vermont statewide travel demand model – a preliminary evaluation

52
A report by the University of Vermont Transportation Research Center Vermont Statewide Travel Demand Model – A Preliminary Evaluation Report #10-007 | May 2010

Upload: others

Post on 12-Sep-2021

4 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

A report by the University of Vermont Transportation Research Center

Vermont Statewide Travel Demand Model – A Preliminary Evaluation

Report #10-007 | May 2010

Page 2: Vermont Statewide Travel Demand Model – A Preliminary Evaluation
Page 3: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

VermontAgencyofTransportation

May2009

Preparedby:AndrewWeeks

TransportationResearchCenterFarrellHall210ColchesterAvenueBurlington,VT05405Phone:(802)656‐1312Website:www.uvm.edu/trc

Page 4: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

ii

Acknowledgements

TheProjectTeamwouldliketothanktheVermontAgencyofTransportation,theChittendenCountyMPO,andStephenLaweatResourceSystemsGroup,Inc.fortheirtimeandinsightsduringthepreparationofthisreport.

Disclaimer

Thecontentsofthisreportreflecttheviewsoftheauthors,whoareresponsibleforthefactsandtheaccuracyofthedatapresentedherein.ThecontentsdonotnecessarilyreflecttheofficialvieworpoliciesoftheUniversityofVermontTransportationResearchCenter.Thisreportdoesnotconstituteastandard,specification,orregulation.

Page 5: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

iii

TableofContents

ListofTables.....................................................................................................................................iv

ListofFigures ...................................................................................................................................iv

ExecutiveSummary.......................................................................................................................... 1

1. Introduction ................................................................................................................................ 3

2. Background ................................................................................................................................. 3

3. ReviewofModelingPractices ..................................................................................................... 4

3.1 StatewideModelingatOtherDepartmentsofTransportation ............................................ 4

3.2 AdvancesinTravelDemandModeling ................................................................................. 6

3.3 OngoingApplicationsofTransportationModeling ............................................................. 7

4. StatewideTravelDemandModelingattheVermontAgencyofTransportation ....................... 8

4.1 ModelStructureandComponents ........................................................................................ 8

4.2 Application,ImprovementandMaintenanceoftheModel ............................................... 12

5. OptionsfortheStatewideModelandModelingProgram ....................................................... 19

5.1 FourOptions ....................................................................................................................... 19

5.2 OptionsandModelStrengthsandWeaknesses ................................................................. 25

6. TravelDemandModelingSoftwarePackages........................................................................... 25

6.1 Cube/Voyager ..................................................................................................................... 26

6.2 TransCAD ............................................................................................................................ 27

6.3 VISUM ................................................................................................................................. 29

6.4 PractitionerExperiences ..................................................................................................... 29

7. PreliminarySummaryandRecommendations.......................................................................... 30

References ..................................................................................................................................... 32

AppendixA ..................................................................................................................................... 35

Page 6: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

iv

ListofTables

TABLE1:STATUSOFSTATEWIDEMODELINGCAPABILITY‐SPRING2005....................................... 5TABLE2:OPTIONSSUMMARY........................................................................................................ 19

TABLE3:MODELSTRENGTHS&WEAKNESSWITHRELEVANTOPTIONS ....................................... 25TABLE4:SUMMARYOFCUBE,TRANSCAD,ANDVISUMFEATURESANDFUNCTIONALITY ........... 28

ListofFigures

FIGURE1:DAILYPERSONTRIPRATESPERHOUSEHOLDBYPURPOSE............................................. 9FIGURE2:TRIPLENGTHFREQUENCYDISTRIBUTION ..................................................................... 10FIGURE3:MODELTRAFFICANALYSISZONES................................................................................. 11

Page 7: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

1

ExecutiveSummary

TheUniversityofVermontTransportationResearchCenter,workingwiththeVermontAgencyofTransportation,isconductinganevaluationoftheexistingstatewidetraveldemandmodeland

toidentifypotentialimprovementsandrefinementstothemodelbasedonplanningpracticesandneeds.Themodel’scurrentsoftwareplatform,Cube/Voyager,isalsoexplored,withacomparisontotwootherwidely‐usedmodelingsoftwarepackages.Theevaluationalsoincludes

aliteraturereviewofstatewidetraveldemandmodelingpracticesinotherstates,includinggeneralmodelstructure,uses,andmodeloperationandmaintenance,aswellasadiscussionofemergingtrendsintraveldemandmodeling.

TheVermontAgencyofTransportationhasoperatedastatewidemodelingprogramfornearly

fifteenyears,beginninginthemid‐1990susingTRANPLANtransportationplanningsoftwareandGISmethods,tothecurrentmodelinCube/Voyager.Thestatewidemodel,whichcoverstheentiregeographicareaofVermontandthousandsofmilesofroadway,hasservedandcontinues

toserveasavaluabletoolthatprovidestravelactivitydataprojectionsfortransportationplanningstudiesandanalysesthroughoutthestate.Thedataprovidedbythemodelisusedbyplannersanddecisionmakerssothatsoundtransportationpoliciescanbedeveloped.

Withthedevelopmentandmanagementoftransportationsystemscontinuingasavitalroleof

theVermontAgencyofTransportationandotherinvestedstateagencies,includingtheplanningofhighways,transit,freightandevaluationsoftheirimpactonsocietyandtheenvironment,the

statewidemodelingprogramwillhaveongoingutilityandisavaluabletoolforVTrans.Thestatewidemodeldoeshavelimitations,however,infunctionalityanddegreeofutilityandreliabilityforpotentialprojectneeds.Importantly,themodelisnotcurrentlycapableof

modelingtransitorrailnetworks,sinceitsprimarytransportationsystemislimitedtohighwaysandotherroadways.

DependingontheidentifiedplanningandpolicyneedsoftheVermontAgencyofTransportationandotherparticipatingstateagenciesthatwouldrequireuseofthestatewidemodel,themodel

couldbenefitfromthefollowingimprovementsandrefinementsinthecomingyears:

Developmentofatransitnetwork(bus,rail,orboth)andafreightrailnetwork. Developmentofanewmodechoicemoduletocomplementanewtransitnetwork. Algorithmsfordisaggregationofvehiclemilesoftravel(byvehicleclass,averagespeed,

roadwaygrade,etc.)forairqualityanalyses. Feedbackloop(s)betweenmodelingsteps,suchasbetweentrafficassignmentandtrip

distribution.

Updatesofinputdata,inclusionofroadwaygradedata,andfurthercalibrationandvalidationofmodelalgorithms,assumptions,andoutputs.

Developmentofanactualyear2030scenariotocover20‐yearforecastsfromyear2010.

AdegreeofintegrationwiththeChittendenCountyMPO’sregionalmodel.

Page 8: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

2

Ataminimum,thereisenoughcausefortheVermontStatewideModeltobemaintainedandperiodicallyupdatedtokeepitviableandavailableforplanningneedsatVTrans.Theoverallmodelcouldbekeptinitscurrentformwithoutsubstantialimprovements,suchastransitand

railnetworks,butshouldremainavailableforforecastsofroadwaytravelandVMTestimates.Mostimportantly,thesuccessandlongevityofthestatewidemodelgreatlydependsontheclearly‐definedproceduresandgoalsofitsutilitybyVTransandotherVermontagencies.As

statedpreviously,thegoalsofastatewidemodelingprogramshouldbedefinedwellinadvanceofmodel‐specificdetails,suchasdataneeds,modelcomponents,computersoftware,andbudgetallocation.

Page 9: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

3

1. Introduction

TheUniversityofVermontTransportationResearchCenter,workingwiththeVermontAgencyofTransportation(VTrans),isconductinganevaluationoftheVermontstatewidetraveldemand

model’sutility,components,andcurrentsoftwareplatform.Thisreportrepresentsthefirstworkingdraftoftheevaluationanditspreliminaryfindings.Thegoalsoftheevaluationareto:

IdentifythecurrentandpotentialusesforthemodelbasedonVTransplanningpracticesandneeds.

Recommendupdatestothemodeltomeetfutureimplementation. Compareitsexistingsoftwareplatformwithtwootherwidely‐usedsoftwarepackages

currentlyavailable.

Alsoprovidedinthisreportarealiteraturereviewofstatewidetraveldemandmodeling

practicesinotherstates,includinggeneralmodelstructure,operation,andmaintenance,andadiscussionofemergingtrendsintraveldemandmodeling.

2. Background

TheVermontAgencyofTransportationisresponsiblefortheplanning,construction,

maintenance,andpolicygoverningthetransportationsystemintheStateofVermont.Thetransportationsystemincludespassengertravelbymodessuchasmotorvehicles,transit,air,andnon‐motorizedtravel,aswellasthemobilityoffreight.ThePolicyandPlanningDivisionof

VTrans,specifically,developsandoverseesthecomprehensivetransportationplanningstrategiesandpolicyframeworkinthestate,andworkswithlocalmunicipalities,theelevenregionalplanningcommissions(RPC),theChittendenCountyMetropolitanPlanningOrganization

(CCMPO),andotherstateagenciestoensurethatthestate’stransportationsystemispartofaresponsibleandefficientsocial,economical,andecologicaldevelopmentpolicyplan.

PartoftheVTransplanningresponsibilitiesincludestheforecastingoffuturetraveldemandandtravelpatternsonthetransportationsystem.Thepurposeoftheforecastsistohelpguidepolicy

andfundingdecisions,suchastheconstructionofanewroadway,thecreationofanewbustransitline,oralandusedevelopmentschemetomakemostefficientuseofthein‐placetransportationinfrastructure.Inordertoperformtheforecasts,andalsotogainbetter

understandingofthecurrenttransportationsystem,VTransreliesontransportationplanningtoolsforanumberofitsplanningtasks.Thesetransportationtoolsrangefromverysmall‐scale,detailedmodels,idealfortaskssuchasroadwaycapacityanalysesandmanagementoftraffic

operations,toalarge‐scale,statewideplanningmodel,idealfordeterminingthetraveldemandloadsontheentiremulti‐modaltransportationsystem.

Page 10: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

4

Foritsplanningresponsibilities,VTransreliesonthestatewidemodeltoevaluatetheeffectsoftransportationprojectsandtoguidefuturedevelopmentpoliciesforsectionsofthetransportationsystem.Thestatewidemodel–identifiedasatraveldemandmodel–coversthe

entireStateofVermontandincorporatesdemographicandeconomicdataofVermont’shouseholds,employmentestimates,andthecharacteristicsofthetransportationsystemitself,includingthetravelmodeopportunitiesandthestate’sroadwaynetworkintoastructuredfour‐

steptransportationplanningmodel.Essentially,thepassengertravelmodelestimatesdemandforuseofthetransportationsystembasedoncharacteristicsofthetransportationnetworkanditspotentialusers,andreportsmeasurabledegreesofuse,includingroadwayvolumes,milesof

travelonroadways,andtransitridership.Thestatewidemodelalsoincludesafreightdemandmodelthatestimatestruckusageofthestate’sroadwaynetworkforthemovementofgoods.

3. ReviewofModelingPractices

3.1 StatewideModelingatOtherDepartmentsofTransportation

Currently,approximatelyone‐halfofthestatesintheUnitedStateshavefunctionalstatewide

models(1).Thestructure,utility,andcostsofthenumerousstatewidemodelsvaryandgreatlydependontheneedsoftheirhostdepartmentsoftransportation.PublisheddocumentssuchasNCHRPSynthesis358:StatewideTravelForecastingModels(1)andStatewideTravelDemand

Modeling:APeerExchange(2)provideinformationaboutcurrentmodelingpracticesinstatesthroughoutthecountrybasedonsurveyresponsesfromstatedepartmentsoftransportation,includingVermont.Table1,takenfromNCHRPSynthesis358,page14,presentsasummaryof

statewidemodelingactivityasof2005.

AscitedinNCHRPSynthesis358,statespredominantlyusetraditionalfour‐stepproceduresatthecoreoftheirpassengertraveldemandmodels,similartourbanorregionalmodels.Tripgenerationisgenerallyperformedbycross‐classificationtechniques,tripratesandlinear

equations,ortour‐basedmultinomiallogitexpressions(1).Themostcommonlyuseddemographicandtraveldatasources,suchastheUSCensus,theCensusTransportationPlanningPackage(CTPP),theNationalHouseholdTravelSurvey(NHTS),andlocalMPOinformation

providereadilyavailabledataforusebystatesintheirmodels,andareusedsolelyorsupplementedbystate‐collecteddata.Furthermore,inmoststateswithfunctionalmodels,gravitymodelapplicationisusedfortripdistribution,whiletherearesomeinstancesofgrowth

factormethods.Modechoiceistypicallyaccomplishedbylogitexpressionsorfixedpercentageshares,andtrafficassignmentisperformedusingestimationsofequilibriumconditions(1).

Moreover,NCHRPSynthesis358findsthatstatedepartmentsoftransportationpredominantlyutilizetheirmodelsforcorridorplanning,systemplanning,bypassstudies,regionalplanning

whilecollaboratingwithanMPO,andproject‐levelforecasts,suchasforenvironmentalimpactpurposes.Approximatelyone‐thirdofstateswithfunctionalmodelsapplytheirmodelsforairqualityanalyses,freightplanning,andeconomicdevelopmentandlong‐terminvestmentstudies.

Page 11: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

5

Thereportalsoindicatesthatstateswithlongerhistoriesofmodelingandgreaterconfidenceinthevalidityofthemodelingoutputstendtomorewidelyusetheirmodelsforabroadrangeoffunctions.

TABLE1:STATUSOFSTATEWIDEMODELINGCAPABILITY‐SPRING2005

StateModel

ConditionCost

DevelopmentTime(years)

Comments

Alabama None Alaska None Arizona None Arkansas None California Operational $200,000 2.4 Colorado None Connecticut Operational $400,000 1 Delaware Operational DistrictofColumbia MPOmodel Florida Operational $1,500,000 4 Georgia Operational $65,000 1 Hawaii None IndividualislandmodelsIdaho Dormant Illinois Dormant Indiana Operational $1,500,000 3 7moreyearsforvariousupgradesIowa Developing $300,000 2 Kansas Developing HasadormantfreightcomponentKentucky Operational $370,000 2 NewmodelunderdevelopmentLouisiana Operational $500,000 CostincludessomeapplicationsMaine Operational $500,000 5 BeingrevisedMaryland None Massachusetts Revising $800,000 Michigan Operational $1,000,000 2 Minnesota Partial Mississippi Developing Missouri Operational $500,000 RevisioncompletionsoonMontana Operational FreightonlyNebraska Dormant BaseyearmodelNevada None NewHampshire Revising $2,000,000 NewJersey Operational $500,000 FreightonlyNewMexico None NewYork None County‐levelODassignmentNorthCarolina None NorthDakota None Ohio Operational $6,000,000 8 Beingrevised;$3,500,000fordataOklahoma None Oregon Operational BeingrevisedPennsylvania Developing RhodeIsland MPOmodel SouthCarolina Operational $25,000 0.5 SouthDakota None FeasibilitystudybeingconductedTennessee Developing BasedonODtableestimationTexas Operational $1,700,000 4 Utah None Vermont Operational $730,000 2.5 Virginia Operational $1,500,000 3 Washington None WestVirginia None Wisconsin Revising $850,000 2.5 Wyoming None Notes:MPOismetropolitanplanningorganization;ODisorigin‐destination.Modifiedfromsource:NCHRPSynthesis358:StatewideTravelForecastingModels(2006),Table1,p.14.[1]Datamovedfrom“Colorado”intheoriginallypublishedtableto“Connecticut”inthisversion.

