Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Participatory model building for exploring water management and climate change futures in the Okanagan… Langsdale, Stacy Marie 2007

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

Item Metadata

Download

Media
831-ubc_2007-267479.pdf [ 21.31MB ]
Metadata
JSON: 831-1.0100360.json
JSON-LD: 831-1.0100360-ld.json
RDF/XML (Pretty): 831-1.0100360-rdf.xml
RDF/JSON: 831-1.0100360-rdf.json
Turtle: 831-1.0100360-turtle.txt
N-Triples: 831-1.0100360-rdf-ntriples.txt
Original Record: 831-1.0100360-source.json
Full Text
831-1.0100360-fulltext.txt
Citation
831-1.0100360.ris

Full Text

PARTICIPATORY MODEL BUILDING FOR EXPLORING WATER MANAGEMENT AND CLIMATE CHANGE FUTURES IN THE OKANAGAN BASIN, BRITISH COLUMBIA, CANADA by STACY MARIE LANGSDALE B.S., University of  Maryland, 1995 M.S., University of  Nevada, Reno, 2001 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Resource Management and Environmental Studies) THE UNIVERSITY OF BRITISH COLUMBIA April 2007 © Stacy Marie Langsdale, 2007 ABSTRACT Studies of  climate change impacts on water resources show that some regions may experience negative impacts and additional strain on the ability to meet future  demand. However, few  practitioners have incorporated climate change into their water planning initiatives. To do so practitioners must first  recognize climate change as a concern, acquire climate impacts information  specific  to their issues and scales, and then assess the potential impacts and adaptation options within the context of  the system. Participatory modeling, in which stakeholders are actively involved in the construction of  a computer model, is an effective  method for  accomplishing these tasks. The collaborative process fosters  a shared learning experience and the model helps assess future  conditions and policies. A year-long participatory modeling exercise was conducted in the Okanagan Basin in south-central British Columbia, Canada. The region's arid, snowmelt-dominated hydrology combined with recent rapid development make its water resources susceptible to climate change impacts. Participants, including water-related professionals,  researchers, and representatives of  non-governmental organizations, assisted in all stages of  model development, from  goal setting and issues identification  to model calibration and testing. The completed model, constructed in STELLA™, conducts thirty-year monthly simulations of  water supply and agricultural, residential, and conservation flow  demands for  a historic period and for  the 2020's and the 2050's, using statistically downscaled climate information  from  the Hadley, CGCM2, and CSIRO general circulation models. The model suggests that climatic changes could impact the system more severely than population growth. Current projections show reduced ability of  the system to meet demand, particularly during the dry month of  August, when demand peaks. Adaptation strategies could play a role in maintaining system reliability. Participants found  both the process and the resulting model valuable. They found  the model to be a relevant and legitimate tool for  exploring long-term water management in the Okanagan when used with the appropriate audience and with minor refinements.  The model could support further  dialogue with the Okanagan community to determine appropriate management options. This methodology is not limited to this case study, but is well-suited for  other applications of  resource management, policy development, collaborative learning and negotiation. TABLE OF CONTENTS Abstract ii Table of  Contents iii List of  Tables vii List of  Figures viii Acknowledgements x Co-Authorship Statement xii CHAPTER 1: LITERATURE REVIEW AND RESEARCH SCOPE 1 1.1 Climate Change and Water Resources Planning 2 1.1.1 Challenges of  incorporating climate change in water resource planning 3 1.1.2 Response options to climate change 5 1.1.3 * Reactive versus anticipatory adaptation 6 1.2 Integrated Assessment 7 1.2.1 Benefits,  challenges and future  directions 9 1.3 Participatory Planning Methods, Including Participatory Modeling 11 1.3.1 Degree of  participation 12 1.3.2 Problem type 14 1.3.3 Profile  of  participants 15 1.3.4 Goal of  participation 16 1.3.5 Challenges in participatory planning 16 1.3.6 Participatory modeling :.17 1.3.7 Research directions in participatory planning 18 1.4 Systems Thinking and System Dynamics 19 1.4.1 Characteristics of  complex systems: rates, levels, and feedback  loops 21 1.4.2 Complex system behaviour 22 1.4.3 General model theory 23 1.4.4 Use of  system dynamics models 24 1.4.5 Challenges .....26 1.5 Managing Uncertainties for  Planning 27 1.5.1 Inherent uncertainty vs. knowledge uncertainty 27 1.5.2 ' Climate change and modeling 28 1.6 Case Studies in Participatory Modeling 28 1.7 The Okanagan Project 31 1.7.1 Previous work 32 1.7.2 Research obj ectives 35 18 References  37 CHAPTER 2: SHARED LEARNING THROUGH GROUP MODEL BUILDING FOR THE MANAGEMENT OF WATER RESOURCES AND CLIMATE CHANGE IN THE OKANAGAN RIVER BASIN, BRITISH COLUMBIA, CANADA 47 2.1 The Okanagan Basin 48 2.1.1 Previous initiatives in the Okanagan Basin 48 2.2 Climate Change in Water Resources Management 50 2.2.1 State of  the field  50 2.2.2 Challenges of  incorporating climate change in water resource planning 51 2.3 Methodology 53 2.3.1 Integrated assessment 53 2.3.2 Participatory modeling 53 2.3.3 Process design: means and ends objectives 55 2.4 The Group Model Building Process 58 2.4.1 Participant recruitment 58 2.4.2 Involving participants in model development 61 2.5 Results 64 2.5.1 Means objectives 65 2.5.2 Ends objectives 68 2.6 Discussion 72 2.6.1 Lessons learned 72 2.6.2 Possible sources of  bias 74 2.6.3 Did this process make a difference?  : 75 2.7 References  77 CHAPTER 3: A SYSTEM DYNAMICS MODEL FOR EXPLORING WATER RESOURCES FUTURES UNDER CLIMATE CHANGE 82 3.1 Climate Change and Water Resources Planning 83 3.2 The Okanagan Basin Study Area 84 3.3 Project History 86 3.4 Methodology 87 3.4.1 Participatory modeling 87 3.4.2 System dynamics 87 3.4.3 Representing uncertainty using scenarios 88 3.5 Description of  the Actual and Modeled System 90 3.5.1 Components of  the Okanagan Sustainable Water Resources Model 91 3.5.2 Dynamics of  the system 97 3.6 Results 101 3.6.1 Managed supply vs. maximum demand 102 3.6.2 Maximum demand versus total allocation 108 3.6.3 Deficit  and adaptation 109 3.6.4 Implications for  future  management 111 3.7 Conclusions ! 112 3.8 References  ...114 I CHAPTER 4: CONCLUDING REMARKS 119 4.1 Project Summary: Objectives & Results 120 4.1.1 The participatory process 121 4.1.2 The Okanagan Sustainable Water Resources Model (OSWRM) 123 4.2 Impact of  Research Results 126 4.3 Lessons and Recommendations 127 4.4 Future Directions 128 4.4.1 Potential for  continued efforts  in the Okanagan 128 4.4.2 New applications for  participatory modeling 130 4.4.3 Advancing the field  : 131 4.5 References  133 APPENDICES 136 Appendix A: Behavioural Research Ethics Board Certificate  of  Approval 136 Appendix B: Workshop Evaluation Forms 138 Appendix C: The Okanagan System Model: Quick Reference  141 Appendix D: Model Level Documentation for  the Okanagan Sustainable Water Resources Model 154 LIST OF TABLES Table 2.1: Summary of  events in the group model building process 63 Table 2.2: Number of  responses to the post-workshop evaluation question "Have your perceptions of  future  water availability in the basin changed due to this exercise?" 70 Table 2.3: Comparison of  responses by event to the evaluation question: "Do you feel  this model is a legitimate and relevant tool to explore long-term water management in the Okanagan?" 70 Table 2.4: Average responses on a scale from  1 (low) to 5 (high) for  the question: "How well do you understand the model's structure?" 71 Table 3.1: Summary of  deficit  years as defined  on the scatter plot (Figure 3.8) for  all climate scenarios, showing (a) Rapid population growth and (b) Slow population growth 105 Table 3.2: Allocations as a percent of  demand, shown as annual totals and for  August, the month with the greatest deficit  in the future  scenarios 109 Table 4.1: Allocations as a percent of  demand, shown as annual totals and for  August, the month with the greatest deficit  in the future  scenarios 125 LIST OF FIGURES Figure 1.1: Figure 1.2: Figure 1.3: Figure 2.1: Figure 2.2: Figure 2.3: Figure 2.4: i Figure 2.5: Figure 2.6: Figure 2.7: Figure 3.1: Degrees of  stakeholder participation in policy making and in scientific  practice (from  van de Kerkhof,  2003:26) 12 Okanagan Basin Map with inset for  location in British Columbia (from  Cohen et al. 2006) '. 32 Flow chart illustrating the progression from  climate models and local records to supply and demand inputs to the Okanagan Sustainable Water Resources Model.. (Sources: Cohen et al. 2004; Neilsen et al. 2001; Merritt and Alila 2006; Neale 2005; 2006) 33 Okanagan Basin Map with inset for  location in British Columbia (From Cohen et al. 2006) 49 Attendance at all events by participants categorized by affiliation.  Plotted as a comparison to the total number of  people, and weighted by the number of events each person attended 59 Percent attendance of  all participants at all six events, categorized by role in Okanagan. Plotted as a comparison to the total number of  people, and weighted by the number of  events each person attended 59 Sketch created by a participant group at Workshop 2 illustrating the important linkages that affect  water supply in the Okanagan Basin 64 c Participant counts according to the number of  events they attended 66 Organization representation counts according to the number of  events attended. 67 Distribution of  attendance for  each of  the five  workshops and the series of  small group meetings, grouped by participants' roles in Okanagan water management. 73 Okanagan Basin Map showing the delineation of  the Uplands water supply model region that includes all managed tributaries to Okanagan Lake. Numbers correspond to watershed names, which are available in the Table 1 in Appendix D (Original map from  Merritt and Alila 2003) 85 Figure 3.2: Graph (a) shows results as an array of  discrete states, while graph (b) shows only the average of  the five  scenarios, along with error bars that provide the range of  possible conditions 89 Figure 3.3: Flow chart illustrating the progression from  climate models and local records to supply and demand inputs to the Okanagan Sustainable Water Resources Model. (Sources: Cohen et al. 2004; Neilsen et al. 2001; Merritt and Alila 2006; Neale 2005; 2006) 94 Figure 3.4: Upland streamflow  for  historic and 2050's climate scenarios. Two years are shown: 1976-77 for  the historic scenario, and 2055-56 for  the 2050's scenarios. 95 Figure 3.5: Causal Loop Diagram of  the Okanagan Basin water resources system 98 Figure 3.6: Population projections for  the Uplands water users 101 Figure 3.7: Thirty-year annual averages of  total managed supply and maximum demand from  the Uplands, showing trends through time for  multiple climate scenarios with (a) rapid population growth, and (b) slow population growth 103 Figure 3.8: Annual total managed supply and total maximum demand for  the Hadley A2 climate scenario and rapid population growth among Uplands water users. ...105 Figure 3.9: Thirty-year average monthly managed Uplands supply and maximum demand profiles  with demand from  the three major sectors revealed 107 Figure 3.10: Thirty-year average annual summary comparing total demand (all three sectors) and total water allocated in the Uplands for  the rapid population growth scenarios 109 Figure 4.1: Annual total managed supply and total maximum demand for  the Hadley A2 climate scenario and rapid population growth among Uplands water users. ...124 Figure 4.2: Thirty-year average annual summary comparing total demand (all three sectors) and total water allocated in the Uplands for  the rapid population growth scenarios 124 ACKNOWLEDGEMENTS My graduate experience was both enjoyable and rewarding because of  my colleagues, mentors, friends,  and family.  Several people had a hand in getting me to UBC, to RMES and onto the Okanagan project. Dr. McCuen first  encouraged me to consider a doctorate when I was applying to graduate programs in 1998. John Tracy also encouraged me on this path when I was completing a Masters under his supervision, but more critically, he introduced me to system dynamics for  watershed planning and stakeholder communication, a topic which excited me enough to study further.  Then, Ziad Shawwash invited me to UBC with financial  support. Once at UBC, Hans Schreier, Les Lavkulich, and Barbara Lence provided direction and support in my transitions to the Okanagan project and the Resource Management, Environmental Studies Program. i I'd like to thank Stewart Cohen for  preparing fertile  ground among the Okanagan community and for  enthusiastically supporting this project from  the first  day I suggested it. Stewart opened my eyes to the climate change issue and taught me the importance of  recognizing those who deserve credit. j Craig Forster generously contributed his time and expertise in designing and facilitating  the stakeholder workshops, in developing the model's user interface,  and in reviewing manuscripts. Thanks for  inviting me to observe a workshop and meet with you in Salt Lake, and especially thanks for  all your mentorship that helped me to think more critically about my research and to envision my career options. I am so lucky that Murray Journeay introduced us. Tina Neale has been both a colleague and a friend.  Thanks for  all the brainstorming sessions and for  being a sounding board when I needed to advice or to talk things through. And for teaching me how to use WORD effectively!  Your support was invaluable. Jeff  Carmichael provided valuable advice on how to make this project both academically rigorous as well as realistic. Thanks for  all the advice and encouragement along this journey and for  facilitating  at all the workshops. Thanks to Allyson Beall for  generously supporting this project with ideas on workshop design, feedback  on manuscripts, and for  facilitating  at all the workshops. Thanks to Andy Ford for  inviting me to meet with you and your research group in Pullman, where I met Allyson. Thanks to all the Okanagan group model building participants. Their enthusiasm really made this work worthwhile and inspired me to continue. Thanks to those who helped with the workshops: Jessica Durfee  provided the ice cream game and took spectacular meeting documentation, Alison Shaw assisted in creating the evaluation forms,  and Natasha Schorb took notes. Natural Resources Canada provided funding  this project through the Climate Change Impacts and Adaptation Program. Thanks to all those that helped me through the writing process. Sonja Klinsky reviewed my draft.  Patricia Keen walked with me on the final  lap. Haim Behar gave me many survival skills by training me in the art of  tai chi. Liena Vayzman taught me life  organization tools when I needed them most. And special thanks to my family  that supported me along the way, including my parents and my loving husband, Daniel. CO-AUTHORSHIP STATEMENT Chapter 3 was submitted to the Integrated Assessment journal and Chapter 2''will be submitted to the Journal of  Water Resources Planning and Management. I am the senior author and originated the research including the concepts and ideas, wrote both manuscript, and took major responsibility for  development of  both papers. My co-authors offered suggestions and are listed in alphabetical order and assisted in the following  ways: • Allyson Beall advised design of  the workshops and development of  the model, helped with facilitation  at all workshops and reviewed manuscript drafts. • Jeff  Carmichael advised design of  the workshops and development of  the model, helped with facilitation  at all workshops and advised the development of  manuscript drafts. • Stewart Cohen, as project PI, ensured that the group model building process and model fulfilled  the goals of  the overall project. He generated adaptation scenarios and advised the development of  manuscript drafts. • Craig Forster advised design of  the workshops and development of  the model. He generated the interface  level of  the model and provided significant  advice on the presentation of  model results. Tina Neale is also a co-author for  Chapter 2, for  her contributions in generating ideas for  the design of  the stakeholder workshops and providing invaluable support in conducting and evaluating these events. Excerpts of  text in Chapters 1 and 3 were taken from  a paper accepted by the Journal of Contemporary Water Research and Education on which I am the sole author. Because the thesis is in a manuscript-based format,  some text is repeated in multiple chapters. CHAPTER 1: LITERATURE REVIEW AND RESEARCH SCOPE Sustainable development, as defined  by the Brundtland report is "development that meets the needs of  the present without compromising the ability of  future  generations to meet their own needs" (Brundtland 1987). Since water is a renewable resource, we can meet this objective by avoiding actions which degrade water quality and by limiting our use of  water to the rate that supplies are replenished. The American Society of  Civil Engineers Task Committee on Sustainability Criteria defined  sustainable water resource systems as those that are "able to satisfy  the changing demands placed on them, now and on into the future,  without system degradation" (ASCE 1998: iv). This is easy to state but challenging to achieve, particularly when natural hydrologic variability is combined with impacts from  human development and climate change. Several elements are required to effectively  conduct Sustainable Water Resources Planning and Management. First, the planning must be long-term and consider all of  the issues that could affect  the system over the long term, such as climate change, population, economic forces,  and human values. Second, the assessment of  future  conditions must integrate as many factors  as possible, and analyze the issue from  a system context. Third, the assessment needs to involve the local community to enhance the analysis, but also to foster  learning among actual planners, managers and users of  the resource. This dissertation describes an application of  sustainable planning for  water resources in the Okanagan Basin, British Columbia, Canada. The approach follows  the above listed elements to the extent possible. The following  sections provide detail about the Okanagan Study area, climate change in the water management context, and the fields  of  "integrated assessment," "participatory planning," and "system dynamics." Selected case studies that applied participatory integrated assessment to natural resource problems are summarized. 1.1 Climate Change and Water Resources Planning Water managers have always worked towards reducing risk and increasing system capacity to handle ever-widening extreme conditions, but are managers prepared with the tools to adapt to future  climate conditions? In the past, climate was assumed to be relatively stable in the timescale that water managers work in, varying around a stable mean (de Loe and Kreutzwiser 2000). Stakhiv (1996) observes that society is constantly adapting in incremental steps and predicts that climate change will simply be an additional stressor to which we must adapt. This practice is generally low-risk as long as changes occur gradually. In the future, climate changes may occur gradually, over several decades, or they could occur as rapid step changes. Most climate models characterize climate change as occurring slowly and gradually, which justifies  a reactive "wait-and-see" approach. In contrast, Kashyap (2004) suggests that climate change adaptation is not comparable to historic adaptation because the environmental changes will be more rapid and intense than in the past. The IPCC's Fourth Assessment Report confirms  this idea, stating that heat waves and heavy precipitation events will become more frequent  and intense in most areas (IPCC 2007:16). A shift  in the climatic mean, combined with the possibility for  sudden changes, makes the conventional practice of relying on historic data for  estimating future  conditions inadequate. New methods for assessing future  conditions are required to maintain the reliability of  operations over the long term. This is true not only for  water resources, but for  other industries as well. For example, the insurance industry, which relied solely on historic records for  centuries, has also started to consider future  climate conditions for  assessing hazards and setting premiums (Leggett, 2001). The impacts of  climate change on hydrologic systems will vary widely by geographic location. For example, climate models are generally estimating more precipitation globally but some areas will become wetter while others will become drier (Miller and Yates 2005). Many areas will experience warmer temperatures, which will impact hydrologic storage and timing. Snow-melt driven systems will experience a reduced annual snowpack and an earlier spring freshet  (Mote, et al. 2003; Mote 1999; Taylor and Barton 2004). These changes may impact the timing of  available water, reducing the availability of  water during the summer irrigation season. Climate change impact studies are relatively recent, and direct use of  the information  by practitioners is lagging behind. In 2005, only four  U.S. states included climate change in their water resources planning (Viessman and Feather 2006). In Canada, two examples of  climate change being incorporated explicitly into water planning are the Trepanier Landscape Unit Water Plan (Summit 2004), which followed  from  work by Cohen et al. (2004); and work on the Great Lakes - St. Lawrence River Basin (Mortsch and Mills 1996). 1.1.1 Challenges of incorporating climate change in water resource planning There are three prerequisites that must be met for  water managers to be willing and able to incorporate aspects of  climate change into their water planning initiatives. First, the water management community must be informed  and concerned about climate change. Second, information  about potential climate impacts must be translated into terms that are relevant to the water community. Third, the water community must be able to assess the climatic impacts within the system context, including other stressors, changes, and management responses. Each of  these tasks contains specific  challenges. There are several reasons for  the lack of  awareness of  climate change among water resource professionals.  First, detecting the climate signal is challenging, as the signals are often confused  with noise, or not felt  directly (Berkhout et al. 2004). In recent years, before  climate change had discernable, significant  impacts as described in the Intergovernmental Panel on Climate Change's latest report (IPCC 2007), the "signal" was virtual, in the form  of  scientific predictions. This adds a new dimension of  complexity; since there is significant  uncertainty in the character of  climate change, how we should respond is not clear. Humans respond more readily to stressors that we personally experience than those that we learn about indirectly. The creeping nature of  climate change makes it easy to ignore. Because our human nature encourages us to focus  on immediate crises, small incremental changes are disregarded1 and continue freely  until a disaster occurs (Moser and Dilling 2004). Many climate adaptation proponents focus  on extreme events; however, in the case of  water resources management, the series of  small events, or the shift  in patterns to a more non-equilibrium dynamic can be more damaging (Dowlatabadi and Yohe 1999; Scoones 2004). For example, damage due to gradual shifts  was demonstrated in an analysis of  reservoir operation in the Columbia River Basin. Researchers at the Univ. of  Washington's Climate Impacts Group discovered that it was the seasonal shifts  of  inflow  that had the most significant  impacts on system performance  (Payne et al. 2004). Next, the information  typically generated by climate scientists is not directly relevant to local water professionals.  In order to incorporate climate change estimates into their planning processes, climate change information  must be translated into terms that are relevant to their concerns. While the level of  uncertainty in climate estimates makes the information cumbersome, the mismatch of  scales is a larger hurdle. Current global climate models provide information  at large geographic scales and low spatial resolution, but managers handle small geographic areas and require data with relatively high spatial resolution (Lins et al. 1997). Furthermore, climate change data must be translated from  temperature and precipitation to terms reflecting  hydrologic impacts. Finally, even if  the first  two challenges are surmounted, and concerned water professionals acquire relevant information,  there remains the challenge of  integrating the information within the system context. For example, climate change is likely to impact not only the availability of  water supplies, but also water demand. At the same time, other stressors, such as urban development and economic trends may also affect  demand in various sectors. Only • by considering climate change in the system context can we estimate future  conditions and then evaluate effective  responses. To date, most work has focused  on specific  components rather than the entire system. For example, Downing et al. (2003) assessed the impacts of climate change on various water demand sectors with the caveat that the results need to be considered as one element in a larger system. 1.1.2 Response options to climate change There are two primary responses to climate change: mitigation and adaptation. Mitigation refers  to actions for  reducing anthropogenic greenhouse gas emissions to rates at or below the capacity of  the earth to assimilate them. However, we have already committed ourselves to many decades of  elevated greenhouse gas levels in the atmosphere and the resulting climatic changes. Therefore,  adaptation strategies which can reduce vulnerabilities are prudent for coping in the short to medium term. There are several definitions  in the literature for "adaptation." De Loe and Kreutzwiser (2000:164) define  adaptation broadly as "[a]ny adjustment in a system in response to climate stimuli." Rosenzweig and Hillel (1998) provide a more detailed definition,  which includes the idea that adaptation can be social or technical, and that the purpose of  adapting is to either reduce the negative impacts or capitalize on the positive effects  of  climate change. The Adaptation and Impacts Research Division of  the Meteorological Service of  Canada, Environment Canada conducted several studies that became the foundation  for  conducting this research project. The organization's web page describes their stance on the importance of adaptation for  Canadians: Atmospheric change, variability and extremes represent real and present threats to the achievement of  sustainable development in Canada. They will continue to affect  the health, integrity, development and, thereby, the sustainability of  Canadian socio-economic and ecological systems. As such, Canadians' responses to these must include adaptation actions that will reduce their vulnerabilities to atmospheric variability and extremes and that will minimize the negative impacts, maximize positive impacts, and allow them to take advantage of  opportunities that arise as a result of  atmospheric changes (EC website 2002). The hope for  the organization is "to promote and facilitate  adaptation to atmospheric change, variability and extremes and to assist in identifying  the need for  other response options" (EC website 2002). If  we are to achieve a sustainable society, we must make changes that support mitigation and adaptation simultaneously. There are numerous examples of  short-term adaptations that conflict  with mitigation. For example, more intense summer heat waves encourage more use of  air conditioners, but air conditioners increase energy use, which increases greenhouse gas emissions. A better response would be to plant trees which shade buildings and provide carbon sequestration. In this project, I focus  on adaptation, and do not evaluate mitigative potential. Future research should evaluate alternatives in a mitigation framework  as well, to ensure that short-term solutions do not exacerbate long-term problems. In this project, I place higher emphasis on alternatives that reduce water demand than on supply-based alternatives that increase supply. Generally, reductions in use of  materials and energy contribute to a more sustainable community. 1.1.3 Reactive versus anticipatory adaptation Water managers could implement changes now to reduce their vulnerability to future climates, or they could wait and see what the future  brings and then respond to these new conditions. The first  approach, known as "anticipatory adaptation," describes investments made today to reduce future  risk. The second approach, "reactive adaptation," prevents unnecessary investment and delays expenditure, but increases system vulnerability. There are strong arguments supporting each view. The Stern Reveiw on the Economics of  Climate Change (2006) supports anticipatory adaptation, by concluding that minor annual expenditures (1% of  the GDP) are required to manage climate change, but if  this investment is not made, the damages to the economy will be severe (decreasing GDP by 20%). Smith et al. (1995) explain that some of  the first  adaptation debates focused  on whether or not reactive adaptation measures were capable of  offsetting  the negative impacts of  climate change. If  the onset of  climate change is rapid, with frequent  extreme events, there may be insufficient  time for  humans, flora  and fauna  to adapt; "[t]his makes a case for  the necessity of  anticipatory adaptation" (Smith et al. 1995:202). On the other hand, studies of  several U.S. cities in the 1990s determined that water management systems were already incrementally adapting in response to various external factors;  therefore,  these systems were robust and resilient enough to handle future  climate variability and anticipatory adaptation is unnecessary. "The choice of  an anticipatory path requires a profound  investment and behavioural changes that cannot be justified  by climate change analyses completed to date" (Stakhiv 1996:246). Middlekoop et al. (2004) explains how world views (regarding nature, society and risk) combined with management styles largely influence  choices about appropriate adaptation to climate change. In a study of  five  international rivers (the Nile, Zambezi, Indus, Mekong, and Uruguay), basin managers felt  that a combination of  traditional structural and non-structural approaches implemented now would help them to cope in the future  (Riebsame et al. 1995). In contrast, an evaluation of  public water supply in England and Wales concluded that at this point in time, "no specific  actions were necessary to deal with future  climate change" (Arnell and Delaney 2006). 1.2 Integrated Assessment Incorporating climate change in future  assessments and adapting to climate change are important, but are only relevant when taken in context. It may be possible that other stressors will exaggerate, offset,  or overwhelm the impacts of  climate change. Surprising behaviours, such as feedbacks  and. non-linearities may emerge when numerous factors  of  influence combine. Therefore,  assessments of  the future  should ideally consider the whole system, with the full  array of  issues that could change, including climate, economic influences,  population, and human values. Integrated assessment (IA) is a field  of  study that works to do just that. Several definitions  of  IA exist in the literature. All agree that IA is a method or process for evaluating complex problems to increase understanding (Rotmans and van Asselt 2002; Risbey et al. 1996). Toth and Hizsnyik (1998:194) noted that all definitions  share a commonality in that IA "is an interdisciplinary and policy-oriented synthesis of  scientific information,"  however the definitions  differ  in their qualifications.  One of  the most recent definitions,  provided by Rotmans and van Asselt (2002:1), refers  to participatory processes (PPs): IA is "an interdisciplinary and participatory process of  combining, interpreting and communicating knowledge to allow a better understanding of  complex phenomena." In earlier definitions,  PPs were considered one of  several approaches to conducting IA; however, Rotmans and van Asselt (2002) claim that participation is now a necessary component for  conducting high quality IA. The aim is thus to analyse, explore and evaluate past, current and future developments in terms of  plausibility, desirability and feasibility.  Integrated assessments should result in added value compared to insights derived from disciplinary research. IA has the explicit purpose to inform  policy and support decision-making. It is important to realise that integrated assessment is not a pure scientific  activity. It requires involvement of  scientific  experts, stakeholders and decision-makers. Communication between those different actors is at the very heart of  IA (Rotmans and van Asselt 2002:1). This recent shift  to perceiving participation as a vital component of  IA can be attributed to the emergence of  new social and scientific  philosophies. The post-modern and social-constructivism movements of  the late twentieth century recognize that science is means to absolute truth. Often  "doing more science" does not reduce or eliminate uncertainty, particularly in dealing with future  assessments. Science is not purely objective, but is subject to the values of  society and the research community; therefore,  science should not be the exclusive leader of  knowledge. Academic study does not foster  the ability to create a sustainable relationship with our natural world because of  the emphasis on "theories, not values, abstraction rather than consciousness, neat answers instead of  questions, and technical efficiency  over conscience" (Orr, 1991: 99). These views all suggest a need for  PPs to address the gaps left  by science and to ensure that appropriate value judgments are included in IAs (ICIS 1999). A related point is that holism, or systems thinking, is replacing the philosophy of reductionism. Holism encourages analysis of  the system as a functioning  whole. Connectedness, relationships, and context are the tools for  understanding whole systems. In contrast, according to the reductionist or Cartesian view systems are comprehensible only by analyzing components in isolation. In addition, this philosophy considers the world a machine, void of  spirituality, emotions or values (Capra 1996). These "soft"  elements have / direct influence  on how humans interact with their environment, so they need to be taken into account when investigating ways to solve environmental problems. It is not a coincidence that these paradigm shifts  are taking place during our age of globalization marked by growth and development that has increased demand for  resources beyond sustainable limits. Local problems are no longer isolated, but interact across sectors as well as political and geographical boundaries. These have all contributed to the need to evaluate environmental problems from  a systemic point of  view, such as using participatory IA methods. The methods of  LA described in the literature are typically catagorized as either analytical approaches or participatory approaches (ICIS 1999; Rotmans and van Asselt 2002). Analytical approaches include modeling, scenarios, decision support systems, environmental impact assessment and some risk analysis (Rotmans 1998; Toth and Hizsnyik 1998). Until the mid 1990's, IA modeling was the dominant method (Rotmans and van Asselt 1996; Risbey et al. 1996). However, participatory approaches now play a role in most LA activities. These approaches are not exclusive, but may be used in combination in various stages of  the assessment. IA shares commonalities with Technology Assessment, Risk Analysis, and Policy Analysis. "These research areas also address some kind of  complex problem, however, from  a specific point of  view. The essential difference  is that IA aims to integrate knowledge from  an a priori integrated point of  view" (Rotmans 1998:156). Decision support tools are also related to IA in their cross-disciplinary nature. However, to date, these have most often  been developed by experts. Participatory modeling which is a fairly  recent approach, will be discussed later in this chapter, as an example of  a participatory method. 1.2.1 Benefits, challenges and future directions The insights gained from  IA can inform  policy, the public, and disciplinary sciences, resulting in improved decision-making. Rotmans and van Asselt (1996:334-5; also ICIS, 1999:9) list the ways in which IA is able to do this: • IA places the problem in a broader context. • IA can assist with trend analysis, eliminating improbable scenarios and evaluating trade-offs  of  impacts. • IA can assess alternative actions/response options to the problem. • IA provides a framework  to structure scientific  knowledge, and to compare and rank uncertainties. • IA helps translate uncertainties into risk • IA can help to identify  research gaps and prioritize areas of  future research. Participatory processes can contribute to, and improve, all of  these tasks listed. There are several challenges to conducting effective  IA processes. ICIS (1999: 22) provides recommendations for  ensuring the longevity and increasing the effectiveness  of  LA processes: • Increase transparency of  process through better communication and documentation. • Expand capabilities of  analysis through the development of  new research methods. • Involve "end-users" in the process from  the start to ensure a good match between the information  needed and the information  supplied. • Establish quality criteria to test the quality of  LA research, and a "codes of  practice." The last item listed was raised repeatedly in the literature over the past decade. (Dowlaitabadi 1995; Rotmans and van Asselt 2002; Rotmans 1998; van Asselt and Rijkens-Klomp 2002). As the field  is relatively new, there are no guidelines or standards established to date; however, approaches to establishing guidelines have been proposed. The establishment of guidelines would increase the credibility of  IA in both the scientific  and political communities. It would also assist in maintaining a productive relationship with disciplinary science. Credibility is necessary if  IA is to remain an accepted approach to evaluating environmental problems (Rotmans and van Asselt 1996). A set of  quality criteria is proposed in Rotmans and van Asselt (2002:13-15). These criteria are divided into three classes: analytical, methodological, and usability. These first  two help evaluate the internal aspects of  the assessment, and aim to address the question, "Have we done a good job?" Usability, on the other hand, is an external quality, related to the question, "Is the study useful?"  Resource managers and policy makers are often  the "end-users" of  an assessment. Involving these parties in the design of  the process improves the likelihood that the product will be useful  to - and used by them. i IA also has several methodological challenges associated with it. First, there is no optimal spatial and temporal scale with which to conduct the assessment, as complex problems include aspects that operate on different  scales. This leads to a constant struggle between aggregation and disaggregation. Second, LAs contain considerable uncertainty of  different types. These uncertainties, from  both technical and methodological sources, must be managed appropriately. Third, appropriate techniques must be used or developed to blend the qualitative and the quantitative types of  knowledge. Most IA frameworks  treat these two _as mutually exclusive but there are a number of  techniques available to support blending, such as fuzzy  logic (Rotmans 1998:164-6). 1.3 Participatory Planning Methods, Including Participatory Modeling Tell me and I'll forget, Show me and I may remember, Involve me and I will understand. -Author unknown There are several reasons why stakeholders (which include professionals,  interested parties, and the public) should be involved throughout the planning process, rather than just as consumers of  research results. The above verse describes one of  these reasons. Active, personal engagement fosters  learning and changes worldviews better than passive learning activities. Involving stakeholders in model development also supports customization of  the model to their needs as model users. Each stakeholder contributes a different  perspective on the system. The majority of  information  available resides in "mental data bases" and has not been written or measured (Forrester 1987), so direct communication is often  the only way to access this information.  Stakeholder involvement also helps to build confidence  in the model. Throughout the development process, assumptions and uncertainties are communicated. As a result, the users have the information  to interpret the results appropriately. The remainder of  this section describes some of  the distinguishing characteristics of \ participatory processes and provides detail on participatory modeling. 1.3.1 Degree of participation The participatory planning literature includes a wide array of  methods and tools with varying degrees of  involvement of  the participants, from  being receivers of  information  to having decision-making power. Van de Kerkhof  (2003:26) provides a detailed ladder of  participation as shown in Figure 1.1. There are separate scales for  policy making and for  scientific practice. Figure 1.1: Degrees of stakeholder participation in policy making and in scientific practice (from van de Kerkhof,  2003:26).* Stakeholder participation in policy making Stakeholder participation in scientific practice Stakeholder control Delegated power Partnership High degree of participation Mutual learning Co-production Co-ordination Placation Consultation Moderate degree of participation Mediation Anticipation Consultation Information Therapy Manipulation Low degree of participation Information * Concept based on previous work of: Arnstein, S.R. 1969. A ladder of  citizen participation. In: J. of  the American Institute of  Planners 35(4). 