Spatial Input-Output Models: PECAS - WebHome - UrbanSim

urbansim.org
  • No tags were found...

Spatial Input-Output Models: PECAS - WebHome - UrbanSim

Spatial Input-OutputModels: PECASWebinar 4 of an 8-part TMIPWebinar series on land useforecasting methods.Tuesday, May 24, 2011


Paul Waddell, 2011PECAS Background• PECAS (the Production, Exchange and Consumption Allocation System) is anurban and regional modeling tool to support transportation and economicplanning• Developed by Dr. Doug Hunt and Dr. John Abraham, University of Calgary• Contains two principal models:- Activity Allocation (AA): an aggregate, equilibrium Spatial Input-Output Model- Spatial Development (SD): a disaggregate State-Transition model• Developed initially as part of an Oregon Department of Transportation (ODOT)Statewide Modeling project as a replacement for a 1st generation statewidemodel using TRANUS• Recently, CalTrans implemented a contract with UC Davis to supportdevelopment of a California Statewide PECAS model, and to support MPOswithin the state in the development of metropolitan level PECAS modelsTuesday, May 24, 2011


Paul Waddell, 2011Application:Current Applications (from model developers)• Oregon, USA State-wide- part of larger modelling system with micro-simulation components• Ohio, USA State-wide- Model designed and used as basis for data collection• Sacramento Area, USA- Part of larger modelling system with micro-simulation components• Calgary Region, Canada- Design for new urban level modelling system• Edmonton Region, Canada- Design for new urban level modelling system• Baltimore Metropolitan Area- Design for new urban level modelling systemTuesday, May 24, 2011


Paul Waddell, 2011The State of the Practice: Survey of MPOs in 20105Tuesday, May 24, 2011


Paul Waddell, 2011PECAS Overview: Spatial Development (SD)• State transition style model of stochastic change of cells or parcels to alternativeland use• Followed initial version of UrbanSim (1998) parcel and gridcell developer modelusing this approach (later UrbanSim versions moved to other formulations)• Unlike AA, SD is disaggregate, at gridcell or parcel level• Uses pricing (rents) from AA and development costsTuesday, May 24, 2011


Paul Waddell, 2011Theoretical Basis: Input-Output ModelsPECAS’s core (the AA) is a spatial input-output model• This venerable approach represents an economy as a matrix- cells contain values representing the amount of economic activity (production orconsumption) for a particular combination of sectors- equations represent the interlinkages between portions of the economy and allowchanges in one area to be traced through to other areas- tracking the activities and flows by geographic location makes the table spatial• Now a brief review of this approachTuesday, May 24, 2011


Paul Waddell, 2011Example I-O Expenditure Table for RetailIn order to produce a total output of $20000, the retail sector consumes inputsfor its production process. Assume the following inputs are purchased to producethe $20000 of retail output, based on the production process for retail:RetailBasic$5000Retail$2000Services$3000Tuesday, May 24, 2011


Paul Waddell, 2011Example I-O Direct Input Requirements MatrixThis table shows the standardized inputs per dollar of output for each industry,also known as technical coefficients.Basic Retail ServicesBasic0.15 0.33 0.05Retail0.25 0.13 0.25Services0.30 0.20 0.10Tuesday, May 24, 2011


Paul Waddell, 2011Multiplier EffectsBasic Retail ServicesTotal Output $10000 $15000 $25000Induced ConsumptionFirst IterationTuesday, May 24, 2011


Paul Waddell, 2011Economic Flows Can be Split by Region (Spatial I-O)EconomicActivitiesRegion ARegion BBasic Retail Services Basic Retail ServicesFinal Demandand ExportsTotal DemandBasicRegion ARetailServicesBasicRegion BRetailServicesFinal Paymentsand ImportsTotal InputsTuesday, May 24, 2011


Paul Waddell, 2011Economic Flows Can be Further Split into CommoditiesProduced and Consumed (Make and Use Tables)EconomicActivitiesRegion ARegion BBasic Retail Services Basic Retail ServicesFinal Demandand ExportsTotal DemandBasicRegion ARetailServicesBasicRegion BRetailServicesFinal Paymentsand ImportsTotal InputsTuesday, May 24, 2011


