@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Forestry, Faculty of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Cavill, Jacqueline"@en ; dcterms:issued "2008-02-12T22:20:05Z"@en, "2008"@en ; vivo:relatedDegree "Master of Science - MSc"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description "Persistent conflicts between stakeholders and complex trade offs among forest values have created a difficult decision environment for sustainable forest management. Tools developed for decision support in land use planning are essential for managing these challenges. This research study is an interactive assessment of a land use planning Decision Support Tool (DST) in the Invermere Timber Supply Area (TSA), located in the East Kootenay area of British Columbia. The aim of this study is to explore whether stakeholders' initial stated preferences change and whether trade-offs are made between various forest values upon observation of a long-term forecast of these values using a DST. Representatives from various stakeholder groups in the area were assembled for individual sessions to interact with the multi-criteria DST. Participants were required to state their preferences for six forest values using a weighting scheme. The DST developed an output for each forest value based on the participants' preferences. Upon review of the DST output, the participant had the opportunity to alter their initial preferences iteratively until a desirable output was found. The results indicate that participants' preferences changed after reviewing the DST outputs and that participants are willing to make trade-offs between various forest values using a DST to find a desirable solution. However, the preference order of the forest values changed only slightly from the participants' initial to preferred scenarios; instead participants made drastic changes to the weighting of each value to find a desirable output. Participants also stated their willingness to use DSTs for land use planning decision-making, although underlying assumptions built into the model must be improved before stakeholders can trust the tool as an aid for decision-making. Studies such as this can further the development of DSTs to help find desirable decisions for sustainable resource management and to help create a productive and engaging process."@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/333?expand=metadata"@en ; dcterms:extent "34442577 bytes"@en ; dc:format "application/pdf"@en ; skos:note "APPLICATION OF A LAND USE PLANNING DECISION SUPPORT TOOL IN A PUBLIC PARTICIPATORY PROCESS FOR SUSTAINABLE FOREST MANAGEMENT by JACQUELINE IRENE CAVILL B.S.F., University of British Columbia, 2003 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Forestry) THE UNIVERSITY OF BRITISH COLUMBIA December 2007 © Jacqueline Irene Cavill, 2007 ABSTRACT Persistent conflicts between stakeholders and complex tradeoffs among forest values have created a difficult decision environment for sustainable forest management. Tools developed for decision support in land use planning are essential for managing these challenges. This research study is an interactive assessment of a land use planning Decision Support Tool (DST) in the Invermere Timber Supply Area (TSA), located in the East Kootenay area of British Columbia. The aim of this study is to explore whether stakeholders' initial stated preferences change and whether trade-offs are made between various forest values upon observation of a long-term forecast of these values using a DST. Representatives from various stakeholder groups in the area were assembled for individual sessions to interact with the multi-criteria DST. Participants were required to state their preferences for six forest values using a weighting scheme. The DST developed an output for each forest value based on the participants' preferences. Upon review of the DST output, the participant had the opportunity to alter their initial preferences iteratively until a desirable output was found. The results indicate that participants' preferences changed after reviewing the DST outputs and that participants are willing to make trade-offs between various forest values using a DST to find a desirable solution. However, the preference order of the forest values changed only slightly from the participants initial to preferred scenarios; instead participants made drastic changes to the weighting of each value to find a desirable output. Participants also stated their willingness to use DSTs for land use planning decision-making, although underlying assumptions built into the model must be improved before stakeholders can trust the tool as an aid for decision-making. Studies such as this can further the development of DSTs to help find desirable decisions for sustainable resource management and to help create a productive and engaging process. ii TABLE OF CONTENTS ABSTRACT^ TABLE OF CONTENTS^ iii LIST OF TABLES LIST OF FIGURES^ vi ACKNOWLEDGEMENTS viii DEDICATION^ ix 1.0 GENERAL INTRODUCTION^ 2.0 LITERATURE REVIEW 5 2.1 INTRODUCTION^ 5 2.2 PUBLIC PARTICIPATORY PROCESSES IN SUSTAINABLE FOREST MANAGEMENT^ 7 2.3 PUBLIC PREFERENCES 10 2.4 TRADE-OFF ANALYSIS^ 12 2.5 DECISION SUPPORT TOOLS 16 2.5.1 Background 16 2.5.2 Types of Decision Support Tools^ 17 2.5.3 Application of Multi-Criteria Decision Support Tools in Forest Planning^ 21 2.6 CONCLUSION^ 24 3.0 METHODS 26 3.1 INVERMERE TIMBER SUPPLY AREA^ 26 3.1.1 Dunbar/Templeton Landscape Unit 31 3.2 THE MULTI-CRITERIA DECISION SUPPORT TOOL FOR SUSTAINABLE FOREST MANAGEMENT^ 33 3.2.1 Alterations to the Multi-Criteria Decision Support Tool^ 35 3.2.2 The Forest Indicators in the Decision Support Tool 37 3.3 THE DECISION SUPPORT TOOL INTERFACE^ 38 3.4 RESEARCH DESIGN^ 41 3.5 QUESTIONNAIRE DESIGN 41 3.6 SAMPLING - PARTICIPANTS^ 43 3.7 SESSION METHODS^ 46 3.8 DATA ANALYSIS 48 4.0 RESULTS 51 4.1 GENERAL OBSERVATIONS FROM SESSIONS^ 51 4.2 INITIAL FOREST VALUE PREFERENCES AND OPINIONS OF DECISION SUPPORT TOOLS^ 52 4.3 THE OUTPUT SUMMARY FROM THE DECISION SUPPORT TOOL EXERCISE^ 58 4.4 PREFERENCE IMPACTS AND ASSESSMENT OF THE DECISION SUPPORT TOOL 67 4.5 LINKAGES AND RELATIONSHIPS BETWEEN ALL MODES OF DATA COLLECTION 74 5.0 DISCUSSION^ 78 5.1 FOREST VALUE PREFERENCES AND TRADE-OFFS ^ 78 5.1.1 Preference Changes^ 78 5.1.2 No Preference Changes 80 5.1.3 DST Preference Scenarios 82 5.2 ASSESSMENT OF THE DECISION SUPPORT TOOL^ 83 5.3 LIMITATIONS OF THIS STUDY^ 87 5.4 FUTURE RESEARCH^ 89 6.0 CONCLUSION 91 REFERENCES^ 94 APPENDICES 100 APPENDIX 1. RED- AND BLUE-LISTED SPECIES WITH THE POTENTIAL TO OCCUR IN THE INVERMERE TIMBER SUPPLY AREA (TSA)^ 100 APPENDIX 2. QUESTIONNAIRE 1 101 APPENDIX 3. QUESTIONNAIRE 2 108 APPENDIX 4. INTERPOLATION SCHEME FOR EACH INDICATOR^ 112 APPENDIX 5. PARTICIPANTS MAIN CONNECTION TO THE DUNBAR/TEMPLETON LU^ 118 APPENDIX 6. PARTICIPANTS ACTIVITIES IN THE DUNBAR/TEMPLETON LU^ 119 APPENDIX 7. TRADE-OFFS BETWEEN INDICATORS MADE BY EACH STAKEHOLDER GROUP.^ 120 APPENDIX 8. UBC RESEARCH ETHICS BOARD CERTIFICATE OF APPROVAL 120 iv LIST OF TABLES TABLE 1. FUNCTIONS AND LIMITATIONS OF COMMON PUBLIC PARTICIPATORY PROCESSES.^ 9 TABLE 2. A LIST OF THE UNGULATE, LARGE MAMMAL, AND SMALL FURBEARER SPECIES FOUND IN THE INVERMERE TSA.^ 31 TABLE 3. THE CRITERIA AND INDICATORS APPLIED IN THE PLANNING MODEL^ 35 TABLE 4. SUMMARY OF PARTICIPANT'S PRIORITIES TOWARDS RESOURCE VALUES IN THE INVERMERE TSA.^ 53 TABLE 5. PROBLEMS WITH DECISION SUPPORT TOOLS^ 54 TABLE 6. REQUIREMENTS TO FEEL COMFORTABLE WITH THE DST^ 55 TABLE 7. PARTICIPANTS RELATIONSHIP WITH THE STUDY AREA 56 TABLE 8. PARTICIPANTS CONCERNS FOR FOREST VALUES IN THE DUNBAR/TEMPLETON LU^ 57 TABLE 9. EDUCATION LEVEL OF PARTICIPANTS^ 58 TABLE 10. THE NUMBER OF SCENARIOS CONDUCTED BY EACH STAKEHOLDER GROUP^ 58 TABLE 11. THE PREFERRED SCENARIO CHOSEN BY EACH STAKEHOLDER GROUP ^ 59 TABLE 12. CHANGE IN PREFERENCES BY STAKEHOLDER GROUPS AFTER USING THE DST^ 68 TABLE 13. DEGREE OF PREFERENCE CHANGE AS A RESULT OF USING THE DST^ 68 TABLE 14. FOREST GROUP RESPONSES: INDICATORS CONTRIBUTING TO CHANGES IN PARTICIPANTS' INITIAL PREFERENCES.^ 70 TABLE 15. ASSESSMENT OF THE DST 71 TABLE 16. IMPROVEMENTS TO THE DECISION SUPPORT TOOL ACCORDING TO THE STAKEHOLDER GROUPS^ 73 LIST OF FIGURES FIGURE 1. INVERMERE TIMBER SUPPLY AREA^ 27 FIGURE 2. SUMMARY OF THE LAND BASE 29 FIGURE 3. THLB AREA BY AGE CLASS AND MAJOR SPECIES^ 30 FIGURE 4. THE LOCATION OF THE DUNBAR/TEMPLETON LU IN THE INVERMERE TSA ^ 32 FIGURE 5. THE DECISION SUPPORT TOOL INTERFACE^ 39 FIGURE 6. AN EXAMPLE OF THE INTERFACE OUTPUT AFTER FIVE COMPLETED SCENARIOS.^ 49 FIGURE 7A. PREFERENCE CHANGES BETWEEN THE FIRST AND PREFERRED SCENARIOS: PROFIT^ 59 FIGURE 7B. PREFERENCE CHANGES BETWEEN THE FIRST AND PREFERRED SCENARIOS: EMPLOYMENT^ 60 FIGURE 7C. PREFERENCE CHANGES BETWEEN THE FIRST AND PREFERRED SCENARIOS: RECREATION^ 60 FIGURE 7D. PREFERENCE CHANGES BETWEEN THE FIRST AND PREFERRED SCENARIOS: ECOSYSTEMS AT RISK^ 59 FIGURE 7E. PREFERENCE CHANGES BETWEEN THE FIRST AND PREFERRED SCENARIOS: VISUAL QUALITY^ 61 FIGURE 7F. PREFERENCE CHANGES BETWEEN THE FIRST AND PREFERRED SCENARIOS: DOMESTIC WATERSHED^ 61 FIGURE 8A. THE AVERAGE CHANGE IN POINTS FROM THE INITIAL SCENARIO TO THE PREFERRED SCENARIO FOR THE COLLECTIVE GROUP OF PARTICIPANTS^ 63 FIGURE 8B. THE AVERAGE CHANGE IN POINTS FROM THE INITIAL SCENARIO TO THE PREFERRED SCENARIO FOR THE INDUSTRY STAKEHOLDERS^ 64 FIGURE 8C. THE AVERAGE CHANGE IN POINTS FROM THE INITIAL SCENARIO TO THE PREFERRED SCENARIO FOR THE ENGO STAKEHOLDERS^ 64 FIGURE 8D. THE AVERAGE CHANGE IN POINTS FROM THE INITIAL SCENARIO TO THE PREFERRED SCENARIO FOR THE GOVERNMENT STAKEHOLDERS^ 65 FIGURE 8E. THE AVERAGE CHANGE IN POINTS FROM THE INITIAL SCENARIO TO THE PREFERRED SCENARIO FOR THE RECREATION STAKEHOLDERS^ 66 vi FIGURE 8F. THE AVERAGE CHANGE IN POINTS FROM THE INITIAL SCENARIO TO THE PREFERRED SCENARIO FOR THE PRIVATE PROPERTY STAKEHOLDERS^ 67 FIGURE 9. CORRELATION BETWEEN THE NUMBER OF SCENARIOS AND THE SATISFACTION LEVEL^ 77 vii ACKNOWLEDGEMENTS I would like to thank my graduate supervisor, Dr. Thomas Maness for his support, guidance, and good humour throughout my graduate years. I would also like to express my appreciation to my committee members, Dr. Robert Kozak and Dr. Michael Meitner, for their contribution and advice, and my external examiner Dr. Stephen Sheppard for taking an interest in this project. Thank you also to Andre Schuetz for helping with the technical aspects of this project and the participants who offered their valuable time and input to make this study possible. I am very grateful to my parents for their unwavering support throughout my academic journey. Finally, I would also like to acknowledge the other students and staff in my research group, especially Katie Maness for her encouragement and friendship over my graduate years. viii DEDICATION To my Nana, Irene Scoble ix 1.0 GENERAL INTRODUCTION Forestry companies in British Columbia are currently implementing forest certification and ecosystem based management. Forest certification implemented by an independent third party (e.g. Forest Stewardship Council 1996, Canadian Standards Association Z809 2002) provides a sustainable forest management guarantee on wood and paper products. The purpose of ecosystem-based management is to create healthy ecosystems and healthy human communities (Coast Information Team 2001). These emerging concepts require incorporating public participation in the development of forest management plans. Sustainable forest management requires balancing the ecological, economic, and social values over scale and time. However, this balancing act often produces conflicting management objectives. Stakeholders and the general public are routinely asked to state their preferences for various forest management alternatives; often causing conflict between interest groups. To reach resolution stakeholders may be required to make complex trade-offs between different forest values. Complex trade-offs between forest values are very difficult for stakeholders to evaluate, and research has shown that there are a variety of problems in obtaining valid information in these processes (Gregory 2002). Shindler (2000) found that stakeholders' preferences can change, or evolve, when they are able to see the planning outcome that results from their stated preferences. Trade- offs are often explored using simulation models to forecast potential outcomes of various forest management alternatives. Therefore, a land use planning model may be an effective aid for helping stakeholders evaluate trade-offs during the decision-making process. Scientists are using computer modeling to develop Decision Support Tools (DST) to assist with forest management decision-making (Kangas et at 2001; de Steiguer et al 2003, Mendoza and Sprouse 1989, Kangas 1 1992, Kuusipalo and Kangas 1994, Murray and von Gadow 1991). The development and use of these models for sustainable forest management has been popular in the scientific and forestry community over the last decade. However, the application of DSTs for decision-making in a public participatory environment has not been actively studied to date. This research project involves an interactive assessment of the land use planning process in the Invermere TSA, located in the East Kootenay area of BC. The Invermere TSA has a total population of 9,000 and includes the communities of Invermere, Edgewater, Canal Flats, and Windermere. The permanent population is augmented by 40,000 part-time residents arriving primarily from Alberta with the purpose of enjoying the many recreational opportunities that this area offers (Columbia Valley Tourism 2005). There are two major forest licensees that operate in the TSA, including Canadian Forest Products located in Radium Hot Springs and Tembec Forest Products located in Canal Flats. A unique dynamic exists in the TSA due to competition for natural resources between the forest industry, a major employer of the residents in the East Kootenays, and the tourism industry, a major economic generator for the communities in the TSA. As a result, land use planning and forest management can create conflicts of interest and a complex decision environment. A Multi-Criteria Decision-making (MCDM) model has been developed for this area and will be used as a DST in this study (Maness and Farrell 2004). The DST is based on desired outcomes for specific forest values and thresholds that determine the minimum acceptable outcome for each value. The purpose of this study is to explore the application of a multi-criteria DST in a public participatory process and to examine its affect on forest stakeholders' stated preferences. The objectives of this study are: 1) to determine how preferences change when stakeholders are directly involved in the planning process using a Decision Support Tool; 2 2) to explore whether stakeholders will make trade-offs between forest values using a Decision Support Tool; and 3) to explore the usability of the Multi-Criteria Decision Making model in a public participatory setting. Most literature focuses on eliciting stakeholders' preferences for various forest values, but the application of these preferences in forest management decision- making is minimal. Several studies using land use planning models to explore temporal scales have been conducted, however, none address the problem of whether stakeholders are willing to make trade-offs over time (Sheppard 2005). A widely accepted method could help to ensure that stakeholders can accurately measure trade-offs and forecast potential outcomes from trade-offs made. Quantitative trade-off analyses can be used to forecast outcomes, but requires site-specific data and model development (Antle et al. 2002). Multi-criteria DSTs are effective in the decision-making process because alternatives are provided and outcomes are forecasted. Although multi-criteria DSTs are effective tools for public involvement in forest resource planning, these tools have also faced much criticism. Further research is required to determine whether multi-criteria DSTs are useful tools for incorporating the public's preferences for forestry-related applications and whether these tools will help stakeholders evaluate trade-offs between forest values. This thesis is organized into six chapters, including this introduction chapter. Chapter Two provides the necessary background and a review of the literature for this area of study. The literature review provides background information and examples from other studies on four key areas contributing to decision-making in forest resources management, including: public participatory processes, stakeholder preferences, trade-off analysis, and DSTs. The methods used to conduct the study and collect the data are described in Chapter Three. The methods in this study were consistent with exploratory research and applied structured interviews with individual stakeholders. Chapter Four outlines the data 3 analyses conducted and provides the results from this study. This chapter begins with general observations, followed by data analyses related to the results gathered from the two questionnaires and the DST exercise. Chapter Five provides a discussion on the significance of the results, limitations of this study, and potential for future research opportunities. Chapter Six provides some concluding remarks. 