 180-190. Russell, A .D. and Ranasinghe, M . 1991. Decision framework for fast-track construction: A deterministic analysis. Construction Management and Economics. 9, 467-479. 233 Russell, A .D . and Ranasinghe, M . 1992. Analytical approach for economic risk quantification of large engineering projects. Construction Management and Economics. 10, 277-301. Sadar, M . H . 1996. Environmental Impact Assessment. Carlton University Press Inc. Sadar, M . H . and Dirschl, H.J. 1996. Generic environmental impacts identified from water impoundment projects in the western Canadian plains region. Impact Assessment. 14,41-57. Schuman, J.R. 1995. Environmental considerations for assessing dam removal alternatives for river restoration. Regulated Rivers: Research & Management. 11,249-261. Smith, G.L. 1991. Canada's changing impact assessment provisions. Environmental Impact Assessment Review. 11 (1), 5-9. Smith, G . M . 1990. Fisheries Values and Potential Construction Impacts of the Kelly Lake-Cheekye Transmission Line Project. Vancouver: B.C. Hydro. Environmental Resources. Smith, G.R. and Bohn, C M . 1999. Small to medium contractor contingency and Assumption of risk. Journal of Construction Engineering and Management. 125 (2), 101-108. Songer, A.D. , Diekmann, J. and Pecsok, R.S. 1997. Risk analysis for revenue dependent infrastructure projects. Construction Management and Economics. 15, 377-382. Spetzler, C S . and Stael von Holstein, C. 1975. Probability encoding in decision analysis. Management Science. 22(3), 340-358. Strait Crossing Inc. 1992. Draft Environmental Management Plan for the Northumberland Strait Crossing Project. The Institution of Civi l Engineers and the Faculty and Institute of Actuaries. 1998. Risk Analysis and Management for Projects. London: Thomas Telford. Touran, A. 1993. Probabilistic cost estimating with subjective correlations. Journal of Construction Engineering and Management. 119(1), 58-71. Touran, A. , and Suphot, L. 1997. Rank correlations in simulating construction costs. Journal of Construction Engineering and Management. 123 (3), 297-301. Touran, A. , and Wiser, P. 1992. Monte Carlo technique with correlated random variables. Journal of Construction Engineering and Management. 118 (2), 258-272. Uher, T.E. 1996. A probabilistic cost estimating model. Cost Engineering. 38 (4), 33-40. 234 United Nations Environmental Program (UNEP). 1980. Industry and Environment. Special Issue, N o . l . United States Fish and Wildlife Service 2000. Threatened and Endangered Species. Available from http://endangered.fws.gov/endspp.html. Internet. Accessed 13 June 2000. van Dorp, J.R. and Duffey, M.R. 1999. Statistical dependence in risk analysis for project networks using Monte Carlo methods. International Journal of Production Economics. 58, 17-29. Vanegas, J.A. and Smith, G. 1994. Multimedia technologies for civil and construction engineering professional practice. Computing in Civil Engineering, Proceedings of the First Congress held in conjunction with A/E/C Systems '94. 1, 592-599. Wall, D . M . 1997. Distributions and correlations in Monte Carlo simulation. Construction Management and Economics. 15, 241-258. Walther, J.B. 1996. Computer-mediated communication: Impersonal, interpersonal and hyperpersonal interaction. Communication Research. 