Page 12: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

6

Accordingly,theNCHRPSynthesis358reportsthatthemostcommonmeasuresofeffectiveness(MOE)usedfrommodelsasindicatorsofsystemperformanceincludevehiclemilesoftravel(VMT),vehiclehoursoftravel(VHT),volumetocapacityratios(v/c),levelsofcongestion,and

trafficgrowthrates.Additionaloutputsmaybeusedformorespecializedmodeluses,butVMTandVHTarestandard.SomemodelsproduceMOEsforcertainperiodsoftheday,namelypeakhours,whileothersreportdailytotals.

Themaintenanceandresourceneedsfortherangeofstatewidemodelsarealsovaried.While

manystatedepartmentsoftransportationrelyonprofessionalconsultingfirmstodeveloptheirmodels,manyperformroutinemaintenancethemselves(1).Staffallocationsformodelingprogramstypicallyrangefromoneormorepersonnelwithpart‐timeresponsibilities–some

supplementedbyvaryinglevelsofconsultantsupport–uptomultiplefull‐timepersonnel.Moreover,maintenanceisperformedcontinuouslyoratfrequentintervals(onetotwoyears),andstatestypicallyperformmodelupdatesonregularcycles–somecoincidingwithreissuingof

transportationplans(typicallyeveryfiveyears),andothersusingcyclesaslongastenyears.

SurveyfeedbackreportedinStatewideTravelDemandModeling:APeerExchangeprovidesinsightintosomereasonsformodelingprogramfailures.AccordingtoRickDonnellyofParsonsBrinckerhoff,whosesurveyresponsewasquotedintheStatewideTravelDemandModeling:A

PeerExchangefortheOregonstatewidemodel,“vagueorpoorlydefinedgoalsandobjectives,”“higherthanexpectedmaintenanceandapplicationcosts,”and“lackofmanagementsupport”aresomeofthecausesformodelfailureandabandonment(p.74).Mostimportantly,acommon

andreasonabletrendinstatewidemodelingprogramsisthatstatesmustbalancemodelqualitywithtimeandbudgetconstraints(1).

Inordertoensurethatastatewidemodelingprogramiseffectiveandsustainable,bothinterms

oflaborandbudgetallocation,itsobjectivesmustbeclearlydefinedanditsutilitymustbetransparent.Asstatedin“ACriticalReviewofStatewideTravelForecastingPractice”(3),thegoalsofastatewidemodelingprogramshouldbedefinedwellinadvanceofmodel‐specific

details,suchasdataneeds,modelcomponents,computersoftware,andbudgetallocation.Theplanningneedsofthestatedepartmentoftransportationshouldsteerthemodelingprogram.

3.2 AdvancesinTravelDemandModeling

Recentresearchinthesubjectoftraveldemandforecastinghasseenashiftfromtraditionaltrip‐basedmodelingtechniquesofthefour‐stepparadigmtoactivity‐basedforecastingmodels.Thishasbeeningeneralresponsetothelimitationsofthefour‐stepprocess,theneedformodelsto

bemoresensitiveandresponsivetotheeffectsofsystemconditionsandpoliciesonmobilityandlifestylechoices,andthedesiretodisaggregatetraveldemandforecastingtothelevelofmicro‐simulationmodelingtechniques,replacingaggregatezonalcalculations(4,5,6,7,8).Since

activity‐basedmodelingconsidersthedecisionprocessesthatdictatetravelofindividualentities

Page 13: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

7

or“agents,”potentiallyonacontinuoustemporalbasis,itsmethodologyadaptsitselfbettertothedynamicnatureandcomputationalcharacteristicsofmicro‐simulation.Therecould,therefore,beashiftawayfromindividual“trips”astravelunitsto“tours”withmoreadvanced

activity‐basedmodelsimplementedatregionalandstatewidelevels.

Inadditiontoservingthestandardforecastingneedsoftransportationplanning,thenewactivity‐basedmodelshavealsoshowntobevaluabletoolsforevaluatingmorespecifictransportationsystemmanagementpolicies,including:traveldemandmanagementandpeak

spreadingeffects,high‐occupancyvehicle/facilitydemand,tollandcongestionpricing,andtransitfareandparkingpricingpolicies(5,7).Theycanalsobeappliedtoplanningissuessuchastransportationequityandchangesindemographics,suchaspopulationaging,household

compositions,andlaborforcecharacteristics(5,7).

Thedatarequirementsfordevelopment,computationaldemands,validationprocesses,variabilityofresults,andcostsversusbenefitsofactivity‐basedmodeling,thoughimplementedandevaluatedonlimiteduptoregionalscales,haveyethadwidespreadapplicationona

statewidescale.Somestateswithneworrecentlyrevisedstatewidemodelshaveimplementedvaryingdegreesofactivity‐basedtraveldemandprocedures,includingOhio,Oregon,NewHampshire,Michigan,andLouisiana,withOhio’s(stillindevelopment)beingthemostadvanced.

VariousMPOsnationally,suchastheMid‐OhioRegionalPlanningCommission(MORPC),theNewYorkMetropolitanTransportationCouncil(NYMTC),andtheSanFranciscoCountyTransportationAuthority(SFCTA),alsohaveimplementedactivity‐basedproceduresintheirregionalmodels.

However,manytraveldemandpractitionersarestillskepticaloftheaccuracyandlarge‐scaleapplicabilityofthenewmodelingtechniques,aswellastheirdataandresourceneeds.Dependingonthedesiredutilityandnecessarydetailforastatewidetraveldemandmodelat

thistime,aswellasimplementationconstraints,activity‐basedmodelingmaynotyethaveafeasibleapplicationforallstatescurrentlywithtraveldemandmodels.

3.3 OngoingApplicationsofTransportationModeling

AstudypanelfortheFloridaDepartmentofTransportation,aspartofitsevaluationoftransportationmodels,providedaneffectivelistof“currentandemergingissuesin

transportationplanning”thatsummarizesabroadspectrumofpotentialapplicationsfortraveldemandmodelsandothertransportationmodels(9).Thelistincludes:

Capacitydeficienciesandcongestion Interactionbetweentransportationandlanduse

Economicdevelopmentimpactsoftransportation Freightmobilitystrategies Airquality

Systempreservation,maintenance,andoperations Safety Securityandemergencyevacuations

Page 14: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

8

Equity Resourceallocationandprojectselection

TheseplanningchallengeshaveuniversalcontextsandarealsoapplicabletoVermont’stransportationsystems,developedspace,andenvironment.Therefore,itisreasonabletobe

cognizantofthesetopicswhilediscussingthecurrentandfuturefunctionalityandutilityofVermont’sstatewidemodel.

4. StatewideTravelDemandModelingattheVermontAgencyofTransportation

TheVermontAgencyofTransportationcurrentlyhasafunctionalstatewidetraveldemandforecastingmodel.In2007,themodelwastranslatedfromtheoriginalTRANPLANsysteminto

thenewCube/Voyager(Citilabs,Inc.)softwareplatformbyVHB,Inc.withupdatestotheroadwaynetwork,freightmodel,andmodelcalibrationproceduresandresults(fordocumentationaboutthisprocessrefertoVermontStatewideTravelDemandModel

Improvements:UpdatedPassengerandTruckModelsinCube/Voyager,2007).Importantly,duringthisprocessemploymentdatafortripgeneration,truckpercentagesbyregionalplanningcommission(RPC)fornonhome‐basedtrips,andthefrictionfactorfunctionsfortripdistribution

wereallupdated(10).Althoughthemodelhasbeenappliedtoanumberofproject‐leveltasks,suchastheCirc‐WillistonEnvironmentalImpactStatementandcurrentlytheWesternCorridorTransportationManagementPlan,thepotentialexistsforfurtherimprovementsandbroader

utilitytoaidVTransprojectworkandplanningintheVermont(seeAppendixA:A‐1–A‐3,A‐5).

4.1 ModelStructureandComponents

Thestatewidemodelisstructuredasatypicalfour‐steptransportationplanningmodel,

comprisedof:1)tripgeneration,2)tripdistribution,3)modechoice,and4)trafficassignment(foradescriptionofthesefourstepsinageneraltransportationplanningcontextreferto11and

12).Briefly,inthetripgenerationstep,themodeltakesinputsfromdemographicsourcesandcombinedwithtriprates,estimatesthenumberofpersontripsgeneratedperday,categorizedintosixgeneralizedpurposes:home‐basedwork,home‐basedshopping,home‐basedschool,

home‐basedother,nonhome‐based,andtruck(asapercentageofnonhome‐based).Forexample,Figure1showsthetripproductionratesforthefourhome‐basedpurposesbasedonthecross‐classificationofpersonsperhouseholdandautosperhousehold;nonhome‐basedtrip

productionsarebasedonalinearapproximationinsteadofcross‐classification,similartotripattractions.Inthetripdistributionstep,theflowsoftripsbetweeneachorigin/destinationpairofTAZareestimatedbasedonthezone‐to‐zonetraveltimes(impedance).Figure2showsthe

triplength(time)frequenciesbypurpose,whichareoneresultofthetripdistributionprocess.Themodechoicestepsplitsthetotalnumberofpersontripsintothoseusingtheroadwaynetworkandthoseusingtransit;inturn,theroadwaypersontripsareconvertedtovehicletrips

usingvehicleoccupancyfactorsbytrippurpose.Lastly,thevehicletripsareassignedtothe

Page 15: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

9

roadwaynetwork,producingtrafficvolumeandVMTestimatesoneachroadway.Becausethestatewidemodelcurrentlydoesnothaveatransitnetwork,thepersontransittripsestimatedinthemodechoicesteparenotassigned.

FIGURE1:DAILYPERSONTRIPRATESPERHOUSEHOLDBYPURPOSE

0

1

2

3

4

5

6

Home‐basedWork Home‐basedOther Home‐basedSchool Home‐basedShopping

Page 16: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

10

FIGURE2:TRIPLENGTHFREQUENCYDISTRIBUTION

Themodelhasa2000baseyearwithaforecastyearof2020andaten‐yearinterpolationto2030.Itcontains628internaland70externaltrafficanalysiszones(TAZ)asshowninFigure3.TheexternalTAZrepresenttravelto/fromandbetweenregionsbeyondVermont,including

neighboringstates(NewYork,NewHampshire,andMassachusetts),therestofNewEnglandandthenortheast,andCanada.Generally,theinternalTAZcoincidewithVermontmunicipalityboundaries,exceptinmoredenselypopulatedareaswheretheTAZaremoredisaggregate.The

TAZalsoaggregateto2000USCensusblockgroupboundaries.

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%

7.0%

8.0%

9.0%

10.0%

2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60

TripLength(minutes)

Home‐basedWork Home‐basedOther Home‐basedSchool

Home‐basedShopping NonHome‐based

Page 17: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

� � wn t d� � d� d� l e� � � � w� v� o� � � n � t � � � � � o� C� � � � w� oen et � wM� � v� oT� de t �

� � � � � � � � � � S wd� ‐ � 37R77f �

/ / �

� � � � � � Fy� � � � � � � � � � � � � � � � � � � � � � � � � �

� y � wd� � p � t � n kp dw0� � dnWWkW� dM� � r r wdoA � k 9t � /Ucmmm� 9n0Wc� n� 9s � na� nk wn� 9� � n� � ok wn� 9� � � � �� nkwd� � � dnn � kdwWc� � T� � � nkd� Mdswk n� � 9� WWM� � kdn� � � k adw WC� �

,/V� wsw� 9� nk wWk� k �

,j V� wsw� 9� r wn� r � 9� � wk w� 9� ,LV� wsw� 9� A ndw� � wk w� 9� ,GV� wsw� 9� A � 2dw� � d99 � kdw� �

,I V� wsw� 9� A ndw� � d99 � kdw�

Page 18: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

12

(9)rurallocal (11)urbaninterstate (12)urbanprincipalarterial/freeway

(14)urbanotherprincipalarterial (16)urbanminorarterial (17)urbancollector

(19)urbanlocal (20)internalTAZcentroidconnector (21)externalTAZcentroidconnector

4.2 Application,ImprovementandMaintenanceoftheModel

TheVTransstatewidemodelhashadorwillhaveutilityfortransportationstudiesthroughoutthe

state,includingtheCirc‐WillistonDraftEnvironmentalImpactStatement(DEIS)(13),theMorrisvilleBypass,theBenningtonBypass(14),theWesternCorridorTransportationManagementPlan(15),andastateemployeecommuterstudy.Themodelwouldalsohave

applicabilityfortheHighwaySystemPolicyPlan(16),PublicTransportationPolicyPlan(17),RailSystem&PolicyPlan(18),theAssetManagementVisionandWorkPlan(19),theVermontLongRangeTransportationBusinessPlan(LRTBP)(20,21)–thestate’soverallplanforthemulti‐modal

transportationsystem–andtheFinalReportandRecommendationsoftheGovernor’sCommissiononClimateChange(GCCC)(22).

Theusefulnessofthemodelforproject‐levelanalyses,planningscenarioevaluations,andstorageandpresentationoftransportationsystemdatareinforcetheneedforcontinued

maintenanceandapplicationofthetool,aswellasoutreachandtrainingtoinformotherVTransdepartmentsandotherstateagenciesofthemodel’sutility.Thefollowingsectionshighlightanticipatedmodelingneedsbasedonfutureplanninggoalsandtheimprovementslikelyrequired

orrecommendedforthestatewidemodelinordertomeetthosespecificneeds,especiallythosethatarenotcurrentlywithinitscapabilities.

Dependingonthedeterminationofthelong‐rangemodelingobjectivesofVTransandthestatewidemodel’sfutureutility,itispossiblethatthestatewidemodelwouldneedadditional

refinementandupdates.Ifsuchrefinementsandupdatesweretobeimplemented,itisexpectedthattheirpurposewouldprimarilybetomeetsomeoftheplanninggoalsidentifiedintheLRTPBandGCCC,aswellasothermodelingeffortsatVTrans,includingtransitplanningandairquality

analyses.Assuch,someofthecontinuedandfuturemodelinggoalsdiscussedcoulddeterminetheneedforanumberofpotentialstatewidemodelimprovements.

4.2.1 SystemPolicyandEnvironmentalPolicyPlanning

Anumberofapplicationsofthestatewidemodelarelikelybasedonanticipatedmodelingneedsdescribedinstateandagencydocuments,suchastheLRTBPandtheindividualsystempolicy

plans.Thesedocumentsdonotexplicitlyidentifytheapplicationofthestatewidemodeltomeet

Page 19: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

13

theirrecommendationsbutfromapracticalviewpoint,themodelisclearlyatooltoaidcompletionofsomeofthepolicycomponents.TheLRTBPidentifiestransportationopportunities,specificallytheintegrationoflanduseandtransportationplanningandtheevolvingfocuson

corridormanagementplanning,asmethodsforthestatetomoreeffectivelymeettransportationsystemchallenges.ThefinalreportandrecommendationsoftheGCCCcitestheneedforstrengthenedstate‐levelplanninganddecision‐makingtosupportmanagedgrowthaswellas

transportationplanningtofacilitatealternativetravelmodes,transitpriorityandtherehabilitationandmaintenanceofexistinginfrastructure.