4 2.0 LITERATURE REVIEW 2.1 Introduction The majority of land in British Columbia is publicly owned (95%). For this reason public input is an important addition to natural resource management decisions in the province. Public involvement in land use decisions has not been practiced fervidly until recent decades. The Canadian Council of Forest Ministers (1993) and the Montreal Process (1995) each initiated a separate set of criteria and indicators to measure the progress towards sustainable forest management. The Canadian Council of Forest Ministers developed the Canadian Criteria and Indicators Framework to establish directions for stewardship and to measure Canada's progress in sustainable forest management. The Montreal Process was formed in Switzerland in 1994 to create and implement a set of internationally accepted criteria and indicators for the sustainable management of temperate and boreal forests. These initiatives led to the incorporation of forest values in land use planning and elevated the importance of public and stakeholder involvement in the decision-making process. Improvements in public participation were required to ensure the effectiveness and efficiency of decision-making in land use planning. The Commission on Resources and Environment (CORE), developed in 1996, acted as a neutral body to direct regional land use planning and resource allocation in British Columbia. Once CORE had established broad plans for four regions of the Province, some Land and Resource Management Plans (LRMP) were implemented and presently they continue to be developed across the province. The goal of the LRMP process is to foster an inclusive, co-operative process for decision-making. These initiatives provided a platform for public participation and helped develop a formalized process to implement collaborative decision- making. 5 Forestry companies in British Columbia are currently implementing forest certification and ecosystem based management. These emerging concepts require incorporating public participation in the development of forest management plans. Stakeholders and the general public are routinely asked to state their preferences for various types of competing forest values; often causing conflict between interest groups. To reach resolution, stakeholders may be required to make complex trade-offs between different forest values. Trade- offs are often explored using simulation models to forecast potential outcomes of various forest management alternatives. Application of various decision support tools are currently a popular area for research examining the possibilities of various tools as an effective aid for helping stakeholders make trade-offs during the decision-making process. This review provides background on four key areas contributing to decision- making in forest resource management: public participatory processes, stakeholder preferences, trade-off analyses, and decision support tools. Section 2.2 describes forms of public participation used in decision-making and explores the benefits and current limitations of involving the public. Section 2.3 provides details on the issues associated with establishing and eliciting public preferences. Section 2.4 explores trade-off analyses and difficulties with certain trade-offs. This section illustrates different methods for conducting trade-off analysis with the public, such as Contingent Valuation, structured decision- making, and other weighting techniques. Section 2.5 describes different Decision Support Tool (DST) methods used in forest resource planning. This section also outlines the benefits and limitations of multi-criteria Decision Support Tools in forest management decisions and explores applications used in forestry. Concluding remarks are provided in Section 2.6. 6 2.2 Public Participatory Processes in Sustainable Forest Management The emergence of forest certification, environmental awareness, and importance of non-timber forest products has increased the need for public participation. As a result, public participation has become an important step in forest management decision-making. The public can be involved in decision-making in many ways and various types of processes can be used. However, this wide range of applications and techniques is not conducive to an efficient and well-structured public participatory process; thus, the lack of direction and boundaries cause issues to arise. Public participatory processes can assume many functions in forest decision- making. For example, public involvement allows for the consideration of a wide range of values across economic, ecological, and social spectrums and allows for the incorporation of unique local knowledge. Furthermore, public participation permits decision-makers to choose a socially acceptable management direction and to set boundaries on management practices and policies. Public participation has also resulted in developing an environment for mutual learning and resolving conflicts between stakeholders (Duinker 1998). The introduction of certification schemes, criteria and indicators, and ecosystem-based management have all contributed to involving the public and stakeholders in land use decisions. To date, the public has been involved in various decision-making processes related to policy frameworks, certification schemes, legislation and regulations, land use strategies and allocation, environmental assessments, forest management plans, and community forests (Duinker 1998). According to Konisky and Beierle (2001), different types of processes involve different participants, intended outcomes, and decision-making authorities. The structure of the process ranges from public access in open house forums to participant selection in which stakeholders must meet certain criteria to participate. The intended outcome of a participatory process may simply include 7 open communication to foster mutual education or participants may be involved in more active roles of providing recommendations or final solutions. Furthermore, the decision-making authority varies between types of processes ranging from no authority, to an advisory position, to shared decision-making power. Thus, depending on the degree of involvement and type of outcome required, there are many different ways of structuring public processes. Table 1 lists the functions and limitations of common public involvement processes. The lack of direction and regulations surrounding public participation results in some key failings and limitations in the process. For example, temporal and budgetary constraints, uncertainty, and loss of control are all significant disadvantages related to involving the public (Blouin 1998). Firstly, time and budget constraints limit the ability of participants to engage effectively and equitably during the process (Hamersley Chambers, and Beckley 2003) and, due to arising conflicts and insufficient facilitation, consensus among stakeholders is time-consuming or is never achieved (Gregory 2002). Secondly, participants may express uncertainty in the process and in the decision-making authority. The public is cognizant of insufficient and biased information and inadequate comparisons of management alternatives, and thus, distrust is built in to the process. Thirdly, widespread dissatisfaction exists with respect to the equity of stakeholders input in land use decisions due to a lack of transparency between decision-making and implementation of land use plans (Gregory 2000). Also, decision-makers often do not conduct the public participatory process and may not even be involved in the discussion; consequently, there is a lack of control to direct participants to produce an operable, measurable, and effective management plan. Public process outcomes are usually broad and informal for forest decision-makers to translate to detailed, spatially explicit decisions (Martin et al. 2000). Thus, experts often dominate the process to fill in the information gaps of public views and ensure that the local information is operational (Kakoyannis et al. 2001). 8 Table 1. Functions and Limitations of Common Public Participatory Processes. Source: Environmental Protection Division: Public Consultation Guide (1991 Public Process Function Limitations Public meetings - provide information to large numbers of people and allow for feedback on the issue - not suitable for consensus or discussion among participants Open house - an informal method for disseminating information at the public's leisure - must be well-advertised to ensure satisfactory attendance Workshops - a structured forum where individuals or groups are invited to discuss a common issue and build consensus -usually conducted by a facilitator and involves a small number of participants - can result in confrontations Public committees - obtain insight into different stakeholders interests concerning specific proposals - build consensus between differing views - must include stakeholders representing broad interests, but consist of members willing to work towards agreement Public discussion paper - there are two types: position paper examines a proposed policy; and options paper examines the alternatives Toll free telephone line - allow the public to easily provide feedback or ask questions individually - no opportunity for discussion amongst a group of participants Targeted briefing - closed sessions that occur when the decision-making authority presents information to a specific group - important stakeholders may not be included in the discussion Public seminars - formal events designed to promote the exchange of information on broad issues - no specific targeted audience; some people may dominate the process for their own agenda Site visits - provides the opportunity to visualize management action or issues on the ground Focus groups - used to monitor the public's potential response to a proposed plan by selecting participants to meet and discuss the proposal - not used for decision- making or consensus 9 However, research shows that participants agree on the key principles of a successful public participatory process. Tuler and Webler (1999) used grounded theory and case study interviews to inductively identify principles of \"good\" processes from participants. The participants identified the following key principles for a successful public participatory process: access to the process and to information; power to influence process and outcomes; constructive interactions are promoted; adequate analysis is conducted; and future processes are possible. McCool and Guthrie (2001) conducted a similar study after interviewing forty two participants regarding the characteristics of a successful participatory process. Their results reflected similar dimensions of success, including: writing and implementing a plan; fair representation of participants interests; relationship building; impressing accountability upon participants; and gaining social and political acceptability of decisions made during the process. Public participation can contribute to the overall planning process in forest management. According to Blouin (1998), involving the public in decision- making increases support and credibility of the process, potentially reduces conflict, and educates the public. Increasing the support and credibility of decisions ensures that the public is aware of the rationale behind decision- making and are able to provide their own opinions and local knowledge. Participation provides a forum for learning and disseminating information for all parties involved in the process. Furthermore, these processes can reduce conflict between stakeholders through inclusive decision-making. This type of decision-making process can lead to solid decisions through early communication of differing opinions and interests. 