Modeling Land Use Change in Chittenden County, VT Austin Troy*, Associate Professor, [email protected] Brian Voigt*, PhD Candidate, [email protected].

Modeling Land Use Change Modeling Land Use Change in Chittenden County, VTin Chittenden County, VT

Austin Troy*, Associate Professor, [email protected] Voigt*, PhD Candidate, [email protected]/envnr/countymodel

*University of VermontRubenstein School of Environment and Natural Resources

AgendaAgenda

8-8:30 continental breakfast8:30-8:45 opening remarks8:45-9:30 presentation on UrbanSim model

and discussion of preliminary model runs9:30-9:45 Q&A9:45-10:00 break10:00-11:00 scenario breakout groups 11:00-11:30 presentation of scenario results

and wrap-up

Team and CollaboratorsTeam and Collaborators

Graduate researchers: Brian Voigt, Alexandra Reiss, Brian Miles, Galen Wilkerson, Ken Bagstad

Co-PIs: Adel Sadek, Breck Bowden, Jon EricksonCollaborating organizations:

◦ Chittenden County Regional Planning Commission◦ Chittenden County Metropolitan Planning Organization◦ Resource Systems Groups, Inc, White River Junction, VT◦ University of Washington Center for Urban Simulation

and Policy Analysis: Paul Waddell, Alan Borning, Hana Sevcikova, Liming Wang

Project SupportProject Support Dynamic

Transportation and Land Use Modeling◦ Funder: USDOT Federal

Highway Administration Integrated Land-Use,

Transportation and Environmental Modeling: Complex Systems Approaches and Advanced Policy Applications. ◦ Funder: UVM

Transportation Center◦ Co Lead Investigator:

Adel Sadek

Research QuestionsResearch Questions

What will land use patterns in Chittenden County look like in 20-30 years?

What effect will future urban development patterns have on environmental quality?

How might alternative policies alter these outcomes?

How can we develop a model framework that effectively integrates the (inter)actions of households, employers, developers, transportation, and the environment?

This project …This project …IS

◦ A model development and integration exercise◦ An opportunity to visualize and assess the range

of land use outcomes resulting from alternative policies in Chittenden County

◦ A chance to explore how modeling tools can fit into the regional planning process

IS NOT◦ An adopted planning tool by any agency◦ An attempt to predict future land use policy◦ Part of the regulatory process◦ An analysis of any immediately proposed policies

Modeling FrameworkModeling Framework

UrbanSim: Land use model - www.urbansim.org

TransCAD (Caliper Corp.): four step travel demand model

Suite of indicators and environmental modulesPlans to also integrate with TranSims (FHWA)

and TransModeler (Caliper, Corp.) agent-based transportation models in coming year. This effort led by Adel Sadek, RSG, Inc and Austin Troy.

The Five D’s of UrbanSimThe Five D’s of UrbanSim

Data-intensiveDisaggregatedDynamicDisequilibriumDriven by

trends and forecasts

Model Coordinator

Database

Scenario Data

Control Totals

TDM

Exogenous Data

Output / Indicators

Modeling with UrbanSimModeling with UrbanSim

Model parameters based on statistical analysis of historical data

Integrates market behavior, land policies, infrastructure choices

Simulates household, employment and real estate development decisions◦agent-based for household and

employment location decisions◦grid-based for real estate development

decisions

from Waddell, et al, 2003

UrbanSim Decision MakersUrbanSim Decision Makers

Grid_ID: 60211

Employment_ID: 427

Sector: 2

Employees: 135

Grid_ID:23674

HSHLD_ID: 23

AGE_OF_HEAD: 42

INCOME: $65,000

Workers: 1

KIDS: 3

CARS: 4

Grid_ID:23674

Households: 9

Non-residential_sq_ft: 30,000

Land_value: 425,000

Year_built: 1953

Plan_type: 4

%_water: 14

%_wetland: 4

%_road: 3

Input DataEconomic

land value, employmentStructuresResidential and non-residential, size, year built

