september 2007travel forecasting for new starts2-1 6. standard reporting/qc fta requirements for new...
TRANSCRIPT
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-11
6. Standard 6. Standard Reporting/QCReporting/QC
FTA requirements for New StartsFTA requirements for New Starts ImplementationImplementation Thoughts on good practiceThoughts on good practice
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-22
FTA RequirementsFTA Requirements
Standard Summit reportsStandard Summit reports– No-Build (2030) versus “today”No-Build (2030) versus “today”– TSM alternative versus No-buildTSM alternative versus No-build– Build alternative versus TSM alternativeBuild alternative versus TSM alternative– Build opening year versus “today”Build opening year versus “today”
QC reportsQC reports– Tests on (1) project and (2) IVTTests on (1) project and (2) IVT– Template on “Travel Forecasts”Template on “Travel Forecasts”
Transit assignment resultsTransit assignment results
*
* = new in 2008
*
*
*
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-33
PurposePurpose
Enable routine quality controlEnable routine quality control Support a coherent story for the Support a coherent story for the
projectproject– How is 2030 different from today?How is 2030 different from today?– What would TSM accomplish?What would TSM accomplish?– How much more would the project do?How much more would the project do?– What would the project do in its What would the project do in its
opening year?opening year?
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-44
Changes over Time:Changes over Time:ReportsReports No-Build versus TodayNo-Build versus Today Build Opening Year versus TodayBuild Opening Year versus Today
1.0 Demographics1.0 Demographics
2.0 Travel patterns 2.0 Travel patterns
3.0 Travel times 3.0 Travel times
4.0 Transit trips 4.0 Transit trips
5.0 Transit shares5.0 Transit shares
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-55
1.0 Demographics1.0 Demographics
General formatGeneral format– By districtBy district– Totals, absolute delta, relative deltaTotals, absolute delta, relative delta
ContentsContents
1.1 Population characteristics1.1 Population characteristics
1.2 Employment characteristics1.2 Employment characteristics
1.3 Supporting statistics1.3 Supporting statistics
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-66
Supporting StatisticsSupporting Statistics
Relationships between variablesRelationships between variables– Household size, workers per person, Household size, workers per person,
densities, parking costs, pop/emp densities, parking costs, pop/emp ratio, etc. ratio, etc.
Checks on trip generationChecks on trip generation– ΔΔ p population versus opulation versus ΔΔ trip productions trip productions– ΔΔ employment employment versus versus ΔΔ trip trip
attractionsattractions
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-77
2.0 Person Trips2.0 Person Trips
General formatGeneral format– District-to-district District-to-district – Totals, absolute delta, relative deltaTotals, absolute delta, relative delta
ContentsContents2.1 Total person trips2.1 Total person trips
2.2 Home-based work person trips2.2 Home-based work person trips
2.3 Other person trips (as needed)2.3 Other person trips (as needed)
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-88
3.0 Travel Times3.0 Travel Times
General formatGeneral format– District-to-district District-to-district – Peak and off-peakPeak and off-peak– Averages, relative delta Averages, relative delta
ContentsContents3.1 Highway travel time3.1 Highway travel time3.2 Transit in-vehicle time3.2 Transit in-vehicle time3.3 Transit total weighted time3.3 Transit total weighted time
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-99
Travel Time Travel Time CalculationsCalculations CalculationsCalculations
– District-to-district aggregation of skimsDistrict-to-district aggregation of skims– Divided by district-to-district I-J cells Divided by district-to-district I-J cells
ImpedancesImpedances– Highway: drive-alone timeHighway: drive-alone time– Transit: “best” walk access pathTransit: “best” walk access path– Transit cells: Transit cells: path available in both path available in both
yearsyears
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-1010
4.0 Transit Trips4.0 Transit Trips
General formatGeneral format– District-to-district District-to-district – Totals, absolute delta, relative deltaTotals, absolute delta, relative delta
ContentsContents4.1 Total transit trips (all modes)4.1 Total transit trips (all modes)
4.2 Home-based work transit trips4.2 Home-based work transit trips
4.3 Other transit trips 4.3 Other transit trips (as needed)(as needed)
4.4 Guideway trips (by purpose, 4.4 Guideway trips (by purpose, as as neededneeded))
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-1111
5.0 Transit Shares5.0 Transit Shares
General formatGeneral format– District-to-district District-to-district – Totals, absolute delta, relative deltaTotals, absolute delta, relative delta
ContentsContents5.1 Home-based work transit shares5.1 Home-based work transit shares
5.2 Other transit shares 5.2 Other transit shares (as needed)(as needed)
5.3 Guideway shares (by purpose, 5.3 Guideway shares (by purpose, as as neededneeded))
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-1212
Changes between Changes between Alternatives:Alternatives:ReportsReports TSM versus No-BuildTSM versus No-Build Build versus TSMBuild versus TSM Contents for Contents for eacheach mode choice mode choice
runrun1.0 Transit trips1.0 Transit trips
2.0 User benefits2.0 User benefits
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-1313
1.0 Transit Trips1.0 Transit Trips
General formatGeneral format– District-to-districtDistrict-to-district– Totals, absolute delta, relative deltaTotals, absolute delta, relative delta
Contents for each mode choice runContents for each mode choice run1.1 Transit trips1.1 Transit trips
1.2 Transit trips by “transit 1.2 Transit trips by “transit dependents”dependents”
1.3 Guideway transit trips (as needed)1.3 Guideway transit trips (as needed)
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-1414
2.0 User Benefits2.0 User Benefits
General formatGeneral format– District-to-districtDistrict-to-district
Contents for each mode choice runContents for each mode choice run2.1 Total user benefits2.1 Total user benefits
2.2 User benefits accruing to “transit 2.2 User benefits accruing to “transit dependents”dependents”
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-1515
2.0 User Benefits 2.0 User Benefits (cont.)(cont.)
2.3 User benefits accruing to “base” 2.3 User benefits accruing to “base” ridersriders ““Base” = TSM for 2030 Build versus TSMBase” = TSM for 2030 Build versus TSM ““Base” = No-Build for 2030 TSM versus Base” = No-Build for 2030 TSM versus
No-BuildNo-Build
2.4 “Base” riders’ share of user benefits2.4 “Base” riders’ share of user benefits
