Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
TCRP H-17 Final Report 1.0 INTRODUCTION I.! Project Objectives This is the final project report for TCRP Project H-12. Integrated Urban Models for Simulation of Transit and Land-Use Policies. The stated objectives of this project are to: · Define more clearly the impact of transit access on land-use, · Evaluate the access component of existing land-use models; and . Augment ongoing work by the Travel Modle! Improvement Program (TMIP), and others to incorporate transit access into land-use models. The ultimate objective of this project is to define a framework for enhancing integrated transportation - land-use models. The particular focus is on the representation within these models Of land-use - transit interactions, so that these models cart provide useful. credible and timely inputs into the transportation pawning process. This framework should define long-term research and development needs. It also should identify actions which improve the applications of existing models In the short-term, and which provide the "bndge" to move forward bow practice and research in the medium-term. Accompanying this report are Guidelines for Implementation and Use of integrated models. The Guidelines describe how Amos, State DOTs arid other planning agencies can act today, in order to initiate or build upon existing integrated transportation - Arouse models. I.2 Definitions Several terms are used frequently in this report, working definitions of which are provided below. Given the urban focus of this project. it should be noted that these definitions reflect only urban transportation and land-use planning. In most cases, a broader context is evident, but is beyond the Interest of this project. Models are defined as mathematical algorithms which simulate human activities, such as travel behavior. Once they are calibrated against existing (known) conditions, We models car be used to forecast these activities. For present purposes, however, it is appropriate also to include theory and data, in addition to algorithms, in the definition of a model. This is because the successful development and application of any mode} involves a complex and balanced interaction among all three of these building blocks. The incorporation of "theory" provides a solid
~ ~ - foundation for the model. "Algorithms" provide the means of translating theories into an operational reality. "Data" represent the inputs and the outputs, without which the aigori~ms neither function nor provide mearungful results. Travel demand forecasting models me used to predict travel on a transportation network. Travel demand -- i.e., traveler behavior -- is a Unction of both human activities (generally represented in the models by land-use; see below) and network characteristics (including the services provided and their costs; where costs generally are expressed in teens of time and/or money expended by the user and/or the transporter. alla possibly intangibles such as comfort, etc.). Therefore, travel forecasts will vary according to changes in human activities or network ~ · . characteristics. 3. 4. Land-use models generally are mathematical tools that forecast demographic and economic measures of land-based activities. These measures can be described as meaIiingfill attributes of the population (e.g., age-group cohorts, income, jobs, etc. but also can be expressed in terms of development (e.g., residential dwelling units, commercial floor space, etc.~. They take into account trends and policies regarding demographic and economic development, as well as We supply of developable land. The shape of the transportation network, and its possible influence on the location of human activities are taken into account, but usually in a generalized manner. Market forces (the timing, magnitude and location of development) also are considered but again usually on a broad, often qualitative scale. Among other uses, the results serve as input to a travel demand forecasting model. However, generally there is little or no feedback from travel demand forecasting models into land-use models. Integrated fransporfation - land-use models simulate the interaction of land-use and transportation. They are ~nter~ded to provide a feedback mechanism between land- use models and travel demand forecasting models. For example, a new or improved transportation link can influence the timing, type and extent of development in a particular area, which in turn can influence the need for further network improvements. Similarly, the evolution of land-based human activities influences the demand for improved transportation services, which in turn further influences development. Therefore, integrated models take lar~-use models a step further, by also simulating the location of human activities as a function of transportation accessibility. Similarly, integrated models do not simply provide lar~-use input to travel demarld forecasting models. Some integrated models focus upon residential . . .. .. .. .. employment allocation ~l.e., the movement of people). Recent models provide a more generalized treatment of an area's economic and demographic activities (i.e., the movement of people and goods). - 2
TCRP H-12 Final Report The term "integrated transportation - land-use models" is very cumbersome. Furler, as argued in subsequent sections of this report, understanding the "urban system" involves understanding land-use processes, Jocahon choice processes (within a given land-use) of households and films, activity/~ave] choices, and the fundamental demographic and economic forces driving these various decisions. Given this, it is argued that integrate urban models is both a more general and succinct term, arid it is this term which is generally used throughout this report. Transit accessibility describes how well a given point in an urban area is connected by transit to other points Within the area. "Accessibility" can be defined in many ways: . . the absolute existence of a physical means of traveling between two points, where no other link exists (in this case, by transit), the relative level of service (speed, cost, comfort, etc.) of a D~tiCUi~ transit .. ~ ~ . . .. .. . . . . . . ~ link where alternate transit links exist and which may offer different levels of service; and the relative level of service of a particular transit link, compared with alternate modes (e.g., auto) which provide the same link. I.3 Why Integrated Urban Models? Before proceeding ~ ~ detailed discussion of integrate urban models, the question surely needs to be addressed of why do we need integrated models at all, especially given their historical absence from urban transportation planning in most cities? Traditionally, most urban areas have used some form of four-staae travel dem=d modeling system which simply takes population and employment forecasts generated "elsewhere" as exogenous, fixed inputs. The "land-use forecasting" problem is treated as a separate activity, often undertaken through some combination of trend extrapolation. professional judgement, scenario hypothesizing, and/or wishful thinking concerning implementation of official pearls (where such documents exist). This "disconnect" between transportation arid land-use analysis and modeling is deeply rooted in the profession resulting from marry different contributing factors. These include the following. Many metropolitan areas have organizational structures which treat laud-use planning and control as a separate function from transportation planning and control, resulting in different people in different divisions (if not entirely different departments or
TCRP [1-12 Final Report organizations) dewing with these two activities Land-use planning and transportation planning also usually split along professional lines. Larld-use planning is usually the dome of urban planners, geographers, etc., while transportation planning and modeling are generally the domain of engineers, economists, etc. Political obstacles generally exist. Land-use is often viewed as something which can't (and/or shouldn't) be interfered with. The assumption often seems to be that land-use v~ll evolve however it will evolve, arid it is our job as transportation planners to respond to land-use trends with transportation system changes as best we can. Within this worId-view, there usually seems to be little "return on investment" with respect to integrated land-use - transportation models. 4. The hypothesis seems to exist among some transportation analysts and planners that a significant transportation - land-use interaction does not, In fact, exist. Very loosely, the argument is that given the near-ubiquity of the road systems in most Norm American cities, that the opportunity to influence land-use decisions with transportation investment is minimal at best. A corollary to this is that firms and households are extremely "foot-Ioose" and base their locational decisions on a wide variety of factors, with transportation accessibility being a minor factor (if a factor at all) In these decisions. Under this hypothesis, the need for integrated models is, it is argued, is highly questionable. Over and above all of the organizational professional and ideological issues raised above, the fact is that building credible, operational, integrated models of transportation arid land-use is a very hard problem. The starboard (and very sensible) approach to tackling a difficult problem, especially in the first instance, is to decompose it into smaller, more manageable sub-problems. It is therefore, very understandable that researchers have tended to work on bits and pieces of the problem: Improvements to work trip mode choice models; empirical studies of micro urban design factors on travel demand; residential location choice models; etc. The result is a huge literature (empirical, theoretical and even applied) on a Aide range of subjects pertaining to the integrated modeling problem, but a relative lack of effort in attempting to "put the pieces together" into integrated. comprehensive models. ~, . ~ The general lack of connection between land-use and transportation modeling, at least in North America, cart also be traced to a general sense of disappointment and frustration with early attempts in the ~ 960s and early ~ 970s to build such models, as typified by the oft-quoted "Requiem for Large-Scale Models" Wee, ~ 9733. This, combined with a general loss of interest in Tong-range planning in the 1970s blue to low urban growth in this time period and a focus on short
TCRP H-12 Final Report "transportation system management" strategies) led to a general decline throughout the ~ 970s and much ofthe 1980s In interest tn. and development of, long-range forecasting models in general and land-use models In particular. The result is that relatively little progress was made dunng this time period in the United States web respect to integrated modeling. due to the lack of investment in R&D In this area At the same time, much of the progress which was made during this era occurred "off-shore" (South Amenca' Europe, etc.) where (at least marginally) higher attention to and investment in Arouse models occurred. Over the last several years, interest In integrated transportation - land-use modeling has picked up in the United States. as indicated by: 1. inclusion in IMIP as "Track E", which deals with land-use modeling (discussed filer in Chapter 21; the TMIP Land Use Modeling Conference (Dallas, February, 1995) which was Intended to stimulate debate and research activity in this area tS trunk, et al., ~ 995], several reports/studies (including this one) dealing with the state of the art in this area (e.g., Cambridge Systematics tI992], Southworth ~1995], URS Consultants tI996], Zebras. et al. [1995~; and, most important of all. 4. an increasing number of MPOs are In venous stages of implementing/us~ng land-use models of one form or another (PortIar~d, Houston, Boston, Baltimore. etch. In addition, it is important to note that our technical capabilities with respect to: · computer hardware (both processing speed arid data storage capabilities), . computer software (Geographical Information Systems (GIS), relational database management systems (RDMS), statistical analysis packages, object-oriented programming capabilities, etch; · databases (GIS-based, census, special-purpose surveys, etch; arid . modeling methods (random utility-based models activity-based models, advanced econometric methods for mode} estimation, etc.) have improved by many orders of magnitude since land-use models were first attempted in the ~ 960s. These pioneering efforts of Me ~ 960s were exactly that, and were doomed to "failure" by the lack of data, computing power and modeling techniques available at the time. We are now, how-ever, at a point where we have (or wail verily shortly have) the technical capabilities to undertake large-scare