Biophysicaltopography, soils, wetlands, flood plains, water

Infrastructureroads, transit, travel time to CBD, distance to Interstate

Planning & zoningland use, development constraints

Householdsage of head of household, income, race, # of autos, children

Employmentemployment sector, number of employees

Control Totalspeople: total population, # of householdsjobs: # of jobs by employment sector

DATABASE

Land Price

Real Estate Development

Residential Land Share

Accessibility

Mobility & Transition

Location Choice

UrbanSim Model Algorithms

Importance of AccessibilityImportance of Accessibility

Location choice for employers, households is strongly determined by accessibility

Accessibility changes over time

Development follows those preferences

from Waddell, et al, 2003

Land PriceLand Price

Real Estate DevelopmentReal Estate Development

Residential Land ShareResidential Land Share

AccessibilityAccessibility

Mobility & TransitionMobility & Transition

Location ChoiceLocation Choice

• movers• vacant units• probabilities• site selection

Modeling with UrbanSimModeling with UrbanSim

Travel Demand Model

Travel Demand Model

Time Scenario Input Data

DATABASEDATABASE

Land Price

Residential Land Share

Accessibility

Mobility & Transition

Location Choice

Real Estate Development

Residential Land Share

Real Estate Development

Location Choice

Mobility & Transition

Land Price

Accessibility

Indicators

Households by Grid Cell: 2000

Model OutputModel Output

GRID_ID SECTOR_ID EMPLOYEES

77061 8 13

76187 11 256

75916 9 13

75938 13 18

75434 9 3

69171 8 1

68900 9 22

…

68902 8 3

Output database: defines grid cell state

IndicatorsGraphics

◦maps◦charts◦tables

Serv ices2866

Retail Trade1537

Finance432

Construction422

Wholesale Trade355

Standard IndicatorsStandard Indicators

Transport: VMT, accessibilityLand use: vacancy, non-residential sq ftLand value: residential, commercial,

industrialPopulation: total, density, summarize by

area (e.g. block group, TAZ)Households: count, type, incomeEmployment: count, type, sector

Initial model runs: Initial model runs: assumptionsassumptionsDevelopment constraints: limit development in low

density residential zones and cells with high % wetlands, no building on public lands

Employment sector: base year employment data from VT Sec of State business data and Claritas Business data

Households: base year households synthesized from 1990 US Census and PUMS 5% sample

Population and employment control totals from RPC/MPOStarting and ending year: 1990 - 2020Different versions of model

◦ With and without travel model◦ No constraints, minor constraints, significant constraints

199019952000200520102015

Residential Units Per Grid Cell: 1990 - 2015

Constrained by development type + travel model

Residential Units

Constrained by development type, no travel model

Limited constraints, no travel model

90 = Limited constraints + travel model

114 = Limited constraints, no travel model

90 = Limited constraints + travel model

114 = Limited constraints, no travel model

Population Density by Town: 1995 - 2015

19952000200520102015

People / Acre

Constrained by development type + travel model

Constrained by development type, no travel model

Environment IndicatorsEnvironment Indicators

Developing sub-modules that use UrbanSim output to estimate environmental impacts◦ Nutrient export (done)◦ Habitat fragmentation◦ Stormwater (Breck Bowden/Mary Watzin)*◦ Mobile source emissions (RSG Inc)*◦ Road particulate impacts (Sarah Lovell)*◦ Carbon footprint analysis (Jen Jenkins)*◦ Energy usage (Troy)**part of UVM Transportation Signature Project 1

UrbanSim Nutrient Indicators Limited constraints, no

travel model The total nutrient load

in 2020 would increase to 62150 kg/yr from 1992 levels (58527 kg/yr)

Calculated from Lake Champlain Basin Program nutrient export coefficients

Increase of 5 residential units in one gridcell equals transition to “urban” land use type

Troy et al. 2007 “Updating the Lake Champlain Basin land use data to improve prediction of phosphorus loading”, Lake Champlain Basin Program

Other Planned Other Planned ComponentsComponentsNetwork Robustness analysis (David

Novak)*“Sketch model” for comparison,

communication, and scenario development (Roel Boumans)*

Complexity/ sensitivity analysis for combined model framework (Adel Sadek, Chris Danforth, Peter Dodds, Jun Yu)*