BusandRailTransitSystemPlanning

Inadditiontotheplanninggoalsoutlinedinofficialagencydocuments,discussionswithVTranspersonnelunderscorethepotentialutilityofthestatewidemodeltosupportvariedtasksatthe

agency(seeAppendixA:A‐1–A‐3,A‐5).AtthePolicyandPlanningDivision,thereisarecognizedneedforimprovedmulti‐modalplanning,asrecommendedintheLRTBPandmorespecificallytheHighwayandPublicTransportationSystemPolicyPlans,whichwouldlikelyincludethe

creationofatransitnetworkinthestatewidemodel.Afullydevelopedtransitnetworkwouldallowplannerstoestimateridershiploadsalongexistingandproposedtransitroutessuchasintercitybustransit,toevaluateheadwaysandstops,andtoidentifypotentialroutecapacity

improvementsandextensions.Importantly,anyimprovementinthecapabilitiesanddetailofthemodel’stransitcomponentwouldrequireredevelopmentofthemodechoicemodule,sincethemodelcurrentlyapportionsafixedpercentageasatransitshare(10).Anewmodechoice

modulewouldlikelybealogitmodel,whichpredictstheselectionprobabilityofalternativemodesbycomparingthecosts(monetarycost,time,convenience,etc.)ofthosemodes(11,12).

Thedevelopmentofatransitnetworkinthemodelwouldbeanintensivebutprobablyworthwhiletaskconsideringtheimportanceoftransitmodelingrecognizedinlongrange

planningdocuments.Additionofatransitnetworkwouldrepresentasignificantimprovementtothemodel’sfunctionalityandutility.Thistransitnetworkwouldcertainlyincludebus,which

wouldusethemodel’sroadwaynetwork,andpossiblyrail,whichwouldrequirethedevelopmentofanewnetworkstructure.

RailFreightSystemPlanning

Developmentofarailnetworkaspartofthefreightmodelwouldimprovethefreightmodelcomponentaswellastheoverallmodel’sutility.Agrowingneedordesiretoshiftmorefreight

fromtrucktorailcouldstillprecipitateaneedforarailcomponentinthestatewidemodel.TheVermontStateRail&PolicyPlanidentifiesagoaltoprovidecompetitivefreightandpassengerservice,eventhoughthatreportalsorecognizestheuncertaintyofthefutureofVermont’s

intercitypassengerrailservice(18).However,withapassengerrailcomponent,ifdeemedworthwhile,thestatewidemodelcouldbeusedtoevaluateplanstoexpandandpromotepassengerrailservicebothinVermontandregionally.

Page 20: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

14

Thegeneralrailnetworktopology–linksandnodes–oftherailpassengertransitcomponentcouldbetransferredfromanewlydevelopedrailfreightcomponent,ifthetransitimprovementsinthemodelwereundertakenfirst.Therailnetworkstructurecouldthenbespecifically

configuredforuseinthefreightmodel.

HighwaySystemPlanning

TheHighwaySystemPolicyPlanindicatesthatforsystemmanagementandpreservation“itisdesirabletodefinedifferentsectionsofthehighwaysystembasedonfunctionalityandoveralllevelofimportanceforwhichdifferentperformancestandardsandinvestmentpoliciesare

developed”(p.ES‐5).Thisideaofsub‐networksappliestoperformancemeasures,includingpreservation,safety,mobility,andenvironment/qualityoflifeasoutlinedintheHighwaySystemPolicyPlan,andtotheprioritizationofassets.AsVermontdevelopswithgrowingpopulation

centers,thedesignationsofexistingandhithertounbuiltroadwaysaspartsofeitherthe“primary”or“off‐primary”sub‐networksmayneedtoberevised.Forecastsandanalyseswiththestatewidemodelcouldaidinappropriatelyidentifyingtheevolvinglevelofimportanceof

roadwaysinthenetworkandevaluatecertainperformancemeasures,suchasVMT,traveltimes,averagespeeds,flows,andvolume‐to‐capacityratiosforroadwaylinks.Asanextensionofthesub‐networkdefinitionprocess,thestatewidemodelcouldservetoevaluateanddeterminethe

criticallinksinthesystem,andthepossibleimpactsduetotheclosure(e.g.bridgemaintenance,naturaldisasters,emergencyevacuations,etc.)ofacriticallink.Thiswouldprovidedatafortheplanningofmanagementandresponsestrategies.

ThestatewidemodelcouldalsoservetheTrafficOperationsdepartmentresponsibleforthe

management,analysis,andreportingofVTranstrafficcountsthroughoutthestate,byaidingtrafficgrowthprojectionsandhelpingtoidentifythepotentialeffectsofnetworkdisruptionsandresultingdetours/re‐routing(seeAppendixA:A‐1).Howeverfordetailedanalysesof

roadwaysectionsorcorridorsusing,forexample,micro‐simulationmodelswiththestatewidemodel’sprojectedvolumesasinputs,itwouldfirstbeadvisabletoevaluatetheaccuracyand

applicabilityofthestatewidemodel’strafficvolumesfordirectuseinothermodelinganalysesbeforesuchapplicationswereundertaken.

EnvironmentalPlanning

Theimportanceofenvironmentalpoliciesrelatingtoclimatechangeandairqualityconditionscoulddrivetheneedforfurtherstatewidetraveldemandforecastmodeling.AlthoughVermont

iscurrentlyinattainment,airqualityconditionscouldworsenormorestringentstandardscouldresultinnon‐attainmentforcertaincriteriapollutantssuchasozoneorparticulatematter(21).Moreover,potentialfuturelimitsongreenhousegases,namelycarbondioxide,aswellasVMT

performancemeasureswouldrequireVMTforecastsandpollutantinventoryestimatesforVermont’stransportationsystem.Thestatewidemodelisavaluabletoolforsuchenvironmentalanalyses,primarilybecauseitcanfeednecessaryinputdata–VMTbyfunctionalroadwayclass

andaveragespeeds–tomobileemissionsmodels(currentlytheEPA’sMOBILEmodelandthe

Page 21: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

15

newerMOVESmodel).ItisreasonabletoexpectthatairqualitymodelingandtheresultingenvironmentalpolicieswillcontinuetogainimportanceforVermontandnationally,andtheneedforincreasinglydetailedandaccuratetraveldatafromthestatewidemodelwouldgrow

concurrently.

Therearemethodstoimprovethestatewidemodel’soutputstobetterservemobileemissionsmodeling,specificallytheMOBILEmodelwhichiscurrentlyusedinVermont.AccuratedisaggregationofVMTdatabyvehicleclasscouldimproveemissionsestimates,sincedifferent

typesofon‐roadvehicles,suchaslight‐dutyversusheavyduty,havedifferentoperatingcharacteristicsandemissionsprofiles.However,atypicalfour‐steptraveldemandmodellikethestatewidemodeldoesnotinherentlyconsidervehicletypesinthismanner.Thestatewidemodel

doescurrentlydefineaportionofnonhome‐basedtripsastrucktripsbasedonapercentage(determinedbyvehicleclassificationcounts)foreachRPCinVermont,butthisdoesnottranslatetospecificvehicle‐typeVMTassignedtothehighwaysystemasamodeloutput.Therefore,it

wouldbeworthwhiletodevelopanewmeanstoaccuratelydisaggregatehighwayVMTbyvehicleclass,atleastintogeneralweightclasses(e.g.light‐dutyvehicle,light‐dutytruck,andheavy‐dutytruck)withinthestatewidemodel.Thiscouldinvolveusing24‐hourclassification

countsandconductingroad‐sidesurveys.

OtherVMTbreak‐downsthatcouldbeexploredtoimproveemissionsmodelingwouldincludeVMTbytimeofday(hourly)andVMTbyaveragespeed.Importantly,themeritofdevelopingthesedistributionsspecificallyforVermontinsteadofusingthedefaultvalues(national

averages)suppliedwiththeMOBILEmodelwouldneedtobeinvestigated.Thedefaultvaluesmaybe“goodenough,”anditwouldnotbeworththetimeandcosttodevelopVermont‐specificdistributionsthatwouldnotservetoimproveemissionsestimates,orworse,decreasetheir

accuracy.

GreenhouseGasEmissionsInventoriesandStandards

Thecurrentandfutureutilityofthestatewidemodelshouldbeconsideredinthecontextofpotentialairqualitystandardschangesandpotentialgreenhousegasstandardslegislationfor

Vermontandonafederallevel.InApril2009,theEnvironmentalProtectionAgencyreleaseditsfindingthatgreenhousegasescontributetoairpollutionthatposesathreattopublichealthandwelfare,inresponsetoa2007rulingbytheSupremeCourtoftheUnitedStatesthatrequiredthe

EPAtoreviewthepotentialimpactsofemissionsofsixgreenhousegases(24).

Concurrently,theStateofCaliforniaistakingaleadingreenhousegaslegislation,includingnewcapstandardsforlight‐dutyvehiclesandrulesforGHGreporting/inventories(25,26).Vermont,inadditiontoapproximatelyfifteenotherstates(27),haspasseditsownlegislationtoadopt

California’sGHGstandards,expectingthattheU.S.EPAwillgrantCalifornia’swaivertotheCleanAirActfortheproposedstandards.California’swaiverrequestwasfirstrejectedbytheEPAduringtheBushAdministrationinlate2007,buttheEPAhasrecentlybeendirectedbythe

ObamaAdministration,inearly2009,toreviewandreconsiderthatdecision.

Page 22: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

16

VermontpasseditsownruleinNovember2005,makingitsGHGemissionstandardsforlowemissionvehiclestobeidenticaltothoseofCalifornia.Intherule,titledAirPollutionControlRegulations,SubchapterXI,LowEmissionVehicles–RegulationstoControlGreenhouseGas

EmissionsfromMotorVehicles,GHGstandardsforsmallandlargelight‐dutyvehicleswouldbegraduallyphasedinbetweenmodel‐years2009and2016(28).However,VermontcannotactuntiltheCaliforniawaiverisgranted.

AsVermontadoptsCaliforniastandards,itispossiblethatthestatewillalsoseektofollow

California’sleadonGHGreportingrules.RequirementsforinventoriesofGHGemissionsforthetransportationsectorwouldlikelyemphasizeregularanddetailedreportingofVMTestimatesforGHGcontributionsfromon‐roadsources.Also,anevaluationofGHGemissionimpactswouldbe

desiredforplanningandpolicymakingpurposes(e.g.therecommendationspresentedintheVermontComprehensiveEnergyPlan,2008,(29)andVermontGovernor’sCommissiononClimateChange,2007)(22).

InMay2009,thedevelopmentanddebateofAmericanCleanEnergyandSecurityActof2009

(H.R.2454)hashighlightedtheproposedbill’smulti‐facetedapproachtoreduceconsumptionoffossilfuelsandlimitgreenhousegasemissions.RegardingthelinkbetweenthetransportationsectorandGHGemissions,asummaryofthebilldatedMay16,2009describesthe

TransportationEfficiencysectionofH.R.2454,whichwouldamendtheCleanAirActto“requirestatestoestablishgoalsforgreenhousegasreductionsfromthetransportationsectorandrequiressubmissionoftransportationplanstomeetthosegoalsbyMetropolitanPlanning

Organizationsforareaswithpopulationsexceeding200,000people”(30).

TheVermontstatewidemodelwouldbethemostsuitabletooltoforecastVMTfortheestimationprocessofgreenhousegasemissionsfromon‐roadsourcesinthetransportationsectorandtoevaluateGHG‐relatedimpactsofplanningalternativesbasedontheVMT

projections.CurrentlyinCalifornia,theCaliforniaAirResourcesBoard’s(CARB)travelactivityforecastingapproachfortransportationsectorGHGemissioninventoriesusesVMTestimates

fromtheEMFAC2007modelandregionalplanningorganizations.EMFAC2007estimatestravelactivitybasedonvehiclepopulationandvehicle‐age‐specifictravelaccumulations,whicharematchedtoVMTestimatesfromtheregionalplanningorganizations’transportationmodels(31).

SimilartoCalifornia,VermontshouldmaintainitplanningtoolsandproceduresforVMTforecasting,includingthestatewidemodel.

4.2.2 OtherModelRefinements

Additionalrefinementtaskscouldbeundertakentoimprovethemodel’sfunctionality.Theaccuracyofthemodel’sbaseyearandprojectedvolumesshouldbefurtherassessed.For

regionalairqualitypurposesandplanningprocesses,theneedforreliableestimatesofvehiclemilesoftravel(VMT)wouldbeparamount,andotherusesofthestatewidemodel,suchasanalysisofasectionofitsvolumeoutputsusingmoredetailedmodelingtoolssuchastraffic

simulation,wouldgreatlydependontheaccuracyofitsestimates.Tothisend,thetaskof

Page 23: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

17

assessingandimprovingthereliabilityofthemodel’svolumeprojectionscouldincludefurtherrefinementofthefourplanningsteps,inputdatamodifications,andadditionaliterationsofcalibrationandvalidation.

Four‐StepModelAlgorithms

Furtherrefinementofthemodel’soutputscouldbeaccomplishedbyincorporatingoneormore

feedbackloopswithinthemodel’salgorithmicprocess.Sincetripdistributionestimationsarebasedoninterzonaltraveltimesasarethetravelmodeutilityvalues(traveltimecost,perceivedcost,etc.)inthemodechoicestep,itiscommontofeedthelinktraveltimesestimatedinthe

trafficassignmentstepbackintothetripdistributionandmodechoicesteps.Thisisespeciallyimportantiftrafficassignmentindicatescongestionthatresultsinlinktraveltimessignificantlydifferentfromthoseinitiallyassumed.Theresultingfeedbackwouldproduceaniterativeprocess

thatwouldcontinueuntilanequilibriumconditionhadbeenmet.

InputDataUpdatesandFurtherCalibration&Validation

DuringtheVHB,Inc.updateandmigrationofthestatewidemodeltoCube/Voyagerin2006‐2007,thetraveltimefrictionfactorsfortripdistribution,originallybasedonthe1994Vermonthouseholdsurvey,wereupdatedusingthefunctionsdescribedintheNCHRPReport365(12).

VHB,Inc.alsorevisedthetripattractionregressionequations,butthetripproductionrates,originallybasedonthe1994householdsurvey,werenotchanged.Newdemographic,socio‐economic,andtraveldatafromupcomingupdatesinthe2008‐2009NationalHouseholdTravel

Survey(http://nhts.ornl.gov/nhts2008.shtml)andthe2010USCensuswouldprovideabasistoevaluateandpossiblyrevisecurrenttriprates.Furthermore,withthenewdata,itwouldalsobepossibletoevaluatethefeasibilityandmeritofapplyingregional‐specifictripratesandfriction

factorswithinthemodel,forexamplebycountyorRPC.ItisimportanttonotethatfollowingVHB,Inc.’srecentworkonthemodel,someadditionalcalibrationwasdoneatVTransforTAZin

westernportionsofVermontfortheWesternCorridorTransportationManagementPlanandsubsequenttothat,CambridgeSystematics,Inc.performedfurthercalibrationwhilerunningthemodelforthatstudy.