23 (1), 3-43. Wideman, M.R. 1986. Risk management. Project Management Journal. September, 20-25. Williams, T. 1995. A classified bibliography of recent research relating to project risk management. European Journal of Operational Research. 85 (1), 18-38. Wood, C. 1995. Environmental Impact Assessment: A Comparative Review. Essex: Longman Scientific & Technical. 235 Appendix A Preliminary Risk Register for Large Hydropower Projects This appendix presents a preliminary risk register intended for use in large hydropower projects. It contains a list of probable work packages of a generic hydropower project, that arise or are affected by environmental issues, and are of uncertain cost and duration. The work packages that may arise during all five stages of the project life cycle have been considered. The risk register is the outcome of a comprehensive literature review. Several Environmental Impact Statements and environmental management plans (B.C. Hydro 1976,1990,1991a, 1991b, 1991c, 1994, Fujian Province Electric Power Bureau, and Shanghai Investigation, Design and Research Institute 1995, Hydro-Quebec 1993, Strait Crossing Inc. 1992), Legislation, and other documents (Gleick 1992, Schuman 1995, Smith 1990, Sadar and Dirschl 1996) were reviewed in this study. While most of the work packages have been treated and described taking into account the natural, legislative and social environment of the Province of British Columbia, it is stressed that the register can easily serve as the starting point in environmental risk analysis of a hydropower project in any part of the world. The register which has been prepared in a format suitable for incorporation into a computer system, contains a brief description of the environmental issue(s) that affect the work package as well as the project phase(s) in which it is most likely to occur. Generic work packages have been identified in instances where an environmental issue affects several work packages simultaneously. A generic work package groups together several work packages that share a common characteristic(s) 236 which leads to them being affected by a particular environmental issue. An example would be an instance in which work packages which share the common characteristic of "generates a large amount of noise" being suspended or curtailed during the migration period of important species of wildlife such as moose. Work packages that share the common characteristic of "generates a large amount of noise" would involve excavation, land clearing, aggregate production in quarries and blasting of tunnels. It is noted that work packages that arise due to environmental issues may themselves match the characteristic of a generic work package. This would effectively mean that the work package cost and duration is subject to the uncertainty caused by several environmental issues. The register also includes a suggestion for the risk type in accordance with the risk type definitions provided in Chapter 4 of this thesis. The use of the risk register is illustrated through an example. Consider the construction of a large dam in Montana, USA in an area known as a habitat of the Bald eagle. The Bald eagle is classified as a threatened species (United States Fish and Wildlife Service 2000). Subsequent to preparing a definition of the project and the natural environment, and analyzing public perceptions as set out in Chapter 