Modeling the Dynamics of Urban Development and the Effect of Public Policies

1
Modeling the Dynamics of Urban Development and
the Effect of Public Policies

• The Human Dimension of PRISM
• Marina Alberti
• Alan Borning
• Paul Waddell

2
Outline of Talk

• Scope of the Human Dimension
• The UrbanSim model system
• Land cover change model
• Current research agenda
• Objectives for this Year

3
The Scope of the Human Dimension

• Inputs
• Starting conditions inventories of land use,
  land cover, real estate, business locations, and
  household locations
• Macro-economic and demographic trends
• Local infrastructure investments and
  regulations/pricing (transportation, water,
  sewer)
• Land use policies (growth management,
  comprehensive land use plans, environmental
  regulations)
• Outcomes
• Spatial patterns of land use and land cover
  change
• Spatial patterns of real estate development and
  prices
• Spatial patterns of business and household
  location

4
UrbanSim Modeling Approach

• Model choices of agents
• Discrete choice models (multinomial logit)
• Microsimulate individual agents
• Dynamically simulate annual time steps
• Model market interactions
• Use very disaggregate spatial units (150 Meter
  grid cells)

5
A 150 Meter Grid Cell in the Queen Anne
Neighborhood
6
Classification of Development Types
7
Current UrbanSim Components
8
Factors Considered in Residential Location Model

• Household Characteristics
• Income
• Age
• Presence of children
• Number of workers
• Number of Vehicles
• Housing Characteristics
• Cost
• Quality
• Density
• Neighborhood Characteristics
• Neighborhood housing density
• Neighborhood commercial and industrial space
• Neighborhood retail employment
• Neighborhood land values
• Regional Accessibility to Employment by Transit
  and Auto for
• 0 car households
• 1 car households
• 2 car households

All independent variables are endogenous in the
model system
9
Factors Considered in Employment Location Model

• Employment Characteristics
• Industry Sector
• Nonresidential Space Characteristics
• Cost
• Type of Space
• Density
• Local Characteristics
• Land values
• Agglomeration Economies mix of jobs by sector
• Proximity to Freeways and Arterials
• Regional Accessibility to Population

All independent variables are endogenous in the
model system
10
Factors Considered in Real Estate Development
Model

• Site characteristics
• Existing development characteristics
• Land use plan
• Environmental constraints
• Urban design-scale
• Proximity to highway and arterials
• Proximity to existing development
• Neighborhood land use mix and property values
• Recent development in neighborhood
• Regional
• Access to population and employment
• Travel time to CBD, airport
• Vacancy rates

All independent variables are endogenous in the
model system
11
Factors Considered in Land Price Model

• Site characteristics
• Development type
• Land use plan
• Environmental constraints
• Regional accessibility
• Access to population and employment
• Urban design-scale
• Land use mix and density
• Proximity to highway and arterials
• Vacancy rates

All independent variables are endogenous in the
model system
12
Assessment of Current Status

• Operational urban simulation system
• Open Source software at www.urbansim.org
• Generic SQL Database for read/write
• Interoperates with GIS
• Version 2.0 now completed (complete
  re-engineering)
• Has been applied in Eugene-Springfield, Honolulu,
  Salt Lake City, Houston now starting
• Puget Sound application to be supported by Puget
  Sound Regional Council

13
(No Transcript)
14
Land Cover Change Model

• A new model component under development
• Predicts probability of 30 meter cell changing
  land cover classification during a single year
• Separate model specifications for differing
  conditions
• Cells affected by land use change in immediate
  area
• Urban-rural fringe areas not immediately affected
  by land use change event
• Urban (built up) areas not affected by land use
  change
• Rural (agricultural, forest) areas not affected
  by land use change

15
Land Cover Change
The probability of transition of a pixel of
initial land cover i at time t having the same
land cover class at time t1 (j0) or changing
to one of the other land cover classes (j 1J)
can be written as a multinomial logit
Where Pij is the probability of land cover
at a given grid cell at time t having the same
cover class at time t1 or changing to another
cover class. ?j is a vector of estimated logit
coefficients. J is the number of land cover
states