RoadwayGradeData

Thenetworklinksinthemodelwithfunctionalroadwayclassifications(i.e.allofthelinksexcept

thecentroidconnectors)currentlyhaveadefaultvalueof“1”forthe“TERRAIN”datafield.AccordingtotheVHB,Inc.2007report,adefaultvalueof1,whichindicateslevelterrain,wasassignedtothelinksduetolackofbetter,morespecificterraininformation.Updatingthe

model’snetworklinkswithaccurateterraininformation,orpotentiallymorespecificdatasuchasgrade,couldserveasaninventoryforVermont’sroadsandimprovethemodel’sutilityinfutureon‐roadmobileemissionsandairqualityestimatesusingtheUSEnvironmentalProtection

Agency’sMOVESemissionsmodel,whichconsidersroadwaygrade(aspartofavehicle’soperatingmodeforrunningemissions)initsestimates.ThestatewidemodelcouldprovideVMT

Page 24: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

18

totalsbygrade.However,ifdefiningindividuallinkgradesisdeemedimpractical,categoricalinformation,suchasgeneralterraintype,couldbedefinedforthemodel’snetworklinksandusedtoapproximatemoredetailedgradevaluesasneeded.Terraincategoriescouldinclude

level,rolling,andmountainous,similartothe2000HighwayCapacityManual.

IntegrationwiththeChittendenCountyMPOModel

PossibleintegrationofthestatewidemodelwiththeCCMPO’sregionalmodelofChittendenCountycouldfurtherpromoteanalysesusingthemodelsjointly,tocomplementeachother.TheCCMPOmodeliscontinuallymaintainedandhasrecentlyundergoneanewseriesofupdates

(23).ForcertainprojectsandplanninganalyseswithinChittendenCounty,thestatewidemodelcouldbereliedupontoevaluatetransportationeffectsbeyondtheregionalboundaryoftheCCMPOmodel(seeAppendixA:A‐4).OnesuchexampleistheCirc‐WillistonDEIS.Transferability

ofnetworkandTAZdatafromtheregionalmodeltothestatewidemodel,orviceversawouldhavedesiredutility.Also,byincludingsomeofthesametravelmodesinthestatewidemodelasintheCCMPO–namelyauto,carpool,bus,andpossiblyrail–therecouldbeimproved

transferabilityoftripdatabetweenthetwomodels.Thiswouldalsorelyonrectifyingthetemporalresolutionofthetwomodels,daily(thestatewidemodel)versuspeakhour(theCCMPOmodel).Also,themodelsarecurrentlyindifferentsoftwareplatforms;thestatewide

modelisinCube/VoyagerandtheCCMPOmodelisinTransCAD.Ifthatweretoremainthecasefortheforeseeablefuture,ameansofdatatransfer(import/export)wouldbeneeded,suchaswiththeuseofdatabasefilesorspreadsheets.

4.2.3 Maintenance

Regularmaintenanceofthemodelshouldbedonecontinuously,includingminorcalibrationand

validationprocessesorasneededbyspecificprojectsemployingthemodel.Moresignificantupdatesandrevisionstothemodelincludingupdatestoinputdatasuchasdemographicand

employmentfiguresandroadway,transit,andfreightnetworkcharacteristics,andpossiblythemodel’salgorithmsandlogiccouldbeonascheduledcycleofapproximatelyfivetotenyears.Updateprogramscouldcoincidewiththereleaseofrevisedsystempolicyplansorthe

publicationofupdateddatasources,suchastheNationalHouseholdTravelSurveyandthe2010USCensus.DependingonairqualityattainmentstatusinVermont,airqualityconformityneedscouldpotentiallyrequireaccelerationofthemodel’supdateschedule.

AccordingtotheLRTBP,theIntermodalSurfaceTransportationEfficiencyAct(ISTEA,1991)

stipulatedthatthestatewidetransportationplanmustcoverforecastsforaminimum20‐yearhorizon.ThatstipulationhasprogressedinTEA‐21(1998)andsubsequentlyinSAFETEA‐LU(2005).Currently,thestatewidemodelforecaststo2020and2030,thoughthe2030forecastis

basedonextendingthegrowthfrom2000to2020byanadditionaltenyears(10).Withtheyear2010approaching,moreaccurateforecastsfor2030wouldberequiredandshouldbedevelopedtomaintaincompliancewiththeminimum20‐yearhorizonruleintheLRTBP.

Page 25: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

19

5. OptionsfortheStatewideModelandModelingProgram

VTranshasanumberofoptionsforthefutureofitsstatewidemodel.VTransmustconsiderthefeasibilityofcontinuingitsmodel,givenanexpectedfuturelevelofutilityofthemodel

comparedtothetimeandbudgetresourcestokeepitactiveandcapableofmeetingtheAgency’sneeds.Beyondthatevaluation,therearelong‐termdecisionsconcerningthestateofthemodelandhowitscurrentfunctionalitydoesordoesnotmeetthedesiredplanning

applicationsofthemodel.

Thecurrentandenvisionedlevelandextentofmodelusewouldlargelydictatethespecificsofifandhowthemodeliskeptactive.Ifthemodelatitscurrentleveloffunctionality,givenperiodicupdatestoinputdata,wouldsufficientlyandcost‐effectivelymeettheplanningneedsofVTrans,

thenitwouldbebeneficialforVTranstokeepthemodelactivewithaclearprogramforitsupdateandapplication.Goingfurther,ifthereisadesirewithintheAgencyforincreasedandmorewidespreaduseofthemodel,perhapsfortrafficresearchapplications,roadwayasset

management,performancemeasuredeterminations,andperhapscoordinationofplanningworkwithotherplanningorganizationsinVermontsuchastheCCMPOandRPCs,thentherewouldbecausetodevelopaprogramtorefineandimprovethemodeltomeetthosegoals.

5.1 FourOptions

Thefollowingfouroptionsgivegeneraloverviewsoftheworkitems,relativetimeandbudget

needs,andprosandconsdependingontheAgency’sdecisionforthestatewidemodel.Table2providesasummaryoftheoptions.

5.1.1 OptionI–Discontinuethemodel

ThisoptionwouldhaveVTransshelvethemodelandrelyonconsultant(s)forfulfillingitsforecastingneedsonanas‐neededbasis.Itispossiblethattheconsultant(s)woulddesireaccess

tothemodel–atwhateverconditioninwhichitwouldbeatthetime–fortheprojectmodelingneeds.Inthiscase,VTranswouldhavethemodelinactivebutavailableforitsconsultant(s).Ifneeded,theconsultant(s)wouldthenberesponsibleforanynecessaryupdatestothemodelin

ordertouseitforthespecificanalysis.

TABLE2:OPTIONSSUMMARY

Option WorkItems ResourceRankingI Discontinuationofthemodel

Noprogramforperiodicupdatesofthemodel VTransrelianceonconsultant(s)forforecastmodelingneeds

4th(leasttimeandbudget[1]requirements)

II Programforperiodicupdatesofthemodel VTranssupervisionofmodelupdatesandapplication DevelopmentofprogramforcontinuedmodelusebyVTrans&

DEC/ANR(e.g.traveldemandforecastingforplanningandair

3rd

Page 26: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

20

qualityanalyses)III Programforperiodicupdatesofthemodel

Formulationofdetailedmodelworkplan,includingthetype/extentofimprovementstobemadetothemodelandtimeframe

Implementationofimprovements/refinementstomodelcomponentsandprocedures

VTranssupervisionofmodelupdates,improvementprogram,andapplication

OutreachwithinVTranstoraiseawarenessofmodelfunctionalityandpromotemodeluse

2nd

IV Programforperiodicupdatesofthemodel Formulationofdetailedmodelworkplan,includingthe

type/extentofimprovementstobemadetothemodelandtimeframe

Implementationofimprovements/refinementstomodelcomponentsandprocedures

VTranssupervisionofmodelupdates,improvementprogram,andapplication

OutreachwithinVTranstoraiseawarenessofmodelfunctionalityandpromotemodeluse

CoordinationofmodelingproceduresandplanninggoalswithotherplanningorganizationsinVermont(CCMPOandRPCs)

1st(mosttimeandbudgetrequirements)

Notes:[1]ThisdeterminationgreatlydependsontheamountofworkcontractedbyVTranstoconsultant(s);ifsignificant,thisoptionmayexceedOptionIIinbudgetresourcerequirements.

Asapossibleextensionofthisoption,theconsultant(s)couldprovidethemodelupdatesandan

updatesummarytoVTransattheendoftheproject,suchthatthemodelwouldexperienceselectedupdatesonaproject‐basisinsteadofonadefinedschedule.However,animportantcaveattothisprocesswouldbethatmodelupdateswouldbeonan“adhoc”basis,potentially

performedbymorethanoneconsultant,sotherecouldbeanissuewiththeconsistencyandscopeofthemodelupdates.Therefore,VTranswouldneedtoreviewandverifytheupdates–a

processthatmaynotbeattractiveconsideringthegeneralgoalofthisoption,whichisforVTranstosignificantlyreduceitsnecessarytimeandresourcesforthemodel.

Worktypeandhours

Workandhourswoulddependonaper‐projectbasis,withVTranscontractingandsupervisingaconsultant,orconsultants.

Pros

Forthemostpart,alow‐costoptionsinceVTranswouldnothavetoallocatetimeand

budgetresourcesformaintenanceandupdatingofthemodelingprogram. Onanas‐neededbasis,VTranswouldonlypayforwhatitneedsperproject(i.e.VTrans

wouldnotfundadormantmodel).

Page 27: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

21

Pointofuncertainty:Aprodependingonifthecosttokeepmodelactiveandrunitwouldbegreaterthanthecosttocontractandsuperviseanoutsideconsultantforforecastingwork.

Pointofuncertainty:Amountofprojectworkcontractedtoconsultant(s);alowamountofprojectworkwouldkeepcostsforVTransdown.

Cons

VTranswouldhavetorelyonconsultantsforprojectforecastingneeds. Aresurrectionofthemodelinthefutureforsomemandatoryanalysisneedwould

requiresignificationresources,ofbothtimeandbudget. Pointofuncertainty:Acondependingonifthecosttokeepmodelactiveandrunit

wouldbelessthanthecosttocontractanoutsideconsultantforforecastingwork.Ifthis

werethecase,OptionII,below,maybeamorecost‐effectivechoicesinceahighamountofprojectperformedbyconsultant(s)workcouldbeoverlycostlyforVTrans.

5.1.2 OptionII–Maintainthemodelatitscurrentlevelofcapability,withperiodicupdates

Thisoptionwouldincludeperiodicupdatestothemodel’sinputdata,suchasnewUSCensus,NationalHouseholdTravelSurvey(NHTS),andemploymentdata,yearlyAADTcountsandnewly

builtroadwaysforthehighwaynetwork.Themodelupdateswouldnotincludechangestotheunderlyingmodelproceduresandassumptions.ItisassumedthatVTranswouldstillutilizethemodelforrelevantprojectanalyses,whiletheUVMTRCand/orconsultant(s)wouldbe

contractedtoperformtheperiodicupdatesandnecessarymodelruns.

Thefeasibilityofthisoptionwoulddependontheextenttowhichuseofthemodelinitscurrentformwouldactasreimbursementforthetimeandfundsputintoit.AnupdatedbutdormantmodelwouldbeofnovaluetoVTrans.

Exampleperiodicupdatesofinputdata

ForUSCensusandNHTSdata,updatescouldoccureveryfivetotenyears,dependingondataavailability.USCensusiseverytenyears,butotherdemographicandhousehold

datasources,suchasNHTS,maybeavailableonamorefrequentbasis. Employmentdatacanbeupdatedeveryonetofiveyears,dependingonavailabilityof

newdata.

Newhighwaydata,suchasnewroadwaysand/orcapacityimprovements,canbeupdatedonafive‐yearinterval,orasneededdependingonprojectanalysisneeds.AADTcountdatacouldbeupdatedyearly.

Significantlandusedevelopmentscanbeupdatedonafive‐yearinterval,orasneededdependingonprojectanalysisneeds.

Moredetailed,lessfrequentupdatescouldincludenewhouseholddailypersontrip

rates,whicharecurrentlybasedona1994Vermonthouseholdsurvey.Anewsurveywouldlikelybeneededtodeterminenewthetriprates.

Page 28: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

22

Note:Ingeneral,periodicupdateswouldbeneededatleasteverytenyearstomakethemodel’sbase‐yearcurrentandensureforecastingtoa20‐yearhorizon.Forexample,themodel’sbaseyeariscurrently2000withaforecastto2020.Duringanupdateprocessfor

baseyear2010,theforecastyearwouldhavetobeupdatedto2030. Note:Updatesofinputdataandmodelbaseyearwouldrequirere‐validationofthe

modeland,ifneeded,newcalibrationofmodelparameters.

Worktypeandhours

Workwouldinclude1)modelupdatesand2)modelapplication.Itispossiblethattheentityresponsibleformodelupdateswouldnotperformmodelapplication.Forexample,theUVMTRCcouldperformmodelupdatesregularlywhileaconsultantcouldrunthe

modelonaper‐projectbasis.Oroneentitycouldberesponsibleforalltasks. Workhoursformodelupdatescouldbedeterminedbasedonanupdateschedule.Hours

formodelapplicationwouldbemorevariableandwouldbedeterminedonanas‐needed

basis.Forthepurposesofthispreliminaryevaluation,itisassumedthattheworkhoursrequiredformodeldataupdatesofthisnatureandcorrespondingtasks(checking,re‐validation,andcalibration,ifneeded)wouldbeapproximatelyone‐tenthoftheduration

oftheupdateperiod.Forexample,toupdatedataonaten‐yearinterval,itwouldtakeapproximately12monthsofworktime,everytenyears.Updatesonaone‐yearintervalwouldtakeapproximately1–1½months,everyyear.

Pros

Likelytheleastexpensiveoptiontokeepthemodelactive,up‐to‐date,andavailableforusebyVTrans.

ScheduledupdatescouldbeperformedbytheUVMTRCand/oraconsultant,with

managementbyVTrans.

Cons Themodelwouldnotgainfurtherfunctionalitybeyondthecurrentlevel.

Themodelwouldnotberelevantfordetailedtransit/multi‐modalplanningefforts.

5.1.3 OptionIII–Improvethemodel’scapabilityforVTrans(includingDEC/ANR)analyticalobjectives

ThisoptionwouldincludenecessaryupdatestothemodelasoutlinedintheOptionII,andwouldalsoincludefurtherimprovementsandrefinementstothemodelandtheVTransmodeling

process.UnderOptionII,themodelcouldbeusedinitscurrentstatetoprovideVMTestimatestoDEC/ANRforgreenhousegasandotheron‐roadpollutantemissionanalyses.However,OptionIIIprovidesforfurtherrefinementofmodeloutputswhenprovidingVMTdataforairquality

analyses.

Page 29: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

23

Thefeasibilityofthisoptionwoulddependontheextenttowhichuseoftheimprovedmodelwouldbeaformofreimbursementforthetimeandfundsputintoit.AnimprovedandupdatedbutdormantmodelwouldbeofnovaluetoVTrans.

Exampleimprovementsandrefinements

Transitnetwork,possiblywitharailcomponent. Modesplitmodule. Feedbackbetweenmodelingsteps,namelytrafficassignmentandtripdistribution.

Newhouseholddailypersontripratesorverificationofcurrentrateswithupdatedsurveydata.

