Bus Rapid Transit Followup

I’ll pull together a response to feedback on the controversial Brisbane busway post in the next few days, but meanwhile, Engineer Scotty asks a good clarifying question:

Part of the problem with BRT [Bus Rapid Transit] acceptance in the US, is [that] most visible BRT systems … tend to look and act like rail-based metros. In the US, we speak of BRT lines–the Silver Line in Boston, the Orange Line in LA, EmX in Eugene, OR–and so forth.  The busses which run on BRT are different than the local busses (different branding, different route nomenclature, different fare structures, rapid boarding, longer station spacing, nicer stations, proof-of-payment or turnstiles rather than pay-the-driver-as-you-board)–

Everything about the service is made to look like a rail-based metro service, except for the tires. I think that therein lies the problem–trains are better than busses at being trains, and this sort of distinct corridor service tends to be better served by rail, all else being equal.

There are exceptions, of course–low volume small-city lines like EmX again–but such services ignore the fundamental strength of busses,
which is that they aren’t tied to the physical infrastructure. Other than the capital cost–something which it is usually not wise to optimize for, as governments can generally get capital funding more easily than operating revenue–if you limit the BRT vehicles to the BRT infrastructure, you’re generally getting something that costs more and not less to operate. BRT, to be most useful, needs to be disconnected from the notion of “lines”, where the line refers to the physical infrastructure. Would you agree with this, or am I barking up the wrong tree?

Scotty is onto an important distinction.  In the BRT business, the systems Scotty cites (Eugene, Los Angeles Orange Line, Boston Silver Line) are
called ‘closed’ systems, in that all buses operate only within the
specified infrastructure, with a separate brand.  Each is just one route or in a few cases, like the Silver Line, there may be two or three patterns.

Brisbane is an ‘open’ system, like Ottawa and Pittsburgh, and like the El Monte Transitway in Los Angeles.  The busway infrastructure is branded, and
there are bus routes that remain entirely within it, but there are many
others that branch off of it into the regular street network.  The Busway brand applies to the infrastructure, not the routes.  The buses running in the busway are ordinary Brisbane Transport buses like you’d see on any route in the city.

However, the distinction between open and closed is not the same as the distinction between “emulating light rail” and “emulating heavy rail.”  That distinction is about grade separation, i.e. separation from cross traffic.  Typical recent American applications such as the Los Angeles Orange Line or Eugene’s EmX are emulating light rail in their ability to run onstreet and cross streets at signals, which heavy rail generally doesn’t do.  As I’ve said, I think Brisbane’s busway segment is a fairer imitation
of heavy-rail speed and reliability than anything in North America,
apart from fragmentary busways in Ottawa, Pittsburgh, Los Angeles etc.
Of those, only Ottawa’s is anything close to a complete system, and
it’s compromised, of course, by the lack of a downtown segment.

The two distinctions are quite separate.  Here in Sydney, for example, there are two busways (Liverpool Parramatta T-way and North West T-way) that are entirely at grade with lots of signalised intersections.  The first is currently ‘closed’ — operated by a single line running the whole length — but there are plans to make it ‘open.’  The second has always been open, i.e. served by a range of routes that branch off of it at various points.

34 Responses to Bus Rapid Transit Followup

  1. EngineerScotty November 19, 2009 at 4:46 pm #

    Thanks for the quick response, Jarrett!
    My question was focusing on the open/closed distinction, more than the grade-sparated/at-grade distinction (though the latter is important)–that said, there’s probably a level of service in between, that being an at-grade busway with signal priority at level crossings.
    An interesting question for you, since you spent your youth in Portland during a time when wood products dominated the local economy: Do you remember the Molalla Logging Road? If you don’t, it was kinda like a busway–for log trucks. (Nowadays it is in disrepair and much of it is undriveable; and is often used as a trail). It ran from a log dump in Canby, OR to Mollala OR and into the woods beyond, and featured such things as a grade-separated crossing over OR99E in Canby (this structure still exists), signalized crossings of major thoroughfares in Clackamas County with signal priority given to log trucks (including at grade-crossings of OR213 and OR211 near Molalla). It allowed logs to efficiently move from the point of harvest, in the cascade foothills, to a a transfer point along the Willamette River north of Canby. (Why this level of infrastructure was necessary for movement of non-perishable freight such as timber I’m not sure, but there it was…)
    Getting back to the topic: How does the open nature of Brisbane affect things like fare collection and rapid boarding? Do entirely in-busway routes handle this differently than those that branch out?

  2. Jarrett at HumanTransit.org November 19, 2009 at 5:33 pm #

    No, I don’t remember the Mollala logging road! It sounds like an interesting bit of history, but my parents and school were both too leftist to take me on tours venerating logging infrastructure.
    The Brisbane busway does not feature off-board fare collection, and I’m almost sure it’s because there’s a contactless smartcard on the horizon and they didn’t want to buy ticket machines that would be obsolete in a few years.
    It’s clearly technically possible to equip stations with TVMs and require their use so that buses can board all-doors. Right now everything is pay-boarding, both inside the busway and out. I expect this will become an issue in the next decade, as some inner city stations start to get pretty congested with boarding buses.

