Sprawl Dynamics

Prof. Philip C. EmmiCollege of Architecture +

PlanningUniversity of Utah

RailVolution 2005

SimTropolis

• Dynamic yet Radically Simplified• Metro-Regional Scale• Integrated Urban Land Use &

Transportation Scenario Generator• For the Exploration of Alternative

Urban Futures

Self-Reinforcing Relationship between

Urban Land Development Urban Road Building

Sprawl FeedbackSprawl Feedback

As Urban Land Develops, Traffic and As Urban Land Develops, Traffic and Vehicle Miles Traveled IncreasesVehicle Miles Traveled Increases

1

Increased Road Capacity Increased Road Capacity Induces Intensity IncreasesInduces Intensity Increases

3

At Higher Intensities, More Land is At Higher Intensities, More Land is Needed for People and JobsNeeded for People and Jobs

4

Increased Travel Demand Increased Travel Demand Leads to More RoadsLeads to More Roads

2

Study Area and Time Horizon

• Urbanized Portions of the Salt Lake, Davis and Weber Counties

• Initialized with 1980 data• Model calibration data: 1980 - 2000• Validation experiments

– Calibration: 1980 - 1992.– Project and compare with 2002

observations

• Projection and policy horizon: 2030

URBANLAND

ROADMILES

developingAcres

road building

InitialUrban Land

InitialRoadMi

Ac\Person

PctDensityIncrease

Vehicle Travel

TargetRoadMiles

DesiredRoadMiles\DVMT

PctTripsNotSingleOccupancy

NonAutoTrafficEquiv

RoadGap

RoadBuildingImplement'nLag

PostPolicyDVMTs

SmoothDensityChange

NonAutoSmthStep

ImplementLag2

Chng InAc\Person

BetterCapacityFromExistingRoads

SmoothCapacityChange

ImplementLag1

PEOPLE

Chng InPeople

Developing Land & Building Roads

System EquationsDVMTs X 10^3 = f(Urban Acres)

y = 0.1221*Acres - 3979

R2 = 0.999

14000

16000

18000

20000

22000

24000

26000

28000

30000

140000 160000 180000 200000 220000 240000 260000 280000

Acres/Capita = f(Road MIles)

Acres/Capita = 0.00001435 * RoadMiles + 0.1316R2 = 0.9975

0.165

0.17

0.175

0.18

0.185

0.19

0.195

0.2

0.205

2500 3000 3500 4000 4500 5000

DVMTs = 0.12214 * (URBAN_LAND) – 3979

Road Miles at t+5 per DVMT*10^3 at t)

RoadMiles(t) / DVMT(t+5) = -0.00133* Year + 2.858

0.195

0.2

0.205

0.21

0.215

0.22

0.225

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998

DesiredRoadMiles\DVMT = -0.00133 * TIME + 2.858

Ac\Person = 0.00001435 * ROAD_MILES + 0.1316

TargetRoadMiles = DesiredRoadMiles\DVMT * DVMTs

RoadGap = (TargetRoadMiles - ROAD_MILES)

road_building = (RoadGap / RoadBuilding_Implement'nLag)

developingAcres = (PEOPLE * Chng_In_Ac\Person) +

(Ac\Person * Chng_In_People)

Percent Change in Urban Performance Indicators, 1980 - 2000

Test Experiment #1• Divide the 1980 - 2002 data set into

2 halves• Calibrate the model on data from

1980-1992• Project to 2002, compare with 2002

observations and note the errors• Find the RMS error of the projections

for:– Urban Land, VMTs, Road Miles & Density

• RMS Error is found to be low

Test Experiment #2

• Note slight logarithmic structure to data on Ac\Person as a function of Road Miles.

• Fit a logarithmic function to the data. • Use it in place of a linear function.• Note effects on year 2030 simulations.

– In the baseline scenario, values are 10-15% lower.

– In other scenarios, values are 1 - 3% lower.

Test # 3 (Incomplete)• Note that there is little agreement among

alternative data series on urban land.• Would the model behave differently if another

data series on urban land were used?• Would the model behave differently if it were

calibrated on the CO Front Range?• Does the model’s data determine its

characteristic behavior or its structure?• Does the model’s data or its structure

determine the (in)accuracy of its historical simulations?

Percent Change in Urban Performance Indicators, 1980 - 2030

Four Alternative Future Scenarios

1. Baseline: an extension of historic practice2. Land Use:

Increase new development densities by 20%

3. Transportation: Reduce single-occupancy vehicle trips to 80% of totalIncrease existing road capacity by 12%Speed up TIP implementation from 5 to 3 years

4. Combine Land use & Transportation

Urban Land Consumption:Four Scenarios

Daily Vehicle Miles Traveled

Road Miles

Intensity of Land Use

The Density of New Development

Acres Developed Per Year

Crude Road Congestion Index

Extensions

•Agricultural land preservation•Water use and conservation•Urban forest and canopy cover•Vehicle fuel and energy use•Carbon dioxide emissions•Criteria air pollutants

Conclusions about Urban Dynamic System

Models• They are simple to understand• They are highly accurate• They are easy to manipulate• They illustrate dynamic complexity• They generate alternative policy

scenarios• They facilitate learning about policy

options• They might facilitate policy negotiation