Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

GIS land use modeling in the context of consensus-based regional planning: the development of a GIS-based… Armstrong, Michelle Louise 2000

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

Item Metadata

Download

Media
831-ubc_2000-0158.pdf [ 4.19MB ]
Metadata
JSON: 831-1.0089384.json
JSON-LD: 831-1.0089384-ld.json
RDF/XML (Pretty): 831-1.0089384-rdf.xml
RDF/JSON: 831-1.0089384-rdf.json
Turtle: 831-1.0089384-turtle.txt
N-Triples: 831-1.0089384-rdf-ntriples.txt
Original Record: 831-1.0089384-source.json
Full Text
831-1.0089384-fulltext.txt
Citation
831-1.0089384.ris

Full Text

GIS land use modeling in the context of consensus-based regional planning: The development of a GIS-based land use model for Greater Vancouver BY MICHELLE LOUISE A R M S T R O N G B.A. (Honours), University of Concordia, 1994 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE D E G E E OF MASTERS OF ARTS (PLANNING) In THE FACULTY OF GRADUATE STUDIES (School of Community and Regional Planning) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 2000 © Michelle Louise Armstrong, 2000 UBC Special Collections - Thesis Authorisation Form Page 1 of 1 In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r an a d v a n c e d degree a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r a g r e e t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d by t h e head o f my de pa r tment o r by h i s o r h e r r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . .. The U n i v e r s i t y o f B r i t i s h C o l u m b i a V a n c o u v e r , Canada Department o f http://vvww.library.ubc.ca/spcoll/thesauth.html 3/27/00 A B S T R A C T There is a need for betterinformation and an understanding of the process of land use change for effective planning in the Greater Vancouver region. Achieving this understanding is challenging due to the multiple public sector organizations setting land use and infrastructure planning policy in the region. A review of literature and interviews with representatives of public sector planning organizations in Greater Vancouver reveals that a region-wide, GIS-based, land use computer simulation model can provide the information infrastructure for increasing understanding as well as developing shared learning and enhancing communication and coordination between the numerous planning organizations. In order to do this, the model structure should take advantage of advances in GIS technology and the modeling process must be open and participatory to ensure a successful modeling project. More specifically, seven guiding principles are recommended for the development of a GIS-based land use model for Greater Vancouver. These principles define the key elements that a successful modeling project for Greater Vancouver should either contain or address: (1) Should be flexible and adaptable (2) Should be 'understandable' (3) Should be methodologically sound in that the results should be the logical extension of the inputs (4) Should consider the entire functional region (5) Should have an open and participatory process for development (6) Should have clear objectives for the project (7) Planners should be involved in the modeling process Overall, the most important principle to consider is the need for an open and participatory process for development. An open process creates shared learning, acceptance of the model results, and a forum in which stakeholders can explore and debate ideas and alternative futures. ii T A B L E OF C O N T E N T S ABSTRACT '. II LIST OF TABLES Y 1.0 INTRODUCTION 1 1.1 Problem Statement 1 1.2 Research Objectives 2 1.3 Research Methodology 4 1.4 Thesis Organization 8 2.0 LITERATURE REVIEW OF GIS-BASED LAND USE MODELING AND MODELING PROCESSES 9 2.1 Land Use Modeling for Urban Regions 9 2.2 GIS in Modeling 13 2.3 The Modeling Process and Planning Practice 16 2.4 Consensus Processes ; 19 3.0 CASE STUDIES OF LARGE-SCALE GIS-BASED LAND USE MODELS 24 3.1 Case Study 1: Greater San Diego 24 3.2 Case Study 2: Greater San Francisco Bay Area 28 4.0 STAKEHOLDER ISSUES FOR THE DEVELOPMENT OF A GIS-BASED LAND USE MODEL FOR GREATER VANCOUVER 33 4.1 Background 33 4.2 Research Findings 35 4.2.1 Structure of the Model: Use of GIS and Small Geographic Areas. 35 4.2.1.1 Small geographic areas 36 4.2.1.2 A GIS-based system 40 4.2.2 Utility of a GIS-Based Land Use Model as a Planning Tool 41 4.2.2.1 Increasing certainty v. 43 4.2.2.2 Testing policy and producing "what if scenarios 45 4.2.2.3 Monitoring the success of adopted plans 47 4.2.2.4 Enhancing Communication 48 4.2.3 The Process of developing a land use model for metropolitan Vancouver 49 * * * ill 4.2.3.1 Identifying Stakeholders 50 4.2.3.2 Forming groups to identify needs and build the model 51 4.2.3.3 Managing the model 53 4.2.4 Other Issues 55 4.2.4.1 Accessing the information 55 4.2.4.2 Project funding 55 5.0 R E C O M M E N D E D PRINCIPLES F O R T H E D E V E L O P M E N T O F A LAND U S E SIMULATION M O D E L F O R G R E A T E R V A N C O U V E R 58 6.0 C O N C L U S I O N : 68 BIBLIOGRAPHY 70 APPENDIX A: INTERVIEW QUESTION OUTLINE 73 APPENDIX B: B A C K G R O U N D INFORMATION F O R G R E A T E R V A N C O U V E R P R O T O T Y P E LAND U S E M O D E L 74 iv List of Tables Table 1: Lee's seven "sins" of large-scale urban modeling 10 Table 2: Evolving views of planning and information technology 17 v 1.0 Introduction 1.1 Problem Statement Greater Vancouver is a large urban region with multiple public sector organizations planning for changing land use and infrastructure needs. These public sector organizations include municipalities, regional governments, the provincial government, infrastructure agencies and regional service providers. The planners in these organizations need accurate information on the anticipated growth and changing characteristics of the population so that they can effectively plan for the future needs of their communities. However, in the context of multiple organizations undertaking land use and infrastructure planning in the region, accurate information will not alone achieve effective planning. Planners also need to have a shared understanding on how the land use policies their organization chooses to adopt will impact the location and character of growth elsewhere in the region and, vice versa, how the land use policies of an adjacent jurisdiction or one at a higher level will impact the land use patterns within their own boundaries. Unfortunately, the spatial and temporal interplay of land use policies in a multi-jurisdictional context is not well understood in most urban regions including Greater Vancouver. Without a shared understanding of these interrelationships and accurate information to guide policy formation and decision making, the coordination of effective long-range planning in an urban region is seriously disabled. Large-scale urban computer simulation models have been developed in many urban regions to address this need for more accurate land use information. However, these earlier models have been discredited, as being ineffective in achieving desired ends. Either too simple or too complex, these models often failed to provide planners with accessible and credible information about regional land development processes. In addition, the process of model building and scenario testing has typically been undertaken by a select few leaving users of the information with little understanding of the inputs, outputs, and assumptions of the model. Two recent trends in planning practice, however, present new opportunities for urban computer simulation modeling. The first is the introduction and proliferation of Geographic Information Systems (GIS) in planning organizations. The analytical and l display capabilities of GIS are able to address many of the barriers and inadequacies that have hampered other large-scale urban modeling projects. The second is the movement towards a new planning paradigm that favours inclusive and communicative planning processes. Within this view, information technology such as GIS provides the information infrastructure that facilitates these processes. Given the improvements GIS has brought to large-scale urban modeling and the increasing emphasis in planning practice on open and communicative processes, what role could the development of a GIS-based land use model in Greater Vancouver play in coordinating public sector land use and infrastructure organizations? With multiple organizations setting policy in Greater Vancouver, what are the needs and expectations of these stakeholders for the functionality of a regional land use modeling tool? Finally, in the context of consensus-based regional planning in Greater Vancouver, by what process should a large-scale land use modeling tool be developed? 1.2 Research Objectives The purpose of this research is to explore how a GIS-based land use simulation model for Greater Vancouver may be used to help achieve coordination amongst public sector land use and infrastructure organizations and to present principles for the functionality and development of such a model given the context of a consensus-based approach to planning in the region. Stemming from this overall goal are five specific objectives: (1) Explore current issues of urban modeling and the use of GIS in modeling urban regions; (2) Explore current issues of the role of modeling as a coordination tool and its place in planning practice; (3) Examine examples where GIS-based land use modeling has been developed for an urban region; (4) Explore views of public sector land use and infrastructure planning organizations on the utility and process of developing a GIS-based land use model for the Greater Vancouver region; and 2 (5) Present general principles for the development of a GIS-based land use model for the Greater Vancouver region. 1 To further clarify the purpose of this research, it may serve well to outline the basic assumptions that form the point of departure for this research. Most importantly, this research does not seek to explore the validity of using models in planning practice. This research rests on the premise that models are essential tools in planning practice and that their use will continue to be a key element in defining planning as a discipline. There is a dilemma, however, surrounding modeling and particularly forecasting, as a major component of modeling. Swanson and Tayman refer to this dilemma as being between a rock and a hard place: the rock being that "numbers provide the rhetoric of our age...but to forecast in the sense of making an estimate that will turn out to coincide with what is actually going to happen is beyond human capacity"; the hard place being that "forecasting is impossible yet unavoidable" (Swanson & Tayman, 1995). That is, i forecasts are rarely correct but bring discipline to the task of exposing relationships and developing an understanding of the urban processes that lead to land use change in a region. The answer to the dilemma lies in how a model forecast is used and the meaning that one places on that forecast. One view is that forecasts determine the most likely future in terms of population and employment growth so that needs can be assessed and services can be planned to meet growth. Within this view the forecast becomes self-fulfilling to the extent that planners provide land and public infrastructure in anticipation of this forecasted future and as a consequence increase the attractiveness of those areas to growth (Swanson & Tayman, 1995). The deficiency of this view is that there is little analysis and open public debate concerning the consequences of the forecast. The simple provisioning of forecasted infrastructure "needs" to accommodate growth without an exploration of alternative scenarios and the balancing of land use interests has little in common with the discipline of planning. A second view of the meaning of a forecast that a model produces is that it is less important as whether forecasts are right or wrong, it is how they are used in decision-making. Within this view, modeling is a proactive tool that is an instrument for creating a 3 future (Romaniuc, 1994). The forecast in this case is normative as those who have a stake in the way land use changes decide first what future outcomes are desirable and then design actions and policies to achieve them. The value of the forecast is the analytical insights the forecast provides to inform discussion and decision-making rather than achieving a numerical accuracy to "plan" for anticipated growth. Thus, this research assumes that modeling, although not without problems, is essential, and therefore does not seek to validate modeling as having a rightful place in planning practice. The research focuses instead on how GIS technology and changing views of planning processes have presented opportunities to improve the development and use of urban simulation models to understand land use change. Urban simulation models that take advantage of these opportunities will help planners form a deeper understanding of the choices that must be made to plan for the desired future of a region and will also help planners articulate those choices to other planners and decision-makers. 1.3 Research Methodology The research consists of a literature review and primary research in the form of qualitative interviews. The literature review covers the resurgence of large-scale urban computer simulation modeling, the use of GIS in urban simulation modeling, the modeling process in the context of current planning practices, and two case studies that explore recent initiatives in large US urban regions where GIS has been used in the modeling process. The interviews identified issues for the use and development of a land use simulation model specifically for the Greater Vancouver region. Interviews were undertaken with a representatives of public sector land use and infrastructure planning organizations to elicit their views on how the development of a GIS-based land use model for Greater Vancouver may address planning needs in their organization and how such a project should be undertaken to ensure credible results. All these organizations have an interest in understanding the processes of urban growth and land use change. Since it was not possible to interview representatives of all public sector land use and infrastructure planning organizations in the metropolitan Vancouver region due to the 4 scope of this research, a sample of organizations were selected for their views. The sample contains a cross-section of planning organizations in the region. The organizations participating in the interviews represent three broad planning functions: land use planning (municipal governments and one provincial commission); infrastructure planning (regional sewerage services and an electrical service provider); and transportation planning (regional planning and transit service provider). The organizations are listed below by their planning function with the service scope of each organization indicated in parentheses as either local or regional: ^ Land use planning City of Burnaby (local) City of Richmond (local) City of New Westminster (local) City of Surrey (local) District of Langley (local) GVRD Policy and Planning1 (regional) Agricultural Land Commission (regional) GVRD Corporate Strategies (regional) ^ , Infrastructure planning . GVRD Sewerage and Drainage District (regional) BC Hydro (regional) ^ Transportation planning TransLink 2 (regional) GVRD Policy and Planning (regional) 1 The land use planning and information services component of the Strategic Planning Department have since been reorganized to form the current Policy and Planning Group at the GVRD. 2 In October 1998, the transportation planning component of the Strategic Planning Department and BC Transit operations within the Greater Vancouver area were amalgamated into a newly formed regional transportation body named the Greater Vancouver Transportation Authority (GVTA). The organization was officially launched on April 1, 1999 and is now known as TransLink. 5 Map 1 indicates the Greater Vancouver Regional District (GVRD) municipalities that are represented in the interviews. Although the sample attempted to obtain a representative cross section of public sector planning organizations in the region, there are a number of types organizations that are omitted in the above sample and should be noted. Further research on an urban modeling project for Greater Vancouver should consider their input and active participation. First, the federal government's role in land ownership and management, particularly in the area of ports and waterways is recognized. Second, there is a significant amount of native band land within the metropolitan area. Most of these lands are sparsely populated relative to surrounding communities and thus have considerable growth potential. Native band participation should be considered in further research. 6 Third, only municipalities within the Greater Vancouver Regional District (GVRD ) are represented in the interviews. Although there is d i sagreement on how the Vancouver metropolitan region should be defined, most would probably agree with University of British Co lumb ia geographers Robert North and Wal ter Hardwick who conc lude that the "Greater Vancouver Reg ional District (GVRD ) is too smal l an area to cons ider for adequate planning" (Quoted in Sancton, 1994. p.68). Due to time and travel d istance constraints, however, interviewing representatives of the adjacent regional districts, in particular the Fraser Val ley Reg ional District ( FVRD) with its three member municipalities, was not possible. S ince it is bel ieved an effective model would need to cover the entire region wherever functional relationships exist, the views, input and active participation of these organizations would be important in any model ing exerc i se undertaken. Al l of those interviewed except one have a background in planning practice and all function in a planning capacity for their organization. They had a combination of s ome bas ic technical knowledge of land use model ing and an understanding of the information needs of their organization. E leven interviews in total were conducted between August and October of 1997. The interviews were conducted informally with the aid of a quest ion sheet to highlight major areas of d iscuss ion. This approach was taken with a view that representatives of each organization may have a very different approach to the subject reflecting not only organizational differences but also of their personal exper iences in practicing planning in the Vancouver region. It is bel ieved this approach al lowed for more open-ended d i scuss ion on the potential role of model ing and for i ssues to be raised that had not been anticipated. The findings reported are therefore qualitative and seek not to general ize but to provide insight into s ome key e lements that would require further exploration for a success fu l land use model ing project in the Greater Vancouver region. A copy of the interview question outline is located in Append ix A. For confidentiality reasons, the interview participants are identified only by the organization they represent. With the consent of the participant, the interviews were recorded so that responses and quotes could be transcribed accurately. Al l participants consented to the interview process as presented to them. 7 The majority of those interviewed were already familiar with the basic concepts of a GIS-based land use urban simulation model either through their involvement in the development of a prototype in 1996-97 or through exposure as a sitting member of the Technical Advisory Committee (TAC) of the GVRD. For those that were not familiar with the prototype development, they were given a one-page background on the model basics for review before the interview took place. A copy of this background information is available in Appendix B. 1.4 Thesis Organization This paper is presented in six chapters including this introductory chapter. Chapter two presents a literature review on modeling and modeling processes. Chapter three presents two case studies of modeling initiatives in large urban regions. Chapter four follows with a summary of the findings from the interviews with land use stakeholders in the Greater-Vancouver area. Chapter five recommends and discusses principles for the development of a GIS-based land use model for Greater Vancouver. Finally, chapter six concludes the paper by highlighting the key principles. 