  3. EngineerScotty November 19, 2009 at 5:46 pm #

    My folks took me on the road once to see property; plus I grew up in Oregon City so the logging road was something I encountered more than a few times. (If you ever drive down 99E through Canby, there’s a big green overpass near where the Molalla branch railline intersects with the UPRR mainline; that’s the logging road).
    Equipping stations along a transitway (road or rail) with ticket machines is obviously not a problem–but if the busses thereon venture into the hinterlands, do you put a machine at every curbside (or kerbside for those down under) stop? Dispense a receipt to everyone paying cash fare regardless of transfer or not? And once you go proof of payment–fare inspection is a tractable problem in a closed network, but is it feasible to do proof of payment on an entire city bus system? Either you would need exeedingly honest patrons, I would think, entire armies of fare inspectors, or sufficiently draconian penalties to offset the much lower risk of getting caught cheating.
    Or you could go entirely fareless and avoid the problem. 🙂
    One compromise would be divide busses into “trunk” routes and “local” routes; the former only serving major stations (whether on or off the busway) and featuring proof of payment, the latter venturing out into the neighborhoods and operating on a pay-the-driver basis. Note that this distinction isn’t entirely the same as open/closed, as the both types of service can use or not use the busway as convenient.

  4. Alon Levy November 19, 2009 at 6:54 pm #

    Presumably, BRT could run in a transit mall downtown, just like the Calgary LRT. If the transit mall is closed to non-emergency traffic other than buses, and if the buses have signal priority, then the surface segment should not slow traffic down too much.

  5. Jarrett at HumanTransit.org November 19, 2009 at 7:25 pm #


    It's no problem doing proof of payment on only part of a line, as long as everyone knows to keep their proof of payment when riding into that segment.  Typically, you do some focused fare enforcement around where vehicles exit the proof of payment segment.

  6. Jarrett at HumanTransit.org November 19, 2009 at 7:27 pm #


    Sure it can.. The speed downtown in that situation is driven by stop spacing and signals.  The old Portland Mall had stops too close (every 2 blocks i.e. 400 feet) and thus the buses averaged around 7 mph.  But it was reliable!

  7. EngineerScotty November 19, 2009 at 7:59 pm #

    Interestingly enough, the “new” Portland mall has stops spaced further apart–which has prompted quite a few complaints from people who apparently prefer the shorter walk to the faster throughput.
    Whether any statistically valid data has been collected on the subject, I don’t know…

  8. Cap'n Transit November 19, 2009 at 8:45 pm #

    I think Scotty put his finger on the issue here: that the North American lines in question are aping light rail instead of trying to be the best bus system they can be. The focus on “lines” instead of “busways” makes no sense unless you’re either trying to pull a bait-and-switch, or going along with what everyone else is doing.
    Hoffman and Caine had a great article last year that really helped me focus my thinking on this issue. Their distinction between “Light Rail Lite” (Eugene) and “Quickways” (Brisbane). In a comment on Streetsblog, Hoffman further elaborated on the value of “open” busways:
    “To those who would question the costs and feasibility of “all those buses,” the experience of Brisbane is critical. Since most buses fill up off corridor–before they even arrive at the Quickway–the decision was made to just run them pretty much straight into their downtown (with very few intermediate stops); after all, why force people to transfer from one bus to another, especially if the first one was already full? And as to why not have them transfer to a train, the answer was also simple: they have an extensive train network, but people found the direct buses much, much faster than taking the bus to the train (as well as inherently more reliable and less stressful). So the network structure has a direct bearing on how many passengers you attract to the system (Brisbane has seen a 40% growth in system ridership in 3 years due in great part to its new Quickway system)–but they also saw that the vast improvement in travel speeds meant that they could provide much more service with a given size fleet (when a round trip that took two hours now takes less than 40 minutes, you can operate three round trips with the same bus that previously made only one), vastly improving productivity and ACTUALLY PUSHING MANY TRANSIT ROUTES INTO THE BLACK.”
    That effect on profitability is the same one that I’ve observed in the Lincoln Tunnel XBL. With that kind of success, I’m baffled as to why people who actually care about transit continue to insist on building “light rail lite”.

  9. Cap'n Transit November 19, 2009 at 9:17 pm #

    About proof of payment: sure, there are ways to make it work with the Quickway concept, but it’s a completely separate bus improvement that’s been incorporated into the “BRT package,” but there’s no reason it has to be there. Paris has proof of payment on its entire network, much of it without physically separated lanes. NYC Transit’s pilot project in the Bronx has proof-of-payment buses running right next to non-proof-of-payment buses, partly in painted lanes and partly in mixed traffic. The Lincoln Tunnel buses have a grade-separated lane with no proof of payment at all.

  10. stephen November 19, 2009 at 9:40 pm #

    What happens when you have one piece of grade separated infrastructure that would be valuable to either mode i.e. rail or bus who gets priority? Many on this post have questioned why wouldn’t Brisbane just put down rail when going to the trouble of building that infrastructure. A similar question might be, why would anyone go to the trouble of building grade separated rail infrastructure without building it to accommodate rubber tires? If they did, we might see where the real transit allegiances lie.
    For example, the Seattle Downtown Transit Tunnel and HOV ramps onto I-90 that are planned to carry light rail are two examples of transit competing for infrastructure.
    Currently there are buses and light rail trains using the downtown transit tunnel, when light rail trains increase their frequency to every 5 minutes or so there will no longer be room to run joint operations. 5 minute trains will provide much more capacity in the tunnel than the buses could, however, the irony is that when the buses leave the tunnel, ridership in the tunnel will acutally decrease. That’s because transit is not all about capacity, but where that capacity can take you. The buses that currently use the tunnel can reach all over the Seattle region, the train will only serve a limited corridor.
    The only way the tunnel will ever serve as many people with trains as it currently does and could do with buses is if travel on the limited light rail corridor grows. Should transit infrastructure favor future growth over meeting current demand? In my estimation this is the central question around rail versus bus infrastructure.

  11. EngineerScotty November 19, 2009 at 10:16 pm #

    If you’re suggesting rails in the pavement, that’s probably a great idea (assuming one mode’s needs aren’t undermined by this). A high-speed metro probably wouldn’t due to share its track with busses, for numerous reasons, but mixed-mode ROWs are useful in many contexts.

