A report by the University of Vermont Transportation Research Center Vermont Statewide Travel Demand Model A Preliminary Evaluation Report #10-007 May 2010
VermontStatewideTravelDemandModel APreliminaryEvaluation VermontAgencyofTransportation May2009 Preparedby: AndrewWeeks TransportationResearchCenter FarrellHall 210ColchesterAvenue Burlington,VT05405 Phone:(802)656 1312 Website:www.uvm.edu/trc
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 Acknowledgements TheProjectTeamwouldliketothanktheVermontAgencyofTransportation,theChittenden CountyMPO,andStephenLaweatResourceSystemsGroup,Inc.fortheirtimeandinsights duringthepreparationofthisreport. Disclaimer Thecontentsofthisreportreflecttheviewsoftheauthors,whoareresponsibleforthefactsand theaccuracyofthedatapresentedherein.thecontentsdonotnecessarilyreflecttheofficial vieworpoliciesoftheuniversityofvermonttransportationresearchcenter.thisreportdoes notconstituteastandard,specification,orregulation. ii
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 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 iii
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 ListofTables TABLE1:STATUSOFSTATEWIDEMODELINGCAPABILITY SPRING2005... 5 TABLE2:OPTIONSSUMMARY... 19 TABLE3:MODELSTRENGTHS&WEAKNESSWITHRELEVANTOPTIONS... 25 TABLE4:SUMMARYOFCUBE,TRANSCAD,ANDVISUMFEATURESANDFUNCTIONALITY... 28 ListofFigures FIGURE1:DAILYPERSONTRIPRATESPERHOUSEHOLDBYPURPOSE... 9 FIGURE2:TRIPLENGTHFREQUENCYDISTRIBUTION... 10 FIGURE3:MODELTRAFFICANALYSISZONES... 11 iv
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 ExecutiveSummary TheUniversityofVermontTransportationResearchCenter,workingwiththeVermontAgencyof Transportation,isconductinganevaluationoftheexistingstatewidetraveldemandmodeland toidentifypotentialimprovementsandrefinementstothemodelbasedonplanningpractices andneeds.themodel scurrentsoftwareplatform,cube/voyager,isalsoexplored,witha comparisontotwootherwidely usedmodelingsoftwarepackages.theevaluationalsoincludes aliteraturereviewofstatewidetraveldemandmodelingpracticesinotherstates,including generalmodelstructure,uses,andmodeloperationandmaintenance,aswellasadiscussionof emergingtrendsintraveldemandmodeling. TheVermontAgencyofTransportationhasoperatedastatewidemodelingprogramfornearly fifteenyears,beginninginthemid 1990susingTRANPLANtransportationplanningsoftwareand GISmethods,tothecurrentmodelinCube/Voyager.Thestatewidemodel,whichcoversthe entiregeographicareaofvermontandthousandsofmilesofroadway,hasservedandcontinues toserveasavaluabletoolthatprovidestravelactivitydataprojectionsfortransportation planningstudiesandanalysesthroughoutthestate.thedataprovidedbythemodelisusedby plannersanddecisionmakerssothatsoundtransportationpoliciescanbedeveloped. Withthedevelopmentandmanagementoftransportationsystemscontinuingasavitalroleof thevermontagencyoftransportationandotherinvestedstateagencies,includingtheplanning ofhighways,transit,freightandevaluationsoftheirimpactonsocietyandtheenvironment,the statewidemodelingprogramwillhaveongoingutilityandisavaluabletoolforvtrans.the statewidemodeldoeshavelimitations,however,infunctionalityanddegreeofutilityand reliabilityforpotentialprojectneeds.importantly,themodelisnotcurrentlycapableof modelingtransitorrailnetworks,sinceitsprimarytransportationsystemislimitedtohighways andotherroadways. DependingontheidentifiedplanningandpolicyneedsoftheVermontAgencyofTransportation andotherparticipatingstateagenciesthatwouldrequireuseofthestatewidemodel,themodel couldbenefitfromthefollowingimprovementsandrefinementsinthecomingyears: Developmentofatransitnetwork(bus,rail,orboth)andafreightrailnetwork. Developmentofanewmodechoicemoduletocomplementanewtransitnetwork. Algorithmsfordisaggregationofvehiclemilesoftravel(byvehicleclass,averagespeed, roadwaygrade,etc.)forairqualityanalyses. Feedbackloop(s)betweenmodelingsteps,suchasbetweentrafficassignmentandtrip distribution. Updatesofinputdata,inclusionofroadwaygradedata,andfurthercalibrationand validationofmodelalgorithms,assumptions,andoutputs. Developmentofanactualyear2030scenariotocover20 yearforecastsfromyear2010. AdegreeofintegrationwiththeChittendenCountyMPO sregionalmodel. 1
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 Ataminimum,thereisenoughcausefortheVermontStatewideModeltobemaintainedand periodicallyupdatedtokeepitviableandavailableforplanningneedsatvtrans.theoverall modelcouldbekeptinitscurrentformwithoutsubstantialimprovements,suchastransitand railnetworks,butshouldremainavailableforforecastsofroadwaytravelandvmtestimates. Mostimportantly,thesuccessandlongevityofthestatewidemodelgreatlydependsonthe clearly definedproceduresandgoalsofitsutilitybyvtransandothervermontagencies.as statedpreviously,thegoalsofastatewidemodelingprogramshouldbedefinedwellinadvance ofmodel specificdetails,suchasdataneeds,modelcomponents,computersoftware,and budgetallocation. 2
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 1. Introduction TheUniversityofVermontTransportationResearchCenter,workingwiththeVermontAgencyof Transportation(VTrans),isconductinganevaluationoftheVermontstatewidetraveldemand model sutility,components,andcurrentsoftwareplatform.thisreportrepresentsthefirst workingdraftoftheevaluationanditspreliminaryfindings.thegoalsoftheevaluationareto: IdentifythecurrentandpotentialusesforthemodelbasedonVTransplanningpractices andneeds. Recommendupdatestothemodeltomeetfutureimplementation. Compareitsexistingsoftwareplatformwithtwootherwidely usedsoftwarepackages currentlyavailable. Alsoprovidedinthisreportarealiteraturereviewofstatewidetraveldemandmodeling practicesinotherstates,includinggeneralmodelstructure,operation,andmaintenance,anda discussionofemergingtrendsintraveldemandmodeling. 