2.3 Public Preferences Conflicts among stakeholder groups arise due to differing preferences for various forest values. According to Martin et al. (2000), the fundamental basis of conflict among stakeholders is differing preferences in three areas: 1) allocation of land 1 0 between commodity and non-commodity; 2) allocation of land between motorized and non-motorized uses; and 3) importance of multiple use management and ecosystem management. More detailed knowledge of stakeholder preferences may help to resolve conflicts more effectively and efficiently. The study conducted by Martin et al. (2000) investigated a participatory process that incorporated the public's preferences and interests into the decision-making environment. Three different stakeholder groups were asked to rank a set of management alternatives in order to elicit preferences from each participant. The authors found that several stakeholders chose different orders for ranking the alternative management plans. Improved understanding of stakeholders' differing interests provides the opportunity for bridging gaps between these interests. This study indicates that information gathered using this method can be applied to help minimize conflict during land use planning processes. Research incorporating stakeholder values explicitly into the decision-making process is lacking. Most literature focuses on eliciting stakeholders' preferences for various forest values, but the application of these preferences in forest management decision-making is minimal. Ananda and Herath (2003) is an exception wherein the effective incorporation of value preferences into decision making processes is investigated. Multi-Attribute Value Theory (MAVT) was applied to elicit and analyze stakeholder values in regional forest planning in Northeast Victoria, Australia from five stakeholder groups (timber, environment, farmers, recreation, and tour guides). This study elicited preferences through face-to-face detailed surveys. Single attribute value functions were developed for timber production, recreation, and old growth conservation by analyzing the survey data. The multi-attribute value functions were developed by aggregating the single attribute value functions; these functions were used to assess the forest management alternatives. At this point, the respondents applied their preferences to rank the three options. The researchers' quantified key trade-offs in the area from information collected with the MAVT approach, thus, providing an effective method for eliciting public values and evaluating management 11 options. Preferences are derived from people's interests in the land base and vary amongst stakeholders, causing conflict between competing interests. As demonstrated in the study conducted by Ananda and Herath (2003), it is possible to reach decisions through trade-offs derived from stakeholders' preferences. 2.4 Trade-off Analysis Limits on resources and people's differing views and beliefs necessitates trade- offs between forest values. A decision-maker can make a trade-off between two or more forest values by choosing to increase one value in exchange for a decrease in another desirable value. Trade-off analysis is defined as a method of eliciting preferences from stakeholders on alternative management actions to guide the decision-making process through socially acceptable thresholds (Brown et al. 2001, Sheppard et al. 2003). Operationally, decision-makers are provided with information on interrelationships between indicators in the system studied; decision-makers then use their own subjective valuations to decide how to balance or trade-off various competing outcomes (Antle et al. 2002). Trade-off analysis requires a set of criteria. Indicators or weights are used to operationalize the criteria. Indicators make criteria operational by providing a measure for sustainability by using thresholds and targets. Thresholds represent a minimum constraint that cannot be violated, the target is the desired indicator level; the trade-off occurs between the threshold and target of each indicator (Maness 2007). Weights can also make criteria operational. This method often requires the use of a decision model. The model generates information on alternative outcomes across the set of criteria. Stakeholders assign weights to these alternatives according to their priorities (Brown et al. 2001). Difficult trade-offs occur during the development of forest management plans. According to Gregory (2002), value trade-offs in environmental decision-making are difficult for community stakeholders for the following reasons: 12 • multiple value measures; • uncertainty of impacts; • unfamiliar evaluation; • balancing effort and accuracy; • incorporating feelings; and • learning over time; It is difficult for stakeholders to compare different forest values, particularly when the measurement for each value is different (e.g. dollars, hectares, population). Stakeholders also have a difficult time assessing the potential impacts and outcomes for making a particular trade-off because accurate forecasts are challenging and experts tend to describe outcomes too broadly. Furthermore, participants usually have little experience with land use planning decisions, thus, the process is unfamiliar, daunting, and time consuming. Emotions such as anger from past decisions or frustration during the process can be difficult to incorporate in the process, but require consideration. Lastly, adaptive learning is important to the process because stakeholders may change their beliefs or opinions during the process (Gregory 2002). Involving the public in trade-off analysis is important because it incorporates local knowledge and diverse priorities, and increases stakeholders' confidence in the planning process (Antle et al. 2002). Various researchers have studied contingent valuation, structured decision-making, and various weighting techniques by conducting trade-off analysis with the public and stakeholders (Niemi and Whitelaw 1999, Carson 2000, Gregory 2000, Hammond et al. 1999, and Keeney and McDaniels 1999, Sheppard et al. 2003). However, few of these studies have been applied in practice for sustainable forest management and many are regarded as controversial. Contingent valuation measures trade-offs by attempting to establish monetary worth of non-market forest values. If the right to use the resource is not possessed, then the value an individual places on a specific use is the amount 13 the individual is willing to pay; if the right is possessed, then the value is the amount the person is willing to accept for compensation of the use (Niemi and Whitelaw 1999, Carson 2000). Contingent valuation is a controversial method because many natural resources have a passive use; in other words, the consumer does not physically use the good to receive utility, and thus, subjective valuation is applied in these circumstances. Furthermore, using monetary values to measure natural resources is of concern due to technical and ethical considerations (Carson 2000). Individuals may lie about the amount that they are willing to pay or accept in order to benefit and protect the forest value. Individuals have the ability to lie because a real transaction is not occurring; thus, people are not accountable for their answer. A structured decision-making process allows stakeholders to work through trade- offs by balancing competing objectives to facilitate an informed choice (Gregory 2000, Hammond et al. 1999, and Keeney and McDaniels 1999). Gregory (2000) conducted a study in Tillamook Bay, Oregon with the goal of developing a scientific, community-based management plan for the Tillamook Bay watershed. A structured decision-making approach was applied using the following fundamental principles: framing the decision, defining key objectives, establishing alternatives, identifying consequences, and clarifying trade-offs. The Tillamook Bay decision was made, firstly, by requiring participants to form objectives in the context of this decision and not on their respective interests and positions. Participants were required to establish and assess preferred alternatives by creating an 'objectives by alternatives' matrix. This matrix allows easy tracking of the potential consequences (benefits and costs) of each alternative when compared to each objective. Alternatives that are not able to satisfy the set of objectives are eliminated. The pared matrix recognizes key trade-offs amongst competing alternatives demonstrating the costs and benefits between them, further helping participants to make an informed decision. Gregory (2000) found that a structured decision-making approach can lead to a broadly acceptable agreement. Structured decision-making differs from consensus solutions 14 because the techniques used in consensus shift away from participants' divergent views to achieve common ground, whereas the structured decision process directs stakeholders to evaluate trade-offs in an effort to balance competing objectives and interests in order to assist an informed choice (Gregory 2000). Using weightings to measure a set of criteria is another method for evaluating trade-offs. Sheppard et al. (2003) reviewed four procedures for determining stakeholder values including: choice experiments; approval rating; ranking and weighting; and contingent valuation methods. Sheppard et al. (2003) applied these methods in the Lemon Landscape Unit in the Slocan Valley of British Columbia. The purpose of this project was to test stakeholders' willingness to partake in trade-off games. Furthermore, the study evaluated the potential of these trade-off methods to inform the decision-maker regarding where and to what extent the stakeholders are willing to make trade-offs between criteria. The results demonstrated that the public is willing to participate in trade-off games with different levels of confidence. For example, participants found that trade- offs between social (e.g. visual quality) and economic criteria (e.g. timber supply) were made with increased confidence than games requiring choices between biological criteria (e.g. Mule Deer Winter Range) and timber supply. The study also found that these methods could be adapted in further studies to aid in SFM decision support. According to Shindler (2000), public acceptance of forest decision-making will increase when managers provide opportunities for the public to understand the rationale and potential outcomes of forest practices. Gregory (2002) believes that most public involvement initiatives do not sufficiently assist participants in evaluating interests, assessing impacts, and measuring trade-offs. Increased direction through facilitation and implementation of structured decision-making is important to explicitly deal with trade-offs. It is important that a widely accepted method is developed to ensure that stakeholders can accurately measure trade- 15 offs and forecast potential outcomes from trade-offs made. Trade-off analyses can be used to forecast outcomes by describing the behaviour and performance of a system across space and time, but requires site-specific data and model development (Antle et al. 2002). 2.5 Decision Support Tools 2.5.1 Background The public's interest in a broad spectrum of forest values requires natural resource managers to consider a wide range of criteria at different spatial and temporal scales. Spatial differences and time lags reduce insight into the outcomes of management actions, thus creating uncertainty in the planning process. Furthermore, humans are constrained by bias and systematic errors when structuring multiple use forest resource problems. Complex decision problems and scales, uncertainty, and human constraints all contribute to the desire for applying decision support tools in the decision-making process. DST are computerized systems that amalgamate complex databases with operational research models, graphical and tabular displays, and expert input to assist decision-making and to optimize between multiple objectives (Lexer and Brooks 2005, Varma et al 2000). DSTs are intended to only provide information regarding potential forecasts and outcomes; they are not intended to make decisions or provide solutions. The demand for DST is growing. These tools vary from general stand level growth and yield models, and landscape level wildlife habitat models to a combination of computerized models and multi-criteria decision-making (MCDM) techniques to simulate management scenarios (Lexer and Brooks 2005). DSTs are used for a broad range of purposes. For example, some methods analyze uncertainty and risk, whereas other tools manage conflict or account for poor/incomplete information. Some methods have been modified for application 16 in forest management planning (Kangas and Kangas 2005). A general classification of DST, suggested by Belton and Stewart (2002), includes: • value measurement models; • goal, aspiration, or reference level models; and • outranking models. Value measurement models use numerical scores to represent the degree to which one decision option may be preferred to another. An example of a value measurement model is the Analytic Hierarchy Process (AHP). Goal, aspiration, or reference level models establish desirable or satisfactory levels of achievement for each criterion. Goal Programming (GP) is an example of this type of model. Outranking models use alternative courses of action to make pair- wise comparisons. The most frequently used outranking models are ELECTRE and PROMETHEE. The most widely used multi-criteria methods include AHP, GP and MCDM. 2.5.2 Types of Decision Support Tools Analytic Hierarchy Process AHP, designed by Saaty (1980), is a mathematical approach for analyzing complex decision problems with multiple criteria. AHP involves three main steps. Firstly, the problem is structured into a hierarchical set of goals and criteria. Commonly, the hierarchy has an overarching goal with a number of alternatives which are compared to a set of criteria (Schmoldt et al. 2001). Secondly, the criteria are evaluated using pair-wise comparisons based on an appropriate measure with respect to the goal; the measure could be preference, importance, or likelihood. Thirdly, calculations are used to synthesize the pair-wise comparisons to produce a final value for each of the alternatives. The purpose of AHP is to clarify public preferences and evaluate alternative management plans related to public values. This method provides direction for areas of agreement, resulting in the potential for compromises between competing objectives, conflict resolution, and trade-off identification (Kangas 1994). Furthermore, AHP does 17 not require explicit units to describe value (Kuusipalo and Kangas 1994). However, there are problems with AHP. AHP does not allow for in-depth analyses of the comparisons, especially with regards to the uncertainty inherent in the data (Kangas and Kangas 2005). Further, increasingly complex problems involve a higher number of criteria and alternatives, substantially increasing the amount of comparisons made. An increase in the number of comparisons reduces comprehensibility and increases the cost and time of the process. The research applications of AHP are growing but practical applications are limited. Both quantitative and qualitative decision criteria can be analyzed with AHP and it has been applied to a broad range of decision issues. However, applications involved with forest resource planning are few (eg. Mendoza and Sprouse 1989, Kangas 1992, Kuusipalo and Kangas 1994, Murray and von Gadow 1991, among others). AHP has been used to elicit public preferences when choosing a management strategy for a forest area. Kuusipalo and Kangas (1994) applied AHP to account for biodiversity in strategic land use planning for the purpose of resource allocation and priority setting. A set of management strategies were evaluated using AHP to identify the strategy that best meets the requirement of maintaining biodiversity, while producing timber income. Kuusipalo and Kangas found that AHP is a flexible tool for this purpose and provides a suitable measure for land use planning when accounting for biodiversity. Ananda and Herath (2002) examined the effectiveness of AHP when stakeholder preferences are involved with regional forest planning in the context of the Australian Regional Forest Agreement Programme. The results from this study indicate that AHP has the potential to foster a formal public participation environment for decision-making and to improve the transparency and credibility of the process (Ananda and Herath 2002). 