  12. EngineerScotty November 19, 2009 at 10:31 pm #

    Thanks for the link, Cap’n!

  13. Alexander Craghead November 19, 2009 at 11:26 pm #

    It must be remembered that most US BRT applications mimic light rail because they were sold on the premise of being a budget option for light rail. They were not then, and generally are not now, looked at as being an improvement to the bus system. Many are planned from the start to include the notion of conversion to light rail at a later date, if traffic warrants it.
    For a great visual of this, pic up a copy of METRO and check out the bus advertisements. Most of the major ones will be for BRT articulated busses. Note they have streamlined noses, sleek styling, and skirts to hide the tires. These aren’t busses, they’re busses in light rail drag. The manufacturers themselves are essentially marketing their BRT oriented products as pretend light rail for the budget conscious agency.
    This is why many — myself included — have an instant visceral reaction to BRT. The immediate words that come to mind are “pretend,” “fake,” and “mediocre.”
    The end result is that you do build systems that have all the right-of-way costs of light rail, with less of the service quality benefits, for a much cheaper cost.
    You are right, Jarrett, there could be a form of BRT that actually is a significantly improved bus mode, something that makes meaningful use of BRT rather than being a pretend train. I have yet to see this application in the U.S.. Locally, TriMet is considering something nebulously termed as “BRT elements” to be added to some routes in the near future. What this means or what it looks like I don’t know, but I suspect it simply means larger, fancier “stations” replacing some stops and some light prioritization.
    One note of irony: the idea of a trunked system that branches out may not be the pattern for BRT here (if there even is one here yet), but it *is* the pattern for future rail infrastructure. The current Metro RTP proposal includes the idea of having future Portland Streetcar routes being built to MAX weight standards so that they can host dual service, allowing the existing eastside trunks to fan out over greater areas at the periphery. Likewise, in discussions of the Barbur HCT route, a branching concept serving two or three different end points is also mentioned. In the end, the pattern you are advocating here may in fact be the future in Portland, albeit not necessarily in the BRT mode.

  14. EngineerScotty November 20, 2009 at 12:01 am #

    I’ve kinda thought that the Barbur Boulevard corridor–at least from Burlingame (or possibly to Capitol) north to PSU (and the present transit mall) might make a good BRT corridor. (Independent of any LRT that gets built out to Tigard or further SW suburbs such as Tualatin or Sherwood). Several FS lines serve the corridor (the 12, the 54/56); however Barbur is frequently subject to traffic delays as (along with I-5) are the only decent arterials in the area.
    Speaking of Tri-Mets future plans, though, Alexandar–I noticed in a recent document issued by Metro, the proposed Beaverton/Wilsonville rapid transit corridor is mentioned–and the current thinking is to expand WES service to all-day service at 15-minute (or better) intervals. Given how expensive WES is to run (FRA compliant trainsets with 2-man crews)–how does Tri-Met think they would pay for such a plan? (Or is the marginal cost of WES service actually much lower than the burdened cost at the current usage rate?) I’d love all-day service along this line–but have doubts that expanding WES service is the right answer. Thoughts?

  15. Alexander Craghead November 20, 2009 at 10:33 am #

    I have no clue about the WES financial issue. The train sets (providing they didn’t break down all the time) shouldn’t cost that much to operate, but the crewing issue does seem to be impossible to work around.
    I’m also not entirely sure that it is TriMet’s desire to make this change, or if it is Metro’s. The RTP primarily reflects Metro modeling and priorities. I have yet to shake the tree on this issue.

  16. calwatch November 20, 2009 at 2:31 pm #

    The lame thing is that Los Angeles is discontinuing one of the central advantages of the El Monte Transitway, by converting most of the routes into the Silver Line and Silver Streak, forcing transfers at stations with little more nearby than a giant parking lot (Artesia Boulevard and El Monte Station) and eliminating through running for alleged reliability reasons.

  17. Doug Allen November 20, 2009 at 5:08 pm #

    First, about the Molalla Forest Road: there was a similar private road in SW Washington serving the International Paper mill at Chelatchie Prairie near Amboy. Being a fully private road, the log trucks did not need to be street legal or pay state highway taxes. I always assumed the Molalla Forest Road was a similar deal. Logging railroad lite, so-to-speak. This was not uncommon mid-20th century.

  18. EngineerScotty November 20, 2009 at 5:31 pm #

    The Molalla Forest Road was also private (though the ROW of at least part of it has been bought by the government, and the City of Molalla and ODOT are considering using part of it as a bypass).
    However, I am still a but surprised that log trucks travelling on the road were given signal priority over automotive traffic on state highways, in the two places where there was a grade crossing. In that regard, it was like a railroad…
    Times certainly have changed. 🙂

