Public Transport Performance Assessment

Public Transport Performance Assessment

Public Transportation Performance

• Transit passengers
• Transit agencies (i.e. operators)
• Authorities/regulators
• Motorists

0. Summary of Details

Some empirical parameters in Ref. [2]

Exhibit 4-3 Typical Values of Time for Different Types of Travel Page 4-9
Exhibit 4-5 Relative Values of Time for Different Stages of a Trip Page 4-11
Exhibit 4-8 In-vehicle Time Equivalent of Bus Stop Amenities

Exhibit 5-2 Fixed-Route Frequency quality of service (QOS) Page 5-4
Exhibit 5-3 Fixed-Route Hours of Service QOS Page 5-7
Exhibit 5-4 Fixed-Route Service Coverage QOS Page 5-10
Exhibit 5-9 Grade Factor Page 5-14
Exhibit 5-11 Average Pedestrian Street Crossing Delay: Signalized Crossings Page 5-16
Exhibit 5-12 Average Pedestrian Crossing Delay (s): Unsignalized Crossings with No Yielding to Pedestrians Page 5-17
Exhibit 5-16 Fixed-Route Passenger Load QOS (Vehicles Designed for Mostly Seated Passengers) Page 5-23
Exhibit 5-17 Fixed-Route Passenger Load QOS (Vehicles Designed for Mostly Standing Passengers) Page 5-24

1. Influential factors for public transport trip decision-making

• Availability:
  • Spatial, Temporal, Information, Capacity
• Comfort and Convenience

1.1 Spatial Availability

The presence of easily accessible transit services near one's origin and destination (e.g., first-and-last mile issue)

(1) Walking Access

• Empirical studies: 75% - 80% walked 0.25 mi (400 m or 5 min walk) or less to bus stops

(2) Bicycle Access

• Bicycle facility availability and maintenance
• Bicycle parking security
• User demographics
• Climate
• On-board restrictions and rules

(3) Automobile Access

• Park-and-Ride

• Serve lower-density areas

• Case in Singapore: Cessation of park & ride scheme at 01 Dec. 2016.

• Types of Park-and-Ride (Classified by location or function)

  • Informal park-and-ride lots
  • Joint use lots
  • Park-and-pool lots
  • Suburban park-and-ride lots
  • Transit centers
  • Satellite parking lots

• Park-and-Ride Market Areas

1.2 Temporal Availability

(1) Service Frequency

• Top factor affecting overall trip satisfaction
• Short headways (10 min or less): passengers arrive randomly
• Long headways (15 min or more): passengers count on schedules

(2) Service Span

• The longer the span, the greater the variety of trip purposes can be served
• Traditional commute trips (shopping, social visits, medical appointments)
• Additional types of trips (work late, night class, holidays)

1.3 Information Availability

• Timely and correct information
  • How to use transit services
  • Where to access it
  • Where to get-off
  • Any transfers
  • Timetable
  • Service change or disruption
• Information channels
  • Posted maps, schedules, notices, signs
  • Audible announcements, visual displays
  • Online information
  • Smartphone (handphone) apps
  • On-site staffs
• Capacity availability
  • Passengers are forced to wait for the next
    bus or find another trip mode

Some interesting facts:

• In the Seattle area: Users with real-time information reported wait times that were 30% lower than users without it

• In London :

  • 65% reported shorter wait times, although bus frequency did not change;
  • 83% felt that time passed more quickly;
  • 89% agreed that wait time was more acceptable with the information.

• Real-time information system reduced users' overestimation of wait time by 0.7 min.

1.4 Public Transportation Comfort and Convenience

(1) Passenger Load

• For passengers: Crowded in-vehicle condition is more
  onerous than non-crowding conditions, especially when they
  have to stand.
• For transit operators: Slow down transit operations.
• A compromise between passenger comfort and moving as
  many passengers as possible with the least number of
  vehicles.