3, the risk analyst would browse through the risk register in order to carry out risk identification. Each listing in the register would be studied individually in order to determine its applicability to the project in question. In this example, browsing through the risk register (see figure A . l ) will lead the user to identify "Relocation of nests of endangered species" as a work package that is relevant to the project due to the presence of Bald eagles in the project area. Nests that are located in the area which will be inundated by the dam, as well as nests found in areas used for the construction of 237 project components will need to be relocated to areas that will be left undisturbed by the project. As the area has a high concentration of Bald eagles, the relocation of some nests would have to be carried out with certainty. In this case the analyst would use the suggested risk classification of A l (see figure A. l ) . If however, the concentration of Bald eagles is estimated to be very low, which makes the relocation of nests an uncertain event the analyst would wish to change the risk classification to B l in carrying out further analysis of risks. Work Package with Uncertain Cost and Duration Description of Environmental Issues that Create / Contribute to the Work Package Cost and Duration Relevant Project Phases Risk Type Relocation of nests of endangered and threatened species Relocation of nests from areas that will be affected by the project to undisturbed areas is carried out in the case of endangered and threatened bird species Construction, Modification A l Rockblasting Limit blastino operat'™'- * ' v - " " " ^ — Fig. A. 1 - Extract from Risk Register for Large Hydropower Projects The risk register is presented below. 238 C . 0> o a OS > ca o cu 3 * 5 O U a* « « PH •r: t*-1 © 5 5 « .2 ^ "C «^ cn 5 « O |Q U CU a « 0 ' • a PH © Q fi 03 •«-« o U fi OO CD Q o CD o1 !-H O H CD GO . OO a o w ' I O fi P Cd (L) k . a -fi ^ CD tN fi O fi op oo <u Q fi £ Of) tt. oo 5 O fi o cd S o " c t n O " >>T3 fi > o <u ^ t£3 fi o rtd .a 1*3 "S "o o fi CO ^ «3 £ o c oo $ «3 t £ .fi CD *3 • T—i fi cr <u *-* ed £ fi < oo CD S -fi — CD <D ° > fi o 2 X ! P 2 i—H -4—• Cd Kl cd £ £ o « M I l l CD O fi O o o a .5 e-i—i o .a a o a £ •55 S CD oo Q 2 fi 2P OO CD fi O CD O » ^ 4—* a^  DC g S w c3 O H a ^ CD > i .2 OH oo oo rt fi cd JS --fi DH fi CD oo CD - f i <D O «4-l o B CD -fi H ei o '3 o O H B o o o +-> oo oo CD fi oo oo 73 cfl s § oo \fi fi cd 03 O 8-1 OH a 2 £ OH 00 CD .3 a T3 CD oo O c3 >< o fi fi « o ^ • r t O H 8 H ~ ^ cS X o - a fi * O H tt J-. rt Kt U rH U n « Pi w O cd cd T3 CD (D *T3 , 7 3 fi fi oo a CD oo CD 239 0> -H SB 2 5 s CN CN ' CO Q 'co Q °0O Q .SP ' 'SI IQ a .2 « "S es is *- Q U : c I 93 93 o U <u W> PH e S ^ C « S « ii © © U 93 >l o d _ „ 3 S 1 -p GO ?3 GO C (j. g a © * . O 'lo '_' ~ ed 9S M ,2 . g ! S 3 | oo -G "S -2 <'J » u! . <-i—i vi "•^  »> 4) <D <D O o 5 c ^ <U <u > 93 „•> l i s t ! 0) -4-» >-> Vi 93 •— T P P O U d 5 -S « 5 «8 93 > '+J 93 (D o B 2 * S > _ l Vi Vi rrt g n . P O <D > ^ I-I (U I> C 93 ts a 9-a u P3 O -t-» i-sa a S-i p u 1 ^ 3 s o < 93 * c « a 93 PH e it s o C3 O o o o 1 Vl <D IQ 2 o 93 O ' • a o d o c a .2 .SPt5 GO (U 240 Vi Vi ii Vi I 7 3 to Q CQ .SP 'GO Q d SP GO d~ - d .2 ^ .2 tS g ts =j .2 o <u <u O & 5 J2 I Q U O S U CQ UH &o d  o GO 00 « C g X o <u o cu 5 — Q U O S d o . 