« good question of the week: transfer penalties | Main | melbourne: don't miss the rock-star of parking »


TrackBack URL for this entry:

Listed below are links to weblogs that reference dissent of the week: my alleged "bias" against rail:


Feed You can follow this conversation by subscribing to the comment feed for this post.

Robert in Calgary

A few early morning observations.

Emotions are involved. Carl was clearly prompted to write out of emotion.

Everyone has biases of some sort. The question is, is the individual aware of them and factors them into his own thinking. I have a clear bias for brunettes, I'm aware of my bias and I don't let it restrict myself to only asking out brunettes.

You know you're popular when people like Carl are going to fight so they can say you're on "their" side.

While it might be nice for each person if you share their bias, if you're writing interesting articles, it's not mandatory for us to agree on everything for me to get something out of what you're writing. (if I was more awake, I could probably phrase that last one better...)


Dear Jarrett,

- BRT-systems are not less expensive than rail - only when you don't have special infrastructure for them, but I would not call it BRT then. Just the contrary: Rail can stay in for 30 years plus, while you have to renew a busway much more frequent. The energy efficiency of a tram is much better than of a bus, so that overhead wires cost in comparison to the saved money nothing.

- People base decisions on emotions. Munich is home to BMW, and at BMW they live on that. Don't expect that you can re-educate them. When they have the choice between a BRT system and their car they take their car. If they can pick between a rail system and their car, a significant number picks rail. It is status and comfort. And emotion. But it is real. If you would sell your products in a supermarket you can waste a lot of money in marketing trying to sell a product that people do not want, even if it is in your eyes wonderful or perfect. Or you can give them right away a product that they accept and buy.

- You write: "Other cities, notably Karlsruhe and Munich, seem more invested in maintaining a feeling of difference between bus and tram. Berlin displays impulses in both directions."
Well, Karlsruhe (pop. 280.000) has a ridership of 97 mil. trips per year by tram and 14 mil. trips by bus. Still there are fast bus lines, stops where buses meet at the same minute to let people change between the lines etc.
In Munich there is a bus ridership of 171 mil. bus annually, about 600,000 on a weekday, not a small number for 1.3 mil people in the city.
The bus system operates with a modern fleet of about 400 buses. Mostly articulated buses, boarding on all doors, ticket sales by machines on the bus and not by the driver. Signal priority. A lot of bus stops built in the way of curb extensions, bus lanes, buses sharing the seperate right of way with the tram etc.
Also Munich was the first city in the world to introduce low floor buses in daily service during the 1980ies. And it is hard to find a bus older than 8 or 9 years in operation.
Munich tries a lot to make the bus attractive, and to a certain extend we are successful, when you look at the high amount of riders in the bus system. Still we get more riders when a bus is replaced by a tram. And usually a new tram line in Munich is not built because our buses cannot handle the amount of passengers anymore, but because we generate a lot of more riders by rail.
Berlin, by the way, has about 112 trips by bus per inhabitant. Munich has 131 trips by bus and year and inhabitant. Despite a dense rail system.
What are the numbers for cities with BRT systems in the western world? Please impress me!

Martin H. Duke

I think the practical obstacles to making a bus system uncomplicated are pretty steep. I don't think anyone, in any area, confuses the bus and rail service in their region. Meanwhile, in many areas bus service already involves several overlapping brands.

Even in the case where there is one bus operator and therefore space for a high-quality brand, people who aren't paying attention will have trouble distinguishing these.

And of course, here in Seattle, the bus-vs-rail debate is particularly fraught. Most cities have already built rail in the obvious high-density corridors and so the case for additional rail is relatively marginal. In Seattle, the system is still a relatively low-ridership stub with the really high density stuff still to come. Yet there are still lots of people that would rather spend much less to improve low-quality bus service.

Eric Doherty


I think you are explaining one of the advantages of rail incorrectly, or at least not in clear direct language.

"Capacity is one of the best technical reasons to build rail. If you routinely need to move more customers per driver than buses can do, you need rail so that you can run much longer vehicles. Capacity is the reason that the Los Angeles Orange Line probably will need to be rebuilt as light rail at some point."

A more accurate statement would be:

"Cost is one of the best technical reasons to build rail. If you routinely need to move very high volumes of passengers, rail allows you to run much longer vehicles and reduce labour costs for drivers without having unacceptably long wait times between vehicles. Labour cost is the reason that the Los Angeles Orange Line probably be rebuilt as light rail at some point."