2. Background TheVermontAgencyofTransportationisresponsiblefortheplanning,construction, maintenance,andpolicygoverningthetransportationsysteminthestateofvermont.the transportationsystemincludespassengertravelbymodessuchasmotorvehicles,transit,air, andnon motorizedtravel,aswellasthemobilityoffreight.thepolicyandplanningdivisionof VTrans,specifically,developsandoverseesthecomprehensivetransportationplanningstrategies andpolicyframeworkinthestate,andworkswithlocalmunicipalities,theelevenregional planningcommissions(rpc),thechittendencountymetropolitanplanningorganization (CCMPO),andotherstateagenciestoensurethatthestate stransportationsystemispartofa responsibleandefficientsocial,economical,andecologicaldevelopmentpolicyplan. PartoftheVTransplanningresponsibilitiesincludestheforecastingoffuturetraveldemandand travelpatternsonthetransportationsystem.thepurposeoftheforecastsistohelpguidepolicy andfundingdecisions,suchastheconstructionofanewroadway,thecreationofanewbus transitline,oralandusedevelopmentschemetomakemostefficientuseofthein place transportationinfrastructure.inordertoperformtheforecasts,andalsotogainbetter understandingofthecurrenttransportationsystem,vtransreliesontransportationplanning toolsforanumberofitsplanningtasks.thesetransportationtoolsrangefromverysmall scale, detailedmodels,idealfortaskssuchasroadwaycapacityanalysesandmanagementoftraffic operations,toalarge scale,statewideplanningmodel,idealfordeterminingthetraveldemand loadsontheentiremulti modaltransportationsystem. 3
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 Foritsplanningresponsibilities,VTransreliesonthestatewidemodeltoevaluatetheeffectsof transportationprojectsandtoguidefuturedevelopmentpoliciesforsectionsofthe transportationsystem.thestatewidemodel identifiedasatraveldemandmodel coversthe entirestateofvermontandincorporatesdemographicandeconomicdataofvermont s households,employmentestimates,andthecharacteristicsofthetransportationsystemitself, includingthetravelmodeopportunitiesandthestate sroadwaynetworkintoastructuredfoursteptransportationplanningmodel.essentially,thepassengertravelmodelestimatesdemand foruseofthetransportationsystembasedoncharacteristicsofthetransportationnetworkand itspotentialusers,andreportsmeasurabledegreesofuse,includingroadwayvolumes,milesof travelonroadways,andtransitridership.thestatewidemodelalsoincludesafreightdemand modelthatestimatestruckusageofthestate sroadwaynetworkforthemovementofgoods. 3. ReviewofModelingPractices 3.1 StatewideModelingatOtherDepartmentsofTransportation Currently,approximatelyone halfofthestatesintheunitedstateshavefunctionalstatewide models(1).thestructure,utility,andcostsofthenumerousstatewidemodelsvaryandgreatly dependontheneedsoftheirhostdepartmentsoftransportation.publisheddocumentssuchas NCHRPSynthesis358:StatewideTravelForecastingModels(1)andStatewideTravelDemand Modeling:APeerExchange(2)provideinformationaboutcurrentmodelingpracticesinstates throughoutthecountrybasedonsurveyresponsesfromstatedepartmentsoftransportation, includingvermont.table1,takenfromnchrpsynthesis358,page14,presentsasummaryof statewidemodelingactivityasof2005. AscitedinNCHRPSynthesis358,statespredominantlyusetraditionalfour stepproceduresat thecoreoftheirpassengertraveldemandmodels,similartourbanorregionalmodels.trip generationisgenerallyperformedbycross classificationtechniques,tripratesandlinear equations,ortour basedmultinomiallogitexpressions(1).themostcommonlyused demographicandtraveldatasources,suchastheuscensus,thecensustransportationplanning Package(CTPP),theNationalHouseholdTravelSurvey(NHTS),andlocalMPOinformation providereadilyavailabledataforusebystatesintheirmodels,andareusedsolelyor supplementedbystate collecteddata.furthermore,inmoststateswithfunctionalmodels, gravitymodelapplicationisusedfortripdistribution,whiletherearesomeinstancesofgrowth factormethods.modechoiceistypicallyaccomplishedbylogitexpressionsorfixedpercentage shares,andtrafficassignmentisperformedusingestimationsofequilibriumconditions(1). Moreover,NCHRPSynthesis358findsthatstatedepartmentsoftransportationpredominantly utilizetheirmodelsforcorridorplanning,systemplanning,bypassstudies,regionalplanning whilecollaboratingwithanmpo,andproject levelforecasts,suchasforenvironmentalimpact purposes.approximatelyone thirdofstateswithfunctionalmodelsapplytheirmodelsforair qualityanalyses,freightplanning,andeconomicdevelopmentandlong terminvestmentstudies. 4
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 Thereportalsoindicatesthatstateswithlongerhistoriesofmodelingandgreaterconfidencein thevalidityofthemodelingoutputstendtomorewidelyusetheirmodelsforabroadrangeof functions. TABLE1:STATUSOFSTATEWIDEMODELINGCAPABILITY SPRING2005 State Model Development Cost Condition Time(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 Individualislandmodels Idaho Dormant Illinois Dormant Indiana Operational $1,500,000 3 7moreyearsforvariousupgrades Iowa Developing $300,000 2 Kansas Developing Hasadormantfreightcomponent Kentucky Operational $370,000 2 Newmodelunderdevelopment Louisiana Operational $500,000 Costincludessomeapplications Maine Operational $500,000 5 Beingrevised Maryland None Massachusetts Revising $800,000 Michigan Operational $1,000,000 2 Minnesota Partial Mississippi Developing Missouri Operational $500,000 Revisioncompletionsoon Montana Operational Freightonly Nebraska Dormant Baseyearmodel Nevada None NewHampshire Revising $2,000,000 NewJersey Operational $500,000 Freightonly NewMexico None NewYork None County levelodassignment NorthCarolina None NorthDakota None Ohio Operational $6,000,000 8 Beingrevised;$3,500,000fordata Oklahoma None Oregon Operational Beingrevised Pennsylvania Developing RhodeIsland MPOmodel SouthCarolina Operational $25,000 0.5 SouthDakota None Feasibilitystudybeingconducted Tennessee Developing BasedonODtableestimation Texas 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. 5