216-224. Mayer, I. 1997. Debating technologies. A methodological contribution to the design and evaluation of participatory policy analysis. Tilburg University Press: Tilburg, The Netherlands. Methods that contain low levels of  participation serve to educate or inform,  generally targeting a specific  interest group or the general public. As information  travels uni-directionally from  experts to participants, this level does not support IA. Examples include brochures, mailings, press releases, field  trips, information  lines, and briefings  (Mostert 2003). Methods which collect information  from  participants, ranging from  consultation through mediation, are classified  as moderate participation. Examples include public meetings (which / request feedback  and comments), interviews, opinion polls, teaching and training games/gaming simulations, focus  groups, dialectical debates, scientist-stakeholder workshops, and decision seminars (Mostert 2003; van Asselt and Rijkens-Klomp 2002; van de Kerkhof  2003). In high levels of  participation, some of  the responsibility of  the project is shared. Participants may contribute to the design of  the process or to making final  decisions. The process works toward coordination, co-production, or mutual learning. Examples include policy exercises, scenario analysis, negotiations, participatory modeling, policy Delphi, rational discourse, and cooperative discourse. (Mostert 2003; Renn 2003; van Asselt and Rijkens-Klomp 2002; van de Kerkhof  2003). Some methods can be classified  as either moderate or high, depending on how the process is conducted and in which stages the participants are involved. These include planning cells, citizen juries, consensus conferences,  working groups, and participatory decision analysis (Mostert 2003; van Asselt and Rijkens-Klomp 2002; van de Kerkhof  2001; 2003). Moderate degrees of  participation foster  first-order  learning, while high degrees of participation support second-order learning. Van de Kerkhof  (2003) describes the differences between first-order  learning and second-order learning: In first-order  learning, the new insights that the participants generate mainly relate to the empirical and technical level of  analysis. This concerns insights in the 'facts'  and expectations on a specific  topic Second-order learning is achieved when the participants gain new insights in the complex relationship between causal and normative reasoning, which may result in a change in participants' norms and core beliefs  that guide their behaviour and that underlie their conception of  the very nature of  the problem concerned (a paradigm shift)  (van de Kerkhof  2003:60). A high level of  involvement is required to foster  second-order learning, while a certain degree of  distance between the participants and the problem is helpful  to limit the discussion to first-order  issues. Second-order learning is developed through authentic conflict  and argumentation, rather than role-playing or artificial  simulation and conflict.  Methods such as group brainstorming, gaming simulation, and role-playing result in first-order  learning, while dialectical debate and policy Delphi can create second-order learning (van de Kerkhof 2003:61). 1.3.2 Problem type There are several types of  environmental problems and planning processes that are well-defined,  so preclude IA. Examples are those that are one-time decisions, such as facility siting decisions (Stave 2003:304). PPs that are most relevant to these well-defined  problem types are charrettes, public hearings, and opinion polls. Van de Kerkhof  (2001) describes a framework^  that maps out problem types in two dimensions. One dimension describes the level of  certainty or uncertainty associated with the problem. The other dimension refers  to the level of  consensus or disparity between stakeholders on the relevant values at stake. The "structured" problem has high certainty and high consensus. IA is not necessary for  these problem types. When there is consensus, but high uncertainty, the problem is "moderately structured (ends)." When there is diversity of opinion but high certainty, the problem is "moderately structured (means)," and those problems that have both significant  diversity of  opinion (lack of  consensus) and high uncertainty are considered "unstructured" problems. Participatory IA is most useful  in * Framework developed by Hisschemoller (1993) and Hoppe (1989): Hisschemoller, M. 1993. De democratic van problemen. De relatie tussen de inhoud van beleidsproblemen en methoden van politieke besluitvorming (in Dutch). VU Uitgeverij: Amsterdam, The Netherlands. Hoppe, R and A. Peterse (eds.) 1998. Bouwstenen voor argumentatieve beleidsanalyse (in Dutch). Elsevier: The Hague, The Netherlands. unstructured and moderately structured problems. In well-structured problems, participation generally involves learning all of  the available information  and developing full  consensus. Citizen juries or planning cells, consensus conferences,  and participatory planning are examples that typically have consensus as the primary goal (van Asselt and Rijkens-Klomp 2002). The primary goal of  IA is typically mapping diversity rather than reaching consensus, so these methods are not always appropriate for  IA (Pahl-Wostl and Newig 2003). However, consensus can result from  confronting  and challenging the diverse views. Examples of methods that may lead to consensus through a process of  mapping and challenging the diversity of  views are participatory modeling, policy Delphi, and dialectical debate (van de ( Kerkhof  2003; Vennix 1996). 1.3.3 Profile of participants Methods are designed for  use with either the general public or with stakeholders who have a close connection with the issue (Pahl-Wostl and Newig 2003). Both groups of  participants may be helpful  in IA, depending on the purpose of  their involvement. Typically the methods that engage the general public are used for  opinion scoping (some focus  groups and some policy exercises) or for  reaching consensus on specific  issues (citizen juries, consensus conferences,  and decision seminars). Methods that engage participants at high levels or for longer durations engage those with concerns about the issue. Since higher levels of involvement require more commitment, it is easiest to engage those with vested interests in the topic. Unfortunately,  this can easily create bias by attracting stakeholders that have a professional  role in the issue, or have the time to afford  through retirement or other means, while excluding the "average" citizen that uses the resource. Providing financial compensation for  their time, assigning representatives of  interest groups, and using additional mechanisms for  communicating with the stakeholder groups can help to engage stakeholders that would otherwise be excluded from  the process entirely. A related issue is the role of  scientists in the process. Depending on the problem type, scientists can take on different  roles. In structured problems, the scientist is the problem solver. In unstructured problems the scientists may be responsible for  identifying  and clarifying  the issue, while in moderately structured problems the scientist may take the role of  either an advocate or a mediator (van de Kerkhof  2001:12). 1.3.4 Goal of participation There are two philosophies to conducting PPs. The PP can either be seen as a means toward meeting other goals, or it can be viewed as a goal in itself  (van Asselt and Rijkens-Klomp 2002). Participatory IA more often  focuses  on the outcome of  the process rather than on the process itself  since the participatory process contributes to generating the assessment, but may focus  on both aspects. When participation is viewed as a means to other goals, goals of participation may be: (1) to improve decision-making, (2) to improve scientific  practice, or (3) to structure complex problems (Forrester 1999; van de Kerkhof  2003). When the PP is seen as a goal in itself,  then conducting the process well is most important. A number of evaluation criteria are relevant: Was the process fair  and transparent? How were power imbalances addressed? Did the decision-making process incorporate participant responses, or give power to participants? Did the process foster  social learning between and among the scientists and the stakeholders? (Folz and Hill 2001; van de Kerkhof  2003). Did the process accommodate the role of  local or traditional knowledge? Learning is an important objective for  PPs not only as an objective in itself,  but also as a means to improving decision-making, scientific  practice and problem structuring. 1.3.5 Challenges in participatory planning There are a number of  challenges to conducting PPs. First, participants are often  volunteers, so attendance may be inconsistent and the group composition may change. Second, the political and public context forces  a short time period for  the work. This frequently  creates a "good enough" attitude for  the product, limiting ability to develop details. Third, participants may not be willing to change their mental models. Participants may have personal objectives that reduce their willingness to challenge their beliefs.  Finally, "group think" and defensive routines may reduce the quality of  the process (Stave 2002:160-1). 1.3.6 Participatory modeling Several fields  in the literature use different  terms to describe processes in which stakeholders are involved in the development of  a model for  purposes ranging from  shared learning to consensus building. IA literature refers  to "participatory modeling," in fields  such as policy analysis and organizational learning, as well as natural resource applications such as water resources and land management. Videira Costa (2005) provides insight on the differences  of two related terms. "Group model building," from  the system dynamics community, has most often  been applied to organizational messy problems (Richardson and Andersen 1995; Vennix 1996; 1999) but has also been applied to sustainability issues (Stave 2002; 2003). The term "mediated modeling" originated from  the ecological economics community in the late 1990's and has been applied to solving complex environmental problems (van den Belt et al. 1998). A term that has been applied exclusively to water resources management applications is "shared vision planning" (SVP). The methodology was formalized  about fifteen  years ago by the Institute for  Water Resources, U.S. Army Corps of  Engineers. SVP combines traditional water resources planning principles, structured public participation, and integrated computer modeling (Palmer et al., in review). These approaches are appropriate for  complex problems, particularly ones where conflict  is anticipated. The process can meet one or more of  the following  goals: foster  team learning, share information  between stakeholders, foster  future  vision, develop consensus on the behaviour of  the system, and reach consensus on a decision and create commitment to that decision (van den Belt 2004:41-45). Participatory modeling is founded  in the belief  that people's mental models of  how a system behaves are based on numerous unstated assumptions, and often  contain gaps and inconsistencies. Fuzzy, incomplete, and imprecise mental models can lead to ineffective policy making. Unfortunately,  today most policy decision-making processes are not explicit. Even in the modern age of  science and industrialization social policy decisions are based on incompletely-communicated mental models. The assumptions and reasoning behind a decision are not really examinable, even to the decider. The logic, if  there is any, leading to a social policy is unclear to most people affected  by the policy. (Meadows and Robinson 1985:3) Technological tools, including GIS, have proven to be effective  for  learning and deliberation. The active engagement process increases information  retention, inspires creativity, and improves analysis skills (Cloud 2001). The process of  sharing mental models identifies points of  agreement and points of  conflict.  The areas of  conflict  draw attention to the underlying assumptions, and the modeling activity is a tool to make assumptions explicit so they can be clarified  and challenged. The process ideally results in a model that describes the structural aspects of  the system, while the model simulations provide information  about system behaviour (Forrester 1987; Vennix 1996). The model can then be used to explore a range of  future  conditions or assumptions. Participants may engage directly in the modeling process, or the model may be developed in an iterative process with regular opportunities to contribute (van Asselt and Rijkens-Klomp 2002: 172). Participatory modeling includes a range of  levels and timing of  stakeholder participation. The resulting model is more a by-product of  team learning than a tool in itself  (van den Belt 2004; Vennix 1996). Instead of  focusing  on the completed model as the primary goal, objectives of  the process are to "enhance team learning, foster  consensus, and create commitment with the outcomes" (Vennix, 1996:101). Consensus does not need to be actively encouraged; if  the process is performed  successfully,  consensus is a natural outcome. Participatory modeling may occur over two days, or it may iterate over multiple years, depending on the context, available resources, and desired level of  detail. This method is applicable to various stages of  IA, from  problem definition  through implementation and adaptation of  strategies. 1.3.7 Research directions in participatory planning The field  of  participatory planning is still developing and there are areas that need attention. An investigation into the applicability, strengths, and weaknesses would provide guidance on the most appropriate and effective  applications of  the array of  tools and methods available. The literature describes several areas for  further  growth and development: the timing of participation, a new emphasis on building trust among the parties involved, and some reflection  on which projects will actually benefit  from  PPs. Early involvement of  participants helps create a product that is useful  to the end-users of  the information.  Involving participants in the development of  project proposals helps to identify areas of  research that are relevant to the interests of  the community. Involving participants as early as possible not only serves to empower stakeholders and increase their interest and commitment, but also fosters  relationships between the project leaders (researchers, scientists) and the community of  stakeholders. According to William Werick (2004), who worked on a number of  participatory projects during his career at the US Army Corps of Engineers, trust between project leaders and participants is the key to a successful  process. Therefore,  early engagement and patience must be part of  the design. Finally, consideration must be given to the appropriateness of  participation in the current spectrum of  environmental projects. The current philosophy is that all projects should engage the public or selected stakeholders in some way. Thus, the pendulum has swung from projects being exclusively in the hands of  experts, to stakeholders participating in all processes. It is possible that the level of  involvement of  stakeholders should be tailored to the character of  the problem. This would be a topic worthy of  debate among academics, policy-makers, and the public. 1.4 Systems Thinking and System Dynamics System dynamics is a useful  tool for  determining sustainable pathways, because it allows practitioners to examine the whole system as well as detailed components in both the short and the long-term. There are currently weaknesses in our decision-making processes, resulting in ineffective  or counterproductive policies. System dynamics can support more consistent, structured processes that will lead to better management of  resources, and pave the way to a sustainable society. System dynamics models "help to clarify  our processes of thought[,]... help to make explicit the assumptions we are already making[, and]... show the consequences of  the assumptions." Models are not static, but must change as our goals change (Forrester, 1985:133). Systems thinking originates from  the fields  of  process thinking, tektology and general systems theory developed in the early 1900's (Capra 1996). Jay Forrester is considered the father  of  system dynamics, as he developed the theory in the 1950's, drawing on principles of control theory (a.k.a. information  feedback  and servomechanism) and decision theory (Vogstad 2005). Early applications of  system dynamics that have defined  the field  include the study of  urban policy and growth (Forrester 1969); global resource dynamics (Meadows et al, 1972; 2004); and human behaviour in organizations (Senge 1990). More recently, system dynamics has become a tool for  stakeholder participation in environmental decision-making. In a case study in which model was created for  water resources policy analysis in Egypt, the authors concluded that the model's "transparency allows for  more participation and empowerment of  the public and all the stakeholders. The suggested framework  provides a feedback  mechanism that facilitates  reaching solutions that enable the society to move away from  the status quo according to its present priorities and objectives" (Simonovic and Fahmy 1999:303). Systems thinking uses mental models, while system dynamics constructs formal  models. Mental models have the advantages of  being more flexible,  rich in detail, and constructed from  the largest source of  information:  collected experience. However, mental models are often  fuzzy,  incomplete, imprecise, and filled  with unstated assumptions and goals. Furthermore, humans, with all their talents, are poor dynamics simulators. The process of creating formal  models is a technique for  overcoming these weaknesses. Contrasting behaviour predictions with a structured, internally consistent simulation provides the surprise discoveries that promote learning (Radzicki 1997). System dynamics practitioners generally do not emphasize the final  model as a finished product; instead, they propose that the modeling process is critical because that is where the learning occurs (Forrester 1985). In fact,  the "most brilliant model and insights have no impact if  they are not embedded in an effective  learning process" (Jones et al. 2002:202). Model output is much more meaningful  when the user knows the internal relationships that drive the behaviour. 1.4.1 Characteristics of complex systems: rates, levels, and feedback loops The system dynamics paradigm assumes that things are interconnected in complex patterns, that the world is made up of  rates, levels and feedback loops, that information  flows  are intrinsically different  from  physical flows, that nonlinearities and delays are important elements in systems, that behaviour arises out of  system structure (Meadows 1989:70). Complex systems contain dominant non-linear interactions (which can rarely be solved analytically), feedback  loops, and time and space lags (which muddle cause-effect  linkages), as well as discontinuities, thresholds and limits. Two examples of  complex systems are ecological systems and economic systems. Taken together, "linked ecological-economic systems are devilishly complex" (Costanza 1996:981). Systems thinkers are concerned with underlying structure; however, it is rarely apparent. Karash (1999) provides a framework  for  developing understanding of  system structure, through stimulating thinking about the differences  between events, patterns, and structure. An event is the answer to: "What happened?" Patterns are identified  by considering: "Where are the changes? The contrasts? The continuities?" To identify  the underlying structure, the important questions are: "Why and how? What would explain the patterns?" Karash illustrates his point with an example of  being stuck in traffic.  The event is a car accident that is blocking the road. Patterns include the frequent  accidents on this road, daily and weekly traffic  patterns, and that drivers are more aggressive at these times. Insights into the structure include: (1) The old, narrow road's poor sight lines increase accidents when traffic  is heavy. (2) Most traffic  to major north and south routes is funnelled  into this corridor, increasing traffic  load. (3) Drivers assume there is no room for  police to monitor so they speed, which causes more accidents (Karash 1996). 1.4.2 Complex system behaviour One reason that systems thinkers are so concerned with system structure is because of  the premise that the internal connections within the system have much more control over the system than external forces  (Forrester 1987; Meadows 1989; Stave 2003; Sterman 2000). For example, financial  analysts look at global politics and external forces  to make predictions about the domestic stock market. If  systems thinkers were hired on Wall Street, they would spend more time investigating the internal workings of  the market than on "external forces." When Forrester began studying systems, he discovered that many policies were ineffective, or worse - counterproductive. The internal structure of  a system can have strong control over that system's behaviour, particularly when there are negative, or dampening, feedback  loops that dissipate the effects  of  external forces.  Therefore,  systems are often  highly resistant to policy changes (Forrester 1987). Luckily, there are still leverage points for  influencing  a system. Forrester often  said, "People know intuitively where leverage points are. Time after  time I've done an analysis of  a company, and I've figured  out a leverage point. Then, I've gone to the company and discovered that everyone is pushing it in the wrong direction!" (Meadows 1997:78). An example of  this is when the Club of  Rome asked Forrester to show relationships and solutions for  the world's biggest problems - poverty and hunger, environmental destruction, resource depletion, urban deterioration, and unemployment. Forrester's response was that they were all related through one clear leverage point: Growth. The world's problems are the costs of  growth. The world's leaders are correctly fixed  on growth as the solution to the world's problems. However, they are pushing in the wrong direction! (Meadows 1997) Meadows describes nine places to intervene in a system. These points may vary for  each application, but in general, they are, in order of  increasing influence:  (9) Numbers (subsidies, taxes, standards); (8) Material stocks and flows;  (7) Regulating negative feedback  loops; (6) Driving positive feedback  loops; (5) Information  flows;  (4) The rules of  the system (incentives, punishments, constraints); (3) The power of  self-organization  (evolution, technical advance, social revolution); (2) The goals of  the entire system; (1) The mindset or paradigm out of  which the goals, rules, and feedback  structure arise. 1.4.3 General model theory Costanza provides a way of  characterizing models as having high generality (conceptual breadth), high quantitative precision (analytical), and high qualitative realism (impact-analysis). Different  models emphasize or de-emphasize these dimensions as appropriate for the context and purpose of  the model as well as the available information.  Due to real-world constraints it is rarely possible, or useful,  to create a model that maximizes all three areas (Costanza 1996). For example, statistical approaches to flood  frequency  analysis are highly analytical, but very low in breadth and realism. The Rational method'1' for  determining peak discharge in a watershed is also analytical and low in generality, but contains more realism than the statistical model. The system dynamics approach may be customized for  any type of model; however members of  the system dynamics community typically create elegant, highly aggregated models that emphasize the aspects of  realism (capturing system structure), as well as generality (including transdisciplinary information).  One example is the interactive, educational Snake River Explorer model. The high level of  aggregation reduced the model run time to only a few  seconds. The integration of  ground water and surface  water flows  into one model eliminated the need to transfer  output and input between separate models. These characteristics were important to keeping the model user-friendly  (Ford 1996). The Patagonia coastal zone management model provides another example. The "objective was not to describe the system and all trophic relations in great detail, but rather to "scope-out" important linkages between different  parts of  the system" (van den Belt et al. 1998:80). In contrast, highly analytical mathematical models are often  used for  predictive purposes in narrowly defined  situations (such as facility  siting) and are inflexible  and rigid (Cleveland 2001; Ford 1999; Stave 2003). Meadows and Robinson (1985) compares and contrasts system dynamics with econometrics, input-output, and optimization approaches. For * The Rational method is described by Q = CiA,  where Q is the peak discharge in the watershed; C is the runoff coefficient  based on land characteristics; i is the rainfall  intensity; and A is the drainage area. applications of  quantitative environmental analysis, Vogstad (2005) discusses the assumptions contained in cost-benefit  analysis, input-output and life-cycle  accounting, environmental risk assessment, material flow  analysis, and exergy analysis. Bagheri (2005) contrasts different  characteristics of  water resources models: deterministic vs. non-deterministic, lumped vs. distributed, steady vs. dynamic, and simulation vs. optimization. System dynamics models are typically non-deterministic, dynamic, and simulation-based. 1.4.4 Use of system dynamics models Primary uses of  system dynamics models in environmental applications are either: (1) to increase understanding of  the current behaviour of  the system (descriptive); and (2) to gain insight into the potential future  behaviour of  the system (predictive). The models are not predictive in the sense of  forecasting  or point prediction, but provide general understanding and guide future  research efforts  (Ford 1999). A descriptive model helps answer questions such as: "What is creating this pattern of  behaviour? Why did the policy have no effect?"  A model for  exploring plausible futures  can provide some answers to: "What will happen and what can we do about it?" A common objective for  predictive models is to determine what future  we really want and then to identify  the leverage points we can use to get us there (Hughes 1999). Guiding questions that recur throughout the system dynamics literature are: (1) "Does the system have the potential for  overshoot?" (2) "If  so, does it matter? Is overshoot something worth avoiding?" (3) "If  so, is there anything we can do about it?" (Jones et al. 2002; Hughes 1999; van den Belt 2004). Exploring system behaviour can be an interesting academic exercise, but to have an impact on the "real world" the insights need to be communicated to the systems' representative stakeholders, including resource managers and users. Conventional practice has been that "experts" develop models and communicate primarily the results to decision makers, and has had varying degrees of  effectiveness.  A newer, alternate approach is to engage policy-makers directly in the modeling process. Models created in computer code are not well-suited to engaging those who may not be technically proficient.  Participatory modeling requires models with new qualities such as transparency, flexibility,  and speed. For example, Nvule (1993) described his experience with a new system dynamics model and compared it with what they used before.  The previous model in Fortran was "inflexible  and difficult  to use by people not associated with its development. The process of  building consensus among interested parties was a drawn out affair  due to the time needed to assimilate the Fortran model results" (Nvule 1993: 492). The system dynamics approach provides benefits  over many analytical approaches. These models can: (1) Evaluate and compare policies; (2) Provide immediate feedback  to participants on their ideas; (3) Display output in graphs enabling easy run comparison; (4) Engage interest by showing unexpected results; and (5) Help participants understand, stimulate discussion, and build consensus (Stave, 2003). The structure of  a model provides "a consistent and rigorous problem-solving framework" (Stave 2002:143) that can help overcome "some of  the problems inherent in linear thinking and compartmentalized, non-participatory decision making" (van den Belt 2004:11). This provides a neutral framework  for  discussion in which alternatives can be discussed more objectively (Stave 2002). Furthermore, "system dynamics modeling can be effective  because it builds on the reliable part of  our understanding of  systems while compensating for  the > unreliable part" (Forrester 1987:137). Forrester explains that the observed structures and policies are usually where everyone can agree. Predicting the behaviour of  these structures and policies is where mental models fall  short, so is commonly the area of  disagreement. Making the inputs of  the model the parts that are known - the structures and policies, while letting the model simulate and provide the system behaviour, allows more objective and effective  debating to occur (Forrester, 1987; Stave 2002). An effective  model includes only the aspects that have a significant  role in the behaviour of interest. The editing ease of  system dynamics models provides a tool for  distinguishing the critical features  of  a system that define  its behaviour. The modeling exercise may recognize aspects which were previously disregarded as unimportant (Faust et al, 2004; Jones and Seville 2002; Stave 2002). Another technique for  reducing the amount of  unnecessary information  in a model is to aggregate temporal and spatial scales. Aggregation does mask detail and reduce the amount of  information  in the model but it also helps to clarify  model output. Aggregation can preserve general trends and provide quick answers to questions commonly asked by policy makers (Simonovic and Fahmy 1999). It can also provide a broader context for  the problem, changing the problem focus  from  decision makers' specific concerns to the more strategic level (Stave 2002; Stave 2003). 1.4.5 Challenges , There are challenges for  using system dynamics for  learning and decision support processes with stakeholders or the public. Sterman (2000:35) observes that, although "simulation models and virtual worlds may be necessary for  effective  learning in dynamically complex systems, they are not sufficient  to overcome the flaws  in our mental models, scientific reasoning skills, and group processes." A possible risk when people, non-technical users in particular, use a computer model is "video game syndrome." This occurs when users thoughtlessly press buttons, trying as many different  options as they can in the allotted time. Ideally, model users should start by selecting options carefully,  then predicting the results, and only then running the simulation (Sterman 2000). Learning occurs when the modelled result is different  than the expected result, so this middle step is a critical one. All models will be "black boxes" to those who are unfamiliar  with the underlying structure and assumptions used to build the model. System dynamics models, particularly STELLA™ and related software  have graphical model interfaces  that assist both technical and non-technical users in learning the model, with the intention of  avoiding this problem. Regardless, in order to be accessible, models need to be kept as simple as possible, with a layout that is logical and easy to follow. 1.5 Managing Uncertainties for Planning A decade ago, Shackley and Wynne (1996) noted: It is frequently  assumed that scientific  uncertainty is a problem for environmental policy. Many decision makers and advisory scientists believe that policy ideally should rest on reliable, robust, and hence certain scientific knowledge, (p. 275) Today, with the development of  complex systems science as well as the focus  on long-term sustainability, it is clear that additional science cannot always provide all of  the answers. Shepherd et al. (2006) found  that uncertainty was not an obstacle in implementing new policies when political will and enabling factors  were present. However, in other cases, the presence of  any uncertainty has been - and continues to be - an excuse for  delaying action. One cause of  this is the belief  that scientific  certainty is a prerequisite for  building consensus (Shackley and Wynne 1996), and negotiations are delayed until more information  is available. The climate change issue provides a prime example of  this effect.  To date, much of the focus  has been on how to reduce the level of  uncertainty in climate change predictions. Reduction of  some of  the uncertainty is reasonable, but uncertainties that are inherent to the system cannot be eliminated. Once inherent uncertainties dominate, then the focus  should shift  away from  reducing uncertainties and move on to clarifying  and communicating what is known about the system and determining effective  and robust responses. 1.5.1 Inherent uncertainty vs. knowledge uncertainty Diefenderfer  et al. (2005) distinguish two sources of  uncertainty: knowledge  uncertainty  and inherent uncertainty.  Knowledge uncertainty, also referred  to as epistemic uncertainty, is due to incomplete knowledge about the system. In modeling, knowledge uncertainties can be gaps in the model's structure or in the data required to support it. If  a system contained only knowledge uncertainty, then, in theory, complete knowledge about the system could be achieved through further  scientific  investigation. In contrast, inherent uncertainty is a result of  natural variability in processes such as non-linear and chaotic behaviour patterns; so when present, no amount of  research will generate absolute predictions. The belief  that scientific knowledge that supports policy must be certain is a result of  falsely  assuming that all of  the uncertainty present is a type of  knowledge uncertainty. 1.5.2 Climate change and modeling Moser and Dilling (2004) describe the uncertain character of  climate change as one contributing reason why the professional  water community, to date, has been reluctant to consider climate change in their planning activities. This is unfortunately  not an uncommon response when modeling a system - to omit the elements for  which we have limited understanding. For example, when the first  report by the Intergovernmental Panel on Climate Change was generated in 1990, little quantifiable  information  was available on the natural feedbacks  related to a warmer climate. Therefore,  the authors omitted the information  from the computer models (Leggett 2001). Unfortunately,  when analyses lack elements that help define  the behaviour of  the system, policies developed from  these analyses may not be ideal for  the conditions that will be realized. 1.6 Case Studies in Participatory Modeling This section provides an overview of  case studies that used participatory modeling approaches, particularly in natural resources and water planning applications. Most of  the case studies reviewed did not discuss inclusion of  climate change (with the exceptions of Pataki et al. 2006; Ivey et al. 2004; and Kirshen et al. 2004). The majority of  participatory modeling processes to date used system dynamics models (STELLA, Powersim, Vensim) either as the primary tool, or in combination with other IA models and participatory methods (Cardwell et al. in preparation). One reason for  this may be the transparency of  these models, which allows active participation and empowerment of those involved in model construction and use (Simonovic and Fahmy 1999). A common alternative to system dynamics are generic simulation models tailored to water related applications, which also support active engagement of  stakeholders by their interactive graphical interfaces  (Loucks 2006). Examples include WEAP which was used for collaborative planning in the River Njoro Watershed in Kenya (Jenkins et al. 2005), and MIKE BASIN, applied to the San Francisco public water supply (Borden et al. 2006) and to management of  the Lehmi River Basin in Idaho (Borden and Spinazola 2006). Participatory modeling by definition  involves stakeholders in the development process. However, cases that used completed system dynamics models for  education can provide insight into facets  of  stakeholder-model interaction. Two of  these case studies noted the model users' backgrounds and prior knowledge significantly  influenced  their experience. A workshop to test the Snake River Explorer developed by Ford (1996) included faculty, students, and representatives of  agriculture and electric power industry, environmental groups, native tribes, state departments and federal  bureaus. Those who had extensive modeling backgrounds voiced concern about using such a highly aggregated model to educate the public. However, the remaining participants found  that the model's simplicity enhanced their understanding. Stave (2003) developed a high-level model of  the Las Vegas water resource system to educate the public and gain support for  management policies. Although the insights were obvious to water managers, the tool effectively  gave the public an appreciation for  the influence  of  return flows  and the constraints on managing the resource. Forster and Journeay (2004) used a two-stage hybrid of  the participatory modeling method to investigate the sustainability of  water resources in the Gulf  and San Juan island communities off  the Pacific  coast of  British Columbia and Washington State. First, the researchers developed a generic model for  an island system, characterizing groundwater and household dynamics. Then, they engaged island communities in turn through meetings with small groups of  island residents to customize the model to each island and to discuss policy implications. Letcher and Jakeman (2003) conducted a participatory IA modeling process for  water allocation issues of  the Namoi River Catchment in Australia. The authors concluded that the process takes a considerable investment of  resources and therefore  commitment by the participants. Because of  this, the approach is most suitable for  complex problems in which the benefits  will outweigh the costs. Furthermore, they observed that advantages to both parties achieved by the learning process can be realized whether or not the completed model is a success. A collaborative modeling process for  the Middle Rio Grande in New Mexico educated the public about the complexities of  managing the resource by identifying  tradeoffs,  lags and feedbacks.  One concern was the lack of  involvement on the part of  policy makers and their reaction to the results; however, several insights from  the process were still incorporated into the regional plan (Cockerill et al. 2006; Tidwell et al. 2004). Hare et al. (2003) compared four  natural resource management case studies in Switzerland, Zimbabwe, Senegal, and Thailand to determine factors  that influence  process design. They identified:  project goals, democratic participatory goals, researchers' normative beliefs, existing management power structures, stakeholder numbers, and the scale at which the final decisions need to be supported. These factors  can also influence  project outcome. Jones et al. (2002) developed a model of  the forest  industry in the northern-eastern United States with support from  an advisory board of  local stakeholders. The advisory board was supportive, but the modeling process did not inspire commitment because they did not share the same objectives of  the project. Eventually the authors partnered with state forestry  officials  who shared the goal of  sustainable forest  management. A contrasting example is a mediated modeling initiative for  the San Antonio Watershed in Texas (Peterson et al. 2004) which generated remarkable commitment by the local community. Participants requested to assist with planning meeting agendas, enforced  established protocol among disruptive newcomers, and maintained the council after  the facilitators'  year-long commitment ended. Participatory modeling case studies vary considerably in their level of  breadth and in the issues they characterized. Pataki et al. (2006) and Durfee  et al. (2004) describe an application to determine the dominating influences  on air quality in the Salt Lake Valley, Utah. Sumer and Lansey (2004) assessed both surface  and groundwater sources in the Upper San Pedro Basin in Arizona. Costanza and Ruth (1998) developed ecologic-economics models of  the Louisiana coastal wetlands and of  the Patuxant River watershed in Maryland. Van den Belt et al. (1998) developed a model for  the Patagonia coastal zone, which included sectors for ) ] fisheries,  oil pollution from  tanker spills, penguins, and tourism. The model of  the San Antonio watershed by Peterson et al. (2004) concentrated on social and environmental issues. A specific  application of  collaborative modeling has been named "Shared Vision Planning" (SVP). Palmer et al. (in review) describe SVP as a combination of  traditional water resources planning, structured public participation, and the use of  an integrated computer model. Early applications of  SVP used watershed simulation models to train water managers in "Drought Preparedness" workshops (Keyes and Palmer 1993; Keyes et al. 1995; Palmer et al. 1993). Two recent applications of  SVP include the Rappahannock Basin in Virginia (Conner et al. 2004) and a project in the Mississippi headwaters that combined simulation with optimization Cardwell et al. (2004). Participatory modeling case studies that explicitly addressed the issue of  climate change were much rarer. Selected models developed for  SVP case studies were used to assess the impacts of  climate change, but only after  the stakeholder process was completed (IWR 2003). Pataki et al. (2006) included the climate linkage, but it became a minor part of  the project. Another case by Ivey et al. (2004) evaluated the adaptive capacity of  a community in southern Ontario, Canada through interviews and data collected from  stakeholders, but did not include a formal  modeling process. One participatory modeling case study did focus  on climate change. Kirshen et al. (2004) actively engaged stakeholders throughout their multi-year process of  modeling and evaluating climate impacts on water, energy, transportation, and public health sectors in Boston, USA. 1.7 The Okanagan Project The Okanagan Basin in south central British Columbia is one of  the most arid regions in Canada, with annual average precipitation ranging from  less than 300 mm to 450 mm. The long, narrow basin extends 182 km from  the Canada-U.S. Border and covers an area of  8200 km (Figure 1.2). Presently, water resources in the area are under stress due to recent rapid population growth, intensification  of  irrigated agriculture and recreational activities, and extensive logging at higher elevations. Drought conditions in 2003 and 2004 resulted in water shortages and major fires.  These changes have already raised concerns about the reliability of  the Okanagan's water resources. In the future  these trends are likely to continue, and may be exacerbated by additional stresses caused by climate change. Figure 1.2: Okanagan Basin Map with inset for location in British Columbia (from Cohen et al. 2006). 