PS Anyone who thinks Jarrett is biased against rail should really hate my recent post on BRT


Cost is an important factor (and in many parts of the world, the most constraining factor at the present time), but cost equations change. What if, for example, the present financial crisis in the US were to lead to massive pay concessions by transit unions (or their being busted outright)? The cost/benefit tradeoff between bus and rail might change quite a bit.

The capacity issue between rail and bus is more fundamental. It might change in the future--I could see articulated busses in the future with computer-controlled steering and powertrains, allowing busses to be safely entrained (and operated by one driver) without needing a fixed guidway. But such technology is not available today.


I'm probably the least knowledgeable of anyone here, but arguing that because people don't want something so you shouldn't recommend it seems to be surrendering your role as a transit expert. Shouldn't you recommend whatever's most suitable for a given situation, regardless of the public's biases? Otherwise won't transit planning become a short-term political decision, rather than a rational process that takes the long term future of the city into consideration?


@Eric Doherty - yes, cost is a bit factor. But absolute capacity is still a limitation for buses. Taking minimum headways into account, long trains take less space than buses carrying the equivalent number of buses. So buses can never provide as much capacity as trains, irrespective of how many drivers you're willing to pay.

Canadian Veggie

Relevant to this debate: Winnipeg has just voted to switch its BRT system to light rail midway through construction. http://www.globalwinnipeg.com/City+approves+light+rail+plan/3306361/story.html

The costs are significantly higher, but the mayor is better on people's emotions to help carry the project.


OK, that last post had some pretty bad mistakes in it (note to self: PROOFREAD). But I think you get the general idea...


@J B,
it seems you are a supporter of Planned Economy.
What is totally okay. A lot of people in history liked the idea of planned economy and it seems sometimes it makes a lot more sense than the waste of a market economy, where you offer so many choices and many of them are just a waste of energy and resources.

For 40 years they were experimenting with planned economy in a part of my country. Until it collapsed.

But reality is that we live in a market economy. Most people do have a choice, they have cars available. And when you do want to get people out of their cars you have to listen to what they want and do what brings them out of the cars.

BRT works in countries like Peru, Colombia and such, where car ownership is low and the need for some kind of well organized public transit (any transit) is high. There you can fill your buses with BRT systems fast, as long as the tickets are affordable.

But it does not work in western countries. I would be happy if you could give me an example of a successfull BRT-oriented city in a western country, where buses attract more than just the usual 5% to 10% or modal split. Where you have, like in rail oriented cities like Munich, 300 or 400 trips by public transit per inhabitant and year.

You could argue, it is a bit conspiracy and dense cities with high potential have already a rail system. But when I examinded cities in Germany with bus only systems I found a similar density and population but in average a lower ridership of public transit. Despite a lot of effort.

For me, and most planners in Europe, it is quite rational to attract as many riders to public transit as possible, by giving them what makes them come: rail.

And every rail project tells the same story of rising ridership.

That is even true for a complete country like France, where the re-introduction of the tram brought public transit ridership to before unknown numbers. It literally exploded. Even if they had quite neat bus systems before.

Light rail / trams have a lot of advantages. And it is difficult to see why BRT should be superior. The main argument I hear is that a tram could not go around an obstacle. Hey, such situation happen much less than you might expect. There are more technical breakdowns of buses that cause a problem than trams stuck somewhere.

We have in Munich some parts of our bus system that are pretty much like a BRT system. But when you ride the bus there and after it go by tram or metro you know why people prefer rail.


@TransitPlannerMunich. I'm not aware of any of the Munich bus routes that really take significant advantage of geometric improvements. Can you point me to any specific routes?

@Jarrett. I think the last important point to make is that there's a lot of other factors that influence mode decision.

Sure, people love riding rail, and on a given route if you swap a bus for a train, more people will probably start riding it. But, if you have an express bus option that saves you 10 minutes off your rail journey, you're probably going to take the bus! While comfort/style is great, if the amount of money spent on building trackway overrides the improvements in geometry/speed/dedicated trackway/conflict reduction/etc., it might not be worth it.

Here in Baltimore, I can ride the free, cute, smooth-riding, hybrid Charm City Circulator, or I can pay $1.63 to ride on the grungy, dark, jerky, loud, MTA bus. But the MTA driver guns it and gets me to work in 10 minutes instead of 15 or more. I vastly prefer the MTA. Unfortunately, the fact that I have to make exact change every time also tilts my decision--smart cards due soon, though!