18 Goal Programming GP is a computer modeling method based on linear programming used for multi- criteria optimization. This method allows the decision-maker to either accept the compromised solution or revise the goal targets and conduct further iterations until an acceptable solution is reached. During the analysis phase, GP chooses the best solutions from those that graphically display the most realistic and attainable level to the estimated goal targets (Rustagi and Bare 1987). The following characteristics make GP an important tool: 1) the decision-maker is not required to explicitly define weights to state preferences among the objectives, instead making value judgments on the goal levels for various objectives; 2) results of each iteration is illustrated graphically; and 3) basic and non-basic solutions are explored to find the best choice (Rustagi and Bare 1987). GP allows decision-makers to easily understand conflicts and relationships between objectives due to transparency in the process, the graphical output, and the absence of a weighting framework (Rustagi and Bare 1987). However, GP is only capable of generating one solution at a time from a change in the goal targets. van Kooten (1995) used GP in a land use planning problem to examine the economic impacts of allocating public forest land on Vancouver Island, British Columbia in a stakeholder process. van Kooten analyzed the allocation of land for multiple purposes and evaluated the impacts on employment, government revenues, and achievement of Annual Allowable Cut requirements. The goal targets were generated by experts with the assumption of matching the public's views and beliefs. The goals were ranked via two public surveys. The results of this study demonstrated that there would be losses of direct jobs, a reduction in government revenue, and an annual decline in society's welfare under current land use practices. These results occurred in spite of using high values for non- timber uses such as recreation, non-use benefits, and tourist employment. 19 Multi-Criteria Decision Making According to Yu (1997) there are four important elements in decision-making, these include: 1) a set of feasible decision alternatives; 2) a set of criteria; 3) potential outcomes of each feasible alternative; and 4) decision-makers preferences concerning the potential outcomes. Comparing and evaluating complex alternatives, such as those found in forest land use planning, can benefit from the application of MCDM models. MCDM has been created for analysis of multi-criteria decision situations wherein evaluation and comparison of alternatives is complex and planning is affected by conflicting interests (Kangas and Kangas 2005). Multi-criteria DSTs allow for thorough evaluations of multiple criteria and indicators as well as a commensurable comparisons of different criteria. This helps decision-makers explore trade-offs between various forest values and account for uncertainty (de Steiguer et al. 2003). MCDM provides decision-makers with management options, objectives, and goals to define decision problems. This type of DST generally defines a set of alternatives; thus, decision-makers contribute through preferences by providing judgments with scores, criteria weights, and alternative estimates (Mendoza 1995). On the other hand, MCDM models are limited by the ability to forecast outcomes for multiple criteria. Forecasting accurate outcomes is one of the most difficult tasks involved in model development. Nelson (2003) provides three challenges for developing credible forecasts from MCDM models: 1) advanced data management systems are needed to support DSTs (e.g. high storage capability, rapid updates, infinite queries); 2) models must be verified through sensitivity analysis, but it is difficult to understand and replicate these models due to the large number of parameters; and 3) large scale, long term forecasting ability is larger than the credibility of the data. Further research is necessary to develop credible forecasts of alternative outcomes. MCDM models range from complex mathematical models using linear programming or spatial modeling to simple applications using Multi-Criteria 20 Analysis (MCA). MCA is a decision support method developed for complex problems involving trade-offs between multiple objectives. MCA accounts for both quantitative and qualitative data. According to Brown et al. (2001), when using the MCA process participants prioritize criteria with a weighting scheme. These weightings are translated to aggregate scores for each scenario; a matrix is created to assess the performance of each scenario. Several iterations may be required before an agreement is reached on the preferred scenario between stakeholders. 2.5.3 Application of Multi-Criteria Decision Support Tools in Forest Planning Multiple criteria decision support has often been used in forest management applications due to its capability of integrating many forest management elements in a structured and rational manner. Also, multiple use and the presence of multiple stakeholders with individual views and beliefs make multi- criteria DST useful in a public decision-making environment. Multi-criteria DST explicitly address multiple criteria, help structure the problem, focus the discussion, and provide processes that lead to rational, understandable decisions (Belton and Stewart 2002). Multi-criteria DSTs are effective in the decision- making process because alternatives are provided and outcomes are forecasted. Thus, transparency and consensus of the process are improved and uncertainty is decreased. The goal of multi-criteria decision-making tools is to identify possible conflicts, provide an interface to amalgamate value preferences, quantify the impact of alternatives through a defined set of criteria and indicators, and communicate potential outcomes to the public. Firstly, these tools have the potential to improve quality and transparency of decision-making due to the systematic process involved. Secondly, they can contribute to consensus by accommodating mutual understanding between stakeholders, soliciting input from stakeholders, and maintaining dialogue (Costanza and Ruth 1998). Thirdly, uncertainty is reduced when DST are involved because forecasts help decision- 21 makers understand the potential outcomes of certain alternatives. However, according to de Steiguer et al. (2003), it is not known if DST in participatory processes can improve public involvement, collaboration, and acceptance of plans. Although multi-criteria DSTs are effective tools for public involvement in forest resource planning, they have also faced many criticisms. Firstly, DSTs can be overly technical when used for public decision-making (Kangas et al. 2001, McCool and Stankey 2001, and Mendoza and Prabhu 2005). Quantitative DSTs can be too complex for non-specialists to implement or explain to the lay public in SFM planning (Sheppard 2005), thus decreasing credibility with some stakeholders (Kangas et al. 2001, de Steiguer et al. 2003). Secondly, biases may be built into the model by the developer, thus promoting distrust (Martin et al. 2000). Thirdly, models have been referred to as a \"black box\" 1 (Gregory 2002, de Steiguer et al. 2003). Lastly, the generation of original alternatives is no longer a priority and the process becomes interrupted with the use of the model. To ensure that the above criticisms do not impact decision-making, the following principles are important for developing efficient participatory decision support methods in SFM (Sheppard 2005): • broad representation of stakeholders; • open access to stakeholders; • clearly structured decision-making process; • engaging process; • understandable and accurate information; • appropriate scale and detail for participants and resource managers; • focus on assessing sustainability over time; • credibility of the process; ' The term \"black box\" has been used to describe models in which assumptions are only known to the programmer, therefore, when an output is generated the user has little information to validate the answer. 22 • mutual learning and capacity building; and • feasibility. Varma et al. (2000) developed a DST for sustainable forest management using GIS integrated with linear programming and data on decision rule uncertainty. This study is important because there has been minimal research into implementing sets of developed criteria and indicators. According to the authors, the two main goals of this study are to find ways to measure sustainable forest management with respect to spatial and temporal dimensions and to identify means for optimizing land use strategies. The results from the study show that a DST using criteria and indicators can facilitate the elicitation of participants' preferences in decision-making and considers trade-offs through computations made by the model. The authors conclude that this is an efficient method for decision-making; however, periodic revisions are required for continued improvement. There have been few successful models applied in public participation in SFM planning. One example conducted in the Arrow Forest District in British Columbia by Sheppard and Meitner (2005) uses MCA and visualization in public participation. The process involved 3D landscape visualizations to illustrate alternative scenarios and experts evaluated scenarios using weightings based on priorities set by stakeholder groups. The results of this study indicated that common preferences existed among even the most polarized groups. In other words, this DST method appeared to be effective in resolving conflicts as well as promoting an open, transparent, and inclusive process (Sheppard and Meitner 2005). Several of these studies explore temporal scales, however, research addressing the problem of whether stakeholders are willing to make trade-offs over time does not exist (Sheppard 2005). Implementation of multi-criteria DST in SFM planning is a large and complex task. An iterative, adaptive approach is required 23 to successfully develop and implement DST methods for public participation in forest decision-making. 2.6 Conclusion The public participation literature provides detailed knowledge of various methods for involving the public and eliciting stakeholders' preferences. However, there is little information on incorporating preferences directly into the decision-making process for forestry-related applications. As the literature has shown, multi-criteria decision-making techniques have the capability of incorporating preferences directly into the decision-making process. These techniques provide structure and direction for public participatory processes, and help involve public preferences during decision-making; therefore, it is possible to include preferences using these tools. Past research has focused on building multi-criteria DST to aid decision-making for various natural resource issues. However, the literature fails to provide direct applications of multi-criteria DST for forest management in public participation. In recent years, research on trade-off analyses has become increasingly important as non-timber values become better understood and conflicts arise between competing interests in the forest land base. The linkage between trade-off analyses and multi-criteria DSTs applied in practical \"real world\" environments requires further development to help stakeholders find common interests and resolve conflicts. Applying a multi-criteria DST in a public participatory setting warrants further research. Further research is required to determine whether DSTs are useful tools for incorporating public's preferences for forestry related applications and whether stakeholders will make trade-offs between forest values using these tools as an aid. It is anticipated that the results of this study, along with previous findings, will lead to the development of improved applications of multi-criteria decision-making methods in forest management public involvement. Furthermore, the application of DSTs may provide more detailed knowledge on 24 public preferences through trade-offs made and preferences elicited. DSTs have the potential to resolve conflicts more effectively and efficiently, while building awareness on stakeholders' different interests and the trade-offs stakeholders are willing to make between forest values. 25 3.0 METHODS This chapter provides the details on the research methods used in this study. Section 3.1 provides background information on the Invermere Timber Supply Area (TSA) study area and explains the reasons for limiting the study to the Dunbar/Templeton landscape unit (LU). Section 3.2 provides background on the development of the multi-criteria Decision Support Tool (DST) that was applied in this study, as well as information on the alterations made to the tool for its successful application. Section 3.3 describes the DST's user interface developed specifically for this study and the method used to connect the interface to the model. The research design and questionnaire design are explained in Section 3.4 and 3.5, respectively. Section 3.6 describes the sampling methods and Section 3.7 describes the session methods with participants. Lastly, the data analysis methods are explored in Section 3.8. 3.1 Invermere Timber Supply Area The Invermere TSA is located in the East Kootenays of British Columbia within the Southern Interior Forest Region. The size of the TSA is 1.15 million hectares. The area is bound by the Cranbrook TSA to the south, the Golden TSA and Tree Farm License (TFL) 14 to the north, the Rocky Mountains and the Alberta border to the east, and the Purcell Mountains to the west (Figure 1). The Rocky Mountain Trench is a broad, flat valley with numerous rivers and wetlands found within the TSA between the Rocky and Purcell Mountain ranges. The Columbia River flows North through the trench forming the Columbia Wetlands, a complex and rich ecosystem. 