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 Accordingly,theNCHRPSynthesis358reportsthatthemostcommonmeasuresofeffectiveness (MOE)usedfrommodelsasindicatorsofsystemperformanceincludevehiclemilesoftravel (VMT),vehiclehoursoftravel(VHT),volumetocapacityratios(v/c),levelsofcongestion,and trafficgrowthrates.additionaloutputsmaybeusedformorespecializedmodeluses,butvmt andvhtarestandard.somemodelsproducemoesforcertainperiodsoftheday,namelypeak hours,whileothersreportdailytotals. Themaintenanceandresourceneedsfortherangeofstatewidemodelsarealsovaried.While manystatedepartmentsoftransportationrelyonprofessionalconsultingfirmstodeveloptheir models,manyperformroutinemaintenancethemselves(1).staffallocationsformodeling programstypicallyrangefromoneormorepersonnelwithpart timeresponsibilities some supplementedbyvaryinglevelsofconsultantsupport uptomultiplefull timepersonnel. Moreover,maintenanceisperformedcontinuouslyoratfrequentintervals(onetotwoyears), andstatestypicallyperformmodelupdatesonregularcycles somecoincidingwithreissuingof transportationplans(typicallyeveryfiveyears),andothersusingcyclesaslongastenyears. SurveyfeedbackreportedinStatewideTravelDemandModeling:APeerExchangeprovides insightintosomereasonsformodelingprogramfailures.accordingtorickdonnellyofparsons Brinckerhoff,whosesurveyresponsewasquotedintheStatewideTravelDemandModeling:A PeerExchangefortheOregonstatewidemodel, vagueorpoorlydefinedgoalsandobjectives, higherthanexpectedmaintenanceandapplicationcosts, and lackofmanagementsupport aresomeofthecausesformodelfailureandabandonment(p.74).mostimportantly,acommon andreasonabletrendinstatewidemodelingprogramsisthatstatesmustbalancemodelquality withtimeandbudgetconstraints(1). Inordertoensurethatastatewidemodelingprogramiseffectiveandsustainable,bothinterms oflaborandbudgetallocation,itsobjectivesmustbeclearlydefinedanditsutilitymustbe transparent.asstatedin ACriticalReviewofStatewideTravelForecastingPractice (3),the goalsofastatewidemodelingprogramshouldbedefinedwellinadvanceofmodel specific details,suchasdataneeds,modelcomponents,computersoftware,andbudgetallocation.the planningneedsofthestatedepartmentoftransportationshouldsteerthemodelingprogram. 3.2 AdvancesinTravelDemandModeling Recentresearchinthesubjectoftraveldemandforecastinghasseenashiftfromtraditionaltripbasedmodelingtechniquesofthefour stepparadigmtoactivity basedforecastingmodels.this hasbeeningeneralresponsetothelimitationsofthefour stepprocess,theneedformodelsto bemoresensitiveandresponsivetotheeffectsofsystemconditionsandpoliciesonmobilityand lifestylechoices,andthedesiretodisaggregatetraveldemandforecastingtothelevelofmicrosimulationmodelingtechniques,replacingaggregatezonalcalculations(4,5,6,7,8).since activity basedmodelingconsidersthedecisionprocessesthatdictatetravelofindividualentities 6
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 7 or agents, potentiallyonacontinuoustemporalbasis,itsmethodologyadaptsitselfbetterto thedynamicnatureandcomputationalcharacteristicsofmicro simulation.therecould, therefore,beashiftawayfromindividual trips astravelunitsto tours withmoreadvanced activity basedmodelsimplementedatregionalandstatewidelevels. Inadditiontoservingthestandardforecastingneedsoftransportationplanning,thenew activity basedmodelshavealsoshowntobevaluabletoolsforevaluatingmorespecific transportationsystemmanagementpolicies,including:traveldemandmanagementandpeak spreadingeffects,high occupancyvehicle/facilitydemand,tollandcongestionpricing,and transitfareandparkingpricingpolicies(5,7).theycanalsobeappliedtoplanningissuessuchas transportationequityandchangesindemographics,suchaspopulationaging,household compositions,andlaborforcecharacteristics(5,7). Thedatarequirementsfordevelopment,computationaldemands,validationprocesses, variabilityofresults,andcostsversusbenefitsofactivity basedmodeling,thoughimplemented andevaluatedonlimiteduptoregionalscales,haveyethadwidespreadapplicationona statewidescale.somestateswithneworrecentlyrevisedstatewidemodelshaveimplemented varyingdegreesofactivity basedtraveldemandprocedures,includingohio,oregon,new Hampshire,Michigan,andLouisiana,withOhio s(stillindevelopment)beingthemostadvanced. VariousMPOsnationally,suchastheMid OhioRegionalPlanningCommission(MORPC),theNew YorkMetropolitanTransportationCouncil(NYMTC),andtheSanFranciscoCountyTransportation Authority(SFCTA),alsohaveimplementedactivity basedproceduresintheirregionalmodels. However,manytraveldemandpractitionersarestillskepticaloftheaccuracyandlarge scale applicabilityofthenewmodelingtechniques,aswellastheirdataandresourceneeds. Dependingonthedesiredutilityandnecessarydetailforastatewidetraveldemandmodelat thistime,aswellasimplementationconstraints,activity basedmodelingmaynotyethavea feasibleapplicationforallstatescurrentlywithtraveldemandmodels. 3.3 OngoingApplicationsofTransportationModeling AstudypanelfortheFloridaDepartmentofTransportation,aspartofitsevaluationof transportationmodels,providedaneffectivelistof currentandemergingissuesin transportationplanning thatsummarizesabroadspectrumofpotentialapplicationsfortravel demandmodelsandothertransportationmodels(9).thelistincludes: Capacitydeficienciesandcongestion Interactionbetweentransportationandlanduse Economicdevelopmentimpactsoftransportation Freightmobilitystrategies Airquality Systempreservation,maintenance,andoperations Safety Securityandemergencyevacuations
Equity Resourceallocationandprojectselection VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 TheseplanningchallengeshaveuniversalcontextsandarealsoapplicabletoVermont s transportationsystems,developedspace,andenvironment.therefore,itisreasonabletobe cognizantofthesetopicswhilediscussingthecurrentandfuturefunctionalityandutilityof Vermont sstatewidemodel. 