2.5 Negative user benefits2.5 Negative user benefits
2.6 Negative user benefits as a share of 2.6 Negative user benefits as a share of totaltotal
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-1616
2.0 User Benefits 2.0 User Benefits (cont.)(cont.)
2.7 Stratified tables2.7 Stratified tables Build alternative transit trips by change in Build alternative transit trips by change in
transit price, separately for CW-CW and MD-MDtransit price, separately for CW-CW and MD-MD
2.8 Frequency distributions 2.8 Frequency distributions Build transit trips by change in transit priceBuild transit trips by change in transit price User benefits by change in transit priceUser benefits by change in transit price All eight (8) access combinationsAll eight (8) access combinations
2.9 Thematic maps of user benefits by zone 2.9 Thematic maps of user benefits by zone for (1) productions and (2) attractionsfor (1) productions and (2) attractions
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-1717
Quality ControlQuality Control
Summit QC impacts in 2002Summit QC impacts in 2002– Errors in modelsErrors in models– Errors in service plansErrors in service plans– Errors in transit network codingErrors in transit network coding– Base|Build coverage inconsistenciesBase|Build coverage inconsistencies
FTA analysesFTA analyses– User benefits unrelated to the projectUser benefits unrelated to the project– User benefits other than transit IVT deltasUser benefits other than transit IVT deltas
routine
case by case
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-1818
Quality Control Report Quality Control Report 11 Share of UBs directly related to Share of UBs directly related to
projectproject1.1. Best paths by transit access (walk, drive)Best paths by transit access (walk, drive)
2.2. UBs based on local models & best pathsUBs based on local models & best paths
3.3. I-J cells with path on the project (0/1)I-J cells with path on the project (0/1)
4.4. UBs in I-J cells with project pathsUBs in I-J cells with project paths
5.5. Project UB share = step4 UBs / step2 UBsProject UB share = step4 UBs / step2 UBs
Summit district-to-district reportsSummit district-to-district reports Project UB share < 80% Project UB share < 80% ‘splaining ‘splaining
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-1919
Quality Control Report Quality Control Report 22 Share of UBs caused by Share of UBs caused by ΔΔ transit transit
IVTIVT6.6. With best paths from Report 1: replace With best paths from Report 1: replace
guideway-path IVT with IVT from baseline guideway-path IVT with IVT from baseline alternativealternative
7.7. Rerun mode choice and SummitRerun mode choice and Summit
8.8. IVT UB share = step2 UBs / best-path UBsIVT UB share = step2 UBs / best-path UBs
Summit district-to-district reportsSummit district-to-district reports ΔΔIVTIVT share < 80% share < 80% ‘splaining ‘splaining
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-2020
Example: Example: Honolulu Fixed Honolulu Fixed GuidewayGuideway
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-2121
Travel Forecast TemplateTravel Forecast TemplateInputs by Trip PurposeInputs by Trip Purpose
Daily transit trips, base and buildDaily transit trips, base and build Daily person trips (base = build)Daily person trips (base = build) Daily hours of user benefits (UBs)Daily hours of user benefits (UBs) Positive UB hours from coverage changesPositive UB hours from coverage changes
– MD-CW, NT-CW and NT-MD groupsMD-CW, NT-CW and NT-MD groups Daily hours of UBs changed by cappingDaily hours of UBs changed by capping Daily hours of UBs for transit dependentsDaily hours of UBs for transit dependents
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-2222
Travel Forecast TemplateTravel Forecast TemplateInputs for Special MarketsInputs for Special Markets
Project trips per event dayProject trips per event day UB hours per event dayUB hours per event day Transit passenger miles per event dayTransit passenger miles per event day Number of event days per yearNumber of event days per year
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-2323
Travel Forecast TemplateTravel Forecast TemplateOther InputsOther Inputs
Daily project tripsDaily project trips– Standard purposes, no special marketsStandard purposes, no special markets– Lowest socio-economic class Lowest socio-economic class
(representation of transit dependents)(representation of transit dependents) Daily project passenger milesDaily project passenger miles
– Standard purposes, no special marketsStandard purposes, no special markets– Transit dependentsTransit dependents
Project length (miles)Project length (miles) Annualization factor (days/year)Annualization factor (days/year)
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-2424
Travel Forecast TemplateTravel Forecast TemplateOutputs-Standard Outputs-Standard PurposesPurposes Daily new transit tripsDaily new transit trips
– % distribution of total new transit % distribution of total new transit tripstrips
% distribution of daily UBs% distribution of daily UBs % distribution of daily baseline transit % distribution of daily baseline transit
tripstrips % UBs lost to capping% UBs lost to capping % UBs accruing to lowest socio-% UBs accruing to lowest socio-
economic classeconomic class
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-2525
Travel Forecast TemplateTravel Forecast TemplateOutputs-Special MarketsOutputs-Special Markets
% distribution of total new special-% distribution of total new special-market annual transit tripsmarket annual transit trips
% distribution of total special-market % distribution of total special-market annual UBsannual UBs
Minutes of UBs per project tripMinutes of UBs per project trip
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-2626
Travel Forecast TemplateTravel Forecast TemplateOutputs-Quality Control Outputs-Quality Control ChecksChecks Minutes of UBs per daily project tripMinutes of UBs per daily project trip
– Before cappingBefore capping– After cappingAfter capping
% UBs coverage related% UBs coverage related % UBs from special markets% UBs from special markets % of project trips that are new transit % of project trips that are new transit
tripstrips Project’s average trip distance / project Project’s average trip distance / project
lengthlength
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-2727
Transit Assignment ResultsTransit Assignment Results
Production/attraction formatProduction/attraction format Total riders (linked trips) and Total riders (linked trips) and
boardings (unlinked trips) by modeboardings (unlinked trips) by mode Guideway station ONs and OFFs by Guideway station ONs and OFFs by
direction and mode of accessdirection and mode of access Directional transit rider load Directional transit rider load
volumes between stationsvolumes between stations Station to station transit ridersStation to station transit riders
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-2828
Ons, Offs, and Load Ons, Offs, and Load VolumesVolumes
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-2929
Mode of Access by StationMode of Access by Station
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-3030
Thoughts on Good Thoughts on Good PracticePractice Boundaries for summary districtsBoundaries for summary districts Changes over timeChanges over time Changes between alternativesChanges between alternatives Thematic mapsThematic maps Desire-line plotsDesire-line plots
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-3131
Summit DistrictsSummit Districts
Generally, between 15 and 20 for Generally, between 15 and 20 for reportingreporting
Smaller near project; larger elsewhereSmaller near project; larger elsewhere Thematic maps of UBs Thematic maps of UBs district district
boundariesboundaries Corridor (for making-the-case discussion)Corridor (for making-the-case discussion)
– Aggregations of districtsAggregations of districts– Perhaps immediate and broader corridorsPerhaps immediate and broader corridors
Different structures for special analysisDifferent structures for special analysis
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-3232
Changes over TimeChanges over Time
Identify/justify very high or very low % Identify/justify very high or very low % changeschanges
Comparisons to observed growth trendsComparisons to observed growth trends Roadway and transit supply checksRoadway and transit supply checks Roadway and transit speed checksRoadway and transit speed checks Transit passenger PMT and PHT checksTransit passenger PMT and PHT checks
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-3333
Changes between Changes between AlternativesAlternatives
Identify/justify large changes in zone to Identify/justify large changes in zone to zone mode shareszone mode shares
Examine zone-to-zone pairs with major Examine zone-to-zone pairs with major negative user benefits or high “per negative user benefits or high “per rider” positive user benefitsrider” positive user benefits
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-3434
Thematic MapsThematic Maps
Zonal population, employment, trip end Zonal population, employment, trip end and transit mode share changesand transit mode share changes
User benefits per impacted transit trip User benefits per impacted transit trip production or attractionproduction or attraction
Percent of zone’s transit riders with a Percent of zone’s transit riders with a change in user benefitschange in user benefits
Positives separate from negativesPositives separate from negatives
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-3535
Desire Line PlotsDesire Line Plots
District-to-district “bandwidth” lines for District-to-district “bandwidth” lines for those pairs representing 25, 50, and those pairs representing 25, 50, and 75% of all user benefits75% of all user benefits
District-to-district lines for those pairs District-to-district lines for those pairs with some zone-to-zone negative with some zone-to-zone negative benefitsbenefits
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-3636
7. Lifting the Cap7. Lifting the Cap
FTA requirements for New StartsFTA requirements for New Starts BackgroundBackground Analytical approach to adjusting Analytical approach to adjusting
the capthe cap
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-3737
FTA RequirementsFTA Requirements
Cap on per-trip Cap on per-trip transittransit user benefits user benefits– Applied to CW-CW and MD-MD trips I-JApplied to CW-CW and MD-MD trips I-J– (“Off-diagonal” UBs handled separately)(“Off-diagonal” UBs handled separately)– Standard cap = Standard cap = ±45 ±45 weightedweighted minutes minutes
FTA adjustment of capFTA adjustment of cap– Case-by-case considerationCase-by-case consideration– Based on demonstration of actual Based on demonstration of actual
project benefits > 45 weighted project benefits > 45 weighted minutes/tripminutes/trip
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-3838
BackgroundBackground
The creature from the swamp – 2002 The creature from the swamp – 2002 – Early Summit testing of 12-15 projectsEarly Summit testing of 12-15 projects– Large problems with Large problems with mostmost forecasts forecasts
Model properties (bizarre guideway Model properties (bizarre guideway constants)constants)
Alternatives (inconsistent baselines)Alternatives (inconsistent baselines)
– Prospect of few rated projects in 2002Prospect of few rated projects in 2002– ““Cap” to salvage Cap” to salvage somesome project ratings project ratings
Model-related problems Model-related problems large UB/trip large UB/trip No pending New Starts decision No pending New Starts decision
X
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-3939
BackgroundBackground
Near-term applicationNear-term application– Adjust/remove cap where appropriateAdjust/remove cap where appropriate– CriteriaCriteria
Large project-caused user benefits per tripLarge project-caused user benefits per trip Absence of large spurious benefitsAbsence of large spurious benefits
Longer-term FTA aspirationLonger-term FTA aspiration– The swamp dries upThe swamp dries up– The creature moves awayThe creature moves away Soon?