  19. Doug Allen November 20, 2009 at 5:42 pm #

    I can give some insight into TriMet mode choices in the 1970’s. Early in the decade, TriMet hired DeLeuw-Cather to develop a long term planning vision. The resulting “1990 Plan” anticipated a network of exclusive busways and bus lanes serving transit centers, to supplement local bus service. Had it been built, it might have attracted decent ridership.
    However, there was no funding for such an expansive system, but instead a pot of federal money resulting from the demise of the Mt. Hood Freeway proposal. This pot was initially divided three ways, for the Banfield Corridor, the Sunset Corridor, and the Oregon City corridor. There was an assumption by many that the first two would be either busways or HOV projects, and the third might be light rail.
    There was a “Governors Task Force” to consider freeway alternatives, and subsequent revisions to federal law that allowed much of the money to eventually be spent on alternative highway projects.
    However, for political reasons, because it had some proximity to the original Mt. Hood Freeway corridor, the Banfield Corridor was the regional priority, with the Oregon Department of Transportation being the lead agency (since it would be a rubber-tire on pavement project).
    I was a member of the Citizens Advisory Committee, and also a member of Citizens for Better Transit, which had been advocating for a “multi-destinational” transit network, along the lines of what was being advocated by Greg Thompson, Jas Kooner, Rudy Massman, Tom Matoff, John Bakker, etc. We felt that a “branching-tree” busway would be an inferior component of such a transit network, compared with the higher capacity vehicles of a light rail trunk line serving transfer stations where it met cross-town bus service.
    Unfortunately, the cross-town network did not materialize and frequencies on the east-side grid never reached optimal levels. The “Reagan Recession” of the early 1980’s just prior to the opening of Portland’s first light rail line more-or-less derailed plans for better bus service.
    Ironically, three of the subsequent rail lines in Portland have followed the inefficient “branching tree” plan, and hopes for a truly “multi-destinational” network capable of providing fast, auto-competitive transit service capable of serving a large percentage of regional trips, remain about as distant as they did in the late 1980’s.
    Still, I think it is fair to say that the original choice of light rail over busway for the Banfield corridor was not, for the most part, due to some belief in the prestige of rail, but due to technical analysis by Wilbur, Smith Associates consulting to TriMet, as well as TriMet staff, who made the case that not only would a light rail line be insignificantly more expensive than a busway built to the same standards, but that light rail, with high-capacity vehicles, was better suited to the type of network that many hoped to see developed, and would be more efficient to operate given the anticipated ridership and vehicle loadings.

  20. K November 20, 2009 at 6:45 pm #

    A question on the open/closed issue. I thought the most efficient way to run a corridor service was to run a single (or as close to single as possible) service at high frequencies on the corridor and then supplement it with smaller feeder routes like in Curitiba.
    It seems the Brisbane model is being sold as the exact opposite. I’ve had traffic engineer friends from Brisbane tout it explicitly as an advantage.
    Feeder routes come at the cost of a transfer but surely the efficiency of the system, implemented with decent transfers comes out in front? I’m willing to be proven wrong but there’s a wry suspicion in me that maybe this busway was a bait and switch pulled by transport planners on the government.
    Having used the bus system in Brisbane (not the busways) they sure could use some simplification of the route overlap. For a visitor it was pretty daunting.

  21. Ben November 20, 2009 at 7:56 pm #

    I think the thing is we need to identify the purpose of the route rather than the technology used. Like most people identify transit with bus/brt, tram/lrt, and heavy rail. But the thing is, in some cities some regular bus and tram routes are better for longer distance commuting than the brt, lrt, or hrt services provided.
    Like in Toronto (which just opened a busway today), the north-south subway line is designed for long distance commuting because it has spaced out stops – up to 2km (1.25 miles) apart outside of downtown. Meanwhile the east-west line has frequent stops through most of its route, with similar spacing in both downtown and the inner city with stops about every 700m apart (0.4 miles). However, both lines are used for long haul commutes, because they are both subways and that is all people see.
    Rather than letting the technology determine what is the backbone of a transit system, we should look at how it is run. 0nce that is figured out, then we can use the best technology to suit its needs.

  22. Jarrett at HumanTransit.org November 20, 2009 at 8:41 pm #

    @ Doug. Thanks for that history! It filled in several gaps for me.
    @ K. Really important point. Yes, if you’re trying to optimise the throughput of a busway at several points along the line, a lot of branching makes that harder. Brisbane just isn’t in that situation. The busway as built has loads of excess capacity. Note that the main capacity constraint is the platform space at stations, not the busway itself, so you can increase capacity still further by running skip=stop patterns and/or expanding key stations.

  23. Jarrett at HumanTransit.org November 20, 2009 at 8:42 pm #

    @ Ben. Yes, I agree completely.

  24. Jarrett at HumanTransit.org November 20, 2009 at 8:50 pm #

    @ Alexander Craighead. Interesting points. I think we have to be careful about reducing any project to a single motivation. All projects are the result of negotiations among people with different objectives, after all. The fact that some proponents of a BRT project want it to convert to LRT doesn’t mean that there aren’t others who want it to work as BRT; in fact, there always are.
    @ Engineer Scotty. I agree that Barbur could easily be a BRT project, particularly because the corridor naturally branches at the city limits into a Tigard-Sherwood branch and a Tualatin-Wilsonville branch. However, the way the Portland Mall has just been rebuilt leaves an LRT stub pointing provocatively at Barbur, giving it an obvious head-start, so to speak, over BRT options.

