A Parking-constrained All-mode Airport Ground Access Mode Choice Model Ian Harrington Central Transportation Planning Staff Boston, Massachusetts.

A Parking-constrained All-mode Airport Ground Access

Mode Choice Model

Ian Harrington

Central Transportation Planning Staff

Boston, Massachusetts

Objectives of Presentation

• Explain reasons to model airport ground access mode choice

• Describe structure of model• Define model parameters and data sources• Describe features of model

Why Model Airport Mode Choice?

• Logan Airport passengers account for about 0.5% of weekday person-trips in Eastern Massachusetts region (74,000 trips in 2007)

• Logan Airport is located in congested area• Special services for airport passengers• Airport mode choice has different factors• Port authority is interested in more detail

Model Structure

• Create separate model for each market segment: • Resident Business• Resident Non-Business• Non-Resident Business• Non-Resident Non-Business• Model demand for 21 modes, including rental cars• Account for impacts of constrained parking supply• Estimate egress mode distributions as well

Ground Access Modes• Rapid transit: walk and drive access• Silver Line (BRT): walk and drive access• Water shuttle: walk and drive access• Logan Express bus: drop-off and park*• Local bus: walk and drive access• Private bus• Door-to-door: Taxi, Shuttle, and Limo• Driven: Drop-off and Short-term Parking• Self*: Long-term, Economy, and Remote Parking• Rental car**: Logan and Off-airport

Model Parameters

• Passenger characteristics• Trip characteristics• Regional travel times and costs• On-airport travel times

Passenger Characteristics

• Market segment• Household size• Household income• Auto ownership• Employer-paid travel?• Logan flight frequency

Source: 2003 Logan Airport Passenger Survey (some empty cells filled)

Trip Characteristics

• Origin Transportation Analysis Zone• Residential Origin?• Party size• Day of week begun• Trip duration• Luggage carried• Time period (AM peak, midday, PM peak, night)• Airport terminal

Source: 2003 Logan Airport Passenger Survey

Regional Travel Times and Costs

• HOV and SOV time, distance, and tolls

Source: CTPS Regional Model highway assignments

• Transit in-vehicle, wait, and walk times and fares – walk-access and drive-access – for each mode (separate walk-access links required for each mode)

Source: CTPS Regional Model transit assignments

On-airport travel times • Auto to terminals• Auto to parking locations• Auto to rental car locations• Buses to terminals• Rapid transit shuttle to terminals• Rental car shuttles to terminals• Water shuttle bus to terminals• Walk from parking locations to terminals

Source: Travel time runs

Resident Business Model

DAT (6)

Long-term Economy Remote

Parking Nest Drop-off (2) Bus Park Door-Door (3)

Access Mode

WAT (4)

Resident Non-Business Mode

RT Water Bus BRT

WAT Nest

Long-term Economy Remote

Parking Nest

Taxi Limo Shuttle

Door-Door Nest

RT, Bus

Drop-off Short Park

Logan Nest Water, PB BRT, LE

Drop-off Nest Bus Park

Access Mode

Parameters Used in Models

• In-vehicle travel time• Out-of-vehicle travel

time• Auto-access travel time• Transfers• Travel costs• Auto ownership

• Household income

• Employer pays expenses?

• Non-resident origin?

• Flight frequency

• Luggage carried

• Household size

Adding Rental Mode

Rental decision includes many additional factors, so:• Freeze coefficients for non-rental modes before

adding rental modes to model• Add other parameters to represent rental modes in

model• Rental fees• Party size• Origin in or near Central Business District• Trip duration

Non-Resident Business Model

Logan Rental Off-airport Rental

WAT (4) DAT (6)

Taxi Limo Shuttle

Door-Door Nest Drop-off (2)

Non-rental Nest

Access Mode

Non-Resident Non-Business Model

a

RT Water BRT Bus

WAT Nest DAT (6)

Logan Rental Off-airport Rental

Rental Nest Drop-off (2)

Taxi Limo Shuttle

Door-Door Nest

Access Mode

Implied Values of Time

• Values of time implied by in-vehicle time and cost coefficients in Resident models:

Resident Business Resident Non-Business

Self-pay, Self-pay, Employer Self-pay, Self-pay,Low-income High-income Paid Low-income High-income

Transit – Walk $17 $44 $52 $18 $32Door-to-Door $46 $117 $140 $40 $70

Parking – Transit $17 $44 $52 $18 $32

Parking – Logan $29 $75 $90 $29 $52

Drop Off – Transit $9 $13 $19 $8 $18

Drop Off – Logan $23 $33 $49 $15 $34

Implied Values of Time

• Values of time implied by Non-Resident model in-vehicle time and cost coefficients:

Non-Resident BusinessSelf-pay, Self-pay, Employer Non-Resident

Low-income High-income Paid Non-BusinessTransit – Walk $15 $19 $26 $10Door-to-Door $36 $49 $66 $37Drop Off – Transit $21 $21 $42 $4Drop Off – Logan $64 $64 $129 $10Rental $24 $32 $43 $21

Parking Constraint• Estimate survey period passenger vehicle arrivals

by day of week• Use trip durations to estimate vehicle parking

demand by day of week

Example: 1 car arriving on Wednesday for 3 days

Sun Mon Tues Wed Thurs Fri Sat 0 0 0 1 1 1 0

Parking Constraint• Compare estimate to survey period parking data

and create parking space demand model. Example: Wednesday Total = (LT + Econ + (.5 * ST))*.95• If demand exceeds supply, distribute excess

demand from parking to other modes Park .5 Park .25 Taxi .3 Taxi .45 Bus .2 Bus .30

Egress Mode ChoiceEgress trip characteristics can differ and other factors involved in mode choice, so applied results of passenger survey questions:• What mode did you take from Logan after your

arrival?• What mode do you plan to take from Logan after

your return?Example: 53% of AM peak long-term parking accesses are Night long-term parking egresses

Model Application

• Originally estimated and applied with 2003 Logan Airport passenger survey

• Validated with 2007 Logan Airport passenger survey

• Applied to individual survey records with appropriate expansion factors

• Future year applications use same survey records with adjusted expansion factors

Summary

The ground access passenger mode choice

model:• Estimates distribution of Logan Airport

passenger travel by 21 modes• Constrains parking demand to available

supply• Includes the rental car modes• Also estimates egress trip modes

Contact Information

Ian [email protected]

Chief Transportation Planner

Central Transportation Planning Staff to the

Boston Region Metropolitan Planning Organization (www.bostonmpo.org)