(2) Reliability

On-time performance vs regularity of headway

• Bus bunching:

  • Irritating for both passengers and bus drivers
  • Bus services with longer route length, longer run times, a higher number of intermediate stops and shorter headway are prone to bus bunching

• Causes

  • Traffic and environmental conditions
  • Road construction and track maintenance
  • Evenness of passenger demand
  • Vehicle quality
  • Route length and number of stops
  • Staff and vehicle availability
  • Driving skill
  • Operations control strategies
    • Schedule-based and headway-based holding strategies

(3) Travel Time for Passengers

• Access time on both ends of a trip
• Wait time
• Transfer time
• In-vehicle time

Value of Time (VoT)

• The longer the trip, the more value the passengers will place on reducing travel time by a single unit of time
• A unit of time spent during access, wait and transfer is perceived as more onerous than a unit of in-vehicle time

(4) Others

• Safety

  • The potential for being injured

• Security

  • The potential for becoming the victim of a crime

• Cost

  • The cost of using transit service against other travel modes

• Appearance and Comfort

  • Clean and attractive vehicles/Amenities
  • Temperature control
  • Seat and ride comfort

• Customer Relation

  • Driver friendliness and helpfulness

2. Public Transport Service Measuring Methods

Measures of availability for fixed-route public transportation services

• Availability
  • Service Frequency
  • Service Span
  • Access (Walk, Bicycle, Automobile)
• Comfort and Convenience
  • Passenger Load
  • Reliability
  • Travel time

2.1 Service Frequency & Service Span

• Seven levels of Service Frequency: (details in Exhibit 5-2 of Ref [2] in page 5-4)

  • \(\leq\) 5 minutes
  • 5 to 10 minutes
  • 11 to 15 minutes
  • 16 to 30 minutes
  • 31 to 59 minutes
  • 60 minutes
  • \(\geq\) 60 minutes

• Six levels of Service Frequency: (details in Exhibit 5-3 of Ref [2] in page 5-7)

  • \(\leq\) 18 hours
  • 15 to 18 hours
  • 12 to 14 hours
  • 7 to 11 hours
  • 4 to 6 hours
  • \(\geq\) 4 hours

2.2 Access

• Walking is the most common access mode for urban public transit services

  • 50-80% of persons walk 1/4 mile or less to a local bus stop
  • 50% of persons walk 1/2 mile or less to rail and BRT service
  • Terrain, street connectivity, street-crossing difficulty are factors

• Bicycling can extend a stop's market area

  • A person can cover 4 times the distance in the same time, compared to walking

• Autos used in lower-density areas to access transit services when park-and-ride facilities are provided

  • Market area depends on area topography and access road network
  • Typical: 50% of demand comes from within 2.5 miles of lot, 35% of demand comes from upstream up to 10 miles away from lot

2.3 Estimating the Service Coverage Area

(1) Planning level

• For bus stops: air distance within 400 m
• For rail or MRT: air distance within 800 m

(2) Refinement

\[r_s = r_0 \cdot f_{sc} \cdot f_{g} \cdot f_{pop} \cdot f_{px} \]

• where: \(r_s\) transit stop service radius (mi, m)
  \(r_0\) ideal transit stop service radius (mi, m)
  \(r_0=\) 0.25 mi (400 m) for bus stops
  \(r_0=\) 0.5 mi (800 m) for busway and rail stations
  \(f_{sc}\) street connectivity factor
  \(f_g\) grade factor
  \(f_{pop}\) population factor
  \(f_{px}\) pedestrian crossing factor

Refinement for street connectivity: \(f_{sc}\)

• Type 1: Grid, \(f_{sc}=1.00\)
• Type 2: Hybrid, \(f_{sc}=0.85\)
• Type 3: Cul-de-Sac (i.e., dead end), \(f_{sc}=0.45\)

Refinement for terrain: Grade factor \(f_g\)

• 0-5% : \(f_g=1.00\)
• 6-8% : \(f_g=0.95\)
• 9-11% : \(f_g=0.80\)
• 12-15% : \(f_g=0.65\)

Refinement for population characteristics: \(f_{pop}\)

• \(f_{pop}=0.85\), if 20% or more of the boarding volume consists of elderly pedestrians

Refinement for street crossing difficulty: \(f_{px}\)