Paul Waddell, 2011Economic Flows Can be Further Split into CommoditiesProduced and Consumed (Make and Use Tables)Non-Residential FloorspaceHousingLaborGoods and ServicesEconomicActivitiesRegion ARegion BBasic Retail Services Basic Retail ServicesFinal Demandand ExportsTotal DemandBasicRegion ARetailServicesBasicRegion BRetailServicesFinal Paymentsand ImportsTotal InputsTuesday, May 24, 2011


Paul Waddell, 2011PECAS Software Architecture• Base PECAS system consists of two major Java modules (the AA and the SD)and supporting infrastructure• Model runs initiated using DOS shell or Python script• Most data stored and passed between modules in CSV format- Scenario inputs and parameters are set by creating CSV files- Most model outputs are also in many CSV files• Parcel information is stored in a database such as SQL Server or PostGIS• Data preparation requires GIS and statistical software• Loose integration with travel model through squeezed skims in CSV• Runs on a multi-processor server- Calibration can take days for a single run- Multi-decade projections can take hoursTuesday, May 24, 2011


Paul Waddell, 2011Activity Allocation (AA) ModuleInputs and Data Sources (1)• Aggregate economic flow: IMPLAN- Demargined for wholesale and retail• Synthetic households by TAZ- Census PUMS- Census SF 3 summary files- Automated in Python• Synthetic employment (by industry and occupation)- CTPP- InfoUSA- Automated in Python• Technology options- Aggregate economic flow; Census PUMS; cluster analysisSource: Shengyi Gao (et al)Tuesday, May 24, 2011


Paul Waddell, 2011Activity Allocation (AA) ModuleInputs and Data Sources (2)• Floorspace inventory- EIA Space use survey- Synthetic employment- Existing land use• Transport costs- BTS commodity flow survey- Midday skims from the travel model- Logsum of mode choice by trippurpose• Rent• Vacancy rate- Census SF 3 summary files- CoStar data• Imports and exports- BTS commodity flow survey- IMPLAN- TradeViewTM, zepol- DataQuick transactions in 2000(residential and non-residential)- CoStar (non-residential)Source: Shengyi Gao (et al)Tuesday, May 24, 2011


Paul Waddell, 2011Space Development (SD) ModuleInputs and Data Sources (1)• General land use plan- Generalized city/county general landuse plans- 35 land use types• Base parcel database- Existing land use type- Zoning- Year built• Rent modifier- Distance to freeways- Distance to ramps- Distance to highways- Distance to beaches- Distance to parks- Distance to schools- Distance to rail roadsSource: Shengyi Gao (et al)Tuesday, May 24, 2011


Paul Waddell, 2011Space Development (SD) ModuleInputs and Data Sources (2)• Construction cost- RSMeans data• Maintenance cost• Typical FAR• Density rent discount• Demolition costs• Age discount- Multiple sources• Maximum/minimum intensity- Zoning ordinance• Development fees- HCD databaseSource: Shengyi Gao (et al)Tuesday, May 24, 2011


Paul Waddell, 20111. PECAS Overview2. Anatomy of the Systema. Model Designb. Software Architecturec. Estimation, Calibration, and Validation3. Application in Practice4. Comparison and AssessmentTuesday, May 24, 2011


Paul Waddell, 2011Economic InteractionsActivity TotalsConsumptionsLabor and CapitalSupplyActivityLocationsOccupanciesLand andFloorspace SupplyEnvironment(externalities)ActivityInteractionsPrice SignalsSocialImpactsTransportDemandsFlowsTransportSupplyTuesday, May 24, 2011


Paul Waddell, 2011PECAS OverviewActivity TotalsConsumptionsLabor and CapitalSupplyPECASActivityLocationsOccupanciesLand andFloorspace SupplyEnvironment(externalities)ActivityInteractionsPrice SignalsSocialImpactsTransportDemandsFlowsTransportSupplyTuesday, May 24, 2011


