USING GIS IN URBAN PLANNING ANALYSIS

Richard K. Brail, Lyna L. Wiggins

Abstract
Geographic Information Systems are reshaping many of the tasks and analyses conducted by urban and transportation planners. In this paper we explore the connections between GIS and urban and transportation analysis. Four areas of applications of GIS are discussed for urban planning and three areas of applications of GIS are discussed for transportation planning. Note is made of topics within these various areas that should be of particular concern to the research community. The paper concludes with a description of a specific application of GIS functions used in a planning analysis of the potential use of New Jerseys bus transportation in serving welfare clients access job sites, childcare facilities, and job training centers.

1. Urban Planning Analysis

Urban planning has a long history of using analytic tools in assessing the past, current status, and future of cities and regions. Planning also has a long tradition of adapting concepts and methods from other fields. Examples include the broad use of statistics, the extrapolation of concepts such as the gravity model, and more recently, the use of cellular automata concepts. Two classics in urban planning analysis, Urban Land Use Planning and Urban Planning Analysis: Methods and Models, defined the general parameters. Urban Land Use

Planning was first published in 1954 under the sole authorship of F. Stuart Chapin. It has been periodically revised since the first edition. The current edition (Kaiser, Godschalk and Chapin, 1995) outlines the land use planning process with a broad recognition of the role of information technology and GIS. In Urban Planning Analysis: Methods and Models by Krueckeberg and Silvers (1974) there was an integration of statistical, project evaluation and management techniques, and projection models. These two books reinforce the strong connection between planning and urban analysis. Geographic information systems have entered planning. GIS deals with spatial objects, their properties, and their relationships to each other. Planners analyze the past and the present, and then project to the future. GIS, on the other hand, is not in itself future-directed. However, it does a very good job of enabling us to store, process and visualize current and past information. Models are used to assist planners in looking to the future. They permit scenario building and provide an important aid to future-directed decision-making. Planning, however, is more than models and GIS. While databases and spatial information systems are important components of current planning activities, planners deal with constituencies, power relationships, and complex urban and regional problems. The planning process has been analyzed in a variety of fashions, and has been characterized as rational, adaptive, strategic and/or communicative, among others (Kaiser, Godschalk, and Chapin, 1995; Esnard and MacDougall, 1997, Innes, 1995). We may debate the degree of rationality in the planning process and the viability of the rational model. However, if planning is what planning does, then the already demonstrated use of GIS in a wide variety of contexts shows it to be an important component of planning research and practice. Insofar as GIS is, in its nature, a systematic and orderly set of concepts and methods, then it is part of a broader rational process. However, GIS also can contribute to the broader view of planning. GIS as an information organizing and manipulating tool can provide, if used wisely, analytic capacity and communicative activity to a wide audience. Klosterman (1997) has characterized the 1990s as the period of collective design in information technology where processes are designed to facilitate social interaction and discourse in the pursuit of collective goals. In this

context, Internet GIS and the integration of GIS and models are potentially important pieces of collective discussion and design. There are a broad number of success stories where GIS has supported and extended planning analysis and practice in meaningful ways. In this show and tell we will focus on land use and transportation planning.

2. Where GIS Assists Land Use and Transportation Planning

There are a broad number of areas of land use and transportation planning in which GIS has either reshaped the analysis or made a task easier to do. We will explore the following areas in outlining the connections between GIS and urban analysis: 1. Improving urban spatial planning 2. Improving transportation planning 3. Using GIS functions in transportation planning

