bus rapid transit: two kinds of flexibility


On yesterday's post on the watering-down of Bus Rapid Transit proposals in Bristol, UK, a number of comments seem a little vague about how buses are more "flexible" than rail.  For example, Carl writes:

The two primary justifications for BRT [are]:

1. Like rail but cheaper. …
2. More flexible. More flexible means it doesn't need exclusive right of way everywhere (usually the chokepoints), that other traffic can use the lanes, etc.

This a very common way of framing the question, and a very misleading one.

First of all, the notion of "flexibility" used in #2 has nothing to do with the bus/rail distinction, at least if we're talking about surface light rail or streetcars.  You can put streetcars or light rail in mixed traffic and get all the same speed and reliability problems that a bus would deliver in the same situation.  So again, in urban transit:  Speed and reliability are not about vehicle technology; they are about what can get in your way.

But Bus Rapid Transit offers a very different flexibility that in certain situations out-competes rail.  A busway can be designed so that buses from many surface lines can flow into it.  This potentially spreads the usefulness of the busway over a large area without requiring an additional trunk-to-feeder connection.  Connections are unavoidable in good networks, but if there are easy opportunities to eliminate one, it's still worth going for.

This ability to flow through to local lines yields what we call an open busway.  North American open busways include the Ottawa busway network, the Pittsburgh busway, and the Los Angeles El Monte Transitway, but they are much more common overseas, including the developed world's most extensive example in Brisbane, Australia.  This kind of flexibility is impossible to do with rail.

Closed busways, which has none of these benefits but can have more "specailized" vehicles, include the Los Angeles Orange Line.

The flexibility of open busways makes sense only where it matches the pattern of the market. Brisbane is a highly radial city, with a single downtown and densities dropping away as you move away from it.  Outlying nodes of high activity, which could be a strong endpoint for a closed busway or rail line, are scarce.  So the open busway makes perfect sense.  It allows busway service to spread out over a larger area, yielding high frequencies on the inner busway where the demand is higher, and correspondingly lower frequency further out.  It's the kind of flexibility that fits the city.

29 Responses to bus rapid transit: two kinds of flexibility

  1. Aleks Bromfield August 5, 2011 at 4:08 pm #

    Re open busways, don’t forget about Seattle! The DSTT is one of North America’s most heavily used pieces of BRT infrastructure, but no one seems to realize it’s BRT.
    Also, I think it’s important to note that the advantages of open busways can apply to rail as well. This was more obvious back in the streetcar era, when every major street had a trolley, but it’s still obvious in the remnants of those systems in older cities. In Boston, for example, the Green Line works precisely this way. Four local lines (Comm Ave, Beacon St, Riverside, and Huntington) come together into a single central tunnel. Inside the tunnel, it’s a super-high-frequency subway; outside, it’s a frequent local bus in (mostly) dedicated lanes that happens to run on rails.

  2. anonymouse August 5, 2011 at 4:41 pm #

    Interestingly, the El Monte and Harbor transitways in LA have been moving away from an “open” model and more toward a “closed” one with dedicated busway routes (Silver Line and Silver Streak), and only a handful of lines that run through during rush hours, with forced transfers. The problem is that if there’s significant demand on the shared section, having a number of routes overlapping makes for frequent but unreliable service, while having a dedicated route makes for more reliable service, but forces a transfer for anyone going outside that core route.

  3. Brent August 5, 2011 at 5:33 pm #

    I had the same thought as Aleks re: streetcar-subways, pointing also to San Francisco and Philadelphia as examples.
    Interesting that LRT today tends to have almost the opposite service pattern — slower, local at-grade service in the downtown and higher-speed, grade-separated service with wider stop spacing elsewhere.