26 Figure 1. Invermere Timber Supply Area Source: Invermere Timber Supply Area Timber Supply Review #3 Analysis Report v. 3.0, May 12, 2004 Communities located within the Invermere TSA have a total population of 9,000. The major population centres are Invermere, Windermere, Canal Flats, and Edgewater; the smaller communities include Radium Hot Springs, Wilmer, Fairmont Hot Springs, Brisco, and Parsons. The permanent population is augmented by 40,000 part-time residents, arriving primarily from Alberta with the purpose of enjoying the many recreational opportunities offered in the area. Panorama Mountain ski hill (located 18 kilometres west of Invermere), the Hot Springs, protected parks, and many golf and resort destinations attract visitors and recreationists year round. Existing populations of approximately 400 First Nations People reside within the boundaries of the Invermere TSA. Archeological evidence shows that the Ktunaxa people have inhabited the area for 10,000 years. Two First Nations communities are located within the TSA including the Columbia Lake Band in Windermere and the Shuswap Band in Invermere. The Shuswap Band is culturally and linguistically connected to the Shuswap Nation and politically connected with the Ktunaxa Kinbasket Tribal Council. The Ktunaxa Kinbasket Tribal Council has submitted a detailed land 27 claim covering the Southeast of the Province including the Invermere TSA. However, a settlement has not been finalized. According to the results of the 2001 Census, tourism is the largest employment sector in the TSA (33.9%), followed by the public sector (21.3%). However, the tourism industry's low wages account for a low basic sector employment income (15.7%), far below the forest industry, public sector, and those relying on pension and investment income (Brown 2004). The TSA has one of the highest diversity indices in the province; in other words, the area's economy is not reliant on only one or two sectors to maintain the quality of life. Two major forest licensees operate in the TSA, including Canadian Forest Products (Canfor) located in Radium Hot Springs and Tembec Industries located in Canal Flats. The current Annual Allowable Cut (AAC) for the Invermere TSA, effective November 1, 2005, is 598,570 m 3 . Tembec has approximately half of the rights to the AAC and ownership of half of the wood processing capacity found in the TSA. Pulp is the main product from Tembec's mill in Skookumchuk with an estimated annual output capacity of 248,000 metric tonnes, while dimension lumber is the main product from the mill in Canal Flats with an estimated annual capacity of 166 million board feet. Canfor's main product is dimension lumber with an estimated annual output capacity of 185 million board feet (Brown 2004). Canfor and Tembec have each received forest certification from different certifying agencies. Canfor has been approved by both the Sustainable Forest Initiative (SFI) and the Canadian Standards Association (CSA); Tembec has received certification from the Forest Stewardship Council (FSC). The land base classification of the Invermere TSA is summarized in Figure 2. The area is closely divided between the Crown Forested Land Base (CFLB) and the non-forested area with only a small fraction dedicated to non-TSA. The chart on the right illustrates the CFLB broken down into sections, the larger sections 28 1,200.033 Non-TSA rrEr°-, I,^0 3- 800,.00 600,000 400,000 200,00 Crown Forested Land Base 48% 600,000 600,000 ME Parks 14%) 40'0,000^ Irparable (34%) 200,000 - 100,000 - Triter FLarvesting Land Base (42%) Unstable, ESA Non Merch, Low S(tes FFTs, WTPa, Rpar:an (10%) including the THLB (42%), inoperable area (34%), and Parks (14%), and the remaining 10% consisting of unstable, non-merchantable, low sites, (PFT), Wildlife Tree Patches (WTP), and riparian areas (Brown 2004). ',re^e TSA Crown Forested Land Base Figure 2. Summary of the Land Base Source: Invermere Timber Supply Area Timber Supply Review #3 Analysis Report v. 3.0, May 12, 2004 The forests are dominated by lodgepole pine (Pinus contorta) (40.7%), Douglas- fir (Pseudotsuga menziesii) (28.7%), Engelmann spruce (Picea engelmannii) (13%), larch (7.2%), and sub-alpine fir (Abies lasiocarpa) (4.3%). There are approximately 60,000 hectares of lodgepole pine over sixty years old on the Total Harvestable Land Base (THLB) in the TSA (Figure 3); areas outside of the ESSF are susceptible to Mountain Pine Beetle infestation. According to Figure 3, half of the THLB is currently older than 80 years. 29 <=20 21-40^41-6C^61-80^131-130 101-120 121-143^41-2E0 K C -3 - 'es 20.C:3— - 15.COD — - 10.000 — - — - Age Class Lodgepole Pine m SpnicelBalsarilOther. ^rs RI-Larch/Yellow Pine Figure 3. THLB area by age class and major species Source: Invermere Timber Supply Area Timber Supply Review #3 Analysis Report v. 3.0, May 12, 2004 One national park, Kootenay National Park, and eleven provincial parks, Mount Assiniboine Park, Height of the Rockies Wilderness Area, Purcell Wilderness Conservancy Area, Bugaboo Alpine Park, Top of the World Park, Windermere Lake, Whiteswan Lake, Premier Lake, Canal Flats, James Cabot, and Dry Gultch, are located in or directly adjacent to the TSA boundary. The six biogeoclimatic zones existing in the TSA are Ponderosa Pine (PP), Interior Douglas-fir (IDF), Montane Spruce (MS), Interior Cedar-Hemlock (ICH), Engelmann Spruce-Sub-alpine Fir (ESSF), and Alpine Tundra (AT). Abundant and diverse populations of ungulates and large predators thrive in the area due to the variety of available habitat types (Table 2). The Columbia Wetlands support 70% of the bird species known to exist in BC due to important habitat for nesting and migration (Brown 2004). There are eight red-listed (endangered and threatened) species and eighteen blue listed (species of concern) species found within the TSA (Appendix 1). The non-timber issues most significantly influencing forest management in the Invermere TSA include: biodiversity, riparian habitat, domestic and community watershed, fire maintained 30 ecosystems, ungulate winter range, grizzly bear, caribou, Identified Wildlife, viewscapes in scenic corridors, and forest recreation (Brown 2004). Table 2. A list of the ungulate, large mammal, and small furbearer species found in the Invermere TSA. Ungulate species Large mammal species Small furbearer species - Elk - mountain lions - beaver^- badger - Mule deer - wolves - mink - wolverine - Whitetail deer - black bear - muskrat^- bobcat - Moose - grizzly bear - otter - lynx - Rocky mountain - fisher^- squirrel bighorn sheep - marten - fox - Mountain goat - skunk^- raccoon - caribou - weasel The Invermere TSA has 34 landscape units (LU). Usually, a LU is managed by one or more of the companies operating in the area. The LU's range in size between 7,645.8 ha and 84,826.3 ha. The LU's differ in terms of operability, accessibility, ecology, recreational opportunity, visual quality, watersheds, and socially contentious issues. The Dunbar/Templeton LU was chosen for this study because the area represents relevant issues related to sustainable land use planning, information is available, and contention exists between high use recreational areas and timber extraction, particularly with respect to Mountain Pine Beetle salvage blocks. 3.1.1 Dunbar/Templeton Landscape Unit The area of the Dunbar/Templeton LU is 25,192 ha. It is located in the northwest corner of the Invermere TSA. The Dunbar/Templeton LU is bounded by Bugaboo Creek and Driftwood Creek from the north, the height of land between Dunbar Creek and Frances Creek from the south, Columbia River from the east, and the headwaters of the Dunbar and Templeton Creeks to the west (Figure 4). Canfor is the head licensee in the Dunbar/Templeton LU; however, British 31 PARK tAKE CANFOR / TEMBEC BCTS Columbia Timber Sales (BCTS) manages a portion of the landscape along the Northern border of the LU. Canfor shares the land with one woodlot located in the northeast corner of the LU, many private land holdings in the eastern area of the LU, as well as a range tenure holder. Figure 4. The location of the Dunbar/Templeton LU in the Invermere TSA The three biogeoclimatic zones in the Dunbar/Templeton LU are the MSdk (dry cool), ESSFdk, and in the lower elevations near the Columbia River, IDFdm2. According to Canfor's Forest Development Plan (FDP), the forest health issues in the area include Mountain Pine Beetle, Douglas-fir Bark Beetle, Spruce Bark Beetle, Armillaria root disease, and lodgepole pine dwarf mistletoe. A domestic watershed exists in the area with two domestic water intakes along the Templeton River and one intake at Ramer Spring. According to Canfor's FDP, 37 blocks are assigned in the LU; 13 of these blocks are approved by Cutting Permits, and an additional 24 blocks are proposed. The Dunbar/Templeton area 32 is managed for ungulate and grizzly bear habitat; caribou habitat does not exist in the LU. Abundant recreational opportunities exist including camping, fishing, boating, hunting, hiking, and sightseeing. Many lakes have road access and camping facilities at government operated recreational sites. Recreational cottages are found along the shoreline of Dunbar, Lang, Botts, and Twin Lakes. Most of the LU is not within the visible highway corridor except for a small portion facing the Columbia River, which requires visual management. Additionally, visual assessments are conducted from the recreational sites to ensure aesthetics are not impeded. 3.2 The Multi-Criteria Decision Support Tool for Sustainable Forest Management A multiple criteria decision support tool (DST) has been developed at the University of British Columbia to forecast SFM criteria and indicators over long- term time horizons for strategic land use planning (Maness and Ristea 2004). The model uses multiple objective linear programming methodology with fuzzy sets in a \"3T Approach\". The 3T Approach refers to targets, thresholds, and triggers; this method is applied in the model instead of a direct weighting system such as those used in a goal programming approach. Targets are the desired outcomes for each criterion, thresholds are the minimum acceptable outcome for each criterion, and triggers are on-the-ground management activities that change the achievement levels for each criterion. The model is based on a set of criteria and indicators embedded in a hierarchical framework. Overall, the model is solved using a hierarchical planning framework. A hierarchical planning technique was implemented using several connected optimizers, each working in different temporal realms. In this context, hierarchical planning differentiates the planning problem into temporal domains and creates a separate model for each domain. It is useful to divide the planning problem into strategic, tactical, and operational temporal realms because the objectives, as well as the required detail, are quite different in each realm. A 33 hierarchical method amalgamates existing models allowing each model to focus on the objectives that are important at a specific temporal domain. The individual models are executed iteratively. The model uses criteria and indicators based on information from regional GIS databases and potential outputs from the model. A detailed review of the CCFM criteria and indicators, as well as criteria and indicators from a local study in the West Kootenay region of BC, were assessed. Experts conducted workshops to determine which indicators would be appropriate to include in the model. To be approved for the model an indicator was required to meet the operational standards adapted from Bunnell (1997). A full description of the rationale for choosing indicators can be found in Maness (2003). The review concluded that four criteria and nine indicators could be applied in the planning model (Table 3). Targets and thresholds were developed for each indicator based on expert judgment 2 using the model output as a guideline. Each LU was divided into polygons, and data for each indicator was collected for each polygon. However, individual polygons were too small and too numerous to effectively execute the model if they were not amalgamated. Consequently, the individual polygons found in the TSA were aggregated into homogenous and continuous Decision Units (DUs) between 10 and 100 hectares in size. The planning horizon for the DST is 100 years, consisting of 10 periods of 10 years each. 2 A team of experts worked together during group meetings to develop a set of operational criteria and indicators that could be used to develop this model. Each indicator was proposed, discussed, and modified by relevant stakeholders and researchers. 34 Table 3. The Criteria and Indicators Applied in the Planning Model Source: Maness and Ristea (2004) Criterion Indicators Criterion I: Biological richness and its associated values are sustained within the management unit 1. Ecologically distinct ecosystem types are represented in the non-harvestable land base 2. Stand and forest-level habitat elements are represented Criterion II: Forest productivity is sustained 3. Annual removal of forest products relative to the volume determined to be sustainable Criterion III: The flow of economic benefits from the forest is sustained 4. Net profitability is sustained (proxy tax revenues) 5. Total employment in all forest sectors is sustained 6. The provincial government continues to receive a portion of benefits Criterion IV: Forest management supports ongoing opportunities for quality of life benefits 7. Availability of recreation opportunities are sustained 8. Visual quality of managed landscape is acceptable to stakeholders 9. Community watersheds are sustained and protected The development of this model is a significant advancement in DST research as it is the first to solve a complex hierarchical model for land use planning. For further details on the development of this model and subsequent studies conducted, see Maness and Ristea (2004). 3.2.1 Alterations to the Multi-Criteria Decision Support Tool The original model has been altered in order to successfully implement the system as a DST in this study. Firstly, the model analyzes data from a single LU instead of the entire Invermere TSA. Limiting the study to a smaller area ensured that the research was feasible within the time constraints and that participants could easily focus their decision-making. The profit indicator, however, was based on a proportion of the harvest volume available in the TSA because the area of the LU was not large enough to sustain a timber volume to 35 the mill over the specified time horizon. In other words, the timber volume decreases proportionally with the rest of the TSA when the model harvests polygons in the LU. Therefore, it was assumed that the TSA emulates the same attributes (size, species composition, age and diameter classes) as the Dunbar/Templeton LU across the other LU's in the TSA. Secondly, the manufacturing simulator was removed from the model hierarchy in order to improve the execution time of the model. The manufacturing simulator determines the Return to Log (RTL) values for each DU and optimizes the operations to determine the profit indicator. For this study, the profit indicator was solved by associating log diameter distributions with each polygon to determine the volume. A pre-assigned value in dollars ($) was attached to each diameter class. Thirdly, the model originally produced an output with stakeholders preferred targets for each indicator. However, for the purpose of this study the model included a weighting system that interpolated the triggers for each indicator. The scores stored in the model, created by experts, are difficult to understand and to describe to non-experts. The DST used the threshold to describe the \"0\" on the point scale, the minimum achievement for that indicator, while the target described the maximum achievement for that indicator, or \"60\" on the point scale. The trigger was the allocated points assigned to each individual indicator. Lastly, only six of the nine indicators in the model were implemented in this study. The following six criteria were applied: Ecosystems at Risk, Visual Quality, Recreation, Profitability, Employment, and Domestic Watershed. The values were limited to these six indicators in order to isolate the key conflicts in the study area and to promote participants to use critical thinking towards indicators. 