How many buses can a CBD street handle?

* For most areas in New Zealand cities the most common alternative to driving is to take bus public transport
* But NZ Transport planners keep assuming buses have limited capacity and so we must to spend billions on building rail and light rail lines to the city CBD
* Transport planning best practice says streets can support much higher volumes of buses.

Introduction

In our major cities bus public transport is the most common alternative to driving to work for those outside the CBD.  But as both the popularity and importance of bus grows, key PT corridors into and through the CBD have become congested with the increasing number of buses.

It is equally apparent that the core bus corridors through the CBDs have received minimal investment … most investment into public transport goes on trains, track and railway stations. The one exception is the North Shore Busway that has proved a major success.

The success of bus PT and the lack of investment has led to major peak hour bus congestion through CBD streets. This is because bus routes from across the city converge into a single bus corridor through the CBD.  A common theme of proposals to fix this is the limited capacity of a bus-on-street corridor and hence alternatives are proposed … mainly light rail.

So a key question to improve our PT services is what is the maximum capacity of a CBD bus corridor?  To put it another way, how many buses can a street support before we have to switch to another mode to support PT services?  This and the articles will explore the capacity of CBD streets to support peak time bus services.

This article explores what the Transport Planning says about on street bus capacity. It must be recognised that there is a vast amount of best practice and research that support transport planning.  Here in New Zealand, Waka Kaitohi have their own research program but this research is done across the world with perhaps the most authoritative source is the US-based Transport Research Board.

Transit Capacity and Quality of Service Manual – Bus Capacity

The bible for planning public transport capacity comes from the Transport Research Board’s “Transit Capacity and Quality of Service Manual” (TCQSM) published in 2017.  This is used by transport planners across the world to estimate how public transport solutions will perform.  Chapter Six Bus Capacity of the TCQSM spends 136 pages on how to analyse bus capacity. The key section on street bus capacity is outlined below:

Bus Volume and Capacity Relationships

The observed peak-hour bus movements along freeways and city streets, and to or from bus terminals, provide guidelines for estimating the capacity of similar facilities. They also provide a means of checking or verifying more detailed capacity calculations. General planning guidelines are presented in Exhibit 6-16 that match scheduled bus volumes on downtown streets and arterial streets leading to the city center to qualitative descriptions of bus flow along those streets.

These service volumes may be used for planning purposes. More precise values for operations and design purposes should be computed from the capacity relationships and procedures presented later in the chapter.
Transit Capacity and Quality of Service Manual, 3rd Edition – Chapter 6 – Bus Capacity, pages 6-20,6-20

Best practice shows that running buses along a single lane of city street can handle 50 – 80 buses per hour but beyond that the operations become difficult and the service gets unreliable.  Trying to run more than 100 buses an hour on a single downtown street lane should be avoided.

The manual also includes the following about the possible maximum capacity of streets:

A study of bus operations in Manhattan recommended the following desirable maximum a.m. peak hour bus volumes for arterial street bus lanes:
ο Two lanes exclusively for buses: 180 bus/h;
ο One lane exclusively for buses, partial use of adjacent lane: 100 bus/h;
ο One lane exclusively for buses, no use of adjacent lane: 70 bus/h; and
ο Buses in curb lane in mixed traffic: 60 bus/h.
   TCQSM Chapter 6 – Bus Capacity, page 6-37

It should be noted that “Two lanes exclusively for buses” is actually:

• One lane for the bus stop used by stopped buses and
• One lane for buses to travel past the bus stops

A high capacity 180 buses per hour corridor has buses pull off the travelling bus lane into an “Off-Line Bus Stop” for passengers to disembark and board.  Stopped buses do not hold up moving buses which enables much higher bus volumes.  Avoiding having a slow bus hold up faster buses (also known as “Bus bunching”) is part of the reason why higher capacity is achieved with Off-Line Bus Stops but this approach also enables the operation of express buses through the CBD and “Skip-Stop at peak” (more about this below).

If buses can only use one lane for both travelling and stopping then they are using “On-Line Bus Stops” and, in these circumstances, the TCQSM recommends bus capacity would be limited to 70 buses per hour.