8 2.0 Literature Review of GIS-Based Land Use Modeling and Modeling Processes 2.1 Land Use Modeling for Urban Regions Computerized urban land use modeling is a tool used in planning to assist in the development of an understanding of the processes of land use change in a large urban region. Information on the impact of land use policy on the timing, location and character of this change is essential for reviewing, evaluating and formulating policy in decision making. Considering more than one organization is at work in formulating land use policy in an urban region, agreement on information provided by an urban simulation model forms the basis for coordinated and effective planning. For example, planning infrastructure upgrades such as the delivery of sewer services depends on good information about future population and employment patterns. Underestimates in population levels or employment locations may result in inadequate sewer line capacity leading to costly short term programs for expansion, While overestimates may lead to excess capacity thus misallocating scarce resources. Computer land use modeling is an important tool for addressing information needs in planning. Computerized models were first introduced in planning during the 1950s and their sophistication grew with the increasing power of the computer, from basic electronic : data processing to management information systems, relational database management systems and computer-based mapping (Klosterman, 1997). Fueled by the general optimism of the period and the fundamental belief in the efficacy of science and technology, large expenditures were directed to the development of ambitious computer-based models in many US cities (Klosterman, 1994). By the 1970s, it became clear that many of these early models had failed to meet their stated objectives. In 1973, Douglass Lee wrote the landmark article Requiem for Large-Scale Models which criticized the large urban modeling projects of the day. Lee's article was influential and some even credit Requiem as single-handedly setting urban modeling back many years (Wegener, 1994). A review of his criticisms is therefore important to place them in context of today's technology and modeling approach. Lee ascribed 9 seven main criticisms, or what he called "sins", to large-scale urban models which are described briefly in Table 1. Table 1: Lee's seven "sins" of large-scale urban modeling hypercomprehensiveness Early urban models tried to replicate too complex a system and were expected to serve too many purposes at the same time. grossness, The information these models provided was too coarse to be useful to most policy makers. hungriness The models tended to have an enormous appetite for data and a large amount of time was spent collecting and entering data. wrongheadedness Patterns observed at the regional level were systematically applied to specific neighbourhoods without recognizing variations. complicatedness The models were so complex that internal errors were large and results had to be "massaged" to produce reasonable-looking output, seriously undermining their scientific validity. mechanicalness The large models created large rounding errors. expensiveness The models consumed too many resources. (Source: Klosterman, 1994, p. 4) Lee drew numerous conclusions from his review of the models of the day. First, he argued models should be transparent, that is, readily understandable by potential users. If people are not able to understand the relationship between the data inputs and the information outputs, then the model will be perceived as a metaphorical 'black box'. As a consequence, the information these models generate will be viewed with suspicion and the use of the model will rarely extend beyond the sphere of the technical people who produce it. Second, he argued that informed modeling can only be achieved with strong theoretical foundations, objective information and good judgement. Although Lee does not address how this can be achieved, the statement implies a process of model building that involves rigorous review of the model with respect to its basic assumptions 10 and its relationship to planning principles. Objective information and good judgement should not be left to the model builders alone. Approaches to the modeling process will be explored further in a later section. Thirdly, Lee recommended that planners start with a particular problem to guide the model building process, gathering only information that is needed to guide the project. Lee felt the delivery of information on a particular problem if even short term would help propel the development of the model and lend the project credibility. Finally, Lee recommended that models should be simple since in his opinion complex models tend to have large internal errors, their results lack credibility, and they are rarely understood by outside users (Klosterman, 1993). Although the requiem Lee heralded did not actually materialize, his recommendations are noteworthy as they are as applicable to today's models as they were to the models he criticized. However, the problems he identified are more easily avoided today with new innovations in information technology supported by a changing view of planning practice. While models did drop from the planner's agenda for a time during the 1970s, data and information definitely remained. The increasing need for integrated data led to the development of urban information systems and ultimately a resurrected interest in modeling (Batty, 1994). , Models are also likely to be much more important for future planning practice than they have been in the last 20 years (Klosterman, 1994). Research into modeling initiatives point to seven large metropolitan areas of the U.S. that have developed integrated urban models to examine the interaction between land use and transportation. Further research into the current state of large-scale urban models reviewed modeling projects being undertaken in 20 urban centres around the world (Wegener, 1994). Wegener argues that the breadth of this modeling research suggests that Lee's "Requiem" was premature. Lee's "sins", says Wegener, "seem from today's perspective rather ephemeral and in part rendered irrelevant by twenty years of progress in theory, data availability and computer technology since the publication of his paper" (p. 17). Technological innovations such as massive increases in desktop computing capability and the advent of powerful data management, spatial analysis and display capabilities of GIS have made urban models much more sophisticated and accessible to planning (Klosterman, 1994). l l It has also been suggested that Lee's attacks were not necessarily directed at the concept of urban modeling but was a more fundamental criticism of the rational planning paradigm that dominated the period (Klosterman, 1997; Wegener, 1994). Within this view, cities were complex systems made up of hierarchical subsystems and planning was a rational procedure for optimizing the overall system (Wegener, 1993; Klosterman, 1997). This approach to planning fell into poor favour in the 1970s when it was increasingly recognized that planning was not value free and that computer-based information were themself inherently and inevitably political, reinforcing power structures, increasing the power of experts, and transforming policy itself. In the broadest sense, technology was perceived to increase the potential for information abuse and misuse. Batty believes the problems Lee identified were appropriate for the time but supports Wegener's arguments that the limitations imposed by technology have largely disappeared. The major reasons today for the lack of practical applications for these models, he argues, are due to the volatility of the problem context that planning addresses and the inability of these models to withstand such shifts. Thus, since planning is politically motivated, any model developed to assist in planning must be robust enough to accommodate a shift in values or priorities over time (Batty, 1993). Wegener sheds more light on this problem. He argues that models of the past were applied mainly to a very narrow set of planning problems, and have failed to adapt to changing perceptions of problems. He argues that in order for models to establish themselves a firm position in the planning process, they must be flexible enough to provide meaningful answers on the pressing issues facing urban regions. He points to the new models being developed to address environmental issues as breathing new life into modeling and offers the following remarks: Some may find it ironic that it requires the urgency of the environmental debate to grant urban models a new lease of life. It is indeed puzzling to see that even vigorous critics of "rational" models in planning call for just that kind of method for tackling environmental problems. However, the new respect for models is more than just another twist in the intellectual debate about rationality in societal planning. It heralds the twilight of postmodernity in the face of growing risks of ecological disaster. 12 Urban models have a renewed chance because they stand for rationality, and rationality is again needed (Wegener, 1994). 2 . 2 GIS in Modeling i The increasing use and availability of GIS for planning functions presents opportunities for modeling in a number of ways. First, the ability of GIS to perform complex analysis on spatially disaggregate data affords a greater sophistication in modeling urban regions. Second, research suggests that the use of GIS's graphic base makes modeling more understandable, especially to non-experts, than in the past and thus lends more credibility to the information produced. Finally, the proliferation of GIS in planning practice, the increasing availability and accessibility of data and the broadening GIS expertise gives GIS the potential to become a useful tool to facilitate communication and coordination between the multiple public sector planning organizations operating in an urban region. GIS technology can provide the basis for information sharing and encourage participation in a modeling exercise. Although GIS continues to be most commonly used in planning practice as a mapping and archival database system containing information on land use, transportation and infrastructure, the most powerful use of GIS is for spatial analysis. Analysis is performed by layering multiple map datasets together, all of which correspond to a common base map. Typical data sets represent information on land use (i.e. location of industrial, commercial and residential land), street networks, sewer and water systems, economic and demographic data. These data sets or 'layers' may represent existing, planned or hypothetical conditions. It is the layering of these data sets where the spatial analysis capabilities are most powerful. Common analytical functions include the generation of buffers (to study delimited zones on either side of a line feature such as a road, transit way, or sewer line), area or 'polygon' overlays (combination of maps to attribute areas with multiple characteristics), network analysis (calculating routes based on length of line features), and suitability analysis (selecting criteria, performing overlays, and generating alternative scenarios) (Esnard, 1997). Given these functions, GIS is well suited to exploring the implications of alternative public policy and land use decisions. These same elements make GIS attractive for modeling. 13 On the technical side, GIS effectively addresses many of the past problems associated with large-scale urban models. Douglass Lee himself in Retrospective on Large-Scale Urban Models, revisits his own conclusions drawn in 'Requiem'. Lee says that had GIS been available in the 1970s, likely the only sins of modeling would have been "hungriness", "expensiveness", and "grossness". He goes further to say that recent innovations in GIS systems have effectively eliminated all sins of modeling, including "expensiveness" with the advent of the desktop computer (Lee, 1993). The increasing power of computers with then decreasing price has furthered the proliferation of GIS. The use of GIS in large-scale urban modeling is still relatively new but there are a number of ongoing efforts to extend the functionality of GIS past pragmatic functions such as mapping to more sophisticated functions such as model-based applications (Batty, 1994). This functionality is limited, however, evidenced by Wegener's 1994 review of operational urban models. Only one model of the 20 models reviewed used GIS technology. In the concluding remarks of his review he refers to developments in GIS as being of great potential benefit to urban modeling but observes ironically that it is the analytical framework of urban modeling that will likely assist in developing the theoretical underpinnings for the use of GIS in modeling: Despite their recent popularity, GIS have so far contributed surprisingly little to methodological analysis. The theoretical vacuousness of GIS can be remedied only if they are linked to the analytical capabilities of urban models. However, the world of GIS and the urban modeling world are still far from each other (Wegener, 1994). Although GIS has been slow to be integrated into modeling exercises, many argue that GIS can not only address many of the technical problems associated with urban modeling but its graphic base can effectively improve the modeling process by making modeling more understandable and accessible (Tayman, 1996). Tayman also argues that the likely further increase in memory and speed of small computers and the availability of the desktop GIS will continue to lower the barriers to making urban models a tool for planning by a widening range of institutions and individuals, including non-experts. Further, research suggests that GIS technology can empower community groups allowing them to participate more fully in planning and policy discussions that affect their neighbourhoods (Sawicki & Craig, 1996). Understanding and communicating geospatial data about small areas is fundamental in allowing active 14 participation at the community level. Ready access to interpretation of data and analysis of the implications of alternative futures will enable community participation. With the help of GIS, assert Sawicki and Craig, planners can and do play a significant role in democratizing data and information technology. Research investigating the effectiveness of GIS in local planning also demonstrated that there is greater confidence in data analysis performed with GIS. The analytical strength of GIS combined with the credibility of the results were found to positively affect improvements in decision making (Budic, 1994). These findings suggest that the use of GIS in modeling has great potential to improve communication and coordinate decision making amongst the multiple land use and infrastructure planning organizations in an urban region. Others have pointed to the potential coordinative effects of using GIS and the opportunities this presents for planning (Innes & Simpson, 1992; Innes, 1991). With GIS, they argue, planners can link data from different sources in order to approach planning issues spatially, and by extension, help citizens and policy-makers to understand issues spatially. In interviews with planning staff from organizations that use GIS technology, staff describes GIS as a way of "breaking down data barriers between departments" and "the glue that binds departments together". These planners thus recognize the role GIS plays in sharing information and enhancing communication and understanding between different work groups. The same should be true for between organizations. Innes also argues that the process of developing a GIS system can have coordinative effects: The effort to design the system, if it is done in an open and participatory way, will force the agencies and communities to develop a common language in which to communicate about the facts. The use of the data will eventually reduce the variation in activities caused by different knowledge or assumptions about actual conditions (Innes, 1991). Thus, evidence suggests that the collaborative design of a GIS-based modeling system, that encourages face-to-face discussions, negotiation and other group processes which bring the players together to define and resolve issues, could play a crucial role in enhancing communication amongst the land use interests. 15 2.3 The Modeling Process and Planning Practice Information technology does present new opportunities for urban land use modeling, but this does not assume that the application of GIS technology will lead necessarily to a successful modeling project. Generally, the ways in which technology is used is a reflection of the values and issues pressing society. Similarly, the way information technology is used in planning reflects much about the values and currently held beliefs of planning practice. It was discussed earlier that criticisms of large-scale urban computerized modeling was not so much an attack on modeling itself but the rational planning model that held science and strict empiricism as the only valid process of decision-making. A review of how these views have changed over time is therefore instructive. In Planning Support Systems: A New Perspective on Computer-Aided Planning, Klosterman examined the changing views of planning and computer-based information in planning practice. The prevailing view of planning, he argues, has changed the way information technology is used, from planning as applied science in the 1960s, to planning as politics in the 1970s, and then to planning as communication in the 1980s. The view of planning in the 1990s, which he calls planning as reasoning together or collective design recognizes the inherently political nature of the planning process and the realization that planning is not the activity of a single individual or organization, but an ongoing process of collective design. Information technology should thus work to facilitate this collective design by encouraging "social interaction, interpersonal1 communication, and debate that attempts to achieve collective goals and deal with common concerns." The table below summarizes Klosterman's presentation of the evolving view of planning and the use of information technology: 16 Table 2: Evolving views of planning and information technology 1960s System Optimization "Planning as applied science" Information technology is viewed as providing the information needed for a value-neutral and politically neutral process of "rational" planning. 1970s Politics "Planning as politics" Information technology is seen as inherently political, reinforcing existing structures of influence, hiding fundamental political choices, and transforming the policy-making process. 1980s Discourse "Planning as communication" Information technology and the content of planners' technical analyses are seen as often less important than the ways in which planners transmit this information to others. 