I'm actually not for planned economies, but I don't think public transit systems should be solely planned around short-term profit (or short-term ridership). (Never thought a West European would call me a Communist!) Perhaps lower mode share in the short term would be offset in the long term. Perhaps, if the product is high-quality, people will use it despite previously claiming they won't. I know you've said that's not the case with high-quality BRT in Europe, but I'm not arguing for or against buses, I'm arguing that just telling people what they want to hear seems to negate the purpose of a transit planner, or any sort of expert in any field.
Unlike East Germany's government, I'm not saying you should force people to accept something they don't want. Obviously it is impossible for a transit planner to do this in a Western democracy. But so long as you're not forcing people, you may as well make an honest recommendation of what works best, and if the public or the government ignores you then that's their problem.

David in Ottawa

As an Ottawan, I can't help but think that Jarrett has a somewhat pro-bus bias. Certainly not the burning bus bias (sorry - we've had some buses spontaneously combust here...) of certain BRT fanatics like our former Regional Chair who spends his retirement touring other cities in Ontario telling them not to build light rail, but there is a bias nonetheless.

If there's one place in the world that ought to switch from BRT to LRT, it's Ottawa. We have ridership levels that are so high that in any sane city there wouldn't be a BRT-LRT debate, but nonetheless Jarrett's take on converting to LRT is distinctly unenthusiastic, accompanied by talk of adding yet more lanes to the downtown surface bus route and bemoaning the lack of a tunnel, as if this would solve anything.

Frankly, this is the problem with BRT supporters, even the milder ones. Ottawa was regaled with tales of how we would convert to LRT when ridership was high enough (nevermind that it already was) and what have we had from the BRT promotion industry this decade? Trenchant, unrelenting opposition. Every attempt to address Ottawa's failing BRT system by replacing the heavily used central portion with light rail was opposed right up until 2008, when they decided the solution was a light rail tunnel, skipping the surface light rail solution (that way the BRT promotion industry can always make the claim that a tunnel was what was needed since they would have looked pretty daft if a surface light rail system succeeded where a surface BRT system had failed). You might think that BRT promoters would want to have a city to showcase the "build-as-BRT-convert-to-LRT" model they espoused a generation ago, but no, they oppose it out of some deep-seated belief that BRT actually is better. We have a former transit general manager - apparently he saddled Brisbane with busway, or somthing - showing up to recycle old idiocies like bus tunnels, freshened up with claims that hybrid buses would make it all work without heavy ventilation and noise (yes, we'll run hybrid buses on 50 km cross-town journeys at over 80km/h... yes, there's a good use for hybrid buses. Oh, and you want to put guidewheels on them too? How delightful. Let's see them last one snowfall.).

All that this kind of nonsense is doing is showing that the 'BRT now, LRT later' argument was a sham, a deliberate deception to gain wider acceptance for BRT when what everyone but a small cadre of bus-obsessed highway engineers wanted was LRT. The BRT promotion industry in Ottawa has been dragged kicking and screaming into being forced by the populace to accept light rail over a relatively minor portion of the network, and they only agreed to it when there was enough money on offer to put a tunnel in place. Had Ottawa just converted to light rail this decade, all these embarrassing pictures of bus jams in Ottawa simply wouldn't exist and the BRT promotion industry would actually have a success story to point to of BRT-to-LRT transition. They could claim that the story they'd be spinning all these years was true - "just look at Ottawa", they could say. Instead, all they have is a monumental failure and BRT is losing credibility fast. Ottawa is now used as an example of the failure of BRT, and rightly so because it is. It serves them right - they were idiotic enough to think BRT pushed to extremes could work.

Other cities see our experience, and want none of it, so they're going for LRT from the outset. Edmonton cancelled any notion of building busways a couple years back, and now Winnipeg is likely going to follow suit. Calgary only implements BRT on the cheap so as not to lock up infrastructure in a non-desirable mode.

Another lesson from Ottawa is that busway infrastructure has this strange way of creating an entrenched bus promotion lobby within the highway engineering consultancy and within the transit agency. The former get addicted on busways while the latter get bedazzled by their own models of moving buses around and constantly "tweaking" routes and schedules to minute changes in demand. It got so bad in Ottawa that when the O-Train pilot project was launched, the rail consultant (from Montreal, naturally, since we had none) charged with getting the project underway was told that no one - NO ONE - at Ottawa City hall knew ANYTHING AT ALL about rail or even the corridor in question. He was told to go talk to the local rail transit advocacy group that had managed to get the O-Train on the agenda, since they were the only ones who actually knew anything.