9 o Sr! -o oo GOO u c ,© u ft 03 73 •*H «B o U o M ) S3 w 73 OH J3 W +H 5 « a * E i vi c v o ft U X ) 3 1) X> 1) T3 £ 2 T3 OJ ii Cl GO CU 2 «J <+H O GO d o ts 3 0) cd GO o Cu GO , 1 1 Cl b o O C J xl ^ GO -o c3 O >-H S-l GO 03 <u fc CU t^n i l 3 <H-H Ci 5 2 •a ^ » -2 oo g | xl ^ x l to b <+-< 3 o cr 3 2 a ffi ^ | g u S ti oo £ is 5 a 3 > 00 £ o u _i> 00 > O VO Cl > " H + H 2 <4-H X e ^ £ H o H d> a ci .2 GO O <u 3 00 T3 fi 2 GO Cl T3 CU 03 o ii ii x i 2 ^ Cu o X> S3 n ii u c & «> 93 0D S. 73 S ft 73 -O OH Si _* * ^ u Cu ii GO i s 'o <D CU V3 C3 00 Cl g +-» x O cd GO •s g 2 2 s g Cu o & oo cl ^ • g ii cd ?3 , £ ^ O GO ii 00 C3 ^H Cu to fc * J O GO ^ 'C ^ a fc Cl c3 <U X J o u Cl M -2 GO 0 S l-l ° 1 5 5 s -d ^ a •B w X l C3 ii > o 52 H 93 X l X ! CU Cu 2 3 2 a E? ii * w o | 52 c l l S u e Ir! so 93 QJ cd t-i Cl 00 o s o ^ a ct, cl o GO « 2 S • -S s . ^ § tH ^ 93 CU GO 'S 0> GO d ow Q GO O d .a *^ H 93 > oo d 93 <U 241 cu 2 o a i » « ON a o £ 3 O O S o CQ 2 13 .2 p d *j Q .2 S • a "3 -2 cd o o S o 1 S.3 JS u o 2 u c _© « h CU 9 05 S u — 05 03 a 2 2 ° , a cu o © S « .2 -3 i- i 50 s <u o |Q U CD <H-I o co CD CD cd CD pfi >> -ct CD CO 3 CO fi H-" u CD CD T3 fi cd o CD 'o' > fi ^ CD OH cd .fl OH CO cd o CD .fl O CD 3 o CO O co o CD cd £ 6 t: cu u fi p & « W) s-I-* o « -a I OH C u -*-© 2 1 * 5 CD fi "C-CD 60 cd cd - O CO CO CD PH T3 CD I-i CO 3 ^ 6 & ° u 2 CD T3 2 3, ° « 7 3 CO CH <U 3 'Eo "53 « cS 6 0 * Cd <4H o ° +5 CO CO CD <w -"fi CO O "fi VH f* 60 C C« _H >> ^  jfi 242 CU a cu o a « CO • Si \J2 £ tf CQ .2 13 .2 -£ c a O o 1 :3 fl .2 o CD co ^ 5; § 8,1 J u o S u CO 0 CD U O ^ U „ - fi 2 "ra .2 S C a o o ra p .2 o CD 2 t3 £ 3 O H 2 2 .2 "ra .2 -fi fi a o o ra B td « 3 1 8 . 1 5 U O ^ U 2 ra .2 £ fi a o o ra p .2 o CD § e l l J fi . g 2 "ra .2 -fi c -a o o ra fi .2 O (U 2 t3 £ 3 <o 1-1 -fi m u o S u fi CU 3 a co O U cu W) C3 PH © C <u W J S w -«H> s « B-rS u *• CO 5 cu o |Q U to ra 11 ra t+H -^ 3 o cu CD OX) JO — fi 9 o - f i <-> 3 CU ed O o . o CO CD OX) cd a cd T3 o -fi AS. o BH fi 00 53 CD OH co til? cu u O -73 - O ! 8 » 13 CD co cd o cd cd o 2 <u cd .a a cu co ra S sS s x i § cu <u ~ cd & cu cu <u a tt T3 ? 3 ra jfi ^ "2 S 0 c 3 * g » co ra f^i cd 2 -tt =3 u g g .fl £ co O cu CU co > cu cd \fi £ ra OH O co cd co fl S co s cu I- l 2 « -x3 +-> CO CO s a * <2 CU <H-H a o O CU LO O fl B 3 cu cu o CU U fi -B C T 1 o b f l B fi cd cu fi c^  l-H CU cd ^ CO CO -H cd co U .3 2 t-1 cd C3 t CU y fl •B _ S B * .2 UD >-, 05 fl I-go PH fl ^ « cu fl cu cd > cu g o cu fij "ra co to .—1 eu 2 cd cd CO O OH co cu CU co C*H O fl O •ri co ^ -2 fi- +-» cd -fi H—» *—H co o B tS OX) T 3 .a cu § Cd rr, O ft o t OH fi Jt* CD fi PH PH CU co co CD S co 1 1 c2 I eu a OB fi jfi op H^-H eu CD .tt I- i CO £ B fi B 2 fi 'co fi •r OH ^ * f i O B PH CU 243 ii a >> H 2 u a> •<? s-a OO a IX! CS CQ „ c 3 .2 3 O cj § His u o ^ o CQ 2 3 .2 fit! _ o cd 3 .2 O <u g 8. | J u o S u „ - fi 2 "3 .2 -3 fi ts o o cd £ t3 £ 3 1 8 . 