1990s Collective Design "Planning as reasoning together" Information technology is seen as providing the information infrastructure that facilitates social interaction, interpersonal communication, and debate that attempts to achieve collective goals and deal with common concerns. (Source: Klosterman, 1997, p.47) Klosterman also presents a number of important practical points on how information technology as a planning support system should support the view of planning as collective design. These points are reviewed in the paragraphs to follow as they are equally applicable to a modeling system. Klosterman's notion of a planning support system is simply a toolbox of technologically relevant applications. Such a system will not consist of GIS alone but must also include other planning tools for urban and regional economic and demographic analysis, forecasting, and transportation planning. In this way, his concept of what a planning support system should be and how it should be designed provides insight into a land use model that aims to incorporate many of the functions of a planning support system. First, although the system should be able to address given short term problems, its function should focus on being primarily a planning system that can address long-range problems and strategic issues. It is recognized that planners are most often dealing with decision problems for the short term and there must be a mechanism to facilitate group 17 interaction and discussion on those issues. As discussed earlier, addressing short-term problems and delivering results may also help to increase the credibility of the project and to secure participation for the longer term strategic issues planners face. Others have observed that stakeholders were attracted to come to the table for immediate and concrete issues suggesting that there should be some issue orientation to the process while continuing to explore general policies and different policy scenarios (Innes, 1996). Most would agree that it is the exploration and discussion of strategic issues where planners can have the most impact on land development and the livability of an urban region. However, planners should be mindful that their organization is often judged by the relative success or failure in its handling of short term problems. Second, Klosterman states that the system should not be a closed black box that receives data inputs and automatically generates plans, forecasts and optimal policy scenarios. Instead, it must "provide the information infrastructure for planning that facilitates interaction among planners, and between planners and other actors, both within and outside the government." He recognizes that without a process for reviewing and discussing the data inputs and outputs, the system will not further communication and understanding among different organizations. The use of GIS plays an important role in that it provides a basis for discussion of the information that most people can understand. This is particularly important when you want to extend the scope of participants, as Klosterman has pointed out, to "outside the government", where expertise in system design and modeling may be significantly less. Other research supports Klosterman's recommendations. The senior planner from San Diego, who was involved with the SANDAG modeling project, argues that forecasts should not be developed in a vacuum, especially if they will be used to evaluate public policy and to develop local and regional plans (Tayman, 1996). He recommends that an extensive forecast review and acceptance process be undertaken with all potential users. Tayman argues, "If users do not understand the nature and role of its [the forecast's] underlying assumptions, then the forecast may be criticized regardless of its empirical accuracy". Thus, he argues that the modeling process and input assumptions being understood and accepted as valid is equally important as the results being reasonable and believable. 18 The third recommendation Klosterman forwards is that the system must be able to support a continuous and interactive process of analysis, design, and evaluation and be able to constantly integrate new information as it is generated. It was stated earlier that past urban models were criticized for not being robust enough to adapt to change in the problem context. Klosterman recognizes that the system must be flexible to adapt to a shift in values and be able to address unforeseen problems and issues as they arise. Planning is an ongoing process and with time the expertise of a group will grow and as a consequence the data and analysis will become more sophisticated. A system that can grow with the group as their level of understanding and expertise increases is attractive. A flexible system also means that the system can be adapted to the particular needs and capabilities of the individuals within a planning organization. In this way, implementation of a system is a process of social interaction between the technology and a particular organizational context (Campbell, 1996 p. 103). A flexible system will not only stand shifts in the problem context but can be adapted to the needs of multiple users. In summary, Klosterman is adamant about the central role an open and accessible information system can play in planning practice: Emerging conceptions of planning as collective design suggest that planners' computer-based tools must be directly accessible to the public and address issues that the public is most interested in (e.g., exploring the likely social, fiscal, and environmental impacts of alternative development proposals) in ways they can readily understand. If this vision is correct, advanced information technologies may finally take their place at the center of professional planning practice, supporting community planning in its fullest and richest forms. (Klosterman, 1997, p.53). 2.4 Consensus Processes Clearly emerging views of planning and information technology as collective design stress the need for open and participatory processes. This is particularly important for planning in a large urban region where many interests have a stake in land development. The planner plays an important role in this respect. The conventional view of the planner (planning as applied science) is to deliver unbiased, professional advice and analysis to the people who need to make the decisions (i.e. the instrumental 19 use of objective information to produce desired outcomes). However, research suggests that formal information and analysis has little impact on decisions (Habermas, 1984). Essentially, planners need to recognize that "being right is not enough". Policy-makers will unlikely set policy in response to new information and analyses provided by planners. Change requires active and ongoing involvement with the appropriate stakeholders. It requires a process that will provide a forum for discussion and debate of the information to reach shared learning on the implications of the information and the choices that must be made. Without such a process change is not likely to occur. Innes also contends that decision-makers do not consciously apply information to make a choice, but that the information frames the choices. Information thus only influences decisions indirectly by changing the mindsets and assumptions of individuals and the practices of those organizations who are setting policy. The process of producing and agreeing on information is therefore crucial. There must be considerable debate among key players for the information to become embedded, in that there is shared meaning and the information is legitimized through discussion of its meanings, its accuracy and its implications. The method of consensus building, or what is also referred to as communicative action, is becoming a more popular method to address complex and controversial public issues where multiple interests are at stake (Innes, 1996). Innes defines consensus building as a method of group deliberation that brings together for face-to-face discussion a significant range of individuals chosen because they represent those with differing stakes in a problem. Important elements of this process are that participants have common information and that all become informed about each other's interests. When the group has an understanding of interests and agreed on facts, they can then explore and create options and ultimately make choices. Processes using a consensus building approach have reported many successes. A 1994 study of 14 cases of processes in growth management throughout California was conducted to advise on the value of consensus building as a coordination technique for growth management initiatives (Innes, 1996). It was determined that in all cases the stakeholders became better informed through the process, and were revealed to value and use their new personal and professional networks to coordinate and collaborate on many projects. In all cases, participants discovered that their interests were 20 interdependent because they all depended on a common economic, fiscal, transportation, or ecological system. Through the consensus building process, participants learned to focus on how to ensure the effective working of that system. In most cases, groups incorporated technical knowledge and analysis into their discussions. Stakeholders could then listen to experts and decide to challenge or confront the data they presented and in the process learn about the data's limitations and decide on its implications. The lengthy deliberations that surrounded this analysis developed collective knowledge and the groups were found to have explored a wide range of factors and their interrelationships, thus increasing understanding of land development process. Innes argues that consensus building is an important tool because it creates social, intellectual and political capital (Innes, 1996). Social capital is created in that personal networks and trust enhance coordination by ensuring informal communication among the stakeholders. Intellectual capital in that there needs to be agreements on data sets, indicators, and an understanding of each other's roles and needs. Her studies revealed that once groups were operating with the same information a principal obstacle to coordination was eliminated. Finally, political capital is created in that new political alliances form to support legislation and ensure implementation. This capital propels further exploration and forms a foundation for approaching any issue that arises. Consensus building processes can thus achieve coordination, not through a top-down exercise of power, but through a horizontal and self-managing process. Even with a well-defined mandate of power at the regional level, the government cannot achieve objectives alone. True cooperation can result only when stakeholders learn that they could produce shared benefits through joint action. Regional planners can be influential in this respect by ensuring an open process, providing technical expertise, and securing the confidence of the group. This last point is important for the regional planner as local government staff are often suspicious of regional government. Local governments sometimes feel coerced to fulfill a regional agenda that they view as detrimental to their own interests. A study in the US examined the negative attitudes of local planners towards regional government to determine how those attitudes affect local planning efforts to implement regional growth management strategies (Baldassare, 1996). The implications of this perception suggest that special efforts must be made to create a better image of regional bodies' responsiveness to local needs. Changing these 21 negative perceptions will help alleviate fears about the loss of local control. This change, of course, takes considerable time and patience and Baldassare calls for "a gradual and flexible approach...to build confidence in regional government". Innes found in her study of consensus building processes that planners were found to have considerable legitimacy among participants, who respected them for essential contributions that they made, especially in the area of technical expertise (Innes, 1996). Innes' observations confirms that planners' technical expertise and knowledge of planning processes will be key in building confidence amongst stakeholders in a process. Planners' use of information technology as a planning support system is thus very important to planning practice and ensuring legitimacy of the profession. There will undoubtedly be fears raised about the increasing use of information technology in planning but it should be recognized that information technology is a reflection of societal values: Those that fear the consequences of the fast-developing information superhighway should be careful not to focus attention on the technology alone; the use and abuse of such systems will simply reflect the dominant beliefs of society and, at a more detailed level, of individual organizations. If we do not like what we see represented in the computing systems being adopted in planning, we should look carefully at the nature of planning practice and the choices we as individuals are making (Campbell, 1996, p.105). Planners also need to be mindful of potential weaknesses in implementation. Research is revealing that stakeholders are reaching consensus on problems and objectives but are having difficulty implementing this consensus. The process of reaching consensus tends to overshadow the process of implementing the agreement (Margerum, 1999). Face-to-face communication has been shown to bind promises to cooperate and thus maintaining good relationships and scheduling regular updates is important for implementing agreed upon goals. In summary, the review of literature provides evidence that new opportunities are emerging for the development and use of urban land use simulation models in planning practice. Technological innovations in computers and GIS allow sophisticated modeling systems to be designed to accurately simulate the complex spatial and temporal 2 2 characteristics of land use growth and change in a large urban region. In addition, despite the greater sophistication of these systems, research suggests the display capabilities and the spatial approach to analysis of GIS-based models are more understandable to people. It would also appear that the evolving view of 'planning as reasoning together' is facilitated by these technological innovations with information technology serving as a basis for interaction, communication and debate on planning issues in the region. 23 3.0 Case Studies of Large-Scale GIS-Based Land Use Models Two case studies of large-scale urban land use models are presented in this section. The first case study looks at the modeling process undertaken by the regional governing body of San Diego, the San Diego Association of Governments (SANDAG). The second case study examines the California Urban Futures Model (CUFM) developed at the Institute of Urban and Regional Development at the University of California at Berkeley. These two modeling projects were selected for further study for a number of reasons. First, they both are large-scale in that they consider the multiple jurisdictions that form an urban region in an attempt to build accurate urban simulation tools. Second, they both innovatively employ GIS technology in their modeling processes to increase understanding and the transparency of the model. Third, they have notable differences in their approach to the modeling process which reveal important considerations for a project undertaken in the Greater Vancouver region. 3.1 Case Study 1: Greater San Diego SANDAG has been producing 25-year forecasts for the region's growth in population, housing and employment since 1971. The forecasts are used by local, state and federal agencies as well as the private sector to guide planning and project decisions. The forecasts also play an important role in the review of their Regional Growth Management Strategy, adopted in 1993, by evaluating different policy scenarios and their impact on urban development patterns (SANDAG, 1994). SANDAG's sophisticated model was chosen as a case study as it makes extensive use of GIS and its modeling process is notable in its consensus-based approach to producing forecasts. The San Diego region has a current population of approximately 2.7 million people spread across 18 cities and county unincorporated areas. SANDAG is a public agency formed voluntarily by local governments to assure overall area-wide planning and coordination for the San Diego region. The Board of Directors has representation from each jurisdiction as well as advisory and liaison members from the State Transportation 24 Department, the US Department of Defense, the Port District, and Tijuana/Baja California. SANDAG's latest forecast (Interim Series 8) covers the period 1990 to the year 2015. The forecast addresses the region's data needs for land use, infrastructure and transportation planning and provides the analytical basis for the Regional Growth Management Strategy. It is used as a key input for the Regional Transportation Plan and is employed in traffic forecasting studies conducted around the region. The forecast is also used to project changes in service levels for public facilities and ascertain the need for new, expanded, or reduced facilities. The forecast is available for use by the private sector for site location studies, determining future demand for products and services, and market penetration analysis. This modeling system, however, delivers more than just a forecast. It also has the ability to simulate the implications of alternative public policy and land-use decisions. To this end, the SANDAG model is being used to evaluate alternative land-use configurations and transportation networks. These simulations include intensifying residential and employment development around rail transit stations and along bus corridors. SANDAG produces its Regional Growth Forecast in two stages. In stage one, forecasts for population and employment growth are prepared for the entire San Diego region. This region-wide forecast is based on a synthesis of two well-known forecasting techniques, the cohort-survival method for population and econometric methods for the 1 economic factors. In stage two, the region-wide forecast is allocated to many smaller geographic areas by a series of subregional allocation models which distribute the regional totals according to attractions and constraints provided by existing and planned land use policies, transportation networks and population and economic concentrations. The smallest geographic units in the SANDAG system result from the multiple overlay of GIS spatial information. The subregional allocation process is undertaken by three models at two levels of geography. First, forecast data is allocated to 204 urban modeling zones. These zones represent groups of census tracts and are contiguous with city boundaries or community plan boundaries. A land use model allocates population and housing units to each of 25 the 204 zones. This allocation is determined by planned land uses, inter-zonal travel times, and the relative proximity to employment concentrations. Residential growth in a zone is limited by its calculated development capacity, which is measured by availability of vacant land, additional allowable residential densities, and redevelopment potential. An employment location model allocates forecasted employment by industry type according to attractions and constraints derived from existing employment concentrations and planned land uses. Second, forecast data is allocated to the smallest units of geography using the 204 zones as control totals. This model also uses travel-time and land-use characteristics as the basis for forecast allocation. The model identifies the most attractive geographic unit with capacity for development and 'fills' it until either the zonal control is reached or the unit reaches capacity. When the unit is filled, the model identifies the next most attractive unit and continues until growth has been completely allocated. The small size of the smallest geographic units allows them to be easily aggregated to approximate any user-defined area, yet the incorporation of a second level of geography ensured the forecast information was consistent across all levels of geography. The GIS layers used in the model include general or community plans, street networks, existing land use, potential redevelopment areas, constrained development areas (e.g. steep slopes, wetlands, airport noise contours), and site-specific projects. In order to verify the information collected, a map set was produced for each local government and reviewed with local planners for accuracy. After the review and editing process was complete, these information layers were then overlaid on each other. The overlay process identified lands that were not suitable for development and thus blocked from receiving allocated growth. These "removed" areas included developed land, constrained land, and land set aside for future streets. The remaining areas represented developable land, including land identified for redevelopment, where the sub-allocation models could place future development. Jeff Tayman, the senior demographer with SANDAG, argues that his organization's approach to their forecasting and modeling process helps build consensus (Tayman, 1996). Each forecast undergoes extensive review and evaluation process before the results are adopted by the SANDAG Board. In addition, all land use inputs are reviewed and corrected by the planning staff of all local governments in the region. The use of 26 GIS overlays as the base architecture of the allocation model was found to increase the understanding of the model by local staff. Throughout the review process, the roles of the land-use information and the modeling process are explained, questions are answered and concerns about the growth patterns portrayed in the forecast are resolved. The reviews generally last 6 months and at the end, the final forecast is produced and adopted for use. Tayman argues that this process of review and acceptance by local planning staff and decision-makers is an essential step for achieving consensus and ensuring that all the players, local and regional, use a single forecast on which to base the development and evaluation of regional and local plans and policies. Only in this way, it is believed, can a growth management strategy be truly effective. As a potent symbol of its success in attracting partnerships and its general philosophy of openness, all SANDAG GIS data files and maps can be downloaded directly from the Internet on SANDAG's Home Page. These are provided free to anyone who has a system to manipulate the data for analysis. Tayman argues that this open sharing of information helps SANDAG in a number of ways. Firstly, it serves to invite constant improvements to the data since SANDAG encourages users to identify inaccuracies in the datasets to them for correction. Secondly, any agency, firm, or individual contemplating starting a GIS program will strive for compatibility with the SANDAG system due to the large knowledge base that currently exists. This fact serves to further reinforce the predominance of the SANDAG system thus increasing potential for enhanced communication and increased coordination amongst the land use interests. Thirdly, the easy access to information increases the credibility of the organization and, in turn, the forecasts it produces. Credible forecasts will increase debates on how forecasts can be either accommodated or averted through coordinated policy action. The use of GIS plays an important role in SANDAG's forecasting methodology. SANDAG has been using GIS for almost twenty-five years to assist in regional and local planning efforts and continues to be on the cutting edge in pioneering innovative approaches to GIS database development, maintenance, analysis and display. Due to the high costs often associated with developing GIS data, and the limited funds available to agencies, SANDAG has formed successful partnerships to develop and purchase data, and has instituted cooperative data sharing agreements with a number of federal, state and local agencies. 