“Chapter 6 Section 2 Fundamentals” of the TCQSM outlines the sources of bus delay that limit capacity:

• Delay Associated with Bus Stops (ie having to stop at a bus stop)
• Deceleration and Acceleration Delay
• Bus Stop Failure (where a bus arrives at a stop only to find that all available loading areas are already being used or are blocked by other buses)
• Boarding Lost Time (When a bus arrives at a large stop, there is a typically a delay, or boarding lost time, from the time the bus doors open and when the first passengers arrives to board the bus)
• Dwell Time (Dwell time is the time a bus spends serving passenger movements, including the time required to open and close the bus doors and boarding lost time)
• Bus Stop Position (The position of a bus stop relative to the traffic lane affects how easily buses can re-enter traffic and continue on their route.)
• Traffic Signal Delay

It goes on to say:

Bus Facility Influences on Delay

In addition to delays associated with individual stops, the roadway facility on which buses operate also contributes delays that affect bus speed, capacity, or both. The most important factors are:
ο Stop spacing-how often a bus must stop as it travels along a facility;
ο Exposure to general traffic-the less exclusive the facility, the more buses are exposed to delays caused by other traffic using the facility;
ο Facility design, in particular, the lack of ability for buses to move around each other or other traffic; and
ο Bus operations-the number of buses scheduled relative to capacity and how buses and routes are organized.
   TCQSM Chapter 6 – Bus Capacity page 6-10

All of the above factors, along with the design of the bus itself, have to be considered when planning an investment in improving a bus corridor. There are a wide range of “treatments” which are improvements to the ability of the bus corridor capacity. The TCQSM finally notes the key point that:

The capacity of a bus facility is determined by the capacity of the critical stop along the facility. The critical stop will be the bus stop used by all buses that has the lowest capacity. There are two exceptions to this general rule:
…
ο Skip-stop operation separates bus routes into groups that stop at different sets of stops along a facility. In this case, the sum of the critical bus stop capacity of each group becomes the starting point for determining bus facility capacity, allowing a nearly two- to four-fold increase capacity, depending on the number of stop patterns used.
TCQSM Chapter 6 – Bus Capacity, page 6-10 (Bold added for emphasis)

The best practice manual highlights the lowest capacity bus stop will dictate the bus capacity of the whole corridor unless providing Off-Line Bus Stops that permit Skip-Stop operation can be implemented which will break the critical bus stop bottleneck.

Transport research on the limits of street bus capacity

To continue the conversation, we’ll use the research paper “On the capacity of bus transit systems” from Transport Reviews V22 (3). Although an older paper it highlights key elements of bus corridor capacity that are relevant to today and New Zealand Cities. Note the order of the quotes has been changed for this discussion:

… in a bus lane having no traffic signals one bus may pass a given point each 3.5 seconds, which means a 1,030-bus/h capacity. If there are traffic signals, the above capacity should be reduced by the ratio of the effective green to the cycle time of the downstream traffic signal. For instance, if the green time is equal to 50% of the cycle time, a capacity of 515 bus/h can be obtained, which is higher than the normal flow of buses at almost any corridor. Therefore, if traffic signals were the only point where buses stop, one bus lane would be sufficient.

However … the actual critical points of bus lanes are intermediate bus stops. For example, if four passengers board each bus at a rate of 3 s/pass and each bus takes 12 s entering, opening and closing doors, and leaving the bus stop, a maximum capacity equal to 150 bus/h can be achieved unless some action is taken. …

Buses spend a large proportion of their running time stationary at stops. Lobo (1997) reported that stop time accounts for about half of the journey time between termini.  Gardner et al (1991) found that the mean bus delay at junctions is 15 s, while for a bus stop range from 45 to 90 s. Stops are, then, the main bottleneck for bus operations.

The reason is simple. Along a route buses spend time on road links, at junctions and at stops. The main task of buses at road links and at junctions is to pass through. By contrast, buses must remain at stops boarding and alighting passengers.

… the importance of stops in the capacity of a transit system was shown, implying that the actual bottlenecks for transit operations are the stops. Therefore, the capacity of a public transport system is determined by the capacity of its stops.

In summary, bus corridor capacity is mainly restricted by the capacity of its bus stops which would be, in the case of major New Zealand Cities, the capacity of its CBD bus stops. Adding the last point from the TCQSM “The capacity of a bus facility is determined by the capacity of the critical stop along the facility”, in fact the bus corridor capacity is limited to the capacity of lowest capacity CBD bus stops.