I'm actually more anti-BRT than I would otherwise be *because* of the shenanigans of the BRT promotion industry in Ottawa. I can see the advantages of BRT in some cases from an "academic" point of view, but political realities are that BRT comes with a lot of undesirable baggage in the form of BRT obsessives who entrench themselves in positions of power and influence and refuse to see reason. The upshot is that I would be VERY VERY careful in recommending BRT anywhere else. I think this is unfortunate, but that what the BRT promotion industry does: throw themselves under the bus.

If Jarrett was really free of a pro-bus bias, he would have concluded as I did that whatever the merits of BRT for Ottawa back in the 1980s, by the early 2000s the gig was up and the City should have spent this decade converting the Transitway to LRT rather than putting forth a few more wayward comments about bus tunnels and turning over yet more lanes to a failed concept.

Chris M

I think partly what people are seeing as Jarret's bias is actually just him seeing things rationally for the contexts he works in. In low density new world cities bus rapid transit has much to recommend it over light rail. Low density means a service that can fan out to various destinations is highly desirbable. It also means there isn't as much as a restriction on the width of rights of way (as you would get in dense European inner cities). Finally with the car dependency it has fostered it means that the road lobby who will oppose full segregation needed to bring out the full advantages of rail are stronger. In that context you need a compelling case (largely capacity based) to make rail the better choice. In Europe the opposite is more commonly true.

The problem for transit planners and transit/urbanist advocates is that the latter don't want to inhabit the low density new world cities they do, but the denser cities of Europe and the US east coast. Rather than seeing transit as a tool for moving people around the city that IS, they see it as one to remake it as the city they WISH IT TO BE. Both perspectives can be valid in different contexts. Sorry I was going somewhere with this, but I lost it.


"if you have an express bus option that saves you 10 minutes off your rail journey" then it's very likely that you also have a car route that also saves you a good deal of time over the rail option. And many (but definitely not all) people would choose to drive instead, because of issues like span of service, flexibility (what if you want to break your journey partway through to, say, go shopping?) and so on.


Chris, actually, in low density new world cities, bus rapid transit has almost nothing to recommend it over light rail - as long as you mean North American cities, that is; where the population is affluent enough to drive.

BRT has produced paltry ridership gains from choice commuters compared to LRT in those low density new world cities - except, again, for the centrally planned poor places in the Southern Hemisphere.


@Chris. I agree. For lower-density areas in North America without very much historical constraint on right-of-ways it breaks down to:

rail vs. bus = romantic vs. practical

Obviously not true in bigger, denser cities, or in Europe. Land use is key here. The existing land use and urban form dictate what the best transit is. I think that's also Jarrett's point on why we should integrate good bus and good rail service conceptually, because there are many circumstances where bus will be the best option.


Montreal has 1 Million weekday riders on the metro, and 1.3 Million weekday riders on the Bus. Montreal Island (where the stm operates) has 2 Million people. You do the math... (source)

I feel that Jarret as a bias, I view him as an advocate for the possibilities of bus travel (The Karlsruhe Model article comes to mind, pretty dismissive of the possibilities because he didn't like some details of their network).

It wouldn't really matter so much, if he wasn't so adamant about saying that he does not prefer any technology over the other. But instead he even uses this argument as a pro for bus.


@Chris, tomtakt:

Your conversation misses on piece of info: what type of BRT?

Much of the BRT build in "low density North American cities" is the light type (no exclusive ROW), which has a dramatically different service and cost profile than does LRT, which has to be exclusive-ROW. Many motorists will squeal just as loudly if traffic lanes are converted to bus-only use as if they are turned into tracks.

But in most lower-density cities, there isn't generally sufficient demand to drive LRT on any give corridor, at least not from a demand-only perpsective. There may be other reasons to build it--such as a desire to transform land-uses, or for environmental reasons, or to attract the won't-ride-the-bus crowd; but many of these reasons are frequently controversial.



Okay, it's great if you get a couple good rail lines, but if it cost you all the improvements in the rest of the system, it still makes getting around without a car hugely difficult.

Improvements that create a better network can be so much more valuable! It's important to consider the real challenges of each transit system and make practical design choices.

My conclusion: it all depends!


@Chris, @Scotty It's kind of like you guys are thinking about a completely different set of cities than I am, I guess. When Phoenix can be north of 40,000 boardings/day, joining Houston as the only one-line LRT system pulling that off, haven't we proven that rail versus bus makes MORE, not less, of a difference in low-density North American cities?