1.7.1 Previous work Prior to commencing the work described here, substantial work had already been conducted in the region that allowed this work to happen. Specifically,  the early studies (1) generated quantitative scenarios for  water supply and demand under climate change; (2) developed positive, trusting relationships with the professional  water management community; and (3) increased awareness and concern among this Okanagan community about potential climate change impacts as well as adaptation opportunities (Cohen and Kulkarni 2001; Cohen et al. 2004). N UNITED STATES Figure 1.3: Flow chart illustrating the progression from climate models and local records to supply and demand inputs to the Okanagan Sustainable Water Resources Model.. (Sources: Cohen et al. 2004; Neilsen et al. 2001; Merritt and Alila 2006; Neale 2005; 2006) Global  Ch mate Models" HadCM3,  COCM2,  CSIROMk2; A2 & B2; 2010-2099 Okanagan climate  stations' The work described in this thesis relied heavily on climate scenarios developed for  both water supply and demand. Sources of  information,  from  the global climate models to the inputs used, are traced in the flow  chart shown as Figure 1.3. Taylor and Barton (2004) generated regional climate scenarios for  the Okanagan by correlating global climate model output to local climate stations using the "delta" method, in which the baseline data is perturbed by averages derived from  the generalized circulation models (GCMs) (Neilsen et al., 2006). These climate scenarios show mean temperature increases between 1.5 and 4 degrees Celsius throughout the year, with generally wetter winters and drier summers. The regional climate scenarios provided the means for  three modeling initiatives in hydrology, residential demand, and crop water demand. Merritt and Alila (2004; 2006) used the regional climate scenarios in the UBC Watershed stream flow  runoff  model to generate hydrologic scenarios. The results show significant  changes to the annual hydrograph from  the historic period (1961-90) to the period from  2010 to 2099. All scenarios show a reduced snowpack, an earlier onset of  the spring freshet  by four  to six weeks, and decreases in summer precipitation. Some scenarios also show more intense (peakier) spring freshets.  Neilsen et al. (2004) used the climate scenarios to model the impact on agricultural crop water demand. Higher temperatures increase both evapotranspiration and the length of  the growing season -two factors  which increase crop water demand. As a result, crop water demand scenarios estimate an increase from  12 to 61 percent, as climate change intensifies  through the decades. Furthermore, Neale (2005) correlated residential outdoor watering with temperature and detached dwellings for  several Okanagan communities, showing that water demand in the residential sector will also increase under climate change. Each of  these results on its own tells us important but limited information.  Only by integrating these stories can we determine the changes to the water resource system in the future. Relationships with stakeholders were developed during the five  years immediately preceding initiation of  this project. Focus groups in early 2001 reacted to single-scenario future hydrographs and generated an extensive list of  potential impacts (Cohen and Kulkarni 2001). In 2002, Shepherd investigated and analyzed several communities considered "early adopters to climate change" because they had implemented demand-side management strategies to monitor and/or reduce water use. She identified  the greatest challenges to implementing strategies as: attitudes, perceptions and values; financial  issues; and politics and policies (Shepherd et al. 2006). During a series of  workshops between fall  2003 and winter 2004, stakeholders investigated the feasibility  of  implementing adaptation strategies at both the local and the regional scales. Through this, we learned about the community's attitudes, perceptions, and policies related to adaptation options, such as the lack of  knowledge about and regulation of  groundwater, the belief  that water conserved by agriculture will support the increasing urban populations, and that water in the tributaries is perceived to be cleaner than the mainstem lake water (Tansey and Langsdale 2004). > Throughout the stakeholder engagement initiatives, as well as collaborations with related projects, researchers communicated their results on climate change to the Okanagan community. Attendance at the outset in 2001 was low, but steadily increased over time as the message spread that the Okanagan's water resources could be significantly  impacted by climate change. By the start of  the group model building work described here, results from the climate scenarios were starting to take a role in the development of  local planning initiatives (Summit 2004). 1.7.2 Research objectives The broad purpose of  both these earlier projects and the work described here is "to ensure that information  is available to Canadian decision and policy makers on: the environmental, social and economic impacts caused by vulnerabilities to atmospheric change, variability and extremes; and viable adaptive responses" (Environment Canada 2002). This project worked toward these goals by encouraging and enabling the region's water professionals  to incorporate climate change in their water resources planning initiatives by (a) translating the climate change information  into terms that are relevant for  the local water community, and (b) engaging them in an exploration of  the relevance of  climate change in the water resources futures.  As a result of  these initiatives, water resource professionals  are enabled to make better decisions regarding management of  their resource, including whether to implement anticipatory adaptation measures or to opt for  the reactionary approach. The research project described in this thesis consisted of  two related components: (1) Developing a system dynamics model to characterize water resource futures  in the Okanagan Basin, and (2) Conducting a participatory process that actively engages the region's water professionals  in the model's development. The participatory modeling process is described in Chapter 2, while a description of  the model and a summary of  results are provided in Chapter 3. ' The purpose of  developing the model was to generate new insight into plausible future conditions of  the Okanagan's water resources through assessing impacts in an integrated system and identifying  dominant feedback  loops, lags, and non-linear relations. Specifically, we explored: (a) The current state of  the system (through studying recent decades); (b) The effect  of  climate change and population growth on the balance of  future  water supply and demand; and (c) The role that adaptation measures might have in future  management. Although the model is valuable on its own, the main purpose of  the model, in this project, is to stimulate and enhance discussion about water management and climate change among members of  the Okanagan community that influence  water management. The purpose of  the participatory process was to create a shared learning experience between the study area's water resource-related stakeholders and the research team. Additional goals were to tailor the model to the needs of  the participants (the model users), and to foster ownership and trust in the model among the participants. Actively engaging stakeholders who are experts in the issues supported accurate characterization of  the system and fostered trust in the model. Participants identified  critical features  of  the system, provided data, and helped with calibration. The participatory process supported learning through hands-on interaction with the model and from  discussions with other stakeholders. Through this process, participants increased awareness of  potential climate change impacts within the system context and of  the complexities in managing the system. These benefits  are steps that support ensuring the reliability of  future  water resources through effective  policies and management decisions. 1.8 References ASCE (1998). Sustainability Criteria for  Water Resource Systems. Reston, American Society of  Civil Engineers, Task Committee on Sustainability Criteria and UNESCO IHP IV Project M-4.3. Working Group. Arnell, N. W. and E. K. Delaney (2006). "Adapting to Climate Change: Public Water Supply in England and Wales." Climatic Change 78: 227-255. Bagheri, A., et al. (2005). "A Systems Dynamics Approach to Analyze Water Resources Systems." XXXIAHR Congress, Seoul, Korea. Berkhout, F., J. Hertin, and D. M. Gann. (2004). Learning to adapt: Organisational adaptation to climate change impacts. Tyndall Centre Working Paper, No. 47. Sussex, University of  Sussex. Borden, C. and B. G. J. Spinazola (2006). "Evaluation of  Diversions Operation Plans to Meet Minimum Fish Flow Requirements Using MIKE BASIN Model Simulations in Lemhi River Basin." in Adaptive Management of  Water Resources: Proceedings of the American Water Resources Association's 2006 Summer Specialty Conference., Missoula, Montana, AWRA. Borden, C., et al. (2006). "Water Supply Assessment Tool for  the San Francisco Public Utility Commission." in Adaptive Management of  Water Resources: Proceedings of the American Water Resources Association's 2006 Summer Specialty Conference. Missoula, Montana, AWRA. Brundtland, G. H., ed. (1987). Our common future:  The World Commission on Environment and Development. Oxford,  Oxford  University Press. Capra, F. (1996). The Web of  Life.  New York, Anchor Books. Cardwell, H. et al. (in preparation). Literature Review of  Computer-Aided Collaborative Decision Making. Draft  Report, version December 2006. Institute for  Water Resources, U.S. Army Corps of  Engineers. Cardwell, H., B. Faber, and K. Spading. (2004). Collaborative Models for  Planning in the Mississippi Headwaters. Proceedings of  the 2004 World Water and Environmental Resources Congress, Salt Lake City, UT, American Society of  Civil Engineers. Cleveland, N. L. (2001). A System Dynamics Assessment of  Water Conservation Alternatives in the Wasatch Area, Utah. Salt Lake City, University of  Utah: 149. Cloud, S. W. (2001). Testing the potential of  system dynamics models for  improving public participation in resource management. Las Vegas, University of  Nevada, Las Vegas: 125. Cockerill, K., H. Passell, and V. Tidwell. (2006). "Cooperative Modeling: Building Bridges between Science and the Public." Journal of  the American Water Resources Association 42(2): 457-471. Cohen, S. and T. Kulkarni, eds. (2001). Water Management & Climate Change in the Okanagan Basin. Vancouver, Environment Canada & University of  British Columbia. Cohen, S., D. Neilsen, and R. Welbourn, eds. (2004). Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report, January 1, 2002-June 30, 2004. Vancouver, Environment Canada, Agriculture and Agri-Food Canada & University of  British Columbia. Connor, J., L. Cartwright, and K. Stephenson. (2004). Collaborative Water Supply Planning: A Shared Vision Approach for  the Rappahannock Basin in Virginia. Proceedings of the World Water and Environmental Resources Congress, Salt Lake City, UT. Costanza, R. (1996). "Ecological Economics: Reintegrating the Study of  Humans and Nature." Ecological Applications 6:4 (Nov 1996): 978-990. Costanza, R. and M. Ruth (1998). "Using Dynamic Modeling to Scope Environmental Problems and Build Consensus." Environmental Management 22(2): 183-195. Diefenderfer,  H. L., R. M. Thom, et al. 2005. A Framework for  Risk Analysis for  Ecological Restoration Projects in the U.S. Army Corps of  Engineers. Economics and Ecological Risk Assessment: Applications to Watershed Management. R. J. F. Bruins and M. T. Heberling. CRC Press, Boca Raton. de Loe, R. C. and R. D. Kreutzwiser (2000). "Climate Variability, Climate Change and Water Resource Management in the Great Lakes." Climatic Change 45: 163-179. Dowlatabadi, H. (1995). "Integrated assessment models of  climate change: An incomplete overview." Energy Policy 23(4/5): 289-296. Dowlatabadi, H. and G. Yohe (1999). "The who-how of  adaptation: towards more realistic impact assessments." Draft. Downing, T. E., et al. (2003). Climate Change and the Demand for  Water. Research Report. Oxford,  Stockholm Environment Institute. Durfee,  J. L., et al. (2004). "The Chicken or the Egg: Does Interdisciplinary Collaboration Enhance Systems Model-Building or Does Systems Model-Building Enhance Interdisciplinary Collaboration?" in Proceedings of  the 22nd International Conference of  the System Dynamics Society, Oxford,  England, UK. Environment Canada (2002). Web page: Adaptation and Impacts Research Group General Information.  Available at: <http://www.msc-smc.ec.gc.ca/airg/general_info_e.cfm >. Faust, L. J., et al. (2003). "Models for  management of  wildlife  populations: lessons from spectacled bears in zoos and grizzly bears in Yellowstone." System Dynamics Review 20(2): 163-178. Folz, H. and J. Hill (2001). "Public Participation Processes: Progressive or Not?" Vancouver, BC. Unpublished paper. Ford, A. (1996). "Testing the Snake River Explorer." System Dynamics Review 12(4): 305-329. Ford, A. (1999). Modeling the Environment: An Introduction to System Dynamics Modeling of  Environmental Systems. Washington, DC, Island Press. Forrester, J. (1999). "The logistics of  public participation in environmental assessment." Int. J. Environment and Pollution 11(3): 316-330. Forrester, J. W. (1985). ""The" model versus a modeling "process"." System Dynamics Review 1(1): 133-134. Forrester, J. W. (1987). "Lessons from  system dynamics modeling." System Dynamics Review 3(2): 136-149. Forster, C. B. and M. Journeay (2004). Exploring the Sustainability of  Island Water Supplies using Dynamic Simulation. White Paper. University of  Utah and Geological Survey of  Canada: 26 pages. Gleick, P. H. (2000). Water: The Potential Consequences of  Climate Variability and Change for  the Water Resources of  the United States. Oakland, Pacific  Institute. Guo, H. C., et al. (2001). "A system dynamics approach for  regional environmental planning and management: A study for  the Lake Erhai Basin." Journal of  Environmental Management 61: 93-111. Hare, M., et al. (2003). "Participatory Modelling in Natural Resources Management: A Comparison of  Four Case Studies." Integrated Assessment 4(2): 62-72. Hughes, B. B. (1999). International Futures: Choices in the Face of  Uncertainty. Boulder, Westview Press. ICIS (1999). Integrated Assessment: A bird's-eye view. Maastricht, The Netherlands, Maastricht University, International Centre for  Integrative Studies. IPCC Working Group II (2007). Fourth Assessment Report - Climate Change 2007: Climate Change Impacts, Adaptation and Vulnerability - Summary for  Policy Makers, Intergovernmental Panel on Climate Change: 23. Ivey, J. L., et al. (2004). "Community Capacity for  Adaptation to Climate Induced Water Shortages: Linking Institutional Complexity and Local Actors." Environmental Management 33(1): 36-47. IWR (2003). The Shared Vision Planning Website: Climate Change, Institute for  Water Resources, U.S. Army Corps of  Engineers. Accessed: February 23, 2006 at http://www.iwr.usace.army.mil/iwr/svp/climatechange.htm. Jenkins, M. W., G. F. Marques, and F. K. Lelo. (2005). WEAP as a participatory tool for shared vision planning in the River Njoro Watershed, Kenya. Impacts of  Global Climate Change: Proceedings of  the 2005 World Water & Environmental Resources Congress, Anchorage, Alaska, ASCE. Jones, A., D. Seville, and D. Meadows. (2002). "Resource sustainability in commodity systems: the sawmill industry in the Northern Forest." System Dynamics Review 18(2): 171-204. Jones, D. and D. Seville (2002). Supporting Effective  Participation in the Climate Change Debate: The Role of  System Dynamics Simulation Modeling. Asheville, NC and Hartland, VT, Sustainability Institute. Karash, R. (1996). "How to Find 'Structure.'" On-line article available at: http://world.std.com/~rkarash/structure . Accessed 21 September 2004. Kashyap, A. (2004). "Water governance: learning by developing adaptive capacity to incorporate climate variability and change." Water Science and Technology 49(7): 141-146. Keyes, A. M. and R. N. Palmer (1993). The Role of  Object Oriented Simulation Models in the Drought Preparedness Studies. Water management in the '90s: proceedings ofthe 20th anniversary conference,  Seattle, WA, American Society of  Civil Engineers. Keyes, A. M., R. N. Palmer, and W. Werick. (1995). Practicing for  Droughts: Guidelines for Virtual Drought Exercise. Proceedings of  the 22nd annual conference  of  Water Resources Planning and Management, Cambridge, MA, ASCE. Kirshen, P., et al. (2004). Infrastructure  Systems, Services and Climate Change: Integrated Impacts and Response Strategies for  the Boston-Metropolitan Area/ Climate's Long-term Impacts on Metro Boston (CLIMB) Final Report, Tufts  University, University of  Maryland, Boston University and Metropolitan Area Planning Council. Leggett, J. (2001). The Carbon War: Global Warming and the End of  the Oil Era. New York, Routledge. Letcher, R. A. and A. J. Jakeman (2003). "Application of  an Adaptive Method for  Integrated Assessment of  Water Allocation Issues in the Namoi River Catchment, Australia." Integrated Assessment 4(2): 73-89. Lettenmaier, D. P., A. W. Wood, et al. (1999). "Water Resources Implications of  Global Warming: A U.S. Regional Perspective." Climatic Change 43: 537-579. Lins, H. F., D. M. Wolock, and G. J. McCabe. (1997). "Scale and Modeling Issues in Water Resources Planning." Climatic Change 37: 63-88. Loucks, D. P. (2006). Generic Simulation Models for  Facilitating Stakeholder Involvement in Water Resources Planning and Management: A Comparison, Evaluation, and Identification  of  Future Needs. Proceedings of  the iEMSs Third Biennial Meeting: "Summit on Environmental Modeling and Software",  Burlington, USA, International Environmental Modelling and Software  Society. Meadows, D. (1997). "Places to Intervene in a System." Whole Earth 91: 78 - 84. Meadows, D. et al. (1972). The Limits to Growth. New York, Universe Books. Meadows, D. H. (1989). "System dynamics meets the press." System Dynamics Review 5(1): 69-80. Meadows, D. H. and J. M. Robinson (1985). The Electronic Oracle: Computer Models and Social Decisions. Chichester, John Wiley & Sons. Meadows, D., et al. (2004). Limits to Growth: The 30-Year Update. White River Junction, VT, Chelsea Green. Merritt, W. and Y. Alila (2004). Chapter 7: Hydrology. Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report, January 1, 2002-June 30, 2004. S. Cohen, D. Neilsen and R. Welbourn. Vancouver, Environment Canada, Agriculture and Agri-Food Canada & University of British Columbia: 63-88. Merritt, W. S., et al. (2006). "Hydrologic response to scenarios of  climate change in the Okanagan Basin, British Columbia." Journal of  Hydrology 326: 79-108. Middelkoop, H., et al. (2004). "Perspectives on flood  management in the Rhine and Meuse Rivers." River Research and Applications 20: 327-342. Miller, K. A. and D. Yates (2005). Climate Change and Water Resources: A Primer for Municipal Water Providers. Boulder, National Center for  Atmospheric Research: 83. Mortsch, L. and B. Mills (1996). Great Lakes-St Lawrence Basin Project on Adapting to the Impacts of  Climate Change and Variability. Downsview, Ontario, Environmental Adaptation Research Group, Atmospheric Environment Service, Environment Canada. Moser, S. C. and L. Dilling (2004). "Making Climate Hot: Communicating the Urgency and Challenge of  Global Climate Change." Environment 46(10): 32-46. Mostert, E. (2003). "The Challenge of  Public Participation." Water Policy 5(2): 179-197. Mote, P. (1999). Impacts of  climate variability and change in the Pacific  Northwest. Seattle, The JISAO/SMA Climate Impacts Group, University of  Washington. Mote, P. W., et al. (2003). "Preparing for  Climatic Change: The Water, Salmon, and Forests of  the Pacific  Northwest." Climatic Change 61: 45-88. Neale, T. (2005). Impacts of  Climate Change and Population Growth on Residential Water Demand in the Okanagan Basin, British Columbia. School of  Environment and Sustainability. Victoria, B.C., Royal Roads University. Neilsen, D., C. A. S. Smith, et al. (2006). "Potential impacts of  climate change on water availability for  crops in the Okanagan Basin, British Columbia." Canadian Journal of Soil Science 86: 921-936. Neilsen, D., et al. (2004). Chapter 8: Crop Water Demand Scenarios for  the Okanagan Basin. Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report, January 1, 2002-June 30, 2004. S. Cohen, D. Neilsen and R. Welbourn. Vancouver, Environment Canada, Agriculture and Agri-Food Canada & University of  British Columbia: 89-114. Nvule, D. N. (1993). Multiparty Model Development Using Object-Oriented Programming. 20th Anniversary Annual Conference  of  Water Resources Planning and Management, ASCE. Orr, D. W. (1991). "What is Education For?" Trumpeter 8(3): 99-102. Pahl-Wostl, C. and J. Newig (2003). Participatory Approaches. Osnabruck, Germany. Palmer, R. N., et al. (in review). "Disciplined Planning, Structured Participation, and Collaborative Modeling: Applying Shared Vision Planning." Journal of  Water Resources Planning and Management. Palmer, R. N., A. M. Keyes, and S. Fisher. (1993). "Empowering Stakeholders through Simulation in Water Resources Planning." in Water Management in the '90s: Proceedings of  the 20th anniversary conference  Water Res. Ping. & Mgt. Conference, Seattle, Washington, American Society of  Civil Engineers. Pataki, D., et al. (2006). BE/CNH: Urban Trace-gas Emissions Study (UTES): Interactions among Canopy Processes, Anthropogenic Emissions, and Social Institutions in the Salt Lake Valley, Utah. 2005 Annual Report. University of  Utah and National Science Foundation. Payne, J. T., et al. (2004). "Mitigating the effects  of  climate change on the water resources of the Columbia River Basin." Climatic Change 62: 233-256. Peterson, T. R., A. Kenimer, and W. E. Grant. (2004). Using Mediated Modeling to Facilitate Collaborative Learning Among Residents of  the San Antonio Watershed, Texas, U.S.A. Mediated Modeling: A System Dynamics Approach to Environmental Consensus Building. M. van den Belt.' Washington D.C., Island Press: 136-163. Radzicki, M. (1997). "Introduction to System Dynamics." U.S. Department of  Energy. Available on-line at: http://www.albany.edu/cpr/sds/DL-IntroSysDyn/inside.htm. Accessed March 2004. Renn, O. (2003). "Participatory Processes for  Designing Environmental Policies." in Conference  Proceedings for  Participatory Methods. European Forum for  Integrated Environmental Assessment and the Dutch Council for  Research on Space, Nature and Environment, Maastricht, The Netherlands. Richardson, G. P. and D. F. Andersen (1995). "Teamwork in Group Model Building." System Dynamics Review 11(2): 113-137. Riebsame, W. E., et al. (1995). Complex River Basins, in K. M. Strzepek and J. B. Smith (eds.). As Climate Changes. Cambridge, Cambridge University Press: 57-91. Risbey, J., M. Kandlikar, and A. Patwardhan. (1996). "Assessing Integrated Assessments." Climatic Change 34: 369-395. Rosenzweig, C. and D. Hillel (1998). Climate Change and the Global Harvest. New York, Oxford  Press. Rotmans, J. (1998). "Methods for  IA: The challenges and opportunities ahead." Environmental Modeling and Assessment 3: 155-179. Rotmans, J. and M. B. A. van Asselt (1996). "Integrated Assessment: A Growing Child on Its Way to Maturity: An Editorial Essay." Climatic Change 34: 327-336. Rotmans, J. and M. B. A. van Asselt (2002). Integrated Assessment: Current Practices and Challenges for  the Future, in H. Abaza and A. Baranzini. (eds.). Implementing Sustainable Development. Cheltenham, UK, Edward Elgar Publishing Ltd. i Scoones, I. (2004). "Climate Change and the Challenge of  Non-equilibrium Thinking." IDS Bulletin 35(3): 114-119. Senge, P. M. (1990). The Fifth  Discipline: The Art & Practice of  the Learning Organization. New York, Currency Doubleday. Shackley, S. and B. Wynne. 1996. "Representing Uncertainty in Global Climate Change Science and Policy: Boundary-Ordering Devices and Authority." Science, Technology and Human Values 21(3): 275-302. Shepherd, P. J. Tansey, and H. Dowlatabadi. (2006). "Context matters: What shapes adaptation to water stress in the Okanagan?" Climatic Change 78: 31-62. Simonovic, S. P. (1999). "A new modeling approach for  water resources policy analysis." Water Resources Research 35(1): 295-304. Smith, J. B., J. J. Carmichael, and J. G. Titus. (1995). Adaptation Policy. As Climate Changes. K. M. Strzepek and J. B. Smith. Cambridge, Cambridge University Press: 201-210. Stakhiv, E. Z. (1996). Managing Water Resources for  Climate Change Adaptation. Adapting to Climate Change. J. B. Smith and others. New York, Springer-Verlag: 243-264. Stave, K. A. (2002). "Using system dynamics to improve public participation in environmental decisions." System Dynamics Review 18(2): 139-167. Stave, K. A. (2003). "A system dynamics model to facilitate  public understanding of  water management options in Las Vegas, Nevada." J. of  Environ Mgt 67: 303-313. Sterman, J. D. (2000). Business Dynamics: Systems Thinking and Modeling for  a Complex World. Boston, McGraw-Hill. Stern, N. (2006). The Stern Review of  the Economics of  Climate Change. Cambridge, New Economics Foundation. r Stiles, J. M. and R. E. Punnett (1999). Including Expert System Decisions in a Numerical Model of  a Multi-Lake System Using STELLA. Sumer, D. Y. and K. Lansey (2004). Evaluation of  Conservation Measures in the Upper San Pedro Basin. World Water and Environmental Resources Congress, Salt Lake City, UT. Summit Environmental Consultants (2004). Trepanier Landscape Unit Water Management Plan, Volume 1. Vernon, Regional District of  Central Okanagan and B.C. Ministry of Sustainable Resource Management. Tansey, J. and S. Langsdale (2004). Chapter 12: Exploring Anticipatory Adaptation in the Okanagan, BC. Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report, January 1, 2002-June 30, 2004. S. Cohen, D. Neilsen and R. Welbourn. Vancouver, Environment Canada, Agriculture and Agri-Food Canada and University of  British Columbia: 165-174. Taylor, B. and M. Barton (2004). "Chapter 5: Climate Change Scenarios." in Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report, January 1, 2002-June 30, 2004. S. Cohen, D. Neilsen and R. Welbourn. Vancouver, Environment Canada, Agriculture and Agri-Food Canada and University of  British Columbia: 47-54. Tidwell, V. C., et al. (2004). "System dynamics modeling for  community-based water planning: Application to the Middle Rio Grande." Aquatic Science 66: 357-372. Toth, F. L. and E. Hizsnyik (1998). "Integrated environmental assessment methods: Evolution and applications." Environmental Modeling and Assessment 3: 193-207. van Asselt, M. B. A. and N. Rijkens-Klomp (2002). "A look in the mirror: reflection  on participation in Integrated Assessment from  a methodological perspective." Global Environmental Change 12: 167-184. van de Kerkhof,  M. (2001). A Survey on the Methodology of  Participatory Integrated Assessment. Laxenburg, Austria, International Institute for  Applied Systems Analysis. van de Kerkhof,  M. (2003). Debating Climate Change: A Study of  Stakeholder Participation in an Integrated Assessment of  Long-Term Climate Policy in the Netherlands. Utrecht, The Netherlands, Lemma Publishers. Draff  copy dated December 2003. van den Belt, M. (2004). Mediated Modeling: A system dynamics approach to environmental consensus building. Washington DC, Island Press. van den Belt, M., L. Deutsch, and A. Jansson. (1998). "A consensus-based simulation model for  management in the Patagonia coastal zone." Ecological Modeling 110: 79-103. Vennix, J. (1999). "Group model-building: tackling messy problems." System Dynamics Review 15(4): 379-401. Vennix, J. A. M. (1996). Group Model Building: Facilitating Team Learning Using System Dynamics. Chichester, John Wiley. Videira, N. (2005). Stakeholder Participation in Environmental Decision-Making: The Role of  Participatory Modeling. Lisboa, Universidade Nova de Lisboa. Viessman, W. and T. D. Feather, eds. (2006). State Water Resources Planning in the United States, American Society of  Civil Engineers. Vogstad, K. O. (2005). Methods for  environmental analysis and decision support: When differences  in analytic paradigms lead to differences  in policy conclusions. Eco-Efficiency  for  Sustainability: Quantified  methods for  decision making, Leiden, The Netherlands, Kluwer Academic Publishers. Werick, W. J.; (2004). "When Should Shared Vision Planning Be Used?" Salt Lake City, UT, Environmental Water Resources Institute of  the American Society of  Civil Engineers. CHAPTER 2: SHARED LEARNING THROUGH GROUP MODEL BUILDING FOR THE MANAGEMENT OF WATER RESOURCES AND CLIMATE CHANGE IN THE OKANAGAN RIVER BASIN, BRITISH COLUMBIA, CANADA* Climate change is affecting  hydrologic patterns in many regions of  the world; however, very few  communities have incorporated climate change into their planning initiatives. The purpose of  this research was to assist the water resources community of  the Okanagan Basin, British Columbia, Canada with incorporating climate change in their planning and policy development and with evaluating their water resources in an integrated system context. This was conducted through a participatory process centered on the development of  a system dynamics model. The study focused  on the possible impacts of  climate change on water resources management and on a wider range of  issues jointly defined  by the participants and researchers. The products of  this process were: (1) A shared learning experience among the participants and the research team; and (2) A simulation model for  increasing knowledge about the system and exploring plausible future  scenarios and adaptation opportunities. This paper describes the group model building process, while Chapter 3 provides more detail about the resulting system dynamics model. The primary objective of  this effort  was to foster  a shared learning experience between and among our research team and the Okanagan water management community. We investigated the plausible effects  on the integrated supply-demand balance and the efficacy  of  adaptive management strategies. The emphasis was on high-level scoping and futures  exploration, rather than on consensus-building and decision-making. The products of  this process enable the Okanagan community to incorporate climate change in their planning activities, and to encourage appropriate adaptation to reduce future  vulnerabilities. Specifically,  the goals of the process were: (a) To create a shared learning experience that broadens people's * A version of  this chapter has been drafted  for  publication. Langsdale, S., A. Beall, J. Carmichael, S. Cohen, C. Forster, and T. Neale. Shared Learning through Group Model Building for  the Management of  Water Resources and Climate Change in the Okanagan River Basin, British Columbia, Canada. Journal  of  Water  Resources Planning  and  Management,  to be submitted. perspectives and clarifies  the complex relationships that define  the water resources system; (b) To develop a tool for  exploring plausible water resources futures  that is tailored to the needs of  the Okanagan community; and (c) To foster  a sense of  ownership and trust in the model among the model building participants. 2.1 The Okanagan Basin The Okanagan Basin in south central British Columbia is one of  the most arid regions in Canada, with annual average precipitation ranging from  less than 300 mm to 450 mm. The long, narrow basin extends 182 km from  the Canada-U.S. Border and covers an area of  8200 2 i M km (Figure 2.1). The major economic industries in the basin are agriculture, forestry  and recreation. Agriculture accounts for  approximately 70 percent of  annual water use in the basin. The Okanagan River is one of  the only tributaries to the Columbia River that still supports viable salmon populations. Rapid development in recent decades, combined with natural hydrologic variability, has increased concern among water resource managers. In the twenty years between 1978 and 1998 the population. in the Central Okanagan doubled and the rest of  the basin also experienced rapid growth that far  exceeded projections (BC Stats 2006; Canada-British Consultative Board 1974). Drought conditions in the summers of  2003 and 2004 caused water shortages and major fires,  leading to a public review of  emergency preparedness (Filmon 2004). 2.1.1 Previous initiatives in the Okanagan Basin Prior water management and climate change research initiatives (Cohen and Kulkarni 2001; Cohen et al. 2004; 2006) provided a foundation  for  conducting the community engagement process described herein. These early dialogues increased concern for  climate change among the region's water management community and established trusting relationships between members of  the research team and key water-related professionals,  elected officials,  and environmental-interests group representatives in the region. Figure 2.1: Okanagan Basin Map with inset for location in British Columbia (From Cohen et al 2006). First, these studies led focus  groups to brainstorm and prioritize a wide array of  potential climate change impacts (Cohen and Kulkarni 2001). This was followed  by community level and basin-scale workshops, where participants discussed the social, financial  and political feasibility  of  implementing climate change adaptation measures (Cohen et al. 2004; 2006). These consultations with the community informed  the research team of  the political climate in the Okanagan, as well as residents' attitudes and beliefs  about adaptation measures, their water resource, and climate change. Water is a "hot topic" in the region, with at least 68 different  governing bodies conducting water-related studies in the basin (Stephens 2006). These early events contributed to increased concern for  climate change in the region. Prior to our efforts,  climate change was not discussed in local water resource circles. By the time the N UNITED STATES group model building project was initiated in January 2005, climate change was the principal issue that drew participants to the table. The products of  these early dialogues were primarily qualitative. The participatory modeling process described herein was initiated in order to continue the learning process. This project built on the earlier studies by continuing dialogue, integrating the sectoral climate impacts into a single model, and creating a means of  evaluating the effectiveness  of  adaptation options. The model supported effective  dialogue at the workshops by providing a common visual language, keeping discussion focused,  and actively engaging each participant. 2.2 Climate Change in Water Resources Management In many geographic areas, climate change will have significant  impacts, so the conventional method of  relying on historic data to estimate future  conditions is inadequate. New methods for  assessing the future  are necessary to maintain the reliability of  water resource systems over the long term. 2.2.1 State of the field The literature contains several case studies in which researchers have conducted impact assessments to identify  the impacts of  climate change on water resource management of  river basins (Lettenmaier et al. 1999; Middelkoop et al. 2001; Mote et al. 1999; Payne et al. 2004; Simonovic and Li 2003). Some climate change assessments also focused  on aspects of adaptation (Beuhler 2003; VanRheenan et al. 2004). Two studies from  Europe interviewed water managers to determine their reaction to adaptation scenarios (Tol et al. 2003; Arnell and Delaney 2006). Within the climate change literature, only two recent case studies have been identified  that interacted with stakeholders during the assessment process. Ivey et al. (2004) evaluated the adaptive capacity of  a community in southern Ontario, Canada through interviewing and collecting information  from  stakeholders. Kirshen et al. (2004) actively engaged stakeholders throughout their multi-year process of  modeling and evaluating climate impacts on water, energy, transportation, and public health sectors in Boston, USA. These climate change impact studies are relatively recent, and direct use of  the information by practitioners is only beginning. In 2005, only four  U.S. states included climate change in their water resources planning (Viessman and Feather 2006). In Canada, two examples of climate change being incorporated explicitly into water planning are the Trepanier Landscape Unit Water Plan (Summit 2004), which followed  from  work by Cohen et al. (2004); and work on the Great Lakes - St. Lawrence River Basin (Mortsch and Mills 1996). 2.2.2 Challenges of incorporating climate change in water resource planning There are three prerequisites that must be met for  water managers to be willing and able to incorporate aspects of  climate change into their water planning initiatives. First, the water management community must be informed  and concerned about climate change. Second, information  about potential climate impacts must be translated into terms that are relevant to the water community. Third, the water community must be able to assess the climatic impacts within the system context, including other stressors, changes, and management responses. Each of  these tasks contains specific  challenges that this project addresses. There are several reasons for  the lack of  awareness of  climate change among water resource professionals.  First, detecting the climate signal is challenging, as the signals are often confused  with noise, or not felt  directly (Berkhout et al. 2004). In areas where climate change has not yet had significant  impact, the "signal" is only virtual, in the form  of  scientific predictions. This adds a new dimension of  complexity; since there is significant  uncertainty in the character of  climate change how we should respond is not clear. Humans respond more readily to stressors that we personally experience than those that we learn about indirectly. The creeping nature of  climate change makes it easy to ignore. Because our human nature encourages us to focus  on immediate crises, small incremental changes are disregarded and continue freely  until a disaster occurs (Moser and Dilling 2004). Many climate adaptation proponents focus  on extreme events; however, in the case of  water resources management, the series of  small events, or the shift  in patterns to a more non-equilibrium dynamic can be more damaging (Dowlatabadi and Yohe 1999; Scoones 2004). For example, damage due to gradual shifts  was demonstrated in an analysis of  reservoir operation in the Columbia River Basin. Researchers at the Univ. of  Washington's Climate Impacts Group discovered that it was the seasonal shifts  of  inflow  that had the most significant  impacts on system performance  (Payne, et al. 2004). The second challenge to enabling water professionals  to incorporate climate change is that the information  typically generated by climate scientists is not directly relevant to local water professionals.  In order to incorporate climate change estimates into their planning processes, climate change information  must be translated into terms that are relevant to their concerns. While the level of  uncertainty in climate estimates makes the information  cumbersome, the mismatch of  scales is a larger hurdle. Current global climate models provide information  at large geographic scales and low spatial resolution, but managers handle small geographic areas and require data with relatively high spatial resolution (Lins et al. 1997). Furthermore, climate change data must be translated from  temperature and precipitation into terms reflecting  hydrologic impacts. Finally, even if  the first  two challenges are surmounted, and concerned water professionals acquire relevant information,  there remains the challenge of  integrating the information within the system context. For example, climate change is likely to impact not only the availability of  water supplies, but also water demand. At the same time, other stressors, such as urban development and economic trends may also affect  demand in various sectors. Only by considering climate change in the system context can we estimate future  conditions and then evaluate effective  responses. To date, most work has focused  on specific  components ^rather than the entire system. For example, Downing et al. (2003) assessed the impacts of climate change on various water demand sectors with the caveat that the results need to be considered as one element in a larger system. 2.3 Methodology This section highlights the literature from  integrated assessment and participatory modeling. The evolution of  the field  of  participatory modeling is provided, along with a discussion of advantages and disadvantages. Finally, the objectives for  this project are defined. 2.3.1 Integrated assessment Integrated assessment is a methodology that examines a problem in the context of  the entire system, ideally incorporating all components that affect  the critical issue. Dowlatabadi and Morgan (1993) observed that researchers often  only captured those components where information  was available. They proposed a new paradigm of  integrating subjective, expert-based knowledge with the better-known' parts of  the system. Engaging stakeholders in the assessment process is one way of  accomplishing this. Rotmans and van Asselt (2002) noted that effective  integrated assessment cannot be completed by scientific  experts on their own, but only through collaboration with stakeholders and decision-makers. Today, stakeholder participation in environmental resource planning and management is common practice. 