Buses Versus Trams: Low-Floor Boarding

For years, Jarrett Walker has been making the point that streetcars do not provide much transportation value over buses. They have higher capacity, because trams can run longer vehicles, but on a medium-ridership bus line without capacity constraints, this is not relevant. As I note in comments, tramways can use rail signaling to get marginal speed advantages, but it’s not a big advantage. They also have slightly lower minimum headways – a single bus route running more often than every 3 minutes will hopelessly bunch, whereas tramways seem capable of higher frequency – but this is again a matter of capacity. And yet, there is an observed rail bias, even when other factors are kept the same: a Transportation Research Bureau report by Edson Tennyson concludes it is 34-43%, using evidence from North American cities in the four decades after WW2. In this post, I would like to propose one mechanism that may produce rail bias even when speed, capacity, and frequency are unchanged.

Low-floor vehicles make it easier to board from raised curbs. Passengers do not need to climb steps as they do on older buses and streetcars, but can walk straight to the vehicle. This speeds up boarding and improves the passenger experience: see p. 14 of a presentation about commuter rail level boarding, p. 64 of a book about low-floor light rail, and a study about bus dwell times in Portland. It’s especially useful for passengers in wheelchairs: as the Portland bus reference notes, the difference between low- and high-floor buses’ dwell times is especially noticeable when there are wheelchair lifts, because operating a wheelchair lift on a low-floor bus is much faster than on a high-floor bus. Low-floor vehicles’ ability to not lose too much time every time a disabled passenger boards or disembarks improves reliability, which in turns allows schedules to be less padded, improving trip times even when no passenger in a wheelchair uses the route.

While high-floor buses and high-floor trains serving low platforms are similar in requiring passengers to climb multiple steps, low-floor buses and trains with level boarding (regardless of whether high or low) are different. Trains run on guideways; they can have the vehicle’s edge exactly level with the platform edge, with narrow horizontal gap. My attempt to measure this on SkyTrain in Vancouver yielded a horizontal gap of about 5 cm, and a vertical gap of perhaps 2 cm; with such gaps, passengers in wheelchairs can board unaided. While SkyTrain is automated rapid transit rather than light rail, nothing about its technology makes the rail-platform gap easier to resolve than on other rail lines.

In contrast, low-floor buses do not have true level boarding. The bus floor is still raised from the curb somewhat, so modern low-floor buses typically kneel, increasing dwell time by a few seconds per stop, to reduce the platform gap. While it’s possible to raise the curbs as at light rail stops, the buses are still buses, and cannot align themselves to be as close to the platform as trains are: without rails, they sway from side to side, so a safety margin is required from the platform edge. As a result, passengers in wheelchairs cannot board unaided without a retractable ramp, which adds considerable dwell time. On board the bus, the more sudden acceleration also requires the wheelchair to be strapped, adding dwell time even further.

Because there is no way to prevent a bus from incurring schedule risk if there’s a passenger in a wheelchair, bus schedules are inherently more vulnerable than train schedules. While few transit passengers are in wheelchairs, passengers who have luggage, walkers, or strollers can get on and off much faster when there’s perfect level boarding, and are slowed by the need to navigate steps or a wide vehicle-platform gap; the schedule will have to take this into account. Low-floor buses reduce this problem, but do not eliminate it.

The limiting factor to bus frequency is not stopping distances, unlike the case of trains. Instead, it is bunching: at very high frequency, small variations in boarding time compound, as fuller buses are harder to board. Soon enough a slightly delayed bus becomes even more delayed as passengers take more time to board, until it bunches with the bus behind it. This effect can be reduced with off-board fare collection, but when the bus is crowded, the combination of narrow passageways and a significant platform gap means that boarding time is nontrivial no matter what.This effect means that all else being equal, a low-floor tramway will be faster and more reliable than a low-floor bus. In practice, all else is not equal, and in particular, in mixed traffic, the bus’s ability to get around obstacles will make it faster. But with well-enforced dedicated lanes, tramways are capable of running reliably with less schedule padding than buses. A familiar experience to North American bus riders – sitting for several minutes as a passenger in a wheelchair boards, and maneuvers awkwardly through the narrow spaces to where the bus driver will strap in the chair – is not an issue on any train with level boarding.