Integrations with ARIES artificial intelligence-based data mining framework (Ferdinando Villa)

* Part of UVM UTC Signature Project 1

The to do list …The to do list …

This is a model in progress; although it’s running, it will be undergoing many improvements with time:◦Fix commercial sector model◦ Improve multi-city zoning inputs◦ Improve non-residential building data and

per employee square footage◦Review the base year database◦Expand development constraints◦Finalize model estimation◦Perform validation: 1990 – 2000◦Develop scenario data

Part 2: Scenario AnalysisPart 2: Scenario Analysis

UrbanSim and Scenario UrbanSim and Scenario AnalysisAnalysis What is a scenario?

◦ Alteration of baseline model inputs and assumptions

Types of changes that can be assessed◦ Zoning◦ Transportation

investments◦ Non-transportation

capital investments◦ State and regional

policy◦ Economic and

demographic changes◦ Behavioral

assumptions

* need TranSims for this analysis

BASE YEAR – business as usual

establish growth

center(s)policy event 1

employment opportunity

employment event

alter transport

infrastructure

investment

increase density

policy event 2

Scenarios: types of things Scenarios: types of things that can be modeledthat can be modeledConstraints to developmentRules for density, use, coverage, zoningMacro-scale transportation network (e.g.

highways, onramps, roundabouts, etc.)Micro-scale transportation network (e.g.

new lanes, turning rules, ITS, speed limits)Placement of public facilities (e.g. hospitals,

schools, courts, parks, arena, airports, etc.)Infrastructure (e.g. sewer, water, electricity)Siting of major employers/employment

centersControl totals

* need TranSims for this analysis

Scenarios: types of things Scenarios: types of things that can be modeledthat can be modeledBuilding assumptions (e.g. square

footage/employee, size of house relative to family size)

Transportation behavioral assumptions (e.g. trip chaining, carpooling, mode share, response of mode share to upgrades in service)

Speculative behavior assumptions (response of developers to new permitting rules, response of commuters and land market to high oil prices, response of shoppers or employers to greatly reduced parking)

* need TranSims for this analysis

Five scenarios for todayFive scenarios for today Based on input from last workshop and

on what we think is technically feasible1. Transportation corridor-oriented

development2. County-wide growth center

implementation3. Investment for increased regional road

connectivity4. Population boom5. Ultra Green

Scenario 1: Transportation Scenario 1: Transportation Corridor-Oriented DevelopmentCorridor-Oriented DevelopmentYour task is to develop specific

recommendations about land use and transportation for three corridors: route 7, route 2 and route 15.

How wide is the “corridor planning area?” This is up to you to decide.

There is a separate map for each corridor. Break into subgroups for each corridor if needed.

Feel free to build on existing corridor plans

Scenario 1: Transportation Scenario 1: Transportation Corridor-Oriented DevelopmentCorridor-Oriented DevelopmentTypes of actions that

could be proposed:◦redefine zoning

district boundaries for the corridors

◦change allowable densities, uses and lot coverage within them

◦Propose major roadway upgrades, like lane additions, roundabouts, divided highways

Scenario 1: Transportation Scenario 1: Transportation Corridor-Oriented DevelopmentCorridor-Oriented Development

◦ Propose pedestrian or bike paths◦ Propose non-infrastructural transportation

upgrades and  investments in the corridors, such as changes to public transportation lines and implementation of intelligent transportation systems to improve traffic flow

◦ Propose other capital investments for these areas, like new transit center/stops, government buildings, auditoriums, schools, parks, sewer lines, etc.