27 In summary, this case study reveals that forecasts using large-scale urban models is not simply a technical process, but very much a planning process in which land use stakeholders are brought into the modeling process to share and debate data inputs and to review and evaluate forecast results. SANDAG's forecasts have met with considerable success by involving local governments in the modeling process from beginning to end and extending its use to not only public agencies, but to private firms and individuals as well. Furthermore, SANDAG's partnerships, open sharing of information, and commitment to a lengthy consultation and review process with its member organizations attest their dedication to a consensus-based approach. 3.2 Case Study 2: Greater San Francisco Bay Area The California Urban Futures Model (CUFM) project was initiated at the request of the California legislature who approached UC Berkeley to design a computerized land use modeling system capable of simulating alternative growth scenarios for the San Francisco Bay Area. The model would help planners and decision makers see clearly how different local, county, and regional development policies shape the location, pattern, and density of development in the San Francisco Bay Region (Landis, 1994). In addition, the model would provide a framework for simulating how realistic development policies applied locally and regionally might alter the pattern of urban development in the Greater Bay Area (Landis, 1992). The CUFM was chosen as a second case study as it also uses GIS for spatial analysis and data integration. In contrast to the SANDAG model, CUFM was developed at an educational institution and was developed primarily as a simulation tool as opposed to the SANDAG model developed primarily for providing information for planning implementation. The CUFM considers 14 counties, an area in which forecasters believe will experience growth of approximately 2.5 inhabitants by 2010. During the development phase of the model, only the nine-county San Francisco Bay Area was considered for study. However, since then 3 other urban areas have been added (Sacramento area, Stockton-Modesto area, and Sant Cruz-Monterey area) as work-related commuting has increased. The economic and social integration of these communities with the Bay area is blurring traditional regional boundaries. When the CUFM project was started, it was 28 reportedly the first time that the original 14 counties had been grouped together as a single region, which was called the Greater San Francisco Bay Region. In 1991, there were more than 150 city and county governments in this defined region. Similar to the SANDAG modeling process, the CUFM model first forecasts how much growth will occur in the region and then secondly determines where that growth will occur. Unlike the SANDAG model, however, CUFM takes a bottom-up approach by forecasting growth at the city or sub-area levej, and then aggregating upward to produce a regional growth total. The methodology supports the idea that local policy initiatives affect not only the location of population and employment growth, but also its size and quality. The separation also allows a trend-model to be used for growth forecasting, or the demand side of the equation, while employing a decision-rule based model for growth allocation, the supply side of the model. The trend-model generates five-year population growth forecasts for each city in the study region. An allocation model then distributes the city-level forecast to small geographic areas named Developable Land Units (DLUs), which describe developed or underdeveloped land that have potential for development or redevelopment. Similar to the SANDAG model, the CUFM relies on highly accurate computer map layers as a source of data and as a mechanism for analysis. These map layers are encoded, updated, and incorporated into the analytical and decision-making components of CUFM through the use of GIS. The Institute's GIS system is an integral part of the model and the transparency it lends to the modeling process makes it easier to communicate results to decision-makers and the general public. The smallest geographic units result from layering GIS datasets that qualify the land with environmental, market and policy attributes. These layers include roads, city boundaries, steep slopes, hydrology, highway buffers, urban buffers, agricultural lands, wetlands, and sewer and water utility service areas. DLUs that are unsuitable for development due to environmental, ownership or public policy reasons are eliminated from consideration. The remaining eligible DLUs are given a ranking for their attractiveness to development based on a set of decision rules. Growth is allocated to the most attractive DLU until it reaches capacity at which point the model searches for the next eligible DLU to allocate growth and so on until the population growth for that city is fully allocated. 29 One interesting component of the model is its simulation of growth "spill-overs". If there is insufficient land in the city to accommodate all the forecasted growth, then the remainder is accumulated for potential re-allocation into a neighbouring municipality or unincorporated area, accurately simulating an observed phenomenon. The coordinator of the CUFM project outlined four principles that guided the development of CUFM (Landis, 1992). The first was that it was foremost a spatial simulation system and not a regional forecasting model. This is in contrast to the SANDAG model whose primary purpose is to produce regional forecasts. Consistent with this principle, the model rejects the use of zonal or county-level growth totals that were too large in area to provide a clear picture of the spatial processes of urban growth. The emphasis on spatial simulation inspired the use of GIS in order to handle the complexity and volume of data that was to be incorporated into the model. GIS also allows the model to simulate specific locations where growth might occur in reality. The second principle was to adopt a policy-relevant approach, that is, the model should be designed to simulate alternative development futures. Not only must the model be able to realistically simulate the impact of a wide range of real policy proposals (e.g. dpwnzoning land, construction of a new transportation line, rezoning land for housing) but must be capable of simulating the impacts of policies undertaken simultaneously by various land use interests. The model thus allows policy simulations to be achieved by adding new spatial features or map layers to the GIS system that change the attractiveness of each DLU to development. The third principle was to deliver a tool useable by and understandable to planners and policy analysts and not just the technicians that produced it. To this effect the growth allocation mechanism in the model was designed around a series of transparent and changeable decision rules, rather than mathematical algorithms. In this way, the model was to be made more accessible to more people. The final principle called for an expandable system that could be updated in a modular fashion. The model was designed as a system of related but independent models with the view that as better data or decision rules become available, they could be easily integrated into the existing framework. 30 The model that resulted was thus specifically designed to simulate how growth planning policies would affect the location, pattern and density of new development in the Bay Region. Once this capability was established, the project team looked at developing a set of realistic yet different growth scenarios that could potentially be adopted and played out. The model was then used to simulate how the different growth policy scenarios would impact development patterns in the region by the year 2010. The purpose of CUFM was to give substance to alternative policy scenarios so that policy makers can make more informed choices. Decisions would be embedded in an understanding of cause and effect in terms of the development patterns. In order to clearly demonstrate the effect of different policy sets, eight policy scenarios were developed in-house by the modeling team. The eight policy sets fall into three broad categories: business-as-usual, environmental protection, and compact growth. The scenarios studied revealed that development policies, if applied uniformly throughout the region, wili have different effects in different locations. The model displays well the inter-dependency of growth policies in that restricting growth in one jurisdiction will, in most cases, cause that growth to spill over into an adjacent jurisdiction. Although the resulting development patterns differed substantially, the modeling team did make clear, however, that the implementation of a majority of the policy scenarios put forward is not politically feasible under the current system of metropolitan governance. The Greater Bay Region currently has no agency to coordinate land use policies among the numerous cities and counties. When the project director of CUFM, John Landis, was asked whether the model was being used outside the university, he regretted that the model had failed to create buy-in from the land use interests in the region and had consequently not been adopted as a planning tool3. The project had remained an academic exercise and had not created debate amongst the land use interests necessary for the development of a shared understanding. Landis contended that this type of planning model needs to reside in and become a tool of a regional planning agency that he believes has closer relationships with local jurisdictions as well as utility and service providers. In contrast to the modeling process of the SANDAG model, the CUFM process did not emphasize a consensus-based approach by implementing a review and evaluation process where 3 Meeting with John Landis, Project Director CUFM Project, at University of California at Berkeley on May 6, 1997. 31 planning organizations were invited to challenge and debate the data inputs and outputs. In some respects, the potential positive impact the CUFM findings could have had on local decision-making was compromised by the research-based imperatives of the project. In summary, the two case studies present examples where GIS-based land use models have been developed to address complex planning issues that traverse multiple land use and infrastructure planning jurisdictions across large urban regions. Although both have contributed much to methodological theory on the development and use of GIS-based, large-scale computerized urban simulation models, the experiences of the two examples bring to light an important consideration for other urban regions undertaking similar modeling projects. If the success of a modeling project will be measured by the positive impact the information has had on planning implementation in the region, then the SANDAG approach would appear to be a clear winner. The SANDAG modeling team focuses on building consensus throughout the modeling process by committing to an extensive review and evaluation process of model inputs and outputs and with open sharing of information to both inside and outside the membership. The next section explores these issues further front the point of view of representatives from public sector land use and infrastructure planning organizations in the Greater Vancouver region. 32 4.0 Stakeholder Issues for the Development of a GIS-Based Land Use Model for Greater Vancouver 4.1 Background The Greater Vancouver region has multiple public sector organizations planning for changing land use and infrastructure needs. Land use planning remains primarily under direct control of local municipalities with some exceptions, most notably the provincial Agricultural Land Commission whose mandate is to protect prime agricultural land in the region from urban development. The majority of significant infrastructure planning projects are undertaken by regional service and utility organizations, addressing needs in the area of electrical generation and transmission (BC Hydro), wastewater removal and treatment (Greater Vancouver Sewerage and Drainage District), and public transportation and other transportation infrastructure (TransLink) amongst others. Land use and infrastructure planning are closely linked in that changes in land use will alter the demand for services and infrastructure, and similarly the supply of infrastructure will create pressure to change land uses. In this way, the policies that each organization adopts, whether on the land use or infrastructure side, will have an impact on the plans of other planning organizations in the region. Given this fact, the ability to increase communication and help develop shared understanding amongst the numerous organizations is very important: This is not an easy task with the large number of public sector land use and infrastructure planning organizations in the Greater Vancouver region. As discussed in the introductory chapter, the geographic definition of Greater Vancouver varies, but most would agree that the area covered by the GVRD is too small an area for effective regional planning. Even so, the GVRD has 20 member municipalities and two electoral areas within its boundaries. The regional district that continues east up the river valley, the Fraser Valley Regional District (FVRD), represents a further three large municipalities. If the definition of the Vancouver region is to be broadened to consider areas of economic and social functional relationships then this number increases again. Also consider that although the GVRD is responsible for planning major regional 33 services in the areas of water delivery, wastewater drainage and treatment, and solid waste removal, the GVRD consists of five distinct legal entities in which services are either directly operated, contracted to the private sector, or handled through municipal partnerships. In this context of multiple planning organizations, the need for communication and coordination of plans is essential for effective regional planning to occur. For this reason, there has been considerable interest from the land use and infrastructure public sector planning field for the development of a land use urban simulation model. In 1996, a land use modeling project was initiated to address specific information needs of the Greater Vancouver Sewerage and Drainage District (GVSDD). The GVSDD required forecasted demographic and economic data for infrastructure modeling and planning. However, due to the long lead times in sewerage planning and implementation, a longer time horizon than was available at that time was needed. In addition, the area of geographic study differed from that in common use by the GVRD. Recognizing that the needs of the GVSDD were specific to that organization, the idea of developing a land use modeling system that could respond to the planning information needs of multiple users was put forward. The development of a prototype model was jointly undertaken by the GVSDD, the GVRD Policy and Planning Department, and The Urban Futures Institute (TUFI). A prototype was developed to test the functionality of a model that combines spatial land use information attached to small geographic areas and simple decision rules to allocate projected growth in dwelling units, population and employment.4 The prototype model uses GIS since it was determined to be the most effective way of integrating the large amounts of spatial information necessary for accurately simulating the complex spatial relationships in the region. The small geographic areas were formed by superimposing a 100 meter by 100 meter grid ('gridcells') over the entire region. These gridcells can be aggregated to reasonably approximate any user-defined area of analysis. Gridcells are ranked in terms of their relative potential for growth, determined by an agreed upon weighting of factors. Factors used in the prototype model were accessibility to employment, availability of sewers, and population density. 4 For more information, consult The Urban Futures Institute report "Documentation of Phase One in the Development of a Prototype Geographical Information System Spatial Land-use Modeling Process for the Metropolitan Vancouver Region", January 1997. 34 GIS datasets give each gridcell land use attributes that further qualifies whether growth can occur in a gridcell, and if so, how much growth each cell can reasonably accept. These attributes include environmental information such as bodies of water, wetlands and steep slopes and policy information such as parks, agricultural land, and residential density. It is possible, therefore, that a gridcell may be attractive for development in terms of its ranking, by virtue of good accessibility and low population density, but rejects allocated growth due to environmental or other development restrictions. Only a change in policy will allow the acceptance of allocated growth into a gridcell where growth was previously denied. The prototype demonstrated that many of the past technical obstacles associated with building a region-wide land use simulation model could be overcome with recent technological advancements. Given the success and interest generated by the prototype, interviews were undertaken with selected public sector land use and infrastructure planning organizations to determine how such a model could be used as a planning tool in the region. More specifically, representatives were asked to identify their needs and expectations as to the functionality of a regional land use model and the process by which such a model should be developed. The findings of the interviews are discussed in the sections to follow. 4.2 Research Findings 4.2.1 Structure of the Model: Use of GIS and Small Geographic Areas Each interview started with an exploration of how a regional land use model that uses GIS and provides detailed information for small geographic areas would address information needs within each public sector planning organization. Planning organizations have different roles and needs in addressing land use and infrastructure issues in the region. The use of GIS and small geographic areas are thought to increase understanding of modeling systems and increase their utility by rendering them more flexible. 