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-4040
Analytical ApproachAnalytical Approach
Sources of benefitsSources of benefits– Based on best paths, Base and BuildBased on best paths, Base and Build– Computation of time savingsComputation of time savings
TSTStottot = build transit trips x (imped = build transit trips x (impedbas bas – – impedimpedbldbld))
TSTSivt ivt = build transit trips x (IVT= build transit trips x (IVTbasbas – IVT – IVTbldbld) ) where:where: imped = total weighted imped = total weighted
impedanceimpedance
IVT = in-vehicle timeIVT = in-vehicle time
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-4141
Analytical ApproachAnalytical Approach
Inspection of time savingsInspection of time savings– Total versus IVT-causedTotal versus IVT-caused– Fraction on I-J paths on projectFraction on I-J paths on project– D-to-D trips with large per-trip time D-to-D trips with large per-trip time
savingssavings– Detailed analysis of illustrative pathsDetailed analysis of illustrative paths
Anticipated outcomeAnticipated outcome– ΔΔIVT IVT usuallyusually the cause of most benefits the cause of most benefits– Project in path for most benefitsProject in path for most benefits
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-4242
Analytical ApproachAnalytical Approach
∞∞
9090
7575
6060
4545
3030
1515
1515 3030 4545 6060 7575 9090 ∞∞
Time Savings: Total versus In-Vehicle Time
Delta IVT
(min.)
Delta Total Weighted Time (min)
capOK
Requires explanation
May need explanation
May need some clean-up
Distribution of time savings across cells determines course of action
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-4343
Analytical ApproachAnalytical Approach
ImplementationImplementation– Custom-written programsCustom-written programs– Reporting features of forecasting Reporting features of forecasting
software?software?– SummitSummit
Track record on cap increasesTrack record on cap increases– Largely IVT-caused: yes increased for allLargely IVT-caused: yes increased for all– Largely service-policy caused: noLargely service-policy caused: no– Largely walk/xfer caused: yesLargely walk/xfer caused: yes
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-4444
Example: NY East-Side AccessExample: NY East-Side Access– Long Island RailroadLong Island Railroad– Now: to Penn Station onlyNow: to Penn Station only– Project: also to Grand Central StationProject: also to Grand Central Station– Analysis Analysis No cap No cap
~All 45+minute changes to east side~All 45+minute changes to east side ~All had much shorter walks, many fewer xfers ~All had much shorter walks, many fewer xfers IVT changes contributed a modest share of UBsIVT changes contributed a modest share of UBs
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-4545
SummarySummary
Cap exists to ward off swamp Cap exists to ward off swamp creaturescreatures
Projects with large per-trip UBsProjects with large per-trip UBs– Demonstration of real project causesDemonstration of real project causes– Usually Usually ΔΔIVT, but not alwaysIVT, but not always
Soon(?): routine QC of forecasts will Soon(?): routine QC of forecasts will eliminate need to capeliminate need to cap
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-4646
8. CTPP-based QC 8. CTPP-based QC ForecastsForecasts
FTA requirements for New StartsFTA requirements for New Starts BackgroundBackground ApplicationApplication ExamplesExamples
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-4747
FTA RequirementsFTA Requirements
No requirement for sketch QC No requirement for sketch QC forecastforecast
Other relevant requirementsOther relevant requirements– Plausible forecastsPlausible forecasts– Analysis of uncertaintiesAnalysis of uncertainties
Potentially desirable applicationsPotentially desirable applications– Starter linesStarter lines– Unfortunate track recordsUnfortunate track records
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-4848
BackgroundBackground
Motivation – to learn from experienceMotivation – to learn from experience– New Starts projects over the past 20 yearsNew Starts projects over the past 20 years– Presumption of available insightsPresumption of available insights– Simple model: markets Simple model: markets ridership experience ridership experience
Purpose – to provide a:Purpose – to provide a:– Synthesis of accumulated general experienceSynthesis of accumulated general experience– Readily available & consistent method and Readily available & consistent method and
datadata– Low-cost source of a Low-cost source of a secondsecond number number– Way to address entirely new park/ride optionsWay to address entirely new park/ride options
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-4949
BackgroundBackground
Aggregate Rail Ridership Aggregate Rail Ridership Forecasting (Forecasting (ARRFARRF) Model) Model– Sponsored by FTA; developed by Sponsored by FTA; developed by
AECOMAECOM– Based on recently built projectsBased on recently built projects
Light rail (11 projects)Light rail (11 projects) Commuter rail (9 projects)Commuter rail (9 projects)
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-5050
ARRF – Key ElementsARRF – Key Elements
1.1. CTPP CTPP workers workersijij
2.2. GISGIS workers workersijij served by rail line served by rail line Home buffers: Home buffers:
6 mi. PNR station 6 mi. PNR station 2 mi. other station2 mi. other station
Work buffers: walk 1 mi.Work buffers: walk 1 mi.