I just crunched some numbers for rail vs. BRT in San Jose. The surface-running parts of light rail on arterial streets outside Downtown, primarily North 1st and Capitol, runs at somewhere between 17 and 22 mph on average and has a stop spacing of just about 0.5 miles, close enough that anywhere along the line you can easily walk to a stop. The "BRT" (it's got some red paint on it) Rapid Bus on El Camino, route 522, actually manages some decent average speeds too, around 15-17 mph in the less heavily congested suburbs. But it does this at the cost of stop spacing, which is around 1-1.4 miles. So, light rail runs faster while providing a more convenient service (more stops), and attracts more ridership, all without grade separation, just running in street medians with signal priority at most, but not all, intersections. And best of all, the travel time really doesn't vary with time of day, so it's just as fast during rush hour. And of course the light rail has a higher capacity and comfort level (ride quality, space) than the bus. I'm actually not sure what my point is here, besides the fact that rail can really be a practical solution in low density American cities.

Alon Levy

@TransitPlannerMunich: Curitiba, the birthplace of BRT, actually has a high car ownership. If I remember correctly, its car ownership isn't much lower than the US average. However, car use in the city itself is low, as people take BRT.

The advantage of BRT in Latin America doesn't come from low car ownership. On the contrary, Latin American cities have high car ownership relative to their income levels. It comes from very low labor costs, and very high technology costs. In 1980s' Curitiba, LRT would have cost 10 times more than full-fat BRT did, and would not save enough labor cost to be justified. In today's first world, where technology is cheap, LRT costs about twice as much as half-assed BRT, and labor costs are much higher, so LRT can be better.


The total US car ownership rate is 780 per 1000 people. Curitiba has, according to commonly cited statistics, 333 per 1000. And has anyone bothered to ask which way the causality goes? Perhaps the implementation of BRT has caused the car ownership in Curitiba to go up, because people want to get out of the crush-loaded buses as soon as they can afford to, while in other cities in Brazil, where people can take a train, there's less incentive to own a car?


I recommend riding Bus 58 (Central Station - Silberhornstraße), or Bus 53 (with an independent right of way between Nordbad and Rotkreuzplatz) or Bus 51 (around Nymphenburg palace or in Fürstenrieder Straße) or the new bus lane of Bus 54 on Candidberg or Bus lines 55 and 155 with its own right of way in Rosenheimer Straße up to Ramersdorf.

What do you want to say with that? Montreal has a successful metro and so in addition a big bus network. About the numbers. Montreal has 50,000 metro riders daily per 100,000 inhabitants and 65,000 boardings in the bus system per 100,000 people. In total 115,000 boardings per 100,000 people a day. Munich has 90,000 metro rider daily per 100,000 inhabitants, 24,000 tram riders daily per 100,000 inhabitants and 43,000 bus riders daily per 100,000 people. In total 157,000 boardings per 100,000 people a day. 36% more than in montreal So Does that prove Montreal can be happy to have a bus system? Or Munich? Or Montreal needs a tram system? Or an expansion of the metro? Or Munich BRT?

@J B:
I did not call you a communist, but your proposal only works in a planned economy. Not only communists use a planned economy. The Nazi war machine relied heavily on a planned economy.
It is the first time I hear that rail is associated by someone with short term profit and ridership. The only way BRT could be successful in a western city is that peak oil is reached and soon driving a car is possible only for a wealthy minority. But then again, we will be happy about all rail built so far.

Jarrett, BRT is the "I can't believe it's not butter" of public transit. You try to suggest to us when we close our eyes BRT almost feels like rail transit. At least while it is not moving. But when I am honest I prefer the real butter, like most people. Especially when the price is quite similar and in the end you cannot even properly bake with the substitute.


@Alon Levy:
Curitiba has about 110,000 public transit boardings per day per 100,000 inhabitants. Munich has almost 160,000 public transit boardings per 100,000 inhabitants per day. So Munich has almost 50% more public transit riders broken down on 100,000 people. And that with a car ownership rate double that of Curitiba.

For me it seems, that whoever has a car in Curitiba will use it. All others take the bus. Like in the US.


Maybe I should add: Curitiba has almost exactly the same population density like Munich. And a 30% larger population.

Jarrett at HumanTransit.org

Did M1EK just call Brisbane, Adelaide, and Auckland "centrally planned
poor cities in the Southern Hemisphere"? None of them is poor;
Brisbane is a natural-resource boomtown. And Aus could use a little
more central planning.


Those "Metrobuses" have some nice improvements, but striping a few sections of dedicated lanes and letting the buses use a mile or so of tram trackage doesn't make it BRT. Those buses still weave around quite a bit from one side of the roadway to another.