Paul Waddell, 2011Model Design DiagramAtlantasource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011Model Design Diagram DetailsData Items Group 7ProductionandConsumptionofCommoditiesbyNon-HouseholdActivitiesCensus Employment by TAZIMPLAN Input-Output TablesData Items Group 8Imports and Exportsof CommoditiesIMPLAN Input-Output TablesBTS Commodity Flow StatisticsBusiness Export StatisticsData Items Group 1Space Parcel DevelopmentDatabaseDevelopment Event RecordsData Items Group 2Production of Laborby HouseholdsCensus PUMS Employment DataData Items Group 3Co-Star Use Non of Residential Spaceby Non-Household Floorspace Use Data ActivitiesData Items Group 6Consumptionof Commoditiesand LaborCensus Households by TAZby Family Expenditure SurveyIMPLAN Social AccountsIMPLANGroupInput-Output9TablesFinancial FlowsTransport ModelGroup 4Census PUMS Employment DataIMPLANUseSocialof LaborAccountsby Non-Household ActivitiesData Items Group 5Use Residential of SpaceFloorspaceby HouseholdsUse DataCensus PUMS and AHSsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011Activity Allocation (AA) Module• Aggregate spatial input-output model• Represents interaction of activities through commodity flows- Food shipping to a processing plant to store- Person driving to work• Travel model provides the yearly description of disutility of movement betweenlocations (congestion) that underly activity interaction- e.g Congestion might move two interdependent industries closer together- e.g. A new highway might drive development of new subdivisions• Connection with SD- Activities occupy floorspace build by the SD- Spatial choices of activities drive prices that motivate SD developerTuesday, May 24, 2011


Paul Waddell, 2011Activity Allocation (AA) ModuleActivities and Commodities• Activities- Industries: 63 (electricity utilities emphasized)- Households: 25, including 5 all seniors household types• Commodities- Goods and services: 60 (including fuel, electricity, GHG permit, agriculture wateruse, etc.)- Labor: 19- Space: 38 (14 residential types; 24 non-residential types)• Zone system- Land use zone: 526- Floorspace zone (TAZ): 5191Counts are from California State model applicationTuesday, May 24, 2011Source: Shengyi Gao (et al)


Paul Waddell, 2011Activity Allocation (AA) ModuleDecision TreeLocation ChoicesProduction/Consumption ChoicesExchange Location ChoicesTuesday, May 24, 2011


Paul Waddell, 2011Economic InteractionsProduction - Exchange – ConsumptionMake and use with exchange zonesselling allocationprocessbuying allocationprocessExchangeZoneTotalConsumptionExchangeZonecommodity flowsExchangeZoneTotal ProductionTotal ProductionTotal ProductionTuesday, May 24, 2011


Paul Waddell, 2011Economic InteractionsProduction - Exchange – Consumption1:productionallocationallocating productionactivity to zones2:technologyselectionallocatingproduction tocommoditiesallocatingconsumption tocommodities3:sellingallocationsbuyingallocationsallocating produced commoditiesto selling locationsallocating consumedcommodities to buying locationsTuesday, May 24, 20113-level nested logit model


Paul Waddell, 2011Activity Allocation (AA) ModuleJoint Discrete UtilityAdditional utility associatedwith location l for activity aAdditional utility associatedwith production option pStochastic error termsQuantity of commodityproduced or consumedunder production option pUtility of exchanging andshipping one unit ofcommodity between l and eTuesday, May 24, 2011Source: Atlanta Regional Commission


Paul Waddell, 2011Space Development (SD) Module• Disaggregate process at the parcel level- Grid cells or parcels• Represent developers’ actions• Connection with AA- From AA: current year space price at LUZ level- To AA: quantity of the spaces for next year AA• Space is a commodity consumed by the activities in the AA model- Unlike other commodities, space cannot be transported- Different activities consume different types of space• e.g. in Atlanta there are 8 PECAS space types (A/D/S/M/O/R/L/H)• Rents are space prices• Zoning rules limit the type of space the can be developed on a parcelTuesday, May 24, 2011Source: Atlanta Regional Commission


Paul Waddell, 2011Space Development (SD) ModuleDevelopment Events• Year-by-year step• Possible development events- E0: no change- En: new space type and quantity- Er: alter or renovate- Ed: derelict• Two step process for each parcel- Selection of development events and update space type- Update space amount• Data needs- Permits- Parcel level data- RentsTuesday, May 24, 2011Source: Atlanta Regional Commission


Paul Waddell, 2011Space Development (SD) ModuleRents• Space prices are rents for the use of space• Per unit of space per unit of time• Rent equation:- Space price at LUZ level in AA (done by AA & SD integration)- Local-level effects due to:• Density of development around the parcel• Age of the structure• Local Effects: distance from (or proximity to) local-level influences• Expressway• Interstate exit• Major road• School• Marta• Green spaceTuesday, May 24, 2011Source: Atlanta Regional Commission