35 All those interviewed were positive towards the development of a GIS-based regional land use model and believed that the model could serve many functions at the local and regional planning level. The perceived strengths of a GIS-based model, however, differed somewhat. For some, the use of small geographic areas containing detailed land use information that can be aggregated to any desired level of analysis was seen as the most important feature of the model while other thought the use of GIS as the basis for the model was the key feature since GIS can "think" and represent graphically complex urban relationships. Only two of the participants addressed the value of a model that covers the entire metropolitan area. Perhaps this simply reflects an implicit recognition that growth crosses municipal boundaries and even regional boundaries. As the representative of one of the regional waste water management agencies, the Greater Vancouver Sewerage and Drainage District (GVSDD) stated, "obviously we don't just stop all our activities at the political boundaries". He believed the region-wide scope of the model would be an essential component since the model would be able to show the impact the policy of one municipality would have on another and how uncoordinated land use and infrastructure plans have cumulative effects, often harmful. Understanding these relationships is achieved most effectively and realistically by looking at an entire urban system. Such a system does not respect municipal nor regional government jurisdictional boundaries. The two main areas of discussion in terms of the structure of the model, however, were the use of small geographic areas and GIS. These are discussed separately in the two sections to follow. 4.2.1.1 Small geographic areas More than half of those interviewed commented on the model's use of small geographic areas as the basic unit. Each geographic area or 'gridcell' represents a 100m by 100m parcei of land and acts as the basic unit to which land use information is attributed. The gridcell was seen as an essential component of the model since it not only meets the ' information needs for small area analysis but gives users the flexibility to aggregate the data to any desired level of analysis. 36 The ability to easily aggregate data is important since each organization studies specific systems whose boundaries often differ considerably. Thus, whether the area of study be a sewer catchment area, a solid waste transfer zone, a traffic zone, a municipality or a town centre, the model allows the user to define how the region has been divided. Said a senior member of the Policy and Planning Department at the GVRD, "the finer the grain, the more useful (the model) is to a broader segment because people can build it up to different aggregations". These views were shared by a planner at BC Hydro's Power Planning Division. "The small area grid is essential", he stated, because "the grid allows you to aggregate to any level." He explained that the spatial relationships important to one organization do not necessarily have meaning to another, and thus the modeling system must be flexible to adapt to the varying needs of different users. The small area grid is particularly useful at the municipal planning level where discreteness of data is quite important. A representative from the City of Surrey's Planning Department said that the model "has to be realistic from the city's point of view". Demographic data such as population levels, housing numbers by structure type and labour force by industry type are presented by the GVRD in 675 traffic zones. Traffic zones may be an adequate level of analysis in studies covering the entire region, however, he maintained they do not allow for detailed analysis at the municipal planning level. A senior planner at the City of Burnaby shared these views and offered an example of the planning analysis undertaken by his department. The City of Burnaby is divided into 38 neighbourhood planning areas which he explained allows for a detailed level of analysis while being large enough to maintain the reliability of their base information. For each planning area, four policy scenarios varying from a laissez-faire to an aggressive regulatory approach are established to define a potential range of growth, or capacity. Currently, this capacity analysis for each neighbourhood is undertaken manually and is very time consuming. Consequently, comprehensive studies are performed infrequently and the planning areas cannot be closely followed to determine the impact of adopted policies. He suggested that if this analysis could be achieved using information generated by a detailed model, the growth patterns in the City could be more closely monitored. The potential of the model for regional planning was also 37 not missed on this planner. He suggested that if this modeling exercise was replicated in each municipality in the region, the regional growth potential under current policies could be readily calculated - "Everyone would know exactly where they stand". One disadvantage of creating a model with small units of analysis, or a 'finer grained model', is the cost. One participant stated that for this very reason many modelers with constrained resources will choose to develop a coarser grain model. In many cases, this smoothing out of data may not affect the end planning result, but the model will not be as flexible and robust. In the modeling experience of this planner, however, "the benefits of getting more discretized information far outweigh the marginal cost of doing so". He believed a fine grain structure is essential for a model that is being developed to address the needs of multiple users but stressed the cost will be an issue. Funding of a modeling project was raised by a number of representatives as an issue and will be discussed in a later section. A second disadvantage of using small geographic areas as a basic unit of the model was forwarded by a senior planner at the GVRD. He expressed concerns about providing detailed land use information at the gridcell level. "The challenge of a fine grain model is that people look at it in a fine grain way...if it produces data on a certain area that does not pass the test of reasonableness for somebody who works or lives in that particular area, it tends to invalidate the model as a whole, whereas if the information is always tested and aggregated in some way, the data will balance out over larger areas than the 100m by 100m grid." When he refers to a "test of reasonableness", he is referring to the comfort level with inaccuracies at the gridcell level and how to minimize these inaccuracies. This is a difficult issue as the purpose of the small geographic areas is not as a unit of analysis but to allow the user to aggregate information to meet their specific needs. Making data available only in an aggregated format may unintentionally encumber the desired flexibility of the model and frustrate the efforts of potential users. Also, suppressing the availability of data at the gridcell level may arouse suspicions in those eager to find fault. It is possible that people may try to discredit the model based on inaccuracies at the gridcell level, but users must be made aware that in any modeling process the results are only as good as the inputs. Educating users on this basic concept of modeling may be a better approach to handling inaccuracies than trying to define and implement a "test of reasonableness". 38 39 4.2.1.2 A GIS-based system One of the characteristics of this land use model that distinguishes it from other large scale urban models is its use of GIS. Half of those interviewed identified the use of GIS as a key component of the model. This is surprising considering that GIS is in use at every organization represented in the interviews except one. There are a couple of possible reasons for this. Although GIS has made an appearance in many planning organizations due to the decreased capital costs of implementing a system, GIS is still used most commonly as a tool for graphic representation and simple spatial database queries and not as a system capable of analysing spatial information and addressing specific policy objectives. Since few planners are adequately trained to navigate these i systems, the only interaction many people have is with the output which represents only a fraction of the information contained in the system and may represent just one of many analytical scenarios. Designing and implementing a program to capitalize on GIS's capabilities beyond mapping is time consuming, requires detailed knowledge of the system, and requires well defined project objectives. Unfortunately, municipal and regional governments currently have few resources to obtain employee training and to undertake GIS-based projects. Another possible reason may simply be that the interviewee may not have had any specific involvement with projects involving GIS even though GIS is being used within their organization for planning analysis. One regional planner who stressed the value of using a GIS system in land use modeling, stated that the ability to measure spatial relationships is essential since establishing where growth is going to be allocated in the region is fundamentally spatially related. Such a system would allow land use and infrastructure planners "to link spatial data from different sources in order to think spatially, and by extension, help policy-makers also think spatially". A second regional planner, who has considerable experience working with large-scale models, stated that the methodology of using GIS is good since the system is sophisticated enough to handle the fine grained information necessary for good results. He also stated that such a system has greater flexibility than other types of large scale urban models thus allowing the model to grow and change as needed. 40 A senior planner in a municipal planning department stated that the visual capabilities of GIS would make the model easier to understand by more people, especially the general public. Referring to the power of graphic representation, she said "when they say (a visual) is worth a thousand words, it is". She recognized that it is often much easier for people to communicate ideas and visions through graphics than through text or charts. "People don't fall in love with numbers, they fall in love with ideas and they fall in love with visions and that is what will motivate them". Another municipal planner also pointed to GIS as a good basis for the model since he believes such a model would be simpler to use and easier to understand than other models. He makes reference to the GVRD transportation model that he maintained conceals too many assumptions. Although all models inherently carry assumptions, GIS-based systems offer relative transparency since the datasets that qualify the land for potential land use change are graphically represented. In this way, GIS makes urban modeling less complicated to the average person. These views were echoed by another planner who also believed that a GIS-based system is superior since its graphic base makes the model assumptions far more understandable. He felt that the assumptions in the current method for allocating growth in the Greater Vancouver region are too obscure an as a result the target numbers for growth that are produced for each local jurisdiction are not defensible. A GIS-based approach appealed to this planner since assumptions in the model can be changed by adding, removing or altering the policy layers, thus changing the attractiveness of the gridcells for development. Ultimately this allows for the exploration of the impact of potential growth patterns of different policy scenarios. The ability of a GIS-based model to change assumptions in an understandable way suggests the model could be an effective tool for policy explorations. 4.2.2 Utility of a GIS-Based Land Use Model as a Planning Tool The second area explored with the representatives of public sector planning organizations was how a GIS-based land use model could be used as a planning tool either within their organization or in the region as a whole. All those interviewed expressed a strong desire that the model should increase understanding in the way land 41 use changes in the Greater Vancouver region. This understanding, they asserted, is needed to help planners and decision-makers see clearly how different provincial, regional and local land use policies shape the location, pattern and density of development. Thus, they had a keen awareness that policies adopted by one land use authority ultimately have impacts on other jurisdictions in the region. New legislation governing regional growth management plans and the consequent requirement for local governments to include regional context statements in their community plans has most likely contributed to this awareness. At the local planning level, municipalities often have a considerable amount of detailed information on land use and transportation patterns in their communities. However, the level of detail and format of this information is not standard across municipalities, making comparison and cumulative effects difficult to ascertain on a regional level. There are many basic questions about land use as it exists today in the region that cannot be readily answered. For example, these questions may include: What percentage of residential land in the region is in the form of single family detached housing? What is the current capacity for growth under existing plans by housing type and population in the region? What has been the change in population densities around transit stations over the last 10 years compared to other areas? A GIS-based model could help answer some of these questions and inform the policy making process. Without an assessment of existing conditions and an understanding of the policies that led to those land development patterns, it is difficult to accurately project conditions in the future. A land use planner at the Agricultural Land Commission (ALC), a Provincial agency, recognized the importance of detailed land use information in reviewing changes in land uses within the Agricultural Land Reserve (ALR). The ALR represents a patchwork of land stretching across the valley crossing municipal and regional boundaries. i "Understanding what is happening internally to the ALR is something that we are very interested in", he said "one might say that its in the Green Zone so its outside the sphere of urban development potential and so there's a relaxed feeling that its taken care of but in fact the dynamics of land use are actually eroding this land base from within". This planner was interested in understanding the internal and external forces that are fragmenting the land base so that he could suggest changes in policy to protect the ALR lands. 42 There were four main areas of interest that the interview participants thought the model should be able to address. The first was to provide a level of certainty for the location and timing of growth so that scarce resources could be allocated more effectively. This is particularly useful for testing investment-oriented policies (e.g., construction of a new wastewater facility or an electrical substation). The second was the ability to perform policy analysis to test the impact of different policies on land development and thus inform the development of new plans or objectives. This would include forecasts for different geographical areas and testing a range of hypothetical "what if scenarios. The third area of interest was a tool to monitor whether the plans and objectives of adopted policies are being met. Closely tracking how well a municipality or a region is doing in meeting objectives can provide crucial information for how new policies may be made more effective. Monitoring can also establish whether targets can realistically be met in the time horizon under study, or whether policy changes are needed to address missed targets. Whatever the approach, the monitoring of plans is essential for successful land use planning policy. Finally, many of those interviewed pointed to the valuable role the model could play in enhancing communication by increasing understanding of the process of land use i change and jointly exploring alternative policy scenarios. Indeed, since the model covers multiple jurisdictions and addresses a wide range of planning objectives (land use, transportation, and infrastructure), communication between interests is essential. A shared understanding of land development will lead to more success in developing policies that re-enforce each other, instead of working against each other. These four areas of interest will be explored further in the sections to follow. 4.2.2.1 Increasing certainty Approximately half of those interviewed pointed to the need for increasing certainty in the timing and direction of growth and land use change as an area in which the model would be useful. Not unsurprisingly, three of these responses came from representatives of organizations that make heavy investments in infrastructure: TransLink, BC Hydro, and the GVSDD. 43 A Power Planning specialist at BC Hydro stated that the size and timing of infrastructure investments are crucial since oversizing and premature construction of infrastructure deplete limited investment capital. Conversely, undersizing infrastructure results in costly emergency measures to meet unexpected demand. He argued that making tough financial decisions on where to spend finite resources is an easier task when a tool is available that can provide greater certainty where these resources will be most needed. His organization has attempted to perform small-area load forecasting, which is based on small-area population and employment forecasts, to analyse where new or upgraded substations will be needed. Unfortunately they have met with little success: "Our ability to do this type of modeling is limited, after 5 years its a wild shot in the dark". He recognized that when utilities and other infrastructure organizations have dramatically different deployment plans, then an infrastructure deficit or surplus is usually the result. These divergent plans can also impact community plans adopted by the local municipality by suppressing or encouraging growth that is inconsistent with those plans; The need for greater certainty was also put forward by the representative with the GVSDD. He emphasized the imperative that his organization deliver sewerage and drainage infrastructure at the appropriate time and with adequate capacity: "We need some short term certainty in the way (growth) is going to occur and then longer term 'what ifs' so we can be opportunistic around planning our facilities". His organization's informational needs are twofold. First, the time horizon for planning and building infrastructure improvements to the GVSDD system is at least three to five years and thus planning short term growth requires a high degree of certainty. Second, longer term planning scenarios that land use stakeholders have agreed upon and support may lead to opportunities to build additional capacity when other non-related upgrades are being undertaken. For example, a road resurfacing project may provide an opportunity to install sewerage infrastructure that is projected to be needed to meet demand in seven years time. The savings in future construction costs may outweigh the costs of premature installation. The senior transportation planner at TransLink commented that the most significant planning work that TransLink undertakes is determining where new transit lines are to be constructed. He pointed out that although travel patterns have changed significantly, 44 TransLink's core business remains moving people from their homes to their workplace. Generally, the greater the population density, the more feasible is the provision of transit: "The intensity of land use is the most important factor we are concerned about". Greater certainty as to where future population and employment concentrations will be located in the region is thus essential for planning transit infrastructure and making good investment decisions. One senior municipal planner expressed some skepticism of the accuracy of growth modeling considering the complex interrelationships of regional scale planning. However, she added that that this skepticism should not preclude the development of a tool of some description to increase understanding. "None of us can afford to make huge mistakes" she says, "so we need to have a system that has us all working primarily on the same wavelength." Her comments suggest that the empirical accuracy of the model may not be as important as bringing public sector stakeholders together to discuss issues openly and coordinate their plans. The "huge mistakes" can only be avoided by having some certainty as to the direction and timing of growth in the region. Achieving certainty is not simply discovering the future, but creating and agreeing on a future through the exploration of alternative futures. This leads to the second area of interest which is the ability to test policy and create 'what if scenarios for investigating alternative futures. 