2.1. Improving Urban Planning With the rapid diffusion of GIS throughout local governments in the United States, most urban planners now have access to the technology. GIS is affecting urban planning in four areas. The first area encompasses applications that expedite access to information used to perform operational tasks. Many of the uses of GIS for planners are operational in nature. For example, one of the most popular GIS applications in planning departments is automated notification. In this application, the GIS software helps expedite the process of creating a mailing list for notification of adjacent property owners (within a regulation-specified distance) about a specific rezoning or variance request. This application is a simple one from a GIS perspective, but can greatly improve the efficiency and accuracy in completing this

task for the planner. Other operational uses include using GIS to assist in answering questions from the public. For example, instead of talking to a planner in person or on the phone, a citizen may find out about the specific zoning for a property of interest, and the description of that zoning category, from an automated GIS application running on a countertop computer. In the future, this zoning information application will be available to citizens through interactive mapping on the Internet. In both of these cases, the planning department should be increasing its operational efficiency through its use of GIS. A second group of GIS applications now being used in urban planning practice are more descriptive in nature. These applications help planners describe and ask questions about the spatial patterns of specific land uses or events of interest. For example, planners are now using GIS to map the locations of building permits. This gives the planner a more complete picture of the spatial distribution of growth. Other applications in this group include mapping crime events, liquor store locations, building abandonment, code violations, brownfield sites, zoning variances, etc. Using these descriptive applications, planners increase their knowledge of specific spatial patterns in their urban area. Using this increased knowledge should give planners the base information to justify programmatic changes. For example, concentrations of code violations in specific neighborhoods may lead to new recommendations for targeted redevelopment efforts. When these simple descriptive GIS applications are expanded to include a temporal component, planners are then also able to describe and evaluate patterns of change in specific land uses and events. This capability provides a third group of GIS applications, those that assist us in monitoring changes in spatial patterns. Adding time to the descriptive analysis should improve the ability of planners to monitor and evaluate change. The temporal data structures in current GIS systems are often weak. Planners are (and should be) more concerned with monitoring changes over time than other professionals within municipal government. For example, the professionals in the property assessment office are focused on sending out the correct tax bills for the current, most up-to-date property values. They are not typically interested in the historical sequence of property value changes. Planners, particularly if their goal is to project into the future,

need to participate in decisions about how GIS data is structured and stored in their municipal GIS. We note here that temporal data structures in GIS are currently weak and should be the focus of increased research (UCGIS, 1997). Applications in the fourth group of GIS applications build more complexity into the use of GIS data by urban planners. This fourth group incorporates more complex spatial analyses that require the integration of multiple GIS layers. Examples of applications in this fourth group of applications include suitability analysis, as well as the use of multiple GIS data sets within operational urban models. These applications, as well as some of the applications described in the proceeding paragraphs, require the sharing of GIS data between departments within local government. For example, for planners to map code violations requires sharing data with the Building Code Enforcement Department. In these Enterprise models of GIS implementation in local government, the Planning Department may be the primary data provider for GIS layers such as zoning and land use, while it will be a data user of information from other departments. It is often noted that one of the strengths of GIS technology is its ability, through the use of the common denominator of space, to integrate information across multiple data providers and users. In an operational Enterprise GIS, planners have access to a wealth of timely and disaggregated spatial data. This quantity, quality, and detail of information was not easily available a decade ago. Besides the sharing of GIS data within local government departments, there is also an increasingly large inventory of GIS data available from regional, state, federal and commercial data sources. We note here that the integration of multiple sources of GIS data with varying scale and accuracy is also an active area of research in Geographic Information Science (UCGIS, 1997). This ease of accessibility to many sources of GIS data is a significant change in our information infrastructure and should be reflected in corresponding changes to our operational urban models. GIS has made model building easier and has changed the structure of certain urban spatial models. Systematic spatial information makes more data more available. On-line census data, the development of topologically based roadway networks, such as TIGER, and the adoption of