4. StatewideTravelDemandModelingattheVermontAgencyof Transportation TheVermontAgencyofTransportationcurrentlyhasafunctionalstatewidetraveldemand forecastingmodel.in2007,themodelwastranslatedfromtheoriginaltranplansysteminto thenewcube/voyager(citilabs,inc.)softwareplatformbyvhb,inc.withupdatestothe roadwaynetwork,freightmodel,andmodelcalibrationproceduresandresults(for documentationaboutthisprocessrefertovermontstatewidetraveldemandmodel Improvements:UpdatedPassengerandTruckModelsinCube/Voyager,2007).Importantly, duringthisprocessemploymentdatafortripgeneration,truckpercentagesbyregionalplanning commission(rpc)fornonhome basedtrips,andthefrictionfactorfunctionsfortripdistribution wereallupdated(10).althoughthemodelhasbeenappliedtoanumberofproject leveltasks, suchasthecirc WillistonEnvironmentalImpactStatementandcurrentlytheWesternCorridor TransportationManagementPlan,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,themodeltakesinputsfromdemographicsourcesand combinedwithtriprates,estimatesthenumberofpersontripsgeneratedperday,categorized intosixgeneralizedpurposes:home basedwork,home basedshopping,home basedschool, home basedother,nonhome based,andtruck(asapercentageofnonhome based).for example,figure1showsthetripproductionratesforthefourhome basedpurposesbasedon thecross classificationofpersonsperhouseholdandautosperhousehold;nonhome basedtrip productionsarebasedonalinearapproximationinsteadofcross classification,similartotrip attractions.inthetripdistributionstep,theflowsoftripsbetweeneachorigin/destinationpair oftazareestimatedbasedonthezone to zonetraveltimes(impedance).figure2showsthe triplength(time)frequenciesbypurpose,whichareoneresultofthetripdistributionprocess. Themodechoicestepsplitsthetotalnumberofpersontripsintothoseusingtheroadway networkandthoseusingtransit;inturn,theroadwaypersontripsareconvertedtovehicletrips usingvehicleoccupancyfactorsbytrippurpose.lastly,thevehicletripsareassignedtothe 8
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 roadwaynetwork,producingtrafficvolumeandvmtestimatesoneachroadway.becausethe statewidemodelcurrentlydoesnothaveatransitnetwork,thepersontransittripsestimatedin themodechoicesteparenotassigned. FIGURE1:DAILYPERSONTRIPRATESPERHOUSEHOLDBYPURPOSE 6 5 4 3 2 1 0 Home basedwork Home basedother Home basedschool Home basedshopping 9
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 FIGURE2:TRIPLENGTHFREQUENCYDISTRIBUTION 10.0% 9.0% 8.0% 7.0% 6.0% 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% 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 Themodelhasa2000baseyearwithaforecastyearof2020andaten yearinterpolationto 2030.Itcontains628internaland70externaltrafficanalysiszones(TAZ)asshowninFigure3. TheexternalTAZrepresenttravelto/fromandbetweenregionsbeyondVermont,including neighboringstates(newyork,newhampshire,andmassachusetts),therestofnewenglandand thenortheast,andcanada.generally,theinternaltazcoincidewithvermontmunicipality boundaries,exceptinmoredenselypopulatedareaswherethetazaremoredisaggregate.the TAZalsoaggregateto2000USCensusblockgroupboundaries. 10
wn t d d d l e w v o n tc oen w et o wm v ot de t S wd R77f 37 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 Mds wk n 9 WWM kdn k adw WC,/V ws w 9 nk wwk k,j V ws w 9 r wn r 9 wk w 9,LV ws w 9 A ndw wk w 9,GV ws w 9 A 2dw d99 kdw,i V ws w 9 A ndw d99 kdw //
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),the MorrisvilleBypass,theBenningtonBypass(14),theWesternCorridorTransportation ManagementPlan(15),andastateemployeecommuterstudy.Themodelwouldalsohave applicabilityforthehighwaysystempolicyplan(16),publictransportationpolicyplan(17),rail System&PolicyPlan(18),theAssetManagementVisionandWorkPlan(19),theVermontLong RangeTransportationBusinessPlan(LRTBP)(20,21) thestate soverallplanforthemulti modal transportationsystem andthefinalreportandrecommendationsofthegovernor s CommissiononClimateChange(GCCC)(22). Theusefulnessofthemodelforproject levelanalyses,planningscenarioevaluations,and storageandpresentationoftransportationsystemdatareinforcetheneedforcontinued maintenanceandapplicationofthetool,aswellasoutreachandtrainingtoinformothervtrans departmentsandotherstateagenciesofthemodel sutility.thefollowingsectionshighlight anticipatedmodelingneedsbasedonfutureplanninggoalsandtheimprovementslikelyrequired orrecommendedforthestatewidemodelinordertomeetthosespecificneeds,especiallythose thatarenotcurrentlywithinitscapabilities. Dependingonthedeterminationofthelong rangemodelingobjectivesofvtransandthe statewidemodel sfutureutility,itispossiblethatthestatewidemodelwouldneedadditional refinementandupdates.ifsuchrefinementsandupdatesweretobeimplemented,itisexpected thattheirpurposewouldprimarilybetomeetsomeoftheplanninggoalsidentifiedinthelrtpb andgccc,aswellasothermodelingeffortsatvtrans,includingtransitplanningandairquality analyses.assuch,someofthecontinuedandfuturemodelinggoalsdiscussedcoulddetermine theneedforanumberofpotentialstatewidemodelimprovements. 4.2.1 SystemPolicyandEnvironmentalPolicyPlanning Anumberofapplicationsofthestatewidemodelarelikelybasedonanticipatedmodelingneeds describedinstateandagencydocuments,suchasthelrtbpandtheindividualsystempolicy plans.thesedocumentsdonotexplicitlyidentifytheapplicationofthestatewidemodeltomeet
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 theirrecommendationsbutfromapracticalviewpoint,themodelisclearlyatooltoaid completionofsomeofthepolicycomponents.thelrtbpidentifiestransportationopportunities, specificallytheintegrationoflanduseandtransportationplanningandtheevolvingfocuson corridormanagementplanning,asmethodsforthestatetomoreeffectivelymeettransportation systemchallenges.thefinalreportandrecommendationsofthegccccitestheneedfor strengthenedstate levelplanninganddecision makingtosupportmanagedgrowthaswellas transportationplanningtofacilitatealternativetravelmodes,transitpriorityandthe rehabilitationandmaintenanceofexistinginfrastructure. BusandRailTransitSystemPlanning Inadditiontotheplanninggoalsoutlinedinofficialagencydocuments,discussionswithVTrans personnelunderscorethepotentialutilityofthestatewidemodeltosupportvariedtasksatthe agency(seeappendixa:a 1 A 3,A 5).AtthePolicyandPlanningDivision,thereisarecognized needforimprovedmulti modalplanning,asrecommendedinthelrtbpandmorespecifically thehighwayandpublictransportationsystempolicyplans,whichwouldlikelyincludethe creationofatransitnetworkinthestatewidemodel.afullydevelopedtransitnetworkwould allowplannerstoestimateridershiploadsalongexistingandproposedtransitroutessuchas intercitybustransit,toevaluateheadwaysandstops,andtoidentifypotentialroutecapacity improvementsandextensions.importantly,anyimprovementinthecapabilitiesanddetailofthe model stransitcomponentwouldrequireredevelopmentofthemodechoicemodule,sincethe modelcurrentlyapportionsafixedpercentageasatransitshare(10).anewmodechoice