Aprogramofmorefrequentupdatessothatthemodelcanaidinbase‐year

determinationofperformancemeasures,AADTestimatesbytrafficresearch,androadwayassetmanagementandsafetydataneeds.

ProcedurefordeterminingVMTbyvehicleclassconsistentwithon‐roademissionsmodel

(e.g.EPAMOBILE’s16vehicleclasses);procedurewouldnotnecessarilyoccurwithinthemodelandinsteadcouldbeapost‐procedurewherethetotalVMTforecastedbythemodelcouldbedisaggregatedbyvehiclefractions(vehiclefractionscouldbedefinedfor

theentirestate,orbyregion,orbyroadwayfunctionalclass,orbysomecombinationofareaandroadwayclass).

Defineroadwaygradeorterraindatatothemodel’shighwaynetworklinks

Developanintra‐agencyoutreachprogramforuseofthemodelinVTrans,sothatotherdivisionsareawareofthemodelanditscurrentuses,andtosolicitdetailedplansforhowmodelcangainbroaderutility.

Worktypeandhours

Formorecompleximprovementstothemodel,suchasatransitnetworkoramodesplitmodule,VTranswouldneedtocontractaconsultanttoperformtheworktasks.Duration

ofworkwouldgreatlydependonthenatureofthemodelimprovement/refinementtasks.Morebasicupdates,asoutlinedinOptionII,couldbeperformedindependentlyandasdescribedinOptionII.

AsmallcommitteeofstakeholderswouldbeneededcomprisedofVTransandotherrelevantstateagencystaffmembers,UVMTRCrepresentatives,andinvolvedprivateconsultants,ifany,todevelopascopeandworkplanforthemodel’supdates,

improvements,andapplication.

Pros Expandsfunctionalityofthemodel,makingitsuitableforabroaderrangeof

applications.

Couldgenerateincreased“awareness”andutilityofthemodel. Couldimproveaccuracyofmodelforecasts.

Cons CostlierthanOptionsIandII.

Page 30: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

24

Improvedfunctionalitywouldnotnecessarilyleadtoincreaseduseofthemodel–increaseduseofthemodelmaybemorerelianton“awareness”ofthemodelbyotherVTransdepartmentsandstateagencies,andmayneedtoprecedemajormodel

improvementstomakethoseimprovementsworthwhileandcost‐effective.

5.1.4 OptionIV–Improvethemodel’scapabilityforVTransandcoordinationwithanalytical/planningobjectivesofotherorganizationssuchasCCMPOandRegionalPlanningCommissions

ThisoptionwouldincludenecessaryupdatestothemodelasoutlinedinOptionII,andwould

likelyincludesomeoralloftheimprovementslistedinOptionIII.Underthisoption,morespecificrefinementswouldlikelyoccurinresponsetomodelingneedsoftheCCMPObeyondChittendenCountyandtheindividualRegionalPlanningCommissions.

Forexample,theupcoming2060MetropolitanTransportationPlan,whichtheCCMPOwill

prepare,isexpectedtoforecastgrowthandtrip‐makingtrendsfortheMPOregionforthenextfiftyyearsandevaluatedevelopmentandtransportationsystemplans.DuetotheanticipatedincreaseoftripsbetweenChittendenCountyanditsneighbors,suchasFranklin,Lamoille,and

AddisonCounties,thereisadesireformoredetailedtraveldemandforecastingforthoseneighboringareas,whicharebeyondtheboundaryoftheMPO’sregionalmodel.Currently,thestatewidemodel,whichspatiallycoverstheareasneededforthepreparationoftheMTP,does

nothavetherequiredforecastingcapabilitytoyear‐2060.Inresponse,improvementstothestatewidemodelcouldincludemoredistancehorizonyearforecasts,suchas2060,andmechanismsfordataconsistencyandtransferwiththeCCMPOregionalmodel.

Worktypeandhours

SimilartoOptionIII,butwithmorerequiredtimeandbudgetresourcesforfurtherimprovementsandcoordinationwiththeCCMPOandtheRPCs.

AsmallcommitteeofstakeholderswouldbeneededcomprisedofVTransandotherrelevantstateagencystaffmembers,CCMPO,RPCandUVMTRCrepresentatives,andinvolvedprivateconsultants,ifany,todevelopascopeandworkplanforthemodel’s

updates,improvements,andapplication.

Pros CreatesadefinedroleforthestatewidemodelintheCCMPOandRPC’splanning

processes.

HelpstoformamoreunifiedplanningprocessforVTrans,theCCMPO,andtheRPCs.

Cons Costliestofthefouroptions,sinceitrequiresthegreatestextentofmodelupdatesand

improvementsaswellasincreasedcoordinationwithorganizationsoutsideofVTrans.

Page 31: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

25

5.2 OptionsandModelStrengthsandWeaknesses

Table3highlightsgeneralstrengthsandweaknessofthecurrentstatewidemodel,andalsoliststheoptionoroptionsthatcouldpotentiallyaddresseachweakness.Anumberoftheweaknesseslistedcouldbeconsideredsuchonlyinthecontextofadesired,moreextensiveutilityofthe

model.Thatis,themodeliscurrentlycapableofperformingitspresentlyintendedfunctions,regardlessoftheweaknessesordeficiencieslistedhere.

TABLE3:MODELSTRENGTHS&WEAKNESSWITHRELEVANTOPTIONS

Strengths Weaknesses CategoryofWeakness

Option(s)toAddressWeaknesses

Generallyup‐to‐datetraveldemandmodelincurrentsoftwareplatform(Cube)

Base‐year2010willsoonbeneeded,witha2030forecastyear

InputDataandRates

OptionII/III/IV

FourStepmodel(widely‐usedandacceptedmodelingmethodology)

Potentiallyout‐datedcalibrationoftriprates,basedon1994Vermonthouseholdsurvey

InputDataandRates

OptionIII/IV(possiblyOptionII)

Well‐developedandwell‐organizedTAZandroadwaynetworklayers

Noroadwaygradedatainnetworklayer

InputDataandRates

OptionIII/IV

Clearandwell‐documentedmodelstructure/procedure

Nofeedbackbetweenmodelsteps

ModelAlgorithm

OptionIII/IV

Broadrangeoftrippurposes(HBW,HBSC,HBSH,HBO,NHB)

Notransitnetwork/assignment ModelComponent

OptionIII/IV

Recentmodelapplicationand"awareness"ofmodelatpublicandprivateentitiesthroughoutthestate(e.g.CCMPO,CAMSYS,VHB)

Nomodesplitmodule(instead,fixedpercentagesharesbasedoncomparisonofhighwaytimesandtransittimes)

ModelComponent

OptionIII/IV

‐‐ Nofreightrailnetwork/assignment

ModelComponent

OptionIII/IV

‐‐ Limitedintra‐agency"awareness"ofmodelfunctionanduses

ModelingProgram

OptionIII/IV

6. TravelDemandModelingSoftwarePackages

Theincreasingprocessingpowerandutilityofpersonalcomputershaveprovidedanopportunityfordevelopmentandmarketingofever‐improvingtraveldemandmodelingsoftwarepackages.Coupledwiththefeaturesofgeographicinformationsystems(GIS),thesemodelsprovidea

valuabletoolforplannersanddecisionmakers.Somecommonlyusedsoftwarepackagestoday,bothnationallyandglobally,includeCubeanditspassengerdemandmodelingmodule,Voyager(anewincarnationofTRANPLAN/TP+softwarelineage),EMME/2,TransCAD,andVISUM.

Threemodelingpackageswerechosenforapreliminaryreview.Cubewasselectedsinceitisthe

nativeplatformofthecurrentVTransmodelandisawell‐establishedmodelingsoftware

Page 32: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

26

package.TransCADrepresentsamajorshareofnationalandglobalmodelingtodayandisusedinVermontfortheChittendenCountyMPOregionalmodel.VISUM,thoughusedtoalesserextentthanCubeandTransCAD,isarobustmodelthatispartofanintegratedsoftwaresuite,including

VISSIM,apremiermicro‐simulationmodel.Table4summarizesthefeaturesandgeneralinformationaboutthethreesoftwarepackages,andthefollowingsectionsdiscussthepackagesingreaterdetail.

Anewerandhithertounvettedpackageforpotentialstatewidemodelingapplicationis

TRANSIMS(TRansportationANalysisandSIMulationSystem),anactivity‐basedtraveldemandmodelrepresentingindividualtravelersinitiallydevelopedattheLosAlamosNationalLaboratory.Itemploysthedisaggregatemodelingtechniquesoftheactivity‐basedparadigmwith

integrateddynamicmicroscopicnetworksimulationmodeling.TRANSIMSisanewandongoingprogrammingeffort,undergoingiterationsandversionupdatesasanopen‐sourceapplication(http://transims‐opensource.org).Itisexcludedfromthisevaluationduetolimitedapplication

thusfar,especiallyonastatewidescale.

6.1 Cube/Voyager

TheVTransstatewidepassengerdemandmodeliscurrentlyintheCube/Voyagerplatform.CubeisdevelopedbyCitilabs,Inc.andisasuiteofspecificmodulestoaccomplishabroadrangeofmodelingandplanningapplicationsrunningintheMicrosoftWindowsoperatingsystem(32).

CubeBaseisthefoundationofthesoftwareandhousestheintegratedArcGIS(developedbyESRI)interfaceandthemenuswiththesoftware’sfunctions.MappingfeaturesandnetworkeditingareaccomplishedusingtheGISinCubeBase.CubeBasealsomakesusesofgeodatabases

throughtheDatabaseManager,allowingquickaccesstomapdataformanipulation,display,anduseinVoyagermodelsteps.IntegratedwithCubeBaseenvironment,theprimarymodelingmodulesinclude:

Voyager–passengerdemandmodeling

Cargo–freightdemandmodeling Land–Landuseforecasting AvenueandDynasim–trafficanalysisandmicro‐simulation

Analyst–travelmatrix/triptableestimation

CubeallowstransportationplanninguserstodeveloptheirplanningmodelsusingtheScenarioandApplicationManagers.ScenarioPlannersorganizesmodelrunsbycategory,suchasforecastyearorplanningalternative(e.g.,theVTransmodelhasthreescenariosinCube:baseyear,year

2020,andyear2030).TheApplicationManagerusesahierarchicalflowchartstructureandscriptingtodefinethevariousstepsintheplanningmodelandprovidesaclearviewofeachpartoftheprocess.TheopenstructureofApplicationManagerallowsfortheapplicationofthe

traditionalfour‐stepprocess,includingthevariousmethodologiesavailabletoaccomplisheachstep,andotherapproaches,suchasactivity‐basedanddiscretechoicemethods.

Page 33: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

27

Multi‐modalpassengerdemandmodelingispossibleinCube/Voyager,includingroadway/highway,transit,pedestrians,andbicycles.CubeCargoincludesthefreightforecastingprocedures.Cube/Avenueprovidestoolsforviewingandanalyzinglinkflows,turningmovement

volumesatintersections,andtrafficoperationsimpacts.Cube/Dynasimisatrafficmicro‐simulationmodelforvideopresentationoftrafficconditions.

6.2 TransCAD

DevelopedbytheCaliperCorporation,TransCADisarobustGISfortransportationplanningapplicationsandoperatesintheMicrosoftWindowsoperatingsystem(33).TransCADcan

performthefour‐stepforecastingprocess,eithermanuallyusingbuilt‐inmenusorthroughtheuseofcustomcodingintheGISDeveloper’sKit(GISDK)interface.TransCADprovidesalternatemethodologiesforeachstepofthetransportationforecastingprocess.Tripgenerationcanbe

accomplishedusingcross‐classificationtechniqueswithuser‐definedtripratesorapplicationofestablishedtripratesfromtheITETripGenerationManualandNCHRPReport365.Gravitymodels,growthfactors,orinterveningopportunitiescanbeappliedfortripdistribution.Likethe

othersteps,modechoicemethodscanbeuser‐definedorperformedusinglogitmodelapplicationfeatures.Trafficassignmenthasanumberofpossiblemethodologies(equilibrium,all‐or‐nothing,incremental,systemoptimal,etc.)andTransCADalsofeaturesanumberoftools

toanalyzeanddisplayassignmentresults.

TheGISisusedfordevelopment,manipulation,anddisplayofmodelcomponents,includingtrafficanalysiszones,networks,routes,andunderlyingdatastructures–tables,matrices,andmaps.Numerouspassengertransportsystemscanbemodeled,includingroadways,transit,

freight,andnon‐motorizedmodes–pedestriansandbicycles.Asapowerfuldatatool,TransCADalsoincludesaplethoraofthemostcurrentUSCensusdataforvaryingspatiallevelsandprovidestoolsforaccessingandviewingtheUSCensusdata.Moreover,routingfunctionality

allowsforlogisticsanalysesandsystemplanning.

Page 34: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

28

TABLE4:SUMMARYOFCUBE,TRANSCAD,ANDVISUMFEATURESANDFUNCTIONALITY

FEATURES/SPECIFICATIONS

Cube TransCAD VISUM

Developer Citilabs Caliper PTV

CurrentVersion 5 5.0 10

VendorSupportandSoftwareServices

Technicalsupport,training,existingmodelconversion,andcustom

moduledevelopment

Technicalsupport,training,andcustomsoftware

developmentbasedonuserneeds

Technicalsupport,training,andcustomsoftware

developmentbasedonuserneeds

VendorStaffOffices

SanFrancisco,CA/WashingtonDC/Philadelphia,PA

Boston,MA/WashingtonDC(PTVAmericaoffices)

Portland,OR/Wilmington,DE/Austin,TX

Micro‐simulationCompanion

Dynasim5 TransModeler2.0VISSIM5(oftenusedasa"stand‐alone"model)and

HCMcompatibility

PCPlatform Windows Windows Windows

SoftwareStructure

"Base"usingArcGISwithmodularextensions,buttightlyintegrated,Extensions:Voyager(passenger),Cargo(freight),Avenue(trafficanalysis),Analyst(triptableestimation),Land(landuse

forecasting),Dynasim

AGISwithallfeaturesareinasingle,integratedplatform;TransModelerseparate

Allfeaturesareinasingle,integratedplatformusingArcGIS,VISSIMintegrated

GISComponent EmbeddedGISfromESRI IntegratednativeGIS EmbeddedGISfromESRI

NetworkEditorGIS‐based(integratedformof

ArcGIS)GIS‐based(integrated,native)

GIS‐based(integratedformofArcGIS)

EaseofUse

Flowchart‐basedApplicationManagerfordevelopingmodelingprocess("wrapper"),ScenarioManager,scripting,toolbars

Menus,scripting,toolbars(no"undo"featureformanytasks),Model/Scenario

Manager

Menus,scripting,toolbars

DataManagementTools

Matrixmanipulation,tablesMatrixmanipulation,tables,USCensusdata/geographic

datatoolsincludedMatrixmanipulation,tables

ModelingTechniques

four‐stepModel,Activity‐basedDemandModel,numerousdiscretechoiceandassignmentoptions

four‐stepModel,Activity‐basedDemandModel,

numerousdiscretechoiceandassignmentoptions

four‐stepModel,Activity‐basedDemandModel,

numerousdiscretechoiceandassignmentoptions

ModelFeedback

Methodsoffeedbackfortripdistribution/modechoice/traffic

assignment,economicandlandusefeedback,canconsidertrafficsignaldata/capacityinassignmentprocess

Methodsoffeedbackfortripdistribution/mode

choice/trafficassignment,canconsidertrafficsignal

data/capacityinassignmentprocess

Feedbackforintegrateddistribution/assignment

loops

ModalModelingPassenger,Freight,Transit,

Pedestrian,BicyclePassenger,Freight,Transit,

Pedestrian,BicyclePassenger,Freight,Transit,

Pedestrian,Bicycle

FEATU

RES/SP

ECIFICATIONS

CommonPracticeforModelImplementation

Codingofcustomapplications/interface,

User/developertrainingandcertification

Codingofcustomapplications/interface,

User/developertrainingandcertification

Codingofcustomapplications/interface

CurrentUsersVermontDOT,MaineDOT,CaliforniaDOT,FloridaDOT,WisconsinDOT,

ARC(GA),etc.