3.3. Model Model total riders total ridersijij on rail on rail
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-5151
ARRF – Key Elements ARRF – Key Elements (cont’d)(cont’d)
LRT modelLRT model– Total ridersTotal ridersijij on rail = function of on rail = function of
WorkersWorkersijij
Workplace densityWorkplace density Direction length of LRT lineDirection length of LRT line
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-5252
ARRF – Key Elements ARRF – Key Elements (cont’d)(cont’d)
CR modelCR model– Total ridersTotal ridersijij on rail = function of on rail = function of
WorkersWorkersijij by income class by income class Average system speedAverage system speed Train-miles per mileTrain-miles per mile Connection to fixed-guideway Connection to fixed-guideway
distributor distributor
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-5353
ApplicationApplication
1.1. Obtain basic input filesObtain basic input files
2.2. Determine the socio-economic characteristics of Determine the socio-economic characteristics of the geographythe geography
3.3. Prepare the CTPP Part 3 flow dataPrepare the CTPP Part 3 flow data
4.4. Determine the relationships between rail stations Determine the relationships between rail stations & geography& geography
5.5. Run Run RailMarketRailMarket program to determine the program to determine the number of work for both live & work nearby rail number of work for both live & work nearby rail stationsstations
6.6. Enter the information from Enter the information from RailMarketRailMarket into the into the model spreadsheetmodel spreadsheet
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-5454
ApplicationApplication
Aggregate test of reasonablenessAggregate test of reasonableness– Guideway ridership onlyGuideway ridership only– Worker-flows as proxy for overall Worker-flows as proxy for overall
marketsmarkets Not a replacement for local modelsNot a replacement for local models
– Unique markets and transit contextsUnique markets and transit contexts– Locally defined purpose and needLocally defined purpose and need– Ridership potential Ridership potential ≠≠ project merit project merit
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-5555
Materials AvailableMaterials Availablefrom FTAfrom FTA Detailed documentationDetailed documentation
– Part-1: Model Application GuidePart-1: Model Application Guide– Part-2: Input Data Development Part-2: Input Data Development
GuideGuide– Part-3: Model Calibration ReportPart-3: Model Calibration Report
RailMarketRailMarket program program Spreadsheets: LRT and CRSpreadsheets: LRT and CR [email protected]@dot.gov
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-5656
Example ARRF Example ARRF ApplicationsApplications Five ExamplesFive Examples
David Schmitt, AECOM ConsultDavid Schmitt, AECOM Consult
Three ExamplesThree Examples Yasasvi Popuri, Cambridge SystematicsYasasvi Popuri, Cambridge Systematics
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-5757
ARRF ApplicationsARRF Applications
Dave SchmittDave Schmitt
AECOM ConsultAECOM Consult
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-5858
Applications – City AApplications – City A
New rail line between CBD and New rail line between CBD and suburban activity centers; strong suburban activity centers; strong corridor bus ridership & servicecorridor bus ridership & service
Compared ARRF LRT model with travel Compared ARRF LRT model with travel demand model resultsdemand model results
ResultsResults– ARRF LRT model results were 100% higher ARRF LRT model results were 100% higher
than travel demand model estimatesthan travel demand model estimates– Stronger motivation to investigate transit Stronger motivation to investigate transit
model parameters; subsequently identified model parameters; subsequently identified issues with walk- and auto-access issues with walk- and auto-access connector methodologyconnector methodology
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-5959
Applications – City A Applications – City A (cont’d)(cont’d)
ConclusionsConclusions– ARRF model may partially explain ARRF model may partially explain
attractiveness of rail over existing attractiveness of rail over existing bus servicebus service
– TDM path-builder probably better at TDM path-builder probably better at evaluating bus/rail competition:evaluating bus/rail competition: Equal service levels for bus & railEqual service levels for bus & rail Buses are just as close or closer to Buses are just as close or closer to
corridor activity centerscorridor activity centers
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-6060
Applications – City BApplications – City B
New rail line between CBD and New rail line between CBD and suburban residential areassuburban residential areas
Used ARRF to develop rationale Used ARRF to develop rationale for alternative-specific constant for alternative-specific constant
Results on next slide…Results on next slide…
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-6161
Ridership Forecasts – Ridership Forecasts – City BCity B
WalkWalk Drive/ Drive/ Drop-OffDrop-Off TotalTotal
ARRFARRF 14,79414,794 6,5486,548 21,34221,342
TDM ModelTDM Model
(no bias)(no bias)11,52011,520 4,5564,556 16,07616,076
TDM ModelTDM Model
(7.5 minute walk, 15 (7.5 minute walk, 15 minute drive)minute drive)
13,14513,145 6,3416,341 19,48719,487
TDM ModelTDM Model
(10 minute walk, 15 (10 minute walk, 15 minute drive)minute drive)
14,77014,770 6,2776,277 21,04721,047
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-6262
Applications – City CApplications – City C
Streetcar in low density urban activity Streetcar in low density urban activity center; existing service is local & center; existing service is local & primarily captive marketprimarily captive market
ARRF LRT model compared with travel ARRF LRT model compared with travel demand model (2000 trip tables, 2030 demand model (2000 trip tables, 2030 networks)networks)
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-6363
Applications – City C Applications – City C (cont’d)(cont’d)
ResultResult– Aggregate model forecast 120% higher than travel Aggregate model forecast 120% higher than travel
demand modeldemand model
ConclusionConclusion– ARRF model may partially explain attractiveness of ARRF model may partially explain attractiveness of
rail over existing service, but does not well-rail over existing service, but does not well-represent benefits of project since:represent benefits of project since:
The project mode is different than calibrated modeThe project mode is different than calibrated mode Lack of choice market not consistent with LRT sample Lack of choice market not consistent with LRT sample
citiescities
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-6464
Applications – City DApplications – City D
Commuter rail between two adjacent Commuter rail between two adjacent metropolitan areas; some express bus metropolitan areas; some express bus service to each CBD, but no service service to each CBD, but no service between CBD’sbetween CBD’s
Commuter rail ARRF model compared Commuter rail ARRF model compared with travel demand model (2000 trip with travel demand model (2000 trip tables, 2030 networks) applied to tables, 2030 networks) applied to eacheach CBDCBD
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-6565
Applications – City D Applications – City D (cont’d)(cont’d)
ResultResult– Aggregate model forecast 130% higher than travel Aggregate model forecast 130% higher than travel
demand modeldemand model
ConclusionConclusion– ARRF model may partially explain ARRF model may partially explain
attractiveness of rail over existing attractiveness of rail over existing commuter bus service, but does not commuter bus service, but does not well-represent benefits of project well-represent benefits of project since lack of service between CBDs since lack of service between CBDs unlike CR sample citiesunlike CR sample cities
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-6666
Applications – City EApplications – City E
New commuter rail line to high mode New commuter rail line to high mode share CBD with established “choice share CBD with established “choice market” commuter bus service from market” commuter bus service from large park and ride facilitieslarge park and ride facilities
Commuter rail ARRF model compared Commuter rail ARRF model compared with travel demand model (2000 trip with travel demand model (2000 trip tables, 2030 networks) appliedtables, 2030 networks) applied
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-6767
Applications – City E Applications – City E (cont’d)(cont’d)
ResultResult– Aggregate model forecast 30% lower Aggregate model forecast 30% lower
than travel demand modelthan travel demand model