16
(No Transcript)
17
Independent Variables
Intensity of development event Devtype
transition 124 Specific attributes of site
Land cover Slope of cell Aspect
of cell Soil quality Parcel
ownership Parcel size Land value
Distance to critical areas Distance to
nearest road Distance to water
infrastructure Distance to critical areas
Distance to nearest road Distance to
CBD Distance to forest source area
Distance to nearest land cover transition
Distance to nearest development event
Restriction on minimum lot size
Continued.
18
Independent Variables
Spatial context of development Built-up
density Road density High erodible soils
of prime farmland Mean patch size of
land use/cover Contagion of land
use/cover Dominant land use/cover Transition to
land cover Residential units recently
added Commercial units recently added Road
capacity recently added Change in mean patch
size Position on the urban gradient Urban Growth
Boundary
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
Current Research Projects

• National Science Foundation, Urban Research
  Initiative, Reusable Modeling Components for
  Simulating Land Use, Transportation, and Land
  Cover, Marina Alberti, Alan Borning, Scott
  Rutherford, Paul Waddell, 439,357, 1999-2001.
• National Science Foundation, The Impact of Urban
  Patterns on Ecosystem Dynamics, Marina Alberti,
  Derek Booth, Kristina Hill, and John Marzluff,
  424,977, 1999-2003.

23
Current Research Projects

• National Science Foundation, Information
  Technology Research Initiative, Interaction and
  Participation in Integrated Land Use,
  Transportation, and Environmental Modeling, Alan
  Borning, Batya Friedman, Mark Gross, David
  Notkin, Zoran Popovic, and Paul Waddell,
  3,500,000, 2001-2006.
• National Science Foundation, Digital Government
  Program, Software Architectures for
  Microsimulation of Urban Development,
  Transportation, and Environmental Impact, Alan
  Borning, David Notkin, and Paul Waddell,
  600,000, 2001-2004.

24
Current Research Projects

• National Science Foundation, Biocomplexity
  Program, Modeling the Interactions between Real
  Estate Development, Land Cover Change, and Bird
  Diversity, Marina Alberti, Mark Handcock, John
  Marzluff, Paul Waddell, 1,128,818, 2001-2004.
• Federal Highway Administration, Case Study on
  the Application of UrbanSim to the Salt Lake City
  Region, Paul Waddell and Alan Borning, 150,000,
  2002-2003.
• Puget Sound Regional Council, Development of a
  Land Use Model, Paul Waddell, Alan Borning,
  Marina Alberti, 150,000, 2002-2003.

25
Proposed Research Objectives for 2002-3(funded
by sources other than PRISM)

• Development of the data and calibration of the
  existing UrbanSim model specification for the
  Central Puget Sound region (King, Kitsap, Pierce,
  and Snohomish counties) pending funding from
  PSRC.
• Development of land cover classification and
  accuracy assessment for Landsat images every two
  years from 1986 to 2001.
• Calibration and testing of the initial version of
  the land cover change model for at least King
  County pending funding from King County DNR.
• Work on an indicators and evaluation component
  for UrbanSim, to support a set of predefined
  indicators, and flexibility to allow users to
  modify and add indicators.

26
PRISM-Specific Objectives

• Developing close collaboration with the Crystal
  Team, to set up the protocols for coupling
  UrbanSim and Crystal, jointly defining the
  specifications for a water demand model, and
  implementing the model as a component within the
  UrbanSim architecture.
• Developing close collaboration with the DHSVM
  Team to set up the protocols for coupling
  UrbanSim and DHSVM through the land cover change
  model and jointly defining specifications to add
  the artificial drainage to the hydrological model
  in urbanizing landscapes.

27
PRISM-Specific Objectives

• Expanding the scope of the UrbanSim Indicators
  and Evaluation component to incorporate more
  environmental indicators, with input from other
  PRISM teams. Some or all of these additional
  indicators would come from other PRISM model
  outputs, so this would also require setting up
  protocols for passing the information from those
  models to the indicator component.
• Restructuring and teaching an Introduction to
  Urban Simulation course, scheduled for Spring
  2003. It will serve as a workshop and
  introduction to urban simulation using UrbanSim,
  focusing on the application of the model to the
  Puget Sound, and development of indicators and
  the creation and evaluation of scenarios.