GIS in local government all contribute to providing the researcher with more and better information. Second, GIS permits the models to incorporate new variables in the estimation and prediction process and has enabled new conceptual frameworks to be explored. One widely known spatial model utilizing GIS extensively is the California Urban Futures Model (CUF), developed by John Landis and others at Berkeley. Landis introduced environmental factors derived from a GIS base into a simulation of urban growth. The first version, CUF-1, used a vector data structure. Through a series of overlays, developable land units in the region were identified. For each unit a propensity to be selected for development was statistically estimated. The overlays included environmental variables such as wetlands and slope, as well as transportation access, housing value and housing density (Landis, 1995). In the second version of the model (CUF-2), Landis moved to a grid (raster) data structure. The use of a grid structure strengthened the model by permitting analytic methods not readily available in a vector format. The model structure, based on a multinomial logit framework, estimates the probability of land use change of one hectare cell as a function of a set of variables including site characteristics, access, community characteristics, policies and relationships to neighboring sites (Landis and Zhang, 1998a; 1998b). Beyond Landis, other researchers have developed GIS-based models. One interesting example is the land development simulation using multicriteria evaluation and cellular automata (CA) concepts by Wu and Webster (1998). Similar to Landis, Wu and Webster use a raster approach, calculating for each cell the transitional probability of a change in a land use classification over time. Batty has been exploring the use of CA theory in simulating urban development (Batty, 1997). He and colleagues have extended CA modeling using a GIS-based framework and developed a very interesting approach simulating urban processes (Batty, Xie, and Sun, 1999). One issue that has arisen in connecting GIS and models is the exact role of each and their relationship to each other. Harris and Batty suggested that: where GIS is to used to support modeling, such systems should be mainly based on using their representational and graphic capabilities to store, derive and communicate data rather than

extend their usage to modeling (1993: 192). They go on to suggest that formal models should not contain GIS-like functions. Yet they also recognize the possibility that a new kind of model-friendly GIS may emerge. There are three basic approaches that encompass the options: (1) operate the GIS and models separately with the outputs of one fed to the other, (2) use GIS functions within models, or (3) use models within GIS (Karimi and Blais, 1997). The decision on how to integrate models and GIS is an important and complex one. For example, Wu and Webster (1998) use GIS functions and call the CA and multicriteria models, written in C, from within the GIS. As the authors indicated, certain display and data transfer functionality inherent in GIS was lost by running external models. The issue of model and GIS integration extends beyond the use of particular spatial functions to assist the model development process. Some urban spatial researchers are using a GIS platform as the centerpiece of the model system. The various model processes outside the GIS are called by the system when needed. Landis has also developed the California Urban and Biodiversity Analysis (CURBA) model (Landis, et al., 1998c). This is a simplified variant of CUFII and is run inside ArcView, a GIS developed by Environmental Systems Research Institute, Inc. Putman, the developer of DRAM/EMPAL model set, has developed METROPILUS, an integrated transportation and land use model running inside a GIS (Putman, 1998). He also has selected ArcView as the core, providing the user with a friendly visual environment within which to do model applications. ArcView itself has an internal object-oriented programming language, Avenue, as well as the capacity to call processes written in other languages such as Visual Basic and C. Since ArcView can read and manipulate GIS layers from a variety of sources, the urban analyst can quickly utilize a variety of data sets as an aid to the model-building effort. Recalling the earlier discussion about the relationship of models and GIS, we suggest that current examples point to a broader recognition of GIS as central to the model process. There are fewer new model efforts that ignore the spatial processing power of GIS. Also, there are substantial efforts to actually make a GIS package the core of the model system. This increasing use of GIS packages in model

