The second generation of the California urban futures model. Part 1: Model logic and theory

In this paper we explore the theory and logic behind the development of the second generation of the California urban futures model, a site-specific urban growth and simulation model. The second-generation model remedies three of the major shortcomings of the first generation. It substitutes a statistical model of urban land-use change, calibrated against historical experience, for an uncalibrated 'developer-driven' model. It includes multiple urban land uses (for example, singlefamily residential, apartments, retail and office, and industrial) and allows them to bid against each other for preferred sites. It allows previously developed sites to be redeveloped into different uses. Finally, in addition to simulating the spatial impacts of regulatory policies, it can also simulate the effects of major infrastructure investments such as highways and transit lines. The CUF model: a work in progress The first generation of the California urban futures (CUF) model (Landis, 1994; 1995; Landis and Zhao, 1994) achieved a number of significant advances in the field of metropolitan growth modeling. It was the first operational urban growth model to be truly disaggregate, that is, it did not rely on zones. It was the first such model to explore the crucially important role of private land developers in determining the future form of metropolitan growth. And it was the first model to incorporate—indeed, depend on—GIS. These advances notwithstanding, the original C U F model also suffered from some significant shortcomings. First, it was limited to residential development: omitted from the model were methods for projecting and/or allocating future industrial, commercial, and public activities. A second limitation, which was a direct extension of the first, was that the model did not allow different activities to bid against each other for appropriate sites. Unless explicitly prohibited, the 'best' sites (that is, those sites which were the most profitable to develop) were always reserved for residential development. Third, and perhaps most critically, the rules used to allocate future development were never calibrated against historical experience. Instead, residential growth was allocated to sites based on the difference between observed housing prices and 'best-practice' housingdevelopment cost functions. Finally, real-estate prices, to the extent that they played a role in the C U F model, were entirely exogenous. New residential development that could not be allocated simply spilled over into other jurisdictions (subject to user-specified limits) rather than feeding back into the allocation process through higher prices. In the second generation of the C U F (CUF-2) model an attempt is made to remedy these shortcomings. Unlike the first generation model, this includes multiple land uses; allows different land uses to bid against each other for preferred sites; is calibrated against recent experience; and incorporates a 'pseudo-pricing dynamic' into the development spillover process. In this three-part series of papers we explore the development, calibration, and use of the CUF-2 model. In this paper (part 1) recent developments in the field of urban modeling are reviewed, the overall structure of the CUF-2 model is explained, and urban growth as a path-dependent process of discrete land-use changes is reconceptualized. 658 J Landis, M Zhang In part 2 the calibration of the land-use change submodel is explained. In part 3 the use of the CUF-2 model for simulating alternative regulatory and investment policies in the San Francisco Bay Area is demonstrated and we conclude with an assessment of the advances made and the remaining deficiencies of the model. Recent developments in urban activity modeling As two recent reviews by Wegener (1994; 1995) make clear, the state of the art of operational urban activity models is advancing at moderate pace. Two developments of recent years—one theoretical, the other data oriented—are combining to help metropolitan growth modelers produce ever more explicit models of urban activity patterns. The first of these is the incorporation of random-utility theory to predict generalized location and travel choices based on observed household, firm, or traveler behaviors. This ability to model metropolitan-scale activities as combinations of individual location and travel choices represents a significant step forward in the theoretical development of urban models. A second advance is based on the growing availability of longitudinal microscale data, and on the growing power of GIS-based software to analyze such data. Activity data tied to specific locations in space (either through Cartesian georeferencing or address matching) are now available in many locations. This makes it possible to represent the metropolitan area as a collection of individual sites or locations, rather than as zones. It also makes it possible to track and analyze actual patterns of metropolitan land use, and demographic or economic change, instead of relying on changes in zonal aggregates. Building on such newly available data sources, Batty (1992), Batty and Xie (1994), and, more recently, Clarke et al (1997) have begun to insert processes of spatial change (such as cellular growth and diffusion) into urban models. These advances notwithstanding, a number of significant challenges remain. The first is the absence of actual, that is, observed land or building prices. Site prices are typically determined endogenously in most urban growth models, either as travel-based shadow prices or indirectly through the process of market clearing. This limits the ability of most metropolitan activity models to simulate the effects of realistic landuse policies and regulations or of nontransportation infrastructure investments. Until urban activity models