modulewouldlikelybealogitmodel,whichpredictstheselectionprobabilityofalternative modesbycomparingthecosts(monetarycost,time,convenience,etc.)ofthosemodes(11,12). Thedevelopmentofatransitnetworkinthemodelwouldbeanintensivebutprobably worthwhiletaskconsideringtheimportanceoftransitmodelingrecognizedinlongrange planningdocuments.additionofatransitnetworkwouldrepresentasignificantimprovementto themodel sfunctionalityandutility.thistransitnetworkwouldcertainlyincludebus,which wouldusethemodel sroadwaynetwork,andpossiblyrail,whichwouldrequirethe developmentofanewnetworkstructure. RailFreightSystemPlanning Developmentofarailnetworkaspartofthefreightmodelwouldimprovethefreightmodel componentaswellastheoverallmodel sutility.agrowingneedordesiretoshiftmorefreight fromtrucktorailcouldstillprecipitateaneedforarailcomponentinthestatewidemodel.the VermontStateRail&PolicyPlanidentifiesagoaltoprovidecompetitivefreightandpassenger service,eventhoughthatreportalsorecognizestheuncertaintyofthefutureofvermont s intercitypassengerrailservice(18).however,withapassengerrailcomponent,ifdeemed worthwhile,thestatewidemodelcouldbeusedtoevaluateplanstoexpandandpromote passengerrailservicebothinvermontandregionally. 13
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 Thegeneralrailnetworktopology linksandnodes oftherailpassengertransitcomponent couldbetransferredfromanewlydevelopedrailfreightcomponent,ifthetransitimprovements inthemodelwereundertakenfirst.therailnetworkstructurecouldthenbespecifically configuredforuseinthefreightmodel. HighwaySystemPlanning TheHighwaySystemPolicyPlanindicatesthatforsystemmanagementandpreservation itis desirabletodefinedifferentsectionsofthehighwaysystembasedonfunctionalityandoverall levelofimportanceforwhichdifferentperformancestandardsandinvestmentpoliciesare developed (p.es 5).Thisideaofsub networksappliestoperformancemeasures,including preservation,safety,mobility,andenvironment/qualityoflifeasoutlinedinthehighwaysystem PolicyPlan,andtotheprioritizationofassets.AsVermontdevelopswithgrowingpopulation centers,thedesignationsofexistingandhithertounbuiltroadwaysaspartsofeitherthe primary or off primary sub networksmayneedtoberevised.forecastsandanalyseswith thestatewidemodelcouldaidinappropriatelyidentifyingtheevolvinglevelofimportanceof roadwaysinthenetworkandevaluatecertainperformancemeasures,suchasvmt,traveltimes, averagespeeds,flows,andvolume to capacityratiosforroadwaylinks.asanextensionofthe sub networkdefinitionprocess,thestatewidemodelcouldservetoevaluateanddeterminethe criticallinksinthesystem,andthepossibleimpactsduetotheclosure(e.g.bridgemaintenance, naturaldisasters,emergencyevacuations,etc.)ofacriticallink.thiswouldprovidedataforthe planningofmanagementandresponsestrategies. ThestatewidemodelcouldalsoservetheTrafficOperationsdepartmentresponsibleforthe management,analysis,andreportingofvtranstrafficcountsthroughoutthestate,byaiding trafficgrowthprojectionsandhelpingtoidentifythepotentialeffectsofnetworkdisruptions andresultingdetours/re routing(seeappendixa:a 1).Howeverfordetailedanalysesof roadwaysectionsorcorridorsusing,forexample,micro simulationmodelswiththestatewide model sprojectedvolumesasinputs,itwouldfirstbeadvisabletoevaluatetheaccuracyand applicabilityofthestatewidemodel strafficvolumesfordirectuseinothermodelinganalyses beforesuchapplicationswereundertaken. EnvironmentalPlanning Theimportanceofenvironmentalpoliciesrelatingtoclimatechangeandairqualityconditions coulddrivetheneedforfurtherstatewidetraveldemandforecastmodeling.althoughvermont iscurrentlyinattainment,airqualityconditionscouldworsenormorestringentstandardscould resultinnon attainmentforcertaincriteriapollutantssuchasozoneorparticulatematter(21). Moreover,potentialfuturelimitsongreenhousegases,namelycarbondioxide,aswellasVMT performancemeasureswouldrequirevmtforecastsandpollutantinventoryestimatesfor Vermont stransportationsystem.thestatewidemodelisavaluabletoolforsuchenvironmental analyses,primarilybecauseitcanfeednecessaryinputdata VMTbyfunctionalroadwayclass andaveragespeeds tomobileemissionsmodels(currentlytheepa smobilemodelandthe 14
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 newermovesmodel).itisreasonabletoexpectthatairqualitymodelingandtheresulting environmentalpolicieswillcontinuetogainimportanceforvermontandnationally,andthe needforincreasinglydetailedandaccuratetraveldatafromthestatewidemodelwouldgrow concurrently. Therearemethodstoimprovethestatewidemodel soutputstobetterservemobileemissions modeling,specificallythemobilemodelwhichiscurrentlyusedinvermont.accurate disaggregationofvmtdatabyvehicleclasscouldimproveemissionsestimates,sincedifferent typesofon roadvehicles,suchaslight dutyversusheavyduty,havedifferentoperating characteristicsandemissionsprofiles.however,atypicalfour steptraveldemandmodellikethe statewidemodeldoesnotinherentlyconsidervehicletypesinthismanner.thestatewidemodel doescurrentlydefineaportionofnonhome basedtripsastrucktripsbasedonapercentage (determinedbyvehicleclassificationcounts)foreachrpcinvermont,butthisdoesnottranslate tospecificvehicle typevmtassignedtothehighwaysystemasamodeloutput.therefore,it wouldbeworthwhiletodevelopanewmeanstoaccuratelydisaggregatehighwayvmtby vehicleclass,atleastintogeneralweightclasses(e.g.light dutyvehicle,light dutytruck,and heavy dutytruck)withinthestatewidemodel.thiscouldinvolveusing24 hourclassification countsandconductingroad sidesurveys. OtherVMTbreak downsthatcouldbeexploredtoimproveemissionsmodelingwouldinclude VMTbytimeofday(hourly)andVMTbyaveragespeed.Importantly,themeritofdeveloping thesedistributionsspecificallyforvermontinsteadofusingthedefaultvalues(national averages)suppliedwiththemobilemodelwouldneedtobeinvestigated.thedefaultvalues maybe goodenough, anditwouldnotbeworththetimeandcosttodevelopvermont specific distributionsthatwouldnotservetoimproveemissionsestimates,orworse,decreasetheir accuracy. GreenhouseGasEmissionsInventoriesandStandards Thecurrentandfutureutilityofthestatewidemodelshouldbeconsideredinthecontextof potentialairqualitystandardschangesandpotentialgreenhousegasstandardslegislationfor Vermontandonafederallevel.InApril2009,theEnvironmentalProtectionAgencyreleasedits findingthatgreenhousegasescontributetoairpollutionthatposesathreattopublichealthand welfare,inresponsetoa2007rulingbythesupremecourtoftheunitedstatesthatrequiredthe EPAtoreviewthepotentialimpactsofemissionsofsixgreenhousegases(24). Concurrently,theStateofCaliforniaistakingaleadingreenhousegaslegislation,includingnew capstandardsforlight dutyvehiclesandrulesforghgreporting/inventories(25,26).vermont,in additiontoapproximatelyfifteenotherstates(27),haspasseditsownlegislationtoadopt California sghgstandards,expectingthattheu.s.epawillgrantcalifornia swaivertotheclean AirActfortheproposedstandards.California swaiverrequestwasfirstrejectedbytheepa duringthebushadministrationinlate2007,buttheepahasrecentlybeendirectedbythe ObamaAdministration,inearly2009,toreviewandreconsiderthatdecision. 15