I also used to board the 53, where the end of the line stop is a basic, bare-bones bus stop...the only characteristic of BRT it has at that point is being articulated. Yes, the turn around lane and special signal on Leopoldstrasse is cool, but they also did a bad job of rerouting the 53 around the construction on Münchner Freiheit. They would not have rerouted a tram--hence not full-service BRT!

Michael D. Setty

Here is a case to consider...

My former employer, the City of Vallejo, in the Northeast San Francisco Bay Area, operates Vallejo Baylink high speed ferry service, supplemented by a number of express bus runs. The fares on the ferry and supplemental buses are identical. In the AM peak period, the ferries operate 5 trips, and the buses operate 5 trips. Overall patronage is about 3,000 daily boardings, roughly 1,000 in each peak and 1,000 during the midday (midday patronage gyrates wildly, depending on the season).

The buses are usually 10-15 minutes faster in each direction than the ferryboats, though they are somewhat less reliable since they suffer from traffic congestion on Interstate 80 despite the availability of HOV lanes most of the way. Despite a similar number of peak period trips, about 90% of total Baylink ridership is on the ferries, not the buses. A slightly larger percentage of commuters use the bus (15%+/-) during peaks, but most riders stick to the ferryboats.

My point is that I think that TransitPlannerMunich I think is mostly correct: where travel volumes justify rail, then rail should be built. I really don't know why BRT advocates engage in overreach though in absolute terms, the corridors where the various flavors of BRT are applicable probably outnumbers those where rail is useful by an order of magnitude in the U.S. (See some of the research at my website, http://www.publictransit.us/index.php?option=com_weblinks&catid=5&Itemid=8, for example).

Of course, the average ridership per potential BRT corridor is much lower than where rail makes sense, but could it be that for some BRT advocates (I DO NOT mean Jarrett!), egos are involved, like the guy at the "BRT Policy Center, Mr. Bill "Better Rapid Transit" Vincent? http://www.gobrt.org/bios.html

Jarrett at HumanTransit.org

@Michael.  Yes, if I were making the Vallejo-city commute, I'd probably take the ferry too, unless it were rough weather.  Then I'd want to know the bus was there, or else I'd be trying to drive and park somewhere in Contra Costa.

But as you note, reliability is part of that, in nice weather.  A bus that might be 15 minutes faster than the ferry if everything works well isn't something I can count on to get me to work 15 minutes sooner.

I'm not arguing for the total unimportance of the pleasure of the journey, only for the need to keep one foot on the solid ground of travel time and reliability as we indulge our emotions with the other. 


I agree that emotions are a huge factor.

Humans love to divide things into groups.

Rail. Bus.
But things arent group, every group you create can lead to more groups.
Commuter rail. Heavy Rail. Light Rail. Trolley. Express Bus. Local Bus.

And the problem with creating these groups is that you have to assign completely arbitrary specifications. A train is on steel. Ok, so what about the mexico city and paris subways? Trains run on guideways. Ok, what about buses that follow tracks?

Where do trackless trolleys come in?

It's pointless, and I dont understand how some people can be so against buses (or rail) because of arbitrary criteria they decided upon. Like capacity.

Buses arent capacity limited, Ive mentioned it before, but you can attach various trailers to one engine, just like a train. There is no technical limitation. If an engine pulling THREE giant trailers is allowed to move at 60mph on an interstate, I think we can manage doing it on an exclusive ROW.

Another example: streetcar.

Imagine for a second, the very modern, sleek, low floor electric trams that roam europe. Now, instead of two tracks, make it one track and add tires (see Verona). It's now a guided bus! But unless you examine the tracks, you think it's a train.

I think Jarrett is very open minded when it comes to buses and rail because he understands that the limitations are constructed, not fixed. (although his thoughts on water transport aren't as open).

Peoples preference to trains over buses are simply because most people were brought up in a system where rail had much more money put into it than buses. 3 minute headways on rail, 30 minute headways on bus. Thats where the bias was born!


An interesting question to get some data points on in the Vallejo example is whether the availability of the bus increases the ridership on the ferry. People obviously prefer the ferry, but if they know that if they miss the ferry, they'll have to wait 15 or 30 minutes for a bus instead of an hour for the next ferry, they're more likely to take the service. So the ridership on the ferry when there is also bus service available might actually be higher as a result. Which gets to the heart of another point: it's not about rail versus bus. It's about rail and bus versus just bus. Rail proponents for the most part want rail and bus used where they're appropriate (and obviously have a bias toward thinking rail is more widely applicable) while some, though certainly not all, BRT proponents seem to think that buses are the solution to all problems, and dismiss their inherent disadvantages as irrelevant and the culturally ingrained ones as easy to overcome.