Scenario 2: Growth Scenario 2: Growth CentersCentersThis scenario is designed

to imagine what it would look like if the growth center concept was implemented to the greatest feasible extent in Chittenden County

Takes established Williston growth center as given; prospective Colchester growth center boundary is mapped but open to alteration in this exercise

It is up to the group to decide how many growth centers there will be and how large

Scenario 2: Growth Scenario 2: Growth CentersCentersPotential types of detail that could be prescribed

◦ Number, size and boundaries of growth center districts

◦ Sub-districts◦ Allowable uses, densities, lot coverage, zoning,

etc. for centers or sub-districts◦ Transportation network upgrades (lanes,

roundabouts, signals, ITS, etc.)◦ Public transit stops or lines◦ Infrastructure investments (sewer, water, parking

etc.)◦ Changes to allowable uses/densities outside

centers◦ Parking regulations◦ Other capital investments for these areas: transit

center/stops, new parking, government buildings, auditoriums, schools, parks, sewer lines, etc

Scenario 3: Scenario 3: Investment for Investment for Increased Regional Road Increased Regional Road ConnectivityConnectivityParticipants will propose

a set of road network enhancements that would increase connectivity between major corridors at the County level

Put differently: where are locations that are harder to access than they should be?

Can go off of list of MPO potential projects or define new project ideas

Results should be drawn on map and described in notes

Scenario 3: Scenario 3: Investment for Investment for Increased Regional Road Increased Regional Road ConnectivityConnectivityPotential types of prescribed projects

◦ New interstate exits/interchanges◦ Through access to PUDs located between

major corridors, but with access only to one◦ Upgrade of smaller order roads to arterial

capacity◦ Adding east-west connectors between major

north-south corridors (e.g. Route 7, Spear and Dorset, 116).

◦ Adding north-south connectors between major east-west corridors Route 15 and Route 2

Scenario 4: Population Scenario 4: Population BoomBoom Model used for this project takes population and

employment growth forecasts as given. Changes to those forecasts have a large impact on

simulated future land use patterns. Participants will imagine a Chittenden County with a far

larger influx of people than our forecasts predict, due to some unexpected influence (e.g. migrations due to global warming).

New forecasts will include estimates of the demographic breakdown of the new population, plus the employment sectors likely to see increases and reductions in jobs.

To accommodate this added growth, however, zoning will have to change.

The participants in this group will present one or two simple county-wide macro-scale zoning change alternatives that would be expected to accommodate that growth in residents and employment.

Scenario 5: Ultra-greenScenario 5: Ultra-greenThis scenario represents a

hypothetical set of regulations and investments designed to maximize the amount of green space in the County while minimizing the population's environmental footprint.

Be creative: don’t get too caught up in feasibility or practicality

Participants are encouraged to focus on a set of coordinated regional actions

Scenario 5: Ultra-greenScenario 5: Ultra-greenPossible types of policies to consider:

◦ urban growth boundary, ◦ significantly increased densities for

established urban centers,◦ new modes of public transportation, ◦ gas and energy taxes, etc. ◦ open space acquisition

You’ll get a map showing all the large undeveloped parcels and their acreage

We’ll give you a hypothetical cost/acre and a budget to work from

General Breakout Session General Breakout Session GuidelinesGuidelinesTry to be as specific as possibleKeep in mind what can feasibly be simulated

in modeling environmentPresent results after about 55 minutes. Leave

5 minutes for brief changeover to full groupDraw on maps with colored pens to illustrate;

we’ll digitally photograph at end and they can be projected for final presentation

Feel free to use any planning documentsDon’t feel constrained by current land use

regulationsDesignate note taker and keep official notes

to submit to organizersBe creative; don’t feel constrained.

Communicating our resultsCommunicating our resultsYour outputs will be extensively

documented (video, notes, maps)We will write up the details of the

scenarios and post them on our website (www.uvm.edu/envnr/countymodel) within two weeks

Where not feasible, we will so note in the online documentation

We will begin working on developing the scenario inputs in parallel with improving the baseline model

Communicating our resultsCommunicating our resultsThe website will contain information about

updates and improvements in the modelsWe will notify workshop participants when

we have a “stable release” of the baseline model for which we anticipate only minor modifications

At this point we will run the five scenarios defined today and then post the results.

The website will provide detailed results and downloadable data (maps, GIS layers, tables) from the baseline and alternative models.

We invite your involvement in this process

THANKS FOR YOUR THANKS FOR YOUR PARTICIPATION!!PARTICIPATION!!

www.uvm.edu/envnr/countywww.uvm.edu/envnr/countymodelmodel