1 J U O ^ O 3 _r 5 .2 3 .2 o fc cd s .2 o u +2 cd rH 3 52 ft "3 55 U O S U r*i ' u o ^ u ,2 3 .2 -J3 fi -a O o cd 3 .2 o CJ 2 ta 5« 3 CJ IH 53 ft *3 oo u o S u 2 "cd .2 -rf fi « cd . S o u -rf cd r s fi ^ O +S fS « £« 00 o DX) CJ ii Cu i o a a W -3 *H S » =H:2 t- is 00 5 Q> O © U bO CJ cj 3 a ^ fi CU H oo _4> fj . O o a l i l t Q. ° - ii ^=3 S 3 ^ ' S cd o oo oo ii g S bO 5 fi ^ . .H cd a T3 oo oo CJ C T3 CJ Cd OH C§ C CD 3 ' ^ ' i ca • a) g ii fi oo 3 o -ct u cj * 3 o- * tH CJ c u x 2 o OH 7 3 ° CJ CJ fi I £ S 2 fi O O D ^ TJ an cd X l CJ "cd f u cj C s & o o o ~ > T3 > ii oo O Cu oo O = £ "3 "3 oo .2 bo cd & 1 <S p, c/a <u 2 CH 2 2, CJ „ CJ Cu cj CD C*H cd oo CJ cd > % cd o f b£) .s >» o CJ tH bfi oo fi cj C /-A ^ 3 O -H3 -q X o 3 oo O ^ C+H ,CJ tn cd cj oo X i <4H H O "2 00 . . o . O CU ow oo 2 -3 53 T3 5H CJ tn cd o _ H—» O oo 2 x g 2 S « X ) CJ fi 3 cd oo o cj OT to •fi cd O £ CS u e p £ 3 * - s ft. w W) fc. CS s y I—I CS -3 CM C! it rt -I S T3 oo G IH — CJ CJ F—H tn 3 cd O ^ "C +5 T 3 CJ w fi trt 1 1 1 « !i y 4-> o cd CD 2 X C*H O Id oo O . Cu CJ oo +i T3 T3 H 00 53 3 CJ o bX) T3 2 c3 +-» cd CD X O oo <4H o "cd 00 O Cu 00 cj %s fc oo O cd CD > ii ccj 00 3 O CJ * a $ <+4 o cd 00 Cu o oo ^ a I CJ b O ' fi O Cu oo cj CJ oo . S . . 4H !2 3 fd x : s O ^ C/l o o o 244 CU p. CO s fl OS CU tf PH CQ 2 "ra .2 -S fi t i CD o ra fi .2 O CJ 1 8 , ! 3 u o S o CN CQ 2 "ra .2 fi fi a 0 o ra p .2 o CD 1 a s j u o S u 2 "ra .2 fi -fi o o ra s i a i 8 8.1.1 U O ^ U CN u o S o fi r- fi ,2 "ra .2 CO •2 2 « A _© s « 05 -<u fl fl 05 es H>* o U cu M ) 95 H* CD 93 PH H* S-© cu © o s « .2 -g U »H w 5 CU O O U CD -fl T 3 pO o CO .2 13 -fi cc! CD 2 s-. H CD H—: i_i CO CD -fij CD -fi O O - f i CD CD o 2. o £ o „ CD fi, >; M O g S ft u 5 B a CD CD CD CD T 3 fi O ra v., .•fi CD ft—. o-fi 00 Cd a rfi o ft o CO — a ' f t O ft H - ' CO - ~ ^ * IT! T> ra ^ Cd CD fi- .2, CD C) — ft CD CD CD < 2 15 £ w _I fi <U fi co 3 fi CD cd X ) CD CQ CD 00 cd M CD cd ft O <4H o fi o i-H CD CD X w .O -J-J "+H cd * i co C4-H o CD fi CD O T3 O t: cu CJ fl & £ fl £.2 6X3 i -93 fl H* Q CU ^ 93 7 3 PH fl cd s-i 00 o s-. ft o o T3 CD CD CQ CD 00 cd CD cd OH Lo' .2 O co ^ CD CD ft ra fi § CD .fi O U T 3 CD Cd - & cd fi CD fi X fi O to CD _a fi !a 'rfi CD co CD > CD CD fi o r—I CD 00 fi •> o CD co CD > > fi cd CD 3 U CD 00 ra CD cd ft co" M .2 O co ^ CD CD •*-> fi § a -fi o o CO CD > O CD CD -fi CD co cd co cd X -Ct co CD a CD C '5 -fi CD (D +^  | & f <4H O fi o CO p~> CD JH CD cd 00 g g CD a . a -fi =2 T3 -5 CD O '5 - H .s ? o w CD o Ufi CD -fl 'c? "g CD o o co r . fi - a S o -a .SP o Si a -3 ra 245 a e 03 I C3 GO CN CQ U O S I .2 "S .2 o S cd = -2 o <u -4-» cd r3 CO l_i i l l CN < P 13 .2 o g tS 3 . 2 O o cj T3 oo -2 o o S u ci o • P H S3 fc. 3 Q •a a a +-GO o U H H WD — C3 C3 Sal I o 03 4) a u. © 2 5 « .2 -3 GO B V o o -° x ^ •p * S D i _ cd fc 3 cd c3 P .2 2 "cd t> x> xi ! H 3 o § CJ .3 3 £ H w I g cj SH o GO 3 O M GO '-3 S cj * 1 tS GO 3 - ° _ H 3 o cS 0 cd GO • g e t s 0 O o B > bO O CJ 15 3 cu O \3 3 > to G M £ * ° P U S cj 3 .2 2 o CJ 3 io W> "P 2 c3 c f S l 3 X ) cd -p "IsCJ DO'S *-> (11 *^  .CJ CJ , * H - VL* cd « -g 1 - 2 ^ CJ O CJ 50 3 60 cd T3 cd 2 2 - 3 2 cd a cd ft.S ft bO <4-< o cj 3 2 'ts tH T 3 bO CJ •=! cj £ 3 3 -3 3 I . a Q -P ' T3 CJ CJ 00^ & £ o o> £ 1 3 ° s § 3 ? 'o •2 S S 3 tH CH 3 O to O CH _ C ! " a 1 CJ 3 cj X 246 CU tf cu o u s > co CU cu 43 tf £ CN CQ fl o • i-H c o ts "—- >.