TCRP H-12 Final Report urban modeling with reasonable expectations concerning both the efficiency and the effectiveness of the exercise [MiTIer and Salvini, ~ 9984. Why is this increased interest in integrated Arouse - transportation modeling occuring? As is discussed farther in Chapter 2. the process-oriented answer is that current Federal policies such as TEA-21 and CAAA marinate it (as did ISTEA), and that MPOs which do not consider larld-use options in combination with the transportation alternatives under analysis do so at their peril. Far more substantive however. is the growing recognition that the land-use - transportation interaction does exist, and it must be understood, modeled and accour~ted for In planning analysis. Urban areas have physical definition only through the transportation system: space has little meaning otherwise. We perceive the interconnections between points (activities) In space through the medium of the transportation system. Build the transportation system differently and people will use it differently, and Hey will organize themselves over space differently. Build Me city differently, and the transportation "needs" will be different. Any apparent "lack of relationship" simply reflects one (current) land-use - transportation combination in which near-ubiquitous automobile-based mobility has "loosened the bonds". It does not mean that other options and outcomes are not possible, nor does it mean that we do not need to explore and analyze how the overall urban system (transportation and land-use) wall evolve over time if we are to understand how best to invest in our transportation system. The achieving of a better understanding bared, eventually, representation) of the land-use transportation interaction is essential to the urban transportation policy debate in the United States (and, for that matter, elsewhere), whether we are talking about roads, transit' or non-motorized modes of travel. One of the most important example is the very basic question of whether constructing a new urban expressway has a net beneficial environmental impact (due to congestion relief and associated increases In average travel speeds) or a net negative impact (due to increased induced sprawl of land- use and travel patterns Increased auto dependency, etc.~. A recent "blue ribbons panel in the United States came to no definitive conclusion on this issue tCSTHCIAQEC, 199~], while a similar study in He United Kingdom clearly endorsed the negative impact case tSACTRA, ~ 9944. The fact that significant scope appears to exist for considered professional opinion to differ so dramatically, and the fact that so much of the debate over such issues is based on subjective arguments (often with strong ideological overtones) points clearly to the failure of currently available analysis methods to provide more definitive Insights into such problems. Transit planning is particularly strongly linked to land-use questions for at least two reasons. First, conventional transit is ordy viable within certain types of land-use/urban forms. In particular, certain minimum levels of trip-end densities are required before fixed-route transit services can be cost-effectively operated tPusharev and Pupae, ~ 977, ~ 9803. Second, transit is open viewed as part
TCRP H-12 Final Report of tithe solution' with respect to urban sustainability. yet this is a surely hopeless proposition without a direct tie (behaviorally and with respect to policy) between transit and land-use. If improved analytical capabilities are to be developed to address such fundamental questions Men hey must clearly be much more wholistic in nature, in that they must address the entire urban system consisting of l~d-use, activities, ~arlsportation arid environmental impacts in an integrated, comprehensive fashion Weiner, ~ 99~. In particular, they must be able to: explore the field range calf system responses (short and long run) to a given policy; ~ ---rig ~ ~~~ or- - x~ -Cal- ~ address a "~11 suite's of policies (transportation, land-use, etc.~; and ~- inco~porate the demographic/socio-economic inputs needed by the disaggregate behavioral models which are generally required to properly address the behaviorsipolicies of interest within urban systems. In addition to their role as forecastinaJpolicy analysis tools integrated transportation - land- use models also have the potential of providing an experimental platform/laboratory for exploring transportation - land-use interactions so that we can better understand the dynamics/relationships involved and hence better craft our policies. At least one reason why the empirical literature discussed In Chapter 3 is so mixed arid why we don't know more than we do about this interaction is because we never have enough empirical data to analyze the problem in sufficiently deep ~rays. In particular, we never observe the "without" case in any particular city (e.g., what would have happened in Toronto if the Yonge Subway had not been built?), or other "alternative historiesifiltures'' (e.g., what would have happened if tight development controls had been implemented in city X twenty, years ago, combined with a pro-active transit investment programed. A credible urban simulation model (if one can be constructed) would be of great benefit In exploring such questions, and, thereby learning more about how our cities work and about the likely efficacy · · . Of various po acres. This argument appears to have a strong circular aspect: if we had good models. we could learn more about our urban processes arid thereby improve our models; which begs the question of how to construct the rood model in the first place. The response to this is that the process is not circular, but rather spiral, In which initial models are hypothesized. Implemented and tested. Based · · ~.. ~ ~ , . on our experience with these initial models we learn "what works and what doesn't". leading to improved model sub-components as the outcome of additional research and development (R&D) efforts. Gradually, on mode] system improves Trough iterative hypothesis formulation and testing (i.e., the scientific method), and gradually the mode] becomes more useful/reliable in terms of providing new insights into urban processes [Miller and Salvini, 1998]. This "model as hypothesis" and "computer as laboratory" approach to the problem has not been an overly common (explicit) paradigm within the transportation modeling cornrnunity, wherein models are often more simply - 7 -