  25. David November 20, 2009 at 9:05 pm #

    “Of those, only Ottawa’s is anything close to a complete system, and it’s compromised, of course, by the lack of a downtown segment.”
    You do realize that this is utter nonsense? This is the same fiction that’s been peddled here in Ottawa for two decades. We narrowly escaped being saddled with a bus tunnel mega project in the early 1990s. Ottawa is compromised by using BRT at all, not by a lack of an overpriced downtown bus tunnel in which transit users can breathe in diesel fumes and have their auditory senses pounded into submission. We’ve already got a short “tunnel” of sorts at St. Laurent Station. That’s more than enough bus tunnel. Sorry for the digression, but advocating the kind of misery that BRT entails in practice is a bit much.
    At any rate, it’s fiction that the system is compromised by a lack of grade separation downtown. This displays an appalling lack of understanding as to what the actual problems with Ottawa’s BRT really are, though perhaps I shouldn’t be surprised since the likes of John Bonsall keep telling the world that this is the problem. The problem with this conclusion is that it fails to explain why Calgary with its downtown on-street LRT, facing even more cross streets, both absolutely and relatively (shorter block lengths), manages to carry more passengers without anything equivalent to the 2-mile bus jams that Ottawa gets. This conclusion fails to explain why buses will have a green light and can’t move due to the block ahead being full – delays from traffic lights don’t fill entire blocks up with buses. The highway engineers who are responsible for BRT never look at the human element. Boarding and alighting are the sources of the delays, not traffic lights. Watching the downtown platforms for half an hour in the evening – something I imagine that all the visitors to our “marvelous” Transitway don’t do – is quite instructive. Because of the proliferation of express routes – the main selling point of BRT in Ottawa – no one knows where their bus will stop at the platforms. People wander back and forth along the crowded platforms to get to where their bus has stopped. Others don’t bother and simply wait for their bus to arrive at the lead position (the older/frailer the user, the more likely this is). That means buses can stop twice, even thrice. Boarding itself takes too long because of the lack of level boarding (onboard fare collection doesn’t help matters either, but that doesn’t tend to apply to express routes where most passengers are passholders). The result is that dwell times routinely exceed 20 seconds per bus, even for the expresses, and often 30, with 60 not being unheard of if the buses are already fairly full. Yet Ottawa sends through 180 buses per hour – or one every 20 seconds. Ahhhh… yes… see the problem now? Average dwell time exceeds average headway! Queues are the inevitable result. The irony of the criticism against the traffic lights is that they actually introduce platooning to the system, allowing things to work better than they might otherwise do. They force triplets of buses to arrive on 55 second intervals, sort of like a, well, umm, A TRAIN! rather than completely randomly.
    Still, I wanted to disprove conclusively the notion that grade separation with 4-lane stations would solve the problem. So one day I asked myself the following question:
    “Is there a station on the Transitway, beyond the influence of downtown, that sees anything like the passenger volumes of the downtown platforms, and how well do they work?”
    Now it’s impossible for any station outside downtown to have the volumes of a downtown station, but this is “good” if they show signs of failure: it means of course that they would be failing at the higher volumes seen downtown (it’s not “good”, of course, for the passengers, but then BRT has never been about that). The reverse, of course, is not true: if the lower-volume non-downtown stations did not fail, it would not be possible to conclude that they wouldn’t under downtown volumes.
    So back to the test. For those not familiar with Ottawa, a couple of stations fit the bill: Campus Station, still “downtown” but fully on the Transitway and – most importantly – no traffic lights beyond it on the outbound direction in the evening. Any outbound queuing at this station could not be attributed to traffic lights. The other station is Tunney’s Pasture Station, a few kilometres west of downtown, serving a major federal government office park.
    And what did I find when I went to these stations and watched? Well, they both showed signs of failure, and I made sure I did this on a sunny midweek afternoon in April to avoid any “bias” from snow, ice, rain or darkness (since BRT promoters are sensitive to unfavourable test conditions, even though LRT manages fine in them elsewhere…). Tunney’s Pasture was particularly interesting because it’s possible to get an overhead view of how it works (or doesn’t). The first thing to note about Tunney’s Pasture is that because it is still relatively close to downtown, many passengers headed west find it advantageous to buy the so-called “express” (in reality single-seat ride) passes, and, moreover, they have a right to flag down express buses coming out of downtown. Because of this, just like downtown – and what makes it a good test case, the drivers of all buses have to pull up to the platform to check to see if there are any passengers to pick up. This is key – no bus driver can skip out stopping here, just as they can’t downtown. There are two designated conceptual platforms on the long platform for westbound and southwestbound routes, but that’s all. Anything more than that would be unworkable because it would require too much passenger movement along the platform, and, as it is, considerable passenger movement along the conceptual platforms is already required.
    Just as downtown, the dwell vs headway issue rears its ugly head. All it takes for the curb lane to fill up is if some of the 90-series buses are present and a couple of the expresses get flagged and start taking on passengers. Once this has happened, all subsequently arriving buses, of all types, and regardless of whether they ultimately take on passengers, queue upstream of the platform to wait their turn to drive up to and past the platform. The passing lane is not of too much help to the queued buses because any one bus driver in the queue doesn’t know what the situation is ahead of the bus ahead of him. The passing lane is of a bit more use for buses that have taken on passengers and are ready to leave before the bus ahead of them has left… but there’s a price to be paid in that passengers might miss their bus if they don’t deign to run up to the bus but rather decide to wait for it to come to them. Indeed, on a couple of occasions I’ve seen a bus start to leave by moving into the passing lane to pass a stopped bus ahead only to be flagged right back to the platform afterwards.
    Fortunately passenger volumes at Tunney’s Pasture are still low enough and the “pulsing” nature of the buses arriving from downtown due to the traffic lights there that usually, before too long, the station clears out, but this would not happen in a downtown bus tunnel. Moreover, recent tunnel cost-cutting measures suggested by long-time BRT promoters, who are now getting desperate at the prospect of seeing their poor baby converted to LRT, have included the idea of dispensing with the passing lanes altogether. That would remove any opportunity for a station to clear out until the most upstream bus had left. The tunnel is a recipe for an underground bus jam.
    I’m afraid the compromise in Ottawa was not downtown. The compromise was BRT itself. The general consensus in Ottawa is that choosing BRT was a big mistake that has held us back. The sooner we are rid of it, the better.