development does present potential problems. Using ArcView from ESRI, for example, means that users must pay for licensing the program and are locked into a single vendor. In an ideal world we would have (and may get) an open GIS either in the public domain or available at a modest cost. Such a system would follow the path of Linux, an Unix-like operating system that is receiving wide distribution. Options exist to avoid using a GIS package at the center, such as utilizing libraries called by a programming language. These options generally mean significant programming overhead in developing original screen interfaces and working with different databases. At a broader level we also must be mindful of the critique of GIS by Harris and Batty (1993). Regardless of whether or not GIS is the centerpiece of the model design, or simply functions that are called when needed, the use of geographical information system concepts and tools is established. 2.2. Improving Transportation Planning In the 1950s and 1960s transportation planning methodology was concerned with highway networks. Through mathematical modeling it was possible to connect land use and transportation plans and to provide a framework for making long-term infrastructure investment decisions. The models could take in land use data, mathematically generate and distribute traffic, and allocate vehicles to roadway segments in the network. It was possible to estimate congestion levels as a basis to planning new and expanded roadways. In more recent years transportation planning has faced a host of new issues, including air quality, a broad commitment to equity issues and to multiple modes, and the rise of information technology. GIS has entered this changing environment. There are three areas in which GIS is affecting transportation planning. First, there is the development of improved databases. This improvement is occurring at two levels the definitional level and the system level. GIS, when implemented in multiple organizations and in different locations, creates a need for common data definitions. In the US the federal government has created a coordinating organi-

zation (Federal Geographic Data Committee) charged with developing common definitional frameworks. The Ground Transportation Committee of the Federal Geographic Data Committee is developing a Transportation Spatial Data Dictionary (Rabkin and Maccalous, 1997). Transportation features form one of the initial seven Framework layers that have been identified as central to the establishment of a National Spatial Data Infrastructure. The committee members working on standards for this important framework layer have noted some of the unique data structure requirements of transportation GIS layers. Notable is the need for adequate representation of linear referencing systems and the need for a datum that would establish a single location-control framework (Dueker and Butler, 1998; Fletcher et al, 1998). Among the research topics related to the creation of the transportation data infrastructure is the automatic extraction of roads from digital images (Shahin, 1997). At the system level in the U.S. we have developed a national topologically correct transportation network database, TIGER. Also, governments at all levels are putting together a range of transportation-focused databases including highways and transit lines, traffic statistics, accident records, and others. GIS is central to these efforts. One central clearinghouse is the Bureau of Transportation Statistics of the U.S. Department of Transportation. Second, there is the availability of real-time information about elements of the transportation system. In the United States the well-funded and highly visible national effort to apply information technology concepts and tools to transportation has been labeled Intelligent Transportation Systems (ITS). The purpose of ITS is to make the transportation system more efficient, more robust, and safer. In general, ITS is a broad name of a set of technologies and programs that interact with the vehicle, the roadway or guideway, or the operator. ITS is a systems of systems, including among others: the Advanced Traffic Management Systems (ATMS), the Advanced Traveler Information Systems (ATIS), Commercial Vehicle Operations (CVO), and Advanced Public Transportation Systems (APTS) (Johnson, 1997). The Advanced Traffic Management Systems (ATMS) includes projects designed to improve efficiency of highway systems, such as better traffic signalization using real-time traffic

data and driver notification of roadway conditions through variable message signs and radio. Advanced Public Transportation Systems (APTS) include Automatic Vehicle Location (AVL) systems, where transit vehicles are tracked on a real time basis, and electronic fare collection (U.S. Department of Transportation, 1996). There is widespread interest in ATIS. These traveler information systems, such as Travlink in Minneapolis, link AVL with a central dispatch and a traveler information system. Commuters are informed of transit conditions through kiosks, signs and videotext terminals in homes and businesses. Such information systems use the Internet as a distribution device to citizens. Pheng and Beimborn (1998) present a broad overview of some current applications of Internet GIS to ATIS. It is possible in Milwaukee using an Internet GIS for residents to explore accessibility of transit service. In Seattle real-time traffic volumes (with video) can be viewed on a map. The Commercial Vehicle Operations component of ITS focuses on making freight operations more efficient with the paperless vehicle with GPS electronic tags, onboard computers and automated fleet/freight administration (Lindley, 1997). The base to ITS is GIS. It is essential to have up-to-date maps that contain a street network with needed attributes capacity, speed, volumes, and signalization. Ideally the voluminous data generated from ITS can enhance the planning process. Real time temporal data on traffic flows and transit vehicle movements are now available, and GIS can provide the display tool. Network models permit rerouting of traffic in the event of accidents and delays. In these ways we can say that ITS uses GIS as a essential building block. In a longer range planning environment, five years or more, ITS is a source of information. There is voluminous data available from real-time traffic monitoring and electronic fare collection. Models can use these extensive datasets for a wide variety of purposes, including specification and calibration. Here models, and related GIS installations, will use ITS data.. Klosterman (1997: 49) has drawn distinctions between information (serving management), knowledge (serving executive decisionmaking), and intelligence (enabling collective design) in the evolution of information technology foci from the 1960s to the 1990s.