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 VermontpasseditsownruleinNovember2005,makingitsGHGemissionstandardsforlow emissionvehiclestobeidenticaltothoseofcalifornia.intherule,titledairpollutioncontrol Regulations,SubchapterXI,LowEmissionVehicles RegulationstoControlGreenhouseGas EmissionsfromMotorVehicles,GHGstandardsforsmallandlargelight dutyvehicleswouldbe graduallyphasedinbetweenmodel years2009and2016(28).however,vermontcannotact untilthecaliforniawaiverisgranted. AsVermontadoptsCaliforniastandards,itispossiblethatthestatewillalsoseektofollow California sleadonghgreportingrules.requirementsforinventoriesofghgemissionsforthe transportationsectorwouldlikelyemphasizeregularanddetailedreportingofvmtestimatesfor GHGcontributionsfromon roadsources.also,anevaluationofghgemissionimpactswouldbe desiredforplanningandpolicymakingpurposes(e.g.therecommendationspresentedinthe VermontComprehensiveEnergyPlan,2008,(29)andVermontGovernor scommissiononclimate Change,2007)(22). InMay2009,thedevelopmentanddebateofAmericanCleanEnergyandSecurityActof2009 (H.R.2454)hashighlightedtheproposedbill smulti facetedapproachtoreduceconsumptionof fossilfuelsandlimitgreenhousegasemissions.regardingthelinkbetweenthetransportation sectorandghgemissions,asummaryofthebilldatedmay16,2009describesthe TransportationEfficiencysectionofH.R.2454,whichwouldamendtheCleanAirActto require statestoestablishgoalsforgreenhousegasreductionsfromthetransportationsectorand requiressubmissionoftransportationplanstomeetthosegoalsbymetropolitanplanning Organizationsforareaswithpopulationsexceeding200,000people (30). TheVermontstatewidemodelwouldbethemostsuitabletooltoforecastVMTforthe estimationprocessofgreenhousegasemissionsfromon roadsourcesinthetransportation sectorandtoevaluateghg relatedimpactsofplanningalternativesbasedonthevmt projections.currentlyincalifornia,thecaliforniaairresourcesboard s(carb)travelactivity forecastingapproachfortransportationsectorghgemissioninventoriesusesvmtestimates fromtheemfac2007modelandregionalplanningorganizations.emfac2007estimatestravel activitybasedonvehiclepopulationandvehicle age specifictravelaccumulations,whichare matchedtovmtestimatesfromtheregionalplanningorganizations transportationmodels(31). SimilartoCalifornia,VermontshouldmaintainitplanningtoolsandproceduresforVMT forecasting,includingthestatewidemodel. 4.2.2 OtherModelRefinements Additionalrefinementtaskscouldbeundertakentoimprovethemodel sfunctionality.the accuracyofthemodel sbaseyearandprojectedvolumesshouldbefurtherassessed.for regionalairqualitypurposesandplanningprocesses,theneedforreliableestimatesofvehicle milesoftravel(vmt)wouldbeparamount,andotherusesofthestatewidemodel,suchas analysisofasectionofitsvolumeoutputsusingmoredetailedmodelingtoolssuchastraffic simulation,wouldgreatlydependontheaccuracyofitsestimates.tothisend,thetaskof 16
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 assessingandimprovingthereliabilityofthemodel svolumeprojectionscouldincludefurther refinementofthefourplanningsteps,inputdatamodifications,andadditionaliterationsof calibrationandvalidation. Four StepModelAlgorithms Furtherrefinementofthemodel soutputscouldbeaccomplishedbyincorporatingoneormore feedbackloopswithinthemodel salgorithmicprocess.sincetripdistributionestimationsare basedoninterzonaltraveltimesasarethetravelmodeutilityvalues(traveltimecost,perceived cost,etc.)inthemodechoicestep,itiscommontofeedthelinktraveltimesestimatedinthe trafficassignmentstepbackintothetripdistributionandmodechoicesteps.thisisespecially importantiftrafficassignmentindicatescongestionthatresultsinlinktraveltimessignificantly differentfromthoseinitiallyassumed.theresultingfeedbackwouldproduceaniterativeprocess thatwouldcontinueuntilanequilibriumconditionhadbeenmet. InputDataUpdatesandFurtherCalibration&Validation DuringtheVHB,Inc.updateandmigrationofthestatewidemodeltoCube/Voyagerin2006 2007,thetraveltimefrictionfactorsfortripdistribution,originallybasedonthe1994Vermont householdsurvey,wereupdatedusingthefunctionsdescribedinthenchrpreport365(12). VHB,Inc.alsorevisedthetripattractionregressionequations,butthetripproductionrates, originallybasedonthe1994householdsurvey,werenotchanged.newdemographic,socioeconomic,andtraveldatafromupcomingupdatesinthe2008 2009NationalHouseholdTravel Survey(http://nhts.ornl.gov/nhts2008.shtml)andthe2010USCensuswouldprovideabasisto evaluateandpossiblyrevisecurrenttriprates.furthermore,withthenewdata,itwouldalsobe possibletoevaluatethefeasibilityandmeritofapplyingregional specifictripratesandfriction factorswithinthemodel,forexamplebycountyorrpc.itisimportanttonotethatfollowing VHB,Inc. srecentworkonthemodel,someadditionalcalibrationwasdoneatvtransfortazin westernportionsofvermontforthewesterncorridortransportationmanagementplanand subsequenttothat,cambridgesystematics,inc.performedfurthercalibrationwhilerunningthe modelforthatstudy. RoadwayGradeData Thenetworklinksinthemodelwithfunctionalroadwayclassifications(i.e.allofthelinksexcept thecentroidconnectors)currentlyhaveadefaultvalueof 1 forthe TERRAIN datafield. AccordingtotheVHB,Inc.2007report,adefaultvalueof1,whichindicateslevelterrain,was assignedtothelinksduetolackofbetter,morespecificterraininformation.updatingthe model snetworklinkswithaccurateterraininformation,orpotentiallymorespecificdatasuchas grade,couldserveasaninventoryforvermont sroadsandimprovethemodel sutilityinfuture on roadmobileemissionsandairqualityestimatesusingtheusenvironmentalprotection Agency smovesemissionsmodel,whichconsidersroadwaygrade(aspartofavehicle s operatingmodeforrunningemissions)initsestimates.thestatewidemodelcouldprovidevmt 17