  26. Jarrett at HumanTransit.org November 20, 2009 at 9:23 pm #

    @ David. When you have a heavily peaked market, as both Ottawa and Brisbane are, LRT options will look good as efficient ways to move the high volumes on the peak, but BRT often looks better the rest of the day, when the frequency matters more than the capacity. There’s no question that on the peak, when many service patterns are operating in the busway and people are waiting to hail particular ones, you can get the problems you describe.
    And of course, only BRT can offer the direct services to many branches, which do have a political value even though I think the value can be overstated.
    Re downtown Ottawa, I watched it during the peak on one sunny afternoon and concluded that it doesn’t work because buses don’t have clear access to the next lane out from the curb. I worked on several studies of double-width transit lanes working on the model of the Portland transit mall, and they move 180 buses/hour without causing anything like the jamming observed in Ottawa. They do this by dividing the stops into four types, generally bundles of routes going to the same general area, so that only 1/4 of the buses stop at each stop and the other buses pass them. Buses stop in the curb lane but travel in the next lane out, so you never get the mess that Ottawa sees.
    In the longer run, all the strategies used to reduce dwell times on light rail can be used on buses, such as off-board proof-of-payment fare collection that allows passengers to board and alight at any door.
    Rail will solve the problem in a different way, and I wish Ottawa the best in that transition. I have no particular passion for BRT; I’m just interested in what works.

  27. Jarrett at HumanTransit.org November 20, 2009 at 9:33 pm #

    @ Stephen. Really interesting comment about the Seattle Bus Tunnel. I agree that its passenger throughput may decline initially if and when the last buses are removed, although this should only occur in the context of an LRT network sufficient to replace many of those buses with feeders, and thus serve the same markets. Note that the bus tunnel has traditionally been used by bus routes that have a high likelihood of being replaced, entirely or in part, by a robust rail system including Bellevue, U-District, and Northridge. Meanwhile, the inner city LRT segments are carrying passengers on LRT whose previous choice was a bus on the surface.
    The Portland Mall is a good example of infrastructure being progressively upgraded from bus-only to bus+rail, with as yet no serious prospect of the buses being removed because the remaining buses serve mostly inner-city fabric where no LRT project is in the offing, though streetcars may be someday.

  28. David November 21, 2009 at 9:45 am #

    The 90-series trunk routes in Ottawa get good ridership throughout the day, well above what many American LRT systems get at peak. Buses are often standing room only by the time they get downtown. Lower daytime volumes do not justify BRT in Ottawa. Indeed, the all-too-common problem of buses arriving in pairs occurs here throughout the day. We’re getting LRT frequencies at BRT costs and quality of service.
    Skip-stopping would introduce an inordinate amount of weaving. Instead of 180 buses going in one lane in one direction, there would be that many buses shifting lanes on every other block (alternate blocks would be the only realistic way to do it – grouping in 4s would be even more of a nightmare). Then just add some snow to the mix and the fun really starts. And what exactly happens when a curb lane in a single block fills up completely, which it will? Then you’ve got buses in the next lane over with nowhere to go – they’d have to hang back at intersections with green lights waiting for the curb lane in the next block to open up. Since no other bus system in North America or Australia carries the kind of passenger volumes that Ottawa does, the problems just do not rear their heads. What might work at low passenger volumes in Portland is not guaranteed to work in Ottawa.
    A lack of an extra lane to pull out into isn’t the problem in downtown Ottawa. As I went to some pains to describe, we’re already seeing signs of the same problems at much lower-volume 4-lane Transitway stations outside downtown. The extra lanes just do not help nearly as much as claimed. Hurdman Station, the major hub with asphalt as far as the eye can see, also experiences problems from too many buses queuing and extended dwell times. Even the system’s designers have figured they could build a tunnel with only one lane per direction. They said this because even they know that the passing lanes just do not get used much in stations where every bus has to stop anyway.
    Even if it does work, the requirement to use 2 lanes per direction puts the lie to the claim of 10,000+ pphpd per lane for BRT. If you’re going to put 4 lanes to work for transit, you might as well just make it as two 2-way systems rather than having crazy amounts of weaving. Four lanes downtown is a lot of real estate. With 4 lanes in downtown Ottawa for LRT (plus 2 lanes for platforms, like we already have) we could move over 50,000/hour/direction.
    Lack of off-board payment, again, as I noted, isn’t really an issue for the express buses since most users are passholders. Even without that issue dwell times are still too long on express buses, never mind anything else. Once a bus gets moderately full, it’s game over. Internal circulation is abysmal, and stepping up takes extra time, especially in winter.
    Indeed, winter is an entire issue unto itself. In winter, not only can stepping up onto a bus take extra time but so can stepping off, since passengers have to take care not to slip or lose their footing as they transfer weight. Hopping over snowbanks is another source of delay. Not only do we get bus jams downtown regularly in the afternoon, we now get them regularly on snowy mornings too. We dump copious quantities of salt on the Transitway to keep it clear of snow and ice, which erodes infrastructure at an alarming rate, all because buses have poor traction (they require hard rubber tires but snow tires are usually soft rubber and no one has yet to invent a bus snow tire) and rear-wheel drive artics tend to jackknife. We’ve had days when 20 artics were stuck in snowbanks and wrapped around light poles. That’s more artics than many other transit systems have in total! The so-called “flexibility” of buses becomes a sick joke when artic after artic gets stuck in a jackknife as a result of each trying to get past their stuck colleagues. A couple million dollars are spent annually towing stricken buses. Considering the problems that winter causes to bus operations, it’s verging on being insane that BRT was even in the running for rapid transit in Ottawa (we’ve known this from the time when the streetcars were replaced with buses; the newly-introduced buses ran into problems never experienced by the streetcars). Rail systems just do not suffer anything like the number of problems in winter that bus systems do. Level boarding, even if the platforms are somewhat snowbound (~10cm) avoids the problems of having to step up/down with buses. If the trains are kept running while the snow is falling it just can’t accumulate on the rails, and we won’t need to poison our waterways with salt to keep the trains moving.
    If you’re interested in “what works”, then at the volumes Ottawa sees and the conditions present here call for something other than BRT, because BRT simply does not work and no amount of tinkering is going to make it work. Above 5000 pphpd BRT should not even be considered. “BRT” might just as well stand for “busted rapid transit” given the way it works. It wasn’t until I left Ottawa for Calgary for awhile that it dawned on me just how stupid our BRT system really was compared to LRT-based systems.
    A final interesting tidbit is to listen to conversations people have about the transit system when they’re stuck in it. In places like Calgary the complaints are about the system not being extended fast enough and rush hour crowding. In Ottawa, people complain about the system still being run on buses. Few Ottawans speak of the Transitway with any pride, unlike the CTrain in Calgary which has become part of that city’s image. Technically, once it is converted to LRT, “the Transitway” could live on in name since it is nominally technology-neutral, but I suspect the term is now regarded so poorly that it may be irredeemable. The common view is that BRT and the Transitway was a big mistake that we’re still paying for. It was interesting that virtually no one supported the concept of a bus tunnel or a shared bus and rail tunnel as a solution to the downtown transit problems when our TMP was recently revised. They just wanted the buses replaced.