ConclusionConclusion– Existing commuter (“choice”) Existing commuter (“choice”)
market in corridor stronger than CR market in corridor stronger than CR sample citiessample cities
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-6868
Applications in Milwaukee, Kansas Applications in Milwaukee, Kansas City,City,
and St. Louisand St. Louis
Yasasvi PopuriYasasvi Popuri
Cambridge SystematicsCambridge Systematics
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-6969
OverviewOverview
Recent ARRF applicationsRecent ARRF applications– Kenosha-Racine-Milwaukee (WI)Kenosha-Racine-Milwaukee (WI)
– Kansas City, St. Louis, MOKansas City, St. Louis, MO
– Madison, WIMadison, WI
– Indianapolis, INIndianapolis, IN
General insights from ARRF applicationGeneral insights from ARRF application– Second data pointSecond data point
– Support for model re-evaluationSupport for model re-evaluation
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-7070
Kenosha-Racine-Milwaukee
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-7171
KRM ServiceKRM Service
ARRF vs. current UPN ridership.ARRF vs. current UPN ridership.– UPN:UPN: 26,000 daily trips 26,000 daily trips– ARRF estimate:ARRF estimate: 21,000 daily trips 21,000 daily trips
Wisconsin OnlyWisconsin Only– Service is close to ARRF defaultsService is close to ARRF defaults– ARRF estimate:ARRF estimate: 2,800 daily riders 2,800 daily riders
Entire Corridor in WI and ILEntire Corridor in WI and IL– Existing UPN Existing UPN andand proposed KRM service proposed KRM service– ARRF forecasts: Entire corridor ARRF forecasts: Entire corridor minusminus Existing Existing
UPNUPN– ARRF estimate:ARRF estimate: 4,500 daily riders 4,500 daily riders
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-7272
KRM Service KRM Service ARRF vs. Model ARRF vs. Model ResultsResults
KRM WI KRM WI Only Only
(RCI = 0.5)(RCI = 0.5)
KRM – WI & KRM – WI & ILIL
(RCI = 1)(RCI = 1)
UPNUPN
(Obs = (Obs = 26,000)26,000)
ARRFARRF 2,8002,800 4,5004,500 21,00021,000
KRM Model-KRM Model-20022002 4,4004,400 5,9005,900 25,07525,075
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-7373
KRM ServiceKRM ServiceConclusionsConclusions
ARRF for existing UPN lower than ARRF for existing UPN lower than observedobserved– Existing UPN service (not a “new” New Start)Existing UPN service (not a “new” New Start)
ARRF for KRM lower than modelARRF for KRM lower than model– KRM as a natural “extension” of UPNKRM as a natural “extension” of UPN
Take a second look at the model:Take a second look at the model:– Trip interchangesTrip interchanges– Transit trip interchangesTransit trip interchanges– PnR and walk access patternsPnR and walk access patterns
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-7474
Kansas City – I-70 Kansas City – I-70 CorridorCorridor
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-7575
Kansas City – I-70 Kansas City – I-70 CorridorCorridor
Traditional suburb-to-CBD commuter rail Traditional suburb-to-CBD commuter rail – Low proposed level of rail serviceLow proposed level of rail service– Location of downtown terminalLocation of downtown terminal– ““What if” scenarios for frequencyWhat if” scenarios for frequency
ARRF offered a set of “second data ARRF offered a set of “second data points”points”– Introduced after model forecasts were completedIntroduced after model forecasts were completed– Independent estimate of model results and Independent estimate of model results and
sensitivitysensitivity
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-7676
Kansas City – I-70 Kansas City – I-70 CorridorCorridor
3 Peak 3 Peak Direction Direction
TrainsTrains
3 Peak + 3 3 Peak + 3 Reverse Reverse TrainsTrains
4 Peak + 4 4 Peak + 4 Reverse Reverse TrainsTrains
ARRF ARRF Riverfront Riverfront TerminalTerminal
815815 1,0201,020 1,1251,125
ARRF ARRF Downtown Downtown TerminalTerminal
1,0601,060 1,3301,330 1,4601,460
Model results with Riverfront TerminalModel results with Riverfront Terminal– 4 peak trains, two-way, and $2.50 fare: 4 peak trains, two-way, and $2.50 fare: 1,244 1,244
ridersriders
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-7777
Kansas City – I-70 Kansas City – I-70 Corridor Corridor ConclusionsConclusions
ARRF used as a second data pointARRF used as a second data point ARRF used for testing ridership ARRF used for testing ridership
sensitivity to:sensitivity to:– Terminal locationTerminal location– Frequency of serviceFrequency of service
Reasonable agreement b/w ARRF and Reasonable agreement b/w ARRF and modelmodel
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-7878
St. Louis Alternatives Analysis
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-7979
St. Louis St. Louis Alternatives AnalysisAlternatives Analysis Data point in ARRF calibrationData point in ARRF calibration On-going study On-going study LRTLRT version of ARRF: markets only, no LOS version of ARRF: markets only, no LOS ARRF applied for 2000 and 2006 MetroLinkARRF applied for 2000 and 2006 MetroLink
– Lower ARRF forecasts for 2000 and 2006Lower ARRF forecasts for 2000 and 2006– Strong impact of special generator trips (ball games, Strong impact of special generator trips (ball games,
airport trips)airport trips)
2002 MetroLink
2006 MetroLink
ARRF 30,600 60,400Observed 42,000 ~80,000
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-8080
St. LouisSt. LouisAlternatives AnalysisAlternatives Analysis
The “build” alignment treated as an The “build” alignment treated as an incremental version to the 2006 MetroLink.incremental version to the 2006 MetroLink.– Growth of 43% from 2006 no-build ridershipGrowth of 43% from 2006 no-build ridership– Ridership of 26,000 without any adjustmentsRidership of 26,000 without any adjustments
Model currently being refinedModel currently being refined– Lower forecast than ARRFLower forecast than ARRF– Examine speeds on competing bus routesExamine speeds on competing bus routes– Examine highway speedsExamine highway speeds
2006 MetroLink2006 MetroLink +
Proposed AlignmentARRF 60,400 86,200Observed ~80,000
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-8181
ARRF InsightsARRF Insights
Market assessment toolMarket assessment tool Second data pointSecond data point Sets the stage to re-evaluate model Sets the stage to re-evaluate model
– Total trips produced Total trips produced – Trip interchange patternsTrip interchange patterns– Transit trip patternsTransit trip patterns– Sensitivity to frequencySensitivity to frequency– Speeds on competing transit and highway facilitiesSpeeds on competing transit and highway facilities
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-8282
9. Alternative-Specific 9. Alternative-Specific EffectsEffects
FTA requirements for New StartsFTA requirements for New Starts ImplementationImplementation Three examplesThree examples
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-8383
FTA RequirementsFTA Requirements
MotivationsMotivations– Correct the “new New Starts” disadvantageCorrect the “new New Starts” disadvantage– Respond to recent evidenceRespond to recent evidence
Positive guideway constants in “well-scrubbed” Positive guideway constants in “well-scrubbed” modelsmodels
Higher-than-explainable BRT ridership impacts (LA, Higher-than-explainable BRT ridership impacts (LA, KC)KC)
ApproachApproach– Attributes rather than modesAttributes rather than modes
– Differential constant and CDifferential constant and Civtivt discount discount
– Post-mode-choice applicationPost-mode-choice application
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-8484
FTA RequirementsFTA Requirements
ApplicabilityApplicability– 2007: “new New Starts” guideways2007: “new New Starts” guideways– 2008: guideway elements of 2008: guideway elements of
baselinesbaselines Stations: amenities and brandingStations: amenities and branding Vehicles: amenities and brandingVehicles: amenities and branding Dynamic arrival informationDynamic arrival information Exclusive runningExclusive running Other attributesOther attributes
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-8585
Attributes, Not ModesAttributes, Not Modes
Attributes not found in modelsAttributes not found in models AlternativeAlternative-specific conditions-specific conditions
– Important: missing attributesImportant: missing attributes– Not important: labels like “BRT” or Not important: labels like “BRT” or
“CR”“CR”– Effects on ridership and mobility Effects on ridership and mobility
benefitsbenefits Guideway-only vs. guideway+local Guideway-only vs. guideway+local
busbus
Someone, please, define these “modes.”
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-8686
Constant and CConstant and Civtivt Discount – FTADiscount – FTAMaximum Maximum
Alternative-Specific Alternative-Specific Effect versus Local Bus Effect versus Local Bus
(mins.)(mins.)