36 3.2.2 The Forest Indicators in the Decision Support Tool The six indicators were carefully chosen in this study to ensure that environmental, social, and economic values were all covered. Two indicators were chosen for each value in order to reduce any bias towards a specific value. The Ecosystems at Risk and Domestic Watershed indicators meet some of the environmental values in the Dunbar/Templeton LU. The Visual Quality and Recreation indicators cover the social values and the Profitability and Employment indicators meet the economic values in this study. The Ecosystems at Risk indicator is based on the percentage of Old Growth Management Areas (OGMA) reserved in the Dunbar/Templeton LU. The Ecosystems at Risk indicator was previously referred to as Old Growth, however, the name was altered to account for overlap between defined biodiversity classifications and old growth. Furthermore, old growth is used to measure this indicator instead of the biodiversity classes because the model originally divided the three classes into separate indicators. Thus, there would have been a requirement to dedicate three biodiversity indicators which would bias the study towards this indicator. The Visual Quality indicator is based on the Visual Quality Objective (VQO) as provided by prescriptions in visual management areas. The Recreation indicator is based on the recreational significance, sensitivity to development, and proximity to water. The Profitability indicator is based on the diameter distributions amongst polygons. A value is assigned to each polygon according to species and diameter size. The model determines the value of the harvested volume after the model has chosen the specific polygons for harvesting. The volume entering the mill and log sales is measured in dollars ($). It is assumed in this project that the cost of raw logs is the same as the cost of milled logs. 37 The Employment indicator is based on the harvest volume and a pre-determined cut level in which the mill can no longer operate. The minimum harvestable volume is located at the point when the mill opens and all the forest values on the interface are set to 10 points each. The indicators were set to the balanced case because it was deemed necessary to find a point where all values were given equal weighting. The minimum harvest volume was found at 189,131 m 3 , which is the volume at which the mill remains open. If the volume was set any higher, the mill would close down at certain periods or close down entirely over the 100 years. It is assumed that harvesting, log transport, road construction and maintenance, and silviculture employment all occur regardless of whether the mill is in operation or not; although, once the harvested volume reaches the minimum harvest volume, the mill opens and timber processing employment is added to the total harvesting employment. When the Employment indicator is maximized the maximum harvestable area is 210,000 m 3 . It is important to distinguish that the Profitability and Employment indicator operate differently. For example, the mill could be operating inefficiently by cutting smaller volume timber, thus producing less profit, but continuing operations at the same or higher levels of employment. In this case, profit decreases and employment increases. The Domestic Watershed indicator is based on the number of hectares of Equivalent Clearcut Area (ECA) by watershed type; ECA is a calculated term that reflects the cumulative effect of harvesting within a watershed. An increase in the percent cover of stands greater than 6 m in height is considered to have a better ECA than a stand with an increase in stands less than 6 m in height. 3.3 The Decision Support Tool Interface The portal to the model is a user interface created uniquely for this study with the .NET framework. The interface consists of three different sections (Figure 5). 38 r^IRan 0:11 60 Ecosystems At Risk Ryer TreeVisual quality Over Tone^i^Recreation Over Tone ri n act^- .CetlatO ocereno. UMW') OCel0e00 ml =MVP: L acenam Rotitabdity Over Tme 10 4 2 1 0 °.^in lime (10 Year periods] 0 0■- 01(1m111.01, 0,010 Tone (10 year periods) ri or-rrenamcor..tocno Tme (10 year periods] ° Coi—r4M•11j1,0,[00,0 Trne (10 year periods)Tme (10year periods) 80 )13- 160 100 100 ,780 '60 e° 180 100 !4t) 020 &2o can 1 40 2 20 New Participant^I Participant, !lest Profitability Employment Multi Criteria Decision Making Tool Dunbar/Templeton LU Visual Quality Scenario Number: Told Desired Condition: Points Remaining: Recreation^Ecosystems At Risk Domestic Watershed Scenario 1. Scenario 2. Scenario 3 Employment Over Time 100 : 80 • 160 Ito Tone (10 year periods) Scenario t: Scan:402: Scenario a Somme, t Saxon:IS , Scenario r Scenario 7: Scenario a Scenario a' Scenario 10 Domestic Watershed Over Tene 100 80 160 1 4 20; S Stn,tiano );), S c ^16 4 Schr.Sra, Ssstwol Sorsre10 7 5otnsos Se-es-41017 Scenario 1^0 Scenario 2: Scenario Scamp Scenario 6: Scenario Scenario : •^• Scenano 77..7.. Scenario 1 Scenario 2 Scenario 3 Scenario a Scenario Scenario Data Ready Figure 5. The Decision Support Tool Interface The first section is located at the top of the window. This section displays the title of the model, as well as four functional fields and buttons. On the top left of the screen, there is a 'New Participant' button for use only by the DST's facilitator. This button, when clicked, produces a fresh screen and a unique ID number for each new participant. Below this button is the Participant ID field which indicates the ID of the current participant. This field allows the researcher to scroll through the final screenshot of each participant. The top right of the window includes the 'Scenario ID' field, the 'Total Desired Condition' calculator, the 'Points Remaining' calculator, and the Run #' field. The 'Scenario ID' field provides a unique ID to each scenario executed throughout the entire study. The 'Total Desired Condition' calculator refers to the points allocated by the participant for each scenario. The participant is required to allocate sixty points, hence, this field maintains a running total to help the 39 participant keep track. This field must sum to \"60\" before the model is executed. The 'Points Remaining' field sums the points the participant is required to allocate before the model is executed. This field must sum to \"0\" before the model is run. The 'Run #' field displays the number of scenarios executed by an individual. The participant applies the 'RUN' button to execute the model once the scenario is developed and all points have been allocated. The middle of the window is divided into six columns; each column is dedicated to an indicator. Ten rows of six empty fields (one field in each column) are labeled vertically 1 through 10. These fields are used for direct input of each participants individual point allocation according to their own preferences. All fields are initially filled with a zero before each scenario is entered. The purpose of entering a zero in these fields demonstrates that the fields are numerical, and ensures that participants begin with a clean, unbiased position for their scenario. Successive rows are built into the interface to help participants identify and distinguish changes made between scenarios in order to make an informed choice for subsequent scenarios. The bottom section of the screen provides a graph in each column summarizing the model's output for each indicator. The output is based on the participant's scenario as a whole and, therefore, the points issued for one indicator effect the other indicators. The graph illustrates the percent achieved of the total potential along the y-axis and the time horizon (ten year intervals over 100 years) along the x-axis. Each scenario, as executed by the model, is described with a continuous line on the graph in different colours for each scenario. The graph for the profitability indicator is shown differently. If shown as the others, the graph for the Profitability indicator would show large fluctuations over time because the forest company may harvest more timber in certain years than other years. Instead, the total profit discounted over the 100 years is calculated for the entire TSA by multiplying the number of hectares by the value of the cutblock per 40 hectare. This value is displayed in the graph window. The y-axis on the graph is displayed in $(millions) over time. 3.4 Research Design This project uses an exploratory research approach. Exploratory studies have the potential to further the understanding of a topic, and to develop methods that may be used in later studies (Babbie 2004). Exploratory research is conducted when the topic examined is relatively new and unfamiliar. The intent of the study is to gain a better understanding of stakeholders' preferences and the potential of the DST through specific observations, not on the basis of general principles; thus, a hypothesis was not tested. A cross-sectional research design was implemented for this study. A cross-sectional study was used because the objectives of the study did not require data collection to occur over a period of time. Furthermore, a cross-sectional design creates a realistic decision-making environment because many individual stakeholders participate in specific forest planning issues over relatively short timeframes. Individual face-to-face sessions were conducted with each participant. Each session ran approximately two hours and involved a pre-survey, DST exercise, and a post survey. The methods for data collection included: 1) a pre and post survey; 2) screenshots of each participant's final interface; and 3) reports generated in Microsoft Access from the database. After the pre-survey, the DST exercise was executed followed by the post-survey. 3.5 Questionnaire Design Two written questionnaires were provided to each participant during the individual face-to-face sessions. The participants were allocated time to complete the questionnaires on-site. The questionnaires provided structure and operationalized the key concepts in this study. The study employed a pre- and 41 post-questionnaire technique. Administering a questionnaire before and after the DST exercise helped to determine and explore any changes in participant's preferences as a result of the DST output. The pre-survey (Appendix 2) was administered preceding the DST exercise and the post-survey (Appendix 3) was administered following the DST exercise. The purpose of the pre-survey (Questionnaire 1) was to gather information on the participant's background, level of knowledge of the Dunbar/Templeton LU and its forest management, as well as preconceptions towards DSTs. Overall, the pre- survey consisted of four sections, with the majority of questions using a close- ended format. Some questions applied different five point scales to assess the participants' responses regarding their preferences for different forest values. The end points for the various questions ranged from not at all important to extremely important, not at all satisfied to extremely satisfied, and little/no knowledge to extremely knowledgeable. The other questions in this survey required participants to choose from an exhaustive list of multiple choices regarding their use and activities in the Dunbar/Templeton LU and their perceptions of DSTs. The purpose of the post-survey (Questionnaire 2) was to understand how value preferences changed as a result of using the multi-criteria DST, as well as an assessment of the model from the participants' perspective. Questionnaire 2 consisted of two sections and was much shorter in length because the majority of the questions were open-ended. Four questions in this survey were open-ended to allow the participant to provide detailed, in-depth responses regarding the use of the multi-criteria DST (Babbie 2004). This survey also included a matrix question applying the Likert scale (strongly agree to strongly disagree) to evaluate a series of statements regarding the participants assessment of the DST. 42 To ensure that participants answered all of the questions in the survey, the following layout procedures were used in the questionnaires: the demographic information was located at the end of the first questionnaire, the surveys were short, there were few open-ended questions, and the survey was visually appealing. It is important to include the demographic information near the end of the survey to ensure that the participant does not believe that this is another generic form. The questionnaires are short as not to overwhelm the participant, particularly because two surveys are involved in this study (Fink and Kosecoff 1998). Also, open-ended questions can be overwhelming and tedious for participants and if there are too many the participant may not be inclined to respond thoroughly; therefore, these questions are few in number (Babbie 2004). Furthermore, a cover letter was provided to each participant before the consultation session began. The letter introduced the study, described why and how the participants were selected, and that their identities would remain anonymous and confidential. Participants were made aware of their privacy rights from the initial recruitment letter, as well as the confirmation letter sent by e-mail. The ethical considerations in face-to-face interviews included the confidentiality of participant's individual responses and the anonymity of each participant (Babble 2004). The processes implemented to ensure confidentiality and anonymity included: employing only one researcher to complete the data collection and data entry phases; providing an ID number for each participant; instructing participants to leave their names off the questionnaires; and excluding videotaping from the sessions (Babbie 2004). 