2.3.2 Participatory modeling Stakeholder participation in the development of  computer models for  environmental decision-making is a relatively new field  but is rapidly becoming the status quo. Participatory modeling processes have developed from  a number of  different  fields;  as a result, there are several names and variations on the parameters by which this work is conducted for  purposes ranging from  shared learning to consensus building. Literature in the integrated assessment field  uses the term "participatory modeling" to describe applications in policy analysis and organizational learning, as well as natural resource applications such as water resources and land management. According to Videira (2005), the term "group model building" from  the system dynamics community has most often  been applied to organizational messy problems (Richardson and Anderson 1995; Vennix 1996; 1999) but has also been applied to sustainability issues (Stave 2002; 2003). The term "mediated modeling" originated from  the ecological economics community in the late 1990's and has been applied to solving complex environmental problems (van den Belt etal. 1998). Participatory modeling has several advantages (Cockerill et al. 2004; 2006; van den Belt, 2004; Vennix 1996): • Participants learn the underlying model structure, not only the results. This fosters  a higher level of  learning. • The "black box effect"  is reduced; assumptions and uncertainties are transparent. • The model can be customized to the needs of  the users, capturing critical issues and presenting the information  in a format  that is clear to them. • Increases ownership and trust of  the resulting model, and therefore  the probability of  continued use. Furthermore, there are several ways in which a participatory modeling process enhances communication and the opportunity for  consensus building: • System dynamics provides a common language that circumvents industry-specific  jargon. • The model organizes ideas into a visual language that helps keep discussion focused,  reducing tangents and circumlocution. • The system dynamics approach encourages a broader perspective and reduces anchoring on individual agendas. • Individual perspectives are acknowledged and respected by the group when they are included in the model. Once the participants see that their ideas are captured, they no longer feel  the need to "soap box" about them. Disadvantages of  participatory modeling, as compared with conventional modeling, are the increased requirements of  financial  resources, time and logistical planning needed to engage the participants in the process. Participatory modeling approaches are particularly appropriate for  complex problems, especially ones where conflict  is anticipated. The process can meet one or more of  the following  goals: foster  team learning, share information  between stakeholders, foster  future vision, develop consensus, and generate commitment. Specific  characteristics of  the applications vary widely, depending oh the context, available resources, and desired level of detail. The process may take days to years, and may result in a simple scoping model or a detailed management support model. Often  the model is not the primary goal, but is simply a means to enhance team learning, foster  consensus, and create commitment with the outcomes (van den Belt 2004; Vennix 1996). An approach applied exclusively to water resources planning and management applications evolved from  work on drought planning that began in the mid 1970's. The methodology, formalized  as "shared vision planning" (SVP) in the early 1990s has been applied to a number of  studies over the past fifteen  years by the Institute for  Water Resources, U.S. Army Corps of  Engineers. SVP combines traditional water resources planning principles, structured public participation, and integrated computer modeling (Palmer et al., in review). Participants are empowered by: (1) influencing  the modeling process directly; (2) increasing their knowledge of  the physical system and appreciation of  the perspectives of  fellow  participants; (3) being able to use the model independently; and (4) the resulting confidence  in the results (Palmer et al. 1993). Six of  the case studies did consider climate change, but not until after the stakeholder participation phase (IWR 2003). 2.3.3 Process design: means and ends objectives In designing the group model building process, we identified  primary objectives and actions that could support these primary objectives. The primary objectives are "ends objectives," while the actions that support them are "means objectives." The ends objectives are: Ends 1: To create a shared learning experience that broadens participants' perspectives and clarifies  the complex relationships that define  the water resources system; To develop a tool for  exploring plausible water resources futures  that is tailored to the needs of  the Okanagan community; and To foster  a sense of  ownership and trust in the model among the model building participants. Three means objectives for  this process were: (1) Gathering a diverse group of  stakeholders; (2) Engaging the participants actively in model development; and (3) Using the system dynamics model as a tool to enhance communication. Means 1: Gather a diverse group of  stakeholders, including the basin's water-related professionals,  environmental NGO's, and researchers, that have a role in managing the regions' water resources. By bringing together people with a variety of  perspectives, discussion at the model building sessions can include a variety of  issues and reveal different  concerns and attitudes across the basin. The gatherings thus become a learning experience for  both participants and facilitators. An additional benefit  of  gathering a diverse group is the networking opportunity among people who are stewards of  the basin's water resources. The relationships developed during the process can build community and support cooperation in times of  future  drought and water conflict.  This benefit  is experienced subtly and over time and is thus difficult  to measure. < Means 2: Actively engage stakeholders in the modeling process. Having the participants take an active role in the modeling process helps both Ends 2 and 3. Ownership of  the model can only be developed by working directly with it, and the participants must guide the process to ensure that the model suits their needs. Means 3: Use the system dynamics model as a mechanism to enhance communication Ends 2: Ends 3: The primary strength of  a participatory modeling process is that the system dynamics modeling tool serves as a mechanism for  supporting effective  communication among a group with varying levels and types of  expertise. When communication is not effective,  negotiation frequently  breaks down because of  misunderstandings from  unclear verbiage, unstated assumptions, and/or failed  logic. The system dynamics model supports dialogue through its explicit, common language that surpasses technical jargon. It also forces  people to recognize and clarify  unstated assumptions, thereby helping to question the logic based on these assumptions. Although the focus  of  this process was not on developing consensus, we established an environment that supported the initial steps of  negotiation. Fisher and Ury (1981) explain that effective  negotiation avoids the confrontation  resulting from  personal agendas. Frequently parties enter dialogue having already determined their best solutions. This, however, is not ideal, nor is it usually effective.  If,  instead, parties begin by communicating their values and objectives, they can work together to identify  win-win solutions. In the ideal use of  the tool, emphasis is placed on first  understanding the system of  mutual interest to the parties involved in the process. Only after  this should alternatives be discussed. Using a system dynamics model to enhance communication fosters  a shared learning experience with particular emphasis on understanding the inner workings of  the complex system. Van de Kerkhof  (2004) reports that in first-order  learning people simply memorize facts,  while in second-order learning people learn processes and relationships. In this work we wanted to foster  second-order learning so that the participants would gain an appreciation for  how human activities as well as climate change will impact water supply and water demand. These relations provide a foundation  for  comprehending why some adaptive strategies would be more effective  than others, thus encouraging responsible management policies. The philosophy of  system dynamics, referred  to as "systems thinking," also encourages second-order learning. Therefore,  we selected a system dynamics general software  platform to use in this process. System dynamics software  packages are flexible  enough to integrate various formats  of  information.  Commercially available software  packages (STELLA™ was used in this project) contain a graphical, transparent modeling language and provide tools for building interactive, user-friendly  interfaces  that are ideal for  engaging participant groups with a range of  technical proficiency. Brief  written surveys were administered at the beginning and end of  the last two workshops that aimed to measure these goals. Participants completed the surveys anonymously and submitted them immediately after  completion. Copies of  these surveys are attached in Appendix B. Responses are summarized in the Results section. 2.4 The Group Model Building Process This section highlights how the design of  the participatory process worked toward meeting our objectives. Additional details about the process are available in Langsdale et al. (2006). 2.4.1 Participant recruitment Invitations to join the participatory process were issued to selected individuals and representatives of  organizations, rather than advertising publicly, because we wanted to r create a diverse and balanced representation of  the various organizations and responsibilities related to water resources management in the Okanagan Basin. The invitation list was * compiled from  the contacts we had established in previous phases of  the project, by targeting key organizations and representatives, and through referrals.  All participants were invited to attend all of  the meetings detailed in Section 2.4.2 and Table 2.1. Figures 2.2 and 2.3 show an aggregation of  the levels of  participation attained throughout the series of  workshops and meetings. In Figure 2.2, participants are categorized by their affiliation  and their major role in Okanagan water management. Affiliations  are categorized as: First Nations, Federal, Provincial (BC), Regional District, or Local Governments; Environmental Non-Governmental Organizations (E-NGO); Academia; Irrigation Districts; Figure 2.2: Attendance at all events by participants categorized by affiliation.  Plotted as a comparison to the totai number of people, and weighted by the number of events each person attended. 0 No. Attendees m Weighted Participation 30% 25% 20% 15% 10% / s  S  s  / jf.  s  y • y  • Figure 2.3: Percent attendance of all participants at all six events, categorized by role in Okanagan. Plotted as a comparison to the total number of people, and weighted by the number of events each person attended. a No. Attendees a Weighted Participation 30% ^ J & J? Agricultural Association (BC Fruit Growers Association); Consultancy; or Local Initiative (The Okanagan Partnership). The groups most strongly represented were the Provincial and Local governments and environmental NGOs, each representing greater than 15 percent of total attendance. In Figure 2.3, participants are categorized by their occupational role according to these types: Elected officials;  Community and resource planners; Practitioners in operations and water management; Technical scientists and engineers that provide support from  within an agency; Researchers; Advocates (for  environment, business or agriculture); Administrative or Communicative roles within an organization; and Senior Managers (who have technical expertise but have advanced to management). The data reveals that roughly half  of  the total participation was comprised of  technical experts and advocates, each of  which represented 25 percent of  total attendance. Figures 2.2 and 2.3 plot the number of  participants as a percent of  the total and an indication of  active participation. Active (or "weighted") participation accounts for  the number of events that each person attended according to the following  relation: Weighted.Participation.  = ——— x 100% '=i where, a, = ^ {Participant i x No.Events.Attended t) i i = affiliation  type or role n = total number of  affiliations  (11) or roles (8) i In Figures 2.2 and 2.3, both the numbers of  participants and the weighted participation values are shown as percentages so that the results may be more easily compared. When the active participation bar is higher than the number of  participants bar, then attendance by members of  this group was higher than average. For example, attendance by both the agricultural association (BC Fruit Grower's Association) and the Provincial Government was higher than average. 2.4.2 Involving participants in model development The model development process was divided into six phases, based on the five  phases used in the UTES Airshed Study (Forster, personal communication). Members who joined the participatory modeling process were invited to contribute to all six phases to maximize their opportunities to contribute and to stay informed  on model progress. Each of  the events listed in Table 2.1 emphasized one of  these phases. The six phases are: (1) Visioning, (2) System Mapping, (3) Structure Construction & Refining.  (4) Quantitative Information  Gathering, (5) Model Calibration, and (6) Futures Exploration. All of  the workshops were held in Kelowna because of  its central location in the Okanagan. The small group meetings were held in various locations throughout the basin, each hosted by one or more participants at their place of  employment. We engaged the participants through plenary dialogue, hands-on activities, and interaction with the system dynamics model. I chaired each meeting, with support from  at least two others in each workshop who facilitated  sessions, group activities, and model explorations. In the first  and second workshops, I asked the participants to brainstorm objectives for  the model including research questions to investigate through use of  the model. The objectives and the research questions were closely related. The research questions are summarized as: (1) Characterize the current system. What is the current (annual and seasonal) water supply and what is current use or demand? Characterize a basin-wide water budget. (2) Characterize the future  system: What are the sustainable limits to growth in the Okanagan? What ranges of  change are possible to supply, demand, and water quality due to climate change, population growth, increased population density, drought, and land use impacts (from  development, forestry). (3) What policy levers are effective  and how effective  are they? What are the trade-offs  we will have to make for  competing demands? Can we maintain our current quality of  life? (4) Make the results easily communicated and understandable so that the model can be used to educate decision makers and the general public. Participant groups also actively supported model development when they created sketches of the important elements that influence  water supply and demand in the basin. An example is shown in Figure 2.4. The themes captured by these images served as the foundation  for developing the first  version of  the model. The first  working version of  the model was presented in the third meeting where participants explored both the model's structure level and the user interface  in facilitated  groups with three to four  participants. Participants evaluated the information  captured in the model and made suggestions for  further  refinement. In Workshops 4 and 5, participants interacted directly with the model, working in pairs at each laptop to encourage dialogue and critical thinking about their model explorations. The model interface  directs users to first  simulate and explore historic conditions by reviewing a number of  output graphs. Then, the user enters a futures  area in which he/she can select scenario settings for  simulation period, climate scenario, and population growth rate. Next, the user may test adaptation options and water management policies. Output graphs on the adaptation pages help to describe the options, while more general output graphs reveal the impact on the system's balance of  supply and demand. Images of  the model's user interface  are attached in Appendix C. Table 2.1: Summary of  events in the group model building process Event Date(s) No. of Attend-ees Major Tasks/Objectives for Participants Workshop 1 22 Feb 2005 19 (1) Visioninq: • Learn concepts of Systems Thinking • Brainstorm/Discuss objectives for model and process (2) Svstem Maopinq Workshop 2 15 Apr 2005 13 • Learn capabilities of STELLA™ • Sketch first  version of model on poster-sized paper • Discuss appropriate scales & research questions for model to answer (3) Structure Construction & Refininq Workshop 3 1 June 2005 10 • Review & evaluate the first  computer version of the model, both the model structure and the user interface levels (4) Quantitative Information Gatherinq Small Group Meetings 27 Sept -Oc t 3 2005 19 total • Review updates to model structure • Provide data to fill specific gaps (5 events) • Evaluate if early results are realistic • Recommend priorities for further  model development (5) Model Calibration Workshop 4 7 Dec 2005 17 • Verify historic and future model simulations according to personal experience (using model interface and referring to structure as needed) • Evaluate user interface and selection of alternative management strategies (6) Futures Exploration Workshop 5 26 Jan 2006 13 • Use model to simulate futures and test alternatives management strategies • Discuss insights generated Figure 2.4: Sketch created by a participant group at Workshop 2 illustrating the important linkages that affect  water supply in the Okanagan Basin. 2.5 Results We evaluated the effectiveness  of  the process in achieving our ends and means objectives through written surveys at the final  two workshops. This section discusses the results of  these surveys. 2.5.1 Means objectives Means 1: Bring together a diverse group of  people with a variety of  experiences, expertise and values. The composition of  participants shown in Figures 2.2 and 2.3, show that a diversity of organisations and roles participated in the process. No formal  analysis was conducted to determine the composition of  affected  and interested parties who should have been invited. Instead, we relied on the experience gained in previous phases of  the Okanagan project (see Section 1.7.1) and the relationships that had been established with local partners to determine who to invite. Means 2: Actively engage the participants in the model development process. Opportunities to contribute to model development took place primarily through the series of workshops and meetings. Therefore,  participation in model construction can be measured by the activities that took place in these events and by participants' level of  commitment, as measured by their frequency  of  attendance. The workshops and meetings allowed participation throughout the entire modeling process, from  setting objectives through to calibration and testing. Throughout model development, the participants provided a wealth of  information  about the details and management of  the actual system. They helped to scope issues, evaluated the appropriateness of  time and spatial scales, and advised both the model structure and calibration of  data. Previous efforts  to develop hydrologic, residential and crop water demand scenarios provided an important foundation  for  modeling the system, as described in Chapter 3. It was the information  provided by the participants, however, that helped to complete the picture by framing  the significance  of  the information  in an Okanagan Basin context. The participants provided information  on: 1) the volume of  storage available and how the reservoirs are managed throughout the year, 2) data regarding sources of  water used by different  residents, and 3) the pathways for  return flows  from  various treatment facilities.  Interesting, and sometimes conflicting,  stories regarding the requirements, regulations, and (lack of) enforcement  of  instream flows  were also provided. New ideas about conservation strategies under consideration were also outlined. The participants, many of  whom manage the system on a daily basis, were the best source of  information  on these issues. A second measure of  how active and committed the participants were in the process is the consistency of  attendance. Because participation was voluntary, some fluctuation  in attendance was expected. Figure 2.5 shows that the twelve most committed participants attended three to four  events out of  a total possible of  six. The figure  also shows that a large proportion of  all attendees only attended one event. There are several reasons for  this behaviour. First, some people attended only the first  event, then chose to or were unable to attend subsequent events. A second reason for  the large number of  single-event attendees is the round of  one- to two-hour meetings in September 2005 took place at several locations within offices  around the basin. Several people who were employed at the meeting locations were invited by their colleagues to attend the session. The short format  and location allowed people with moderate levels of  interest, or simply curiosity, to attend. With some exceptions, most of  these "drop-ins" did not continue. A third reason for  the high number of  single-event participants was that invitees spread the word and invited their colleagues, who joined the process late. In fact,  the final  session was the first  attended for  six out of  the thirteen present. Figure 2.5: Participant counts according to the number of events they attended. 4 events (6) \ Some efforts  may help to reduce the instability in attendance. Keeping the group size to a minimum (while still achieving diversity) can increase individual responsibility and ownership to the process. Identifying  and emphasizing the benefits  that participants will gain from  their involvement may help keep their attendance a priority. Financial compensation can also be used to secure a commitment. An alternate approach is to encourage representation by organizations, such that individuals may share responsibilities with their colleagues. There were several instances of  this within this project. When participation was evaluated by organization rather than by individual, single-event attendance dropped from  29 out of  51 (57 percent) to 13 out of  34 (38 percent).-(See Figure 2.6.) Figure 2.6: Organization representation counts according to the number of events attended. 5 events 2 events (8) Means 3: Use the system dynamics model as a tool to support and enhance communication In the group model building process, significant  time was spent capturing the essential structure of  the system. Water management alternatives were discussed primarily in the last two workshops. Once model building began, there was very little direct confrontation  or conflict  between participants, as the focus  remained on the model. There was certainly a difference  between the plenary session in the first  workshop, before  model construction began, and plenary sessions at later workshops. At the first  workshop the discussion was rather conceptual, with speakers contributing both verbal descriptions and imagery to convey their values and opinions about the future  of  the basin. Differences  in opinion caused some tension among speakers, as well as between the speakers and the facilitator.  Certainly, part of the differences  can be attributed to the personalities present in the room. However, in later workshops, discussion remained on issues specifically  related to the water resources in the basin and centered on (a) learning about what the model included, and (b) suggestions for improving the model. This suggests that the presence of  the model in the workshops helped to redirect personal agendas and foster  effective  communication as required for  negotiation (Fisher and Ury 1981) In the final  plenary session of  the workshop series, participants shared and discussed the insights that they learned from  exploring the model. These included: (a) It is difficult  to manage the system when instream flows  for  sockeye (in downstream portions of  the river network) are given priority; (b) Transferring  all of  the diversion points from  the tributary streams to Okanagan Lake (to improve aquatic habitat in the tributary streams) causes the lake level to drop significantly;  and (c) Residential (domestic and urban municipal) demand, even under projected rapid growth rates, remains less significant  than anticipated in terms of overall percentage of  demand. This is largely because the significant  increase in agricultural demand due to climate change continues to dominate the demand profile. 2.5.2 Ends objectives Ends 1: To create a shared learning experience that broadens participants' perspectives and clarifies  the complex relations that define  the water resources system. Gathering a diverse group of  stakeholders together and providing opportunities to share their perspectives helped to broaden people's understanding beyond their own experiences. Furthermore, through developing and interacting with the model, participants clarified  their understanding of  the linkages and processes that define  the water resource system. When discussing the model we emphasized causal relationships rather than facts  to support second-order learning (van de Kerkhof  2004). Participants found  both the process and the resulting model valuable. Participants assessed their own learning experience in evaluations given at the final  two workshops using both numeric scales and written comments. Copies of  these evaluation forms  are attached in the appendix. Numerical responses for  the two workshops are combined and summarized in Table 2.2. Only five  of  the twenty-five  people at the final  workshops attended both events. In the post-session evaluations, we asked, "Have your perceptions of  future  water availability in the basin changed due to this exercise?" The mode of  responses was in the middle with "some change", while the median value was slightly lower. Because all of  the participants entered the session with previous knowledge about the system (indeed many assist in managing the water resources daily) the weighting towards the left  side of  Table 2.2 was not surprising. In written comments on the evaluations, five  responses noted that the results described a picture that was worse than they previously thought, while two responses noted it was better than they previously thought. Survey comments suggest that prior knowledge influenced  participants' experience of learning. Two of  the level "1 - no chaiige" responses (see Table 2.2) noted that they had "a pretty good sense" of  the system beforehand.  In contrast, an environmental NGO representative said the experience increased his/her learning "[ijmmensely ~ The whole scope of  the exercise was mind expanding." Several water resource professionals  felt  that the experience increased their clarity of  system linkages. One participant noted a better appreciation for  the "challenges inherent in holistic management" and for  the value of  "drawing together varied aspects of  water." Another respondent showed that he/she grasped the dynamic nature of  the system particularly well by observing that the concept of  "limits to growth," which was expressed in the early stages, as "How many people can the basin support?" is tempered by our ability to manage and adapt our way through drought. Table 2.2: Number of responses to the post-workshop evaluation question "Have your perceptions of future water availability in the basin changed due to this exercise?" 1 No Change 2 3 Some Change 4 5 Major Change No Response Workshop 4 2 3 4 0 1 6 Workshop 5 2 4 5 0 0 5 Total 4 7 9 0 1 11 Ends 2: To tailor the model to meet the needs of  the community for  the purposes of planning and exploration The suitability of  the tool for  future  exploration will be best evaluated through use over time. However, during the final  two workshops, we did ask the participants if  they saw potential in the tool. In our survey, we asked, "Do you feel  this model is a legitimate and relevant tool to explore long-term water management in the Okanagan?" All respondents agreed with this statement. Several felt  that it had the potential for  being a useful  tool, noting either that it would depend on the user, or that additional updates were needed. As shown in Table 2.3, people's trust in the model increased slightly from  Workshop 4 to Workshop 5. This is reasonable because the model was more refined  in the final  event than in the previous one. Several people commented that it was a good communication and/or education tool for visualizing the future  and discussing policy. This is in alignment with the original purpose of the model. Table 2.3: Comparison of responses by event to the evaluation question: "Do you feel this model is a legitimate and relevant tool to explore long-term water management in the Okanagan?" No Yes Maybe/Has Potential No Response Workshop 4 0 3 7 5 Workshop 5 0 5 5 5 Total 0 8 12 10 Ends 3: To foster  a sense of  ownership and trust in the model among the participants. ) The positive responses summarized in Table 2.3 suggest that the participants had developed some level of  trust in the model being appropriate for  future  water explorations. The second issue, ownership, was measured indirectly by evaluating how familiar  the participants were with the model. A commitment to learning about and contributing to the model can both result from  and generate a sense of  ownership of  the model. Therefore,  the level of understanding of  the model can indicate a feeling  of  ownership. In the process of  learning the model's potential and its limitations and assumptions, the participants learn how to interpret the results most effectively.  If  participants understand and trust the model they may be inclined to more seriously consider the results and to incorporate them into their own work. We used surveys to gather feedback  from  the participants as to what extent they learned about the model. At the two final  workshops, we asked the participants to self-evaluate  "How well do you understand the model's structure?" in both pre- and post-session surveys. The < v scale ranged from  1 (not at all) to 5 (very well - 1 could teach it to someone else). The middle option was defined  as 3 (I know what issues are included). Responses ranged from  1 to 4. Average values at each stage are provided in Table 2.4. We speculate that the similar pre-evaluation scores for  the two workshops could be attributed to the large turn-over in participation between these two events. Average understanding increased substantially at both workshops between the pre- and post-evaluations. The smaller increase in understanding at the fifth  workshop could be partly attributed to the increased complexity of  the model that the participants had to learn. However, these changes could also be due to the different composition of  each group. For the five  participants who did attend both events, it was impossible to determine whether their responses changed, as evaluations were anonymous. Table 2.4: Average responses on a scale from 1 (low) to 5 (high) for the question: "How well do you understand the model's structure?" Pre-Evaluation Post-Evaluation Change Workshop 4 2.3 3.2 + 0.9 Workshop 5 2.5 3.1 + 0.6 There were two challenges to fostering  ownership. First, as described above, participation in the workshops and meetings was inconsistent. Second, participants did not code the model themselves, but provided information  to the facilitators,  who then constructed the model outside of  the workshops. Based on these evaluation questions and comments at the workshop sessions, the participants generally trusted the model, however, their sense of  ownership was limited by their limited interaction with it. 2.6 Discussion 2.6.1 Lessons learned Bringing individuals into a volunteer process is challenging. In the Okanagan Basin of British Columbia, there are currently many water planning initiatives competing for  the time and attention of  water-related professionals.  We made a concerted effort  to accommodate everyone's schedules. For example, we intentionally did not meet over the summer months, when many people take vacations, but waited until the end of  September to hold the next event. However, late September is the time of  both tree fruit  harvesting and salmon migration field  work. Maintaining communication with participants regarding scheduling can avoid some of  these conflicts.  Participants have different  motivations for  attending, so may have varied levels of  interest at different  stages of  the process. For example, we anticipated that the elected officials  and environmental advocates would be more interested in the early stages, when objectives were being established, and the final  stages, when results were generated and discussed; rather than during the "nuts and bolts" of  the model construction. In fact,  there was more participation from  these two groups early and late in the process as shown in Figure 2.7. This figure  also shows that technical experts dominated participation at the September small group meetings, although all participants were invited to attend. This outcome likely resulted from  our need to fill  particular data gaps, combined with participant self-selection. Figure 2.7: Distribution of attendance for each of the five workshops and the series of small group meetings, grouped by participants' roles in Okanagan water management. W1 • W2 0 W 3 B Sept Meeting B W 4 0 W 5 Elected Official  Pjanning Oper & Mgt Tech-Sci/Engr Sr. Management In this process, we chose to hold five  full-day  workshops throughout the year. We found  this to be the minimum required to enable participants to follow  and contribute to model development. More frequent,  shorter meetings might have been more effective  at keeping the participants updated on model progress, and would have created more opportunities for people to participate. However, in this case, the spacing of  the workshops helped us to minimize travel expenses for  facilitators,  and allotted sufficient  time to both revise the model and conduct event planning. Better familiarity  with and ownership of  the model could have been fostered  without the long summer break when the bulk of  the model was assembled. During the workshop sessions, participants raised many questions about model assumptions and requested model documentation for  their own reference.  Providing model documentation to the participants before  or at each session would have been an asset. The documentation will be made available to all participants through publication in Cohen and Neale (2006). Throughout the process, we encouraged the participants, rather than the project team, to define  the process. Our intent was to encourage them to take ownership of  the model. This was effective  for  most of  the process, but some ground rules are necessary to make sure that everyone is "on the same page." Better clarification  of  the process will also help invitees to know what to expect, which will help their decision to participate. Some examples of  ground rules that are important to determine and reiterate throughout the process are: (1) the purpose of  the model and its intended use; (2) the expected outcome of  process; and (3) the intended model user. In this project, the group model building participants were the model users. Evaluations are critical to understanding the impact of  the process. Although we did conduct a number of  evaluations along the way, a more structured format  to the evaluation process would have more effectively  recorded participant learning, the effectiveness  of  the process, and individual experiences. Evaluations given at the beginning and end of  a participatory modeling process could measure evidence of  learning about the characteristics of  the system. This project measured whether participants thought they learned, but did not measure learning directly. It is challenging to know what learning will occur at the outset of  the process, and you want to be careful  not to introduce bias through the instrument of  measure, whereas the evaluation sheet is informative.  One solution is to ask generally about the critical issues and solutions surrounding the topic. What the respondents describe as the critical issues and solutions can provide insight as to how informed  they are and how wide their "lens" is. Their responses can help to distinguish if  they just know a specific  component, or if they are aware of  the greater system. Do they know "the" solution, or do they appreciate the complexity of  managing the resource?.With the use of  a identifying  code to keep anonymity, changes in understanding (i.e., learning) can be measured from  the beginning to the end of the process. 2.6.2 Possible sources of bias There are a number of  factors  that could have introduced bias into this process, influencing the results that we achieved. First of  all, the personalities of  the facilitators,  particularly the chair, define  the atmosphere and the attitudes that are present in the room. I chaired all of  the sessions, and tried to be welcoming and appreciative of  participants' efforts,  as well as maintaining an energetic and positive attitude about the process. Any experienced public speaker is well aware of  how their attitude affects  their audience, and this phenomenon certainly could have played a role in the workshops. All participants were volunteers, so those who came must have felt  that it was worth their time and money to attend. Therefore,  self-selection  created bias such that all those who were present already believed in the value of  the process. Once a part of  it, there were elements that could have motivations why they wanted to process to succeed. They may have felt  the need to justify  their time to their employer; or it could have been to please the research team -and me in particular, as the nice doctoral student whose future  partially relied on the success of  the process. The evaluation of  whether the tool is appropriate also contains bias because most people who evaluated it had some level of  contribution to it. The responses to the evaluation were intended to reveal if  the participants felt  it was appropriate, and not serve as an objective evaluation of  the tool. 2.6.3 Did this process make a difference? Through the evaluations and direct communication with the participants, we heard that learning occurred through model construction, but also simply from  bringing together a diversity of  experience and expertise. Through discussions, many participants gained an appreciation for  the values of  others. They also gained insight on the complexity of  water management. Additionally, the single-disciplinary researchers who participated gained new perspective from  the relatively simple, high-level, multi-disciplinary model. The process did make a difference  to the participants, who were positive about the experience. Throughout the process, participants recommended that this work be shared with a wider community, particularly to elected officials  and the public. For example, several participants expressed interest in collaborating with related initiatives in the region, including Smart Growth on the Ground (Siu 2006), an organization that fosters  sustainable community design and planning within communities, and the Okanagan QUEST model, a tool to explore urban futures  (Carmichael et al. 2004; Tansey et al. 2002; Robinson and Tansey 2006). Furthermore, the resulting model can be used to continue dialogue with the community regarding the evaluation and selection of  appropriate adaptation strategies for  reducing the negative impacts of  climate change on their ability to manage water resources in the basin. Participants' reported satisfaction  with the process supports evidence that participatory modeling is an effective  tool for  fostering  communication and supporting effective consensus-building. The long-term significance  of  this participatory modeling process to policy development cannot be measured during the timeframe  of  this project. Decisions are complex and are not based on single factors.  Furthermore, decisions are made by many more people than just those who participated in our model building exercise. Since we do not have a control group, we may never be able to measure exactly what changed as a result of  our efforts. 2.7 References Arnell, N. W. and E. K. Delaney (2006). "Adapting to Climate Change: Public Water Supply in England and Wales." Climatic Change 78: 227-255. BC Stats (2006). "Sub-provincial population estimates." Available at: http://www.bcstats.gov.bc.ca/data/pop/pop/estspop.asp. Berkhout, F., J. Hertin, and D. M. Gann. (2004). Learning to adapt: Organisational adaptation to climate change impacts. Tyndall Centre Working Paper, No. 47. Sussex, University of  Sussex. Beuhler, M. (2003). "Potential impacts of  global warming on water resources in southern California."  Water Science and Technology 47(7-8): 165-168. Canada-British Columbia Consultative Board. (1974). Okanagan Basin Agreement: Summary Report of  the Consultative Board including the Comprehensive Framework Plan. Carmichael, J., J. Tansey, and J. Robinson. (2004). "An integrated assessment modeling tool." Global Environmental Change 14:171-183. Cockerill, K., H. Passell, and V. Tidwell. (2006). "Cooperative Modeling: Building Bridges between Science and the Public." Journal of  the American Water Resources Association 42(2): 457-471. Cockerill, K., V.C. Tidwell, and H. Passell. (2004). "Assessing Public Perceptions of Computer-Based Models." Environmental Management 34(5): 609-619. Cohen, S. and T. Kulkarni, eds. (2001). Water Management & Climate Change in the Okanagan Basin. Vancouver, Environment Canada & University of  British Columbia. Cohen, S. and T. Neale, eds. (2006). Participatory Integrated Assessment of  Water Management and Climate Change in the Okanagan Basin, British Columbia. Vancouver, Environment Canada and University of  British Columbia. Cohen, S., et al. (2006). "Learning with Local Help: Expanding the Dialogue on Climate Change and Water Management in the Okanagan Region, British Columbia, Canada." Climatic Change 75: 331-358. Cohen, S., D. Neilsen, and R. Welbourn, eds. (2004). Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report, January 1, 2002-June 30, 2004. Vancouver, Environment Canada, Agriculture and Agri-Food Canada & University of  British Columbia. Dowlatabadi, H. and G. Yohe (1999). "The who-how of  adaptation: towards more realistic impact assessments." Draft. Dowlatabadi, H. and M. G. Morgan (1993). "Integrated Assessment of  Climate Change." Science 259: 1813,1932. Downing, T. E., et al. (2003). Climate Change and the Demand for  Water, Research Report. Oxford,  Stockholm Environment Institute. Filmon, G. and Review Team (2004). Firestorm 2003 Provincial Review. Victoria, Government of  British Columbia, Fisher, R. and W. Ury (1981). Getting to Yes: Negotiating Agreement Without Giving In. Boston, Houghton Mifflin. Ivey, J. L., et al. (2004). "Community Capacity for  Adaptation to Climate Induced Water Shortages: Linking Institutional Complexity and Local Actors." Environmental Management 33(1): 36-47. IWR (2003). The Shared Vision Planning Website: Climate Change, Institute for  Water Resources, U.S. Army Corps of  Engineers. Accessed: February 23, 2006 at http://www.iwr.usace.armv.mil/iwr/svp/climatechange.htm. Kirshen, P., et al. (2004). Infrastructure  Systems, Services and Climate Change: Integrated Impacts and Response Strategies for  the Boston Metropolitan Area/ Climate's Long-term Impacts on Metro Boston (CLIMB) Final Report, Tufts  University, University of  Maryland, Boston University and Metropolitan Area Planning Council. Langsdale, S., et al. (2006). Ch. 4: Shared Learning Through Group Model Building. Participatory Integrated Assessment of  Water Management and Climate Change in the Okanagan, British Columbia, Canada: Final Report. S. Cohen and T. Neale. Vancouver, UBC and Environment Canada. Lettenmaier, D. P., et al. (1999). "Water Resources Implications of  Global Warming: A U.S. Regional Perspective." Climatic Change 43: 537-579. Lins, H. F., D. M. Wolock, and G. J. McCabe. (1997). "Scale and Modeling Issues in Water Resources Planning." Climatic Change 37: 63-88. Middelkoop, H., et al. (2001). "Impact of  Climate Change on Hydrological Regimes and Water Resources Management in the Rhine Basin." Climatic Change 49: 105-128. Mortsch, L. and B. Mills (1996). Great Lakes-St Lawrence Basin Project on Adapting to the Impacts of  Climate Change and Variability. Downsview, Ontario, Environmental Adaptation Research Group, Atmospheric Environment Service, Environment Canada. Moser, S. C. and L. Dilling (2004). "Making Climate Hot: Communicating the Urgency and Challenge of  Global Climate Change." Environment 46(10): 32-46. Mote, P. (1999). Impacts of  climate variability and change in the Pacific  Northwest. Seattle, The JISAO/SMA Climate Impacts Group, University of  Washington. Palmer, R. N., A. M. Keyes, and S. Fisher. (1993). "Empowering Stakeholders through Simulation in Water Resources Planning." in Water Management in the '90s: proceedings of  the 20th anniversary conference  Water Res. Ping. & Mgt. Conference, Seattle, Washington, American Society of  Civil Engineers. Palmer, R. N., et al. (in review). "Disciplined Planning, Structured Participation, and Collaborative Modeling: Applying Shared Vision Planning." Journal of  Water Resources Planning and Management. Payne, J. T., et al. (2004). "Mitigating the effects  of  climate change on the water resources of the Columbia River Basin." Climatic Change 62: 233-256. Richardson, G. P. and D. F. Andersen (1995). "Teamwork in Group Model Building." System Dynamics Review 11 (2): 113-137. Robinson, J. and J. Tansey (2006). "Co-Production, emergent properties and strong interactive social research: The Georgia Basin Futures Project." Science and Public Policy 33(2): 151-160. Rotmans, J. and M. B. A. van Asselt (2002). Integrated Assessment: Current Practices and Challenges for  the Future. Implementing Sustainable Development. H. Abaza and A. Baranzini. Cheltenham, UK, Edward Elgar Publishing Ltd. Scoones, I. (2004). "Climate Change and the Challenge of  Non-equilibrium Thinking." IDS Bulletin 35(3): 114-119. Simonovic, S. P. and L. Li (2003). "Methodology for  Assessment of  Climate Change Impacts on Large-Scale Flood Protection System." Journal of  Water Resources Planning and Management 129(5): 361-371. Siu, J. (2006). Residential Policy, Planning and Design - Incorporating Climate Change into Long-term Sustainability in Oliver. Participatory Integrated Assessment of  Water Management and Climate Change in the Okanagan Basin, British Columbia. S. Cohen and T. Neale. Vancouver, Environment Canada and University of  British Columbia: 107-111. Stave, K. A. (2002). "Using system dynamics to improve public participation in environmental decisions." System Dynamics Review 18(2): 139-167. Stave, K. A. (2003). "A system dynamics model to facilitate  public understanding of  water management options in Las Vegas, Nevada." J. of  Environ Mgt 67: 303-313. Stephens, K., et al. (2005). Water Balance Management in the Okanagan: Now What Do We Do? Conference  Proceedings 2005: Water - Our Limiting Resource, Towards Sustainable Water Management in the Okanagan, Kelowna, B.C., Canadian Water Resources Association, British Columbia Branch. / Summit Environmental Consultants (2004). Trepanier Landscape Unit Water Management Plan, Volume 1. Vernon, Regional District of  Central Okanagan and B.C. Ministry of Sustainable Resource Management. Tansey, J., et al. (2002). "The future  is not what it used to be: participatory integrated assessment in the Georgia Basin." Global Environmental Change 12: 97-104. Tol, R. S. J., et al. (2003). "Adapting to Climate: A Case Study on Riverine Flood Risks in the Netherlands." Risk Analysis 23(3): 575-583. van de Kerkhof,  M. (2004). Debating Climate Change: A Study of  Stakeholder Participation in an Integrated Assessment of  Long-Term Climate Policy in the Netherlands. Utrecht, The Netherlands, Lemma Publishers. van den Belt, M. (2004). Mediated Modeling: A system dynamics approach to environmental consensus building. Washington DC, Island Press. van den Belt, M., L. Deutsch, and A. Jansson. (1998). "A consensus-based simulation model for  management in the Patagonia coastal zone." Ecological Modeling 110: 79-103. VanRheenen, N. T., A. W. Wood, et al. (2004). "Potential Implications of  PCM Climate \ Change Scenarios for  Sacramento-San Joaquin River Basin Hydrology and Water Resources." Climatic Change 62: 257-281. Vennix, J. (1999). "Group model-building: tackling messy problems." System Dynamics Review 15(4): 379-401. Vennix, J. A. M. (1996). Group Model Building: Facilitating Team Learning Using System Dynamics. Chichester, John Wiley. Videira, N. (2005). Stakeholder Participation in Environmental Decision-Making: The Role of  Participatory Modeling. Lisboa, Universidade Nova de Lisboa. Viessman, W. and T. D. Feather, eds. (2006). State Water Resources Planning in the United States, American Society of  Civil Engineers. CHAPTER 3: A SYSTEM DYNAMICS MODEL FOR EXPLORING WATER RESOURCES FUTURES UNDER CLIMATE CHANGE* The purpose of  this initiative was to enable and support a water resources community in incorporating climate change projections into their planning and policy development and in evaluating their water resources within a system context. This was conducted through a participatory integrated assessment centered on the development of  a system dynamics model. The products of  this process were: (1) A simulation model of  the water resource system, incorporating future  projections of  climate change and population growth, as well as adaptation options; and (2) A shared learning experience for  both the participants and the research team. This paper describes the model structure and an analysis of  model output, while the shared learning process is described in Langsdale et al. (2006) and in Chapter 2 of this thesis. The model was created to investigate several questions: (a) What is the current state of  the system? (b) What effect  will climate change and population growth have on the future  water supply and demand balance? (c) What role could adaptation measures have on improving or maintaining water resource system reliability despite increased stress from  climate change and population growth? Following a brief  introduction to "participatory modeling," this paper presents quantitative results for  (a) and (b) and discusses insights related to point (c). The model was constructed as a high level scoping model, with greater emphasis on capturing the structure of  the system, rather than on calibrating data. This is appropriate to the objectives, as the model has not been and will not be used for  design purposes. The significance  of  the results presented here is provided in the general trends, rather than in any specific  numerical values. * A version of  this chapter has been submitted for  publication. Langsdale, S., Beall, A., Carmichael, J., Cohen, S., Forster, C. An Exploration of  Water Resources Futures Under Climate Change Using System Dynamics Modeling. Integrated  Assessment, submitted. 3.1 Climate Change and Water Resources Planning Are water managers prepared for  operating under future  climate conditions? Water managers have always worked towards reducing risk and increasing system capacity to handle ever-widening extreme conditions, so some argue that they are ready for  climate change. Stakhiv (1996) agrees that society is constantly adapting in incremental steps and that climate change will simply be an additional stressor to which we must adapt. However, these "common-place adaptations" that have been part of  daily management practices assumed that climate was relatively stable, varying around a stable mean (de Loe and Kreutzwiser, 2000). It is unknown whether future  climate changes will occur gradually, over several decades, or if there will be sudden shifts.  Kashyap (2004) suggests that climate change adaptation is not comparable to historic adaptation, because the environmental changes will be more rapid and intense than in the past. However, most climate modeling characterizes climate change as occurring slowly and gradually, justifying  a reactive or "wait-and-see" approach. Regardless, the conventional practice of  relying on historic data to estimate future  conditions is inadequate. New methods for  assessing the future  are necessary to maintain the reliability of water resource systems over the long term. Certainly the future  contains many unknowns, so an effective  assessment needs to integrate all known stressors on the system and support the development of  strategies that are flexible  and resilient under a wide range of  future conditions. In many regions, climate change will be one of  the significant  stressors, so it must be included in any planning initiative. Few communities have done this, however, as assessments of  climate change on water resources in Canada are rare (Mortsch and Mills 1996; Cohen et al. 2006 are exceptions). The U.S. is similar, with only four  U.S. states including climate change in their water resources planning in 2005 (Viessman and Feather 2006). I For water managers to incorporate climate change issues into their planning processes, climate change information  must be translated into terms that are relevant to their concerns. Current challenges include the level of  uncertainty in climate change estimates and the mismatch of  both spatial and temporal scales. While current global climate models provide information  at large geographic scales and low spatial resolution, managers handle small geographic areas and require data with relatively high spatial resolution (Lins et al. 1997). A third challenge is that of  translating climate change data into terms of  hydrologic impacts. The presence of  these issues has, to date, deterred practitioners from  bringing climate change into the water management forum.  Unfortunately,  projections of  future  conditions that do neglect climate change could be grossly inaccurate, and managers who rely on this information  may be unwittingly and unnecessarily allowing vulnerabilities in their systems. While a community can often  endure single-year events without permanent losses, a prolonged deficit  in the water balance could deplete water storage in reservoirs and groundwater aquifers,  and even collapse industries dependent on water. 3.2 The Okanagan Basin Study Area The Okanagan Basin in south-central British Columbia is one of  the most arid regions in Canada, with annual average precipitation ranging from  less than 300 mm to 450 mm. The long, narrow basin extends 182 km from  the Canada-U.S. Border and covers an area of  8200 km (See Figure 3.1). The major economic industries in the basin are agriculture, forestry  and recreation. Agriculture accounts for  approximately 70 percent of  annual water use in the basin. The Okanagan River is one of  the only tributaries to the Columbia River that still supports viable salmon populations. In recent decades, rapid development combined with natural hydrologic variability increased concern among water resource managers. In the twenty years between 1978 and 1998 the population in the Central Okanagan Regional District doubled and the rest of  the basin also experienced rapid growth that far  exceeded projections (BC Stats 2006; Canada-British Consultative Board 1974). Drought conditions in the summers of  2003 and 2004 caused water shortages and major fires,  leading to a public review of  emergency preparedness (Filmon 2004). Figure 3.1: Okanagan Basin Map showing the delineation of the Uplands water supply model region that includes all managed tributaries to Okanagan Lake. Numbers correspond to watershed names, which are available in the Table 1 in Appendix D (Original map from Merritt and Alila 2003). CrtyiMtcfc l&k m Tfe  sr sle South End to 20 Kilom«t«r$ Uplands 3.3 Project History Several previous research initiatives focusing  on both the physical and social aspects of  the system established a sound foundation  on which to build this project. Stakeholder dialogue activities between 2001 and 2004 began communications and developed trust with parties responsible for  or interested in water management in the Okanagan. In addition, these activities increased awareness and concern about potential climate change impacts as well as adaptation opportunities (Cohen and Kulkarni 2001; Cohen et al. 2004; 2006). As a result, one Okanagan community included climate change scenarios in their water resources planning document (Summit Environmental Consultants 2004). Taylor and Barton (2004) statistically downscaled six global climate models to create a range of  plausible scenarios for  the Okanagan. These climate scenarios show mean temperature increases between 1.5 and 4 degrees Celsius throughout the year, and generally wetter winters and drier summers. Merritt and Alila (2004) and Merritt et al. (2006) incorporated these climate scenarios within simulations by the UBC Watershed stream flow  runoff  model (Quick 1995) to generate hydrologic scenarios. The results show significant  changes to the annual hydrograph from  the historic period (1961-90) to the period between 2010 and 2100. All scenarios show a reduced snowpack, an earlier onset of  the spring freshet  (by as much as four  to six weeks in the 2080's), and decreases in summer precipitation. Some scenarios also show more intense spring freshets.  Neilsen et al. (2004b) used the climate scenarios to model the impact on agricultural crop water demand. Higher temperatures increase both evapotranspiration and the length of  the growing season - two factors  which increase crop water demand. As a result, crop water demand could increase by 12 to 61 percent, as climate change intensifies  through the decades. Furthermore, Neale (2005; 2006) correlated residential outdoor watering with temperature and detached dwellings for  several Okanagan communities, showing that water demand in the residential sector will also increase under climate change in the absence of  conservation measures. Each of  these results on its own provides important information,  but only reflects  part of  the picture. By considering these impacts together - in the system context - we can determine the increased risk to the water resource system in the future. 3.4 Methodology This section reviews the methodology of  participatory modeling and system dynamics modeling. 3.4.1 Participatory modeling Participatory modeling is a recently established approach for  conducting integrated assessment but it has already been applied to a variety of  fields  such as policy analysis and organizational learning, as well as environmental resource applications such as water resources and land management. Participatory modeling is founded  on the belief  that mental models are based on numerous unstated assumptions and often  contain gaps and inconsistencies. The process of  sharing these mental models exposes points of  agreement and points of  conflict.  Effective  conflict  negotiation illuminates hidden assumptions so that they may be clarified  and challenged (Fisher and Ury 1981). Participatory model development can focus  on characterizing system structure, while model simulations reveal system behaviour, which is less intuitive and which is often  the source of  confusion  (Forrester 1987; Vennix 1996). The model can then be used to explore a range of  future  conditions or assumptions. Participants may engage directly in the modeling process, or the model may be developed in an iterative process with regular opportunities to contribute (van Asselt and Rijkens-Klomp 2002). 3.4.2 System dynamics Models used for  collaborative modeling in water resources applications include system dynamics platforms  like STELLA™ (Cardwell et al. 2004; Costanza and Ruth 1998; Langsdale et al. 2006; Palmer et al. in review), and Studio Expert (Tidwell et al. 2004). Other types of  models which have been used include MIKE-BASIN (Borden and Spinazola 2006; Borden et al. 2006); the Water Evaluation And Planning system model (WEAP) (Jenkins et al. 2005); and OASIS with OCL (Hydrologies 2003). System dynamics software  packages are blank slates and can be applied to any problem, while MIKE-BASIN, WEAP, and OASIS are all limited to water resources applications. System dynamics was developed for  the purpose of  characterizing complex, non-linear systems through capturing interrelations, feedback  loops and delays. Modern system dynamics software  packages are ideal for  use with a participant group of  varying levels of technical proficiency  because of  their graphically-based model level and user interface. These models can easily manage both clearly-defined  and poorly-defined  components in the same model. Similarly, they can capture quantitative, physical parts of  the system, such as hydrology, as well as intangible parts of  the system, such as policies and human responses, so they are quite appropriate for  participatory modeling applications. Case studies where the system dynamics approach was applied to environmental issues include: the Louisiana coastal wetlands, the South African  fynbos  ecosystems and the Patuxent River watershed in Maryland, USA (Costanza and Ruth 1998); water resources management in Switzerland, Senegal and Thailand, and vegetation management in Zimbabwe (Hare et al. 2003); water allocation issues in the Namoi River, Australia (Letcher and Jakeman 2003); transportation and air quality in Las Vegas, USA (Stave-2002); and Patagonia coastal zone management (van den Belt et al. 1998). 3.4.3 Representing uncertainty using scenarios Scenarios are defined  as "plausible combinations of  circumstances that can be used to describe a future  set of  conditions" (Smith et al. 1996). The scenario approach provides an alternative to the convention of  aggregating results into an average value and then representing the uncertainty with error bars or statistics, as shown in Figure 3.2. Whether the audience is technically-trained or not, expressing results in terms of  scenarios provides a clearer picture of  the range of  future  states possible. Figure 3.2(a) shows five  equally plausible scenarios, while Figure 3.2(b) reports the average of  the five  scenarios and uses error bars to represent the extent of  the individual scenarios. So, both figures  represent essentially the same feasible  region or decision space. However, Figure 3.2(a) more clearly conveys that any of  the five  states are equally plausible, while Figure 3.2(b) implies that the average condition is most likely and that the probability of  occurrence decreases toward the limits of  the feasible  region. When users of  this information  are presented with results as an average, there is a temptation is to focus  primarily on the single point, and the importance of the range of  uncertainty is lost. Figure 3.2: Graph (a) shows results as an array of discrete states, while graph (b) shows only the average of the five scenarios, along with error bars that provide the range of possible conditions. (a) Scenario Results (b) Average Condition 0.0 0.2 0.4 0.6 0.8 1.0 X X Note that in Figure 3.2, the model uncertainty, based on knowledge uncertainty is not represented. As long as the inherent uncertainties are significantly  larger than the model uncertainties, then the feasible  region defined  by the complete range of  scenarios will fully encompass the region defined  by model uncertainties (Langsdale, accepted) The simplest form  of  a scenario analysis generates an array of  equally-plausible scenarios. In / circumstances when different  probabilities of  occurrence are relevant, Bayesian Belief Networks may be used to estimate the likelihood of  each scenario. Bayesian Belief  Networks characterize the cause-effect  relationships in a system using conditional probabilities (Ghabayen and McKee, 2006). The use of  scenarios, whether equally plausible or containing assigned probabilities, helps to alleviate several challenges. First, using a framework  of  discrete scenarios helps to spell out the sources of  uncertainty and highlights their inherent nature, which may reduce people's intolerance of  uncertainty. Next, explicitly displaying the range of  conditions helps to prevent anchoring and overconfidence  in a single point (as in the average state). The message that the future  could be any of  the number of  states is continually reinforced.  Finally, the discrete scenarios also present the results in a format  that is manageable and can be readily used for further  assessment. Presenting the scenarios through a decision support tool can further  this goal by providing a framework  for  stakeholders to manage and evaluate the scenarios effectively. 3.5 Description of the Actual and Modeled System The Okanagan Sustainable Water Balance Model (OSWBM) simulates future  conditions by projecting current conditions and overlaying the effects  of  population growth and climate change on water supply and demand. The purpose of  the model is primarily for  supporting stakeholder dialogue surrounding the issue of  how climate change could play a role in future water management and is not intended to optimize design or guide real-time operation. The model can help dialogue participants to learn about the complexities of  managing water resources for  multiple uses, simulate a range of  plausible water resources futures,  assess adaptation strategies (and portfolios  of  strategies), identify  data gaps, and prioritize areas of future  research. Here, we provide detail about the Okanagan water resources system and how it was characterized in the Okanagan Sustainable Water Resources Model (OSWRM) using a STELLA™ platform.  First, major features  and components of  the model are described. Then, relationships between these components, which provide more insight into behaviour, are described through the use of  a Causal Loop Diagram. In this text, the term "demand" refers  to the volume of  water requested by a user group for consumptive or non-consumptive use. The magnitude of  demand is not necessarily equal to existing water rights, nor is it always the amount allocated. Modeled demands for  agricultural or residential diversions are based on current use patterns in absence of  conservation measures or any water shortage restrictions and are referred  to as "maximum demand." The maximum demand is not the maximum possible, but is simply the current trajectory based on normal year conditions. Instream demand and conservation targets are defined  by policies with fixed  monthly targets. When shortages occur, allocations will be less than maximum demand. Residential demand includes domestic and other municipal demands. Most out-of-stream water use in the Okanagan can be classified  as either agricultural or residential applications. Water to support non-domestic municipal use, such as watering of  parks or golf  courses, is either averaged into per capita residential use values or counted as agricultural use. Industrial use is very low in the region, and therefore  was not separated from  residential use in this study. The terms "municipal" and "urban" are less representative because of  the water allocation structure in the Okanagan: municipalities frequently  serve both residential and agricultural customers. 3.5.1 Components of the Okanagan Sustainable Water Resources Model 3.5.1.1 Spatial scales OSWRM describes nearly the entire basin, from  the northernmost extent to the mouth of Osoyoos Lake (see Figure 3.1). As most people work at the community or regional level, they are typically not aware of  whole-basin issues, or how their area interacts with the larger scale. A comprehensive study in the 1970's recommended basin-scale management of  the water resource (Canada-British Columbia Consultative Board 1974). Except for  the formation  of  the Okanagan Basin Water Board, which until recently has had limited scope and influence,  there has been little progress on realizing basin-scale management. Modeling the entire basin provides an avenue for  exploration and discussion of  the larger perspective. Participants suggested several spatial scales, from  modeling the basin as a whole, to describing all 60 watersheds. A third suggestion was to divide the basin into a few  regions based on topography and climate. This last idea was the foundation  for  dividing the basin into three major regions (Figure 3.1) according to water source type, which participants found  to be appropriate when shown at the following  workshop. The three sources are: all of the tributary watersheds to Okanagan Lake on which there are human controls (Uplands), Okanagan Lake as well as a few  small, unmanaged watersheds contributing to the lake (Valley), and all watersheds that contribute to the mainstem downstream of  the Okanagan Lake dam at Penticton (South End). These major sub-basins have areal extents of  5200, 800, and 1500 km2 respectively and have distinct climates, topography, and water use patterns. Feedback between these sub-basins is minor, limited to some water cycling by return flows. Otherwise, the relationship between these areas is defined  by water that flows  through from the Uplands, to Okanagan Lake, and finally  into the South End. This paper describes results for  water supply and use from  the Uplands. The Uplands region comprises 70 percent of  the total land area modeled, so results for  the Uplands dominate in an aggregation of  results for  the basin. Also, because water is used multiple times through the basin, an analysis of  the Uplands provides a clear and accurate picture of  the relative magnitudes of  instream and out-of-stream  demands. 3.5.1.2 Time scales Simulations use monthly timesteps in thirty-year blocks of  either a historic period (based on 1961-90 data) or one of  two future  periods (2010-2039 or 2040-2069). The data gap between 1990 and 2010 was a consequence of  our reliance on_data from  established climate models and previous work that predetermined our simulation periods. The future  periods, referred  to as the 2020's and 2050's by climate modelers, and the historic years were those used by researchers in the previous phases of  this project (see Merritt and Alila 2004; Taylor and Barton 2004). Monthly timesteps were chosen to capture the seasonal climate shifts  while maintaining simulation efficiency. 3.5.1.3 Hydrology Figure 3.3 summarizes how climate change information  was translated into hydrologic impacts that were directly relevant to the balance of  water resources and use in the OSWRM. Taylor and Barton (2004) used the delta method to downscale three global climate models (Hadley, CSIRO, and CGCM2) and two emissions scenarios (A2 - high growth in global greenhouse gas emissions; B2 - moderate growth in emissions (IPCC-WG3 2000)) using local temperature and precipitation data. Merritt and Alila (2004; Merritt et al. 2006) generated hydrologic streamflow  scenarios for  these six climate scenarios. Because all future scenarios are adjustments to the 1961-1990 historic climate data, the pattern is repeated in each time block (Figure 3.4) Included in OSWRM are three climate scenarios, referred  to as Hadley A2, CSIRO B2, and CGCM2 B2, for  the 2020's and 2050's time blocks. These scenarios were selected because they provided the widest range of  behaviour, and thus the widest range of  possible future  conditions among the scenarios that Taylor and Barton developed. Generally, future  climate scenarios predict an annual streamflow  hydrograph that has an earlier, more intense spring freshet  than in the historic record. In addition to surface  water sources, some Okanagan communities rely on groundwater and diversions from  two adjacent river basins. These were characterized in OSWRM. However, aquifer  studies have only begun recently, so neither the aquifer  volume nor the sustainable yield is known. As a result, the groundwater "aquifers"  in the model can provide information about the relative state, but not the absolute state of  the groundwater stock. OSWRM assumes that as the communities dependent on these sources increase in population, withdraws will increase proportionately. These contributions are a small portion of  the total managed supply, and we assume that maximum sustainable yield (groundwater) or legal limits (adjacent river diversions) will not be exceeded in the range of  values tested. However, if  this assumption is false,  then the results could slightly overestimate future  managed supply. Figure 3.3: Flow chart illustrating the progression from climate models and local records to supply and demand inputs to the Okanagan Sustainable Water Resources Model. (Sources: Cohen et al. 2004; Neilsen et al. 2001; Merritt and Alila 2006; Neale 2005; 2006). Global  Climate  Model's Had  CMS,  CGCM2,  CSIROMk2; A2 & B2, 2010-2099 Okanagan climate  stations Figure 3.4: Upland streamflow for historic and 2050's climate scenarios. Two years are shown: 1976-77 for the historic scenario, and 2055-56 for the 2050's scenarios. Simulation Month 3.5.1.4 Agricultural demand Agricultural water demand was based on Neilsen et al. (2004b). The model described in Neilsen et al. generated estimates for  crop water demand for  major water purveyors by relating demand to climatic and location-based factors.  In OSWRM we aggregated this output according to water source and normalized by area and crop type. Each water source region has a single average per land area irrigation demand profile  per crop and climate scenario. The normalization of  the data allowed us to create options for  users to simulate changes both in total land in production and in crop type mix. The values for  agricultural demand were derived by applying water delivery factors  on the crop water demand estimates. Nielsen (2004b) assumed that an additional 33 percent above crop water demand is required for  transporting water through the soil medium. Thus, irrigating with a rate that is 133 percent of  crop water demand is considered the minimum required to satisfy  crop needs. This rate is theoretically possible if  maximum efficiency  can be achieved through technologies like drip irrigation combined with irrigation scheduling. To estimate actual, current irrigation rates, an additional factor  of  30 percent was applied to account for  losses from  irrigation technologies such as overhead sprinklers and unlined ditches (van der Gulik and Stephens 2005). These factors  combine for  a total of  73 percent above crop water demand. 3.5.1.5 Residential demand Residential demand, based on work by Neale (2005; 2006), uses correlations of  temperature and outdoor water use, average residents per dwelling, proportions of  detached and multi-unit dwellings, as well as average savings realized by a number of  demand side management strategies (discussed in detail below). Data generated for  selected communities was extrapolated to OSWRM's regions. 3.5.1.6 Instream flow demand and conservation flow targets Instream flow  requirements are included in both the tributaries to Okanagan Lake and the mainstem lakes/river chain south of  Penticton. Because water is diverted out of  the tributary streams, we assume that instream flow  demands downstream cannot be satisfied  by water earmarked for  diversion. Instead, instream flow  demands are exclusive from  the out-of-stream demands. „ . In the tributaries to Okanagan Lake, conservation flow  targets defined  for  several streams as monthly percentages of  mean annual discharge (Northwest Hydraulic Consultants 2001) were extrapolated to all tributaries. The "normal" conservation flow  target is automatically modified  in dry years when not enough water is present in the system to satisfy  the target. "Normal" instream demand remains constant because it is based on established policy parameters; however, in practice, this target is modified  during droughts. 3.5.1.7 Adaptation and policy options A variety of  water conservation measures for  the agricultural and residential sectors are included, such as metering, xeriscaping, and technology upgrades. Policy options are provided for  drought management, enabling the user to select different  priorities for  water allocations. Some of  the policy options included on the basic user interface  include: • Implementing agricultural conservation and selecting a level of  efficiency • Implementing residential conservation strategies, including public education, xeriscaping, plumbing retrofit,  and metering. • Modifying  residential development patterns, including housing occupancy rate and the ratio of  apartments to multi-unit dwellings. • Modifying  sector allocation rules applied during water shortages. • Implementing a policy to satisfy  all  Upland water shortages with Okanagan Lake water Advanced options include increasing the capacity of  storage in the Uplands and adjusting the irrigated land area for  each crop type. A complete list of  adaptation and policy options is available in the model documentation (Table 10 of  Appendix D). 3.5.2 Dynamics of the system ( Here we describe the actual and modeled system through key linkages that define  the behaviour of  the aspects of  interest. Since our main objective is to explore the balance between supply and demands, we characterize the aspects that will increase or decrease the supply and/or the demand. Figure 3.5: Causal Loop Diagram of the Okanagan Basin water resources system. 3.5.2.1 The causal loop diagram One tool for  illustrating a complex system is a "Causal Loop Diagram" (CLD, Figure 3.5). CLD's are particularly useful  for  identifying  feedback  loops and for  clarifying  the factors  that control system behaviour. Since the purpose of  OSWRM was to gain a better understanding of  the water balance under a variety of  times and conditions, we chose "Water Deficit"  as the state variable to indicate the condition of  the system. "Water Deficit"  is directly influenced by "Water Available" and "Total Water Need." The arrows that connect these elements show the relationship, and the +/- signs indicate the direction of  influence.  For example, the positive link from  Total Water Need to Water Deficit  means that as Total Water Need increases, Water Deficit  will also increase. The negative link from  Water Available to Water Deficit  means that as the amount of  Water Available increases, the Water Deficit  decreases; there is an inverse relationship. Similarly, as the Water Deficit  increases, Water Use is forced to decrease. 3.5.2.2 Water deficit The Water Deficit  is the shortage in water relative to water demand, as expressed by the equation below. In the CLD, the Water Deficit  represents an aggregate for  the whole basin. The parameter is always zero or positive, as states of  water surplus are ignored. More severe deficit  conditions are represented by larger magnitudes. Water Deficit  = MAX (Maximum Water Demand - Managed Supply, 0) "Managed Supply" aggregates surface  water, groundwater, and water diverted from  adjacent river basins, and includes the delay created by reservoirs. "Maximum Water Demand" aggregates the basin's agricultural, residential, and ecological demands. Forest evapotranspiration is captured as land cover in the UBC watershed model, so is already subtracted from  streamflow. 3.5.2.3 Balancing feedback loops There are several balancing loops that work to alleviate Water Deficit  either by increasing supply or by reducing demand. Supply is increased through additional imports and/or groundwater pumping. In the actual system, we know basin residents have supplementary groundwater wells. However, these are unregulated and there is little information  on the magnitude, location, or frequency  of  use. It should be noted that, although groundwater pumping increases supply in the short term, it is probable that surface  water and groundwater are closely linked; therefore,  groundwater pumping may reduce the amount of  surface  water available over the longer-term. In OSWRM, only certain communities rely on these supplemental sources, and their contribution increases only as the populations in the communities increase. This balancing feedback  of  tapping supplementary supplies during drought is not captured. When deficit  is present, mechanisms exist for  reducing allocations to each of  the three use sectors. Decisions about prioritizing water use and thus, implementation of  these mechanisms, is decided at the local scale, often  by individual purveyors. Extended periods of water deficit  may encourage implementation of  conservation measures. 3.5.2.4 A reinforcing  feedback loop Water is reused multiple times on its journey between precipitating onto the ground and exiting to Osoyoos Lake. This phenomenon is captured by a weak reinforcing  loop. Water is returned to the system post treatment, or through irrigation returns. Several communities reclaim treated water from  residential sources and use it for  watering golf  courses and municipal parks. The increase in water available reduces the water deficit,  which allows for increased water use. Additional water consumption increases the volume of  water returned to the system. The reinforcing  strength of  this loop is highly limited by exit pathways, such as flows  downstream, and losses to evapotranspiration or to deep aquifers. 3.5.2.5 External drivers - climate and population Without external drivers, the system could achieve dynamic equilibrium. However, the external influences  of  a climate change and population growth disrupt the system. Climate change can affect  the water deficit  through multiple influence  points - decreased precipitation reduces streamflow,  and increased temperatures increase agricultural irrigation requirements and residential outdoor watering. In this analysis, we assume population is affected  only by factors  outside of  our system and that it will continue to increase over time. Therefore,  without significant  water reduction or conservation strategies, residential water demands will continue to increase. Residential growth rate projections used in this work are based on community and regional plans, as well as work by Neale (2006:2005; also see Appendix D). These rates are significantly  lower than the growth rates of  recent decades. Figure 3.6 compares the population projections for  the three growth rates defined  by the community plans (rapid, moderate, slow) and for  the rates from  recent history (1961-1990) referred  to as "continued trend." The historic growth rate was significantly  higher than projections, so the future estimates of  population growth may also be underestimated. For that reason, we emphasize the rapid growth scenario in the results presented in this paper. Figure 3.6: Population projections for the Uplands water users. Year - - -Continued Trend ———Rapid —— -Moderate— -Slow 3.6 Results The results presented in this paper all focus  on the Uplands portion of  the basin, defined primarily as the managed tributaries to Okanagan Lake and users of  this water source. Section 3.6.1 describes projections of  maximum future  demand compared with managed supply, at a number of  scales from  thirty-year aggregations to monthly averages. Section 3.6.2 shows feasible  allocations associated with the amount of  supply available in these future  scenarios. Finally, Section 3.6.3 discusses the role of  adaptation as a means of  making a smooth transition to the future  as described by these plausible scenarios. 3.6.1 Managed supply vs. maximum demand Figure 3.7 compares the total managed supply with the total maximum demand in the Uplands Region of  OSWRM from  the historic to the future  simulations. These results are aggregated and reported as annual averages, with one value for  each of  the three thirty-year simulation periods. Figure 3.7 presents both rapid (a) and slow (b) growth rates to show the sensitivity of  the system to population growth. The remaining figures  present only the rapid growth scenario unless otherwise noted. "Managed supply" combines stream flow  in the tributary streams with the supplemental supplies and return flows,  and includes timing adjustments from  the reservoirs. The "No Climate Change" scenario supply data lines show a slight increase over time. This increase can be fully  attributed to these supplemental sources and return flows,  which are dependent on population. Note that all three climate scenarios contain these minor increases in supply, although they are superimposed by the more dominant decreasing trend that is a direct result of  the decrease in basin precipitation due to climate change. In Figure 3.7, "Total demand" includes the three major sectors: agricultural, residential, and conservation flows.  The maximum demand values are based on projections of  current use patterns and do not assume any increases in efficiency  such as implementation of conservation measures. Maximum demand is independent of  supply and may be greater than available supply, even in the historic period. In all of  these scenarios the conservation flow target remains constant through time, so changes in demand are all a result of  changes to residential and agricultural demands. Agricultural land under production and crop types are also constant; any increase in agricultural demand is due to climate change. Figure 3.7: Thirty-year annual averages of total managed supply and maximum demand from the Uplands, showing trends through time for multiple climate scenarios with (a) rapid population growth, and (b) slow population growth. Historic 2020's (a) Rapid Population Growth 2050's — O — No Climate Change Scenario Mgd Supply —515— No Climate Change Scenario Total Demand — * — Hadley A2 Mgd Supply — » H a d l e y A2 Total Demand • CGCM B2 Mgd Supply — a CGCM B2 Total Demand • CSIRO B2 Mgd Supply — e CSIRO B2 Total Demand 900 1 fc  850 E -3 800 •Q 3 O C 750 o | 700 v o> 650 S » 600 TO a 550 o co 500 —E3 \ \ ^^^  " " - ~x Historic 2020's (b) Slow Population Growth 2050's —£3— No Climate Change Scenario Mgd Supply — * — No Climate Change Scenario Total Demand — * Hadley A2 Mgd Supply Hadley A2 Total Demand -m CGCM B2 Mgd Supply -a CGCM B2 Total Demand - • CSIRO B2 Mgd Supply -e CSIRO B2 Total Demand In the historic period, the average managed supply in the Uplands exceeded the average maximum demand. All of  the future  scenarios in Figure 3.7 show decreases in supply and increases in demand over the long term. The CGCM B2 scenario does show an increase in supply in the 2020's, but the large decrease in supply in the 2050's still leads to a decrease overall. Average annual demand exceeds supply by the 2050's in the Hadley A2 and the CSIRO B2 scenarios in both the rapid and slow growth scenarios. The CGCM B2 scenario is the least severe, but still shows a smaller gap between supply and demand in the 2050's. 3.6.1.1 Annual variability The thirty-year annual averages shown in Figure 3.7 show the long-term trends, but conceal the presence of  shortages due to annual climate variability. Figure 3.8 and Table 3.1 show annual variability which reveals the magnitude and frequency  of  annual water shortages. The scatter plot (Figure 3.8) presents managed supply versus maximum demand for  each year of simulation from  the historic period through the 2050's for  a single climate scenario (Hadley A2). The dashed line represents the supply-demand equality, which is the approximate division between conditions of  deficit  or surplus. Points located above this threshold represent a deficit  in the annual water budget. In the thirty-year historic period, there are three years in deficit  (one out often).  