4.2.2.2 Testing policy and producing "what if scenarios The second major area that interview participants thought the model should be able to address is the testing of policy and the ability to produce 'what if scenarios. One regional planner said the model should have the ability "to do explorations of the future based upon different sets of assumptions". Scenario testing, he argued, would allow the user to ask the model the likely land development patterns resulting from a hypothetical land use policy or set of policies. These explorations would also help the user understand more clearly the relationships between policy decisions and land development and provide an opportunity to explore alternative futures. Another planner stressed the importance of a model that allows the exploration of alternative futures and not one that simply calculates and outputs a future based on the 45 input assumptions. "We don't want a crystal ball...the ability to explore is of paramount importance." These explorations thus increase understanding and allow policies and actions to be effectively designed to achieve the desired outcomes. Instead of simply discovering the future, the crystal ball effect, the model becomes a tool for exploring and creating a future. He also pointed out that explorations of worst case scenarios can "give incentives to change a dystopic future". These dark windows into an undesirable outcome, he argued, may help to qualify the value of the decisions at hand and give greater urgency to stakeholders to work together in achieving a greater common goal. One planner offered an example where scenario testing could be very useful. "It would be interesting to develop criteria to demonstrate those areas that could accommodate multiple uses... I am not thinking of a transformation of uses...but rather more layering of uses where layering does not now occur." He referred to the example of a commercial area that is redeveloped to introduce residential as a new component while retaining its current level of commercial activity. The cumulative effects of targeted "layering" of uses across the entire region could be quite dramatic. If growth can be accommodated by intensifying land uses in already developed areas, this may relieve pressure for development of ecologically or socially important lands such as wetlands, agricultural and recreational lands. A municipal planner pointed to the value of scenario testing. She would also like to explore how small changes to existing land uses applied across all municipalities could translate into a significant number of new homes without sacrificing livability. Some suggested examples of strategies to be studied could be the promotion of secondary suites in low-density residential areas and the reclamation of "grey" land, such as surface parking, for residential development. Testing these types of scenarios, she argues, may go a long way in convincing people that with coordinated land-use planning growth can be accommodated without a significant loss in quality of life. As discussed in an earlier section, a model that uses GIS and small geographic areas is well suited to policy analysis. Each gridcell represented in the model is the result of land use attributes and decision rules that qualify the attractiveness of that area to growth and land use change. Policy layers can be added and removed to either lower or increase the attractiveness of an area to land use change. Examples of policies that hinder development are steep slopes, wetland protection areas or density caps, 46 whereas policies that encourage development may be the removal of density caps or the introduction of mixed uses in single use areas. The results of a Change in policy can be analysed and compared to other policy scenarios tested. 4.2.2.3 Monitoring the success of adopted plans At least half of the participants also pointed to a third area of need where a GIS-based model may be useful. The model could be used as a monitoring tool to study how well stated objectives and targets of adopted policies or plans are being achieved. Targets are markers to provide a point of comparison, with the final results almost always falling short or exceeding the targets prescribed. A significantly missed target can prompt two kinds of responses. Either targets are changed to reflect what is believed to be the best possible result, or, changes are made to the policy scenario to ensure subsequent targets are met. Thus, the ability to monitor plans allows for their evolution over time based on the success of varying policies and the further exploration of different hypothetical scenarios. Too often, however, there is no mechanism in place to monitor the effectiveness of those plans. One planner stated that the model could "provide an independent, dispassionate point of view and show that what [a local planning organization] is proposing is taking us off course and now we are going to have to make a big effort everywhere else to keep us still on course". Her comment suggests that there is a need for an analytical tool that can monitor plans and inform without bias the discussion as to the impact of new policies on adopted plans. Essentially, empirical evidence presented to a stakeholder who is proposing a potentially damaging policy may have greater credibility and could potentially have greater success in sobering the discussion than unsubstantiated accusations and objections presented by other stakeholders. In essence, if there is buy-in into the model, then it would be difficult to argue with the numerical evidence it supplies as to the impacts of a proposed policy. One planner stressed that not only should regional planning strategies be monitored but municipal and even neighbourhood strategies as well. He saw monitoring job creation in the municipal town centres at the municipal level as being just as important as 47 monitoring the effectiveness of environmental protection policies at the regional level. He argued that the closer the monitoring is engaged at the municipal level, the greater the likelihood that regional targets will be met as well. Provincial legislation requires that municipal plans, or Official Community Plans (OCPs), be revised every five years. Every OCP must contain a consistency clause, or Regional Context Statement, that confirms that the policies and objectives outlined in the plan are consistent with the objectives set in the regional plan (Livable Region Strategic Plan). For this reason, this planner asserts "it is the success of the municipal plans that will largely determine the success of the regional plan." 4.2.2.4 Enhancing Communication Finally, approximately half the interview participants pointed out that the model could serve as a tool to facilitate communication between organizations involved in public sector land use and infrastructure planning. They believed the model could provide the basis for bringing representatives from all organizations together to develop base data and assumptions, explore land development relationships, examine alternative policy scenarios, and monitor the relative success of adopted plans. Their interaction would help develop a shared understanding of the process of land use change and provide the basis for enhanced communication and coordination. Communication between public sector planning organizations in the Greater Vancouver region is hindered in two ways. First, there is a lack of understanding of each organization's interests, roles, and needs. Mutual understanding and appreciation is not necessarily easily achieved when the organizations have conflicting objectives and their individual representatives have different backgrounds and experiences. A planner with the ALC commented on the difficulty he often encounters in communicating the value of protecting agricultural land. He commented "politicians, planners and land developers are at least a generation or two removed from any direct farming...so their orientation from the outset is urban. Very few agricultural people have a feel for planning and zoning and fewer and fewer planning and zoning people have a feel for agriculture". Increased interaction would help broaden understanding amongst stakeholder interests in the region. Second, communication is frustrated by the lack or perceived lack of an 48 impartial forum for the open exploration of ideas and issues. Open communication is difficult in an environment where organizational agenda's appear to be at work. One planner stated that he did not believe the model information would directly influence decision making arguing that "it is always the political decision that prevails", but nevertheless values the information being available for people to debate and adds that the information could frame the discussion. A regional planner also forwarded the idea that decision makers will not necessarily use information to make rational decisions and concluded therefore that the model should be viewed primarily as a communication tool or a facilitating tool for the planner. A GVRD planner also noted that group dynamics is very important in undertaking a modeling exercise and the model should be about "people talking and thinking about [land use issues] together". He pointed out that in the past, regional modeling exercises were undertaken internally within the organisation and were very much a product of the GVRD Policy and Planning Department whereas the success of the LRSP (the regional plan) lies with the member municipalities and the Province becoming participants of the process and taking ownership of the Plan. Thus although some planners expressed doubt as to whether or not information will necessarily influence decision-making, they agreed that the model could enhance communication by bringing the interests together to debate land use issues, to learn about each others organizations and to form ongoing working relationships. In this environment, they hope, shared learning is possible and a level of trust will result. 4.2.3 The Process of developing a land use model for metropolitan Vancouver The third area of exploration with the interview participants was the process of developing a land use model for metropolitan Vancouver. Overall, the interviews revealed that the process by which a region wide model would be developed was seen as very important. The participants agreed that without the participation of the land use stakeholders in some capacity, the process of developing the model would be flawed, 49 and the validity of the information produced by the model would be continually challenged. Although no participant put forward a clear process, their ideas can be summarized into a number of important elements. The process would need to identify the organisations that have an interest or stake in land development. Representatives of each • stakeholder organization would then need to be invited to form a group to drive the development of the model. One of the first tasks of the group would be to develop an understanding of the roles and needs of each so that clear objectives for the modeling project can be put forward. The building of the model requires reaching agreement on base data, input assumptions, and a data verification process that would test the model for reasonable results. Other issues that would need to be addressed in the process are who will manage and maintain the model, how will the model and data outputs be accessed and how will the modeling project be funded. 4.2.3.7 Identifying Stakeholders Overall, the municipalities were considered to be the prime public sector stakeholders. Municipalities continue to have almost complete power over local land development, subdivision of land and the density of land development. Their participation was therefore considered very important. Speaking more generally, one senior planner with the GVRD expressed fear that "[The model] will die if multiple interests are not involved". She stressed that bringing participants to the table and keeping them there will be a challenge and there must be a perceived benefit from the point of view of the municipalities before they will likely opt to participate. Further to this she suggested the need for a rallying point (for example, a discussion of a specific policy) to bring organizations to the table. Most however, while stressing the need for a municipal emphasis, favoured encouraging the participation of other public sector planning interests that influence the way in which land is developed in the region. "As many people as possible" should be involved said one planner. Another municipal planner went even further and suggested that private sector organizations should be involved so that their needs are addressed. The use of 50 the model by the private sector, he suggested, could also provide a much needed revenue stream for the project. The argument was also put forward that things that are "free" are not valued as much as things that people have provided input and invested time in. Without participation by the potential users of the model, this planner warned, this project could become an academic exercise in which considerable work is achieved but is not adopted into practice because its value is not recognized or its validity is questioned. One planner observed that people who work in isolation without a wider sphere of input tend "to fall in love with their own work". Without the work baring scrutiny to an outside audience, she argued, there might be errors in assumptions or needs are not properly addressed. Creating ownership of the model is important in that a working group of interests guide its development and influence the development of objectives. Thus, in the context of multiple land use stakeholders in the region, the success of the model will depend upon securing the interest and participation of these interests. Their participation will produce the buy-in into the project required to ensure the results are usable and considered valid by all. 4.2.3.2 Forming groups to identify needs and build the model Everyone interviewed recognized the land use interests, with particular emphasis on the municipalities need to be involved in the development of the model. Forming a group of stakeholder representatives is then needed to drive the development of the model. One planner suggested a two-tiered system that would consist of a working group and a steering or advisory committee. The smaller working group would assemble individuals with technical expertise to provide meaningful input at the detailed functional level of the model and the larger advisory committee with representative(s) from each stakeholder group would study the broader issues of model development. The forming of these groups, he said, would have to be a balancing act between inclusion and too large a group to the point that they are ineffective. He suggested that invitations to participate on the core modeling team should be extended to as many people as possible but should emphasize technical expertise, whereas the steering committee should secure 51 representation from all stakeholders in the region. He also stressed the need for political support from the outset recommending model demonstrations to political organizations such as City Councils, Regional District Boards, and municipal associations (i.e. UBCM), to be undertaken periodically to spark interest and engender a high level support for the model development. This system calls for top down support but with an inclusive process to ensure stakeholders' interests and needs are being addressed. A regional planner pointed to the already established Technical Advisory Committee (TAC), a sub-committee of the GVRD Board, as a good opportunity to link the modeling project to the technical community. In addition, TAC reports to the GVRD Board on a regular basis thus acting as a good forum for exposing elected officials in the region to the issues associated with the project. Other comments supported the idea of a flexible and voluntary modeling team. One planner pointed out that the role of each organization in the process will vary since each will have a different level of expertise and knowledge about urban modeling. Municipalities that have GIS systems and expertise would likely be quite active. Due to this difference in expertise, he asserted, a strict equal input approach would inevitably bog down the project in futile processes. He argued strongly, however, that the model building should be an interactive process in which the municipalities provide the base information and agree on the standardization of that information. Those interviewed also asserted that the process of model development should include identifying the planning needs of the participants and defining objectives for the project. One planner warned that excessive attention to the development of a model could impair the project team's ability to achieve objectives along the way. This in turn, he argued, not only impacts the schedule and the budget but the credibility of the project as a whole. In order to avoid this problem, the advisory groups must have a clear understanding of the planning needs of all participants and be able to agree on and prioritize needs and communicate them to the model builders. They must also be able to define the specific outputs and projected timelines for delivery. He recommended that project be run like "any business with a good strong customer focus, understanding the needs and requirements of the customer and making sure they are satisfied." 52 A second planner also recognized that the ideal of an equal input approach is not possible. Although he believed all municipalities need to be involved in order for the model to be useful, he stated that many municipalities do not have the resources to be actually inputting and refining the data. Their role, he said, would therefore be more likely to review the assumptions, inputs and outputs. Others suggested that the municipalities should input the data as they have direct access to the approvals that are given for development and future development. Municipal inputting would ensure the accuracy of the data. One planner stated that the regional plan should be a "living document" and be in a constant state of revision to respond better to changing conditions. In recent years, information management programs that track land rezonings and permit applications have become available in standardized software packages that can then be customized to the specifics of an organisation. Most of these programs are linked to a spatial reference and mapping function which may provide opportunities to link land use information to a region-wide land use model. The issue of standardization was raised a number of times. They pointed to a need for consistency in the way information is gathered and represented so that, as one planner explained, "municipalities are comparing apples with apples and oranges with oranges". Agreement on base data will need to include the issue of standardization so all the interests are working with common information. 4.2.3.3 Managing the model The issue of the most appropriate organization to manage the model was also explored with stakeholders. Although everyone agreed that a single organization would need to take overall responsibility for the model, opinions differed as to which organization or type of organization would best assume this role. One regional planner pointed out that the main issue is that of trust: the organization must be one that will not usurp the process and is perceived by the stakeholders to be trustworthy and objective. 