Paul Waddell, 2011Space Development (SD) ModuleSimulation of TransitionsParcel-by-parcel microsimulationindustrialcommercialmid density residentialmore the sameno changederelictquantitySource: Atlanta Regional Commissionzoning dictates set of alternativesTuesday, May 24, 2011


Paul Waddell, 2011Space Development (SD) ModuleDecision TreeDevelopment EventsBuilding TypesDevelopment QuantitiesTuesday, May 24, 2011


Paul Waddell, 2011Space Development (SD) ModuleJoint Discrete UtilityRent less amortizedconstruction cost per unitspaceAdditional Rent lessdevelopment costs per unitlandSpace quantity (buildingsize)Land quantity (parcel size)Stochastic error termsTuesday, May 24, 2011Source: Atlanta Regional Commission


Paul Waddell, 2011Space Development (SD) ModuleParcel Level Data and Derived Floorspace• For each parcel:- Area of the parcel- Existing space type- Existing space quantity (buildingfloorspace)- Structure year- Zoning rules (allowable uses anddensity range)- Cost and fees (associated withdevelopment of each permittedspace type and quantity)• Challenges (20 Counties: everydataset is different)- Parcel features and ID- Parcel attributes (building floorspace,space type…)- Geocoded points for Clayton…- Combine parcel with tax assessors’data- Updates• 20-county parcels are cleaned andloaded- About 2 million parcels are cleaned• Benefit other planning projectsTuesday, May 24, 2011Source: Atlanta Regional Commission


Paul Waddell, 2011Space Development (SD) ModuleParcel Level Data and Derived Floorspace20-County parcel featuresSource: Atlanta Regional CommissionCounty ParcelsBarrow 28,184Bartow 42,167Carroll 50,633Cherokee 93,866Clayton 88,723Cobb 228,690Coweta 55,348DeKalb 230,888Douglas 39,140Fayette 42,808Forsyth 77,639Fulton 341,017Gwinnett 260,371Hall 77,103Henry 72,839Newton 44,374Paulding 59,670Rockdale 34,780Spalding 29,616Walton 36,561Total 1,934,417Tuesday, May 24, 2011


Paul Waddell, 2011Space Development (SD) ModuleParcel Level Data and Derived Floorspace• Why do we need the derived space?- The quality of the parcel space data: very inconsistent- The (in) consistency between employment and space- Mixed use issues• Derived using NAICS employment and Landpro• New space totals at LUZ and disaggregate to TAZ• Then, evaluation…Tuesday, May 24, 2011Source: Atlanta Regional Commission


Paul Waddell, 2011Space Development (SD) ModuleParcel Level Data and Derived Floorspace• FloorSpace Synthesizer Tool• Based on existing space type, quantity and zoning…• CalibrationTuesday, May 24, 2011Source: Atlanta Regional Commission


Paul Waddell, 2011Space Development (SD) ModuleParcel Level Data and Derived Floorspace• Calibration tasks:- Evolution of initial synthesizedresults- Directing synthesizeddevelopment to actual built-onparcels- Directing the correct space typeto the parcel- Directing the synthesized builtspace to developed parcel inamount resembling actualquantitiesTuesday, May 24, 2011Source: Atlanta Regional Commission


Paul Waddell, 2011Space Development (SD) ModuleParcel Level Data and Derived FloorspaceSynthesizedResultsTuesday, May 24, 2011Source: Atlanta Regional Commission


Paul Waddell, 2011Full PECAS System Through TimeTravel Model(aggregate now disaggregate later)Model InputsCSV, Excel,DatabaseScriptingBatch filesPythonAA core codeJavaModel Outputs.CSVSD core codeJavaGeovisualizationGISMapserverLU SynthesizerJavaParcel DatabaseDatabasesource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011Integrating PECAS and Travel ModelTuesday, May 24, 2011