By the 2050's, both the Hadley A2 and the CSIRO B2 scenarios show a deficit  frequency  of  about two out of  three years, whether population growth is slow or rapid. CGCM B2 is less severe/ Table 3.1 summarizes the years in deficit for  each of  the scenarios and time periods. By the 2050's period, the climate change scenarios estimate shortages occurring every 14 to 22 years out of  30 if  rapid population growth occurs. Slow population growth has little effect^  with shortages still occurring every 11 to 21 years out of  30. The "moderate adaptation portfolio"  includes a selection of strategies which have already been implemented, or are being considered, in one or more communities within the Okanagan Basin, and simulates their implementation at the basin-wide scale (see Section 3.6.3). Figure 3.8: Annual total managed supply and total maximum demand for the Hadley A2 climate scenario and rapid population growth among Uplands water users. 1000 900 ra CD i 800 _ 8 700 f E o 2 600 3 o c o = 500 400 300 A A AA A A A A A »• x * • • historic » Hadley A2 2020 A Hadley A2 2050 Cemand = Supply 300 400 500 600 700 800 900 1000 1100 1200 Managed Supply [Million cubic metres per year] Table 3.1: Summary of deficit years as defined on the scatter plot (Figure 3.8) for all climate scenarios, showing (a) Rapid population growth and (b) Slow population growth. (a) Rapid Population Growth Scenarios No. of years (out of 30) where demand equals or exceeds supply Historic 2020's 2050's No CC 3 6 10 Hadley A2 11 22 CGCM B2 8 14 CSIRO B2 14 21 Hadley A2 Mod Adapt — 9 19 (b) Slow population growth scenarios No. of years (out of 30) where demand equals or exceeds supply Historic 2020's 2050's No CC 3 5 6 Hadley A2 9 19 CGCM B2 7 11 CSIRO B2 — 14 21 3.6.1.2 Intra-annual variability Water supply and demand in the Okanagan are unequally distributed through the year, so some of  the years that are not in deficit  overall may still experience summer shortages. Figure 3.9 shows supply and demand by month, including a breakdown of  the three major demand sectors. Managed water supply currently peaks in March as a result of  the spring freshet.  Total demand also peaks in March, but out-of-stream  demands peak in July and August. Instream conservation flow  targets roughly follow  the historic natural pattern of supply, and the monthly targets are held constant each year. In all of  the future  climate change scenarios, the spring freshet  occurs slightly earlier in the 2020's and 2050's. Managed supply reflects  this, as is shown by the increases in April supply through time (Figure 3.9). Because the conservation flow  targets are based on the historic peak flow,  a slight offset  in timing of  peak flow  emerges. Future residential and agricultural demands increase through the irrigation months (March - October). By the 2020's, the thirty-year averages of  demand in July, August, and September exceed the system's capacity to allocate supply. By the 2050's, the deficit  during these months is exacerbated and extends from  June through October. * In Figure 3.9, conservation demand remains constant through the future  periods (because it is defined  by policy). Agricultural demand increases with climate change, and residential demand, which historically was rather minor, becomes a more notable, although still small, portion of  the profile  by the 2050's with rapid population growth. Figure 3.9: Thirty-year average monthly managed Uplands supply and maximum demand profiles with demand from the three major sectors revealed. 250 225 ( a ) Historic Period ^ m UL Resid Demand CZ3 UL Ag Demand f*w i Normal Cons Target -o— Managed Supply 250 M M O N D 250 225 S 200 o b 175 a> Q. 150 R b E 12b o 3 100 3 C 75 o I 50 25 0 250 225 5 200 0 t 175 2L 160 E 125 o n 100 3 C 75 0 I 50 25 (b) Hadley A2 2020's with rapid population growth UL Resid Demand UL Ag Demand Normal Cons Target Managed Supply M M O N (c) Hadley A2 2050's with rapid population growth UL Resid Demand UL Ag Demand Normal Cons Target Managed Supply 3.6.2 Maximum demand versus total allocation The volume of  water that can be allocated to meet demands is limited by the amount of managed supply available each month. When water shortages occur, water allocations are determined by drought policies and management decisions. The graph in Figure 3.10 shows maximum demand and total allocation over the thirty-year simulation period for  several scenarios. OSWRM allocates water to the three sectors based on interpretations of  the current drought policies and management practices that were described by the local stakeholders that participated in the model building sessions. For example, residential outdoor watering restrictions are standard practice in the region, so it is the first  sector to be cut. On average, percent reductions across sectors are similar, with a slight priority granted for  conservation flows  and slightly greater reductions in the residential sector. In the future  climate change scenarios, both agricultural and residential demand levels during the summer increase. At the same time, supplies are generally decreasing overall. Spring melt occurs earlier, and summers are drier, which makes meeting summer and fall  demand even more challenging. Critical months by the 2050's extend from  June through October, with August becoming the most severe. The difference  between the allocation curves and the demand curves in Figure 3.10 shows how much demand can be satisfied.  These are expressed as percentages for  both annual totals and August values in Table 3.2. Table 3.2 provides a summary of  the percentage of  demand met (through allocations) both as an annual summary and for  August only. In the historic simulation, 98 percent of  annual demand was satisfied,  while 95 percent of  August demand was satisfied.  All of  the future  scenarios show reduced capacity of  the system to meet demand. Figure 3.10: Thirty-year average annual summary comparing total demand (all three sectors) and total water allocated in the Uplands for the rapid population growth scenarios. - -S3" - No Climate Change Scenario Total Demand — J t ! — No Climate Change Scenario Total Alloc — H a d l e y A2 Total Demand Hadley A2 Total Alloc CGCM B2 Total Demand - a CGCM B2 Total Alloc - • — CSIRO B2 Total Demand -e CSIRO B2 Total Alloc - Had A2 Mod Adapt Total Demand - Had A2 Mod Adapt Total Alloc Historic 2020's 2050's Table 3.2: Allocations as a percent of demand, shown as annual totals and for August, the month with the greatest deficit in the future scenarios. Percent of Demand Met (Allocations/Demand) Annual Totals August Only 2020's 2050's 2020's 2050's CGCM B2 93% 82% 84% 59% Had A2 90% 74% 79% 50% CSIRO B2 76% 72% 49% 45% Had A2 Mod Adapt 86% 78% 81% 55% 3.6.3 Deficit and adaptation Reductions in demand can be forced  during dry years, which will occur with increasing frequency  and severity, or they can be voluntary and anticipatory, through the use of anticipatory conservation strategies. Various conservation and adaptation strategies that reduce consumption could decrease the total demand as shown, and may help lessen the frequency  and severity of  forced  allocation reductions. The scenarios described thus far  include several assumptions. First of  all, the future "maximum" demand assumes that no additional conservation strategies will be implemented. Furthermore, the scenarios assume agricultural demand will continue to increase as a function  of  crop water demand without regard to water rights limitations. Residential water demand will continue to increase as a function  of  the exponentially increasing population. Therefore,  conservation measures and/or regard for  legal limits may reduce the severity and frequency  of  deficit. Figure 3.10 and Table 3.2 include the results of  a "moderate adaptation scenario." This scenario takes some of  the current adaptation trends in the region, and extends them to the entire basin. Residential demand management includes public education and metering with charges by increasing block rate. Combined, the strategies may slow residential demand by about 40 percent (Neale 2006). The moderate adaptation scenario also includes a reduction in all agricultural demand, by six percent. (A six percent reduction in total agricultural demand results from  an increase in efficiency  of  25 percent of  the maximum reduction theoretically possible while maintaining current states of  irrigated area and crop mix. The conversion is explained in Section 3.5.1.4.) These reductions are not from  current levels, but are reductions from  the future  scenarios without adaptation, as shown in Figure 3.10. Note that future  average allocations are also slightly reduced in the adaptation scenarios as compared with allocations in the associated (Hadley A2) no adaptation scenarios. The decrease in average allocation is partly due to the decrease in managed supply and partly to the decrease in demand.. The decrease in managed supply is due partly to the decrease in return flows  from residential indoor use and partly from  storage release adjustments. The wastewater return flows  are significant,  so as residents use less, the return flows  are much less. The remaining supplementary sources are minor. As a result, the reliability of  meeting demand in the future does not increase significantly.  In fact,  it appears to drop slightly in the 2020's, but shows a slight improvement from  the no adaptation scenario in the 2050's. Alternate ways of  allocating water to the three main sectors are nearly limitless. OSWRM provides an opportunity for  users to explore this feasibility  space through numerous optional settings, as described in Section 3.5.1.7. Cohen and Langsdale (2006) show simulation results from  a number of  adaptation scenarios, providing some insight into the effectiveness of  a range of  options available. It is theoretically possible to define  this feasible  region (i.e. the range of  possible combinations of  adaptation measures that would produce satisfactory results); however, the boundaries of  the region are subjective, dependent on perceptions of risk, changing water technology (e.g. irrigation), and future  development choices. Therefore, defining  boundaries would be quite challenging. This task is beyond the scope of  this project, but is recommended for  future  work. The purpose of  this modeling initiative, couched in a participatory process, was not to identify  "the" solution, but to guide the local professional water community in exploring the effectiveness  of  available policies to support desired future conditions such as increased reliability of  the water resource. 3.6.4 Implications for future management All future  climate change scenarios, from  2010 through 2069, show a significant  decrease in water,supply from  the 1961-1990 condition. This decrease may be slightly offset  by additional groundwater pumping and diversions from  adjacent river systems, however, the limitations of  these sources are currently unknown. Simultaneously, out-of-stream (agricultural and residential) demands are projected to increase significantly.  Residential water demand is more sensitive to population growth than to climate change, though climate change does accelerate the effect  of  an increasing population. In contrast, the area of agricultural land in production is quite stable in the Okanagan, but crop water demand is highly sensitive to changes in climate, so irrigation may intensify.  Conservation flows  are policy-based, and are assumed to remain constant throughout the simulation time period. All of  the climate scenarios show that long-term average allocations may remain close to the levels in the historic simulation, even though the average supply is decreasing. This is because the allocation decreases in the increasingly frequent  dry years are offset  by the allocations increases to match the greater future  demand in wet years. Although the magnitude of  allocations remains relatively constant, the reliability of  the water supply to meet demand decreases from  a historic rate of  98 percent to a range of  72 to 82 percent in the 2050's. Most of  this future  deficit  is due to the impacts of  climate change. Population growth contributed to only a small portion of  this reduced reliability, as is evidenced by the "no climate change" scenario. Future simulations without climate change result in allocations that are 94 percent of  maximum demand in the 2050's. Satisfying  demand during the dry season, when irrigation demand peaks, will become increasingly difficult.  August may be the most challenging month, with allocations reduced from  95 percent in the historic simulation to between 45 and 59 percent of  demand in the 2050's. Conservation measures may reduce this deficit,  however, the "moderate adaptation scenario" which incorporates some of  the adaptation measures currently being implemented in the region and extrapolates them for  the whole region, does not significantly  reduce the deficit.  The potential effectiveness  of  the conservation measures may be slightly dampened by the reduction of  return flows  from  residential indoor use. Indoor water use is, in effect,  not a consumptive use. Stricter conservation measures, limiting future  residential development, and limiting increases in agricultural demand, may be required to prevent future  water conflicts.  The agricultural sector may be forced  to implement efficient  irrigation technologies, change crop types, and/or reduce land under production. As a complement to conservation measures, expansion of  supplies may play a role. Cohen and Langsdale (2006) showed that expanded use of  Okanagan Lake may be feasible  if  care is taken to avoid depleting this resource. Expanding groundwater use has not been fully  explored. Caution should be applied with expansion of  either groundwater or Okanagan Lake as they are not new resources, but are hydrologically connected to current sources. 3.7 Conclusions The OSWRM is a highly aggregated scoping model intended for  the purpose of  exploring a variety of  future  scenarios. Analyzing the results of  these scenarios helped to illuminate dominant and controlling system characteristics, such as feedback  loops. Identifying parameters to which the system is sensitive also helps to select priorities for  future  research. For example, reductions in residential water use caused a notable decrease in wastewater ) return flows  (from  residential indoor water use). Refining  the portion of  residents on sewer, and the portion of  water which is returned may increase the accuracy of  the model results. Although the model was supported by numerical data, the focus  of  the modeling exercise was on identifying  the important causal relationships (qualitative characterization) rather than on providing a fully  calibrated and validated quantitative assessment. Therefore,  when reviewing the results it is more important to focus  on the general trends of  the future scenarios than on specific  values. OSWRM is useful  for  quickly testing a number of  different climate change, population growth, and adaptation scenarios, however, when the Okanagan community is ready to move forward  towards design and policy-setting, then more detailed and rigorous studies are recommended. The results presented in this paper focus  on the Uplands portion of  the basin for  clarity of presentation. At the outset of  this work, the common belief  was that the dry southern portion of  the basin, downstream from  more than half  of  the basin's population, would be most vulnerable to all stressors on the water resources. Intuitively, one would expect that water shortages in the upstream end of  the basin would increase in severity as you move downstream. However, to date, watersheds in the Upland tributaries have proven to be most sensitive to drought. The 2003 drought caused severe conflict  and resulted in the development of  an operating agreement in the Trout Creek watershed, while the South End felt  little or no impact. One reason for  the lack of  sensitivity is that the South End community's two main water sources are quite buffered  from  climate variation. Surface water supplies used for  irrigation are managed by a large extent through operation of  the dam on Okanagan Lake, and groundwater used for  domestic purposes is typically a stable resource, not immediately impacted by drought. However, the question remains whether long-term strains in the Upland region will eventually trickle down to the South End. Current research on the characterization of  the groundwater aquifers  will provide some clues to this puzzle. Ultimately, it will be the decisions that the residents and water managers make that will have significant  influence  over the future  of  water resources in the Okanagan Basin. 3.8 References BC Stats (2006). "Sub-provincial population estimates." Available at: http://www.bcstats.gov.bc.ca/data/pop/pop/estspop.asp. Borden, C. and B.G.J. Spinazola (2006). Evaluation of  Diversions Operation Plans to Meet Minimum Fish Flow Requirements Using MIKE BASIN Model Simulations in Lemhi River Basin. Adaptive Management of  Water Resources. Proceedings of  the American Water Resources Association's 2006 Summer Specialty Conference., Missoula, Montana, AWRA. Borden, C., et al. (2006). Water Supply Assessment Tool for  the San Francisco Public Utility Commission. Adaptive Management of  Water Resources. Proceedings of  the American Water Resources Association's 2006 Summer Specialty Conference., Missoula, Montana, AWRA. Canada-British Columbia Consultative Board. (1974). Okanagan Basin Agreement: Summary Report of  the Consultative Board including the Comprehensive Framework Plan. Cardwell, H., B. Faber, and K. Spading. (2004). Collaborative Models for  Planning in the Mississippi Headwaters. Proceedings of  the 2004 World Water and Environmental Resources Congress, Salt Lake City, UT, American Society of  Civil Engineers. Cohen, S. and T. Kulkarni, eds. (2001). Water Management & Climate Change in the Okanagan Basin. Vancouver, Environment Canada & University of  British Columbia. Cohen, S. and S. Langsdale. (2006). Appendix E: Model Output, in S. Cohen and T. Neale (eds.) Participatory Integrated Assessment of  Water Management and Climate Change in the Okanagan, British Columbia, Canada: Final Report, April 1, 2004 to June 30, 2006. Vancouver, Environment Canada & UBC, p. 195-221. Cohen, S., et al. (2006). "Learning with Local Help: Expanding the Dialogue on Climate Change and Water Management in the Okanagan Region, British Columbia, Canada." Climatic Change 75: 331-358. Cohen, S., D. Neilsen, and R. Welbourn, eds. (2004). Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report, January 1, 2002-June 30, 2004. Vancouver, Environment Canada, Agriculture and Agri-Food Canada & University of  British Columbia. / Costanza, R. and M. Ruth (1998). "Using Dynamic Modeling to Scope Environmental Problems and Build Consensus." Environmental Management 22(2): 183-195. de Loe, R. C. and R. D. Kreutzwiser (2000). "Climate Variability, Climate Change and Water Resource Management in the Great Lakes." Climatic Change 45: 163-179. Filmon, G. and Review Team (2004). Firestorm 2003 Provincial Review. Victoria, Government of  British Columbia. Fisher, R. and W. Ury (1981). Getting to Yes: Negotiating Agreement Without Giving In. Boston, Houghton Mifflin. Forrester, J. W. (1987). "Lessons from  system dynamics modeling." System Dynamics Review 3(2): 136-149. Ghabayen, S. and M. McKee. 2006. A Bayesian Belief  Network Model for  Managing the Uncertainty in Gaza Water and Wastewater Systems. Adaptive Management of  Water Resources: AWRA Summer Specialty Conference,  American Water Resources Association, Missoula, Montana. Hare, M., et al. (2003). "Participatory Modelling in Natural Resources Management: A Comparison of  Four Case Studies." Integrated Assessment 4(2): 62-72. Hydrologies, I. (2003). What is Oasis with OCL? IPCC Working Group III (2000). Emissions Scenarios. A Special Report of  Working Group III of  the Intergovernmental Panel on Climate Change. Cambridge, U.K.: 599. Jenkins, M. W., G. F. Marques, and F. K. Lelo. (2005). WEAP as a participatory tool for shared vision planning in the River Njoro Watershed, Kenya. Impacts of  Global Climate Change: Proceedings of  the 2005 World Water & Environmental Resources Congress, Anchorage, Alaska, ASCE. Kashyap, ~A. (2004). "Water governance: learning by developing adaptive capacity to incorporate climate variability and change." Water Science and Technology 49(7): 141-146. Langsdale, S. Accepted.  Communication of  Climate Change Uncertainty to Stakeholders Using the Scenarios Approach. J.  of  Contemporary  Water  Research and  Education. Universities Council on Water Resources. Langsdale, S., et al. (2006). Ch. 5: Exploring Water Resources Futures with a System Dynamics Model. Participatory Integrated Assessment of  Water Management and Climate Change in the Okanagan Basin, British Columbia, Canada: Final Report. S. Cohen and T. Neale. Vancouver, UBC and Environment Canada. Letcher, R. A. and A. J. Jakeman (2003). "Application of  an Adaptive Method for  Integrated Assessment of  Water Allocation Issues in the Namoi River Catchment, Australia." Integrated Assessment 4(2): 73-89. Lins, H. F., D. M. Wolock, G. J. McCabe. (1997). "Scale and Modeling Issues in Water Resources Planning." Climatic Change 37: 63-88. Merritt, W. and Y. Alila (2004). Chapter 7: Hydrology, in Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report, January 1, 2002-June 30, 2004. S. Cohen, D. Neilsen and R. Welbourn. Vancouver, Environment Canada, Agriculture and Agri-Food Canada & University of British Columbia: 63-88. Merritt, W. and Alila (2003). Chapter 3: Hydrology, in S. Cohen and T. Neale (eds.) Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Interim Report, January 1, 2002 - March 31, 2003. Vancouver, Environment Canada & University of  British Columbia: 27-57. Merritt, W. S., et al. (2006). "Hydrologic response to scenarios of  climate change in the Okanagan Basin, British Columbia." Journal of  Hydrology 326: 79-108. Mortsch, L. and B. Mills (1996). Great Lakes-St Lawrence Basin Project on Adapting to the Impacts of  Climate Change and Variability. Downsview, Ontario, Environmental > Adaptation Research Group, Atmospheric Environment Service, Environment Canada. Neale, T. (2005). Impacts of  Climate Change and Population Growth on Residential Water Demand in the Okanagan Basin, British Columbia. School of  Environment and Sustainability. Victoria, B.C., Royal Roads University. Neale, T. (2006). Ch 2: Urban Water Futures: Exploring Development, Management and Climate Change Impacts on Urban Water Demand. Participatory Integrated Assessment of  Water Management and Climate Change in the Okanagan, British Columbia, Canada: Final Report. S. Cohen and T. Neale. Vancouver, Environment Canada & UBC. Neilsen, D., et al. (2004). Chapter 8: Crop Water Demand Scenarios for  the Okanagan Basin, in S. Cohen, D. Neilsen and R. Welbourn (eds). Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report, January 1? 2002-June 30, 2004.. Vancouver, Environment Canada, Agriculture and Agri-Food Canada & University of  British Columbia: 89-114. Northwest Hydraulic Consultants. (2001). Hydrology, Water Use and Conservation Flows for  Kokanee Salmon and Rainbow Trout in the Okanagan Lake Basin, B.C. Victoria, BC Fisheries. Palmer, R. N., et al. (in review). "Disciplined Planning, Structured Participation, and Collaborative Modeling: Applying Shared Vision Planning." Journal of  Water Resources Planning and Management. Quick, M. C. (1995). The UBC Watershed Model. Computer Models of  Watershed Hydrology. V.P. Singh. Highlands Ranch, CO, Water Resources Publications: 233-280. Smith, J. B., S. Huq, et al., Eds. 1996. Vulnerability and Adaptation to Climate Change: Interim Results from  the U.S. Country Studies Program. Kluwer Academic, Dordrecht. Stakhiv, E. Z. (1996). Managing Water Resources for  Climate Change Adaptation. Adapting to Climate Change, edited by J. B. Smith and others. New York, Springer-Verlag: 243-264. Stave, K. A. (2002). "Using system dynamics to improve public participation in environmental decisions." System Dynamics Review 18(2): 139-167. Summit Environmental Consultants. (2004). Trepanier Landscape Unit Water Management Plan, Volume 1. Vernon, Regional District of  Central Okanagan and B.C. Ministry of Sustainable Resource Management. Taylor, B. and M. Barton (2004). Chapter 5: Climate Change Scenarios, in S. Cohen, D. Neilsen and R. Welbourn (eds). Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report, January 1, 2002-June 30, 2004. Vancouver, Environment Canada, Agriculture and Agri-Food Canada and University of  British Columbia: 47-54. Tidwell, V. C., et al. (2004). "System dynamics modeling for  community-based water planning: Application to the Middle Rio Grande." Aquatic Science 66: 357-372. van Asselt, M. B. A. and N. Rijkens-Klomp (2002). "A look in the mirror: reflection  on participation in Integrated Assessment from  a methodological perspective." Global Environmental Change 12: 167-184. van den Belt, M., L. Deutsch, and A. Jansson. (1998). "A consensus-based simulation model for  management in the Patagonia coastal zone." Ecological Modeling 110: 79-103. van der Gulik, T. and K. Stephens (2005). Personal communication. December 5, 2005. Vancouver, BC. Vennix, J. A. M. (1996). Group Model Building: Facilitating Team Learning Using System Dynamics. Chichester, John Wiley. Viessman, W. and T. D. Feather, eds. (2006). State Water Resources Planning in the United States, American Society of  Civil Engineers. CHAPTER 4: CONCLUDING REMARKS This dissertation describes a participatory modeling project which explores the issue of climate change related to water resources in the Okanagan Basin, British Columbia, Canada. Chapter 1 begins with a description of  how the water resources planning and management field  is responding to climate change, including challenges and response options. The chapter continues with a review of  current literature in the fields  of  integrated assessment, participatory planning, system dynamics, and the combination of  these fields  which forms participatory modeling. Finally, the Okanagan Basin case study is introduced. Chapter 2 describes the participatory processes applied in the Okanagan case study as well as the response of  the participants, while Chapter 3 describes the system dynamics model that resulted from  the process and presents an analysis of  model simulated results for  multiple future  scenarios. This chapter provides an overview of  the work, and discusses objectives, results and conclusions. I conclude with possibilities for  continuing research in this project, for  new applications of  this methodology, and for  advancing the methodology. This project shows that participatory modeling can be an effective  tool for  creating and structuring dialogue on the issue of  climate change in natural resource planning and management. Gathering a diverse group of  stakeholders with different  perspectives and experiences brings to the table a breadth of  ideas necessary to have informative  dialogue, while development of  and interaction with the model provides a common language with which to discuss and characterize the complex system as described by participants. These advantages work toward keeping the dialogue focused  on the issue at hand. Exploring model output helps participants to question their own understanding of  the system, which fosters learning about processes and relationships, rather than just facts.  This more sophisticated type of  knowledge is a prerequisite for  developing effective  policies to manage the system. Although it was not measured in this project, this form  of  dialogue among diverse stakeholders, supported by an interactive model, may inspire creativity surrounding solutions to the problem. The model results show that climate change may, in fact,  significantly  impact the Okanagan's water resource balance by increasing the frequency  and magnitude of  water shortages. The scenarios show that the impacts of  climate change on the system are significantly  more intense than the impacts due to anticipated levels of  population growth. The combined impacts from  climate change and population growth create water deficit conditions that exceed historic conditions. Anticipatory adaptation measures have the potential to reduce the intensity of  future  water shortages. A continuation of  model-supported dialogue with the Okanagan's water resource stakeholders could assist in the evaluation of management alternatives and decision making. Published case studies (Cardwell et al. 2004; Cockerill et al. 2006; Connor et al. 2004; Costanza and Ruth 1998; Jenkins et al. 2005; Jones et al. 2002; Keyes and Palmer 1993; Kirshen et al. 2004; Letcher and Jakeman 2003; Palmer et al. 1993; Peterson et al. 2004; van den Belt et al. 1998; van den Belt 2004) show that the participatory modeling approach is an effective  tool for  engaging communities in dialogue about managing natural resources applications. The present work shows that the method is effective  for  engaging participants in a policy dialogue and an interactive learning experience on the topic of  climate change. 4.1 Project Summary: Objectives & Results The two major components of  this thesis were to conduct a participatory process and to develop a system dynamics model. The participatory process had three primary objectives: (1) To create a shared learning experience among the participants and the research team; (2) To tailor the model to suit the needs of  the community; and (3) To foster  a sense of ownership of  the model among the participants. The purpose of  constructing the model was to explore water resources futures  in the Okanagan Basin, with specific  attention to the influence  of  climate change and population growth, and to the role of  adaptation strategies on maintaining reliability in the water resources system. 4.1.1 The participatory process Participatory modeling was selected as the methodological approach because it is an effective tool for  both characterizing the system and communicating the information  to relevant parties. The expert stakeholders who participated in the modeling process were the best resource for  certain numerical data. I could have acquired this information  from  them regardless of  their participation in the process. However, the benefit  of  their active involvement was that it created the opportunity for  all participants to share their perspectives and knowledge of  the water resources system. The workshop activities were designed to draw out this knowledge and educate the modeling team on the important components that needed to be characterized to assess the current and future  state of  the region's water resources and management options. These activities included generating system diagrams and a historical timeline, discussing relevant issues, and providing feedback  on the model. As a result, the modellers created a richer description of  the system than they would have in absence of  the participatory advisors. In addition, the interaction with the local community provided a more rewarding experience for  the research team. The effectiveness  of  the participatory process in meeting our objectives was measured through surveys in the final  two workshops. Gathering a diverse group of  stakeholders together and providing opportunities to share their perspectives helped to broaden people's understanding beyond their own experiences. Furthermore, by developing and interacting with the model, the participants clarified  their understanding of  the linkages and processes that define  the water resource system. In discussing the model, we emphasized causal relationships rather than facts,  to support second-order learning (Folz and Hill 2001; van de Kerkhof  2004). Participants found  both the process and the resulting model valuable. When asked if  their perceptions of  future  water availability in the basin changed due to the exercise, the majority of  respondents selected a middle response ("some change"), while the median value was slightly lower. Those with the least knowledge about the water resource system experienced the most learning. Several participants noted a new appreciation for  the complexity of  the system and in managing it effectively. Information  and feedback  gathered from  the participants throughout the workshop series supported the development of  a model that was relevant to those who manage the region's water resources. All participants surveyed at the final  two workshops responded that they felt the model is, or has potential to be, a legitimate and relevant tool to explore long term water management in the Okanagan. The majority of  respondents described parameters for effective  use of  the model; that it would be useful  with the appropriate audience or after additional refinements  were made. Several people commented that it was a good communication and/or education tool for  visualizing the future  and discussing policy. Because the model had become rather complex by the fifth  and final  workshop, and because half  of  the attendees had not attended previous events, they spent considerable time familiarizing  themselves with the model. This illuminates the fact  that the investment of  time and resources required to bring newcomers up to speed makes sharing attendance responsibilities with colleagues less than ideal. More structured guidance would have helped lead the participants through specific  explorations in the time allotted. Simplifying  the model's user interface  is recommended for  increasing the efficiency  and effectiveness  of  any future  model exploration sessions. In the early stages of  designing the participatory modeling process, I wanted the participants to be as active as possible in the development of  the model. For practical reasons, the participants did not build the model directly, but they did contribute substantial information that supported construction of  the model. By the final  two workshops, participants felt,  on average, that they were familiar  with the issues included in the model. The opportunity for participation provided by the six sessions over a twelve-month period, combined with the low rate of  continuity in participation, may not have been sufficient  for  the participants to feel  that it was "their" model. However, it was not practical to have the participants constructing the computer version of  the model. The limitations on time and financial resources, as well as the interest and skills of  the participants, defined  the scope of participation in our final  design of  the sessions. 4.1.2 The Okanagan Sustainable Water Resources Model (OSWRM) The OSWRM is a highly aggregated scoping model intended for  the purpose of  exploring a variety of  future  scenarios. The modeling exercise focused  more on qualitative characterization than on quantitative calibration, so when reviewing results should also focus more on the general trends of  the future  scenarios than on specific  values. The model is useful  for  quickly testing a number of  different  climate change, population growth, and adaptation scenarios; but not appropriate for  policy or engineering design. This section highlights findings  from  the generation of  future  scenarios and describes implications for future  management. In the future  scenarios, water balance is shown to be more sensitive to climate change than population growth. Climate change may decrease supply while increasing water demand in multiple sectors. Figure 4.1 shows the simulated decreases in supply and increases in demand through time, for  the Hadley A2 climate scenario. Furthermore, climate change may exaggerate the offset  in timing between supply and demand peaks, as the spring freshet  will occur earlier in the year. This effect  will increase the challenge of  meeting demand during the summer irrigation season. An analysis of  model results compared total water demand (the amount that irrigators, residents, and the aquatic ecosystem would use if  they could) with total water allocation (the amount that could be provided to meet this demand). The volume of  water that can be ) allocated is limited by the amount of  managed supply available in each monthly timestep. When water shortages occur, water allocations are determined by drought policies and management decisions. The graph in Figure 4.2 shows maximum demand and total allocation over the thirty-year simulation period for  several scenarios. Note that in the future  climate change scenarios both agricultural and residential demand levels during the summer increase. At the same time, supplies are generally decreasing overall. Spring melt occurs earlier, and summers are drier, which makes meeting summer and fall  demand even more challenging. Critical months by the 2050's extend from  June through October, with August becoming the most severe. Figure 4.1: Annual total managed supply and total maximum demand for the Hadley A2 climate scenario and rapid population growth among Uplands water users. 1000 900 § > E fc m Q . _ S 700 S £> 0 0> H E E 3 JQ 1 3 x 1-5 ° = 500 4  600 300 A A A* A A £ A A A A A * 1 • / - S I • * • Hstoric a Hadley A2 2020 A Hadley A 2 2050 Demand = Supply 300 400 500 600 700 800 900 1000 1100 1200 Managed Supply [Million cubic metres per year] Figure 4.2: Thirty-year average annual summary comparing total demand (all three sectors) and total water allocated in the Uplands for the rapid population growth scenarios. 900 - - 53 - - No Climate Change Scenario Total Demand - - No Climate Change Scenario Total Alloc — Hadley A2 Total Demand - Hadley A2 Total Alloc — CGCM B2 Total Demand - CGCM B2 Total Alloc — CSIRO B2 Total Demand — e — - CSIRO B2 Total Alloc — & - Had A2 Mod Adapt Total Demand — • - Had A2 Mod Adapt Total Alloc Historic 2020's 2050's The difference  between the allocation curves and the demand curves in Figure 4.2 shows how much demand can be satisfied.  These are expressed as percentages for  both annual totals and August values in Table 4.1. Table 4.1 provides a summary of  the percentage of  demand met (through allocations) both as an annual summary and for  August only. In the historic simulation, 98 percent of  annual demand was satisfied,  while 95 percent of  August demand was satisfied.  All of  the future  scenarios show reduced capacity of  the system to meet demand. Table 4.1: Allocations as a percent of demand, shown as annual totals and for August, the month with the greatest deficit in the future scenarios. Percent of Demand Met (Allocations/Demand) Annua Totals August Only 2020's 2050's 2020's 2050's CGCM B2 93% 82% 84% 59% Had A2 90% 74% 79% 50% CSIRO B2 76% 72% 49% 45% Had A2 Mod Adapt 86% 78% 81% 55% All of  the climate scenarios show that long-term average allocations may remain close to the levels in the historic simulation. However, as the future  simulations show decreasing overall supply and increasing demand, the reliability of  the water supply to meet demand decreases from  a historic rate of  98 percent to a range of  72 to 82 percent in the 2050's. Most of  this future  deficit  is due to the impacts of  climate change. Population growth contributed to only a small portion of  this reduced reliability, as is evidenced by the "no climate change" scenario. Future simulations without climate change result in allocations that are 94 percent of maximum demand in the 2050's. Satisfying  demand during the dry season, when irrigation demand peaks, will become increasingly difficult.  August may be the worst month, with allocations reduced from  95 percent in the historic simulation, to between 45 and 59 percent of  demand in the 2050's. Chapter 3 presents a "moderate adaptation portfolio"  that includes a selection of  strategies which have already been implemented, or are being considered, in one or more communities within the Okanagan Basin, and simulates their implementation at the basin-wide scale. The adaptation portfolio  includes agricultural efficiency  improvements, residential metering, public education, as well as slight shifts  in the urban development patterns of  housing type and occupancy rate. These adaptations do reduce the magnitude of  deficit  in each simulation; however, they do not fully  compensate for  the climate change impacts. More proactive adaptation measures will be required to prevent conflict  or harm due to water deficit  in dry years in the future.  A wider array of  adaptation measures is presented in Cohen and Langsdale (2006). 4.2 Impact of Research Results Before  starting this process, work by Cohen et al (2001; 2004) had led the Okanagan water resource community to confront  what impact climate change could have on their snow-melt dominated system, and to start evaluating some of  the response options that they could take. This process brought the community further  along the path of  considering climate change in their water resources planning and management by providing a tool that can integrate both supply and demand scenarios. The active participation of  stakeholders allowed the community to learn from  one another and to contribute to the stories that the model tells us of current and future  conditions. The extremely positive response by the participants suggests that the experience will be remembered as they continue to play a role in the region's water management. The learning that took place, whether it changed their initial ideas, or reinforced  what they already knew, should give them more confidence  and motivation to include climate change in future  work. Throughout the process, participants recommended that this work be shared with a wider community, particularly to elected officials  and the public. The resulting model can be used to continue dialogue with the community regarding the evaluation and selection of appropriate adaptation strategies for  reducing the negative impacts of  climate change on their ability to manage water resources in the basin. The satisfaction  that was expressed by the participants with the process supports evidence that participatory modeling is an effective tool for  fostering  communication and supporting effective  consensus-building. The long-term significance  of  this participatory modeling process to policy development cannot be measured during the timeframe  of  this phase of  the project. Decisions are complex, and not based on single factors.  Furthermore, decisions are made by many more people than just those who participated in the model building exercise. Since I do not have a control group, I may never be able to measure exactly what changed as a result of  these efforts. 4.3 Lessons and Recommendations This section describes several lessons that I learned through this experience, and provides recommendations for  future  applications of  this methodology. Fostering commitment from  volunteers is always challenging. In the Okanagan Basin, there are currently many water planning initiatives competing for  the time and attention of  water-related professionals.  