53 Over half immediately identified the GVRD5 as the most appropriate managing organization. One regional representative stated that the GVRD had proven experience managing large-scale regional models through the development of its regional transportation model. He points out that the GVRD maintains the transportation model that is used by the member municipalities and some private sector firms. Multiple interests, he adds, also work together in the calibration of the model and the municipalities have the ability to run the model in house with their own sets of assumptions. Another representative from a regional land use interest believed the GVRD would be the organization best suited to managing the modeling project since they have already established relationships with many of the public sector planning organizations in the region. The GVRD also would have expertise and resources to standardize the varying datasets for inputting. Two municipal representatives also felt that the GVRD should control the process but with the condition that all input from the different interests are given full consideration. They stressed that the GVRD must manage the model without an agenda to ensure the credibility of the model. The consultation with the municipalities, they asserted, must look seriously at numbers resulting from municipally generated growth scenarios that may not necessarily support regional plan objectives. One planner feared that there may be some resistance within the GVRD to the model if the growth numbers produced challenge previous target capacities for some municipalities. He argued therefore that there needs to be open discussion as to how to proceed should the numbers produced by the municipality were not consistent with those produced by the GVRD. A number of representatives from municipalities said that it was commonly perceived within municipal organizations that the target numbers for growth used by the GVRD in the regional plan were not negotiated but imposed. Fostering an environment of trust through open and honest discussion and full information sharing would be necessary for the GVRD to successfully manage the model. Two planners representing regional service providers suggested that a non-government and non-political body would be more appropriate than the GVRD in order to bring greater impartiality to the modeling process. One planner warned that a GVRD-managed model would hinder the level of trust necessary for the project to be 5 The GVRD was identified without specifically identifying a Department of the GVRD 54 f successful. The model could be perceived as a tool to control the municipalities, he argued, and in which the growth numbers produced through the modeling exercise are predetermined. They suggested that a university may be better suited to this role and point to a university's resources, research base, and their relative neutrality. 4.2.4 Other Issues 4.2.4.1 Accessing the information How the model would be accessed was viewed as an important issue by the representatives although their views differed on the degree of information that should be accessible and to whom. All believed that access to the model would secure the trust necessary to ensure the model's legitimacy. A number of representatives suggested that municipalities could access the model via the Internet to submit revisions and to perform in-house land use analyses. Easy access to the model, he argued would help create ownership of the model, maintain the relevancy of the base data, and promote and secure ongoing participation in the modeling project. Two representatives suggested that data should be made available to other interests outside public sector planning organizations to further legitimize the model. As discussed in an earlier section, reservations were expressed for providing full access to the model. Full access to the model would give users access to forecasts at the block level, a level at which he does not have confidence the numbers would pass the "test of reasonableness". He recognized, however, that suppressing the data complicates the principle of accessibility and could potentially impair the level of trust. 4.2.4.2 Project funding The majority of participants pointed out that the development and maintenance of the model with the desired level of detail and complexity could be expensive. The development stage would require the conversion of data into a common standard, the digitization of municipal information where digital maps are not available, and a data review and verification process. Maintaining and running the model would require updating information and executing the work program as decided by the modeling 55 advisory group, including the development of scenarios, changing assumptions, and allocating forecasts. Securing adequate funding on an ongoing basis will thus be essential for the project. A number of ideas were suggested on how the model could produce a revenue stream. One municipal planner suggested a system in which the contributors to the model would have free and open access, while non-contributors who benefit from the model would pay a user fee. As an example, he outlined a possible scenario where the contributors were the municipalities providing and inputting the data, the region managing the data, and the utilities providing verification data, and the customers were private commercial ventures. Selling data to private firms could provide a valuable funding stream, but he recognized that the model must therefore be responsive to the needs of potential customers. A second municipal planner also suggested a cost recovery system funded by fees for private users. He asserted that outside funding and interest would make the model much more resistant to budgetary cuts. He also emphasized that the model has to be "good enough" to attractive private interests. A regional planner suggested a slightly different approach in which the organizations that do not provide funding for the modeling project up front would be subject to a user fee. Organizations should be encouraged, she asserted, to dedicate money up front as this will more likely secure their involvement in the model development process. She also added that the model has to be good enough and its value recognized in order for other organizations, including those outside public sector planning, to consider paying for it. In summary, the interviews with representatives of public sector land use and infrastructure planning organizations in the Greater Vancouver region revealed overall support for the development of a land use simulation model. The methodology of using GIS and small geographic areas in the structure of the model were thought to be important, bringing greater flexibility and clarity to the modeling process. Depending on their organizational perspective, they pointed to a range of planning issues that could be explored at varying levels of analysis. Municipal planners tended to look at how the model could be used to study forecasted growth in population and employment in selected neigbourhood planning areas and to explore the impact of alternative land use 56 scenarios on urban development patterns within their municipality. Infrastructure and service provider planners were quick to point out the potential benefit of the model in providing a greater level of certainty as to future land development patterns, especially over the longer term. Regional planners gravitated towards the ability to monitor adopted plans and to explore different policy scenarios that could help communicate the impact of municipal policy on the achievement of regional objectives. In addition, the interviews revealed that the process of the model's development was considered to be equally important as the structure and function of the model. They described an open and participatory modeling process in an impartial forum where shared understanding can be developed. This dialog must explore the interests, roles and needs of the various land use interests, must seek agreement on the inputs, outputs and assumptions of the model, and must allow for the results to be tested, challenged, and reviewed. They also warned against potential pitfalls in developing a large-scale urban land use model. The interviews raised many important considerations for the development of an urban land use simulation model. In the chapter to follow, these ideas are weaved with the research presented in the previous chapter to recommend principles for a land use modeling project in Greater Vancouver. 57 5.0 Recommended Principles for the Development of a Land Use Simulation Model for Greater Vancouver This research reveals that there are many important considerations in the development of a land use simulation model for Greater Vancouver. Recent literature indicates that the increasing availability and use of GIS in planning practice has helped overcome many of the traditional barriers to large-scale urban modeling and can address the need for more open and participatory planning processes. The two case studies illustrated the successful use of GIS in modeling large urban regions and the importance of planning processes in model development and review. In addition, the interviews with representatives of public sector land use and infrastructure planning organizations in Greater Vancouver revealed many important considerations for the development and use of a region-wide land use simulation model. All research indicates that a regional land use model that takes advantage of GIS technology and whose development affords an open and inclusive process, can serve as a tool to increase understanding amongst stakeholders, enhance their communication and help coordinate their planning efforts. In the context of consensus-based regional planning in Greater Vancouver, a tool that can draw stakeholders into debate to address common concerns and reach agreement on common goals represents a good opportunity to overcome some of the challenges of effective regional land use and infrastructure planning. From this body of research, a number of guiding principles for the development of a GIS-based land use model for Greater Vancouver can be drawn. These guiding principles are meant to define the key elements that a successful modeling project for Greater Vancouver should either contain or address, and which relate both to the structure of the model and the process of its development. The principles are listed below, followed by a more detailed explanation on each. The chapter to follow offers concluding remarks on which principles are considered essential for a successful modeling project. 58 (8) Should be flexible and adaptable (9) Should be 'understandable' (10) Should be methodologically sound in that the results should be the logical extension of the inputs (11) Should consider the entire functional region (12) Should have an open and participatory process for development (13) Should have clear objectives for the project (14) Planners should be involved in the modeling process (1) Should be flexible and adaptable This principle asserts that the structure of a land use simulation model should have the flexibility to allow the model to address a number of planning problems at different levels of analysis and should have the adaptability to be improved, altered and changed over time. Planning is an ongoing process, and thus a planning tool must be flexible to adapt to changing problems, shifts in values, and increasing expertise. These goals can be achieved to a large extent through the use of small geographic areas and GIS technology in the model. The use of small geographic areas in land use modeling ensures a high degree of flexibility, allowing aggregations to any user-defined area. The SANDAG modeling project in particular emphasized the use of small units of geography to both increase spatial accuracy and to allow for aggregations to different geographic areas of analysis. SANDAG recognized the importance of flexibility in the model structure to meet the planning analysis and information needs of its 18 member organizations. The interviews with public sector land use and infrastructure planning organizations in Greater Vancouver revealed strong support for the use of small geographic units. Region-wide data is only currently provided by the GVRD at the traffic zone level, a unit of geography that is not adequately discrete to allow for small area analysis, particularly at the 59 municipal and neighbourhood planning level. The ability to conduct planning analysis at the neighbourhood level was raised as a potentially powerful tool for studying the cumulative effects of targeted local policies on regional land use. One interviewed regional planner pointed to the inherent margin of error in each individual small area, or 'gridcell', and raised concern as to the release of information at the gridcell level as opposed to in an aggregated format. He argued the data should pass a "test of reasonableness" before being more broadly released. Since the data is not intended to be accurate at the gridcell level and since limiting access to gridcell level data would hamper flexibility in the use of the data, one possible solution is to establish minimum aggregations at which "reasonableness" is consistent. This could be in the form of a disclaimer for those wishing to perform analysis combined with suggested minimum aggregation levels at which the modeling group believes the outputs are reasonable. The challenge would then be to find agreement on the definition of "reasonableness". In any case, the results of a model are only as accurate as the data inputs and as a general rule one shouldn't attempt to perform analysis below the accuracy of the data. The use of GIS technology has also increased the flexibility and adaptability of land use modeling. GIS allows datasets to be changed and improved upon over time. The 'layering' capability of GIS means that new assumptions and inputs can easily be brought into consideration without changing the overall structure of the model. In addition, the improving sophistication of these systems and their increasing presence in planning organizations means that much information exists to draw on for extended analysis. (2) Should be 'understandable' The model should be relatively transparent so that potential users can understand the assumptions as well as the relationship between the inputs and outputs. One of the major criticisms of past large-scale urban models was that they were too often 'black boxes' to anyone who was not directly involved with the model's development. Users outside of the model building team were not able to understand the 60 relationship between the data inputs and the information outputs, raising questions as to the reliability of the model and the credibility of its results. If models are to be used effectively as planning tools, they must be able to embrace a larger scope of participants whose expertise and knowledge of modeling may be significantly less than the creators of the model. Research indicates that the use of GIS positively impacts modeling as it brings greater transparency to the structure and process of modeling. GIS provides a basis for discussion of the information that most people can understand. This transparency allows a deeper level of participation as well as a broadening one to embrace a larger variety of participants in the modeling process. Considering these findings, many of the past 'sins' of modeling have been diminished or eliminated by the introduction of GIS to modeling processes. In the SANDAG experience, the use of GIS overlays in the structure of the model was found to increase the understanding of the model by the users, primarily local staff. This understanding was reinforced by the review process in which the datasets were reviewed and corrected in detail, familiarizing staff with the data in their planning organization and how it is applied to the regional land use planning model. In addition to using GIS in their system, the designers of the CUFM model consciously designed their model around a series of changeable decision rules, rather than mathematical algorithms, with a view to making the model accessible to more people. The interviews revealed that stakeholders believed GIS would make land use modeling more understandable with its graphical representation of spatial relationships and with its ability to help planners and others 'think spatially'. They also believed that less assumptions are concealed in a land use model that uses GIS since those assumptions can be changed or altered in an understandable way by adding or removing data layers. (3) Should be methodologically sound in that the results should be the logical extension of the inputs 61 The model should have sound theoretical foundations in that the results should accurately describe the logical extension of inputs at a reasonably small level of geography and based on an explicit set of assumptions and rules. i The designers of the CUFM model specifically rejected using the established zonal system of geography since they considered the zones too large in area to provide a clear understanding of the spatial processes governing urban growth in the region. Since their purpose in developing the model was to help planners and decision makers examine the impact on land development resulting from alternative growth scenarios, the input data needed to be detailed enough to realistically simulate the impact of a wide range of real policy proposals. The designers were also faced with the difficult task of simulating the regional impacts of policies undertaken simultaneously by various land use interests. The model's simulation of spill-over growth into unincorporated areas when adjacent communities have reached capacity is a tribute to the modeling team's commitment to an accurate and theoretically sound simulation. The SANDAG model takes a similar approach in that the designers realized that smaller geographic units than the existing zones of analysis were needed to attain a level of analysis at which the information would be useful. The model was developed to produce periodic forecasts used in land use, infrastructure and transportation planning, but is also used as an analytical tool to evaluate the implications of alternative policy scenarios. To further ensure accuracy, the data inputs and outputs are extensively discussed and reviewed with the local authorities to establish that the results are reasonable and believable. The interviews with land use and infrastructure planning organizations also confirmed the need for methodological credibility and accuracy in a land use model developed for Greater Vancouver. The representatives pointed to a number of planning information needs suggesting that the model must be able to perform multiple functions. Primarily they expressed a desire that the model increase understanding of the process of land use change in Greater Vancouver by exposing relationships between land use policy decisions and the location, pattern and density of development. This increased understanding would give planners greater certainty of the location and timing of growth for infrastructure investment, allow plans and objectives to be monitored to ascertain whether targets are being met, and to explore alternative futures through the simulation 62 of land development patterns resulting from hypothetical policy scenarios. Clearly, the model must have sound theoretical foundations to deliver on these needs. The use of GIS to handle the complexity of spatial information will be key to the model's ability to simulate specific locations for land development. The use of small geographic areas derived from a 100m by 100m grid approximates the block-level land use activities. The interviews revealed a desire to understand how a policy applied in one municipality, or even in a neighbourhood, could potentially impact the region considering the application of that policy across all neighbourhoods. A number of the stakeholders felt that this type of concrete analysis would give greater meaning to the concepts of 'livability' and 'protection of the Green Zone' as outlined in the adopted regional plan (LRSP). (4) Should consider the entire functional region The model should consider all the activities and functional relationships in the region. The more complete the large-scale context, the more accurate the small area projections and analysis will be. In the case of SANDAG, the model was housed within the regional planning agency that has 18 cities and county unincorporated areas covering the entire San Diego metropolitan area. In addition, their Board of Directors had secured liaison members from state and federal departments whose activities had an impact on land development in the area. The builders of the CUFM model also recognized the need for a comprehensive approach and considered an area covering 14 counties, which was reportedly the first time these counties had been grouped into a defined regional entity. This broad grouping was considered essential for developing an effective urban simulation tool for policy exploration and scenario testing. Although only two of the representatives of public sector land use and infrastructure planning organizations in the interviews addressed directly the value of a model that covers an entire urban region, most expressed a desire to have as many organizations participating in the process as possible, including municipalities, regional bodies, service providers, and private interests as Well. In addition, many of the participants 63 acknowledged the impacts the policy decisions of one organization can have on another and on the region as a whole and wanted the ability to explore those interrelationships within the land use simulation system. (5) Should have an open and participatory process for development The process of model development should be open and seek the input and active participation of the multiple land use and infrastructure planning organizations in the Greater Vancouver region. Large-scale urban models of the past were developed by a handful of people behind closed doors, leaving potential users of the model results with little understanding of the structure of the model, the assumptions, inputs and outputs. In such an environment, there is little or no buy-in into the model and any argument put forward supported by analysis using the model is challenged by anyone not in agreement. The view of planning in the 1990s has emphasized open and participatory planning processes, where stakeholders can communicate and debate information and ideas to increase understanding and coordinate to achieve common goals. Planning is not, then, a product of a single organization but an ongoing process of collective design. By embracing this approach, a land use modeling project can provide the forum for face-to-face interaction and debate between the key stakeholders. The interaction of these organizations will ensure a wider acceptance of the model results as well as building social, intellectual and political capital for ongoing projects. SANDAG's land use model produces forecasts for population, housing, and employment, and are used by local, state and federal agencies to plan for services and infrastructure throughout the region. The forecasts are also used to explore alternative policy scenarios for the development of the region. The model's success in influencing decision-making and coordinating the region's planning efforts lies in its commitment to the development process in which stakeholders of all levels are included from beginning to end, sharing information, debating the data inputs and outputs and reviewing the forecast results. The SANDAG case clearly illustrates that stakeholder involvement in the process to ensure understanding and acceptance of the results is equal in importance to the results being accurate and reasonable. 