Paul Waddell, 2011PECAS 3-Stage Calibration Approach• Stage 1 - the S1 parameters- Consider each module separately- Based on specific, separate dataset- Often ‘disaggregate data’- Often statistical estimation- Fixed for remainder of calibration• Stage 2 - the S2 parameters- Consider each module separately- Based on module hitting targets- Often ‘aggregate data’- Some also S3 parameters- Specialized software developed• Stage 3 - the S3 parameters- Consider all modules linked together- Based on module hitting targets- ‘Aggregate data’- Certain S2 parameters also S3parameters, process updates thesein response to total model behaviour- Specialized software developedTuesday, May 24, 2011


Paul Waddell, 20111. PECAS overview2. Anatomy of the System3. Application in Practice4. Comparison and AssessmentTuesday, May 24, 2011


Paul Waddell, 2011Sacramento Blueprint StudyTuesday, May 24, 2011


Paul Waddell, 2011Sacramento Blueprint StudyTuesday, May 24, 2011


Paul Waddell, 2011Sacramento Blueprint StudyTuesday, May 24, 2011


Paul Waddell, 2011SACOG Equity AnalysisFIGURE 2 Household and employment location in the PRB scenario6source: Equity Analysis of Land Use and Transport Plans Using an Integrated Spatial Model. Rodier, Abraham, Dix, and Hunt. UCD-ITS-RR09-46Tuesday, May 24, 2011


Paul Waddell, 2011SACOG Equity AnalysisTuesday, May 24, 2011FIGURE 1 Household and employment location in the BAU scenariosource: Equity Analysis of Land Use and Transport Plans Using an Integrated Spatial Model. Rodier, Abraham, Dix, and Hunt. UCD-ITS-RR09-465


Paul Waddell, 2011SACOG Equity AnalysisFIGURE 3 Percent Change in Dwelling Units by Type Between the BAU and the PRBSFD=single family dwelling units; MFD=multi family dwelling unitssource: Equity Analysis of Land Use and Transport Plans Using an Integrated Spatial Model. Rodier, Abraham, Dix, and Hunt. UCD-ITS-RR09-46Tuesday, May 24, 2011


Paul Waddell, 2011SACOG Equity Analysis20source: Equity Analysis of Land Use and Transport Plans Using an Integrated Spatial Model. Rodier, Abraham, Dix, and Hunt. UCD-ITS-RR09-46Tuesday, May 24, 2011


Paul Waddell, 2011SACOG Equity AnalysisFIGURE 7 Percentage change in annual worker and industry transport cost from the BAU tothe PRB scenarioTABLE 1 Average Annual Transport Cost (TC) by and across Labor Group(s) (2000 U.S.nominal dollars)Labor GroupChange inTC (dollars)PercentageChange in TCBAU: TC asIncome ShareAgriculture -326 -11.8 6.0 5.2Construction -303 -11.1 5.8 5.2Educators -170 -6.6 5.8 5.6Entertainers -372 -14.2 5.7 5.0Food -250 -9.9 5.5 5.0Health -306 -11.9 5.3 4.8Maintenance & repair -300 -11.1 5.9 5.3Managers -339 -13.0 5.5 4.8Non-retail sales -426 -15.9 5.7 4.9Office & administrative -323 -12.7 5.4 4.8Production -293 -10.9 5.9 5.3Professionals -351 -13.4 5.4 4.8Retail sales -256 -9.9 5.5 5.0Service -306 -12.0 5.4 4.9Transport -281 -10.6 5.9 5.4Total -307 -11.8 5.5 5.0PRB: TC asIncome Sharesource: Equity Analysis of Land Use and Transport Plans Using an Integrated Spatial Model. Rodier, Abraham, Dix, and Hunt. UCD-ITS-RR09-46Tuesday, May 24, 2011


where such homes are typically located have driven up average rents for the highest income classand mitigated declines relative to the regional mean. It also appears that the low and middle Paul Waddell, 2011income household categories have benefited from the significantly increased supply of multifamilyhousing and lower transport costs in the areas in which they are located (i.e., the innersuburbs and central business areas). Upper income households are also more likely to be owneroccupied,and thus receive less benefit from reductions in rent than do lower income households.Note that this sort of reduction in rents will not necessarily lead to an increase in consumersurplus in AA, since AA also represents the greater preference for single family dwelling units.The PRB scenario reduces opportunities for housing, which reduces consumer surplus, but alsoreduces rents for housing, which increases consumer surplus. AA represents both of these andweighs them against each other.SACOG Equity AnalysisTABLE 2 Change in Average Annual Rent by and across Household Class(es) (2000 U.S.nominal dollars)Income Class ($1,000) Total Change (dollars) Percentage changeless than 10 -1,248 -6.410 to 19 -1,299 -6.020 to 39 -1,702 -8.040 to 49 -1,833 -7.950 to 99 -1,933 -6.7100 to 199 -309 -0.7200+ 505 1.0Total -1,526 -6.122source: Equity Analysis of Land Use and Transport Plans Using an Integrated Spatial Model. Rodier, Abraham, Dix, and Hunt. UCD-ITS-RR09-46Tuesday, May 24, 2011