10

Following Klosterman a product of ITS research may enable collective design. The concept of intelligence as collective design is found in PLANiTS (Planning and Analysis Integration for Intelligent Transportation Systems), an ITS decision support system (Khattak, Kanafani and Cayford, 1996). The system is designed to support collective responses and analysis for transportation planning. Third, GIS has improved transportation planning analysis. These improvements have occurred at a variety of scales, for a range of purposes, and with many different tools. At the regional scale GIS can organize and manage demographic and transportation data. Given sufficient data at a detailed parcel or block level, a GIS can permit more detailed traffic impact analyses or better public transit planning. In general, the analysis of transportation systems and the corollary planning of capital and operational improvements have generated a broad body of research and a host of projection and impact models. The traditional four-step urban transportation planning process was defined in the 1950s. Trip generation equations converted land use and economic activity into trips emitted from traffic zones. Trip distribution models, often based on gravity models, allocated trips from zones to other zones. Modal split models proportioned the trips to different travel alternatives, and traffic assignment models put autos on to a computerized version of the roadway network. A number of computerized transportation planning model packages following these steps were built and widely used. The early networks used in these packages connected nodes (where streets crossed) with straight lines, or links. These networks abstracted from reality: there were no curved roads. On the other hand, GIS data structures permit realistic networks with curved roads. The data structures for GIS road networks and transportation models are different. While transportation networks are constructed with a link-node structure, GIS uses an arc-node topology. This fundamental difference has slowed the integration of GIS and transportation model networks. However, we can expect a full integration of GIS and transportation databases and models in the next few years. Also, there have significant developments on the GIS side in the creation of network functions, such as minimum path solutions. We have suggested that GIS will integrate with the traditional urban

11

transportation planning process. GIS functionality is also changing the questions we can answer in transportation planning. The simplest example of this is the combination of two basic GIS functions geocoding and buffers. We will demonstrate the power of these tools in the context of a broad social policy issue in the United States. 2.3. GIS Functions in Transportation Planning One of the current late 1990s social initiatives in the United States is the movement of welfare clients into work opportunities. Mandated in federal law, able-bodied welfare clients are to look for and hopefully find jobs. While this broad national effort is politically popular and has received wide support, it is still unclear how well the policy will work. One of the major issues in this welfare-to-work movement (W2W) is how to get clients to jobs. GIS can be a useful analytic tool for W2W. An example in New Jersey demonstrates the potential of GIS in assisting the W2W program (Brail, et al., 1997). The study examined the potential use of New Jerseys bus transportation network in serving welfare clients. The study used two standard GIS functions, address matching and buffers. Address matching was employed to assign two-dimensional spatial coordinates for each of four address data sets residences of potential participants, job sites, childcare facilities, and job training centers. Using GIS maps of bus routes, buffers were created and the number of individuals and potential destinations located within discrete distances from transit lines were calculated. The study found, not surprisingly, that there was a great overlap between welfare client location and bus lines. Fully 94 percent of clients lived within one-half mile of a bus line. In this example, transportation planners benefited directly from GIS functionality. This kind of buffer analysis can be replicated in relatively easy fashion. Of course there are some interesting problems with this simple analysis. Address databases, particularly for job locations, are of varying quality. Additionally, the bus transportation data used in this study was limited to bus routes, and did not include stop or frequency information. Finally, all data in this report reflect single temporal snapshots of the clients, firms and centers. The client list will

12

change, new firms will be added and subtracted, child and day care centers may open or close, and transit routes may be altered. This is a highly dynamic environment that will require continual monitoring. Also, this analysis does not mean that the W2W problem is anywhere near being solved. Necessary steps in the process involve getting detailed information on transit schedules, headways and stops, determining the role and availability of paratransit options, and building real-time scheduling systems based on GIS.

