The UPLAN Urban Growth Model

Bob Johnston

rajohnston@ucdavis.edu

ANALYSIS OF THE IMPACTS OF URBAN GROWTH DERIVED FROM THE LAND USE COVERAGE: UPLAN, A GIS MODEL

Objective –

            We need an urban model that will project several land uses in size increments roughly matching the development parcel sizes. Residential densities must be represented, to get runoff and habitat impacts right. The model must be inexpensive, to cover bioregions and watersheds.

            1. UPLAN is a GIS-based model for testing urban growth scenarios. It is written in Avenue for running in ArcView3.1/Spatial Analyst on a personal computer. We intend it to be general enough to be run for any urban region or county in the U.S. The software will be posted for downloading on the ICE Web site. It runs in raster (grid) format at any cell size, down to what the data will bear.

            2. For general urban footprints, UPLAN is probably alright for any region. If you want to get realistic locations for several land uses and realistic travel behavior, larger regions with multiple growth centers should be modeled with spatial competition models, such as TRANUS, MEPLAN, METROSIM, AND URBANSIM, and then UPLAN.

            3. UPLAN is interactive, that is the user can change the population growth rate, or other basic assumptions, such as: employees per household, households per acre, and employees per acre. The user can also change the assumed proportions of land use types, such as high-density commercial v. low-density commercial, or high-, medium-, and low-density residential. Proportions of employment land use types can also be changed. All variables will have default values.

            4. UPLAN uses input data layers that are widely available in the U.S.: DEM slopes, STATSGO soils, surface water features, wetlands, floodplains, GAP vegetation, roads, freeway ramps, passenger rail stations, ports, airports, incorporated cities, spheres of influence (areas around cities to be serviced within 10-15 years), land ownership, existing urbanized lands for the base year, and general land use plans. We operate on 25m grid cells, so small urban infill sites and individual rural residential sites can be represented. 

            5. UPLAN uses transportation and utility services variables as attractiveness factors (freeway ramps, roads, rail stations, city and sphere of influence, port, airport), with slope as a discouragement factor. It then allocates land uses in order of bidding ability in the market (industrial, commercial high density, residential high density, commercial low density, residential low density, residential very low density).

            6. The user can set various environmental and social constraints to growth, such as steep slopes, areas with shallow groundwater, wetlands or surface water bodies with a buffer of any size, etc. One can also specify various levels of general land use plan compliance, ranging from none, to using only the industrial designations, to one-way zoning, to two-way zoning. Useful experiments include turning agricultural zoning on and off in a county. Another key experiment is to designate various habitat patch combinations as protected areas, to project the effects on urban growth.

            7. Policy tests that can be undertaken include: general plan changes, urban growth boundaries, habitat/open space preserves, riverway/floodplain protection, new freeways and roads, and new rail transit lines.

            8. We will post UPLAN on the ICE Web site, so it can be used remotely and map image files downloaded. We hope people all over the U.S. and world will try it out on our Sacramento datasets and then apply it to their regions. We will require that software improvements made likewise be posted on our Web site.

            9. A prototype project is the Sacramento region, where we will run scenarios for 2020 and 2040, in collaboration with various citizen and business groups in the region. We will also cooperate with SACOG, the MPO in the region, if they wish to use this model. Scenarios will include a Trend Scenario, two or more Sustainable Development scenarios, and others as desired. 

            10. The growth impact models work from the urban layer for the future year and the natural data layers. Models include: runoff volumes, surface water quality, potential for groundwater pollution, costs from flooding, costs from wildfires, and county fiscal effects.

            11. We hope to gain support to apply UPLAN to all the counties in California for 25-, 50-, and 100-year growth projections. This simple exercise will alert interest groups and officials to the consequences of continued low-density suburbanization. A key part of this project will be a habitat richness model, developed at ICE. This model HEPLAN (Habitat Evaluation Planner), will permit the evaluation of habitat quality for each growth scenario in each county. (could be SHEPLAN for State Habitat Evaluation Planner...).

ANALYSIS OF THE IMPACTS OF URBAN GROWTH DERIVED FROM THE ECONOMIC INTERACTIONS IN THE REGION: URBAN ECONOMIC MODELS

Objective –

            Many other impacts of an urban pattern derive from the travel and locational decisions of the households and firms. These spatial interactions can be simulated with urban models of the spatial competition type. These urban models can give measures of vehicle-miles-of travel by speed, which can be fed into the California mobile emissions models. We also get traveler surplus (net benefits) and locator surplus from these models, as well as travel costs, all by household income class and type of economic activity (for firms).

            1. In a related project, we have applied two true urban economic models, MEPLAN and TRANUS, to Sacramento region datasets. These statistical models represent supply and demand for travel and land use in 58 zones, linked by road and transit networks. We use UPLAN to disaggregate the land use projections from these models.

            2. We can also set environmental constraints in UPLAN and carry them "backward" to one of the urban models to limit acres available in the 58 zones in its input files. We then run the urban model and feed its acres of land consumed for each land use type for each zone to UPLAN and run it to get spatially detailed land use allocations. The constrained lands are then discretely held out from development in the GIS.

            3. This set of linked models is vastly superior to running UPLAN by itself, but is very expensive to implement. We could perform a similar exercise using the existing  urban models in the Bay Area and in San Diego Co., if those MPOs wished. SANDAG actually has a GIS capability similar to UPLAN, we believe.

            4. The urban economic models that we have applied in the Sacramento region give measures of housing costs, land costs, land rents, and travel costs, so one can project the effects of various habitat protection programs on housing and travel costs. We are trying to improve the models, to get a measure of consumer welfare for households and firms, specified as locator surplus.

            5. With this complex model set, we are attempting to evaluate habitat protection schemes and other policy scenarios, such as transit with compact growth, multiple employment centers with local transit, regionwide paratransit, reuse of the air bases, and maximum rural homesites regionwide. 

THE ROLE OF MODELS:

Models may be used for: 

1. Analysis of past and present spatial patterns of phenomena (with maps and descriptive statistics), 

2. Projection of the most likely future patterns of these conditions, and 

3. Prescription of desired future conditions and requisite policies (and testing of these policy sets).   

            Analysis and projection can help to identify common ground among user groups, as they come to understand and accept current problems and future likely problems. Regarding prescription, we believe our role to not be one of advocacy of particular viewpoints, but rather to be one of clarifying value choices, especially the values that are traded off when one selects any particular policy set. In order for models to help us to understand these tradeoffs, the models must be complex enough to represent a great variety of social, economic, and environmental phenomena. 

            In addition to being "truth-telling" devices, models also help us ask the right questions. In this heuristic role, models allow us to learn together about the urban region and to test prescriptive concepts. Models can greatly facilitate bargaining, by bringing all interest groups into the planning process and allowing the quick testing of the ideas of all participants. Models that run in a few minutes or less on PCs,  can be used in real-time analyses in meetings, creating a detailed real-time discourse.

            Models can be useful in all these roles, because they are systematic assemblages of our assumptions about how the world works. They provide a consistent framework for our discussions and analyses. Models do not provide answers, they just illustrate various points of view. Since they employ graphical outputs, models can greatly help get interest groups to meet together and bargain, because of the evocative methods of analysis and portrayal used.

            Models generally contain errors of at least 1% per year of projection, much of this caused by input error (aggregate population forecasts input to start the model). In long-range forecasting, we ignore the input error (we assume under- or over-projection of population just means we are projecting for a nearer or farther year), and concentrate on the differences across alternative futures (their rankings). Typical planning is for 20-year time horizons. We will perform 50-year scenarios, to project growth impacts more clearly.