64 The CUFM case study demonstrated that although the modeling team achieved their objectives of building an accurate and policy relevant spatial simulation model and that was understandable to more people than those that developed it, the model was unable to secure buy-in from public sector planning organizations. The project did not implement an open process where the stakeholders could participate in the modeling process. The lack of communication and coordination between the CUFM team and the land use stakeholders in the San Francisco Bay Area impaired the ability of the land use modeling exercise to influence decision making in any significant way. The public sector land use and infrastructure planning organizations recognized that the Vancouver region has multiple land use interests, both public and private, whose policies impact the way land is developed both inside and outside each jurisdiction. They believed therefore that the process of developing the model was very important for the ultimate success of the project. Most favoured the participation of representatives from all levels of government, with an emphasis on municipal participation, to build support for the project and to produce the necessary buy-in. They argued that participants would need to identify needs and objectives, and review and debate model base data, input assumptions and outputs. Most also acknowledged that participants in the process would contribute in different capacities, some offering technical expertise, while others would likely play an advisory role only. The contribution would depend on the organization's resources and level of expertise in modeling and modeling processes. Regardless of the level of participation, stakeholders emphasized the need for an open and transparent process with full information sharing. Although a few participants disagreed, most identified the GVRD as the most appropriate organization to manage and maintain the model. Clearly, the most important issue on all sides is that of trust, all desiring an organization that can provide an impartial forum and that will not usurp the process to forward its own agenda. (6) Should have clear objectives for the project A modeling project should establish clear objectives to maintain the credibility and the momentum of the project. One of the perils of large-scale and data-intensive modeling projects is that they can become derailed, both financially and politically, by a failure to 65 deliver results. This common failure in large-scale modeling projects has led to recommendations that planners start with a particular problem to guide the model building process. This problem should be short term and easily obtainable. Meeting an objective, however small, will keep the entire project moving forward and will lend credibility to the project. The need for addressing specific problems or issues was also recognized by a number of the interview participants, who believed that a rallying point could help bring people to the table and build credibility. They recognized from experience that model building teams can sometimes become lost in their work which may impair their ability to achieve objectives along the way, thus impacting the schedule, budget, and credibility of the project as a whole. To this end, many participants suggested that a steering committee or advisory group should be formed. This group would clearly identify the planning needs of the land use stakeholders and then agree on and set out clear objectives for the project. These objectives must be prioritized and effectively communicated to the model builders, complete with details of the deliverables and anticipated timelines for completion. (7) Planners should be involved in the modeling process The process of model development should draw on the skills of planners to inform and guide the process. The conventional view of the planner as the purveyor of unbiased professional advice and analysis for decision-making is no longer a widely held belief. Planners lend their technical expertise and knowledge of urban development to inform decision-making, but they also form and guide planning processes, acting for example, as advocates, facilitators and conflict mediators. Planners can enhance the process by ensuring that a multiplicity of viewpoints can be heard and by structuring an impartial forum where information is shared and where issues and ideas are debated openly. The combination of planners' expertise, technical and process-oriented, gives structure to the process and lends credibility to the project. The task of building and maintaining the confidence of the stakeholders will be essential in fostering an environment where shared learning and understanding is possible. 66 Information technology is increasingly evident in the planning profession. The increasing availability of GIS in planning departments and the greater expertise in using such systems to undertake a wide range of geographic and spatial analysis place planners in a good position to educate and communicate with the public and other interests on pressing planning issues. In order to do so, planners must be able to interpret land use modeling systems to enhance their own understanding of the processes of land use change and to be able to communicate that knowledge to others. In contrast to the CUFM project, the SANDAG modeling project was led by a team of practicing planners who implemented a clear process where the local land use jurisdictions could provide their input and review the results. The SANDAG team benefited from established good relationships with local jurisdictions and solidified those ties through full information sharing and a consensus-building approach to the modeling process. 67 6.0 Conclusion: The seven guiding principles presented in the previous chapter address the major elements to be considered in a land use modeling project undertaken in Greater Vancouver. Although all the principles presented are recommended, some hold greater importance for the ultimate success of a modeling project in Greater Vancouver. To this end, three concluding observations are offered in the paragraphs to follow. First, in view of the research presented in this paper, the most important principle is the open and participatory process for development. The SANDAG experience reinforces the idea that the process of model development is equally important as the information the model provides. This assertion is demonstrated by the fact that both the CUFM and SANDAG land use simulation models are understandable and demonstrated that they are methodologically sound by accurately simulating land use change in these large urban regions. Where they differ is in their process of development. The open and participatory process adopted by SANDAG clearly demonstrates the success of this approach in creating shared learning, acceptance of the model results, and a forum in which stakeholders can explore and debate ideas and alternative futures. In addition, the interviews with planning organization representatives in the region revealed that there is an awareness of the importance of this issue. The participants stressed the need for an open process guided by an organization that they could trust to provide an impartial forum for discussion and debate. i Second, the principles that the land use simulation model should be understandable and should be methodologically sound are necessary for the model to have credibility and for the information to influence decision-making. Without these basic principles addressed, the model would be in danger as being perceived as a black box and its legitimacy constantly challenged. The third remark is that the remaining recommended principles demand attention but are less essential than the three identified above. They should be used to steer the project and ensure the appropriate questions are being asked but their limited application due to prevailing conditions should not serve to discourage the development of a land use simulation model for Greater Vancouver. The benefits of dialog and 68 debate on the model and the modeling process will build the social, intellectual, and political capital to achieve success when the conditions become more favourable. i 69 Bibliography Baldassare, Mark, and Joshua Hassol. 1996. Possible Planning Roles for Regional Government. Journal of the American Planning Association. 62: 17-29. Batty, Michael. 1994. A Chronicle of Scientific Planning: The Anglo-American Modeling Experience. Journal of the American Planning Association. 60: 7-16. Beauregard, Robert. 1995. Quoted in Plan Canada. Vol.38, No.6. Blueprint for the San Diego Region: A Progress Report on the Regional Growth Management Strategy. 1995. San Diego: San Diego Association of Governments. Budic, Zorica. 1994. Effectiveness of Geographic Information Systems in Local Planning. Journal of the American Planning Association. 60: 244-262. Campbell, Heather. 1996. A Social Interactionist Perspective on Computer Implementation. Journal of the American Planning Association. 62: 99-107. Chim, Jim, Greater Vancouver Regional District Spatial Forecasting Requirements. July 1996. Esnard, Ann-Margaret and Bruce MacDougall. 1997. Common Ground for Integrating Planning Theory and GIS Topics. Journal of Planning Education and Research. 17:55-62. GISSAM: Geographical Information System Spatial Activity Model. 1997. Burnaby, BC: Greater Vancouver Regional District. Innes, Judith. 1991. Implementing State Growth Management in the US: Strategies for Coordination. Institute of Urban and Regional Development. University of California at Berkeley. July 1991. Innes, Judith and David Simpson. 1992. Implementing Geographic Information Systems for Planning: Lessons from the History of Technological Innovation. Institute of Urban and Regional Development. University of California at Berkeley. October 1992. Innes, Judith and John Landis. 1993. Issues in Growth Control Management. Institute of Urban and Regional Development. University of California at Berkeley. January 1993. 70 Innes, Judith. 1995. Planning Theory's Emerging Paradigm: Communicative Action and Interactive Practice. Journal of Planning Theory's Emerging Paradigm: Communicative Action and Interactive Practice, p. 183-190. Innes, Judith. 1996. Planning Through Consensus Building: A New View of the Comprehensive Planning Ideal. Journal of the American Planning Association. 62: 460-472. Innes, Judith. 1995. Quoted in Plan Canada. Vol.38, No.6. Isserman, Andrew. 1994. Projection, Forecast, and Plan: On the Future of Population Forecasting. Journal of the American Planning Association. 60: 208-21. Klosterman, Richard, E. 1994. Large-Scale Urban Models: Retrospect and Prospect. Journal of the American Planning Association. 60: 3-6. Klosterman, Richard E. 1997. Planning Support Systems: A New Perspective on Computer-Aided Planning. Journal of Planning Education and Research. 17:45-54. Landis, John. 1992. BASS II: A New Generation of Metropolitan Simulation Models. Institute of Urban and Regional Development. University of California at Berkeley. April 1992. Landis, John. 1993. CUF Model Simulation Results: Alternative Futures for the Greater Bay Area Region. Institute of Urban and Regional Development. University of California at Berkeley. April 1993. Landis, John. 1994. Future Tense. Planning. 60,2: 22-5. Lee, Douglass. 1994. Retrospective on Large-Scale Urban Models. Journal of the American Planning Association. 60: 35-40. Livable Region Strategic Plan. 1996. Burnaby, BC: Greater Vancouver Regional District. Margerum, Richard. 1999. Getting Past Yes: From Capital Creation to Action. Journal of the American Planning Association. 65: 81-192. Sancton, Andrew. 1995. Governing Canada's City-Regions: Adapting Form to Function. Montreal, PQ: The Institute for Research on Public Policy (IRPP). SANDAG Home Page. August 1996. 71 Sawicki, David and Craig William. 1996. The Democratization of Data: Bridging the Gap for Community Groups. Journal of the American Planning Association. 62: 512-523. Series 8 Regional Growth Forecast - Subregional Allocation, Vol.1 - OverView. 1994. San Diego, California: San Diego Association of Governments. Swanson, David and Jeff Tayman. 1996. Between a Rock and a Hard Place. Population Research and Policy Review. 14,2: 233-249. Tayman, Jeff. 1996. The Accuracy of Small-Area Population Forecasts Based On A Spatial Interaction Land-Use Modeling System. Journal of the American Planning Association. 62: 85-98. Wegener, Michael. 1995. Operational Urban Models: State of the Art. Journal of the American Planning Association. 60: 17-30. 72 Appendix A: INTERVIEW QUESTION OUTLINE A GIS-BASED LAND USE MODELING SYSTEM FOR METROPOLITAN VANCOUVER MASTER'S THESIS RESEARCH by: Michelle Armstrong School of Community & Regional Planning University of British Columbia Question outline (1) Should a small-area GIS land use modeling system that covers the entire metropolitan Vancouver region and accommodates different users be developed? Why or why not? (2) What should such a model do? (3) What do you consider to be the most important result of the model? Why? (4) How should a model that covers multiple jurisdictions and to be used by multiple interests be developed? (5) Who should be involved in developing the model? (6) Who should control the model? (Where is home? Who is host?) (7) To what extent should regional planning policy be an assumption of the model? (8) How should the model and data output be accessed? (9) Are there any other issues you feel are important in developing a modeling system to be used by multiple interests for land use planning? 73 APPENDIX B: BACKGROUND INFORMATION FOR GREATER VANCOUVER PROTOTYPE LAND USE MODEL Documentation of Phase One in the Development of a Prototype Geographical Information System Spatial Land-use Modeling Process for the Metropolitan Vancouver Region Understanding how much and where future growth in population and employment can be accommodated in a large urban area is essential for effective long-range planning at the local, regional and inter-regional level. Large-scale urban models have been developed in most major urban centres in North America to evaluate and project the relationship between growth and changes in land use, transportation and infrastructure requirements. Historically, these models have been relatively successful in providing information about how much growth will occur but often fell short in accurately describing where forecasted growth will be accommodated. The traditional barriers to spatial modeling, however, are being eliminated with the increasing use and availability of geographic information systems (GIS). A joint project undertaken by the Greater Vancouver Sewage and Drainage District (GVSDD), the GVRD Strategic Planning Department (SPD) and The Urban Futures Institute (TUFT) is developing a spatial land use modeling process that employs die spatial analysis and display capabilities of GIS to produce small area projections of dwelling units, population and employment. At its current stage of development, the modeling process utilizes control total projections for the Greater Vancouver Regional District produced by SPD and distributes these totals to small spatial units (100 metre by 100 metre polygons) within the Regional District The projected values for these polygons can then be aggregated to larger spatial units such as the GVSDD's catchment areas. The allocation of regional control totals to polygons is based on the relative characteristics of the polygons. Areas that are not eligible for future development (for natural reasons, such as steep slopes or water, or for policy reasons, such as parks and green zones) are first excluded from the projection. The remaining polygons are then ranked in terms of attractiveness for development. At the current stage of development, the ranking is based on a weighting of relative accessibility to employment, availability of sewers, and population density. Land use planning is currendy reflected in the option to use the 5 year target figures from the SPD Livable Regional Strategic Plan 0L-RSP) as maximums for households, population and employment Such an approach to modeling is referred to as a "functional" allocation, where allocation of projected variables is based on functional relationships between attributes of polygons: as many attributes that could be used (e.g., slope, soil conditions, distance, etc.) are spatial, GIS are particularly suited to this approach. This approach is in contrast to a "trend" allocation, where future allocation of variables is based on historical shares and/or trends. Trend approaches have predominated in the past, prior to the widespread availability of GIS. In the trend approach, the spatial units used are determined by the data sets used to identify and calibrate trends: census tracts are most commonly the spatial units used. Where fixed spatial units and trend to target values for projected future values are acceptable, the trend approach offers significant advantages in terms of modest data requirements, computational ease, and hardware requirements, as GIS is not required. However, if non-trend analysis, relationship modeling, and ability to create a range of aggregations of spatial units is required, then a GIS approach becomes essential. A model using fine spatial units that would allow forecasted information to be aggregated to any spatial area would make more sense. The advances in the computational power of the desktop computer and the spatial analysis capabilities of GIS software means that a new generation of urban models can manage multiple data files attached to literally hundreds of thousands fine spatial units. Such a powerful model can realistically simulate the impact of different policies on land development in metropolitan Vancouver. The long term impact of developing this type of model could be great, delivering better information to decision makers. 7 4 

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.831.1-0089384/manifest

Comment

Related Items