  29. Jarrett at HumanTransit.org November 21, 2009 at 2:08 pm #

    @ David. I appreciate the detail in your comment, but it sounds like you’re thinking on an Ottawa-Calgary axis in which Calgary automatically becomes the answer to everything that doesn’t work in Ottawa. Calgary has been a particular success for light rail, but there are other solutions to the problems you were describing, and other possible outcomes.
    The two-lane busway is a well-established concept that works well in all of Portland’s weathers, which include ice and light snow though not deep snow. If it weren’t working, the redesign of the mall built last year wouldn’t have retained it. In fact they went much further, and actually built LRT in the same weaving pattern. I spent several years on a re-think of downtown Minneapolis which is now creating exactly the same arrangement on two of its major downtown streets. They do have winter, last I checked, and aren’t concerned about that. Indeed, I can’t see how winter conditions are worse for a weaving arrangement than for the horrible single-file arrangement we see in Ottawa, as long as the streets are sanded.

  30. EngineerScotty November 21, 2009 at 2:39 pm #

    David a question: Ottawa is doing 180 busses per hour, when the busses are artic; you make the comment that “about 5000 pphpd, BRT shouldn’t be considered”.
    180 busses per hour, assuming artics full to design load, isn’t 5000 pphpd, its closer to 20,000.
    At any rate, Portland’s mall sees 180+ busses per hour, plus 10 trains per hour, without any problems. One big difference is the weather (we don’t get heavy snow like Ontario does); the other big difference is that we’re not running artics, just standard 40′ busses. With standard board-in-the-front-and-pay-the driver, exit-in-the-back operation; no high-volume payment and boarding configurations.
    Every fourth block or so is a train stop, which consumes the entire block, and the trains weave to the right and busses to the left. Other blocks are bus blocks; and as Jarrett mentions, these are divided into 4 zones. There have been a few complains about having to walk up to two blocks up or down the mall to reach a particular stop, but with the 200′ (60m) blocks in Portland, it isn’t that big of a deal, unless it rains. There’s no signal priority on the mall–and it works fine.

  31. Alon Levy November 21, 2009 at 5:27 pm #

    When the buses bunch, only the first bus fills to design load.