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 totalsbygrade.however,ifdefiningindividuallinkgradesisdeemedimpractical,categorical information,suchasgeneralterraintype,couldbedefinedforthemodel snetworklinksand usedtoapproximatemoredetailedgradevaluesasneeded.terraincategoriescouldinclude level,rolling,andmountainous,similartothe2000highwaycapacitymanual. IntegrationwiththeChittendenCountyMPOModel PossibleintegrationofthestatewidemodelwiththeCCMPO sregionalmodelofchittenden Countycouldfurtherpromoteanalysesusingthemodelsjointly,tocomplementeachother.The CCMPOmodeliscontinuallymaintainedandhasrecentlyundergoneanewseriesofupdates (23).ForcertainprojectsandplanninganalyseswithinChittendenCounty,thestatewidemodel couldbereliedupontoevaluatetransportationeffectsbeyondtheregionalboundaryofthe CCMPOmodel(seeAppendixA:A 4).OnesuchexampleistheCirc WillistonDEIS.Transferability ofnetworkandtazdatafromtheregionalmodeltothestatewidemodel,orviceversawould havedesiredutility.also,byincludingsomeofthesametravelmodesinthestatewidemodelas intheccmpo namelyauto,carpool,bus,andpossiblyrail therecouldbeimproved transferabilityoftripdatabetweenthetwomodels.thiswouldalsorelyonrectifyingthe temporalresolutionofthetwomodels,daily(thestatewidemodel)versuspeakhour(the CCMPOmodel).Also,themodelsarecurrentlyindifferentsoftwareplatforms;thestatewide modelisincube/voyagerandtheccmpomodelisintranscad.ifthatweretoremainthecase fortheforeseeablefuture,ameansofdatatransfer(import/export)wouldbeneeded,suchas withtheuseofdatabasefilesorspreadsheets. 4.2.3 Maintenance Regularmaintenanceofthemodelshouldbedonecontinuously,includingminorcalibrationand validationprocessesorasneededbyspecificprojectsemployingthemodel.moresignificant updatesandrevisionstothemodelincludingupdatestoinputdatasuchasdemographicand employmentfiguresandroadway,transit,andfreightnetworkcharacteristics,andpossiblythe model salgorithmsandlogiccouldbeonascheduledcycleofapproximatelyfivetotenyears. Updateprogramscouldcoincidewiththereleaseofrevisedsystempolicyplansorthe publicationofupdateddatasources,suchasthenationalhouseholdtravelsurveyandthe2010 USCensus.DependingonairqualityattainmentstatusinVermont,airqualityconformityneeds couldpotentiallyrequireaccelerationofthemodel supdateschedule. AccordingtotheLRTBP,theIntermodalSurfaceTransportationEfficiencyAct(ISTEA,1991) stipulatedthatthestatewidetransportationplanmustcoverforecastsforaminimum20 year horizon.thatstipulationhasprogressedintea 21(1998)andsubsequentlyinSAFETEA LU (2005).Currently,thestatewidemodelforecaststo2020and2030,thoughthe2030forecastis basedonextendingthegrowthfrom2000to2020byanadditionaltenyears(10).withtheyear 2010approaching,moreaccurateforecastsfor2030wouldberequiredandshouldbedeveloped tomaintaincompliancewiththeminimum20 yearhorizonruleinthelrtbp. 18
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 5. OptionsfortheStatewideModelandModelingProgram VTranshasanumberofoptionsforthefutureofitsstatewidemodel.VTransmustconsiderthe feasibilityofcontinuingitsmodel,givenanexpectedfuturelevelofutilityofthemodel comparedtothetimeandbudgetresourcestokeepitactiveandcapableofmeetingthe Agency sneeds.beyondthatevaluation,therearelong termdecisionsconcerningthestateof themodelandhowitscurrentfunctionalitydoesordoesnotmeetthedesiredplanning applicationsofthemodel. Thecurrentandenvisionedlevelandextentofmodelusewouldlargelydictatethespecificsofif andhowthemodeliskeptactive.ifthemodelatitscurrentleveloffunctionality,givenperiodic updatestoinputdata,wouldsufficientlyandcost effectivelymeettheplanningneedsofvtrans, thenitwouldbebeneficialforvtranstokeepthemodelactivewithaclearprogramforits updateandapplication.goingfurther,ifthereisadesirewithintheagencyforincreasedand morewidespreaduseofthemodel,perhapsfortrafficresearchapplications,roadwayasset management,performancemeasuredeterminations,andperhapscoordinationofplanningwork withotherplanningorganizationsinvermontsuchastheccmpoandrpcs,thentherewouldbe causetodevelopaprogramtorefineandimprovethemodeltomeetthosegoals. 5.1 FourOptions Thefollowingfouroptionsgivegeneraloverviewsoftheworkitems,relativetimeandbudget needs,andprosandconsdependingontheagency sdecisionforthestatewidemodel.table2 providesasummaryoftheoptions. 5.1.1 OptionI Discontinuethemodel ThisoptionwouldhaveVTransshelvethemodelandrelyonconsultant(s)forfulfillingits forecastingneedsonanas neededbasis.itispossiblethattheconsultant(s)woulddesireaccess tothemodel atwhateverconditioninwhichitwouldbeatthetime fortheprojectmodeling needs.inthiscase,vtranswouldhavethemodelinactivebutavailableforitsconsultant(s).if needed,theconsultant(s)wouldthenberesponsibleforanynecessaryupdatestothemodelin ordertouseitforthespecificanalysis. TABLE2:OPTIONSSUMMARY Option WorkItems ResourceRanking I Discontinuationofthemodel 4 th Noprogramforperiodicupdatesofthemodel (leasttimeand VTransrelianceonconsultant(s)forforecastmodelingneeds budget [1] requirements) II Programforperiodicupdatesofthemodel 3 rd VTranssupervisionofmodelupdatesandapplication DevelopmentofprogramforcontinuedmodelusebyVTrans& DEC/ANR(e.g.traveldemandforecastingforplanningandair 19