Maximum Maximum Guideway Guideway
Civt Civt DiscountDiscount
Guideway Attributes that Are Guideway Attributes that Are Different from Local BusDifferent from Local Bus
Guideway(sGuideway(s) ) onlyonly
Guideway(sGuideway(s) + local ) + local
busbus
Any Any GuidewayGuideway
Guideway-like characteristicsGuideway-like characteristics 88 33 Civt x 0.85Civt x 0.85
-- Reliability of vehicle arrival, travel -- Reliability of vehicle arrival, travel timetime
44 22 Civt x Civt x 0.900.90
-- Branding/visibility/learnability-- Branding/visibility/learnability 22 11 ----
-- Schedule-free service-- Schedule-free service 22 00 ----
-- Ride quality-- Ride quality ---- ---- Civt x Civt x 0.950.95
Span of Span of goodgood service service 33 00 ----
Passenger facilitiesPassenger facilities 44 33 ----
-- Amenities at stations/stops-- Amenities at stations/stops 33 22 ----
-- Dynamic schedule information-- Dynamic schedule information 11 11 ----
Vehicle amenitiesVehicle amenities ---- ---- Civt x 0.95Civt x 0.95
Availability of seatAvailability of seat ---- ---- Civt x 0.95Civt x 0.95
Maximum effectMaximum effect 1515 66 Civt x 0.75Civt x 0.75
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-8787
Post Mode ChoicePost Mode Choice
Track recordTrack record– FTA reviews of 19 New Starts forecastsFTA reviews of 19 New Starts forecasts– Starter lines Starter lines highest risk of highest risk of
overestimateoverestimate ConsequentlyConsequently
– ASE adjustments for user benefits onlyASE adjustments for user benefits only– No change in total or guideway ridershipNo change in total or guideway ridership– No change in walk/bus/auto access modesNo change in walk/bus/auto access modes– JudgmentJudgment on sizing of park-ride lots on sizing of park-ride lots
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-8888
ImplementationImplementation
Identification of appropriate ASEsIdentification of appropriate ASEs– Sponsor: description of service Sponsor: description of service
attributesattributes– FTA: determination of ASE valuesFTA: determination of ASE values
Detection of paths using the Detection of paths using the projectproject
Detection of guideway-only pathsDetection of guideway-only paths
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-8989
Paths using the ProjectPaths using the Project
With all-or-nothing impedancesWith all-or-nothing impedances– For each I-J:For each I-J:
B B = If = If bldbldIVTIVTgdwygdwy > > basbasIVTIVTgdwygdwy 1, else 0 1, else 0
ASEASE = B x = B x bldbldtripstripstrntrn x ASC x ASC
+ ASD+ ASDivtivt x ( x (bldbldIVTIVTgdwygdwy - - basbasIVTIVTgdwygdwy))
With probability-weighted With probability-weighted impedancesimpedances– Assignment basedAssignment based– ExampleExample
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-9090
Guideway-only PathsGuideway-only Paths
Imprecise conventions in most modelsImprecise conventions in most models– Zone walk-access percents for Zone walk-access percents for anyany transit transit– I-J paths for I-J paths for anyany transit transit– Unable to isolate guideway-only pathsUnable to isolate guideway-only paths
FTA conventionFTA convention– Guideway-only ASC for PnR/KnR onlyGuideway-only ASC for PnR/KnR only– Case-by-case exceptions for (1) model Case-by-case exceptions for (1) model
structure or (2) small zones near projectstructure or (2) small zones near project
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-9191
Guideway-only PathsGuideway-only Paths
Portland streetcarPortland streetcar Subdivision of Subdivision of
zoneszones Low chance of bus Low chance of bus
component if component if IVTIVTbusbus = 0 = 0
FTA agreement on FTA agreement on ASCASCgdwygdwy for walk- for walk-accessaccess
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-9292
ExamplesExamples
ASEs for BRT with an Older ASEs for BRT with an Older PathbuilderPathbuilder
Jeff Bruggeman, AECOM ConsultJeff Bruggeman, AECOM Consult
ASEs for Streetcar with Multi-pathsASEs for Streetcar with Multi-paths Jennifer John, Portland Tri-Met Jennifer John, Portland Tri-Met
ASEs for BRT with Some RefinementsASEs for BRT with Some Refinements Tom Maziarz, Hartford CRCOGTom Maziarz, Hartford CRCOG
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-9393
BRT and Path Choice BRT and Path Choice with an Older Pathbuilderwith an Older Pathbuilder
Jeffrey M. BruggemanJeffrey M. Bruggeman
AECOM ConsultAECOM Consult
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-9494
Nature of AlternativesNature of Alternatives
System with BRT projects to be System with BRT projects to be built parallel to or within right-of-built parallel to or within right-of-way of major highwaysway of major highways
Baseline alternative to be created Baseline alternative to be created with same operating planwith same operating plan
Benefits to flow exclusively from Benefits to flow exclusively from improved speeds on BRT rights-improved speeds on BRT rights-of-wayof-way
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-9595
Anticipated ResultsAnticipated Results
Buses on BRT guideway Buses on BRT guideway equivalent to a premium transit equivalent to a premium transit mode (so, coded as a different mode (so, coded as a different “mode” in the network)“mode” in the network)
Paths with identical out-of-vehicle Paths with identical out-of-vehicle times between baseline and buildtimes between baseline and build
Travel time benefits only on the Travel time benefits only on the BRT guideway onlyBRT guideway only
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-9696
Nature of ProblemNature of Problem
Some BRT routes circulate before and/or Some BRT routes circulate before and/or after travel on BRT guidewayafter travel on BRT guideway
Some arterial routes diverge at BRT Some arterial routes diverge at BRT entry points with some bus-trips using entry points with some bus-trips using the BRT facilities and others continuing the BRT facilities and others continuing on an arterialon an arterial
Path builder finds off-guideway Path builder finds off-guideway differences between baseline and build, differences between baseline and build, causing negative user benefitscausing negative user benefits
Cause: breaking of combined headwaysCause: breaking of combined headways
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-9797
Nature of the ProblemNature of the Problem
ON OFF
Headway?