  4. Aleks Bromfield August 5, 2011 at 7:55 pm #

    Anonymouse: With short enough headways, it’s not a problem. I’ve spent a lot of time riding the Green Line (and some time in Seattle’s DSTT), and it’s never more than a 60-90 second wait for a train going *somewhere*. It’s only when you want to go to one of the branches that you have to wait forever, 😉
    Brent: It’s no mystery why modern LRT has wider stop spacing in the suburbs. The streetcar suburbs that were around when the Green Line was built are practically the inner city today. Even in Boston and San Francisco, newer lines (like the Braintree branch of the Red Line) have much wider stop spacing, matching the sprawling development of car-oriented suburbs.
    The lack of grade separation downtown is inexcusable, but sometimes you do whatever you can to save money…

  5. zefwagner August 5, 2011 at 9:47 pm #

    The Portland MAX works this way on the eastside, with the Red, Blue, and Green lines converging along I-84. Unfortunately, this area of greatest frequency is almost entirely in the freeway right-of-way, severely limiting access and development potential.
    As much as the “grid” in Portland is laudable, I think it could use a bit more of this “open” bus system. For example, most of Portland is low-density, but the Belmont and Hawthorne corridors are getting moderately dense. Due to the grid system, however, the same 15-30 minute frequency bus serves both low and moderate density neighborhoods. Why not have several feeder buses combine into trunk routes down Burnside, Belmont, Hawthorne, and Division as they head towards downtown? I’m imagining a lot of L-shaped routes coming from north and south, then all heading west.
    The 71, 72, and 73 express buses in Seattle are also a fine example. They come in from 3 different low-density neighborhoods, meet in the dense University District, then go express to downtown.
    Unfortunately, RapidRide in Seattle is a closed BRT-lite system. I think the West Seattle RapidRide would be ideal for an open system, since West Seattle has several distinct nodes separated by low-density housing. Ballard would also work as an open system, for that matter. RapidRide could go up 8th, 15th, 24th, and 32nd, with all of them converging to cross the Ballard Bridge. That way it could replace the 15, 17, and 18, instead of just the 15.
    The main reason people might not like an open system is that it might be confusing. It doesn’t have to be though. Just give it a name like Ballard A, Ballard B, etc. Anyone taking it from downtown knows it goes to Ballard, and anyone going farther just needs to remember which letter is theirs.

  6. Jarrett August 5, 2011 at 10:47 pm #

    Zef. The system you describe would work well in Seattle because of the Ballard chokepoint. Less so in Portland because the spreading branches would get tangled up in each other where the street network is still a grid. When a simple grid line has dramatic variations in demand, especially lower demand near the outer end, the appropriate solution is the shortline. Most of the long east-west grid lines in Portland, for example, have a shortline around E 100th even if their long form continues to 200th or beyond.

  7. voony August 6, 2011 at 12:18 am #

    “You can put streetcars or light rail in mixed traffic and get all the same speed and reliability problems that a bus would deliver in the same situation”
    now, imagine the case of a 3 lanes road, 2 lanes of general traffic, and the middle lane used by transit only to bypass general traffic..in both direction at the same time.
    It is certainly possible to do that with a guided transit, but it is looking for operational headache…when the “lateral flexibility” of the bus allow it to occupy or free the middle lane at wish
    It is one of the reason why Nantes, a pionner city with tram in France, has chosen a
    BRT, because this kind of “chokepoint” could have cost way too much to address with a guided solution

  8. zefwagner August 6, 2011 at 12:19 am #

    True, although Portland still has the bridges as choke points–there are just more of them. Mainly I am just concerned because the inner SE (the river to about 50th) is getting pretty lousy frequencies, enough to make the grid system pretty useless since connections take a really long time. The Belmont area is especially hard hit. There are no north-south bus lines from 11th to 39th, and the 15 to downtown (part of the “frequent” network) only gets 15 minute headways during peak times and worse at all other times! Very frustrating as I try to live car-free in what should be a transit-rich neighborhood.