The paper, after reviewing four different bus capacity models, then has a section that compares the capacity of a bus lane or busway for buses with a light rail system using trams. I include the whole section in full because it illustrates the logic of a possible high volume bus corridor:

3 Example of capacity of different transit systems

In this Section, a comparative analysis based on the passenger capacity (for a reasonable level of service) of two technologies of public transport is presented. The technologies compared are a bus lane or busway for conventional buses, and a LRT system consisting of articulated trams. The data come from a study case in Santiago. These are based on an engineering study for the extension of the only busway in the city. The data can be summarised as follows:
ο Bus flow on the corridor: 140 bus/h
ο Average boarding rate of passengers: pb = 4 pass/veh
ο Average alighting rate of passengers: pa = 2 pass/veh
ο Running speed between stops: Vr = 40 km/h
ο Acceleration rate of vehicles: a = 0.9 m/s2
ο Deceleration rate of vehicles: f = 1.2 m/s2
ο Dead time per stop: 0 = 1.0 s
ο Marginal boarding time: b = 3 s/pass
ο Marginal alighting time: a = 2 s/pass
ο Saturation flow of the busway: s = 900 veh/h
ο Vehicles have one-way doors.

3.1 Buses on bus lanes or busways

The mean boarding and alighting rates at the corridor and the application of IRENE suggest that a 2-berth stop can provide an absolute capacity equal to 174 vehicles per hour. However, as the bus flow on the corridor is 140, the saturation degree is 0.81 and a mean queue of 1.8 buses will develop upstream the stop area. On the other hand, a 3-berth stop provides an absolute capacity of 225 bus/h with a saturation degree equal to 0.62, so some possibility of queues is still possible (one bus 40% of the time). 

However, in the instance that a multiple-berth stop is supplied, no more than two adjacent berths are recommended because the gain in efficiency of each additional berth drops sharply (see Tyler, 1982; TRB, 2000). As a consequence, a multiple station made of two 2-berth split stops is preferred. In such a station each stop can easily accommodate almost 105 veh/h for a practical degree of saturation equal to 0.6 (i.e., 0.6 times 174 bus/h); therefore, the whole capacity of the station will be 210 veh/h.

The busway will operate with 12-m conventional 2-door buses with a seat capacity equal to 44 passengers. It is assumed 25% standees; therefore, the mean occupancy of vehicles is 55 passengers. According to the HCM this means a level of service D, for it states that this level is based on local bus operations where short trips at relatively slow speed allow standees. However, the same source states that express bus services on expressways and busways should not allow standees (TRB, 2000).

As a result, the passenger capacity of the busway for conventional buses with 2-berth split stops will be 11,550 pass/h per direction (210 veh/h times 55 pass/veh). Under the same operational assumptions, a bus lane – so that buses cannot overtake at stops – with 2-berth single stops can accommodate almost 6,000 pass/h per direction (105 veh/h times 55 pass/veh). These results broadly agree with Gardner et al (1991) and TRL (1993) in which several bus corridors of developing cities where studied. Thus, TRL (1993) states that a standard 2-berth bus stop can accommodate up to 111 bus/h or 7,400 pass/h-direction. However, 4-berth high capacity bus stop can accommodate more than 161 bus/h or 10,600 pass/h-direction.

3.2 LRT with articulated trams

The capacity of an LRT system will depend on the capacity of the transfer stations and the characteristics of the vehicles. For the latter, we assumed the CITADIS 202 of ALSTOM. This consists of a 22-m articulated tram with a seat capacity equal to 44 passengers, the smallest vehicle available in this system, to be compared with a bus. Similar to buses, it is assumed that 25% are standees; therefore, the mean occupancy of these vehicles is 55 passengers. 

The capacity of the stop will depend on the number of trams that can simultaneously be attended. In contrast with the busway, it is not practical to provide multiple-berth or split stops along the route, due to the difficulties in overtaking manoeuvres that are common to all rail-based vehicles. Therefore, the system should operate as a one-berth stop station serving one articulated tram each time (as a metro system). However, as the trams have 4 doors, it is assumed that 2 doors are used for boarding and the other 2 for alighting. Thus, the boarding and alighting rates per vehicle will be half of the corresponding to buses.