Paul Waddell, 2011SACOG Equity AnalysisTuesday, May 24, 2011Table 3 shows that the total annual value of owned homes decreases across all income bracketswith the exception of the highest income category. This is consistent with the general decreasesand increases in rental values by income class and indicates a change in residential propertyvalue.TABLE 3 Total Annual Value of Owned Homes (2000 U.S. nominal dollars)Table 3 shows Household that the Income total annual ($1,000) value BAU of owned ($100,000) homes decreases PRB ($100,000) across all income Percentage brackets Changewith thelessexceptionthan 10of the highest income7,840category. This is consistent7,788with the general decreases-0.7and increases in rental values by income class and indicates a change in residential property10 to 19 13,384 13,201 -1.4value.20 to 39 38,520 37,578 -2.4TABLE 340 Total to 49 Annual Value of Owned 20,868 Homes (2000 U.S. 20,258 nominal dollars) -2.9Household 50 to Income 99 ($1,000) BAU ($100,000) 124,620 PRB ($100,000) 121,462 Percentage -2.5 Changeless than 100 to 199 7,840 78,739 7,788 78,122 -0.7 -0.810 to 19200 or more 13,384 15,298 13,20115,415 -1.4 0.820 to 39 Total 38,520 299,268 37,578293,823 -2.4 -1.840 to 49 20,868 20,258 -2.950 to 99 The results suggest that lower 124,620 transport and housing 121,462 costs in the PRB -2.5 scenario have driven downthe region’s cost of living, and thus average annual wages (see Table 4). Average wage income is100 to 199 78,739 78,122 -0.8reduced by $783 (a percentage reduction of 1.6%). By labor occupation category, average200 or more reduction ranges from a low 15,298 of $50 to a high 15,415 of approximately $1,000 0.8 (percentage reductions ofTotal 0.1% to 2.0%, respectively). 299,268 Agricultural and 293,823 construction workers, -1.8 typically lower income jobs,experience some of the lowest reductions and professional, sales, and administrative labor groups,The results typically suggest higher that income, lower transport experience and some housing of the costs highest in the reductions. PRB scenario have driven downthe region’s cost of living, and thus average annual wages (see Table 4). Average wage income isreduced TABLE by $783 4 Change (a percentage in Average reduction Annual of 1.6%). Wage Income By labor by occupation and across category, Labor Group(s) average (2000U.S. nominal dollars)reduction ranges from a low of $50 to a high of approximately $1,000 (percentage reductions of0.1% to Labor 2.0%, Group respectively). Agricultural Total Change and construction (dollars) workers, Percentage typically Change lower income jobs,experienceAgriculturesome of the lowest reductions-50and professional, sales,-0.1and administrative labor groups,Construction -282 -0.6typically higher income, experience some of the highest reductions.Educators -802 -1.8Entertainers -925 -1.9TABLE 4 Change in Average Annual Wage Income by and across Labor Group(s) (2000U.S. nominalFood workersdollars)-752 -1.6Health workers -847 -1.7Labor Group Total Change (dollars) Percentage ChangeMaintenance & repair -731 -1.6Agriculture -50 -0.1Managers -922 -1.9Construction -282 -0.6Non-retail sales -951 -2.0Educators -802 -1.8Office & administrative -892 -1.9Entertainers -925 -1.9Production -670 -1.4Food workers -752 -1.6Professionals -980 -2.0Health workers -847 -1.7Retail sales -759 -1.6Maintenance & repair -731 -1.6Service -749 -1.5Managers -922 -1.9Transport -719 -1.6Non-retailTotalsales -951-783-2.0-1.6Office & administrative -892 -1.9Production -670 -1.4Professionals -980 -2.0Retail sales -759 23 -1.6Service -749 -1.5Transport -719 -1.6source: Equity Analysis of Land Total Use and Transport Plans -783 Using an Integrated Spatial -1.6 Model. Rodier, Abraham, Dix, and Hunt. UCD-ITS-RR09-46