3. The Many Roles of GIS

We have seen a variety of ways in which GIS has been connected to urban planning analysis. First, GIS acts to organize spatial and attribute data and as an information system. Local governments can organize their data around a multi-jurisdictional and multi-purpose GIS. Second, GIS also operates as visualization and summation support for larger data warehousing enterprises found in the Integrated Intelligent Transportation System (IITS), as outlined by Scherer and Smith (1999). The challenge of gluing together disparate databases within a data warehouse has relegated GIS to a special role within a larger framework. Third, we can employ GIS as a stand-alone set of spatial storage and manipulation tools useful at a variety of scales and for a variety of purposes. We have demonstrated one such effort in the W2W study. Two classic GIS functions, address matching and buffers, were used together in an important task keying back to broad societal objectives. Fourth, GIS feeds model development in land use and transportation. We have seen how GIS functionality is embedded in CUF-1 and CUF-2 by Landis and in the CA model of Batty and colleagues. Finally, there are GIS-centric models at differing scales. The Spatial Analyst and Network Analyst extensions of ArcView contain sophisticated models that do many useful tasks. The extension of the DRAM/EMPAL model set by Putman runs inside ArcView, as does the CURBA model of Landis.

13

With all of these uses of GIS, is there a down side? We have moved beyond the question of whether or not GIS and model building should be kept separate. The simple answer to the question of how GIS and models relate is that it depends. If the researcher wants CA or environmental variables in the model, then GIS functions handling proximity or overlays are very useful. If the researcher is working at a household or land parcel scale, and can handle demographic, spatial and environmental variables where the GIS, at most, provides data, then there is much less use of this tool. Also, we cannot ignore the real need to focus on theoretical issues surrounding model development (Anas, Arnott and Small, 1998; Waddell,1998). Urban spatial theory will go multiple paths, and these paths may or may not include GIS functionality. The central concern is whether or not a fascination with GIS shapes research in ways that may be inappropriate. GIS, with its various functions and wonderful maps, should not be a distraction or a distortion. We will close with one example where using GIS functionality drives a model structure. Landis, in CUF-2, used a grid cell model at 100 meter by 100-meter resolution to capture land use change in northern California. By using grid cells he was able to utilize a rich set of explanatory variables. However, behavior is being imputed to cells, rather than households (Waddell, 1998). Also, there is the question of the size of the grid cell. Would the results vary markedly if a different size cell were modeled? Landis has developed a superb model structure, and made a decision on a research design that, as we all do, makes tradeoffs. We have shown that GIS is an established and important component in urban planning analysis. The rapid evolution of model functionality inside GIS and the rapid adoption of GIS in urban research both suggest that we will see increasing integration. We should be cautious, however, in overselling GIS, and leave its development and use to evolve as needed. We all need to recognize there is a large gap between seduction by an attractive tool and commitment to its larger use in appropriate situations.