The application of IRENE to this system indicates an absolute stop capacity of 147 vehicles per hour. This means a practical capacity of 88 veh/h (0.6 times 147 veh/h). Therefore, the whole capacity at the station will be almost 90 veh/h.

As a result, the passenger capacity of the LRT with 22-m articulated trams will be almost 5,000 pass/h per direction (90 veh/h times 55 pass/veh). However, as multiple-berth or split stop stations are not feasible along a tram route (except at terminal stations), the capacity can be increased with longer vehicles. Thus, the transport capacity can be up to 7,500 pass/h if a 43-m long tram is used (see Fernández, 2000).

To summarise, the transport capacity under identical operational assumptions for the bus and LTR systems is shown in Table 3.

The paper then makes some concluding recommendations and remarks:

… Simulation experiments with IRENE have provided capacity values for various combinations of operational variables and number of berths (Gibson and Fernández, 1995). The capacity values derived from these combinations are shown in Table 4 for 2-door vehicles. As can be seen in the table, for given operational parameters, the practical capacity of a multiple-berth stop ranges from 60 to 130 vehicles per hour for 2 berths and 80 to 160 for 3 berths. This confirms that the sole increase in the number of berths does not bring in much benefit. It also shows that disordered behaviour produces a loss in capacity in relation to an ordered operation.

…

5 Concluding remarks

It has been argued in this paper that the capacity of any transit system depends on the capacity of its bottlenecks. Thus, it was demonstrated that the active constraints in terms of capacity for transit operations are the stops, as junctions and link provide, in general, more capacity than stops. …

The capacity of stops depends on the particular conditions at each stop. Within these the most important are: frequency of vehicles (flow) and arrival pattern of vehicles; passenger demand and arrival pattern of passengers; and facilities to leave the stop once the transfer operations have finished. Thus, the behaviour of the vehicle flow and passenger demand define the boarding and alighting rates per vehicle. These rates, in turn, determine the capacity and performance of the stop. As a consequence, the capacity of the system will be limited by the capacity of the critical stop at the critical period; that is, the stop in which the boarding or alighting rate is the highest.

The capacity of critical stops can be improved with appropriate design and this paper has provided some recommendations based on simulation studies. However, this design should respond to the particular operational conditions at the stop. These particular conditions will determine physical issues such as:
ο the number of berths at the stop;
ο the need of single or split stops;
ο platform space for passengers and cage dimensions for vehicles;
ο overtaking and exit facilities from the stop; and
ο comfort, accessibility, and information for passengers.

In summary, the design of stops must regard every physical and operational detail to provide an effective public transport system. Otherwise, the system will not be attractive enough to their potential users

Conclusion

To date, the debate to fix bus congestion in our CBD streets seems to be based largely on the assumption that investing in high quality bus facilities will not provide the same level of capacity as, say, investing in a new and expensive light rail service.  In New Zealand, bus based solutions have been and continue to be eliminated as possible Mass Rapid Transit solutions on the basis of what is essentially prejudice they cannot meet the high capacity of peak demand.

This article shows that transport planning best practice says buses can reach high levels of bus capacity even on CBD streets.  As outlined in the Transit Capacity and Quality of Service Manual, 3rd Edition, bus volumes of 180 buses per hour are possible if built around two dedicated bus lanes (one for the bus to stop and one for the bus to pass the stop).  The paper “On the capacity of bus transit systems”, also says high bus volumes are theoretically possible from high quality dedicated bus lanes and bus stops:

As a result, the passenger capacity of the busway for conventional buses with 2-berth split stops will be 11,550 pass/h per direction (210 veh/h times 55 pass/veh). Under the same operational assumptions, a bus lane – so that buses cannot overtake at stops – with 2-berth single stops can accommodate almost 6,000 pass/h per direction (105 veh/h times 55 pass/veh)

Of course, these theoretical models are difficult to implement in the real world.  It may take hundreds of millions or even one or two billion dollars to do this.  The reality is, of course, this is the sort of money already being proposed to enable CBD light rail rapid transit.  Before we start spending billions and months to digging up our CBD streets to lay tracks, we should first really understand what is potentially possible for bus based rapid transit.