  32. David November 24, 2009 at 12:03 pm #

    @Jarrett
    This city is unusually full of BRT-loving transit professionals (that would be engineers, since none of the actual planners think much of BRT) and they have every incentive in the world to figure out ways to make BRT work better. They’re either a completely uncreative bunch (which, unfortunately, I wouldn’t rule out) or they’ve determined that nothing else is going to make it work. The best they’ve come up with is to keep curtailing the number of routes and replacing standard buses with artics. They’ve now been forced due to political forces beyond their control (a pissed off population) to give up on the concept of a tunnel.
    I compare to Calgary for several reasons. For one, I’m familiar with it. But more to the point Calgary is a fellow Canadian “second tier” city facing similar operating conditions and similar passenger volumes (indeed, it now carries more than Ottawa). The difference
    I just fail to see how weaving is going to be any better. You still have to find a way to put 180 buses through a single point in space, and you’re now doing this while turning rather than just driving in a straight line. Assuming stops in alternating blocks, in any given block during an hour each lane will be visited by 90 buses, which then has to trade places with the buses in the other lane in the next block. It’s not like the passing lane is going to be empty and buses can just turn into it. Buses in the passing lane are going to have to stop and wait for buses in the curb lane to pull out into it, or buses in the curb lane will have to wait for the passing lane to clear. Either way someone is going to have to wait. It also sounds like a recipe for accidents. I can sort of understand the idea in theory but in practice both lanes are just going to be jammed up. Once the number of buses queued for one station exceeds the length of lane space allotted to them (i.e. the length of the block upstream of the station)*, they’ll have to wait in the next lane over until the queuing/platform lane opens up. And that waiting in the next lane will in turn jam up that lane. The problems are still at the stations with dwell times.
    And what of the poor drivers who are expected to carry this out? They already have enough grief to put up with and you’re going to saddle them with this too? Driver turnover could not be expected to go anywhere but up, and it’s already pretty bad. Up go OC Transpo’s labour costs once more.
    Rather than simplifying things by switching to LRT, you’re just escalating the complexity so as to maintain a high-cost system.
    As for snow, well snow just slows everything down when it comes to buses (and cars too). And it’s not just snow – with sand and salt and vehicles it becomes slush. Stopping distances are increased and acceleration is reduced, as is speed. That increases dwell times (if measured from the time when the front of the bus first arrives at the stop to the time the back of the bus passes the stop, plus some minimum following interval of maybe 1-2 seconds). As for why weaving will be worse, with rear wheels stuck in 3-4″ of slush next to a snowbank, those artics could spin out and jackknife while trying to turn – the rear wheels want to keep going straight once they finally grab a piece of asphalt but the bus is trying to go left. Since the front wheels don’t have much traction or weight on them and the road surface on which they’re trying to turn is a slushy mess, the bus slides on its front tires and jackknifes out into the road. We see this today when a car or taxi or other bus gets stuck and a bus has to try to go around it. The City of Ottawa has had these problems on arterial roads with bus lay-bys for stops; turning out into traffic proved to be a problem for the artics in many of these (in addition to the fact that motorists won’t stop) so they’ve been gradually removed and filled in. Even the 40 footers are not immune to getting stuck, though they don’t jackknife.
    Let’s be serious about this. No other city on the continent takes snow clearing as seriously as the City of Ottawa. We’ve got fleets of snowclearing equipment that would put entire states to shame. Transit routes get absolute priority. The Transitway has massive wide shoulders for snow storage and it is salted to an inch of its life while extensive drainage systems carry away all this muck. Snowbanks are cleared away overnight from all major transit routes. And despite this we still cannot prevent massive delays and the occasional jackknifing in the downtown on a day of a moderate or better snowfall. That’s how vulnerable BRT is to snow. And LRT? All that needs to be done to keep it running in snow is to keep a pair of 2″ wide rails separated by 4’6″ relatively clear of muck, something that happens by default if the trains are running during the snowfall. And even if they aren’t fitting them with plows (especially directly ahead of the wheels) is an option while you can get away with quite a lot if the train is long enough and the wheel wells have sanders. Ottawa’s O-Train (which is at the top of this page) operates reliably every winter while the BRT system falls to pieces, and yet its track only sees snow clearing once or twice a winter, usually just to move snowbanks further from the track, since the O-Train has a plow (not visible in your picture since you’ve cut it off too high). It’s a no-brainer, really. BRT is a bad enough choice anywhere else, but in a winter city like Ottawa it is downright stupid.
    * Don’t delude yourself for one moment into thinking that a block can’t possibly fill up with buses waiting for one stop in an alternating stop system – to the west of downtown there’s a bit of a hub station at LeBreton (a 4-lane station) which is separated from downtown by something like 500 m, and yet that entire length routinely fills up with buses waiting to go through LeBreton to pick up passengers. LeBreton even has a divided platform structure going west sort of like you’re advocating with different buses stopping either side of Booth Street, and that doesn’t prevent the bus jams from occurring.
    @EngineerScotty
    180 artics can’t carry 20,000 pphpd. That’s ~110 per bus, or over 6 per lineal metre. Even supposing one could jam that many people into a bus, the dwell times would be astronomical.
    I mention 5000 pphpd because if you’re expecting that volume then you’re well past a level at which LRT makes sense.

  33. EngineerScotty November 24, 2009 at 12:21 pm #

    David,
    The snow may well be a deal-breaker for BRT in Ottawa–Portland only gets occasional snow. In most cases, the response of TriMet (the transit agency) is chains–bus service slows down but the busses do stay on the roads. (Portland, especially the West side, is rather hilly; which aggravates weather-related issues).
    But we do 180 busses *PLUS* ten LRT trains, per hour, just as Jarrett describes, with no problem. Bus traffic jams are simply not an issue; its when busses leave the Mall and enter the regular street network that service is most likely to be disrupted. You may think it won’t work, but simulations (and actual experience!) suggest otherwise.
    Not having artics probably helps, as we can fit four stopping zones per block, with 40′ (10m) busses.
    Going to LRT (or to mixed-mode operation) may well be the correct decision for Ottawa. But it sounds like quite a bit of the problems of the Ottawa busway are related to operation practice.

  34. David November 24, 2009 at 5:44 pm #

    How many passengers is this transit mall in Portland carrying?
    You say it works and I have to accept that but when I try to envision doing this here I just can’t see it working. All it would take is for one block to fill up with buses waiting to stop at the station in that block (i.e. it backs up to the previous alternate station). I have absolutely no reason to believe this won’t happen based on current experience. A second lane for passing isn’t going to help you then because any more buses trying to join that queue won’t have anywhere to wait but in the passing lane in the previous block. My impression is further bolstered by the fact that where we do have 4-lane stations and higher-than-average passenger volumes we see signs of failure and in some cases actual failure.
    Just watch this YouTube video taken at Hurdman Station (looking east down the East Transitway, Southeast Transitway is to the right and downtown is behind), which is entirely on the Transitway and where there is way more than one passing lane available. As major as this station is, its passenger volumes are nothing like as large as downtown. There are plenty of occasions when buses cannot enter the station behind the camera operator not because it’s not their turn but because the way is blocked by buses stopped at the station itself. At 4:20 there’s just the sort of cross-over weave that you guys are suggesting. Look at the queue that has formed on the East Transitway by 5:00! The SUV at 7:30 probably belongs to one of our transit engineers, err, “planners”.
    http://www.youtube.com/watch?v=JVUOkdFrIEk
    Here’s another by the same guy (no, not me, and I don’t know him either). This is at the Rideau Centre. You can see as people run around back and forth to catch their buses. It’s interesting towards the end as he goes for a walk along the platform and you can see that the 90-series buses dump their passengers before they even get to the platform at all. Moves like this on the part of the drivers speed things up to be sure, but that’s not the way it’s supposed to work either. BRT is just not designed for humans and the sentiments of this videographer in his comments are pretty typical of the views of people when you ask them. BRT just has to go. It was an interesting experiment but it’s time to declare it a failure and move on to a system that actually works.
    http://www.youtube.com/watch?v=cLOGuPoftAI