  1. Fbfree

    Buses at platforms also have minimal vehicle/platform gaps. Given an internal layout with decent circulation and enough doors, low dwell-time is achievable. Recently, while visiting Mexico City, I experienced far worse dwell times on the Metro (line 9) than on the Metrobus, largely because passengers were further from a door on a crush-loaded subway car than on a crush-loaded bus.

    • Alon Levy

      Were these dwell times at equal level of crowding? A crush-loaded subway car has way more passengers than a crush-loaded bus.

  2. Sascha Claus

    The need to get the bus as close to the platform as possible led to the development of the Kassel kerb. As the bus still has to steer towards the kerb/platform edge, it’s floor (at least at the first door) remains above the platform, and it still has to kneel.
    The need to strap down a wheelchair is driven by overcautious regulations or insurance premiums; not by the acceleration. I’ve never seen a bus driver securing a wheelchair on this side of the pond; only signs for the wheelchair users to fasten their brakes.

  3. Nathanael

    This is definitely part of it. Another part is ride quality (buses typically have very unpleasant ride quality, and this is actually worse for low-floor buses).

  4. Ptrqc

    Very northamerican point of view…. Having worked in transit both in Europe and North america, I have not seen the quality of european buses in north america. European buses and infrastructure do offer LRT quality boarding (mostly in BRT like settings), due to above mentioned Kassel kurb (or other variations), full low floor buses (versus north american partial low floor), more doors, all door boarding, etc. More and more hybrid solutions such as the VanHool Exquicity and models developed by other manufacturers under european bus of the future program bridge the gap with LRT. Double articulated and plus size (15m) standard buses offer more choice and european regulations do not require buses to be overbuilt (such as in north america)

  5. Pingback: NTSB Finally Takes an Interest in Cycling Safety — Still Misses the Point | Streetsblog.net
  6. staybailey

    Despite the arguments against the inherency of this rail advantage it’s pretty clear that trains tend to meet these criteria for easy boarding and buses, even BRT, tend not to.

    But I doubt that this accounts for much more than 10% of the difference. My suspicion is that most of this can be theoretically explained by two ideas. First, people’s mode choice for typical trips tends to be sticky. It requires a major jolt for people to switch modes. In the North American case, people often won’t ever try transit (except maybe in exceptional cases like as a tourist or going to the airport) and therefore don’t know whether it would be useful for them.

    Second, at least in North America, rail is perceived, largely for historical reasons, to be superior to buses. This perception of substantial superiority more readily jolts people to try transit. Buses largely lack this ability and so you see giant ridership differences.

  7. Pingback: Today’s Headlines | Streetsblog Ohio
  8. Pingback: Urban Links VII – urbnblog
  9. michael.r.james

    Slightly off-topic but I thought this picture from today’s Guardian is worth a thousand words on the issue of buses-v-rail. Well, you do need to know that the solitary passenger on this 2am London NightTube (inaugurated last Friday night, ie. Sat morning), was a bus driver returning home at the end of his shift.

    The night tube will make a significant difference to his weekend commute. What used to be a 90-minute journey to Finsbury Park by bus would now be just 25 minutes on the tube.


  10. Boris Lipkin

    Interesting post Alon. I’ve been thinking a lot about this question of why people seem to like trains more than buses. You make some good points about trams but I wonder if the larger vehicles also lead to generally less crowded conditions on board (or at least the perception of it). With buses often having seats more crammed in and with aisles being narrow, it always seems like even moderately busy routes are crowded. With trams it feels like there is a bit more space and comfort for riders. Maybe the ride is also a bit smoother. All of this stuff is hard to measure but could play a big part in how people choose to take transit or not.

  11. Nathanael

    The Kansas City Streetcar, which I rode recently, is an informative example of the inherent superiority and greater popularity of trams over buses. It runs in *mixed traffic*, which is not desirable at all. Despite this, it is way, way more popular than any bus on the route could possibly be, because it’s nicer. It’s now routinely filling, including standing room.

    It would take at least twice as many buses to hold the passengers, probably three times as many; you wouldn’t be able to run them frequently enough without bunching; and the fact is that the passengers wouldn’t be there if it was a bus.

    (Kansas City has pretty low traffic, which means the mixed traffic isn’t hurting it as much as it would in a busy city.)

    The level boarding is a big part of it.

  12. Pingback: N-S MetroLink Plan Flops - The Kirkwood Gadfly

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.