Paul Waddell, 2011SACOG Equity AnalysisTABLE 5 Total and Average Consumer or Producer Surplus for PRB Scenario Relative tothe BAU Scenario (2000 U.S. nominal dollars)Industry Activities Total ($100,000) Average(per million dollars of production)Agriculture 254 13,819Construction 944 8,783Manufacturing 962 5,588Transport 249 12,336Communication 483 9,630Wholesale trade 996 8,532Retail 5,354 20,345Restaurants 2,281 51,192Financial 1,961 18,934Real estate 1,330 6,804Business services 1,200 15,477Automotive services 308 13,994Amusement services 197 46,647Education 717 36,163Personal services 697 35,366Non-profit organizations 565 48,809Professional services 1,213 17,099Government 2,916 15,501Total 22,626 15,028Household Income Class ($1,000) Total ($100,000) Average per Householdless than 10 731 1,00810 to 19 1,226 1,07420 to 39 1,617 64740 to 49 254 22950 to 99 -1,966 -442100 to 199 -1,384 -668200+ -151 -454Total 327 27source: Equity Analysis of Land Use and Transport Plans Using an Integrated Spatial Model. Rodier, Abraham, Dix, and Hunt. UCD-ITS-RR09-46Tuesday, May 24, 2011


Paul Waddell, 2011SACOG Equity AnalysisTABLE 6 Total and Change in Annual Values of Space Categories (2000 U.S. nominaldollars)BAU Total PRB Total Total Change Average($100,000)($100,000) Change($100,000)Industry SpaceAgriculture & Mining 43 48 5 0.3Industrial 3,424 3,504 79 0.1Office 22,561 22,729 169 0.1Retail 24,205 24,240 35 0.0Medical 26,152 26,200 48 0.1Primary School 7,434 7,436 1 0.0Colleges & Education 2,653 2,655 1 0.0Government Office 31,015 31,002 -13 0.0Total 117,488 117,813 325 0.0Residential SpaceLuxury SFD 195,707 185,408 -10,299 549.0Standard SFD 153,245 152,531 -714 -243.0Owned MFD 8,976 9,322 345 -1017.0Rented MFD 26,510 27,069 559 -1537.0Total 384,438 374,330 -10,108 -820.0SFD=single-family development; MFD=multi-family developmentsource: Equity Analysis of Land Use and Transport Plans Using an Integrated Spatial Model. Rodier, Abraham, Dix, and Hunt. UCD-ITS-RR09-46Tuesday, May 24, 2011


Paul Waddell, 2011CalSIM• California Statewide Integrated Model• Integrated PECAS land use model and new statewide activity-basedtransportation model• Spurred by California SB375: land use related reductions from autos and lighttrucks• Funded by CalTrans in conjunction with metropolitan-level upgrades• Massive data collection and imputation effort• Timeline- Transportation model built and calibrated during 2010- Land use model calibration ongoing- Metropolitan models ready by 2015• Preliminary resultsTuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMConstrained 0 iteration model (supply/demand not matched)108642021 0 1 2 3 4 5 6 7 8 9AGRICULTURE (Animals)PRODUCTAGRICULTURE (ForestryandFishing)PRODUCTAGRICULTURE (Plants)PRODUCTCONSTRUCTION (space)PRODUCTHomeToOtherHomeToRecHomeToShopLaborMANUFACTURING (ChemicalsPlasticRubberGlassCement)PRODUCTMANUFACTURING (ComputersElectronics)PRODUCTMANUFACTURING (ElectricalEquipmentAppliance)PRODUCTMANUFACTURING (Food)PRODUCTMANUFACTURING (Machinery)PRODUCTMANUFACTURING (MetalSteel)PRODUCTMANUFACTURING (PetroCoalproduction)PRODUCTMANUFACTURING (PulpandPaper)PRODUCTMANUFACTURING (Textiles)PRODUCTMANUFACTURING (Transportation Equipment)PRODUCTMANUFACTURING (WoodProductsPrintingFurnitureMisc)PRODUCTMINING ANDEXTRACTION PRODUCTOtherToOtherOtherToRecreationSCRAPTransportationCommUtilandWholesaleWorkToOthersource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIM Develop Target space quantity transitions 10 counties selected to represent low medand high growth situations, plus SanFrancisco as a special county Low: Sacramento, San Diego, Orange County Med: Amador, Inyo, Shasta High: Fresno, Imperial, Placer San Franciscosource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 2011CalSIMsource: Developing California Integrated Land Use/Transportation Model. Gao, Lehmer, Wang, McCoy, Johnston, Abraham, and Hunt. Presented at TRB 2010Tuesday, May 24, 2011