Furthermore, individuals have different  motivations for  attending, so may have varied levels of  interest at different  stages of  the process. Strongly encouraging consistent attendance and maintaining communication with participants regarding scheduling i may help to foster  more commitment to the process. More frequent,  shorter-duration meetings could be more effective  at keeping the participants informed  on model progress and at providing opportunities for  participation. This participatory modeling process occurred in five  full-day  workshops held over a twelve-month period. However, the desire for  more frequent  meetings must be balanced by logistical realities. A minimum amount of  time is required between meetings for  model revisions and event planning. The time that facilitators,  modelers and participants can give to the process must be considered. Also, financial  resources of  the project are often  a limiting factor. Providing model documentation to the participants before  or at each session would have been an asset. During the workshop sessions, participants raised many questions about model assumptions and requested model documentation for  their own reference. Facilitators should repeatedly inform  participants of  ground rules of  the modeling process to keep everyone "on the same page." Clearly articulating ground rules of  the process can also help invitees decide whether or not to participate. Some examples of  ground rules that could have been stated more frequently  in this project include: (a) the purpose of  the model and its intended use; (b) the expected outcome of  process; and (c) the intended model users. Evaluations are critical to understanding the impact of  the process. Although we did conduct a number of  evaluations along the way, a more structured format  that uses more precise measures of  actual learning or attitude change) would have more effectively  recorded participant learning, the effectiveness  of  the process, and individual experiences. Evaluations given at the beginning and the end of  a participatory modeling process could measure evidence of  learning about the characteristics of  the system. However, it is challenging to know a priori how and where learning will occur, making it difficult  to know what to measure at the outset. 4.4 Future Directions This section presents ideas for  continuing the participatory modeling work in the Okanagan case study region, for  expanding applications of  participatory modeling to other areas related to climate change, and for  advancing the participatory modeling field. 4.4.1 Potential for continued efforts  in the Okanagan The severity of  future  climate scenarios indicates that continued evaluation of  management options is advisable for  ensuring reliable water resources into the future  in the study area. Furthermore, the participants' attendance at the modeling events, and their enthusiasm for  the topic showed that there is sufficient  interest in the topic to support additional efforts.  Specific actions that could be taken to maintain stakeholder dialogue and outreach are to: (1) Continue engagement of  the participatory modeling group to provide further opportunity to evaluate adaptation strategies and discuss policy implications. Facilitation should guide the exploration process through a structured process such as decision analysis, and could result in the identification  of  feasible  adaptation portfolios. (2) Use the model as an educational tool among a wider audience, targeting policy makers and the general public. Dissemination of  knowledge through the community may encourage voluntary water conservation, encourage effective  water management policies, and foster  public support for  those policies. A number of  model refinements  would increase the effectiveness  of  the use of  the model in these stakeholder engagements. Suggested refinements  include the following  tasks: (1) Calibrate the Valley region of  the model. Add Okanagan Lake evaporation. (2) Calibrate the South End portion of  the model. A major obstacle to accurately modeling this region is characterizing the managed releases from  Okanagan Lake which are defined  by a complex set of  rules that help balance both quantity and quality demands of  multiple uses in the lake and in the river downstream. Some of these multiple uses include supporting aquatic species, providing flood  control management, and maintaining recreational access, all of  which must be managed within operational constraints. The coarse monthly time step, combined with the lack of  forecasting,  made calculating realistic lake releases challenging. Output from  the prospective Fish Water Management Tool (Alexander et al. 2005) for  the Hadley A2 2050's scenario, which will be available in spring 2007, could provide assistance on this issue. (3) Increase the level of  detail in the adaptation strategies. In particular, specific agricultural adaptation strategies need to be identified,  and their benefits  quantified. (4) Simplify  the user interface  by reducing the amount of  navigation required. Consider creating predefined  adaptation portfolios  to reduce the number of  choices the user must make. Provide a directory of  all adaptation options available. Reduce the number of  output graphs and ensure that the graphs are effective  and concise indicators of  the state of  the system. (5) Expand the model to include information  about the groundwater resource. Important linkages to be captured in the model are the rates of  surface  water - groundwater interaction, and the relationship ^between current groundwater use and sustainable yield of  the aquifer.  Current research initiatives, including the Groundwater Assessment in the Okanagan Basin led by Natural Resources Canada, and a Canadian Water Network project led by Dr. Diana Allen at Simon Fraser University is beginning to provide this information,  so this may be a feasible  project in the near future. (6) Add financial  costs of  implementing and operating adaptation measures, building on adaptation cost studies provided by Hrasko and McNeill (2006). (7) Add a spatial component to residential growth, such that limits of  development at chosen population densities are tested. (8) Incorporate the effects  of  land use changes (residential development, forestry,  etc.) on hydrology. 4.4.2 New applications for participatory modeling The participatory modeling approach is not limited to local resource management applications, but could support other aspects of  the climate change dialogue. The significantly  delayed and geographically-displaced feedbacks  make system dynamics models ideal tools for  characterizing and exploring the climate change problem. New climate-related applications could include working with local citizen groups to increase awareness of  climate change and to encourage behavioural changes. A participatory modeling process could help citizens understand how their life  choices affect  the global climate. At a larger scale, the methodology could assist in global climate policy-making, such as the Kyoto protocol and emissions-trading schemes. Pioneers in the system dynamics community have proven that the model is well-suited to characterizing global issues. Examples include the Limits to Growth studies by Meadows et al. (1972; 2004) which examined issues of  population, food,  industry, resources, and pollution; and a behavioural climate-economy model by Fiddaman (2002) that conducted a policy analysis and the role of  Kyoto. Given that a global climate model has already been developed, it may be feasible  to use this model as a foundation  for  a participatory process involving representatives of  climate negotiating parties. The model could serve as a means to validate or refute  various worldviews and test assumptions that negotiating parties make about policy impacts and effectiveness.  The potential benefits  of smoother negotiations and more effective,  global-scale policies could easily outweigh the investment costs. High stakes, global policies for  complex problems are being negotiated regardless, so an investment to ensure these policies are fair  and effective  is prudent. 4.4.3 Advancing the field Participatory modeling applications to natural resource planning and management are relatively recent, but are rapidly becoming more prevalent. The methodology developed from several different  research fields,  and to date, no research has been completed assessing the various tools and processes available and for  what case studies they are most applicable. "[I]n the face  of  the proliferation  and ad-hoc experimentation with participatory methods, there is still a lack of  accumulated knowledge on the strengths and weaknesses of  each tool in meeting the objectives set by the evolving policy context ... [Additional research on the comparison and complementarity between participatory modeling and other deliberative methods" is recommended (Videira 2005: vii-viii). To fill  this gap, an extensive analysis of case studies in the literature and an evaluation of  the strengths and weaknesses of  the range of  tools and approaches currently in use are required. Specific  tasks could include: (1) Creating a typology for  classifying  the case studies according to model and decision-making process; (2) Critically comparing simulation model platforms  applicable to facilitating stakeholder involvement in water resources planning and management and their applicability to case study typologies; (3) Critically comparing participatory decision-making processes and their applicability to case study typologies; (4) Assessing the compatibility of  the models and the decision-making processes. This work would help to unify  the now seemingly disjointed projects in existence, provide guidance to practitioners regarding model and process selection, and identify  knowledge gaps and areas of  future  research. As clearly shown through this research, the participatory modeling methodology offers  to increase our capacity to explore complex problems and to prepare for  uncertain futures. Didactic investigation of  this field  will increase our ability to train others to conduct these processes. As the approach spreads to a wider array of  applications and to new audiences, society will be better equipped to collaboratively negotiate local solutions for  the global problems looming before  us. 4.5 References Alexander, C. A. D., B. Symonds, et al. (2005). The Okanagan Fish/Water Management Tool (v. 1.0.001): Guidelines for  Apprentice Water Managers. Kamloops, BC, Canadian Okanagan Basin Technical Working Group. Cardwell, H., B. Faber, and K. Spading.. (2004). Collaborative Models for  Planning in the Mississippi Headwaters. Proceedings of  the 2004 World Water and Environmental Resources Congress, Salt Lake City, UT, American Society of  Civil Engineers. Cockerill, K., H. Passell, and V. Tidwell. (2006). "Cooperative Modeling: Building Bridges Between Science and the Public." Journal of  the American Water Resources Association 42(2): 457-471. Cohen, S. and S. Langsdale (2006). Appendix E: Model Output, in: S. Cohen and T. Neale (eds.). Participatory Integrated Assessment of  Water Management and Climate Change in the Okanagan Basin, British Columbia. Vancouver, Environment Canada and University of  British Columbia: 203-229. Connor, J., L. Cartwright, and K. Stephenson. (2004). Collaborative Water Supply Planning: A Shared Vision Approach for  the Rappahannock Basin in Virginia. Proceedings of the World Water and Environmental Resources Congress, Salt Lake City, UT. Costanza, R. and M. Ruth (1998). "Using Dynamic Modeling to Scope Environmental Problems and Build Consensus." Environmental Management 22(2): 183-195. Fiddaman, T. S. (2002). "Exploring policy options with a behavioral climate-economy model." System Dynamics Review 18(2): 243-267. Folz, H. and J. Hill (2001). Public Participation Processes: Progressive or Not? Vancouver, BC, UBC. Hrasko, B. and R. McNeill. (2006). Costs of  adaptation measures, in: S. Cohen and T. Neale (eds.). Participatory Integrated Assessment of  Water Management and Climate Change in the Okanagan Basin, British Columbia. Vancouver, Environment Canada and University of  British Columbia: 37-48. Jenkins, M. W., G. F. Marques, and F. K. Lelo. (2005). WEAP as a participatory tool for shared vision planning in the River Njoro Watershed, Kenya. Impacts of  Global Climate Change: Proceedings of  the 2005 World Water & Environmental Resources Congress, Anchorage, Alaska, ASCE. Jones, A., D. Seville, and D. Meadows. (2002). "Resource sustainability in commodity systems: the sawmill industry in the Northern Forest." System Dynamics Review 18(2): 171-204. Keyes, A. M. and R. N. Palmer (1993). The Role of  Object Oriented Simulation Models in the Drought Preparedness Studies. Water management in the '90s: proceedings of  the 20th anniversary conference,  Seattle, WA, American Society of  Civil Engineers. Kirshen, P., et al. (2004). Infrastructure  Systems, Services and Climate Change: Integrated Impacts and Response Strategies for  the Boston Metropolitan Area/ Climate's Long-term Impacts on Metro Boston (CLIMB) Final Report, Tufts  University, University of  Maryland, Boston University and Metropolitan Area Planning Council. Letcher, R. A. and A. J. Jakeman (2003). "Application of  an Adaptive Method for  Integrated Assessment of  Water Allocation Issues in the Namoi River Catchment, Australia." Integrated Assessment 4(2): 73-89. Meadows, D., et al. (1972). The Limits to Growth. New York, Universe Books. Meadows, D., et al. (2004). Limits to Growth: The 30-Year Update. White River Junction, VT, Chelsea Green. Palmer, R. N., A. M. Keyes, and S. Fisher. (1993). Empowering Stakeholders Through Simulation in Water Resources Planning. Water Management in the '90s: Proceedings of  the 20th anniversary conference  Water Res. Ping. & Mgt. Conference, Seattle, Washington, American Society of  Civil Engineers. Peterson, T. R., A. Kenimer, and W. E. Grant. (2004). Using Mediated Modeling to Facilitate Collaborative Learning Among Residents of  the San Antonio Watershed, Texas, U.S.A. Mediated Modeling: A System Dynamics Approach-to Environmental Consensus Building. M. van den Belt. Washington D.C., Island Press: 136-163. van de Kerkhof,  M. (2004). Debating Climate Change: A Study of  Stakeholder Participation in an Integrated Assessment of  Long-Term Climate Policy in the Netherlands. Utrecht, The Netherlands, Lemma Publishers. van den Belt, M. (2004). Mediated Modeling: A system dynamics approach to environmental consensus building. Washington DC, Island Press. van den Belt, M., L. Deutsch, and A. Jansson. (1998). "A consensus-based simulation model for  management in the Patagonia coastal zone." Ecological Modeling 110: 79-103. Videira, N. (2005). Stakeholder Participation in Environmental Decision-Making: The Role of  Participatory Modeling. Lisboa, Universidade Nova de Lisboa. APPENDICES Appendix A: Behavioural Research Ethics Board Certificate of Approval Appendix B: Workshop Evaluation Forms Workshop 5 Pre-Evaluation How many previous events (workshops in Feb, April, June and Dec 2005, and meetings in Sept/Oct 2005) have you attended (not including this event)? 0 1 2 3 4 5 What role do you play in water resource planning or management in the basin (water manager / provincial govt / elected official  / NGO / other)? How well do you understand the model's structure? No knowledge I know which issues Very well - 1 could of  model are included teach it to others 1 2 3 4 5 Do you feel  your participation in the sessions contributed to the development of  the model? Yes No Undecided a. If  yes, what ideas or data were you able to contribute? b. If  no, what minimized your participation in the development of  the model? Has your participation in the sessions increased your knowledge and understanding of the state of  water resources in the Okanagan? Yes No Undecided a. If  yes, what have you learned through these sessions (either procedural, collegial, or substantive)? Workshop 5 Post-Evaluation How many previous events (workshops in Feb, April, June 2005 and meetings in Sept/Oct 2005) have you attended (not including this event)? 0 1 2 3 4 5 What role do you play in water resource planning or management in the basin? How well do you understand the model's structure? No knowledge I know which issues Very well - 1 could of  model are included teach it to others 1 2 3 , 4 5 Has your participation in the sessions increased your knowledge and understanding of the state of  water resources in the Okanagan? a. If  yes, what have you learned through these sessions (either procedural, collegial, or substantive)? Have your perceptions of  future  water availability in the basin changed due to this exercise? No change Some change Major changes 1 2 3 4 5 a. Please explain. Do you feel  the model captures the major factors  that influence  water resources in the Okanagan, now and in the future? Not at all Some factors  All factors 1 2 3 4 5 a. Can you identify  important factors  that are missing? ^ Do you feel  this model is a legitimate and relevant tool to explore long-term water management in the Okanagan? Briefly  explain (use the back if  necessary). Appendix C: The Okanagan System Model: Quick Reference The Okanagan System Model: Quick Reference ><• u * THE OKANAGAN  SYSTEM MODEL if!-HH^MMM:i^ il:;i^ 1 i 1; 1:iV:i'!:" ' •' ' January 2006 1111^ ^ • l^fcS- • 9-- Ate^ ?3 s . ,'t,  ' Size'tfitMinJow  iv -lit  I  fit  T>ox '- ' '•-f' 1 , (7^ Background'^ 7]:V - [ Learn the System j tj j This  Is  Home!  • ~~ ''  Oult  } • ~ " ' ' - — » uxm-p&ii • • HD^S^aiia^ldicae.- 1 IGOanth." V3-.J ' . -»'. A Brief  Introduction  to  the  Okanagan  Water Management  & Climate  Change  System  Model Version  date:  January  2006 Stacy  Langsdale University  of  British  Columbia TABLE OF CONTENTS I. Purpose of the Model II. Methodology: Participatory Modeling & System Dynamics III. Major Data Sources IV. Navigating the Interface 1. Background 2. Learn the System 2. Review Simulated History 3. Explore the Future 3.1 Future Scenario Options 3.2 View Consequences 3.3 Test Policy Options I. PURPOSE OF THE MODEL The purpose of  this initiative is to assist the Okanagan community in fostering  sustainable water for  their future.  To do this, they need an aggregated picture of  the stressors that may play a role, as well as information  about the effectiveness  of  various strategies to reduce vulnerabilities in water resources system. In the Okanagan, climate change and population growth are the major stressors expected to negatively impact the balance of  supply and demand. Previous growth estimates were too conservative—the population of  the central Okanagan doubled in the 25-yr period from 1975 to 2000— and this rapid rate is likely to continue. Bringing climate change into the conversation carries additional challenges. People are uncomfortable  with the level of  un-certainty in climate change estimates. Furthermore, most work has been conducted at the global scale, with results being reported in terms of  long-term average changes in tempera-ture or in precipitation. Regardless of  the challenges, bringing climate change into the con-versation increases the opportunity to prepare for  the future. One mandate of  the Adaptation & Impacts Research Division of  Environment Canada is to encourage awareness among communities of  the importance of  adapting to climate change. One step in this process is translating the results of  global climate models into impacts that are relevant to local communities. Prior to the modeling phase, researchers at Environment Canada (B. Taylor, M. Barton) downscaled the results of  selected global climate models (GCMs) to the Okanagan Basin. Three models provided a range of  plausible impacts, mak-ing the range of  uncertainty more explicit. Next, these climate change scenarios (expressed as temperature and precipitation changes) were used to simulate changes in the hydrologic system as well as crop water and residential demand. Thus, the climate change information was translated into terms that were more meaningful  to the local water community. This information  became the foundation  of  the Okanagan Model. In creating this model, we wanted to provide information  about the current and future  sup-ply-demand balance of  the Okanagan's water resources, and to allow water professionals and related interests to explore the effectiveness  of  an array of  alternative management op-tions, but also to learn why they are more effective,  through increased understanding of  the system. II. METHODOLOGY: PARTICIPATORY MODELING & SYSTEM DYNAMICS A model is useful  for  the purposes outlined above because the Okanagan's resources are a complex system. The human capacity for  understanding complex systems is surprisingly limited. Even when a person can understand each relationship between parts separately, a System Dynamics model is useful  in that it can track each of  these relationships simultane-ously. Furthermore, it can track delays and feedback  loops that frequently  are the sources of  unexpected results. This model is a decision support tool, not a decision maker. It does not optimize, nor is it appropriate for  use in system design or operation. It is, however, very useful  for  gaining an appreciation for  the system as a whole, and for  identifying  general trends and probable be-haviour. Anyone who has made an important decision in his/her life  knows that decisions are not based exclusively on facts  and data. Values and beliefs,  historic and cultural contexts, and attitude about risk all play important roles in decisions. The field  of  Integrated Assessment acknowledges that science doesn't have all the answers and that local experts also play an important role. Since the purpose of  our initiative is to provide support for  decision mak-ing, it is important to include the end-users in the modeling process. Therefore,  model de-velopment took place within a "Group Model Building" process, involving local experts, water professionals  and related interests from  the initial stages through to delivering the product. This approach is called "Participatory Modeling" within the Integrated Assessment litera-ture, and "Group Model Building" within the System Dynamics community. Both aca-demic areas have documented several recent case studies using this approach, but few  that incorporate climate change. III. MAJOR DATA SOURCES Climate Scenarios Global  CC  models  downscaled  (Taylor,  Barton, 2003) Hydrologic Scenarios UBC  Watershed  Model  (Merritt  & Alila,  2005) Crop Water Demand Agriculture  Canada  Model  (Neilsen  et al., 2005) Residential Demand & Cons. Strate-gies -Master's  Thesis,  Royal Roads  (T.  Neale,  2006) Conservation Flow Targets Trout  Creek  WUP  Guidelines  (Epp  & NHC  Report) \ Okanagan Lake Operational Rules FWMT  Apprentice Guidebook  (Alexander,  Symonds  & Hyatt,  2005) & Symonds—personal  communication Return flow  pathways Epp, Reeder,  Cotsworth  —personal  communication IV. NAVIGATING THE INTERFACE This section walks you through some of  the main areas of  the interface  level, to help you get a sense of  the navigational "map." This will show you some of  the key features  of  the model and prepare you for  exploring it. 1. Background m m v t Select "Home" Figure 1: Home Page Background Page: Select "Map of Water Supply Source Areas" Note the model has been di-vided into three water source areas: • Uplands = Tributaries to Okanagan Lake • Valley = Okanagan Lake • South End = Sources from the outlet of  Okanagan Lake 2 Penticton to the inflow  to Osoyoos Lake. Version 0 57 January 2006 When you open the model, you will see the "Home Page:" Figure 4: Learning about the System Home Page. Check that the History button is on, and then click on the purple "Review History" but-ton. 2. Learn the System Figure 3: Returned to the Home lBefare  Starting.. CPbraseSize  tSe  iiinJow  to  I'll  th Learning About the System On this page, select whether you want to explore historic conditions, or simulate the fu-ture In the model, this page is re-to as "Learning." 3. Review Simulated History i OtMrtafcan JJ5J.STM Either the "Go Back" (U-turn) button or the "30-Year History Graphs" will take you back to Figure 5. Select to run the historic simulation (1961-90) Click on these triangles to flip pages in the graph up or down. Figure 6: Detail for  the 5-Year Drought 1986 -1990. •>•» laoajBiBB Review 30-Year History • 1: 3ass Case :ntows(Upiand Tni>s] 2:6a&« C m snaowsjOK Uk»; 3: ess s C m :r*o*s;3c«m EndWWi 5-*ear Example i Periods Figure 5: 30-Year Page with 1 st graph of  pad showing. •» < • Afl  h ^  r 1986 to 1990 Drought L ^ J 5-Y»ait , Example : P»tio<ls | I Normal'tm' to 19€S) Plood i l*>71 to 1975) ) JL Upland Inflows (million cubic meters per month) and Supply Demand Ratios Next, we will discuss the Future Section. Return to the "Review 30-Year History Page" and then select the button labeled "Learning" to navigate back to the Learning About the System Home Page. 4. Explore the Future . „ ] . . . , n . - . W f f Learning About the System 1 a :  Quit  , Figure 7: Returned to Learning Ab [13 To run future  simulations, turn on "Future" and then select "Future Settings" )ut the System Home Page. Pie Ed* Interface Run Help »y reviewing the Future trios (Population. Climat* I pecting the different rios for the two different Years (2010 20401. Use immary Graphs to got a i of how the different rios might affect  the Select  Initial  Year Future Scenarios Population ' Climate start  *> :xvi First, choose the 30-yr simulation period: 2010—2039 or 2040-2069. Future Settings Graphs of Consequences •explore  Policy  Options Agricultural W i l l i U s . O Upper Tributaries O South End Tributaries O Oltanagan Lars nmary Graphs liilllal»Mai#S«lttM» Next, choose the future scenario conditions for population growth rates and climate change. These two pages are illustrated on Page 10. #smu'ss.i.i Figure 8: Future Settings Page. This is the Command Centre for  your setting for  simula-tions of  the future.  *Remember this page is called "Future Settings" as you will return here often. 4.1 Future Scenario Options Here we select the rate of  population growth into the future  (Zero, Slow, Moderate, or Rapid) and "turn on" climate change. The population growth rates in each setting vary by regional district, based on official  estimates. This plot is a compara-tive graph. When you run additional simula-tions, previous output will be maintained. Plotted here are the range of  future  popula-tion scenarios for 2010—2039. * Comparative graphs can only plot a single parameter. Multi-parameter graphs will clear with each new simulation. Figure 9: Population Growth Settings Page. Is Selected Inflow  Oala for  Climate Scenario mem oer monthlUpland Tribal: SOOn . - J 1 L Turn on the Climate Change by clicking on the switch's small round knob. Until you do this, all simulations will main-tain base case conditions. Staphs of Conseque m cubic meters per month) jyj """•"< 1 ); Future  Setting s ) This model contains only one scenario— Hadley A2, which is one of  the more severe of  the moderate scenar-ios. The graph shown is comparing the inflow hydrograph for  two simulations—Base Case (1) and Hadley A2 (2). Figure 10: Climate Change Settings Page. 4.2 View Consequences Most pages have a navigation button labeled, "Graphs of  Consequences." Selecting this button brings you to the menu page shown in Figure 11. Each menu item leads you to a page of  output graphs related to that topic. The first  item, "Supply Demand Issues" is shown here (Figure 12). This page provides supply and demand balances for  the Uplands and Valley areas. BOBijflgi * -AB-'Hftg/ Scenario & Policy Consequences graphs Okanagan  Lake Residential  Sector Agricu ltur e  Sector I Groundwater  Is sues  | Uyj  Settings jf  RbvIbw  History  j Figure 11: Menu for  Consequen ;es Graphs. • STtlMJ-M | S Okanagan J58.STM - . . • ; , ' . ' ( " IE2HI ! '- - RISE WM I Sl/gifX? 'IS '-''I Supply and Demand are ex-pressed as a normalized dif-ference.  Thus, positive val-ues represents a surplus con-dition, while negative values reveal episodes of  shortages. This graph compares results for  the Base Case (1) and Hadley A2 (2) scenarios. Figure 12: Consequences Relating to the Balance of  Supply and Demand in the Uplands When you are finished explor-ing the consequences pages, select "Future Settings" to re-turn to the Command Centre. 4.3 Test Policy Options If  you are not happy with what the future  could bring, then do something about it! The model provides numerous policy options that could affect  water management in the future. The Urban/Residential Use Options are shown here in Figure 14. S 1111 AS 8.1.1 I Fh Edt Interface Run Help I Oka«aftan Q57.STM A n ar» v Future Settings 1 ) start by reviewing t h . Future Scenarios (Population, c l imat . i lo i i l C ig t h . different icsnarios for t h . two different Initial Years 12010. 20401. U s . Summary Staphs to get a by inspectin  l s e n s , of how scenarios ml| system. Ftrtma  Scenarios population ' ; Climate Summar y S taphs \ \ Mo. MM Darwei Sottas i men.... Graphs of Consequences Explore  Policy  Options Agricultural Water Us* Q Upper Tributaries O Soutii End Tributaries Figure 14: Policy Options in Urban/Residential Water Use. Figure 13: Back to the Future page to Explore Policy Options. Urban/Residential Water Use Settings j Rub [stop] | msmsm*: 1 Total UL Res Dsmand mem per month: 1 Residential Conservation Options are here. Select one and run a simulation to see the ef-fects. Appendix D: Model Level Documentation for the Okanagan Sustainable Water Resources Model Model Level Documentation for  the Okanagan Sustainable Water Resources Model Version 1.62.1 April 2007 Stacy Langsdale Institute for Resources, Environment and Sustainability University of British Columbia I. Overview V The Okanagan Sustainable Water Balance Model (OSWBM) simulates future conditions by projecting current conditions and overlaying the effects  of  population growth and climate change on water supply and demand. The purpose of  the model is primarily as a tool to support stakeholder dialogue on the issue of  long-term water management. The OSWBM is not intended to support design or operation. The model can support users in learning about the complexities of  managing water resources for  multiple uses, generate a range of  plausible water resources futures, assist users in assessing adaptation strategies (and portfolios  of  strategies), identify data gaps, and prioritize areas of  future  research. Error! Reference  source not found,  is a causal loop diagram of  the Okanagan water resources system. This diagram shows the important relationships and feedback loops that drive the behaviour of  the system. Figure 1: Causal loop diagram of  the Okanagan water resources system, with boundaries of modeled system noted. II. Climate Change Scenarios The information  in the model regarding climate change is based on work by Taylor and Barton, who down-scaled global climate model output to the Okanagan scale. The climate scenarios are based on the results of  three global climate modeling teams, with two emissions types from  each. Simulations are generated in thirty-year time blocks, identified  as the 2020's (2010-2039) and the 2050's (2040-2069). Data was also generated for  the 2080's (2070-2099), but was not added to this model, as the participant group felt  it was unnecessary. Planning initiatives typically consider twenty-year projections, and occasionally fifty-year  projections. The participants expressed concern that longer projections would contain too much uncertainty to be informative.  The new climate condition is held constant over each thirty-year period. This version of  the model includes three climate scenarios, which were selected to show a range of  possible futures:  Hadley A2, CGCM B2, and CSIRO B2. Details about these climate scenarios can be found  in Cohen et al. (2004). The model does not include output directly from  global climate models, but incorporates translations of  the climate scenarios into supply and demand impacts. In this way, the model captures the linkages from  climate change to hydrologic inflows and to and agricultural water demand. In addition, residential outdoor use is correlated to the temperature component of  climate change. Temperature scenarios are also based on the thirty-year scenario periods. These temperature scenarios were smoothed to create a gradual temperature change over the thirty-year period, rather than having a uniform  temperature shift  in each thirty-year time block. Smoothing was performed  on the temperature shift  due to climate change only; monthly and annual variability was maintained. Figure 2 shows how the climate change scenarios were utilized and translated and input into the model. Figure 2: Flow chart illustrating the progression from  climate models and local records to supply and demand inputs to the Okanagan Sustainable Water Resources Model. (Sources: Cohen et al. 2004; Neilsen et al. 2001; Merritt and Alila 2006; Neale 2005; 2006). Global  Climate  Models HadCM3,  CGCM2,  CSlROMk2; A2 & B2; 2010-2099 Okanagan climate  stations Water Supply Hydrologic Scenarios Hydrologic scenarios were generated for  all of  the subwatersheds in the Okanagan Basin. Details on this work are published in Merritt and Alila (2004). These scenarios were based on the Climate Scenarios generated by Taylor and Barton (2004) and were generated using the UBC Watershed stream flow  runoff  model (Quick 1995). Climate scenarios produced information  about precipitation and temperature changes. This information  was fed  into the UBC Watershed model to generate 30-year hydrographs for  each watershed. In most of  the watersheds, particularly those in the higher elevations that receive a significant  portion of  precipitation as snow, increased temperatures decrease snowpack and induce an earlier spring freshet. The UBC Watershed model was used to simulate natural streamflow,  with no management of  dams. In each of  the 50+ watersheds, the simulation was based on the location of  the stream gage. No groundwater interaction was modeled. Calibration focused  on matching peak flows  rather than minimum flows  (Merritt and Alila 2006; 2004). The watershed data are aggregated into three groups: Uplands,  Valley,  and South End.  Uplands  include all managed watersheds that are tributaries to Okanagan Lake. Valley  includes a few  small unmanaged watersheds that are tributaries to Okanagan Lake. South  End  includes all watersheds that feed  into Okanagan River and the mainstem lakes between the outfall  of  Okanagan Lake and the inflow  to Osoyoos Lake. Figure 3 and Table 1 shows the placement of  each watershed into the three water source groupings. Most simulations were based on the historic period 1961-1990. However, several watersheds did not have complete data records during this period. Most commonly, data was missing from  1961-1968. In these cases, I correlated watersheds with missing data to adjacent watersheds of  similar size. Because the use of  the data is in an aggregated state, the accuracy of  calibration to each stream is less critical then maintaining the order of  magnitude of  the aggregated total. Figure 3: Map of  the Okanagan Basin showing the delineation of  the three model regions defined  by water source. Cneyhatk Lake £Bs aewr/eir TMrric Oam South End 10 20 Kilometers Uplands Watershed Identification  by Water Source Group Uplands Valley South End No. Name No. Name No. Name 1 Deep Creek 5 — 26 Shingle Creek 2 Irish Creek 7 — 27 Skaha and Felis Creeks 3 Newport and Bradley Creeks 9 — 28 Marron River 4 Equesis Creek 11 — 29 — 6 Naswhito Creek 13 Faulkner and Keefe Creeks 30 Park Rill Creek 8 Whiteman Creek 15 Smith and Westbank Creeks 31 10 Shorts Creek 17 Drought Creek 32 Reed Creek 12 Lambly Creek 19 McCall Creek 33 Hester Creek 14 MacDonald Creek 21 . — 34 Testalinden 16 Powers Creek 25 Madeleine Creek 40 — 18 Trepanier Creek 56 — 41 — 20 Peachland Creek 60 - - 42 Wolf Cub Creek 22 Eneas Creek 43 Vaseaux Creek 23 Prairie Creek 44 Irrigation Creek" 24 Trout Creek 45 Shuttleworth Creek 51 Penticton Creek 46 — 52 Strutt, Naramata, Robinson Creeks 47 McLean and Matheson Creeks 53 Chute Creek 48 Gillies Creek (and others) 54 — 49 Ellis Creek 55 Bellevue Creek 50 — 57 Mission Creek 58 — 59 Kelowna Creek 61 Vernon Creek 62 — Figure 4: Two year hydrograph of  aggregated flow  in upland streams, showing shift  from  base case (1981-82), to a decreased, earlier peak in Hadley A2 - 2020's (2030-31) and Hadley A2 -2050's (2060-61). : Base Case lnf lows[Upland Tribs] 2: HadleyA2 Inflows 2010 to 2039 mem. . . 3: Hadley A2 Inflows 2040 to 2069 mem. Page 2 V —!-» 1 l 240.00 246.00 258.00 252.00 Months t j Inflow Values for 2 Cl imate Scenarios (Base Case and Hadley A2) in mi l l ion cubic meters per mon th Diversions from Adjacent Basins There are at least three sources from  which water is transferred  to streams in the Okanagan - Kettle, Shuswap (from  Duteax Creek) and Alocin (from  the Nicola Basin). All of  these are relatively minor. Kettle and Shuswap are captured in the model. Alocin is neglected because the.participant/expert advisory committee suggested it was very minor. (The relative magnitudes of  the three diversions were not compared, however.) Kelowna has water rights to divert up to 3780 acre-feet  (4.66 million cubic metres) of water per year from  the Kettle River basin, and this water is taken during the spring and summer months. Information  about the Kettle diversion was not available, so I approximated this diversion to provide 500 acre-feet  (0.62 million cubic metres) per month to the Okanagan from  May through October, matching the months of  peak demand. Diversions from  the Shuswap River basin (taken from  Duteax Creek) are not recorded, but I did have an estimate of  the number of  users on this source. Therefore, the model calculates the Shuswap diversion to be equal to the demand by the populations of  residents who currently depend on the Shuswap as their water source. The Shuswap also supports some agricultural customers, so the estimate may be low. The diversion increases as the populations in these communities increase. The total water rights on the Shuswap are equal to 41,000 ac-ft  or 50 million cubic meters. The maximum diversion simulated falls  well below the total Shuswap water right.1 Groundwater Groundwater contributes to the total water available in the Uplands. Since there is presently a lack of  information  about the interaction between groundwater and surface  water, and about the sustainable yield of  the aquifers,  in the model I assume that groundwater supplies can meet present and future  demand. In the same manner as the Shuswap diversion, future  use is based on current use, with selected population growth rates applied. As with the Shuswap diversion, groundwater pumping increases as the populations of  the communities that currently dependent upon this source grow. Groundwater is also an important water source in the South End. Here, however, the model does not merge groundwater and surface  water. Because almost all residents are on groundwater and almost all other users depend on surface  water, the sources are kept separate. This is not reflective  of  the physical reality of  the system - in fact, I expect that there is groundwater/surface  water interaction. 1 A variation developed for  Alexander and Hyatt for  the Fish Water Management Tool increased the Shuswap diversion to supplement water shortages in the Uplands, up to the limit by the water right. Figure 5: STELLA component for  treatment of  groundwater in the Upland sector, UL Groundwater Aquifers mem <5 EQ Use( water infiltrating to UL mem per month UL Groundwater Pumping mem per month W W to UL G W mem per month Return Flows to Uplands In the Okanagan, water is not only used once, but cycles through several uses before flowing  south into Osoyoos Lake and across the international border. In the Uplands, water is returned in two ways, both actively and passively: (1) Domestic water is treated and re-distributed for  irrigation of  golf  courses, (2) Outdoor irrigation (agricultural and residential applications) drains to Upland surface  water sources. Additional information  on return flows  is provided in a following  section. IV. Water Demand Sectors Residential Demand Population Growth Calculations All population data in the model is based on recent population statistics for  each Regional District. Planning areas within each Regional District were considered because the political boundary lines are not aligned with the watershed boundary, and some residents do not depend on Okanagan water supplies. The identification  of populations outside of  the Okanagan was estimated using spatial maps as well as information  provided by Al Cotworth. See Table 2. Only populations served by Okanagan water sources were included in the model. Data was acquired through each Regional District's web sites. The model calculates population for  the initial simulation year using the historic or future  selected growth rate between the data year and the initial simulation year. In the model, options for  initial years are 1961, 2010, and 2040. Recent population data is used to calculate initial populations using: Pn=PMrY° where Px = population at time x n = year of  interest o — year of  known value r = annual growth  rate between time n and  o. For historic simulations I did not simply define  the initial starting populations because it was simpler to follow  the same structure developed for  the future  simulations. Historic growth rates from  1961 to 2001-04 were calculated from  BC Stats census data to be 1.98% for  NORD; 4.22% for  CORD; and 1.80% for  RDOS. Determining these growth rates was complicated by the fact  that I am only modeling portions of NORD and RDOS, and because the political boundaries for  certain communities changed during this forty-year  period. However, I think these are reasonable estimates for  the purposes of  the modeling exercise. Table 3 shows the small errors between the simulated and recorded populations for  1961. Table 2: Current population data for  Regional Districts, and the portion of  the population served by Okanagan Basin water sources. Recent Population Data Reported Portion on Modeled Sources Population % # NORD, 2002 census data Vernon 33494 100 33494 Armstrong 4256 100 4256 Coldstream 9106 100 9106 Enderby 2818 0 0 Lumby 1618 0 0 Spallumcheen 5134 15 770 Elec Area B 5303 100 5303 Elec Area C 3627 100 3627 Elec Area D 2840 0 0 Elec Area E 938 0 0 Elec Area F 4093 0 0 TOTAL 73227 56556 CORD, 2004 census data Westside 37693 100 37693 Central OK East 4100 100 4100 City of Kelowna 105621 100 105621 Lake Country 10064 100 10064 Peachland 5077 100 5077 TOTAL 162555 162555 RDOS. 2001 census data 2001 Penticton 30985 100 30985 Summerland 10713 100 10713 Oliver 4224 100 4224 Osoyoos 4295 100 4295 Princeton 2610 0 0 Keremeos 1197 0 0 Elec Area A 1897 100 1897 Elec Area B 1241 0 0 Elec Area C 4721 100 4721 Elec Area D 6604 100 6604 Elec Area E 1996 100 1996 Elec Area F 1989 100 1989 Elec Area G 2194 0 0 Elec Area H 1969 0 0 TOTAL 76635 67424 Table 3: Comparison of  historic populations and those simulated by the model using calculated growth rates from  the historic to the present. 