Heres a verona bus. At first glance, it's light rail/tram. At second glance, it's light rail/tram.

But look closely, it's a bus. At least with the definition that a bus run on tires.


Michael D. Setty

From Anonymouse:
An interesting question to get some data points on in the Vallejo example is whether the availability of the bus increases the ridership on the ferry. People obviously prefer the ferry, but if they know that if they miss the ferry, they'll have to wait 15 or 30 minutes for a bus instead of an hour for the next ferry, they're more likely to take the service. So the ridership on the ferry when there is also bus service available might actually be higher as a result.

There is no question that the Baylink ferry benefits from having the bus available. The buses allows 20-minute headways combined with the ferries during peak periods, and are also operated when there is insufficient demand to dispatch a 300-passenger boat at a round-trip cost of around $2,000 out of pocket. Now that the ferry fleet is approaching middle age, there are more breakdowns and the buses provide a vital safety net.

One thing we did learn from the Vallejo Baylink ferries: commuters and other local residents will ride the bus when necessary, but seasonal tourist day trippers will take a bus when Hell freezes over.

I suppose some people would say cancel the ferries because you could subsidize a lot more bus runs, but the tourist market would be lost (as would the development potential--which is still there despite the decade-long fumblings of the bankrupt City of Vallejo, but that wasn't OUR department!)

Alon Levy

Anon, is there a link to the 333 figure? In Sao Paulo, there are government statistics saying the city has nearly 600 cars per 1,000 people, and a 55% transit mode share: follow this link and let the Flash graphic load.

Anyway, Curitiba does not have high car use. It has high car ownership, but its gas usage is below that of the rest of urban Brazil. If the BRT system sucked, you'd expect to see the opposite situation: car use and car ownership would be at similar levels, as anyone who owned a car would drive it.

Eric Doherty


"yes, cost is a bit factor. But absolute capacity is still a limitation for buses. Taking minimum headways into account, long trains take less space than buses"

Maybe, but not by much. And how often is that a real factor in the US or Canada? Victoria BC Canada ran into a problem with a shortage of bus stop space for conventional buses, they just went to double deck buses.

If space at station platforms, or length of vehicles in general was a big factor I would expect to see a lot more double deck buses and rail vehicles (yes, very common with commuter rail since the frequency is so low).

The fact is that it is usually not that hard to expand station platforms on the surface, and signal timing is not usually a limitation to the length of platoon that can get through an intersection if you have good signal priority.

My suspicion is that the labour cost issue usually kicks in first. And the longer the bus, the higher the ridership before rail and BRT costs converge.

Jarrett at HumanTransit.org

@all.  The whole "is it cost or is it capacity?" question seems to be a quibble.  What matters is cost per unit of capacity, i.e. "operating cost per passenger" which usually tracks with "number of employees per passenger" in the high-wage developed world, which usually means bigger one-driver vehicles rather than smaller ones.  Keep your eye on that (and one foot on the floor!).

Chris M

@engineer scotty. I was talking about BRT as a general concept, ideally as the fully segregated Brisbane example, but with the ability to compromise this when politically necessarry. Of course car users will be less than pleased whatever they lose lanes for, but in cities where this is a severe political issue and joint running is required then Jarret's point about rails inability to deviate from the track becomes important. In a city that fully or almost fully segregates its rail lines (like Munich) this will rarely be an issue.

M1EK - 40,000 daily boardings is not great by global standards, even for such a limited system. But carrying more passengers isn't always reason to invest in rail unless. What if BRT would have carried 30,000 passengers for half the cost? Two BRT lines could have been built carrying 60,000 passengers for the same amount of money and serving more of the city. Fewer choice riders would have been attracted, but regional mobility would have been improved much more.

The way I see it saying that mode choice should be either or is silly anyway.


Munich's Line 53 that you did not identify as BRT has a total length of about 13 km in each direction. 25% of it is an indepentend right of way. And it has a ridership of 10 million people annually.

That is aleady about 10% of the ridership you find in all of Ottawa's OC Transpo system. And that on a single line here.

By the way, it has bus priority on the complete line.

About the BRT-metropolis Curitiba: it has 60 km of dedicated busways in a total network of 1,100 km and a larger population and size of the city than Munich. In Munich you have on the surface (tram and bus) 74 km of own right of way, usually in the center of the streets.