TCRP H- 12 Final Renort viewed as static instruments within which current "best practice/Icnowledge" are embedded. It is, however, essential to the evolutionary improvement of our understanding and our models over time. Finally, as discussed further ~ Chapters 4 and 5, "land-use" models are much more than that. In particular, Hey also ideally are demographic simulators, generating not only the population, but the attributes of the population required for behavioral location choice and travel demand modeling. GouTias and Kitamura tI992], among others argue eloquently that we must have this demographic simulation capability to support the effechve implementation of We new generation of activity-based models which are generally recognized as being required to address the full range of transportation policies and issues facing twenty-first century cities. Certainly, a strong case can also be made that the relatively limited impact which even disaggregate logit mode choice models have had on travel demand modeling practice can be traced in large measure to the difficulty of projecting the population socio-econom~c attributes required for these models to operate most effectively Emitter 1996]. I.4 Objectives For Integrated Urban Models Given the foregoing discussion of the need for integrated urban models, one can identify a number of objectives which these models should strive to achieve. In particular, integrated urban models should be: . . Theoretically sound, in particular, they should be based upon the determinants of the "transportation - land-use" connection. Among the key factors to be addressed are the traveler's determinants of mode choice for a particular trip and the developer's determinants of the use of a particular piece of land. Result-driven, but respectful of due process and other practicalities (such as the input data that are, or are likely to be, available). Responsive to the issues faced currently by MPOs transit operators and other urban transportation planners -- notably the current fiscal environment, Federal legislation (TEA-2 l, CAAA, etc.), the dynamics of local zoning regulations and authorities, the public's desire for accountability in public spending, emerging travel patterns (which may or may not support transit) and the private sector's growing role in supplying public services. Of particular interest is the growing transfer of responsibility to individuals ~ d, increasingly, their employers -- to resolve their own karlsportation needs with minimal public involvement. Exceptions, of course, exist. A particularly noteworthy one is perhaps Birch, et al. [ 1974~. - 8
TCRP H-12 Final Report Cognizant of Me regional, state. national and global demographic and economic inter-relationships that determine the dynamics of urban form and development. Practical to operate, win meariingfi~l outputs and a traceable, defensible process. · Sufficiently flexible to accom~nodate We differing scales and magnitudes of different . . - cities ant regions. Presentable in an understandable way to decision-makers and the public. I.5 Report Organization Chapter 2 discusses Me planning context for integrated urban modeling. Chapter 3 summarizes our empincal understanding of land-use - transportation lands in particular, transits interactions. Together, these two chapters provide Me basis for defining integrated urban modeling needs and desired capabilities. - Chapter 4 then defines art ttidealt' integrated urban modeling system which' if developed' should be capable of incorporating our best understanding of transportation - land-use interactions and of addressing our plaruiing analysis arid forecasting needs. The conceptualization of this "ideal" model is view-cd as a critical step in this project, since it both provides a benchmark against which current models can be compared and a target towards which model research and development efforts can be directed. Chapter 5 presents the detailed review of six currently operational models, where these models have been carefi~, chosen to be representative of the current state of practice. Emphasis ithin the review is placed on the strengths and weaknesses of the overall state of practice, both respect to short-term implementation and as a basis for evolution into the "ideal" model. Chapter 6 then details a multi-year research and development (R&D) program designed to improve significantly our operational integrated modeling capabilities. Throughout this report special consideration is given to the role which public transit plays, both within the integrated urban models under discussion and, more important, within the transportation - lar~d-use interaction. This focus is most explicit in the context-setting Chapters 2 arid 3. In the subsequent discussions of the integrated models (both actual and desired) in Chanters , · . · ~ . · . · . . .. ~-o transit is clscussecl where it is appropriate to Go SO, OUt. overall, a more general, "multimodal', point of view tends to be adopted. This does not reflect any reduction in our concern for the role of transit with these models, but rather is consistent with a fiJndamental tenet of this report that a unholistic arid comprehensive approach to modeling urban systems must be adopted. Thus, in order to understand arid mode] public ~ansit's role within the system properly, one must mode} the overall system (roads, land-use, etc.) "properly" as well. 9
ACRE H-12 Final Report - 10