CCMPO(VT),MichiganDOT,OhioDOT,IndianaDOT,North

CarolinaDOT,SCAG(CA),NYMTC(NY),etc.

WashingtonDOT,NewYorkStateDOT,CapitalDistrictTransportationCommittee

MPO(NY),etc.Summaryinformationcompiledfromsources:9,32,33,34,35,36,37.

Page 35: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

29

6.3 VISUM

DevelopedbyPTV,VISUMisusedbothnationallyandgloballyfortraveldemandmodelingandtransportationplanning,thoughtoalesserextentthaneitherCubeorTransCAD.VISUMisofferedinanumberofpackagelevels,dependingonuserneeds,andincludesaversionofthe

micro‐simulationmodel,VISSIM,aspartofthePTVVisionSuite(34).Itiscapableofimplementingabroadrangeofplanningmodelmethodologiesandincorporatesallstandardmodesoftransportation–drive‐aloneauto,carpool,publictransit,bicycles,pedestrians,and

trucks.VISUMusesanintegratedGISfromESRI(similartoCube),fordisplayandnetworkediting.Furthermore,ithastheadvantageofintegrationwithVISSIM,awidelyusedandrespectedmicro‐simulationmodel.

Moreadvancedfunctionalityincludesactivity‐basedmodeling,dynamicassignmentprocedures

withlinkandintersectionmodels,anddetailedtransitassignmentwithoperationalanalyses.Activity‐basedmodelingispossibleusingVISEM,asystemthatperformstraveldemandestimatesastripmatricesforinputintoVISUM,thusreplacingthetraditionalstepsoftrip

generation,tripdistribution,andmodechoice.VISUMoffersanumberof“postassignment”analysisfeaturesincludingtraveltimeisochrones,nodeflows,subareanetworkisolation,trafficsignaleditor,andintersectionmodeling.VISUMalsoreadilyimportsmodelsfromothersoftware

packages,includingCube,TransCAD,EMME/2,andTModel2.

6.4 PractitionerExperiences

Inmid‐2003,theFloridaDepartmentofTransportation,whichatthetimewasusingTRANPLAN

asthesoftwareenginefortheFloridaStandardUrbanTransportationModelStructure(FSUTMS),reportedthefindingsofitsevaluationofnumerousmodelingsoftwarepackagesand

recommendationsforthenextiterationofitsstatewidemodel(9).ThereportconsideredmodelingsoftwarepackagesincludingVISUM,Cube/Voyager,EMME/2,andTransCADinthecontextofeachprogram’srobustness,operationalstrengthsandweakness,applicabilityto

modelinginFloridaaspartofFSUTMStoolbox,andabilitytomeettherequirementsofFlorida’sfuturemodelingneeds.Practitionersurveysprovideduseropinionsandratingsforthesoftwarepackages,andCube/VoyagerandTransCADwereshortlistedbythestudy’ssteeringcommittee

fordetailedevaluation,includingreviewofsoftwarespecificationsanddiagnosticsofvariousmodelfunctions.Ultimately,thestudy’smodeltaskforcevotedinmid‐2003toadoptTransCADinsteadofCube/VoyagerasthenewsoftwareplatformforFSUTMS.However,documentation

indicatesthatapproximatelyoneyearaftertherecommendationofTransCADandthestartofthetransitionalprocess,themodeltaskforcedecidedinOctober2004tonegotiatewithCitilabsfortheimplementationofCube/VoyagerforFSUTMS,essentiallyreversingthe2003

recommendationforTransCAD(35).SlowdevelopmentoftheTransCADengineforFSUTMSandanunsatisfactorybusinessrelationshipwiththedeveloperwerecitedasreasonsforthedecision.Currently,theFloridaDepartmentofTransportationhasimplementedandisusingCube/Voyager

forFSUTMS.

Page 36: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

30

TheCityofIrvineDepartmentofPublicWorks’AdvancedTransportationDivisionalsoperformedamodelingsoftwareevaluation,includingCube/VoyagerandTransCAD.RecommendationsincludedusingCube/Voyagerforshort‐termupgradestotheexistingmodelinTRANPLANand

implementationofTransCADforlong‐termnewmodeldevelopment(36).

Inrecentyears,theNorthCarolinaDepartmentofTransportation,theSouthernCaliforniaAssociationofGovernments(SCAG),andtheNorthCentralTexasCouncilofGovernment(NCTCOG)havetransitionedtheirmodelsfromTRANPLANtoTransCAD(37).TheOrangeCounty

TransportationAuthority(OCTA)usesTransCADfornewmodeldevelopment.

Currently,otherusersofCube/VoyagerincludeDepartmentsofTransportationinCalifornia,Maine,Florida,andWisconsin,andtheAtlantaRegionalCommission(ARC)inGeorgia.OtherusersofTransCADincludeDepartmentsofTransportationinOhio,Massachusetts,Michigan,

Iowa,Indiana,andtheChittendenCountyMPO(CCMPO)inVermont.Similartomanyconsultingfirms,ResourceSystemsGroup,Inc.,basedinWhiteRiverJunction,Vermont,isversedinabroadrangeoftraveldemandmodelingsoftwarepackages,includingCube,TransCAD,andVISUM(see

AppendixA:A‐6).Thisisabrieflistofactivemodelingsoftwareusersconsideringtheubiquityofthemodels,bothdomesticallyandinternationally.

TransCADisactivelyusedforplanning,studies,andresearchinVermontbytheCCMPOandtheUniversityofVermontTransportationResearchCenter.In2005,ResourceSystemsGroup,Inc.

completeda2000baseyearregionalTransCADmodelofChittendenCounty,andcontinuallymaintainsandupdatesthemodelfortheCCMPO(23).Themodelperformstraveldemandforecastsfor2005to2030atfive‐yearintervals,andisusedfortheMetropolitanTransportation

Plan(38),aswellasotherplanningstudies.Similarly,theTransportationResearchCentercurrentlyusesTransCADasananalyticaltoolforanumberofresearchprojects,includingtheNetworkRobustnessIndex:AComprehensiveSpatial‐BasedMeasureforTransportation

InfrastructureManagement(39)andTransportationImpactsofTransit‐OrientedDevelopmentinRuralTowns(40).

7. SummaryandRecommendations

Thestatewidemodelisavaluabletraveldemandforecastingandtransportationanalysistoolfor

Vermont,anditshouldbemaintained,updated,andrefinedsothatitcancontinuetoserveinthatregard.ThemodelhasbeensuccessfullyappliedtoplanningstudiesinthestateandhasanapparentroleintheLongRangeTransportationBusinessPlan,theHighwaySystemPolicyPlan,

andinairqualityanalysesforenvironmentalplanning.Withnewtransitandrailcomponents,themodelwouldalsobepositionedtocontributetothegoalslistedinthePublicTransportationPolicyPlanandtheStateRail&PolicyPlan.

Thestatewidemodelhasrecentlyundergoneaseriesofupdatesandamigrationtoanew

softwareplatformin2007,butfurtherrefinementsofthemodelwouldbebeneficial.Themodelshouldbecontinuallymaintained,includingcalibrationandvalidationprocesses,and

Page 37: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

31

improvementstoitsfour‐stepalgorithmsshouldbeexplored,suchasanewmodechoicemoduleandtheinclusionofafeedbackloopbetweenassignment,distribution,andmodechoice.ThefurtherdisaggregationofitsVMTandaveragespeedoutputsshouldalsobeconsideredto

potentiallyimproveemissionsmodelingforairqualityanalyses.

Thethreetraveldemandmodelingsoftwarepackagesreviewedinthisreportmeetthecurrentrequirementsandpotentialfutureusesofthestatewidemodel.Fundamentally,theirfunctionalitiesaresimilar,butwithuniqueinterfacesandformatting,sothedecisiontouseone

insteadofotherswouldprimarilydependonuserpreferenceorconformitywithothersoftware(i.e.,anArcGISusermayprefertheGISinterfaceandoptionsofCube,andpossiblyVISUM,insteadofTransCAD).Animportantadditionalfactortoconsideristhatthereiscurrentlyabase

ofpractitionersandresearchersinVermontemployingTransCAD.

Ataminimum,thereisenoughcausefortheVermontStatewideModeltobemaintainedandperiodicallyupdatedtokeepitviableandavailableforplanningneedsatVTrans.Theoverallmodelcouldbekeptinitscurrentformwithoutsubstantialimprovements,suchastransitand

railnetworks,butshouldremainavailableforforecastsofroadwaytravelandVMTestimates.Mostimportantly,thesuccessandlongevityofthestatewidemodelgreatlydependsontheclearly‐definedproceduresandgoalsofitsutilitybyVTransandotherVermontagencies.As

statedpreviously,thegoalsofastatewidemodelingprogramshouldbedefinedwellinadvanceofmodel‐specificdetails,suchasdataneeds,modelcomponents,computersoftware,andbudgetallocation.

Page 38: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

32

References

1. Horowitz,A.J.,etal.NCHRPSynthesis358:StatewideTravelForecastingModels.TransportationResearchBoardoftheNationalAcademies,Washington,D.C.,2006.

2. Giaimo,G.T.andSchiffer,R.StatewideTravelDemandModeling:APeerExchange,LongboatKey,Florida,September23‐24,2004.TransportationResearchCircular,No.E‐C075.TransportationResearchBoardoftheNationalAcademies,Washington,D.C.,August2005.

3. Horowitz,A.J.andFarmer,D.D.“ACriticalReviewofStatewideTravelForecastingPractice,”TransportationResearchRecord1685,TransportationResearchBoard,NationalResearchCouncil,Washington,D.C.,1999,pp.13–20.

4. Bowman,J.L.andBen‐Akiva,M.“Activity‐BasedTravelForecasting”Activity‐BasedTravelForecastingConferenceProceedings,NewOrleans,LA,June2‐5,1996.

5. Davidson,W.,etal.“Synthesisoffirstpracticesandoperationalresearchapproachesinactivity‐basedtraveldemandmodeling.”TransportationResearchPartA:PolicyandPractice,2007,vol.41,issue5,pages464‐488.

6. Vovsha,P.,Bradley,M.andBowman,J.“Activity‐BasedTravelForecastingModelsintheUnitedStates:ProgressSince1995andProspectsfortheFuture.”InProgressinActivity‐BasedAnalysis(H.Timmermans,ed.),Elsevier,Amsterdam,Netherlands,2005,pp.389–414.

7. Vovsha,P.andBradley,M.“AdvancedActivity‐BasedModelsinContextofPlanningDecisions.”TransportationResearchRecord:JournaloftheTransportationResearchBoard,No.1981.TransportationResearchBoardoftheNationalAcademies,WashingtonD.C.,2006,pp.34‐41.

8. MetropolitanTravelForecasting,CurrentPracticeandFutureDirection,SpecialReport288.TransportationResearchBoardoftheNationalAcademies,Washington,D.C.,2007.

9. Pendyala,R.M.,etal.“EvaluationofTransportationModelsfortheStatewideModelTaskForce.”FloridaDepartmentofTransportationResearchCenter,2003.

10. VHB/VanasseHangenBrustlin,Inc.VermontStatewideTravelDemandModelImprovements:UpdatedPassengerandTruckModelsinCube/Voyager.VermontAgencyofTransportation,June2007.

11. Horowitz,A.J.,GuidebookonStatewideTravelForecasting,ReportFHWA‐HEP‐99‐007,FederalHighwayAdministration,Washington,D.C.,July1999.

12. Martin,W.A.,etal.NCHRPReport365:TravelEstimationTechniquesforUrbanPlanning.TransportationResearchBoardNationalResearchCouncil,Washington,D.C.,1998.

13. Circ‐WillistonEIS:(http://www.circeis.org/).

14. BenningtonBypass:(http://www.aot.state.vt.us/progdev/Sections/Design/bennBypass/Benn.htm).

15. WesternCorridorTransportationManagementPlan:(http://www.vtwesterncorridor.org/).

16. CambridgeSystematics,Inc.Vermont’sHighwaySystemPolicyPlan.VermontAgencyofTransportation,June2004.

Page 39: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

33

17. TranSystems.Vermont’sPublicTransportationPolicyPlan.VermontAgencyofTransportation,February2007.

18. ParsonsBrinckerhoff.StateRail&PolicyPlan,2006,StateofVermont.VermontAgencyofTransportation,December2006.

19. CambridgeSystematics,Inc.VTransAssetManagementVisionandWorkPlan.VermontAgencyofTransportation,June2002.

20. CambridgeSystematics,Inc.VermontLongRangeTransportationPlan.VermontAgencyofTransportation,January2002.

21. ResourceSystemsGroup,Inc.VermontLongRangeTransportationBusinessPlan,PublicReviewDraft.VermontAgencyofTransportation,October2008.

22. FinalReportandRecommendationoftheGovernor’sCommissiononClimateChange,PresentedtoGovernorJamesH.Douglas.October2007.

23. ResourceSystemsGroup,Inc.CCMPORegionalTransportationModelDocumentation:2000BaseYearModel.ChittendenCountyMetropolitanPlanOrganization,January2008.(Retrievedfrom:http://www.ccmpo.org/modeling/release_history.html).

24. U.S.EnvironmentalProtectionAgency.“EPAFindsGreenhouseGasesPoseThreattoPublicHealth,Welfare/ProposedFindingComesinResponseto2007SupremeCourtRuling.”April17,2009.www.epa.gov(Retrievedfrom:http://yosemite.epa.gov/opa/admpress.nsf/0/0EF7DF675805295D8525759B00566924).

25. PewCenteronGlobalClimateChange.“CaliforniaVehicleStandards.”www.pewclimate.org(Retrievedfrom:http://www.pewclimate.org/what_s_being_done/in_the_states/vehicle_ghg_standard‐moreinfo.cfm).

26. OfficeoftheGovernorofCalifornia.“Gov.SchwarzeneggerSignsLandmarkLegislationtoReduceGreenhouseGasEmissions.”gov.ca.gov(Retrievedfrom:http://gov.ca.gov/press‐release/4111/).

27. PewCenteronGlobalClimateChange.“StateLegislationfromAroundtheCountry.”www.pewclimate.org(Retrievedfrom:http://www.pewclimate.org/what_s_being_done/in_the_states/state_legislation.cfm).

28. VermontAgencyofNaturalResources.“AirPollutionControlRegulations,SubchapterXI,LowEmissionVehicles–RegulationstoControlGreenhouseGasEmissionsfromMotorVehicles.”November7,2005.(Retrievedfrom:http://www.anr.state.vt.us/air/htm/ProposedAmendments.htm#cars).