ON OFFHeadway
? BRT guideway
Arterial street
BRT all-stops
Local bus
BRT integrated
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-9898
Flagging Guideway Flagging Guideway PathsPaths Same mode-code for all bus routes Same mode-code for all bus routes
in network and path-buildingin network and path-building Transit line file parsed to determine Transit line file parsed to determine
usage of guidewayusage of guideway Pseudo-fare links coded for each Pseudo-fare links coded for each
guideway link traversed by each guideway link traversed by each lineline
Result: detection of both “guideway Result: detection of both “guideway all-stops” and “guideway express” all-stops” and “guideway express” routesroutes
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-9999
Path ConditioningPath Conditioning
Single path built with all routes Single path built with all routes treated as local bustreated as local bus
Fare links skimmedFare links skimmed as they are as they are encountered on pathencountered on path
Fare conditioning programFare conditioning program– Paths using at least one fare links – BRT Paths using at least one fare links – BRT – Paths using no fare links – local or Paths using no fare links – local or
expressexpress
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-100100
Mode ChoiceMode Choice
Mode choice model considers Mode choice model considers express, local, and BRT choicesexpress, local, and BRT choices
Express-bus choice is from a Express-bus choice is from a separate path-building step for separate path-building step for expresses that do not use the expresses that do not use the BRT guidewayBRT guideway
““Silver bullet” added to utility Silver bullet” added to utility expression for BRT pathsexpression for BRT paths
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-101101
Current FTA NeedsCurrent FTA Needs
Latest FTA guidance does not allow Latest FTA guidance does not allow for “silver bullet” within model to for “silver bullet” within model to create higher ridershipcreate higher ridership
Technique would have to be changed Technique would have to be changed to skim amount of path time on to skim amount of path time on guidewayguideway
Could be accommodated by updating Could be accommodated by updating fare flags to reflect link impedancefare flags to reflect link impedance
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-102102
Isolation of Trips and Isolation of Trips and Travel Time on a Project Travel Time on a Project In a Multipath In a Multipath EnvironmentEnvironment
Jennifer JohnJennifer John
TriMet TriMet
Portland, OregonPortland, Oregon
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-103103
Multi-path AssignmentMulti-path Assignmentin EMME/2in EMME/2
Several paths between an origin and a Several paths between an origin and a destination are identified destination are identified
Trips are assigned to the paths using Trips are assigned to the paths using probabilities based on weights in the probabilities based on weights in the assignment parameters (in-vehicle, out-assignment parameters (in-vehicle, out-of-vehicle, walk) and headways of the of-vehicle, walk) and headways of the routesroutes
Assignment creates weighted travel time Assignment creates weighted travel time based on proportion of trips on each pathbased on proportion of trips on each path
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-104104
Portland Streetcar Portland Streetcar ExampleExample
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-105105
Isolating Streetcar TripsIsolating Streetcar Trips
Multipath assignment results in Multipath assignment results in trips on LRT, bus, and streetcar.trips on LRT, bus, and streetcar.
The project is a streetcar extension.The project is a streetcar extension. How do we isolate information How do we isolate information
about the trips that use the project?about the trips that use the project?– How do we isolate this information How do we isolate this information
from travel on the existing part of the from travel on the existing part of the Streetcar route?Streetcar route?
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-106106
Isolating Streetcar TripsIsolating Streetcar Trips
In EMME/2In EMME/2– Code Streetcar as unique mode in Code Streetcar as unique mode in
network and transit line codingnetwork and transit line coding– Flag Streetcar routeFlag Streetcar route
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-107107
Flagging the Streetcar Flagging the Streetcar “Project”“Project”
Total StreetcarTrips for theFull route =
10,500
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-108108
Flagging the Streetcar Flagging the Streetcar “Project”“Project”
How manyStreetcar tripsAre on ONLY
The new Portion (A-B)Of the route?
A
B
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-109109
Additional Options Additional Options AssignmentAssignment
Select Line/SegmentSelect Line/Segment– ResultsResults
O/D Matrix of trips for flagged route (or O/D Matrix of trips for flagged route (or portion of route as in the case of an portion of route as in the case of an extension of an existing line)extension of an existing line)
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-110110
Multipath Assignment Multipath Assignment Results: No FlagsResults: No Flags
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-111111
Additional Option Additional Option Assignment Results: Assignment Results: Project FlaggedProject Flagged
Total StreetcarTrips for the
Selected Segment
Of route =8,100
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-112112
Using the Assignment Using the Assignment ResultsResults
Total boardings for the project as Total boardings for the project as isolated in the assignments are isolated in the assignments are used to apply ASE calculationsused to apply ASE calculations– Maximum value of 15 for Guideway Maximum value of 15 for Guideway
ONLYONLY– Maximum value of 6 for Guideway + Maximum value of 6 for Guideway +
local buslocal bus
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-113113
Guideway ONLY TripsGuideway ONLY Trips
Portion of the O/D results matrixPortion of the O/D results matrix– Isolated by a submatrix of the Isolated by a submatrix of the
additional options assignment resultsadditional options assignment results– Includes only zone-pairs where direct Includes only zone-pairs where direct
walk to/from a streetcar stop is walk to/from a streetcar stop is certaincertain at at bothboth ends of the trip ends of the trip
– Would be eligible for up to 15 minutes Would be eligible for up to 15 minutes (as determined by FTA) of ASE(as determined by FTA) of ASE
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-114114
Guideway ONLY TripsGuideway ONLY Trips
TAZs in YellowHave direct
Walk access to Project
(no transfers)
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-115115
Guideway + Local BusGuideway + Local Bus
Portion of the O/D results matrixPortion of the O/D results matrix– Remainder of trips in the matrix that Remainder of trips in the matrix that
do not have direct walk on/off do not have direct walk on/off access to the projectaccess to the project
– Would be eligible for up to 6 minutes Would be eligible for up to 6 minutes (as determined by FTA) of ASE(as determined by FTA) of ASE
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-116116
Isolating Project Isolating Project In-Vehicle Travel In-Vehicle Travel TimesTimes Used in CUsed in Civtivt Calculations Calculations Project coded as unique modeProject coded as unique mode
– In EMME/2 transit times can be In EMME/2 transit times can be saved out specific to a modesaved out specific to a mode
– Use these transit times along with Use these transit times along with the o/d matrix from the additional the o/d matrix from the additional option assignment to identify travel option assignment to identify travel time on projecttime on project
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-117117
ConclusionsConclusions
Multi-path environment and ASEsMulti-path environment and ASEs– Feasible applicationFeasible application– In EMME/2, the key is the additional In EMME/2, the key is the additional
option assignmentoption assignment– Full credit for walk-only ASEs also Full credit for walk-only ASEs also
depends on small zonesdepends on small zones
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-118118
ASEs for the ASEs for the New Britain – Hartford BuswayNew Britain – Hartford Busway
Tom Maziarz & Ming ZhaoTom Maziarz & Ming ZhaoCapitol Region Council of Capitol Region Council of
GovernmentsGovernments
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-119119
Keys to Hartford’s ApproachKeys to Hartford’s ApproachStart with advantage: Start with advantage: 6 transit skims produced for MC6 transit skims produced for MC
• Walk to guideway – Walk from guideway Walk to guideway – Walk from guideway (WG-WG)(WG-WG)• Walk to guideway – Walk from bus Walk to guideway – Walk from bus (WG-WB)(WG-WB)• Walk to bus – Walk from guideway Walk to bus – Walk from guideway (WB-WG)(WB-WG)• Walk to bus – Walk from bus Walk to bus – Walk from bus (WB-WB)(WB-WB)• DriveDrive to transit (PNR) – Walk from guideway to transit (PNR) – Walk from guideway (DT-WG)(DT-WG)• Drive Drive to transit (PNR) – Walk from guideway to transit (PNR) – Walk from guideway (DT-WB)(DT-WB)
Goal: Goal: Consider individual trip attributes of each O-D pathConsider individual trip attributes of each O-D pathAssign weight for Assign weight for each factoreach factor & for & for each path typeeach path type
• ReliabilityReliability• Branding & LearnabilityBranding & Learnability• Station amenitiesStation amenities• Schedule-free serviceSchedule-free service• Long span of serviceLong span of service• Dynamic schedule informationDynamic schedule information
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-120120
Keys to Hartford Approach:Keys to Hartford Approach:
Breakdown & isolate path types using combination of: Breakdown & isolate path types using combination of: – 6 skims6 skims– Key data from skims Key data from skims (some newly created)(some newly created)
– Creative analysis of skim data Creative analysis of skim data • Total IVTTotal IVT• IVT on BRT routes IVT on BRT routes (requires separate coding)(requires separate coding)• # Transfers ….# Transfers ….