3.6 Sampling - Participants The study does not rely on statistical descriptions of large populations, nor is it a random sample. All residents of the Invermere TSA did not have an equal chance of selection in the sample because the study required the participation of 43 actual stakeholders interested in the Dunbar/Templeton LU. The selection of \"real\" stakeholders abets the simulation of actual practices. In this study, the population of interest includes various stakeholders in the Invermere area. Hence, a quota sampling method was employed insofar as certain stakeholder groups were represented on the basis of pre-specified characteristics (Babbie 2004). However, the quota sampling method requires that a total sample with the same distribution of characteristics existing in the study population is accrued. This requirement was not met in the study; as a result, the sample population could include some biases. To ensure that bias towards a specific stakeholder group was reduced and various statistics could be calculated, the sample size was made the same for each stakeholder sub-group. The DST was based on six different forest indicators; therefore, stakeholder sectors were chosen to participate in this study based on interests that coincided with the corresponding values in the study. The stakeholder groups chosen to participate included: the forest industry; Environmental Non-Government Organizations (ENGO's); the provincial government; recreation groups; and private property owners (including various licensees). Participants were residents or had a business based directly in the Invermere TSA, as well as a direct interest in the Dunbar/Templeton LU. This study is an exploratory study using a non-probability sampling technique. A purposive, stratified, snowball sampling method was employed for selecting participants. The researcher employed purposive sampling by selecting a sample based on the knowledge of the population, its fundamentals and the purpose of the study (Babbie, 2004). A purposive technique was important for this project because the researcher was able to use judgment to sample from the membership of stakeholder organizations that have specific interests in the LU, but was not required to sample from all stakeholder groups in the Invermere TSA. Furthermore, the purposive technique permits the study of a small sector 44 of a larger population in which many members of the sector are easily identified, but the inventory of them all is not feasible (Babbie 2004). Knowledge of the population was minimal; however, contact was made with the Radium Canfor Public Advisory Group (PAG). PAG members identified as representatives of one of the specified stakeholder sectors were chosen to initiate the snowball sampling method. Implementation of the snowball sampling technique began with participants providing further contacts in their respective interest groups; these contacts provided others, and so on. The snowball technique was employed for this study because the researcher was unfamiliar with this area and complete lists of various stakeholder groups with interests in the Dunbar/Templeton LU were not available. The purposive, snowball method used in this study allows for a realistic replication of public participatory processes because the sampling technique selects target groups (the stakeholders) instead of people from the general population with no interest in the Dunbar/Templeton LU. The exact number of participants involved in this study was not specified before the data collection was initiated, this was determined during collection of the data. The researcher finished once the data collected became repetitive and no additional information was found. The sample size appropriate for this study was twenty participants. Twenty participants were included to involve an equal number of representatives from the five identified stakeholder groups. Four representatives with similar interests in the Dunbar/Templeton LU were enough to ensure that additional information was not required. The responses seemed consistent and no new information would have been gained by increasing the individual sample sizes. 45 3.7 Session Methods The sessions were conducted at the Radium Hot Springs Resort. A specified meeting room was used to minimize distractions or interruptions during the course of the session. Each participant was provided a computer monitor, mouse, and keyboard. The participant was stationed in front of the computer screen during the DST exercise and the facilitator sat beside the participant during the course of the exercise to provide assistance and explanations of DST outputs. Upon arrival, the participant was required to produce a signed consent form and was verbally briefed on the proceedings of the session by the facilitator. At this point, the pre-survey was administered. Following completion of Questionnaire 1, the researcher explained and demonstrated the use of the DST to the participant. The participant was required to allocate sixty points across all six indicators according to his/her preferences; these were referred to as the participant's initial preferences. The participant began in Row 1, and once Row 1 was filled with the allotted points, the scenario was developed and the model was ready for execution. The participant allocated the sixty points with the assumption that the points allocated to a certain indicator were the desired condition that would be fixed over the 100 year time period. Every time the participant created a new scenario, there was a model execution time of approximately 15 -20 seconds. The interface stored all of the changes in the participant's decision-making made between scenarios. This helped to explore how planning models affect public stakeholders initial preferences. Participants only have sixty points to allocate among all six indicators; therefore, each indicator is weighted more heavily in the interpolation scheme for points between 0 to 15. The interpolation scheme refers to the translation of the sixty point scale to the potential achievement scores used in the model. Indicators were weighted more heavily for points in the low range because the addition or 46 removal of points from the participants' initial scenario creates a large difference in the lower point range. This occurs because once points are allocated across six indicators, participants are working with weighting in the lower point range. Therefore, point changes in the lower range are more meaningful than point changes to indicators with higher point allocations. The points used in the interface corresponded to a percentage ranking based on expert judgment, as well as response curves for each indicator from model executions. An interpolation scheme, detailed in Appendix 4, was used to calculate the achievement potentials for each interface point. The maximum potentials for each indicator translated into the \"60\" on the interface point scale. Thus, if an indicator was set at \"60\", it was maximized and all other indicators were set to \"0\". If an indicator was maximized it is the intent of the participant to protect 100% of that particular indicator or maximize profit or employment at the expense of the other indicators. Each indicator had a different interpolation scheme; however, the rationale behind each scheme was the same. For example, if the participant chose to dedicate five points to the Recreation indicator, then the interpolation scheme dictated that the participant wanted to achieve 20% of the model's achievement potential for Recreation. The interpolation between the points on the interface and the achievements in the model was necessary in order to create cohesion and interconnectedness between the interface and the model. Each row referred to a single scenario. A scenario refers to the results of a single model execution after sixty points have been allocated across the six forest indicators. Once the row had been filled, the participant selected the RUN button to execute the model. A 20 second delay occurred while the model prepared the output. The output was displayed in the line graphs for each indicator provided directly under the scenario rows on the interface (Figure 6). The facilitator described the output displayed in the graphs. The participant was then instructed to re-weight each indicator and re-run the model until he/she was 47 satisfied with the results. After viewing the results, the participant was asked the following: Upon review of the results, would you like to create a new scenario to get closer to the results you expected? The DST exercise was completed once the participants chose a preferred scenario from one of the outputs of the model. The participant indicated with a checkmark on the interface which scenario output he/she preferred. Following completion of the DST exercise, the post-survey was administered as the final step in the session. 3.8 Data Analysis The data collected from the screenshots of the DST's weighting interface and the answers to the questionnaires were used in the analysis. The screenshot analysis explored possible connections between the participants' initial preferences and the final output. The screenshots tracked the points allocated for each indicator and all executed scenarios. The data from the screenshots were entered into a Microsoft Access database. Correlations were generated from this database to make comparisons between the trade-offs applied by participants in order to reach satisfaction with the DST output. The points allocated for each forest value was calculated between scenarios. These results were averaged among stakeholder groups for each forest value. The results were used to compare differences and similarities of trade-offs between stakeholder sectors. 48 Run):15 5 0 Participant: test 60 New Participant Scenario Number: Total Desired Condition: Points Remaining: Multi Criteria Decision Making Tool Dunbar/Templeton LU Es 13.33 %t 5 )111 61.67 ^1 1 0 46.67% I lo^)11110 46.67 ZErEi 13.33 X ( 5̂ AMU ceriano • A.^0 zceriano ^ c;,enano . n- 0 cenario ^ oenario Profitability Over Tone 13 33 % 8 7 11 6 E 4 10 9 1111 Scenario 1: 1 5 ) Scenario 2: [ 20 Scenario 3: o Scenario 4: [ 20 Scenario 5: I s) 0 Scenario 7: [a) Scenario 8: Scenario 9: Scenario 10: ) Scenario 7: Scenario a 1 T. I Scenario 9: Scenario 10 i;60 :a .640 1 20 100 _as Li SFM Multi-Criteria Decision Making Tool Profitability 3 2 1 0 ^ N in.1 In 0 N CO 0 0 Time (10 year periods) Scenario 6:^0 o N 01 sr III CO f•-• CO 0 0 Time (10 year periods) Scenario 1: [ 10 Scenario ^ Scenario a 'T] 111 Scenario 4: •Scenario ^ IN Scenario 6:^0 Visual Quality Over Tone 100 _80 60 z.40 ex, °^N^lf) CO f•-• 03 0) 0 Time (10 year periods) Recreation S cerierio'9. S carieriL) 10. MOS %MU 100 . _80 I \\ E 60 z.40 &20 0 •-• N 'I' ID CON CO 0 0 Time (10 year periods) Scenario 1: Scenario 2: ^ Scenario 3: Scenario 4: Scenario 5: Scenario 6: Scenario 7: ^ Scenario 8: ^ Scenario 9: Scenario 10: o0 N̂ nt Ul CON' 0) 0)0 Time (10 year periods)^•- Scerkyro 1 Scenario 2^—1 IN Eicerrairro 3 ^j iriScenario 4: Scenario 5: Scenario 9:^0 S c er,ario 7:^0 1 Scenario 8. 7' Scenario 9. tn t Scenario 1 Oi^ Domestic Watershed Over lime 100 _80 Esc) 140 '220 o Ô N sr^0) 0) 0 Time (10 year periods) Employment^Visual Quality Ecosystems At Risk Domestic Watershed :a 140 20 _80 E,60 Employment Over Time 100 1 5 1 20 10 L U 15 [ 20 ) 0 [a ) [^0^) Recreation Over Tone^Ecosystems At Risk Over Tone Scenario Data Ready Figure 6. An Example of the Interface Output after Five Completed Scenarios. As stated previously, this study is exploratory. Therefore, finding statistical significance was not necessary, and hence the low participant numbers. Instead the purpose was to find relationships between the participant's preferences and the DST output. Thus, the data analysis focused on simple descriptive statistics for the close-ended question items on the surveys and the screenshots, as well as a content analysis of the open-ended question items on the surveys. The open-ended question items found in the questionnaires were analyzed using content analysis to examine the dominant themes of each stakeholder group. It was necessary to create a coding scheme for each open-ended question; this scheme defined the main themes found in the responses. The coding scheme for this analysis used manifest content because it is more reliable and specific than latent content due to the use of concrete terms (Babbie 2004). Furthermore, a quantitative content analysis was employed by recording the numerical frequency of certain words and phrases found in the responses. A more detailed description of the data analysis is found in the following Results section. 50 4.0 RESULTS This study is exploratory; therefore, the purpose is to find strong relationships between participant preferences and the Decision Support Tools (DST) output. In this chapter, general observations from the sessions are outlined in Section 4.1. Sections 4.2, 4.3, and 4.4 summarize the data from Questionnaire 1, the DST exercise, and Questionnaire 2, respectively. Section 4.5 provides linkages and relationships between all three modes of data collection. 4.1 General Observations from Sessions Overall, the participants' reaction towards the DST exercise was positive and enthusiastic; however, the majority of participants expressed some concerns and frustrations during the session. This section describes general observations by the session facilitator that were not explicitly captured in the data. Participants least comfortable with computers were the most likely to under appreciate the value of the model as a tool for decision-making. Property owners were very keen participants 3; however, the participants in this sub-group had very little to no experience using a computer. This resulted in noticeable apprehension and held back these participants from using the tool to its full potential. Further, property owners conducted the least amount of scenarios amongst all five stakeholder groups. Upon review of the DST's initial output; most participants admittedly stated that they were trying to manipulate the model to find the response they were looking for. One participant commented that he felt he was managing the model after reviewing the first output rather than managing his preferences. Participants were required to make trade-offs between the forest indicators if they were not satisfied with the DST output; therefore, feelings of manipulation may be 3 Property owners were keen participants because the study concentrated on a small area allowing them to easily visualize consequences and relate outcomes that may affect their investment. 51 connected to the discomfort associated with making a hard trade-off from their initial preferences. Finally, many participants did not agree with the data used to create the recreation indicator. As the summary reports indicate, this discrepancy in the data greatly impacted the way participants responded to this indicator. For example, some participants stopped allocating points to the recreation indicator altogether. Most participants wanted to keep the mill running; this decision greatly affected the recreation indicator. In the model, recreation is the only indicator found in every polygon in the Total Harvestable Land Base (THLB); therefore, more areas in the THLB are cut when the mill is open. As a result, the recreation indicator experiences a larger impact than the other indicators in the model. There was a general consensus among participants that the set-up for the recreation indicator was not appropriate. Participants believe that people adapt and continue to recreate as the landscape changes (or is harvested) over time. Furthermore, many participants noted that clearcuts can enhance rather than hinder the recreational experience for certain activities. 