29. VermontDepartmentofPublicService.“VermontComprehensiveEnergyPlan2009,[An]Updatetothe2005Twenty‐YearElectricPlan,PublicReviewDraft.”May2008.(Retrievedfrom:http://publicservice.vermont.gov/planning/CEP%20%20WEB%20DRAFT%20FINAL%206‐4‐08.pdf).

30. DemocraticStaffoftheCommitteeonEnergyandCommerce.MemorandumtoMembersoftheCommitteeonEnergyandCommerce,U.S.HouseofRepresentatives,regardingmask‐upsessiontoconsiderH.R.2454,theAmericanCleanEnergyandSecurityActof2009.May16,2009.(Retrievedfrom:

Page 40: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

34

http://energycommerce.house.gov/index.php?option=com_content&view=article&id=1625:chairmen‐waxman‐and‐markey‐introduce‐the‐american‐clean‐energy‐and‐security‐act&catid=141:full‐committee&Itemid=85)

31. CaliforniaAirResourcesBoard.“EMFAC2007version2.30,CalculatingemissioninventoriesforvehiclesinCalifornia,User’sGuide.”November2006.(Retrievedfrom:http://www.arb.ca.gov/msei/onroad/latest_version.htm).

32. Cube/Voyager,version5.©1999‐2008Citilabs,Inc.(http://www.citilabs.com).

33. TransCAD,version5.0.©1994‐2008CaliperCorporation.(http://www.caliper.com).

34. VISUM,version10.©1998‐2008PTV/PTVAmerica,Inc.(http://www.ptvamerica.com).

35. MinutesoftheFloridaModelTaskForceMeeting(2004).Orlando,Florida,MeetingofOctober6,2004.

36. Chen,J.X.TravelDemandModelingSoftwareEvaluation.PresentedattheSCAGModelingTaskForceMeeting,January24,2007.

37. Cervenka,K.TransCADModelingatNCTCOG:HowWeDidIt.PresentedattheFloridaModelTaskForceMeeting,November12,2003.

38. ChittendenCountyMetropolitanPlanningOrganization.2025ChittendenCountyMetropolitanTransportationPlan.Adopted,January2005.(Retrievedfrom:http://www.ccmpo.org/MTP/).

39. UniversityofVermontTransportationResearchCenter.NetworkRobustnessIndex:AComprehensiveSpatial‐BasedMeasureforTransportationInfrastructureManagement.(Retrievedfrom:http://www.uvm.edu/~transctr/?Page=022318.html&SM=Research_Submenu.html).

40. UniversityofVermontTransportationResearchCenter.TransportationImpactsofTransit‐OrientedDevelopmentinRuralTowns.(Retrievedfrom:http://www.uvm.edu/~transctr/?Page=022755.html&SM=Research_Submenu.html).

Page 41: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

35

AppendixA

MeetingSummariesandCorrespondence

Page 42: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

36

A‐1MeetingDate:November6,2008Location:VermontAgencyofTransportation,Montpelier,VTPartIAttendees:MaureenCarr,VTrans,TrafficOperations ([email protected],802‐828‐3091)JohnBlodgett,VTrans,TrafficOperations ([email protected],802‐828‐3972)RichardWatts,UVMTRCAndrewWeeks,UVMTRCMaureenandJohnexplainedwhattheirusualtasksareatVTrans,whichincludemanagement,analysis,andreportingofVTranscountsthroughoutthestate.Theyalsomakeprojectionsforfutureyearvolumesbasedonhistoricaltrends.Theyprovidedandexplainedsomeoftheiroffice’sAADTandseasonalvolumetrendsreports.Whenaskedhowthestatewidemodelcouldaidintheirtasks,Maureenindicatedthattheycouldmakeuseofthemodelforthefollowing:

trafficgrowthprojectionsforfutureyearprojects effectsoftrafficnetworkdisruptions,suchasdetours/re‐routingduetoabridgeclosure,

orsomeothernetworklinkloss(thisdatawouldthenbeusedfordetailedanalysesalongdetourroutesusingsimulationsoftware)

bettervolumeestimationsfor“local”roadwayclasses,whicharenotroutinelycountedTheTRCwillcoordinatewithMaureenandJohnasthestatewidemodelevaluationprogresses.PartIIAttendees:CostaPappis,VTrans,PolicyandPlanning ([email protected],802‐828‐5790)RichardWatts,UVMTRCAndrewWeeks,UVMTRCCostaprovidedthestatewidemodelfilesanddocumentationpreparedbythepreviousmodeler,includingadevelopmenttimelineofthemodel,set‐upinstructions,listsofthefilestructure,andcontactinformationofthoseinvolvedinitsdevelopmentanduse.Costadescribedsomeprevioususesofthestatewidemodel,includingenvironmentalimpactstatements(e.g.Circ‐WillistonEIS),theBenningtonBypass,theMorristownBypass,theWesternCorridorStudy,andtheStateEmployeeCommutingStudy.SimilartoMaureen’scomments,Costaindicatedaninterestinusingthemodelfordetour/re‐routingevaluationsandITS.CostasaidthatVTransneedstodevelopamodelingworkplanforthenext3‐4years,likelyaspartoftheLongRangeTransportationBusinessPlan,andwouldincluderecommendationsfromtheTRC.Costahighlightedsomeneededimprovementstothestatewidemodel:

transitnetworktoanalyzemulti‐modalalternatives

Page 43: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

37

integrationwiththeCCMPOmodelshouldbeexplored,suchastransferabilityofnetworkandTAZdata;travelmodesshouldmatchthoseintheCCMPOmodelfordatacoordinationwiththeregionalmodel

preparationforpotentialnon‐attainmentinareasofVermont(theVermontDepartmentofEnvironmentalConservationcurrentlyoverseesairqualitywork,butVTranswouldneedtoperformanalysesforconformitydeterminationandtheSIPshouldnon‐attainmentoccur)

potentialfreightmodelchanges,includingarailcomponent(thisdependsontheoutcomeofVTransfreightplanRFPcurrentlyactive)

CostasaidthatsinceVTransisin“preservationmode”therewouldlesslikelybeagreatneedformodeluseinmajorcapitalimprovementstudies,butalternativelyfortransitimprovementstudies.Afterthemeeting,wespokeinformallytoGinaCampoli([email protected],802‐828‐5756)whomentionedthepotentialmodelusefortheGovernor’sClimateChangeCommission(fortransportation‐relatedpollutantandgreenhousegasemissionsandenergypolicies).TheTRCwillcoordinateprimarilywithCostaasthestatewidemodelevaluationprogresses.ActionitemsforTRC

BegintoreviewVTransstatewidemodelanddocumentation PreparedraftrecommendationsforpotentialuseandupdatesoftheVTransstatewide

model Continuemodelingsoftwarepackageevaluationandreviewofmodelingpracticesat

otherstatedepartmentsoftransportation

Page 44: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

38

A‐2MeetingDate:December10,2008Location:VermontAgencyofTransportation,Montpelier,VTAttendees:ClayPoitras,VTrans,PolicyandPlanning ([email protected],802‐828‐3968)ScottBascom,VTrans,PolicyandPlanning ([email protected],802‐828‐5748)CostaPappis,VTrans,PolicyandPlanning ([email protected],802‐828‐5790) AndrewWeeks,UVMTRC

ScottismanagingtheHighwaySystemPolicyPlan,whichispartoftheMulti‐modalPolicyPlan.

Costamanagescorridorplanning,publictransitandfreightplanning,withthePublicTransportationSystemPolicyPlanandRailPolicyPlanaspartoftheMulti‐modalPolicyPlan.

Claywouldliketothinkoutside‐the‐boxaboutwhatotheragenciescouldusethemodel’soutput.ANR/DECwoulduseVMTandrelatedinformationforairqualityneeds.Note:WhataboutVermontAssociationofPlanning,DevelopmentAgencies(VAPDA)andAgencyofCommerceandCommunityDevelopment,DepartmentofPublicSafety?

ChittendenCountyiscoveredbytheCCMPOregionalmodel.Note:ForotherareasofVermont,woulditbemoreworthwhiletoisolateanddetailasectionofthestatewidemodelforregionaluse,insteadofbringingtheentirestatewidemodeltothatlevelofdetail?

TheAssetManagementPlancouldpotentiallymakeuseofthemodelforidentificationofroadwaysforfunding/workpriority–aperformancereviewofthenetworktodetermineinvestmentqueuing.Contact:BartSelle([email protected],802‐828‐2757).

ChuckGallagher([email protected],802‐828‐3889)andAlecPortalupi([email protected],802‐828‐3889)fromOperationsmayalsohaveadditionalusesforthatdivision.

Scottshowedinterestedintheideaofidentifyinga“primary”or“critical”roadwaynetworkforthestate,similartoTRC’sworkontheNetworkRobustnessIndex.Usingthestatewidemodelasananalysistoolfordeterminingcriticalroadwayscouldbebeneficial.Note:OneexamplediscussedwasapossiblebridgefailureordisasterevenalongI‐89inthevicinityofWaterbury.Whatwouldbetheregionalimpactandhowcouldtrafficbediverted?ThiswouldalsorelatetotheHighwaySystemPolicyandAssetManagementPlans.(Seealso:possiblereplacementofCrownPointBridge,NYSDOT.)

Regardingsoftwarepackages,CostaindicatedthatitwouldbeimportanttoconsiderCCMPO’spractices(i.e.TransCAD).AnotherimportationconsiderationisthatVTransusesArcGISexclusively,whichistheGISinCUBE.

Page 45: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

39

A‐3PhoneConversationwithClayPoitras,VTransDate:November20,2008DiscussionClay’sgeneralcommentsaboutthestatewidemodel:

Heisconfidentofthelong‐termneedanduseofthemodel.

Therewilllikelynotbemuchneedforthemodelinthenext12‐18months,buthewantstokeepitfunctionaluntilalong‐termplanforuseisdeveloped.

Althoughthemodeliscurrentahighwaymodel,therewillnotbemanybighighwaycapacityprojectsinthefuturethatwouldrequirethemodel.Theemphasiswouldbeonexistinghighwayinfrastructureimprovements.

Clayrefersthemodelas“macro”andclearlystatesthatitbecomesincreasinglylimitedinutilityasthespatialresolutionandanalysisdetailincrease

Therewillbeincreasedemphasisoncorridormanagementandplanning,suchasfor5to10‐milesegmentsofroadways,whichwouldrequireadditionalmicro‐analysiscapabilities(simulation/HCMmethodologies).Thestatewidemodelcouldpotentiallyfeedvolumestothemicro‐analyses.SeeWesternCorridorStudy.

Claycitedanumberofareasthatthemodelcouldbeapplied,perhapsnotimmediately,butatsomepointinthefuture:

Projectdevelopmentandalternativeplanning(contact:KevinMarshia,VTrans)

Airqualityanalyses,possiblenon‐attainment(contact:GinaCampoli,VTrans)

Transportationplanningassistancefortheregionalplanningcommissions(RPC)–specificallythelocaltransportationplanners;abridgedcontactlistisincluded

Corridormanagementstudies(contact:CostaPappis,VTrans);EleniChurchill,CCMPO,canbecontactedregardingtheWesternCorridorStudy

Modalpolicyplan,typicallyat5‐yeariterations(contact:ScottBascom,VTrans)

Possiblenewinterchangecapacityimprovement–I‐89nearHinesburgRd./KennedyDr.inSouthBurlington

ContinueduseforEISwork,suchasCirc‐WillistonDEIS

CollaborationwithCCMPOtosupplementspatiallimitationsoftheCCMPOmodel(contact:EleniChurchill,CCMPO)

ClaysaiditispossibletocontactJulieMurphy,VHB,withtechnicalquestionsaboutthemodel;howeverVHBisnotcurrentlyundercontractwithVTrans.

Page 46: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

40

Clayidentifiedthestatewidemodel’scurrentutilitylimitationsduetonotransitnetworkorrailcomponentinthefreightmodel;Costa’sfreightmodelRFPmaydealwiththelatterlimitation.

Claythoughtitworthwhiletoexplorecurrentmodelingtrends,evolution,andlikelytoolsthatwillprimarilybeusedinthecomingyears.

Clayrecognizedthathiring‐outtoconsultantsforstatewidemodelingtasks(applicationofthemodeland/ormaintenanceonthemodel)isanoption.

Claybelievesthatanimportantquestiontosetastageforthisentireevaluationis:“whyshouldtheStateofVermontandVTranscontinuewithastatewidemodel?”

Clayagreedthattheevaluationprocessismovingintherightdirection–thatweshouldcontinuetodevelopasetoflikelyusesofthemodelinthefuture,identifyanydeficienciesinthemodeltoaddressinordertofacilitatethatsetofmodeluses,andthenmakeanactionplantoaddressthosedeficiencies.

ClaysaidthatwecanprovideupdatestohimandCostaatanypointduringtheongoingprocess,inamannerofourconvenience.

ContactInformationprovidedbyClay

VTransProjectDevelopment:KevinMarshia,Roadway,Traffic&[email protected](802)828‐2664

VTransPolicy&Planning:

ScottBascom,PlanningProjectManager(Scottwillbemanagingtheupdatetothestatewidehighwaymodalplan.)[email protected](802)828‐5748

CostaPappis,PlanningProjectManager(CostamanagesourCorridorManagementStudiesandwillbemanagingthependingstatewidefreightstudy.)

GinaCampoli,AOTEnvironmentalPolicyManager(airquality&other)[email protected](802)828‐5756

RegionalTransportationPlanners:

ContactinformationforregionalplanningcommissionsandtheCCMPOisattached.SuggestconsultingEleniChurchillinadditiontoDaveRobertsattheCCMPO.Eleniismanagingthewesterncorridorstudy.Inaddition–asaformermemberofVTransandacolleague‐sheisverywellfoundedinthestatewidemodel,howitworksanditslimitations.

Page 47: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

41

A‐4ConversationwithEleniChurchill,CCMPODate:December09,2008attheTRCEleniworkedatVTransfor12yearsfromabout1994to2006,managingthestatewidemodel.DuringhertimeatVTrans,shecoordinatedcloselywithJulieMurphyatVHB,theon‐callconsultant,workingonthemodel.DiscussionEleni’sgeneralcommentsaboutthestatewidemodel:

Shefeelsthatitisimportanttokeepthemodelalivewithmaintenanceandupgrades,mainlyfortransitmodeling(modelimprovementsneeded,though)andpotentialairqualityneedsinthefuture,incaseofnonattainmentinVermont.

FortheWesternCorridorStudy,whichEleniismanagingatCCMPO,thestatewidemodelreceivedadditionalcalibrationalongthewesternportionofVermont,goingbeyondtheupdatesthatweredonebyVHBwhenthemodelwasbroughtintoCUBE.CharlieMarkperformedsomeofthatadditionalcalibrationwhilehewasamodeleratVTrans,andEdBromageatCambridgeSystematics,whoisrunningthemodelingfortheWesternCorridorStudy,alsoperformedmorecalibrationforwesternVermont.EleniwillrequesttheupdatedmodelfilesfromCambridgeSystematicsforTRCuse.