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 qualityanalyses) III Programforperiodicupdatesofthemodel 2 nd Formulationofdetailedmodelworkplan,includingthe type/extentofimprovementstobemadetothemodeland timeframe Implementationofimprovements/refinementstomodel componentsandprocedures VTranssupervisionofmodelupdates,improvementprogram,and application OutreachwithinVTranstoraiseawarenessofmodelfunctionality andpromotemodeluse IV Programforperiodicupdatesofthemodel 1 st Formulationofdetailedmodelworkplan,includingthe (mosttimeandbudget type/extentofimprovementstobemadetothemodeland requirements) timeframe Implementationofimprovements/refinementstomodel componentsandprocedures VTranssupervisionofmodelupdates,improvementprogram,and application OutreachwithinVTranstoraiseawarenessofmodelfunctionality andpromotemodeluse Coordinationofmodelingproceduresandplanninggoalswith otherplanningorganizationsinvermont(ccmpoandrpcs) Notes:[1]ThisdeterminationgreatlydependsontheamountofworkcontractedbyVTranstoconsultant(s);if significant,thisoptionmayexceedoptioniiinbudgetresourcerequirements. Asapossibleextensionofthisoption,theconsultant(s)couldprovidethemodelupdatesandan updatesummarytovtransattheendoftheproject,suchthatthemodelwouldexperience selectedupdatesonaproject basisinsteadofonadefinedschedule.however,animportant caveattothisprocesswouldbethatmodelupdateswouldbeonan adhoc basis,potentially performedbymorethanoneconsultant,sotherecouldbeanissuewiththeconsistencyand scopeofthemodelupdates.therefore,vtranswouldneedtoreviewandverifytheupdates a processthatmaynotbeattractiveconsideringthegeneralgoalofthisoption,whichisforvtrans tosignificantlyreduceitsnecessarytimeandresourcesforthemodel. Worktypeandhours Workandhourswoulddependonaper projectbasis,withvtranscontractingand supervisingaconsultant,orconsultants. Pros Forthemostpart,alow costoptionsincevtranswouldnothavetoallocatetimeand budgetresourcesformaintenanceandupdatingofthemodelingprogram. Onanas neededbasis,vtranswouldonlypayforwhatitneedsperproject(i.e.vtrans wouldnotfundadormantmodel). 20
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 21 Pointofuncertainty:Aprodependingonifthecosttokeepmodelactiveandrunitwould begreaterthanthecosttocontractandsuperviseanoutsideconsultantforforecasting work. Pointofuncertainty:Amountofprojectworkcontractedtoconsultant(s);alowamount ofprojectworkwouldkeepcostsforvtransdown. Cons VTranswouldhavetorelyonconsultantsforprojectforecastingneeds. Aresurrectionofthemodelinthefutureforsomemandatoryanalysisneedwould requiresignificationresources,ofbothtimeandbudget. Pointofuncertainty:Acondependingonifthecosttokeepmodelactiveandrunit wouldbelessthanthecosttocontractanoutsideconsultantforforecastingwork.ifthis werethecase,optionii,below,maybeamorecost effectivechoicesinceahighamount ofprojectperformedbyconsultant(s)workcouldbeoverlycostlyforvtrans. 5.1.2 OptionII Maintainthemodelatitscurrentlevelofcapability,withperiodicupdates Thisoptionwouldincludeperiodicupdatestothemodel sinputdata,suchasnewuscensus, NationalHouseholdTravelSurvey(NHTS),andemploymentdata,yearlyAADTcountsandnewly builtroadwaysforthehighwaynetwork.themodelupdateswouldnotincludechangestothe underlyingmodelproceduresandassumptions.itisassumedthatvtranswouldstillutilizethe modelforrelevantprojectanalyses,whiletheuvmtrcand/orconsultant(s)wouldbe contractedtoperformtheperiodicupdatesandnecessarymodelruns. Thefeasibilityofthisoptionwoulddependontheextenttowhichuseofthemodelinitscurrent formwouldactasreimbursementforthetimeandfundsputintoit.anupdatedbutdormant modelwouldbeofnovaluetovtrans. Exampleperiodicupdatesofinputdata ForUSCensusandNHTSdata,updatescouldoccureveryfivetotenyears,dependingon dataavailability.uscensusiseverytenyears,butotherdemographicandhousehold datasources,suchasnhts,maybeavailableonamorefrequentbasis. Employmentdatacanbeupdatedeveryonetofiveyears,dependingonavailabilityof newdata. Newhighwaydata,suchasnewroadwaysand/orcapacityimprovements,canbe updatedonafive yearinterval,orasneededdependingonprojectanalysisneeds.aadt countdatacouldbeupdatedyearly. Significantlandusedevelopmentscanbeupdatedonafive yearinterval,orasneeded dependingonprojectanalysisneeds. Moredetailed,lessfrequentupdatescouldincludenewhouseholddailypersontrip rates,whicharecurrentlybasedona1994vermonthouseholdsurvey.anewsurvey wouldlikelybeneededtodeterminenewthetriprates.
VermontStatewideTravelDemandModel APreliminaryEvaluation UVMTRCReport#10 007 Note:Ingeneral,periodicupdateswouldbeneededatleasteverytenyearstomakethe model sbase yearcurrentandensureforecastingtoa20 yearhorizon.forexample,the model sbaseyeariscurrently2000withaforecastto2020.duringanupdateprocessfor baseyear2010,theforecastyearwouldhavetobeupdatedto2030. Note:Updatesofinputdataandmodelbaseyearwouldrequirere validationofthe modeland,ifneeded,newcalibrationofmodelparameters. Worktypeandhours Workwouldinclude1)modelupdatesand2)modelapplication.Itispossiblethatthe entityresponsibleformodelupdateswouldnotperformmodelapplication.forexample, theuvmtrccouldperformmodelupdatesregularlywhileaconsultantcouldrunthe modelonaper projectbasis.oroneentitycouldberesponsibleforalltasks. Workhoursformodelupdatescouldbedeterminedbasedonanupdateschedule.Hours formodelapplicationwouldbemorevariableandwouldbedeterminedonanas needed basis.forthepurposesofthispreliminaryevaluation,itisassumedthattheworkhours requiredformodeldataupdatesofthisnatureandcorrespondingtasks(checking,revalidation,andcalibration,ifneeded)wouldbeapproximatelyone tenthoftheduration oftheupdateperiod.forexample,toupdatedataonaten yearinterval,itwouldtake approximately12monthsofworktime,everytenyears.updatesonaone yearinterval wouldtakeapproximately1 1½months,everyyear. Pros Likelytheleastexpensiveoptiontokeepthemodelactive,up to date,andavailablefor usebyvtrans. ScheduledupdatescouldbeperformedbytheUVMTRCand/oraconsultant,with managementbyvtrans. Cons Themodelwouldnotgainfurtherfunctionalitybeyondthecurrentlevel. Themodelwouldnotberelevantfordetailedtransit/multi modalplanningefforts. 5.1.3 OptionIII Improvethemodel scapabilityforvtrans(includingdec/anr)analytical objectives ThisoptionwouldincludenecessaryupdatestothemodelasoutlinedintheOptionII,andwould alsoincludefurtherimprovementsandrefinementstothemodelandthevtransmodeling process.underoptionii,themodelcouldbeusedinitscurrentstatetoprovidevmtestimates todec/anrforgreenhousegasandotheron roadpollutantemissionanalyses.however,option IIIprovidesforfurtherrefinementofmodeloutputswhenprovidingVMTdataforairquality analyses. 22