\J ., , . B 5^  O O ft o u o S t3 rS fl" fl o „ o -S fl '-ts Q 2 S - t3 £ c S o u o S a o ss -cu 3 , Q CA o co cu fl W l l & « ^ *» PH s , fl CU 1 o o 5 » .2 -g U £« co H cu © Q U CC! o cd w > & ^ CD O O co k g CD CD * — CD ft O -fl .g O-H co 00 7? O cd O ft 5P fl ftfi o o w • ZJ F L CO CD .2 T3 T3 3 § T3 ft 2 § Cj 60 CD . f l —• 60 cd cd o O fa ° CD cd co fl T3 CD fl 60 a o fl S -a '55 <3 >; 3 ^ fl cd -< CD O 247 CU CH >> H U SO cu o fc. c S 8 . « JS CN CQ = 5 o .o +3 ' - 3 O 03 3 2 « •» S 3 oo O O S u Cl Cl o .o 3 ' - 3 O 03 § 2 « 5 o o 3 'ti O 03 § 2 u 2 o 2 3 'Cj O 03 § 2 u 2 e & '"is eu 3 ' ' ft 3 U NH 6J3 — cs cs ^ e <u iw 5 d o S « •2 a 1 o - l *E £ oo O CU o IQ U £ «2 S 3 O lo cj ^ <u iS £ O cj 5 oo g CJ JJ o 5 0) o > fi 3 (+H , T3 O < " fi U 2 '+3 c o 3 2 a s . § ^ * -g £ £ " B O 3 C 3 '5 2 2 -3 - 5 3 \ 3 7 3 O 3 T i *-< O « u CJ i— i r X o3 ' S W X oo 03 m a t ! oo Cj T3 £ - 3 6 0 ^ 3 S O o 'co H 3HX> cu u M ft* « , <3 3 13 « CS -o I OH C o 1 1 * 3 CJ 50 03 fi 'S - 3 oo 03 «_> 0 M ° fi fi - 3 •2 I oo cj fi 3 £ o cj 248 CU a, >> H Vi u cu «? s-PH w C S 3 <U tf £ « 5 o o fl '+-> o g § 5 fl c o o fl '+J o cd 2 trj o 2 -fl l u i o , o *? 'a o cd 2 ts « S P .2 +-; o cd 2 c£ u 2 c ^ .2 <u s -fl w Vi O V cu eu 05 s S S3 cu t*5 -B S « .2 -tt 5 o- £ Vi B <U O Q U cd -f l 5 s cd sp < * CD fl - o cd co — -fl CO o rt O a co CD S <+-< o o o a cd ^ CO 00 co C CD H CD X <+4 o CD a o I-l CD co CD I 8 b OH 3 IES I fl I" H-» cd fl co fl CO O CD cf fl c*o .2 fl fl .2 CO CD O »- -c 22 tt -fl CD H tf 00 _ f l 'CD fl -fl o -fl fl-cd 2 fl 6 > 2 =3 * co —j CD Cd n 00 •nfl co 00 * -fl fl CD bo cd r_ CD co ^ o CO ,CD fl I CD CD 'G fl •3.1 B ta T3 CD S 2 -2 OH o o co H co fl •-H fl T3 CD CD 3 CO 2 o -H OH X ) c cd co CD fl "O co CD £ co -fl G cd ?= s i CD fd A cfr CD cd 1 2 §•§ o 2 05 t: CU CD fl & « 93 fl Q CD " 93 -a PH S H* 5 8 - H * * o co cd +-> fl O -T5 CD 3 -9 o CD co CD o cd CD -fl -fl fl fl O £ 5 CD H—» CD Cr-I o fl c3 O co 2 2 o "S U 3 co . f l 'co cd X ) 00 CD •S E ^ 2 CD fl CO O <+H CD O C+H fl 0 .2 "3 n H—» co CD 2 ° tt B « ? co -f l -f l fl co O fa cd U c S ^ 249 ii a >> H 12 a cs 00 ii |_- 8 £ s o o 3 O 03 2 LC fi -3 CJ S C N c 5 o .o 3 '+-; O 03 3 ^ o o o 2 O c3 § 2 C N < fi g .2 .2 O 03 2 ie o 2 S s ° 2 O 03 2 m § 2 C N « 5 ° 2 ~ '-3 O 03 2 ipj 3 3^ § 2 s 2 'cs <u 3 , -« O cs o u 4» DJD CS a cs PM © fi o fc. • M  w o c .2 3 'E h oo 5 ii o Q U JS ii 03 I fi CO T 303 CJ • 3 o O 43 "53 <g 2 CL) ^ c l a) . ^ H CJ o " 3 N o C o B S _ £ a O CJ o 03 CJ CJ fi T3 •4—» 00 O 00 - d fi i3 CJ tH 2 CJ 13 00 fi +-CJ o OH CJ 8 a c2 | 2 « g CJ fi T3 00 03 CJ ol fi CJ OH "c3 fi 00 2 CJ H" »H 03 T 3 03 00 fi HH CJ o OH CJ 8 OH CJ . 3 .H H-> «3 2 a 2 3 X I « S fi OH o3 CJ _ -3 S <-! '-3 o 00 2j CJ 3 _2 'H-» O 2 H-> 00 3 O o £P.fi oo fi ~ fi CJ o 3 3 CJ so 2 S 3 o x " M S 3 q3 S fi — O T3 CJ OH Z, -3 a 3 ~ S H 2 O CJ CJ _ 3 "S £ 3 o S ' - 3 o 3'-3 00 OH O < a a S 3 g 3 u o g CJ ca o H 2 | 2 fi g OH u ^ c2 " 3 cj ^3 u X 3 oo _ -T3 d O -CJ 3 CJ 1 — 1 CJ 03 5 3 „ CJ 00 CJ J3 £ -3 « o T 3 "5 -S «* .2,41 « OH fj § CJ CJ i l - a l i « t; u u C S 3 ft « 6X1 fc. . <3 3 I-go cs *o PH a X cs fc. 1 ^ U oo 3 " o cj 'o" tH OH fi O 00 — .a to 'oo 3 2 CJ o oo I H—» CJ CJ 3 o 3 O "3 CJ tH CJ 3 O • rH 00 00 00 fi S .2 2 -g C J O "CJ bp u 'oo 3 CJ o Q PH o o 2 "> <+H o c cj a a 2 "8 CJ r v > ZT " 5 o CJ p bO 3 C cj ts CJ Q bO _3 3 fi 00 ^ .2 ° "S •3 3 fi 2 M a u tH bO 3 cj o tH 2 'js 3 oo 250 251 <u a >, H co 2 u cu o I* a cu CO cs "3 tf PH C N < a c o . o fi "-fl O 03 o 2 c s o o fi •+-> o co § £ u S s e « = cu 3 «s CO O u « H* u CJ PH i -o •6* ifi CU 5 « .2 -g '•c £ £ fl CU O Q U at t cu u , fl s £ .2 & « , « fl i y i—i IPH fl ha -t -S 2 S '£ O Q O CO CO B O O X> o fl o J3 +-» CO fl 00 fl ' § Ji "o -fl 