  9. Paul K McGregor August 6, 2011 at 6:49 am #

    Where does the Adelaide O-Bahn fixed guideway bus system fall in the scheme of BRT definitions? It gives you the fixed guideway of a rail system and the flexibility of a bus system.
    Also here in Utah, they use a dedicated bus priority lane to bypass congested areas. However, the limiting factor for this scheme is that when they approach intersections that have left turn lanes for automobiles, they are only allowed one lane for both directions so there is a signal at each end that holds the bus if there is already a bus traveling through the lane. Right now the bus operates every 15 minutes so it is not a problem but it could become a problem if they wanted to make the service more frequent.

  10. Jarrett at HumanTransit.org August 6, 2011 at 12:34 pm #

    Adelaide is an open guided busway. While Adelaide uses an old
    mechanical technology, optical guidance offers the potential for a
    much more versatile product with some ability to improve ride and
    definitely precision level boarding, while retaining the open busway's
    branching capability.

  11. Carl August 6, 2011 at 4:28 pm #

    Hasn’t the Pittsburgh busway been a huge disappointment, carrying far less riders than the tram lines it replaced?
    And isn’t a big problem of the open busways, especially when they radiate from a downtown, how to unload the passengers and handle stops? While the theoretical capacity of the busway may be 600 buses/hour, in practice boarding and disembarking passengers becomes the bottleneck. Isn’t that the big problem of the Ottawa system?
    Even in Seattle’s DSTT with no conflicting traffic at all, it’s the stops that determine the capacity. I guess the ideal design of an closed transitway is one with no stops and radiating lines at either end.

  12. Tim (in Brisbane) August 6, 2011 at 5:04 pm #

    ^ “While the theoretical capacity of the busway may be 600 buses/hour, in practice boarding and disembarking passengers becomes the bottleneck.”
    That’s certainly the bottleneck in Brisbane, especially at the Cultural Center stop which is too short (for geographic reasons). The transit authority has somewhat alleviated it by using pricing to force most riders onto rapid embarkation swipe-cards and banning the purchase of tickets on many peak hour buses.
    But it’s still a major flaw. The only long-term solution I can see is making fewer buses use that section of the busway at peak times (it can be bypassed). But that reduces simplicity.

  13. Jarrett at HumanTransit.org August 6, 2011 at 8:33 pm #

    Carl. Skip stop arrangements and long platforms are the solution to
    stop capacity. Unlike rail, buses can stop without blocking the line
    so long as there's a passing lane. To my knowledge Ottawa Busway
    works fine except for downtown where it doesn't exist!

  14. zefwagner August 6, 2011 at 9:50 pm #

    The Portland Bus Mall and the Seattle Transit Tunnel both use skip stops and it works beautifully.

  15. Jacob Mason August 7, 2011 at 6:04 am #

    How would you characterize the express bus lane on the highway approach to the Lincoln Tunnel which operate in the AM each day? They certainly get things out of the way of transit and allow use by many many different local & express routes and are used by around 1,800 individual buses to bypass traffic getting into NYC each morning.

  16. CroMagnon August 7, 2011 at 8:23 am #

    ^Correct me if I’m wrong, but isn’t the 1800 figure for the segment which has no stops? The capacity for a stopless segment can be extremely high, if the buses disperse somewhat on the far end.

  17. Cap'n Transit August 7, 2011 at 8:41 am #

    Yes, Jarrett, I don’t understand why you consistently ignore the Lincoln Tunnel XBL every time you talk about busways. It’s hands-down the most successful busway in North America.

  18. Jarrett at HumanTransit.org August 7, 2011 at 10:09 am #

    Cap'n.  Sorry, don't mean to be ignoring.  There are numerous very effective pieces of bus-exclusive access to major cities.  In the context of this blog's conversations I tend to focus more on the busways that are fully comparable to rail, in having bidirectional operations, fulltime existence, considerable length, multiple in-line stations etc. 