Basic Concept: Basic Concept: break down a path into 3 components break down a path into 3 components – IVT IVT ON the guideway ON the guideway (requires separate coding)(requires separate coding)– IVTIVT ON BRT loop ON BRT loop (requires separate coding)(requires separate coding)– IVT IVT OFF guidewayOFF guidewayProcess yields: Process yields: 12 path types that 12 path types that use the use the
guidewayguideway
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-121121
Components used to classify path types
BRT loop Walk from BRT bus
Off Guideway IVT
On Guideway IVT
On BRT Loop IVT
Feeder bus with transfer
Walk to bus
Walk to bus
BRT bus (no transfer)
dashed line also illustrates “BRT bus route” (mode 5)
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-122122
Examples of Path Types
Walk from guideway
Walk to guideway
On Guideway
Most basic path type: 100% On-Guideway
Identifiers:• From WG-WG skim
Alternate ID:• Total IVT = Guideway IVT
Key ID for all 12 path types: • Guideway IVT > 0
WG–WG trip
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-123123
BRT loopWalk from BRT bus
Walk to bus
On Guideway
On BRT Loop
More complex path type: On-Guideway + On-BRT Loop
WG–WBRT trip
Identifiers:• From WG-WB skim
• WB could be feeder bus or BRT bus• BRT bus egress could be on or off loop
Additional Identifiers:Total IVT = G-way IVT + Loop IVTTotal IVT = Mode 5 IVT (BRT bus)
Transfers = 0
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-124124
BRT loopWalk from BRT bus
Walk to bus
Off Guideway
On Guideway
On BRT Loop
Another complex path type:Off Guideway + On-Guideway + On-BRT Loop
BRT bus
Identifiers:• From WB-WB skim• Total IVT > G-way IVT + Loop IVT• Total IVT = Mode 5 IVT (BRT bus)
• Transfers = 0
WB–WBRT trip(Single seat trip)
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-125125
Final example of path type:Off Guideway + On-Guideway + On-BRT Loop
BRT loop Walk from BRT bus
Walk to bus
Off Guideway
On Guideway
On BRT Loop
Feeder bus with transfer
Identifiers:• From WB-WB skim (same as previous slide)
• Total IVT > G-way IVT + Loop IVT (same)
Total IVT > Mode 5 IVT (BRT bus)
Transfers = 1
WB–WBRT trip(with transfer)
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-126126
Process skim tables to develop trip tables for each path type
• WG-WG trips• WG–WBRT trips• WG-WB trips• .....• ……• …… (12 classes)
Off-G’way IVT
BRT Loop IVT
Guideway IVT
Total IVT skim
• Identify O-D pairs that meet criteria for path type
• Select trips for that O-D pair & enter into trip table for that path type
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-127127
Results: # HBW Trips by Path Type(only trips that use all or part of guideway)
1 2 3 4 5 6 71-seat ride 1-seat ride 1-seat ride 1-seat ride transfer transfer transferWG-WG WG-WBRT WB-WG WB-WBRT WG-WB WB-WG WB-WBRT
Trips 87 209 221 620 276 1,439 1,749
Access guidewaystation
guidewaystation
BRTbus
BRTbus
guidewaystation
feederbus
feederbus
Egress guidewaystation
loopstation
guidewaystation
loopstation
feederbus
guidewaystation
loopstation
8 9 10 11 121-seat ride 1-seat ride 1-seat ride 1-seat ride transferDG-WG DG-WBRT DB-WG DB-WBRT DG-WB
Trips 316 720 140 483 462
Access guidewaystation
guidewaystation
BRTbus
BRTbus
guidewaystation
Egress guidewaystation
loopstation
guidewaystation
loopstation
feederbus
Walk Access
Drive Access
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-128128
Assigning attribute weights (constants)
Walk Access Trips1 2 3 4 5 6 7
100% BRT 100% BRT1-seat ride 1-seat ride 1-seat ride 1-seat ride transfer transfer transfer
WG-WG WG-WBRT WB-WG WB-WBRT WG-WB WB-WG WB-WBRT
Reliability 2.0 2.0 1.0 1.0 0.0 0.0 0.0Branding-Learn. 2.0 2.0 0.5 0.5 0.0 0.0 0.0
Schedule-free serv 2.0 2.0 ---- ---- ---- ---- ---- Span of service 2.0 2.0 ---- ---- ---- ---- ----
Station amenities 1.0 1.0 0.5 0.5 0.5 0.5 0.5Dynamic sched info 1.0 1.0 0.0 0.0 0.0 0.0 0.0
Total Constant 10.0 10.0 2.0 2.0 0.5 0.5 0.5
Trips 87 209 221 620 276 1,439 1,749
Benefits (hrs) 14.5 34.8 7.4 20.7 2.3 12.0 14.6constant x trips/60
Attribute Weights in Minutes
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-129129
Entirely on Guideway or Loop
Receive full benefitof guideway (To degree that busway achieves full functionality)
• Reliability: exclusive ROW• Branding : distinct stations,
etc.• Freq serv: 2-4 min h’dways any
station-station trip• Span of serv: 18 hours • Sta. amenities: covered
platforms, visible, secure• Dynamic schedule info: at
guideway & loop stations
1 2100% BRT 100% BRT1-seat ride 1-seat ride
WG-WG WG-WBrt Max
Reliability 2.0 2.0 4
Branding 2.0 2.0 2
Sched-free serv 2.0 2.0 2
Span of serv 2.0 2.0 3
Sta. amenities 1.0 1.0 3
Dyn. sched info 1.0 1.0 1
Total 10.0 10.0 15
Trips 87 209
Benefits 14.5 34.8
Benefits (hours) = Constant x trips/60 min
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-130130
Start off-busway on a ‘BRT’ bus Receive some benefit of G’way • Reliability: lose some reliability
with off-busway segment, but start at less congested end
• Branding : some benefit from BRT marketing, bus branding , etc.
• Freq serv: NO – single route• Span of serv: NO – single route• Sta. Amenities: one end of trip• Dynamic schedule info: NO
3 4
1-seat ride 1-seat ride
WB-WG WB-WBRT
Reliability 1.0 1.0Branding-Learn. 0.5 0.5Sched-free serv ---- ---- Span of service ---- ---- Sta. amenities 0.5 0.5
Dyn. sched info 0.0 0.0
Total 2.0 2.0
Trips 221 620
Benefits 7.4 20.7
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-131131
• All use guideway• But all involve transfer
& feeder bus
• Most guideway benefits compromised by dependence on feeder bus & transfer
• Only 0.5 min credit
5 6 7
transfer transfer transfer
WG-WB WB-WG WB-WBrt
Reliability 0.0 0.0 0.0Branding-Learn. 0.0 0.0 0.0Sched-free serv ---- ---- ---- Span of service ---- ---- ---- Sta. amenities 0.5 0.5 0.5
Dyn. sched info 0.0 0.0 0.0
Total 0.5 0.5 0.5
Trips 276 1,439 1,749
Benefits 2.3 12.0 14.6
September 2007September 2007 Travel Forecasting for New StartsTravel Forecasting for New Starts 2-2-132132
Final Thoughts: Will it work with other models?
BRT loopWalk from BRT bus
Walk to bus
Off Guideway
On Guideway
On BRT LoopTRANSFER
Identifiers:• From WB-WB skim• Total IVT > G-way IVT + Loop IVT• Total IVT > Mode 5 IVT• Transfers = 1
Experience in Hartford• Suggests potential to adapt to other areas• Creative use of coding and & analysis of skims allowed
• Better definition of path types• Better assessment of guideway related benefits