4.2 Initial Forest Value Preferences and Opinions of Decision Support Tools Questionnaire 1 was administered at the beginning of each session to gain an overall sense of participants' perceptions towards different forest values in the area, as a well as a general understanding of peoples attitudes and knowledge towards DSTs. The first section of this survey implemented questions on forest value preferences and opinions on sustainability. The purpose of Question 1 was to gain a sense of each participant's preferences towards various resource values within the Invermere TSA. The participant ranked each resource value based on its overall importance to society, the level of health and well-being derived from the resource value, and the level of knowledge the participant had about the 52 value. These questions were answered on a scale from 1 (low importance /satisfaction/knowledge) to 5 (extremely important/satisfied/knowledgeable). Table 4 provides a frequency distribution of participants' highest priorities (ranking of 4 or 5) for each resource value in the Invermere TSA. Table 4. Summary of Participant's Priorities towards Resource Values in the Invermere TSA. Importance Health & Well- Being Knowledge Cultural/Historical 65% 57.9% (19) 55% Ecosystem Health and Biodiversity 90% 73.7% (19) 75% Jobs 80% 75% (19) 80% Recreation/Tourism 90% 78.9% (19) 75% Timber Supply 65% 55.6% (18) 80% Visual Quality 50% 26.3% (19) 90% Water 95% 89.5% (19) 80% The majority of participants found all resource values to be important or extremely important; however, variation exists between resource values. For example, 80% or more participants agreed that water, recreation, jobs, and ecosystem health and biodiversity are important indicators. Whereas fewer participants rated the cultural/historical, timber supply and visual quality resource values as important. Furthermore, the majority of participants are satisfied or extremely satisfied with the level of health and well-being derived from each resource value, except for visual quality. The data indicates 75% or more of the participants have a good knowledge of the resource values, however, the cultural/historical resource value is found to be the lowest with only 55% of participants having an extensive knowledge of this value. The second question requires participants to provide their opinion on forest management practices; 65% of participants agreed that forest management in the Invermere TSA supports sustainable resources management. However, one quarter of the participants do not feel that forest management in the Invermere 53 TSA is sustainable. These participants cited over-harvesting as the main practice inconsistent with sustainable forest management; some participants elaborated on this point citing the mountain pine beetle salvage effort as a contributor to over-harvesting. The final question in the forest values and sustainability section required participants to rate the six forest values found in the model from 1 (most important) to 6 (least important). Using frequency distributions for each rating unit, the collective group rated the resource values from most to least important as the following: 1) ecosystems at risk, 2) water, 3) employment, 4) recreation, 5) profit, and 6) visual quality. The purpose of the second section in Questionnaire 1 was to gain more perspective on participants' opinions towards DSTs before conducting the exercise. The survey found that 80% of participants were either familiar with or had heard about DSTs, whereas the remaining 20% were unfamiliar with DSTs. Overall, 90% of participants agreed that DSTs should be used as a helpful aid in decision-making, but should only be used as a guide for making decisions, not as a means to determine the final solution. Participants were asked to identify the issues and problems that they felt were associated with DSTs. The responses varied amongst four key problems. The problem regarding assumptions received the most attention, with 50% of participants agreeing that too many assumptions are made by model developers (Table 5). Table 5. Problems with Decision Support Tools DST Problem Frequency Distribution (% of Participants) Assumptions 50% Bias 35% Complexity 25% Manipulation 30% None of the Above 15% Don't know 30% 54 Participants also provided feedback on factors that would make them feel more comfortable using DSTs. Again, the response was heterogeneous; however, 70% of participants agreed that it is necessary to understand and assess the assumptions of the model, and 80% of participants agreed that it is necessary to understand the general idea of why and how the model was built. Table 6 provides a breakdown of the frequency distribution of the responses to this question. Table 6. Requirements to feel comfortable with the DST Requirement Frequency Distribution (% of Participants) Trust the output 45% Trust development institution 45% Understand assumptions 70% Understand why and how the model was built 80% Experience using the model 30% None, I am comfortable 0% None, I will never be comfortable 0% The third section of Questionnaire 1 focuses on participants' knowledge of the area and their activities in the Dunbar/Templeton LU. The majority of participants felt that they were knowledgeable or extremely knowledgeable about the different resource values in the Dunbar/Templeton LU. This is expected as the recruitment criteria ensured participants had a vested interest in the Dunbar/Templeton LU. Participants listed the environment, forestry, the provincial government, recreation (motorized and non-motorized), and tourism as the main connections to the forest in the Dunbar/Templeton LU (Appendix 5). Furthermore, participants listed hiking, fishing, and camping as the most frequently practised recreational activities in the Dunbar/Templeton LU (Appendix 6). Table 7 provides a breakdown of the participants' property ownership, working relationship, and license activity on the Dunbar/Templeton LU. 55 Table 7. Participants Relationship with the Study Area Activity Frequency Distribution (% of Participants) Property Ownership 15% Work Nearby or In the LU 60% Licensee 35% Water license 15% Non-replaceable forest license 5% Firewood permit 5% Trapping/Hunting license 10% Finally, participants were required to state any major concerns in the Dunbar/Templeton LU in an open-ended statement. Table 8 provides a quantitative content analysis, recording the numerical frequency of manifest phrases related to concerns for different resource values in the Dunbar/Templeton LU. This data shows participants are highly concerned with problems related to timber supply and forest health in the Dunbar/Templeton LU. Over-harvesting is mentioned as a concern for a declining timber supply and unbalanced age classes, whereas the key concern related to forest health is the Mountain Pine Beetle infestation. Other areas for concern include wildlife protection, recreation pressures, water protection, and profitability in the area. The final section of Questionnaire 1 collected personal background information from each participant. In total, 15 males and 5 females participated in this study ranging between the ages of 36-85 years old. Participants live in various locations throughout the Invermere TSA including: Radium Hot Springs, Edgewater, Fairmont Hot Springs, the City of Invermere, Cranbrook, Golden, Wilmer, and Brisco. Table 9 outlines the education level of the participants; 75% of participants have a university or college degree. 56 Table 8. Participants Concerns for Forest Values in the Dunbar/Templeton LU 4 Profit Timber supply Recreation Water Wildlife Forest Healthy Total Forest Sector Responses Industry 2 (7.4%) 3 (11.1%) 0 (0%) 0 (0%) 0 (0%) 2 (7.4%) 7 (25.9%) ENGO 0 (0%) 1 (3.7%) 1 (3.7%) 1 (3.7%) 1 (3.7%) 1 (3.7%) 5 (18.5%) Government 0 (0%) 2 (7.4%) 0 (0%) 1 (3.7%) 2 (7.4%) 1 (3.7%) 6 (22.2%) Recreation 0 (0%) 1 (3.7%) 2 (7.4%) 1 (3.7%) 1 (3.7%) 2 (7.4%) 7 (25.9%) Private property 0 (0%) 0 (0%) 1 (3.7%) 0 (0%) 1 (3.7%) 0 (0%) 2 (7.4%) Total Pressure Type Response 2 (7.4%) 7 (25.9%) 4 (14.8%) 3 (11.1%) 5 (18.5%) 6 (22.2%) 27 (100%) 4 One participants response to this question was unclear in the survey data; therefore, the sample size for this analysis is 19 participants. 5 Forest health refers to disease and insect issues, including Mountain Pine Beetle infestations. Table 9. Education level of Participants Education Frequency Distribution (% of Participants) some high school 5% high school 5% some university/college 10% University/college 75% graduate degree 5% 4.3 The Output Summary from the Decision Support Tool Exercise The number of scenarios conducted by participants ranged from 1 to 19. The average number of scenarios conducted by the collective group was 7.25 (standard deviation = 4.62), while the mode was 9 scenarios. Table 10 shows the number of scenarios conducted by each stakeholder group. The recreation group had the highest average at 9 scenarios, and the Private Property group had the lowest average at 3.25 scenarios. It was possible to compare standard deviations between stakeholder groups because the sample size is the same for each group. The standard deviation for the industry, ENGO, and Private Property group is low; thus, these groups conducted a similar number of scenarios within their stakeholder group. The government and recreation groups had larger standard deviations; thus, the number of scenarios conducted by participants in these groups varied. Table 10. The Number of Scenarios Conducted by Each Stakeholder Group Forest Stakeholder Group Average No. of Scenarios Standard Deviation Minimum No. of Scenarios Maximum No. of Scenarios Industry 7.5 1.73 6 9 ENGO 8.25 2.363 5 10 Government 8.25 7.805 2 19 Recreation 9 5.888 3 17 Private Property 3.25 1.708 1 5 58 The average preferred scenario is 5.6 for the collective group. The median and mode are both 5. The standard deviation is 4.096. The minimum chosen preferred scenario is 1 and the maximum preferred scenario is 17. On average, participants did not choose the initial scenario as the preferred scenario and instead chose a preferred scenario amongst the final scenarios conducted. The average difference between the total number of scenarios conducted and the preferred scenario is 1.65, and the median and mode is 1. The standard deviation between the difference is 1.69. Again, the Industry, ENGO, and Private Property groups have small standard deviations, whereas the Government and Recreation groups have large standard deviations. Table 11. The Preferred Scenario Chosen by Each Stakeholder Group Forest Stakeholder Group Preferred Scenario Standard Deviation Minimum Preferred Scenarios Maximum Preferred Scenarios Industry 6 0.816 5 7 ENGO 6.25 1.5 5 8 Government 5.5 5.916 1 14 Recreation 8.25 6.291 2 17 Private Property 2 1.155 1 3 The trade-offs between the first scenario and the preferred scenario for each indicator is shown in the following Figures 7a to 7f for the collective group as an increase, decrease, or no change. 40% 50% 10% 0 Decreased ■ Increased 0 Same Figure 7a. Preference Changes Between the First and Preferred Scenarios: Profit 59 25%^25% 50% o Decreased Increased ^ Same Figure 7b. Preferences Changes Between the First and Preferred Scenarios: Employment La Decreased mi Increased ^ Same Figure 7c. Preferences Changes Between the First and Preferred Scenarios: Recreation 20%^15% 65% 0 Decreased ■ Increased ^ Same Figure 7d. Preferences Changes Between the First and Preferred Scenarios: Ecosystems At Risk 60 45% 50% 45% 5% o Decreased is Increased ^ Same Figure 7e. Preferences Changes Between the First and Preferred Scenarios: Visual Quality 15% Decreased Increased ^ Same Figure 7f. Preferences Changes Between the First and Preferred Scenarios: Domestic Watershed The charts clearly illustrate that certain indicators impacted participants initial preferences more than others. The Employment indicator and Ecosystems at Risk indicators were the only indicators in which the majority of participants increased their point allocation between their initial and preferred scenarios. The majority of participants decreased the point allocation from their initial preference scheme to the preferred scenario for the Profit and Recreation indicators. The Visual Quality indicator is the only indicator wherein the majority of participants kept their point allocation the same for both the initial scenario and the preferred scenario. The Domestic Watershed indicator did not indicate any decisive preference change, as both a decrease and increase in the point allocation between the first and preferred scenarios had a strong representation. 61 The pie charts for the individual stakeholder groups (Appendix 7) show the trade- offs made between indicators for each group. There were general similarities amongst stakeholder groups upon review of the DST output. All stakeholder groups decreased the Profit indicator, except the Recreation group. The Recreation group was divided on whether they increased or decreased the profit indicator. Overall, the Employment indicator was increased; however, the Industry group decreased the emphasis on Employment from their initial scenarios and 50% of the Government group decreased the initial points allocated to Employment. All groups decreased the Recreation indicator. All stakeholder groups increased Ecosystems at Risk. However, the Recreation group was again divided on whether they increased or decreased the Recreation indicator. The collective data for the Domestic Watershed indicator does not illustrate a clear trade-off; interestingly, the data is very different when analyzing the sub-groups. All participants in the Industry group increased the points allocated to the Domestic Watershed indicator. The ENGO group was divided on whether to increase or decrease the emphasis on this indicator, whereas the other groups decreased the points on this indicator. The emphasis on the visual quality indicator either decreased or stayed the same; however, the Recreation group was divided between increasing and decreasing the points. Figures 8a to 8f describe the average change in points for each indicator between the first and preferred scenario. This information will help explore the degree of preference change for each indicator. Figure 8a illustrates this change for the collective group of participants whereas, the remaining graphs show this change for each forest stakeholder group. 62 k`.^c'•`•^0^.^e^0 ,e,0q),o,^\\10.^co^,,,co e^\\).(1'^.0^0.4,_ 0c,,.\\\\ ek0 ca Initial Sc ill Preferred Sc Figure 8a. The Average Change in Points from the Initial Scenario to the Preferred Scenario for the Collective Group of Participants The most substantial increase between the initial and preferred scenario is shown by the Ecosystems at Risk indicator. The largest decrease from the initial to the preferred scenario is the Recreation indicator. Initially, the most points were given to the Ecosystems at Risk indicator; the Employment and Domestic Watershed indicators were also given high points. In the preferred scenario the most points were again given to the Ecosystems at Risk, Employment, and Domestic Watershed indicators. The difference in the preferred scenario is that more points were taken from the Profit, Visual Quality, and Recreation indicators and added to the Employment and Ecosystems at Risk indicators; the Domestic Watershed indicator did not change. Figures 8b to 8f illustrate the difference between stakeholder groups and the collective group. 63 .11 1^[I. ,,,^,, 4c• 0 P\\. QC 0^ e^.).(6.^2 r^\\ .0 s *q: ' 0 Initial Scenario D Referred Scenario Figure 8b. The Average Change in Points from the Initial Scenario to the Preferred Scenario for the Industry Stakeholders The most substantial increase between the initial and preferred scenarios for the industry stakeholders is shown by the Ecosystems at Risk indicator; the Domestic Watershed indicator has a significant increase as well. The largest decrease from the initial to the preferred scenario is the Profit indicator. Initially, the most points were given to the Employment and Profit indicators. In the preferred scenario, the most points were given to the Ecosystems at Risk and Employment indicators; the Domestic Watershed indicator was close behind. The Industry stakeholders made negligible change to the Employment indicator. C '50_ 30 25 20 15 10 5 0 6\\^0-^6°^,\\co .),ez\"^