  19. In Brisbane August 7, 2011 at 3:04 pm #

    “Outlying nodes of high activity, which could be a strong endpoint for a closed busway or rail line, are scarce. So the open busway makes perfect sense. It allows busway service to spread out over a larger area, yielding high frequencies on the inner busway where the demand is higher, and correspondingly lower frequency further out.”
    I’m somewhat skeptical of this comment. I would agree that the direct trip is fantastic, but most people who use the busway (I think the figure is somewhat like 80%) don’t get on/off at busway stations- most are brought into it from the suburbs where the frequency is mind-numbingly low! (this is being progressively fixed with the BUZ program)
    There are also suburban nodes which may in the future accommodate a feeder and transfer model- places like Garden City/Carindale/Indooroopilly/Chermside/Brookside and Wynnum.
    I think in the future the core section of the busway could be upgraded to metro, which would allow the suburban buses to be turned back at stations and allow much increased frequency in the suburbs (where the usual bus frequency is very poor- a bus every half-hour).

  20. Interurbans August 9, 2011 at 5:08 pm #

    The person who wrote this is more on theory than one who actually rides transit. In Los Angeles BRT’s are a failure. The famous Orange bus line does not have the capacity needed and it cannot be increased due to blocking traffic on cross streets, It takes twice as long and cost twice as much per passenger mile to make the trip as a LRT train would. The busses are small and cramped inside with a very rough, bumpy uncomfortable ride even on the new exclusive pavement. The Silver BRT on the Harbor freeway caries only 10,000 to 15,000 riders per day compared the Blue LRT Line that it parallels only a few miles away which carries more than 70,000 riders at half the cost per mile. The Pittsburg Bus ways carry only a small ridership again compared to the LRT line paralleling close by. This is even with a fare structure that makes it cost more to use the LRT. Overall busses just don’t measure up when it comes to high capacity service. They are fine for express or local service, but not when the ridership reaches over 15 to 20 thousand riders per day.

  21. Jarrett at HumanTransit.org August 9, 2011 at 5:40 pm #

    Interurbans:  Brisbane busways carry about 60 million trips per year, which works out to almost 200,000 per day.  I don't have breakdowns for SE vs N busways, but the inner part of SE has to be at least 60% of the total. 
    As I've said many times, the compromised implementation of busways in the US to date says nothing about the potential of busways.

  22. www.facebook.com/profile.php?id=609058067 August 10, 2011 at 5:35 am #

    I was going to comment on the watered-down BRT post, but this one may be more appropriate.
    It is very true that light rail / trams are vulnerable to the same kind of compromise as BRT. Here in Helsinki trams are so commonplace that bad design decisions are common. Mixed running, winding alignments and short dead end branches are all being planned and built.
    Internationally the risk of compromise does seem to be larger for BRT. I think this is not purely for cultural reasons. BRT solutions are often sold on lower costs. The high cost of rail means that it is taken more seriously. Buses also suffer less from mixed running. After all they can bypass obstructions and have similar driving characteristics to cars. As there is less (perceived) damage done, compromise is more attractive. On the cultural side the idea of rail vehicles mixing with car traffic is often foreign while buses mixing with cars are perfectly normal.
    So there are some technical or financial differences related to this wrong kind of flexibility even if cultural differences are probably more significant.

  23. anonymouse August 10, 2011 at 4:35 pm #

    And once again, I feel compelled to mention that because buses can avoid obstructions, there is an expectation that they will do so, and this leads to bus systems being compromised in this way and suffering from an accumulation of thousands of little delays and inconveniences, where a train system might be more carefully planned to avoid that because the train can’t, for example, change lanes to get around a double parked car.

  24. Pete Brown August 10, 2011 at 4:47 pm #

    The UK Dept of Transport published a report on busway capacity in 1991. It says:
    “A basic 2-lane two-way busway serving a corridor with busy bus stops, but without special operational measures, can accommodate 11-15,000 passengers/hour/direction (p/h/d), depending upon local conditions. High capacity configurations can offer substantially superior performance. Five of the eight busways studied accommodate passenger flows of 15,000 p/h/d or more in the critical boarding direction* and of these, three carry 18,000 p/h/d or more. The highest recorded passenger flow was 26,000 p/h/d in the predominantly alighting direction in Porto Alegre”.
    * the report highlights slower bus boarding times as an issue where inbound bus stops have predominantly boarding passengers in the morning peak, with the same occurrence in the evening outbound direction, on systems with on-board fare collection.