can be calibrated against the characteristics of observed land-use transactions, including both location and price, their usefulness for real-world applications will remain suspect. A second challenge concerns the modeling of economic activity, particularly jobs. Most existing metropolitan models adhere to some form of export-base framework. The total demand for regionally exported goods and services is determined exogenously as a model input. Once determined (and converted to jobs), regional economic activity is then disaggregated sectorally into primary and secondary industries, as well as disaggregated spatially by zone. The processes by which actual job changes really occur—through the growth or shrinkage of existing industries, through the birth of new industries or the decline of older ones—are ignored. Similarly ignored are the nontransportation cost reasons why industries locate where they do. A third challenge concerns how best to represent and integrate supply-side characteristics. Missing from most urban growth models are a wide variety of supply-side considerations, including the suitability of the built and natural landscape to support new or additional development; the capability of the landscape to support development at different intensities; and the characteristics and behaviors of builders and developers. Because of these limitations, none of the present urban models can be used reliably (1) Transportation costs typically enter as employee journey-to-work costs and, in a more limited number of models, as goods shipment costs. The second generation of the California urban futures model 659 for project-based environmental, social, or fiscal impact analysis, or for cumulative impact assessment. A related problem, identified twenty-five years ago by Lee (1973), is the insensitivity of most urban models to many types of regional and local development policies. Present urban models are useful for analyzing the travel, congestion, air quality, and perhaps regional economic implications of new highway and/or transit policies, but not for much else. Because they do not include site-level environmental or regulatory information, these metropolitan activity models are limited in their ability to analyze (or simulate) the sorts of zoning changes, environmental regulations, and particular infrastructure investments of most interest to local governments. The logic of the CUF-2 model The first generation of the C U F model was designed to fill some of these gaps. As noted by Wegener, it sacrificed comprehensiveness and theoretical elegance in favor of spatial detail and an ability to simulate local policy initiatives. The newest version of the model maintains the same policy focus and level of spatial detail as the original, while plugging some of its theoretical holes. Conceptually simpler than its predecessor, the CUF-2 model is also much more data intensive. Figure 1 (see over) shows its four main components. (a) Activity projection. This first component of the CUF-2 model consists of a series of econometric models used to project future population, households, and employment by jurisdiction at ten-year intervals. The equations used to project population and households are the same as those used in the first C U F model (Landis, 1994). Employment projection is new to the CUF-2 model. Using County Business Patterns data from 1981, 1989, and 1993, we generated three-digit employment estimates for every zip code. These estimates were then aggregated by city and sector. Separate econometric models were developed for each sector. The development of these estimates and the calibration of the different sector models is detailed in Landis et al (1998). (b) The spatial database. The second component of the CUF-2 model consists of a GISbased database of developable land units (DLUs). As in the prior version, DLUs are potentially developable or redevelopable sites. In the original C U F model, DLUs were generated as the spatial union of multiple GIS layers. The resulting DLU database consisted of thousands of unique polygons, each defined as different along some attributable dimension than its immediate neighbors. In this version, DLUs consist of one-hectare (100 m x 100 m) grid cells, which may or may not be uniquely different from adjacent grid cells. The shift from polygons to grid cells was made for a number of reasons. First, the analytic power of grid-based GIS procedures has advanced considerably since the first version of the C U F model. Second, data on existing and historical land uses are collected and tabulated at the hectare level by the Association of Bay Area Governments (ABAG). These data are essential to the development and calibration of the land-use change (LUC) submodel, described below. Finally, the hectare grid cell has about the right level of resolution for analyzing patterns of development and land-use change. But not, as we note above, the site-specific implications of such investments. (3) Thirteen sectors appropriate to the San Francisco Bay Area economy were identified: (1) agricultural services; (2) food manufacturing; (3) high-tech manufacturing; (4) other manufacturing; (5) tourism; (6) construction; (7) finance, insurance, or real estate; (8) professional services; (9) transportation and wholesaling; (10) public utilities; (11) retail trade; (12) consumer services; and (13) others not previously included. (4) This data set was not available in digital form prior to 1993. 660 J Landis, M Zhang (a) Activity projection. Project household and commercial and industrial job growth by city and county. (b) Spatial database. Assemble different data layers (for example, wetlands, farmland, slope, city boundaries with the hectare grid cell lattice) (c) Land-use change model. Estimate multinomial logit model of historical land-use change: P (hectare land-use change) = f (site and area characteristics) 0.85