1961 Population Data NORD CORD RDOS BC Stats Simulated Error 25694 27460 33486 25078 27485 33030 2.4% -0.1% 1.4% BASIN TOTALS 86640 85593 1.2% Annual population growth rates for  future  simulations are the same as those used by Neale (2005), and are based on BC Stats PEOPLE 27, Official  Community Plans and other planning documents. These rates are shown in Table 4 with historic rates shown for  comparison. Table 4: Historic and Projected Annual Population Growth Rates [%] by Regional District. Historic (1961-90) Future Slow Future Moderate Future Rapid NORD 2.0 1.0 1.4 2.0 CORD 4.2 1.0 1.7 2.4 RDOS 1.8 0.5 1.0 2.0 Because the model simulates in monthly timesteps, these annual rates (Rannuai) are converted to monthly growth rates (Rmonthiy) using: monthly = + ^annual ~ ^ During simulations, populations increase according to the selected growth rate using a classic exponential growth relation. The relation is shown in the population growth equation, where n and o are in months rather than years, and n- o = 1. In STELLA™ language, population growth is expressed as shown in Figure 6. Figure 6: Exponential growth in the STELLA™ language. Converting RD populations to populations by water source The Regional District boundaries are not the same as the three zones in the model Population growth is calculated by RD to make use of  the different  growth rates. This sector reorganizes the populations according to their water source. Data regarding water sources for  different  communities was provided by local experts, particularly Al Cotsworth, Toby Pike and Andrew Reeder (Table 5). Table 5: Division of  residential water use according to water source. % Resid. Population on Each Water Source % Pop on Source Area South OK South Shuswap Ground- Upland Okanagan End Osoyoos Uplands Lake End Tribs water SW Lake SW Lake Total Total Total NORD Vernon 10 0 90 0 0 0 100 0 0 Armstrong 95 5 0 0 0 0 100 0 0 Coldstream 40 0 60 0 0 0 100 0 0 Spallumcheen 10 90 0 0 0 0 100 0 0 Elec Area B 100 0 0 0 0 0 100 0 0 Elec Area C 75 0 25 0 0 0 100 0 0 CORD Westside 0 0 100 0 0 0 -100 0 0 Central OK East 0 0 0 100 0 0 0 100 0 City of Kelowna 0 0 50 50 0 0 50 50 0 Lake Country 0 0 85 15 0 0 85 15 0 Peachland 0 0 100 0 0 0 100 0 0 RDOS Penticton 0 0 40 60 0 0 40 60 0 Summerland 0 0 100 0 0 0 100 0 0 Oliver 0 100 0 0 0 0 0 0 100 Osoyoos 0 100 0 0 0 0 0 0 100 Elec Area A 0 100 0 0 0 0 0 0 100 Elec Area C 0 100 0 0 0 0 0 0 100 Elec Area D 0 0 40 60 0 0 40 60 0 Elec Area E 0 0 ' 100 0 0 0 100 0 0 Elec Area F 0 10 0 90 0 0 0 90 10 Calculating Residential Water Use Residential water demand calculations are based on an analysis of  water use records in the communities of  Kelowna, Penticton, and Oliver by Neale (2005; 2006). The analysis assumed that indoor use is relatively constant throughout the year, while outdoor watering occurs primarily between April and October. Water use also increases in the summer due to a significant  influx  of  tourists. This effect  is not distinguished, but it is captured in the outdoor use peak since relations were developed from  actual data.2 Indoor water use is calculated as a per capita rate. The historic and default  rate used is 400 L/day/person which captures all residential, municipal and industrial use. The relation is simply: Indoor  Use  = Population  x Use.Rate Neale showed that outdoor water use is a function  of  both housing type and maximum daily temperature. Assuming that the majority of  residential outdoor water use is applied to lawns and landscape maintenance, residents of  multiple unit dwellings have no significant  outdoor water use. (Most apartment complexes do have some landscaping, but this is minor when divided by the large number of  residents.) Linear correlations were developed between water use and daily maximum temperatures averaged monthly. These relationships were used to estimate increases in watering in the future  associated with increasing temperatures due to global warming. I applied the relation for  Kelowna to both the Uplands and Okanagan Lake water source areas, and applied the relation for  Oliver to the South End. The relations varied widely between the three study areas, with use increasing substantially from  the northernmost location (Kelowna) to the drier, southernmost location (Oliver). The central proximity of  Kelowna and Oliver to their associated water source areas makes the extrapolations reasonable, however, the extrapolations from  single communities to entire regions should be reassessed as more information  becomes available. The linear correlations from  Neale (2005) for  outdoor water use per dwelling are: U s e d , , m n g = a • 7™ax + h [Equation 1] where Usedweiiing = cubic metres of  water use per month per detached dwelling, Tmax = daily maximum temperature values averaged monthly, in degrees Celsius, and the coefficients  a and b are defined  in Table 6. 2 Because outdoor use includes tourism demand, simulations of  outdoor use should not drop to zero. This result is possible with the scenario of  100% apartments and no single family  dwellings. Water Source Area a [m3/month/°C/dwelling] b [m3/month/dwelling] Uplands & OK Lake (Kelowna) 3.0612 -34.465 South End (Oliver) 10.979 N -145.35 Next, total outdoor water use for  each area of  interest [cubic metres per month] is determined by: WaterUse outdoo r =Use dweUin g • Population  (SDD.Ratio)-1 Occupancy.Rate where SDD.Ratio  = the ratio of  ground-oriented, single detached dwellings to total dwellings, Occupancy.Rate = the average number of  people per household. At present, Regional Districts in the Okanagan have reported between 2.3 and 2.5 occupants per household and 31% apartments (69% ground-oriented dwellings). Due to changing population demographics, these values are expected to change to 2.1-2.2 occupants per dwelling in 2031, and apartments are expected to increase to 34%, decreasing SDDs to 66% by 2069. Residential Demand Side Management Strategies Neale (2005) reviewed case studies in the literature to estimate the efficacy  of  a variety of  demand side management strategies. The strategies evaluated and included in the model are listed in Table 7 with the expected reduction in water use. Table 7: Average water savings from  demand side management options (from  Neale et al. 2007; Neale 2007; 2005) DSM Option Water Savings Indoor Outdoor DSM Option 1: PUBLIC EDUCATION Sustained public awareness program including a part-time staff  person and printed brochures etc. (source: Hrasko 2003) 10% 10% DSM Option 2: METERING WITH CUC Water meter installation and volume-based billing. Water rate is a constant unit charge (CUC), or the same charge for each additional unit of water consumed, (sources Stephens et al. 1992; Mayer et al. 2004) 20% 20% DSM Option 3: METERING WITH IBR Water meter installation and billing with an increasing block rate structure (IBR). Volume-based water charges increase when water use exceeds pre-defined thresholds, (sources: Gleick et al. 2003; Herrington 2001) 32% 32% DSM Option 4: XERISCAPING Xeriscaping bylaws are implemented, similar to the landscape ordinances used in several US jurisdictions, requiring all new and renovated landscaping to conform to xeriscaping principles, (sources: Richard 1993; Brandes and Fergasun 2003; Gleick et al. 2003; Kunzler 2004; Xeriscape Colorado 2005) 0% 50% DSM Option 5: HIGH EFFICIENCY FIXTURES & APPLIANCES Bylaws are implemented requiring water efficient  appliances and fixtures to be installed in all new and renovated dwellings, (sources: DeOreo et al. 2001; Gleick et al. 2003; Mayer et al. 2004) 40% 0% DSM Option 6: COMBINED METERING WITH IBR, XERISCAPING AND HIGH EFFICIENCY APPLIANCES & FIXTURES Options 3-5 are implemented. Water savings for Xeriscaping and High Efficiency are reduced by the percentage water savings for Metering with IBR to account for voluntary adoption of these mechanisms under the metering program. 59% 66% Agricultural Demand Agricultural water demand is based on scenarios developed by the Crop Water Demand (CWD) model by Neilsen et al. (2006; 2004). Neilsen correlated crop water demand to a number of  factors: • Crop type • Solar radiation (function  of  latitude) • Canopy development ( • Length of  growing season • Maximum daily temperature • Growing degree days (GDD) The CWD model generated scenarios that correspond to the climate change scenarios. Output is given in terms of  total water required by each major water purveyor per unit of  time. Total land in production and the crop type mix were based on current conditions and assumed to be constant in both the historic and the future  simulations (Cohen et al. 2004, p 91). To provide the option for  varying the total land in production and the mix of  crop types in OWSRM, I converted the CWD volumes to rates of  application per land area and per crop type (Pasture, Vineyard, Tree Fruits, and Other Crops). Rates were calculated for  the major water purveyors and extrapolated to the total land in production by water source (Upland tribs, Okanagan Lake, South End). The variation of  crop water demand rates between purveyors is illustrated in Figure 7. Note the rates are relatively consistent. A sub-set of  the 50+ major water purveyors were carefully  selected to ensure diversity in geographic location and to ensure adequate representation of  each crop type. Because the variation between major water purveyors in each area was relatively minor, the additional effort  of  testing each and every purveyor would not have changed the value of  the average significantly, particularly relative to the accuracy required for  this modeling exercise. Crop water demand does vary somewhat between locations (Figure 8) and varies considerably between crop types (Figure 9). Although the timing and magnitudes of "Pasture" and "Cropland" are nearly identical, "Tree Fruits" and "Vineyard" show very different  requirements. The lower water requirement of  wine grapes has stimulated considerable discussion among the community. However, this option may not be socially or economically viable. Figure 7: Comparison of  crop water demand [m] for  seven major water purveyors in the Uplands during the first  two years of  the historic scenario (1961-62) 0.300 0.250 in 0) £ 0.200 <a •o c M E 0.150 ® Q I— 2 IB 5 0.100 a o i— o 0.050 0.000 Jan-61 Nov-61 Mar-62 Jun-62 Sep-62 May-61 Aug-61 Figure 8: Comparison of  crop water demand [m] for  the three different  water source types. Note the increasing trend from  Uplands down to the warmer South End. Uplands - * - Valley South End Month Figure 9: Comparison of  crop water demand rates [m] for  the four  categories of  crops in the Uplands. 0.300 0.250 ¥ g 0.200 ra T3 c ra 0.150 <l> a a V) 3 S 0.100 0.050 0.000 Jan-61 Apr-61 Aug-61 Nov-61 Feb-62 Jun-62 Sep-62 Dec-62 Month •^—Cropland - » Pasture —•—Vineyard —"-—Tree Fruits Figure 10: Comparison of  Pasture crop water demand scenarios. Note increasing trend from historic base case to 2020's and 2050's. Page 2 Months ^ ? In OSWRM, the appropriate crop water demand scenario is selected according to the simulation settings. Next, these values are used to estimate the agricultural water demand in each month. Multiplying the crop water demand rate [m] by the area of land dedicated to that crop gives the total crop water demand in m3/month. However, this rate is the minimum required with 100% irrigation efficiency.  Actual irrigation also includes watering for  frost  protection and evaporative cooling. In the South East Kelowna Irrigation District (SEKID) between 1976 and 1990, which was prior to metering, actual agricultural use was 30% higher than crop water demand. This value is in agreement with expert judgment (van der Gulik, Jan 2006 meeting) and estimates for  Oliver. This factor,  labeled "Ag demand above CWD" in the model, is applied to all four  crops and all three water source areas. To summarize, values from  Neilsen include a 30% increase over what the crop requires, which covers delivery needs (moisture absorbed by soil surrounding plants, for  example). The values I input in the model include an additional -30% for inefficiencies  in the irrigation systems. Instream Flow/Conservation/Fish Demand Conservation Flow Requirements in Tributaries to Okanagan Lake In the Okanagan, instream flow  recommendations to support aquatic life  have been made since the early 1970's, however, targets have not been enforced.  Trout Creek, as a result of  the 2003 drought, is one exception. In 2001, B.C. Fisheries released a report that outlined "Conservation Flow Targets" for  twenty-one3 tributaries to Okanagan Lake that support rainbow trout or Kokanee salmon. These targets vary each month, and are based on a percent of  Mean Annual Discharge (MAD). The MAD is calculated for  each stream, based on long-term average historic data. The target varies each month according to the percentages shown in Table 8. The percentages are the same for  each of  the 21 streams. "Conservation flows"  are intended to be sufficient  to support aquatic life.  This is in contrast to "minimum flows"  which are only the lower threshold of  what is needed to support aquatic life,  and are in the order of  10% of  MAD. In OSWRM, all of  the upland tributaries are aggregated into one parameter. I calculated the MAD for  this aggregated flow  using the 30 years of  historic (1961-90) monthly data (with data gaps patched). The long term average flow  worked out to be 65.1 million cubic metres per month, or 24.77 cubic metres per second. (I checked this by calculating the MAD for  individual creeks and comparing with those reported in the NHC report.) 3 There are forty-six  named tributaries of  Okanagan Lake. Month(s) Percent of MAD January - March 20 April 100 May 200 June 100 July 40 August 30 September 25 October - December 20 Note that if  aquatic ecosystem requirements are only applicable to 21 of  the tributaries, then instream flow  needs are overestimated in the OSWRM. However, even if  this were true, then the current structure captures the relative balance of demands present in each of  the critical streams. Future model revisions should further  explore this issue. Applying the percentages to the MAD of  65.1 million cubic metres per month gives us the target values as shown in Figure 11. In model simulations, the MAD value is held constant, even if  climate change impacts the hydrologic regime. This reflects the reality that recalculating the MAD will be a result of  a policy decision, not an automatic, annual effort.  The figure  shows conservation targets compared to upland stream inflows  in five  years (1978-82) with varying hydrologic conditions. As you can see, it may not be possible for  water managers to meet conservation targets in dry years. When the ideal target cannot be met, "dry year" targets are calculated based on current flows  as an alternative to the standard targets. Default  values are set at 50% of  current monthly inflow  for  all months, however, model users may adjust this percentage for  both the peak flow  months (April - June) and the low flow  months (July - March). In practice, as these targets are not enforced  (except on Trout Creek), the magnitude of  the reduced targets is up to the discretion of  water managers. Early model versions automatically adjusted the conservation target during dry years, and showed a reduced target in the total demand (Figure 12). However, updates in fall  2006 revised this such that the normal year conservation target is used for showing total demand. Figure 11: Modeled standard conservation targets shown with five  years of  varying hydrologic conditions (1978 - flood;  1979-80 - drought; 1981-82 - average). 1: Base Case Inf lows[Upland Tribs] 2: UL Cons Flow Target Normal Year mem per month Page 1 Months Untitled Figure 12: Conservation targets with modifications.  Conservation target is no more than 50% of  inflows.  Greatest reductions are in drought years (1979-80). Months Untitled Page 2 Base Case Inf lows[Upland Tribs] 2: UL Cons Flow Target Selected mem per month Fish Flow Requirements in South End Fish flow  requirements included in the model are defined  by the operational rules for management of  Okanagan Lake, as spelled out in the Fish Water Management Tools Documentation (Alexander et al. 2005). The targets ranges are specified  and are most restrictive between August and mid-October, as shown in Figure 13. Figure 13: Instream flow  targets for  Sockeye in Okanagan River near Oliver over a 12-month period (Jan - Dec) J ® 1: SE Sockeye Max Targets at Oliver m3 per sec 2: SE Sockeye Min Targets at Oliver m3 per sec 1 1 i / / S 2 A 1.00 3.75 6.50 9.25 12.00 Page 1 Months t Untitled V. Wastewater/Return Flow Pathways Water that has been used for  irrigation (both agricultural and residential outdoor) evaporates, transpires and infiltrates  into the soil and groundwater. For residents on sewers, only a fraction  is lost through evaporation (assumed 5%); the remainder flows to a wastewater treatment plant (WWTP). There are several WWTPs in the Okanagan, and each "disposes" of  the treated water in different  ways. Current populations in each of  these communities were used to determine, for  the basin as a whole, what percentage of  water from  residential indoor use goes where. Table 9 shows the return flow  pathways for  each community, and the percentages of  total indoor water use that takes this pathway. The model assumes all residents are on sewer, but this is not the reality. Residents on septic release to groundwater. Table 9: List of  wastewater pathways by community (information  provided by Phil Epp and treatment plant web sites). Residential return flow pathway Communities served Population represented (% of  all modeled region) Uplands groundwater (ex: infiltration ponds) Coldstream (50%); Spallumcheen; NORD Electoral Areas B & C; Lake Country; RDOS Electoral Areas E & F 15% Okanagan Lake Vernon (10%); Westside and Central Okanagan East Electoral Areas, including Native Reserves; City of Kelowna; Peachland; Summerland 60% South End surface water Penticton (90%) 5% Reclaimed for irrigation / Reuse Vernon (90%); Armstrong; Coldstream (50%); Penticton (10%); Oliver; Osoyoos 20% Water Balance Calculations Uplands Hydrology Sector UL  Reservoirs stock represents the aggregate of  the live storage of  all upland storage reservoirs and lakes, as well as supplementary sources. The stock behaves like a single reservoir for  the whole region, regardless of  the fact  that in the real system, many of  these sources do not combine. The major relations in the sector are shown in Figure 14. Inflows  to the UL Reservoirs stock, which are detailed above, include the natural stream flow  scenario (either the base case or a climate scenario), labeled "UL Natural Inflow,"  and the total additional flows  (groundwater, Kettle and Shuswap), labeled "Additional Sources." The net inflow  describes the stream flow  after'land  cover processes take effect,  such as forest  evapotranspiration. Here, a rule is applied that the additional sources are cut off  when the reservoirs are full,  or more specifically,  when UL Emergency Spill Ratio > 1. This is an actual policy for  the diversions. Outflows  from  the UL Reservoirs stock include infiltration  to the aquifer,  the sum of out of  stream diversions, and water that remains instream and flows  to Okanagan Lake. Each of  these components is defined  in other sectors, and described above. There are several assumptions related to this sector: Streamflow  losses can be significant,  particularly in dry years, such that what is released from  storage is not equal to what arrives at a destination point downstream. Loss rates vary between stream reaches, and with the current and pre-existing hydrologic condition. As stream reaches can be both losing and gaining, not enough information  was available to aggregate this phenomenon to the scale of  the Uplands region, so it was neglected in this version of  the model. As additional information  becomes available, it could be added. Evaporation from  storage areas is neglected, on Bob Hrasko's recommendation, because the surface  area of  the lakes is relatively small, and they are located in higher elevations at lower temperatures. The total capacity of  storage reservoirs on the tributaires to Okanagan Lake was calculated by information  provided by Don McKee. Assuming 85% of  licensed storage is actually built, there is 280,000 ac-ft  of  storage in the tributaries. Instream flow  requirements apply to the entire stream reach, which includes areas downstream of  diversion points. Return flows  may supplement instream flow  in the lower reaches of  some streams, however this study did not examine the system at this level of  detail, so I simply assume that instream flow  needs cannot be met by any of  the water that is diverted for  agricultural and domestic uses. Figure 14: Main components of  the Uplands Hydrology Sector, where the water balance for  this region is calculated. © Selected Inflow Data or Climate Scenario rricm per month GW Recharge Rate streamflow factor Total Additional Sources mem per month UL Emergency Spill Ratio Total UL Diversions Final Alloc mem per month UL Instream Flow Final Alloc mem per month UL Lower River Flow mem per month Upland Storage Outflow Calculations Sector Calculations for  releases from  the Upland Reservoir stock are based on operations for the upland reservoirs. Inflows  and total demands are considered to determine what amount of  water the managers would like to release. Next, the state of  the reservoir is considered. The upland reservoirs are not mandated to operate for  flood  control. Through the late fall  and winter, storage may be low, and releases are approximately equal to the rate of  inflow,  maintaining instream flow.  The reservoirs may fill  very gradually through these months, but do not fill  significantly  until the spring during the freshet.  At this time, all demands can be easily satisfied,  and the remainder is captured in the reservoirs. Water is stored in the reservoirs until required late in the irrigation season, to supplement the diminishing surface  water supplies. The amount requested from  the reservoirs is the "UL Pre-release." Between the months of  November through April, the pre-release is equal to the total amount of water requested by agricultural, residential, and instream users. In the remaining months, the pre-release is equal to the maximum of  the total amount requested and the natural inflows. The condition of  the reservoir is defined  by the max and min storage targets shown in Figure 15. Here, they are shown expressed as percentages, but are converted to units of  volume by using the value of  the total storage capacity. The extra step permits the user to easily simulate a change in storage volume while maintaining the operational rules. These storage targets are not absolute limits, but act as guides to help the model to operate in a more realistic manner. During the simulations, the stock may surpass these targets. Figure 15: UL Reservoir Storage Target ranges as a percentage of  the total capacity. These targets help direct management decisions for  releasing (summer) and filling  (during spring The actual value of  the reservoir stock is then compared with the values of  the max and min targets (as units of  volume) to determine if  the reservoir needs to hold back water and fill  or to release excess water. First the "UL Emergency Spill Ratio" is calculated by dividing the actual volume of  the reservoir by the maximum reservoir target. The ratio is limited to a minimum value of  1. If  the value exceeds 1, then additional water must be spilled. Rules for  spilling are defined  in the "Spill Rules Factor," as shown in Figure 16. As the ratio increases, the intensity of  spilling also increases. Figure 16: Spill Rules Factor Rule curve 10.00: Spill Rulesi Factor •. 1.000: txioo s. 1 000 j<t i - i n UL_Eme UL Emergency: l|s|||Si® 1 000 1.050 1 100 1 150 1.200 1 250 1 300 1.350 1 400 1 450 1 500 1.550 1.800 iMillPliP 1.000 1:500. 2.000: 3.000 4 000 5.000: 6.000, 7 000 8.000 9 000 10:00 10.00' 10.00. Dital 'omls Edit Output To Equation Delete Graph Cancel OK j Similarly, the "UL Cutoff  Ratio" is defined  as the ratio of  the actual volume of  the reservoir to the minimum target volume. This parameter is limited to a minimum value of  1. If  the value falls  below 1, then the reservoir volume is low, and the amount of  the water requested for  release may be adjusted, according to the Cutoff Rules Factor, as illustrated by the graph in Figure 17. As values fall  below 1, the cutoff  rules factor  also decreases. When the ratio is zero, then the cutoff  rules factor is also zero, such that no water will be released from  the reservoir when it is empty. Figure 17: UL Cutoff  Rules Factor, as a function  of  the UL Cutoff  Ratio. UL Cuttoff Ratio Cutoff  Rules Factor 0.400 0.200 0.450 0.300 0.500 0.400 0.550 0.500 O.SOO 0.800 0.G50 0.650 0.700 0.700 0.750 0.750 0.800 0.800 0.850 0.850 0.900 0.900 0.950 : 0.950 1.000 1.000 Data Points Edit Ou • ! To Equation Delete Graph Finally, the value of  the "UL Managed Outflow"  is calculated simply as the UL Pre-release multiplied by the two Rules Factors. Also in this sector, the model calculates the value of  the "UL Supply Demand Balance," as: T r r „ , ^ , „ , ULManaged.Outflow-ULTotal.Demand UL.Supply.Demand.Balance  = ULTotal.Demand By this definition,  when the calculated ratio is equal to zero, then only enough water to satisfy  all demands is available. Positive values indicate more than the minimum amount of  water is available to satisfy  demands, while negative values represent deficit  conditions. This value is the important indicator for  determining if  allocations will be short of  demands, and if  restrictions must be implemented. Uplands Demand Reductions for Shortages Sector The main purpose of  this sector is to implement established policies for  reducing use during periods of  water shortage. For example, today there are established policies for  reducing residential use during periods of  drought. This is a tiered structure, where watering is restricted to a certain number of  days per week. As shortages become more severe, the watering restrictions become more intense; watering is permitted on a fewer  number of  days per week. This policy mechanism is captured in the Residential Outdoor Restrictions Factor as shown in Figure 18. When the UL Supply Demand Ratio is low, the Restrictions Factor takes a value greater less than one. The factor  is multiplied directly against the original amount of  residential outdoor water demanded, providing the value for  "Res Outdoor Restricted." More specifically,  outdoor water restrictions are implemented when supplies are only 30% above demands. This is more prudent than waiting until you already have shortages (when the ratio = 0.0) to implement restrictions. Figure 18: Residential Outdoor Restrictions Factor defined  as a function  of  the UL Supply Demand Balance Ratio. , 11.000 Res Outdoor i: Restrictions Factor 0.000 -0.800 •Jurao UL_Supply_Demand_Balance_ UL Supply . Demand. ;••"< Balance .-..Res Outdoor ' 3 R estrictions Factor -0 800 0.000 -0.700 0.000 -o.eoo 0.000 -0.500 0.000 . -0.400 0.300 -0.300 0.300 -0.200 0.300 -0.100 0.300 0.000 0.700 0.100 0.700 0.200 0.700 0.300 0.700 0 400 1 000 iitiiiiMii 5 Data Points: (5T »|§BB»ti liiillli memmmm^BBm^m iffttttttttttt^ ^ To Equation Delete Graph-*};'; . ; ' . Cancel OK Residential Indoor Demand and Agricultural Demand have parallel mechanisms for managing drought. The thresholds and intensities of  restrictions vary for  each use. Residential Indoor restrictions are implemented last, as shown in Figure 19. Agricultural restrictions are not implemented until the ratio is at -0.2, however, water is entirely cut off  if  the value drops as low as -0.8 Figure 19: Residential Indoor Restrictions Factor defined  as a function  of  the UL Supply Demand Balance Ratio. - UL Supply ':2 Demand.-.-^ f Balance. •1.000 -0.900: -0.800 -0.700 -0.600 -0.500 -0.400 •0.300 -0.200 -0.100 0:000 0:100 0.200 UL_Supply_Demand_Balance_ . Data Points: Edit Output: Res Indoor- •' Restrictions Factor 0 250 0.250 : 0 500 0.500 0 750 0.750 1.000 1.000 1.000 1.000 1.000 1.000 1.000 21 To Equation Delete Graph -..£:;' Cancel OK Figure 20: Agriculture Restrictions Factor defined  as a function  of  the UL Supply Demand Balance Ratio. The last step in this sector is to total the new reduced demand. The restricted levels for  agriculture and residential uses are summed with the conservation flow  target that has already adjusted for  drought periods. This new level of  demand that includes adjustments from  drought policies is represented in the converter, "UL Total Reduced Demand." Note that there is an alternative management option included in the model where, instead of  restricting use during low flow  periods, all diversion demands can be satisfied  by pumping water from  Okanagan Lake. This mechanism appears in the last Upland Sector, named, "UPLANDS ALLOCATION SUMMARY." Uplands Demand Allocations Sector The purpose of  this sector is to allocate water to the various users. The first  step in this process is to determine if  the water shortage policies implemented in the previous sector were sufficient  to reduce the total demand level to the amount of  supply available. In a formula  that is very similar to the first  supply demand ratio calculated, we determine the new supply demand balance as: TT T  . , „ , ^ , , ULManaged.Outflow-ULTotal.Reduced.Demand UL.Adjusted.,Supply.Demand.Balance  = 2 ULTotal.  Re duced.Demand Keep in mind that these reductions are only relevant for  months when there are shortages. For the timesteps that had sufficient  water supplies to meet all demands from  the start, these adjustments will have no effect. If  the Adjusted Supply Demand Balance is negative (meaning that we still cannot satisfy  all demands) then we have two options. The model user chooses to either have the simulation stop, at which point they adjust the drought policies, or have the model automatically reduce all demands by the percentage that the system is short of meeting them. The default  value is for  the model to reduce all demands. The "Drought Management Auto or Manual Mode Switch" is where the model user controls this setting. If  manual mode is selected, the "UL Shortages Messenger" will be active. Otherwise, the model will continue on its merry way cutting the allocation to the required level. This is done through the "UL Auto Demand Adjuster." This converter simply takes a value of  the managed supply divided by the reduced demand, but is restricted to a maximum value of  1. This adjuster is used to determine the allocation levels of  all of  the uses, including conservation demand. A final  step conducted in this sector is to calculate the "UL Instream Flow Allocated." In cases of  shortages, this instream flow  will be equal to the "UL Cons Flow" so this step appears to be unnecessary at first  glance. However, in all of  the timesteps with more than sufficient  water available, the actual amount of  water retained in the stream will be greater than the conservation flow  allocation. The formula  is simply: UL Instream Flow Allocated = UL Managed Outflow  - UL Total Div Allocated UPLANDS ALLOCATION SUMMARY Sector In this sector, final  decisions are made regarding allocations. If  the default  setting is used, such that water diversions are reduced during deficit  situations, then the allocations are equal to those calculated in the previous sector. However, there is a user option where any deficit  in the Uplands can be satisfied  with water from Okanagan Lake. In this case, there are no reductions, and allocations are equal to what was originally demanded. Converters with names that include "Final Alloc" are used to include any and all of  the reductions and policies described in other sectors. • ; Valley Hydrology Sector Figure 21: Supply and Demand balance calculation for  the Valley Sector. Okanagan Lake is the water source labeled as "Valley." Sources for  Okanagan Lake include flows  from  the Uplands region, additional natural inflows  for  the minor, unmanaged tributaries, some return flows  from  residential use, and groundwater recharge from  the Uplands.. The value for  inflows  from  the Uplands is equal to the instream flow  allocated for  the Uplands. Lake outflows  include simply diverted flows  for  residential and agricultural use, and outflow  downstream to Okanagan River, representing releases from  Penticton dam. Okanagan Lake Condition Sector The purpose of  this sector is to convert the volume of  Okanagan Lake into terms of stage. We assume that the surface  area of  the lake remains at a constant 341 msm, and that at a volume of  26,000 mem, the stage is at 341.9 m. Valley Diversion Tracker Sector In this sector, we make a significant  assumption that all demands can be met. The justification  is that current diversions from  Okanagan Lake for  agricultural and residential purposes are quite small, as compared with the other regions. Furthermore, I assume that the diversions can be compensated through lake management. Okanagan Lake Dam Operation Sector The information  for  this sector is based on information  provided in The Okanagan Fish. Water Management Tool: Guidelines for  Apprentice Water Managers (Alexander et al., 2005: p. 32). The lake is managed for  a number of  different purposes throughout the chain of  lakes and rivers that connect them, with requirements that vary throughout the year. Purposes include flood  control, kokanee shore spawning and incubation, sockeye incubation, mitigation of  the temperature-oxygen squeeze on sockeye juveniles, agricultural and domestic water intakes, recreational navigation and river recreation. Many of  these needs can be accommodated by stage targets. In the real system, forecasting  plays a significant  role, particularly related to the snowpack and the expected volume and timing of  the spring freshet.  This version of  the model does not have a forecasting  component. The monthly timestep also creates challenges for imitating lake management, which in reality may be adjusted on an hourly basis. Four things are considered to determine how much water to release from  the lake each timestep. First, net inflows  are determined. Theoretically, releasing the same volume would maintain lake levels. Next, lake stage and monthly targets are considered, and the model calculates how much water needs to be released to achieve the target. Third, the minimum and maximum outflow^  thresholds are incorporated and adjust the outflow  value if  needed. Fourth, flow  for  sockeye is considered. Finally, the user must decide wheter the model will aim to meet the lake level targets or meet the sockeye flow  targets. The user setting will determine which value the model uses to calculate releases from  the lake, labeled, "Valley Outflow  FINAL." VII. Adaptation Strategies The following  table summarizes the adaptation strategies and policies available to the user on the user interface.  The "Basic Options" are available in the standard interface,  while the "Advanced Options" can be found  by navigating to the Programmer's Interface. Table 10: Adaptation and policy options included in the model on the user interface Description of option Default value Range I.  Basic  Options Agricultural use in the Uplands Implement ag conservation? How effective  will aq conservation be at reducing water use? No 50% Yes/No 0-100% (max) Agricultural use in the Valley (Okanagan Lake source) Implement ag conservation? How effective  will aq conservation be at reducing water use? No 50% Yes/No 0-100% (max) Agricultural use in the South End Implement ag conservation? How effective  will aq conservation be at reducing water use? No 50% Yes/No 0-100% (max) Residential conservation for whole basin Public education Xeriscaping Plumbing retrofit Metering with constant unit charge (CUC) or increasing block rate (IBR) Off Off Off None On/Off On/Off On/Off None, CUC, or IBR Residential development patterns across the basin Housing occupancy rate (average people per dwelling) Ratio of apartments (multiple-unit dwellings) in the housinq stock 2.3 0.31 1.0 - 5.0 people 0.00 - 1.00 Instream flow needs at Oliver (sockeye as an indicator species) Operate lake outflow to optimize for sockeye instead of optimizing for Okanagan lake level? No Yes/No Drought policies regarding allocation to user groups. Graphs define appropriate levels of restrictions for the full range of values of water deficit. Ag restrictions Res outdoor restrictions Res indoor restrictions 1.00 - 0.00 1.00 - 0.00 1.00 - 0.25 Graph Graph Graph Drought policies for conservation flow targets - these targets are active only when the standard target cannot be met % of inflow in peak flow months (Apr-- Jun) % of inflow in low flow months (Jul - Mar) 50 Sfc-50% 25 - 75% 50 - 100% Drought management simulation mode When shortages occur, either the model will allocate water according to the restriction graphs, or the user can be alerted and asked to reduce allocations manually. Auto Auto or Manual Water recycling and reuse Recycling and reuse presently occurring in the Okanagan can be omitted to test the effects  of this practice on the system. On (Re-use) Re-use/No re-use Supplementation from Okanagan Lake for Upland users Anytime shortages in the Uplands occur, the Ag and Res demands are supplemented with water from Okanaqan Lake Off  (not supplemented) On/Off II.  Advanced  Options  (in  Programmer's  Interface) Hydrology Upland Storage - total capacity (mem) Uplands groundwater recharge - transit time (months) Uplands groundwater recharge - streamflow factor (how much of natural flow enters aquifer) 350 6 0.00 200 - 500 1 - 3 6 0.00 - 1.00 Population growth In the basic user options, population growth scenarios have preset rates for each Regional District. Here, the user can change the multiple values varies from 0.0 to 5.0% Description of option Default value Range definitions of the rapid, moderate and slow growth rates for the three Regional Districts. Agricultural land use by crop type in the Uplands. Note that the initial total is 215 million square metres. Changing values will change the total land in production, unless user maintains this total. Pasture and Forage Tree Fruits Wine Grapes Other 32 59 9 115 0 - 100 0 - 100 0 - 5 0 0 - 1 5 0 Indoor water use in the Uplands Loss from indoor use 5% 0 - 10% Indoor water use return flow pathways in the Uplands. Where does water go after indoor use? Okanagan Lake Recycled Groundwater South End surface water (Rivers downstream of Okanagan Lake) 60% 20% 15% 5% 0 - 100% 0 - 100% 0 - 100% 0 - 100% Outdoor water use return flow pathways in the Uplands.. Where does the water go after ag or residential irrigation? Evapotranspiration losses Groundwater recharge Surface water recharge 50% 17% 33% 0 - 100% 0 - 100% 0 - 100% References Alexander, C. A. D., B. Symonds, et al. (2005). The Qkanagan Fish/Water Management Tool (v. 1.0.001): Guidelines for  Apprentice Water Managers. Kamloops, BC, Canadian Okanagan Basin Technical Working Group. Brandes, O.M., and K. Fergasun. 2003. Flushing  the Future?  Examining Urban Water  Use  in Canada.  Victoria: POLIS Project on Ecological Governance, University of  Victoria. Cohen, S., D. Neilsen and R. Welbourn. (eds.) 2004. Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report. Environment Canada, Agriculture and Agri-Food Canada and University of  British Columbia. DeOreo, W. B., Dietemann, A., Skeel, T., Mayer, P. W., Lewis, D. M., & Smith, J. 2001. "Retrofit  realities". Journal  of  the American Water  Works  Association, 93(3): 58-72. Gleick, P.H., Haasz, D., Henges-Jeck, C., Srinivasan, V., Wolff,  G., Kao Cushing, K., Mann, A. 2003. Waste  Not,  Want  Not:  The  Potential  for  Urban  Water Conservation  in California.  Pacific  Institute for  Studies in Development, Environment, and Security. Oakland, CA. [cited June 10, 2005] Available from http://www.pacinst.org/reports/urban_usage/waste_not_want_not_full_report . pdf. Herrington, P. 2001. "Pricing and efficiency  in the domestic water supply sector". Pricing Water  Economics, Environment  and  Society.  Conference  Proceedings, Sintra, 6-7 September 1999. European Communities. 203-211. Hrasko, R. 2003. Environment Canada Conservation Options Research Report. Kelowna: Earth Tech Canada Inc. Kunzler, C. 2004. "Laws of  the land". American City  & County.  119(11):42-48. Mayer, P. W., Towler, E., DeOreo, W. B., Caldwell, E., Miller, T., Osann, E. R., et al. 2004: National  Multiple  Family  Submetering  and  Allocation  Billing  Program Study.  Boulder, CO: Aquacraft,  Inc. and East Bay Municipal Utility District. Merritt, W. S., Y. Alila, et al. (2006). "Hydrologic response to scenarios of  climate change in the Okanagan Basin, British Columbia." Journal of  Hydrology 326: 79-108. Merritt, W. and Y. Alila. 2004. "Hydrology." Chapter 7 in Cohen, S., D. Neilsen and R. Welbourn. (eds.) 2004. Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report. Environment Canada, Agriculture and Agri-Food Canada and University of  British Columbia. Neale, T., J. Carmichael, and. S. Cohen. Submitted. Urban Water Futures: A multivariate analysis of  population growth and climate change impacts on urban water demand in the Okanagan Basin, BC. Canadian Water Resources Journal. Neale, T. (2005). Impacts of  Climate Change and Population Growth on Residential Water Demand in the Okanagan Basin, British Columbia. School of Environment and Sustainability. Victoria, B.C., Royal Roads University. Neale, T. (2006). Ch 2: Urban Water Futures: Exploring Development, Management and Climate Change Impacts on Urban Water Demand. Participatory Integrated Assessment of  Water Management and Climate Change in the Okanagan, British Columbia, Canada: Final Report. S. Cohen and T. Neale. Vancouver, Environment Canada & UBC. Neilsen, D., C. A. S. Smith, et al. (2006). "Potential impacts of  climate change on water availability for  crops in the Okanagan Basin, British Columbia." Canadian Journal of  Soil Science 86: 921-936. Neilsen, D., et al. (2004). Chapter 8: Crop Water Demand Scenarios for  the Okanagan • Basin. Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report, January 1, 2002-June 30, 2004. S. Cohen, D. Neilsen and R. Welbourn. Vancouver, Environment Canada, Agriculture and Agri-Food Canada & University of  British Columbia: 89-114. Northwest Hydraulic Consultants (2001). Hydrology, Water Use and Conservation Flows for  Kokanee Salmon and Rainbow Trout in the Okanagan Lake Basin, B.C. Victoria, BC Fisheries. Richard, G. 1993. "Pina County's buffers".  Planning.  59(9): 15-18. Stephens, K.A., Van der Gulik, T., and Johnston, C. 1992. "Demand side management: the Okanagan Valley case study". American Water Works Association 1992 Annual Conference. Taylor, B. and M. Barton. 2004. "Climate Change Scenarios." Chapter 5 in Cohen, S., D. Neilsen and R. Welbourn. (eds.) 2004. Expanding the Dialogue on Climate Change & Water Management in the Okanagan Basin, British Columbia. Final Report. Environment Canada, Agriculture and Agri-Food Canada and University of  British Columbia. Quick, M. C. (1995). The UBC Watershed Model. Computer Models of  Watershed Hydrology. V.P. Singh. Highlands Ranch, CO, Water Resources Publications: 233-280. Xeriscape Colorado Inc. 2005. [cited June 10 2005]. Available from http://www.xeriscape.org/index.html. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0100360/manifest

Comment

Related Items