Paul Waddell, 20111. PECAS overview2. Anatomy of the System3. Application in Practice4. AssessmentTuesday, May 24, 2011


Paul Waddell, 2011Strengths of Input-Output Models• I-O Models provide a concise summary of the economic flows in the economy• Multipliers from I-O models are used widely to predict the impact of changes inoutput of a sector on the broader economy - the multiplier effect• With suitable data, national I-O models can be localized to states or possiblylower units of geography- Keep in mind the model represents economic flows between every geographic unitand every sector, as in an international trade model - so the data requirements togenerate a highly disaggregate I-O model are immenseTuesday, May 24, 2011


Paul Waddell, 2011Limitations of Input-Output Models• Wikipedia’s article on Input-Output models provides the following assessment:- “Input-output is conceptually simple. Its extension to a model of equilibrium in thenational economy is also relatively simple and attractive but requires great skill andhigh-quality data. One who wishes to do work with input-output systems must dealskillfully with industry classification, data estimation, and inverting very large, illconditionedmatrices. Moreover, changes in relative prices are not readily handled bythis modeling approach alone.”• I-O model theory does not account for the effects of changes in relative prices onproduction functions of firms, and therefore on the I-O structure• I-O model does not allow flexible substitution among inputs and price adjustment• I-O model deals only with monetary flows in the economy, not quantities ofemployment, households, population, etc.• I-O model is an aggregate, static equilibrium model, with no capacity torepresent effects of heterogeous agents, temporal dynamics, changes inproduction technologyTuesday, May 24, 2011


Paul Waddell, 2011Strengths of the PECAS Model System• Built on a half-century of Input-Output modeling of macro-economies dating toLeontieff’s 1960’s model of U.S. economy, and spatial input-output models ofMEPLAN and TRANUS from approximately 1970• The spatial input-output framework has been used over several decades outsidethe U.S., and is beginning to see more use in the U.S., especially at a statewidescale• Integrates interregional trade with and supports modeling of freight due to therelationship between trade and the movement of goods by mode at a time whenlogistics is becoming increasingly important in many cities• The model development process can be started with IMPLAN, commerciallyavailable data that many U.S. regional planners already use• Has been extended in PECAS to include not only origin and destination marketsbut also exchange markets• Provides a static equilibrium framework, but can be run annually• Is marketed as open source software (but not clear that it is downloadable)Tuesday, May 24, 2011


Paul Waddell, 2011Limitations of PECAS Model System• Theory for price adjustment and its integration with I-O model needs development• Spatial extensions to include production, consumption and exchange locations iscomplex and abstract• Data is not readily available for the large number of assumptions to be madeespecially at the the metropolitan spatial scale, much must be synthesized.• Creation of quantities of population, jobs, and commodity weights for freightmovement are all derived by translating dollars flows to quantities• AA module is an aggregate, static equilibrium model - not microsimulation• SD module is a loosely coupled land transition model at a cell or parcel level, lacksdemand side at comparable level of detail• Model estimation/calibration is difficult and to our knowledge no applications havebeen developed without substantial consulting involvement by developers• There is limited experience with fully operational applications. No MPOs had usedPECAS for official Regional Transportation Plan updates in a 2010 survey byMaricopa Association of Governments; only one reported having used it in theirprojection series.Tuesday, May 24, 2011


Questions and DiscussionPECAS Links:http://www.hbaspecto.comPresenters:Paul WaddellDepartment of City and Regional PlanningUniversity of California Berkeleyemail: waddell at berkeley.eduMike ReillyAssociation of Bay Area Governmentsemail: michaelr at abag.ca.govTuesday, May 24, 2011

More magazines by this user
Similar magazines