252 253 H co 5 CD cu o u B C3 CU i i . £ u -fl tf PH o .o •fl •+-< O cd 2 £ § £ u 2 o o fl '+-> CD cd 2 is § a fl fl fl o .2 * o .2 "£J *-*-» 2 ,2 O C3 2 ,8 Const Modii Const: Modii CQ fl fl o 2 '•>-> CD cd 2 ei c "3 o 2 fl © « h cu S CU ® s U _ _ CS 73 H* § PH , fl cu w HH> s « .2 -g •** fl a * « tS co A cu O Q U c o 1 h X > I-i on i<2 3, " f t 00 c CO CD c o c .2 o 2 -*-> CO C O 00 c C+H © c o <D • — CO CO fl cd CD l - i xj 3 CO cd CO CO CD fl -fl CD O c o fl Xj a § 5 fi CO CD cd S o fi T3 co 6 £ ° « 3 fl &0 fl g CD fi°g>2 I -ft CD fi fl T3 xi-f l ~ S c o fl fl co o co ^ ^ | . 2 O AH -fl fl .2 fl >? 'C fa CD cd £ -2 . ft " CD - f l £ -fl .2 § c O | cd fa # c o ft -fl o O c o fl tt ft 5 "cd - B d ^ § S CD r f l g c o > 2 > o ° ^ CO - H CO fl CD O CD O — i > fl •£> CD - f l fl fa CD CD ft I-i fl O O 2 H H fi J—i CD . . ft ° & CD fl CD B l - l 11 IH CD 00 I-I cd -fl CD xi o <L> -zi CD s * fl <D CD " ! t! 'C a I-I ao S3 cd 3 § cd fl CD C O CO -fl C .2 J> "cd td CD ^ O CJ) CD > 00 g CD -fl > - f l CD CD m -H CD > i CD I i 8 2 -A 2 cd ro c5 O i f l fl-^ tt CO 3 CO S 3 CD » ] - r i fl ? « B ™ fa fl ft - H fl 00 +-> ™ n ? C 3 U O T3 _fl " t t CD ^ cd CD O •2 fl 73 •c cu CU £ fl WD !-, «3 fl cs -a PH fl ° 1 cd s • f l cd O t H CD CD -H cd O ft B a *s s fl -g o S X 0 0 2 X J ' - ' CD P tf £ CD fl -fl j_i fl CD CO • r H fl O O 2 O H 2 .a W & CD ' X T3 T3 s I d > o s CD tf CD rfl td ^ B tt fl cd fi X cd cd CD 3. 254 cu oo 2 a S S 1 8 QJ J -C N 2 .2 3 -tj O cd 2 « o o 'S '-3 O Cd § 2 cs _< v S - fi oo o 5 U _ CJ s S fi cu W -fi ® 5 5 » .2 "g C h W fi so 5 cu o Q U bfi o T J CO > 2 o rf « S .2 ° cl 3 43 _2 cd u — 1 n i S ' d 2 o ci * "S .2 =3 J2 ts fix; i d cd w oo "53 co Cu O h fi fi J> oo O Cu s3 \3 O co - cd fi M -co CO 3 fi •S 43 co co c? bO j j a .s "S w) 3 8 •S 7 3 d • o o o CO oo CO fi fi2 .2 33 'oo O O ir! >-u CO o 43 00 t d g CO i j* t o cd .2 c CO 00 OH CO ~ oo -g 8 : 2 3 co "cd ^ o 0 'Lo g " P S 2 2 OH g S CO CO 00 > cd O CO oo Cl CO 00 O 2P « CO -> "S 8 g -3 co co S C/i fi t o O T3 X j £?.2 CO - 3 fi OH co co oo fc fi B 'cd g tH s o CO fi oo w CO . .1-1 J T 3 <+-< \ 3 CO M >, cd co O -3 i Cu O O £2 3 SrV 3 oo bD co cs "C cu u e ft « on S» cs a cu 1 — 1 CS -o P- s ^ CS bO 3 • i-H H-» 00 _cd X 4^  CO 3 O CO 3 O 'oo O tH CO o 00 255 cu Ii a 5 S cu 43 tf PQ e o Cd a o <D fl Ifl -H fl rtn 3 O S U CQ CD CM ca C o rH td >-. CD OH O CQ cd - H CJ OH cd c cd CD OH fl cs £* CU S ^ •2 93 93 • M CO o U CD M ) 93 u 93 PH H* u o fi cu ® 2 5 « .2 3 C C * 1 CU O Q U x OH CD X " f l CD cd fl r—H cd > CD O Vi CD 60 a X CD CD r H 'vi CD -fl CD cd B o g CD CD tt OH bp 3 X fd g a s s £ cd CD r D - f l - f l 2 2 >-»*2 co cd CD r H CD .2 .fl. Cd g It 8 ° I-l CO fi OH CO X CO O 8 I ° -tt 2 fi o -fl CD _ f i > o fi o CD CD C+H O fi o CD co CD '3 a co OH XH O co Oncf l X j CD r H CD CD CD 60 § § fl CD CD fl fi OH ^cfr-2 o x a '3 -tt 0 .O rD A -S CO a fi co CD § OT tt -^> - f l CD cd 2 S > 2 _ fl CD cd * ^ Ul rH — * ° o ^ ^ 1 E l 1 3 H - S ^ t fl § fi I 5* 8 1 » 3 § 3 & £ fi B fl O fi - fi A " —1 o -fl fl -A a O cd 2 fl .2 u cd fl —~ r , CO CD OH 60 *H y-H CD CD <+H 2 fl fl O CD C+H O CD OH CD CD -fl CD -fl CO fi O '+-» fl fi O o CD 1 •§ § g CD co -fl a + - » r H cd CO CD CD -+-> CD r H CD 60 CD CO CD td i cd tJ O CD o CD r H CO .2 'o CD OH co U - r -£ u o , r H C+H X CO X cd U cd a fl - A B 'a tf fi cd g O OH W fi O o tt o X C+H O r H 8 '3 CD CD fl - H • f l « cd X o +-> •fi C+H fl ° X o u *2 ^  |2 cd r> -H 60 - O O CD f i -^-2 OH 13 CD cd > CD C3 - ~ " tt •-cd o I — tt > -c3 cd (H 1 fl -s a CD cd CO 2 .