• Details in Exhibit 5-11 of Ref. [2] in page 5-16
• Pedestrians start to become impatient once crossing delay exceeds 30 s
• Pedestrian crossing factor \(f_{px}\)
  \[f_{px} = \sqrt{\frac{-0.0005 \, d_{ec}^2 - 0.1157 \, d_{ec} + 100}{100}} \]
  • where: \(f_{px}\) is pedestrian crossing factor, and
    \(d_{ec}\) is pedestrian crossing delay exceeding 30 s (\(d_{ec} = d_p - 30\))
• Average pedestrian delay \(d_p\)
  \[d_p = \frac{(C -g_{walk})^2}{2 \, C} \]
  • where \(d_p\) is average pedestrian delay (s)
    \(C\) is traffic signal cycle length (s)
    \(g_{walk}\) is effective green time for pedestrians (WALK time + 4 s of flashing DON'T WALK)

At unsignalized pedestrian crossings, we estimate the average pedestrian delay by considering the factors (Details in Exhibit 5-12 of Ref. [2] in page 5-17):

• Crossing distance
• Average walking speed
• Traffic volumes (vehicle flow rates)

(3) Estimate Transit-supportive Area

Net acre vs Gross acre

• Net acres consider only the area developed for housing and employment
• Gross acres are the total land areas, including streets, parks, etc.

A household density of 3.0 units per gross acre is a typical minimum residential density for hourly daytime transit service to be feasible.

An employment density of approximately 4 jobs per gross acre produces the same level of ridership as a household density of 3.0 units per gross acre

(4) Service Coverage Area vs. Transit-Supportive Area: Four-step calculation procedure

Step 1 : Assemble data

• Transit stops and station locations from GIS database/public transport data system
• Traffic analysis zone (TAZ) data (households, jobs, and TAZ boundaries) from regional transportation planning model

Step 2 : Determine the service coverage area

• Bus stops are buffered using an adjusted radius (400 m is an ideal radius) and
• Rail stations using an adjusted radius (800m is an ideal radius)

Step 3 : Determine the transit-supportive area

• Identify the TAZ areas with a household density of 3.0 or more per acre
• Identify the TAZ areas with a job density of 4.0 or more per acre

Step 4 : Compare Service Coverage Area to Transit-supportive Area

2.4 Service Coverage of Bicycle Access

• Five times faster than walking

  • A 2-km radius for bus stop and a 4-km radius for rail stations
  • These distances are much greater than the typical transit route spacing

• At system level:

  • Improving bicycle facilities and connectivity

• At a station level:

  • Bicycling conditions on access routes to the stops/stations
  • Bicycle parking capacity
  • Onboard bicycle storage

2.5 Service Coverage of Automobile Access

• One-half (50%) of a park-and-ride users start their trips within 3-5 km of the lot
• The outer market area can be scattered over an area 4 or more times as large as the inner market area
• At planning level, 4-km radius around larger Park & Ride (100 spaces or more) may be used as catchment area

2.6 Passenger Load

Details in Exhibit 5-16 of Ref. [2] in page 5-23 and Exhibit 5-17 of Ref. [2] in page 5-24

• For transit vehicles designed for mostly seated passengers:

  • Load factor (passengers per seat)
    Example: Buses, ferries, commuter rail

• For transit vehicles designed for mostly standing passengers:

  • Average standing passenger space (square meters per
    passenger)
    Example: MTR/metro
  • A trade off between transit agency costs (e.g., increased frequency, larger vehicle) and passenger comfort

2.7 Reliability

(1) On-Time Performance Ratio

Definition: The percent of schedule deviations that fall within a defined range (i.e., between 2 min earlier arrive on time 5 min later of scheduled headway)

Note: Suitable for measuring transit services running at longer headways (e.g., 10 min or more)

(2) Headway Adherence Ratio

Definition: The regularity of transit vehicle arrivals with respect to the scheduled headway for a given time period

Measurement:

\[C_{vh} = \frac{\text{stand deviation of } \boldsymbol{\rm actual} \text{ headways}}{\text{average } \boldsymbol{\rm scheduled} \text{ headway}} \]

Note: Suitable for measuring transit services running at shorter headways (e.g., 10 min or less)

(3) Excess Waiting Time (EWT) of the Luckiest Passenger

Definition: Extra wait time compare to what was promised in the schedule at a station/stop

Example:

• time table

  • 3.55 p.m. : 2% departure
  • 4:00 p.m. : scheduled departure
  • 4:02 p.m. : actual departure
  • 4:10 p.m. : 95% departure

• excess time:

  • excess wait time: 4:00-4:02
  • excess platform wait time: 3.55-4:00
  • potential wait time: 4:00-4:10
  • excess budgeted wait time: 3.55-4:10

2.8 Comfort and Convenience

(1) Safety and Security

• Accident rate (preventable or non-preventable)
• Passenger safety (injuries or fatalities per time period)
• Percent positive drug and alcohol tests
• Number of traffic tickets/percent of buses exceeding speed limit
• Number of station overruns
• Number of fires
• Number of crimes
• Ratio of police officers to transit vehicles
• Percent of vehicles equipped with safety devices

(2) Customer Service

• Direct measurement of agency services
  • Percent of missed phone calls
  • Percent of calls held excessively long
  • Customer service response time
• Tracking customer compliments and complaints
  • Web-based forms/E-mail/Hotline/Postage-paid card
  • Customers should receive feedback promptly
• Customer satisfaction survey

3. Case Study of Singapore

3.1 Government Bus Contracting Model

• The roles of land transport authority (LTA)

  • In possession of all bus assets (e.g., buses, depots, interchanges)
  • In charge of bus service planning and standards setting
  • Retain all fare revenue

• The roles of operators

  • Run bus services in accordance with specified standards
  • Receive a fixed service fee independent of its operational cost
  • Up to 10% of its annual service fee will be rewarded or deducted according to the Quality Incentive Scheme

3.2 LTA assesses the operator's performance in the following aspects

• Bus Service Reliability Framework (BSRF): Excess Wait Time and On-Time Adherence
• First and Last Bus Punctuality
• Bus Maintenance
• Bus Interchange and Bus Depot Maintenance
• But Ticketing System Maintenance

(1) BSRF: Average Excess Waiting Time of Passengers at an AITP

• AITP: Assessment Intermediate Timing Point : \(I = \{1,2, \cdots, i, \cdots, m \}\)

• During a time period, a set of headways : \(J = \{1,2, \cdots, j, \cdots,n \}\)

• Average Scheduled waiting time (ASWT) at AITP \(i\)

\[\text{ASWT} = \frac{\sum_{j=1}^{n} (\text{scheduled headway at AITP } i)^2}{2 \times \sum_{j=1}^{n}\text{scheduled headway at AITP } i} \]

• Average Actual waiting time (AAWT) at AITP \(i\)

\[\text{AAWT} = \frac{\sum_{j=1}^{n} (\text{actual headway at AITP } i)^2}{2 \times \sum_{j=1}^{n}\text{actual headway at AITP } i} \]

• Average Excess Waiting Time (AEWT)

\[\text{AEWT} = \text{AAWT} - \text{ASWT} \]

(2) BSRF: Average Excess Waiting Time of Passengers

(3) BSRF: On-time Performance Ratio

\[\frac{\text{No. of on-time buses}}{\text{Total No. of scheduled buses}} \geq 85\% \]

(3) First and Last Bus Punctuality

• Evaluated for each direction of a bus service on a daily basis
• No earlier than the specified time and no more than 5 min later

(4) Bus Fleet Maintenance

• Scheduled and ad-hoc compliance audits
• Vehicle inspections and a review of repair and servicing records
• Maintenance Audit Rating (MAR) is evaluated on a monthly basis

(5) Bus Interchange and Depot Maintenance

• Maintenance Audit Rating (MAR) is evaluated on a Quarterly basis

(6) Bus Ticketing System Maintenance

• The three-monthly moving commuters' Over-Deduction Claim versus Ridership for the Operator shall be less than 0.0008%
• The availability of the Depot Computer System and WLAN system shall be greater than 97% per month

\[\text{Availability} = \frac{\text{Monthly operating time} - \text{Monthly downtime}}{\text{Monthly operating time}} \]

Reference

[1] Kittelson & Associates Inc. et al., TCRP Report 88: A guidebook of developing a transit performance-measurement system. Washington, D.C: Transportation Research Board, 2003, website

[2] Kittelson & Associates Inc. et al., "Chapter 4: Quality of Service Concepts" and "Chapter 5: Quality of Service Methods" in TCRP Report 165: Transit Capacity and Quality of Service Manual, Third Edition. Washington, D.C.: Transportation Research Board, 2013, website