14

References
Anas, A., Arnott, R. and Small, K.A. (1998), Urban Spatial Structure, Journal of Economic Literature, XXXVI: 1426-1464. Batty, M. (1997), Cellular Automata and Urban Form: A Primer, Journal of the American Planning Association, 63,2:266-274. Batty, M., Xie, Y. and Sun, Z. (1999), Modeling Urban Dynamics Through GIS-Based Cellular Automata. Draft. Brail, R.K., et al. (1997) Assessment of Public Transportation Opportunities for WorkFirst New Jersey Participants. Report prepared for the Office of Policy and Planning, New Jersey Department of Human Services. New Brunswick: Bloustein School of Planning and Public Policy, Rutgers University. Dueker, J. and Butler, A. (1998), GIS-T Enterprise Data Model with Suggested Implementation Choices, URISA Journal, vol. 10. Fletcher, D., Espinoza, J., Mackoy, R.D., Gordon, S., Spear, B. and Vonderohe, A. (1998), The Case for a Unified Linear Reference System, URISA Journal, vol. 10. Harris, B. and Batty, M. (1993), Locational Models, Geographical Information, and Planning Support Systems, Journal of Planning Education and Research, 12: 184-198. Innes, J.E. (1995), Planning Theorys Emerging Paradigm: Communicative Action and Interactive Practice, Journal of Planning Education and Research, 14,3: 183-189. Johnson, C.M. (1997), The National ITS Program: Where Weve Been and Where Were Going, Public Roads, 61,2: 6-14. Kaiser, E.J., Godschalk, D.R., and Chapin, F.S., Jr. (1995), Urban Land Use Planning, Fourth Edition. Urbana: University of Illinois Press. Karimi, H.A. and Blais, J.A.R. (1996), Current and Future Directions in GIS, Computers, Environment and Urban Systems, 20,2: 85-97. Klosterman, R.E. (1997), Planning Support Systems: A New Perspective on Computer-Aided Planning, Journal of Planning Education and Research, 17:45-54. Krueckeberg, D.A. and Silvers, A.L. (1974), Urban Planning

15

Analysis: Methods and Models. New York: John Wiley and Sons. Landis, J. (1995), Imagining Land Use Futures: Applying the California Urban Futures Model, Journal of the American Planning Association, 61,4: 438-457. Landis, J. and Zhang, M. (1998a), The Second Generation of the California Urban Futures Model. Part 1: Model Logic and Theory, Environment and Planning B: Planning and Design, 25: 657-666. Landis, J. and Zhang, M. (1998b), The Second Generation of the California Urban Futures Model. Part 2: Specification and Calibration Results of the Land-Use Change Sub-Model, Environment and Planning B: Planning and Design, 25: 795-824. Landis, J., et al. (1998c) Development and Pilot Application of the California Urban and Biodiversity Analysis (CURBA) Model. Draft Monograph 98-01. Berkeley: Institute of Urban and Regional Development, University of California. Lindley, J. (1997), Building Smart Infrastructure to Serve Travelers and Systems Managers, Public Roads, 61,2: 30-35. Peng, Z. and Beimborn, E.A. (1998), Internet GIS and Its Applications in Transportation, TR News, March-April. Putman, S.H. (1998), Urban Modeling System for Use in Transportation, Land Use, and Air Quality Analyses for Metropolitan Areas, presentation notes, November 20. Rabkin, M. and Maccalous, S. (1997), Transportation Spatial Data Dictionary, prepared for Federal Geographic Data Committee, Ground Transportation Subcommittee. Cambridge, Massachusetts: The Volpe Center, U.S. Department of Transportation. Scherer, W.T. and Smith, B.L. (1999). The Development of Integrated Intelligent Transportation Systems (IITS), presented at the annual meetings of the Transportation Research Board, Washington, D.C., January. Shahin, F.S. (1997). Automatic Extraction of Roads from Digital Images, GIS/LIS Conference Proceedings. University Consortium for Geographic Information Science (1997) Research Agenda, . U.S. Department of Transportation. (1996), Benefits Assessment of Advanced Public Transportation Systems (APTS), prepared for the Office of Mobility Innovation, Federal Transit Administration.

16

Cambridge, Massachusetts: The Volpe Center, U.S. Department of Transportation. Waddell, P. (1998), Simulating the Effects of Metropolitan Growth Management Strategies, presented at the annual meetings of the Association of Collegiate Schools of Planning, Pasadena, California, November. Wu, F. and Webster, C.J. (1998), Simulation of Land Development Through the Integration of Cellular Automata and Multicriteria Evaluation, Environment and Planning B: Planning and Design, 25,1: 103-126.

17

18