^CD 28 2 .S | CD - - H CO Itf P -c 256 257 H o e CS <"> 4) CQ CQ o o CH„2 — o S o c i 3 . 2 .2 o cj cd rtl fi CO O a o w as ^ o S u cd a <s <u fi fc. Q oo s ° , 8 ft I—I 6X1 — CS CS fi cs « PH S *-. p © , fi 0> a •& S « .2 -g .S-fi C ~ u a CO fi <U O IQ U a >-i co CJ co cd H cj CJ 8-S c3 cd CJ cd _ 1 o -P co 3 > " ^ o x CJ ^ H + J fi rfi o CH CJ CO ,—1 2 'Cu fi . 2 , o c> a d , o CJ CJ < £ ex-5 CJ 2 9^ +^ CO § 2 cd -T 3 T 3 fi cj fi co 3 fi CJ cd S o 43 CJ "cd "o fi cd 2 ^ g a < -2 o CJ CJ CJ CS -*H fc. CJ .tS A £ .2 - t H 3 Q CS -o PH fi w fc. *-\> s cd co fi "-J CJ T 3 cd CJ .3 * O —1 3 ^  « . ' e g t ; -2 O U O 258 QJ ft. it Vi 2 u QJ o fi es <« • o» is 8 Cv) C o es £« v fi fc. Q es fi es Vi O U be es u es PH _4 fc. © es HH> fi 4) s fi o fc. • RH > , a « 1 © o c c i Vi fi QJ O IO u QJ oo CJ bO 4 f i O CJ 0) 3 CO O IO CL) CO a Q fi " ccj cd 1^  a a & w .a .a CL) 4 f i " f i CJ Closure CJ CO fi cd o cd "S CJ o t H c CJ <4H o bfl d • H T3 C cd H-> CO CO a> CJ is co fi fi fi O ve 'M o 13 t H t H OH -*-» fi o _ f  c2 To nd ef • cd _fi > -cd CJ fi o ri T3 CJ § 1 co cd fi CJ . 2 a 2 c =3 . 2 CJ H-J t n cd OH bfl +-> o cd fi O H , C * a <HH o CO ti CJ C fi CJ CJ > > O '-3 M ,CJ <H-H _ CJ es fc. v CJ fi M Ml fc. . es s 13 Q , « T3 PH fi l mt * fc. £ u CJ t H fi CO bO fi '§ „ 7 3 cj bO ^ fi o .-fi CJ 3 cd 2 H H 8 OH 259 a H cj a tf PH o » '« ss £-« 3 . *• O U -a *• a * « § u , s « NH 191) _ 93 93 H* § PH fl cu W xi S « .2 g Cu - 0 E £ « t! <*> B cu O Q U CU CJ fl U5 £ fl OA fi . 93 fl I-go PH fl H* 5 8 fi o S 1*5 OO CJ DO X CJ cu X cd >-< co CJ • f l C+H O bO fl 'S CJ CJ • f l •> . o > CH CJ O CJ H x j CJ -H fl CO O O fl CO O IO CJ CJ co CJ co 3 CJ CQ CJ -H fl CO O 260 Appendix B Transformation of Random Variables The transformation procedure that can be used to transform variables distributed uniformly on the interval [0,1 ] into any general probability distribution function and vice versa is described in this appendix. The procedure described follows from Cooke and Waij (1986). Let F(x) be an invertible cumulative distribution function. Therefore, by definition 0<F(x)<\ { B A ) The value of x may range from - oc to + oc, with additional constraints being placed by the specific functional form of F. Consider a random variable v, which is distributed uniformly on the interval [0,1]. It can be written that, ?rob(v<r) = r (B.2) Considering F(x); for every r strictly between 0 and 1, there exists a unique value q of x such that F(q) - r (B.3) 261 Therefore, (B.2) can be written as ?vob(v<F(q)) = F(q) Since F is invertible, i f v<F(q) Then for strictly increasing distributions F-\v)<q (5.5) (5.6) Therefore, (B.4) can be written ?rob(F-\v)<q) = F(q) (A7> which implies that F"'(v) is a variable distributed as F(x). Therefore, the transformation F''(v)> transforms v, which is distributed uniformly on the unit interval into a variable having the distribution function F. Alternatively, any random variable x, having a distribution function F can be transformed into a uniformly distributed variable on the interval [0,1] by the transformation F(x). 262 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0063994/manifest

Comment

Related Items