  25. Pete Brown August 10, 2011 at 4:55 pm #

    Tim (in Brisbane), I have always wondered how the Brisbane Busway stations accomodate waiting passengers when several buses arrive at once. Are there several stop locations along the platform allocated to specific routes? I have seen the term “Lead Stop” used but no explanation as to its meaning.

  26. In Brisbane August 11, 2011 at 6:23 am #

    Hi Pete,
    The Brisbane busways work by having buses pull up at the stop that is free as they arrive. Lead stop is just so people wait at the front.
    The Brisbane busways are quite good, I don’t understand why ridership is being used as a metric to compare systems because the CONTEXT is so different and also you need to fit the right tool for the right circumstance.
    The Brisbane busways are great- it took me a grand total of 1 minute wait tonight (10pm on weeknight) to get my bus. If that was on the Brisbane train system, I’d be waiting half an hour!
    There needs to be more focus on Brisbane trains though- no reason why the scope of hours and frequency could not also be improved. Brisbane train services are a disgrace!

  27. David in Ottawa August 11, 2011 at 8:23 pm #

    One of the things that’s happening in Ottawa is that some of BRT’s defenders are arguing that Ottawa should adopt a “closed” model of operation rather than its current “open” model, which they term is not “real” BRT. They’re doing this, of course, to stave off conversion to LRT.
    Growth in absolute transit ridership and a growth in the service area is gradually forcing an ever more closed model of operation, with low patronage routes being shortened to/consolidated with local feeder routes.

  28. Puma Outlet Store August 22, 2011 at 3:03 am #

    In the broader sense, a red herring is any plot device that purposely tricks the reader into thinking the book or narrative is going one way, when it is actually going another. A red herring in mystery stories is generally a clue found by the protagonist that is either wrongly interrupted by the reader or character or has nothing to do with solving the crime at all. In a romantic plot, the red herring could be an overheard conversation wherein the love interest describes his love for another woman that our protagonist overhears, not realizing he’s talking about how much he loves hanging out with his sister (and not in an incest way).

  29. Nathanael August 31, 2011 at 7:22 pm #

    “In the context of this blog’s conversations I tend to focus more on the busways that are fully comparable to rail, in having bidirectional operations, fulltime existence, considerable length, multiple in-line stations etc.”
    But they’re all crap.
    The ones which aren’t comparable to rail, on the other hand, are often rather good… like the Lincoln Tunnel stuff.
    The Brisbane example is *consistently* the only good example of a busway you can find, and your main argument for why it’s good is the way it’s *not* like rail. You can’t honestly call the US examples “compromised” — they blew their budgets on making them as rail-like as they could.
    Now, for some reason, Brisbane is apparently willing to build a busway and not to make its trains run properly. In contrast, cities in the US with the *same attitude* — focusing on busways but not taking their trains seriously — *STILL* find that the trains do better.
    Heck, the best argument you have for a busway is that it’s an “open system” like the old streetcars feeding into the downtown tunnels in Philadelphia and Boston. Except *even there the streetcars work better*. Ottawa is dumping the busway and going to LRT because it needs a tunnel downtown…. and *it’s cheaper to build a rail tunnel than a bus tunnel and cheaper to maintain it*.
    You get back to the fact that for the applications where you want long stretches of exclusive ROW, rail is practically always cheaper for comparable or better results. Sure, there may be occasional exceptions, and you’ve specified that Brisbane has an exception: the reason the busways were cheaper there is the reuse of a particular road-only bridge. Where there isn’t a special exception like that, busways are just… inefficient. Everywhere.