Sherecognizestheimportanceoftransitmodelingcapabilitiesinthemodel,andtheneedforarailnetwork,formodelingofpassengerrailandfreightrail.Agrowingdesiretoshiftmorefreightfromtrucktorailcouldprecipitateaneedforrailmodelinginthestatewidemodel.

WhileatVTrans,Eleniwasinvolvedinworktoidentifya“primarynetwork”inVermont–asetofcriticallinks/infrastructure–thatshouldreceivepriorityinfundingandmaintenance.Traveltimefromthestatewidemodelwasakeycomponentinidentifyingthe“primarynetwork,”andthemodelcouldbeusedsimilarlyinthefuture.

ElenibroughtahardcopyofVermontStatewideTravelDemandModelImprovements,VHB,June2007forTRCreference.

ElenialsosuggestedgettingintouchwithJulieMurphyatVHB.WhenVHBwasperformingthemodelupdateandmigrationintoCUBE,partorallofitwasdonefirstinTransCAD.ElenithinksitwouldbeworthwhilefortheTRCtoobtainthoseTransCADfiles.

AlsoregardingTransCADversusCUBE,Elenithinksthatitwouldbenicetohavethestateandregionalmodelallinoneplatform.She,ofcourse,recognizestheneedforaclearplanforfutureusesofthestatewidemodelbeforeworryingaboutdetailssuchassoftwarepackages.

Page 48: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

42

A‐5MeetingDate:March12,2009Location:VermontAgencyofTransportation,Montpelier,VTPartIAttendees:BartSelle,VTrans,PolicyandPlanning(AssetManagement)([email protected],802‐828‐2757)AndrewWeeks,UVMTRCBartbrieflydiscussedthegeneralresponsibilitiesofassetmanagementincludingthesystemsofpavement,bridges,buildings,aviation,safetyandmaintenancemanagement.

Assetmanagementhasestablishedprocedurestomeetitsplanningandpolicy‐makingneeds.Traveldemandforecastingdoesnotplayaninherentroleintheassetmanagementprocedures,sinceassetmanagementisameansforeffectiveinvestmentforexistingtransportationinfrastructure.

Asidefromthetraveldemandforecastingapplicationsofthestatewidemodel,Bartidentifiedapotentialneedforestablishingabaseconditionforlikelyperformancemeasures,includingmobility(capacities,v/candtravelspeed).ThereauthorizationofSAFETEA‐LUin2009mayidentifyaplanforstatetransportationagenciestoestablishguidingperformancemeasuresformaintenance,investment,safety,andmobilityinpolicyplanning.

Increasingemphasisonperformancemeasuresandaccountabilityforprojectapprovalandfundingmayrequiremorequantifiableanddefensiblemetricsinthedecision‐makingprocess.Thestatewidemodelcouldserveasatoolinthisregard,providingquantitativemeasuresofmobility.

PartIIAttendees:BruceNyquist,VTrans,ProgramDevelopment(Roadway,TrafficandSafety)([email protected],802‐828‐2696)AndrewWeeks,UVMTRCBrucediscussedplansforVTranstouseFHWA’sSafetyAnalystprogram(http://www.safetyanalyst.org/index.htm)toidentifyprobablehighcrashlocationsinthestate,andtheassociateddataneeds.Hesaidthattheywouldfocuspredominantlyonruralroadways,includingmajorandminorcollectorsandlocalroads.(Inthestatewidemodel,thosethreetypesareidentifiedasfunctionalclasses7,8,and9,andaccountforapproximately18,000linksinthemodelandroughly6,000milesofroadway.Trafficvolume(AADT)needsforSafetyAnalystwouldspatiallyexceedtheTrafficResearchdepartment’scurrentcollectioncapabilities.)

Furthermore,HPMSwouldnotbecapableofprovidingtheneededdata.

Page 49: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

43

TheAADTneedsforSafetyAnalystwouldbecurrent‐yearvolumes–theparticularyearbeingstudiedbyVTrans–andnotnecessarilyfutureyearforecasts.IfthestatewidewereusedtosupplementTrafficResearch’sAADTdatawithanalysis‐yearestimates,implementationofSafetyAnalystmaybemorefeasible,andresultinmoreaccurateestimates.Twoissueswiththestatewidemodelwouldneedtobeaddressedpriortoitsimplementationinthiscase:

Themodelprovidesbaseyearvolumesandfutureyearforecasts.Thecurrentbaseyearis2000,andthefutureyearis2020,whichwouldlikelybeupdatedto2010and2030,respectively,ifthemodelwerekeptactive.Inshort,thestatewidemodel’sanalysisyearsarenotreadilyadjustableduetoitsnatureasatraveldemandforecastingtool.However,itislikelythatVTranswouldneedAADTestimatesforSafetyAnalystforananalysisyearbetweenthestatewidemodel’sbaseyearandforecastyear.Asasolution,aprocedureforinterpolationbetweenbaseyearvolumesandfutureyearvolumeswouldbeneeded,toarriveatvolumeestimatesfortheanalysisyearbeingevaluatedinSafetyAnalyst.

Thedirectuseofvolumeestimatesfromthestatewidemodel,insteadofageneralcomparisonofforecastedvolumeestimatestothebaseconditions,raisestheissueofthemodel’saccuracy.ThisissuewaspreviouslyraisedduringthediscussionwithTrafficResearchinNovember2008.ThiswasalsoidentifiedbyaformerVTransmodeler.Updatedcalibrationandvalidationofthemodelwouldlikelyberequiredtoensurethemodel’saccuracy.Itisalsoprobablethatadditionalrefinementofthemodel’svolumeestimatescouldalsobenecessarybeforebeingusedinaprogramsuchasSafetyAnalyst.Thequestionthenbecomes:isthiseffortworththevolumeestimatesfromthemodelforuseinSafetyAnalyst,orwoulditbemorefeasibletouseamorebasic,butmoreeasilyestimatedandchecked,proceduretoestimatetheneededAADT?

Inadditiontosupplementingvolumedatacollection,alsohavinganinventoryofbaseyearconditionsformobilitymeasures,suchascapacities,v/candtravelspeedscouldbehelpfulforprojectevaluation.Inthisregard,themodelwouldnotbereliedonforforecastedvolumesforfutureyears,butinsteadasatooltobetterdefinebaseyearhighwayconditionsinVermont.Thisobjectivemaynotbeanappropriateapplicationofthecurrentmodel,sinceitismeantasaforecastingtoolforcomparisonoffuturevolumestobaseyearvolumes,orcomparisonofplanningalternatives.However,thepossibilityofemployingthemodelasadataresourceforsafetyanalysesandperformancemeasuredeterminationswouldbenefitVTransdataneedsandobjectiveswhilekeepingthemodelrelevant.

Otheruses:thestatewidemodelisgenerallynotapplicablefortrafficoperationsandproject‐levelanalyses.Thisisthecaseformostmacroscopicstatewidemodels,notjustVermont’s.However,althoughindividualprojectsarenotevaluatedusingthestatewidemodel,theneteffectofdevelopmentandgrowthduetonumerousprojectsinanareacouldbequantifiedusingthemodel(e.g.St.AlbansRoute7corridor).Thisprocedurecouldprovideguidanceinthereviewprocessofdevelopmentprojects.

Page 50: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

44

Otheruses:therearenomajorroadwayprojectsplannedorunderanalysisthatwouldrequirethestatewidemodel.ThemodelhasbeenrecentlyemployedfortheCirc‐WillistonEIS,andcurrentlyfortheWesternCorridorStudy,buttherearenoclearusesformajorroadwayplanninginthenearfuture.

SummaryComments–TheStatewideModelasaDataResource

AcommonthreadinthediscussionswithBartandBrucewastheneedformoretrafficdata,bettertrafficdata.Certainlythecostsandresourcesrequiredtogathertransportationsystemdatainthefieldaresignificant.Itissimplynotpracticaltogatherabroadsetofdatatypes(volumes,speeds,delays,geometryandcontrol,etc.)atalllocations,andcontinuouslyfortrafficconditions.

NCHRPReport446,AGuidebookforPerformance‐BasedTransportationPlanning(TransportationResearchBoard–NationalResearchCouncil,2000)identifiesurbanandstatewidetraveldemandmodelsasusefultoolstosupplementfield‐collecteddataforperformance‐basedevaluationsandplanning.Importantly,however,thereportidentifiestheneedformodelcalibrationwithsurvey/fielddataonaregularbasisinorderforthemodelstoprovideaccurateestimatesofcurrent/baseyeardata.

SoftwareemployedbyVTransfortransportationnetwork/systemevaluationsthatrequireextensivedatainventoriesincludeSafetyAnalyst,asmentionedbyBruce,andHighwayEconomicRequirementsSystem–State(HERS–ST),whichisanassetmanagementtoolprovidedbytheFHWA(http://www.fhwa.dot.gov/infrastructure/asstmgmt/hersindex.cfm).TheRoadwayAssetManagementUnitcitessuchdatainventoriesforSafetyAnalystandHER–STaspartofitsmission.(http://www.aot.state.vt.us/progdev/Sections/Design/Design.htm).Furthermore,theVermontHighwaySystemPolicyPlan(Section5.0pages5‐6and5‐7,andAppendixB)recommendsanevaluationofthefeasibilityofexpandingtheHighwayPerformanceMonitoringSystem(HPMS)datasettomeettheneedsofHERS–STinVermont.

TheOregonDepartmentofTransportationandtheTexasTransportationInstitutehaveworkedjointlytoexploretheuseofHERS–STinassetmanagementprograms,evaluatingitsdataneedsandpotentialforimplementationwiththeODOTstatewidemodel.Theirworkhasidentifiedtheneedformorerobustdatasources,whichcanbeaconsiderablechallengewhenapplyingtheHERS‐STmodel(http://mobility.tamu.edu/resources/odot_op_perf_measures.pdf;http://www.fhwa.dot.gov/infrastructure/asstmgmt/dior09.cfm).

ThestatewidemodelcouldbecomeavaluabletoolforProgramDevelopment,specificallytheRoadwayAssetManagementandTrafficandSafetygroups,toprovidedataforthedeterminationofperformancemeasures.Importantly,inorderforthestatewidemodeltobeeffectiveinthismannerandtoensurethatitsvolumeestimatessufficientlyaccurate,itwouldrequirefrequent,regularupdatestoitsbaseyeardataandcarefulcalibrationandvalidation.

Page 51: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

45

A‐6EmailfromStephenLawe,RSGInc.Date:November11,2008ThefollowingisanemailfromStephenLawe,answeringquestionsabouthisexperienceandfeelingsaboutthetraveldemandmodelingsoftwarepackagescommonlyusedatRSG:

WhichtraveldemandmodelingsoftwarepackagesdoesRSGfrequentlyuse?

Fortraveldemandmodelsweusemostifnotallofthecurrentlyavailablepackages.Thelistbelowincludesnewsoftwareandsomeoldersoftwareaswell:TransCAD,CUBE,VISUM,QRS,TModel,TRANPLAN,TRANSIMS,EMME/2,aswellasapackagecalledITM(integratedtravelmodel)whichIwrote.Wealsouseseveralmicro‐simulationpackagesbutIwillnotbotherwiththosehere.

ApproximatelyhowmanyyearshasRSGusedeach?

NowIwishIhadmadeashorterlist:

TransCAD=sincetheearly1990’s.WewereabetatesterforoneoftheveryfirstversionsofTransCAD.

CUBE=sincethelate1990’s(notethatthisstartedwithTRIPS,TP+,Voyager–allofwhichwewereusingaroundthattime).

VISUM=around2005andwedon’thavesignificantexperiencewithVISUM. QRS=wehavebeenusingthisonandoffforyearswhenourclientsrequestit.Probably

mid1990’swasoneofthefirstuses.Wehadarecentprojectin2004whereweusedQRS.

TModel2=thisisnowlargelyoutofservice.Weuseditareasonableamountstartinginthelate1980’s

TRANPLAN=wehaveastatewidemodelforFloridarunninginTRANPLANrightnow.Weknowthedeveloperofthispackagewellsowecanmodifythecodetomeetspecificneeds.Weprobablyfirstusedthisintheearly1980’s.

TRANSIMS=wefirstusedthisin2006–itisstillreallyindevelopmentandnottobetakenlightlyifbeingconsideredforimplementation.

EMME/2=thisisanotherpackageweusedalongtimeagoforalittlewhile.IknowtherearestillversionsouttherebutIthinkthisisalsobeingphasedout.Weprobablyuseditinthelate1980’s.

ITM=thisisapackageIfirstwrotein1992andplayedwithovertheyears.Iwroteitwhenwewerenotsatisfiedwiththecapabilitiesofotheravailablepackagesandwewantedcontrolofthecodebase.Ihavenottouchedthisinabout3years.

Ofthosepackagesbeingused,hastherebeenashiftinsoftwarepackagepreferenceatRSG?

WhenourclientsallowustochooseweprimarilychooseeitherTransCADorCUBE.IalsolikeVISUMbutwehavehadlessopportunitytouseit.Keepinmindthatmanystateshavealreadymadetheirchoiceofsoftwaresowearesomewhatlimitedinourabilitytochoose.IliketheGIScapabilitiesinTransCAD.Itsmostfrustratingfeatureishowithandlestransitnetworksandalsohowsomeofthematricesareexposedtotheprogramminginterface(GISDK).CUBEisslightlymoreutilitarianbutisalsoagoodpackage.Iliketheinterfaceandtheintegrationwithgeodatabases.Themostfrustratingfeaturetomeistheassumedrequiredloopingintheprogramminginterface.

Page 52: Vermont Statewide Travel Demand Model – A Preliminary Evaluation

VermontStatewideTravelDemandModel–APreliminaryEvaluation

UVMTRCReport#10‐007

46

Havenew/inexperiencedmodelersatRSG(ifany)masteredcertainsoftwarepackagesmorequicklythanothers?Oristherenodiscernibledifferenceinthesoftwares'learningcurves?

Idon’tthinkIhavenoticedabigdifference.Onceyouunderstandwhatthealgorithmsaredoing,pickingupanewsoftwarepackageisreasonablytrivial.Iwouldcertainlynotcallthepackagesintuitivetoanon‐modelerbutIalwaysfeelthatit’stheconceptsratherthanthesoftwareitselfwhichisthelimitingfactor.Infact,thesoftwareoftengivespeopleasensethattheyknowmoreaboutmodelingthantheyactuallydowhichcanbedangerous.

Haveyouhadsufficientsupportfromthesoftwarevendors/developerswhenneeded?

Notreally.Ithinkthisisbecausewearenotaskingtrivialquestions.Usually,weareexposingabuginthesoftwareoraskinghowtodosomethingthatisratheruncommon.ItoftentakessometimetohearbackeventhoughIknowmostofthetechpeoplereasonablywell.Havingsaidthis,Idon’tknowthatthisexperienceisreallyapplicabletopeopleaskingmorecommonquestions.

Whattypesofprojects/useshaveyouemployedvariousTDMmodelingsoftwarepackagesfor?

Well,thereareseveralwaystoanswerthis.Geographically,wehavemodeledareasassmallaspartsofatownandaslargeasthestatesofFlorida,GeorgiaandUtah.Substantively,wehavemodeledroadwayinfrastructure,transitplanning,environmentalissues(airquality,globalclimatechange,waterquality,etc.),impactsonlanduse,environmentaljusticeissues,congestionpricing,andMPOandStatelongrangeplanningexercisesjusttogettheliststarted.