Why just imitate the tram with a tram on tires and one instead of two tracks?

In the long run it is more expensive, cause tracks last much longer than busways, rubber tires are also not cheap.

When you can afford to buy one track for a guided bus and special vehicles, then you can afford some miles of steel extra for a tram system.

It would be as if you would build an extra large bus with 200 seats, add wings, add jet engines, give it the ability to fly and say: it is a bus, an "airbus", so we do not need planes anymore. Look, it still has some tires!

Or define the Paris metro that is operating on tires, as a successful BRT system.


@Alon Levy:
About Curitiba and it presumed high public transit usage:
Curitiba has a population of 1.7 million. In most larger cities you have a mobility of 3.5 to 3.8 ways per person and day. But lets say, people in Curitiba are less mobile and do just 3.3 trips average per day and person, on weekends a bit less. That are about 1.8 billion ways per year by bus, car, feet and bike.
Ridership in public transit is at about 570 million a year, 32% of the modal split of all ways. Even if people would walk 30% and use for 10% or their ways the bike it leaves 30% mobility for the car. When you have a car ownership of 30% then you could clearly say that every car owner will use his/her car very extensively, even if you take into consideration that they have passengers (family, friends) going with them.


@Jarrett, I was responding to another commenter who referred to cities in the New World - do Australia and New Zealand count as New or Old? I was thinking of the Americas only.

Jarrett at HumanTransit.org

You'll sometimes hear the term New World to mean Americas + Aus + NZ.  European colonisation of Aus and NZ was a couple of centuries later than the Americas.  Of course the whole terminology is Eurocentric, but hey, we're speaking a Eurocentric language.

Brent Palmer

I guess you could also count Westernised areas of South Africa as "New World"? You see the very similar urban/suburban development to Australia and New Zealand.


As a resident of Brisbane I have to disagree on some points:

1. "Buses are better than rail in mixed traffic."

If we look at Coronation Drive and the Captain Cook bridge in Brisbane: Compared to heavy rail (or exclusive LRT) the fact that buses are even in mixed traffic slows them down.

No amount of bus maneuverability will save a bus stranded in a sea of congested traffic in peak hour when it matters most. Accidents causing congestion on the road network (happens weekly) also trap them in traffic.

LRT or trams usually come with priority and segregating measures. Heavy rail has an exclusive Right-of-Way.

2. While buses may look like rail, light rail is also evolving towards metro standards. Automated LRT for example. I know of driverless trains, but not of driverless buses. The capacity of rail is higher and more scalable, and less labour needs to be used.

3. On the versatility of the buses not requiring a connection being a plus- I don't agree with this. The same effect can be achieved for rail if timed feeder bus services and interchange were done like in Toronto.

Brisbane's busway, per kilometre, is comparable to the cost of heavy rail, but with half the capacity of heavy rail. That's quite a cost for "versatility" especially when it can be emulated using timed feeder bus services to rail.

In time the volume of passengers on the busway will grow such that a feeder bus operation will be required, defeating the very logic of the direct trip.

The strength of rail is line haul work. The strength of bus is collection of passengers in low density areas. Put the two together and you have a transfer-based network. This is why I think the bus-or-rail debate is silly.

The role of rail should be line haul, the role of buses should be feeding passengers into the rail system.

Zoltan Connell


32% public transit and 30% private car sounds pretty good for a city of its size.


@Zoltan Connell:
That is, because in Brazil you have a high ridership of people who do not have access to a car. So anyone is happy about any well organized public transport.

In other Brazilian cities of a similar size (like Recife) you find a similar or higher public transit ridership.

Eric Doherty


"What matters is cost per unit of capacity, i.e. "operating cost per passenger"

You are missing a huge segment of cost on some systems (thinking about the 'Skytrain' automated light metro lines in Vancouver Canada). Unless that is just a typo, and i.e. should have been e.g.

What matters is the overall cost per passenger, including all capital costs and interest. Focusing only on operating costs amounts to a serious bias in favor of capital intensive rail projects.

And maximum capacity matters too, once in a while, where street space or other surface rights of way are a limiting factor.


Thank you for this post, Jarrett.

You still exhibit a pro-bus bias, however, and now I can tell you why.

The only justification for buses being more "versatile" is that they can run on the existing leftover asphalt infrastructure.

The exact infrastructure which is obsolete and unsustainable, being heavily oil dependent.

In the future, we're going to have to replace most of it with non-oil-dependent structures -- and that probably means rails rather than roads, given that asphalt is literally made out of oil (and the main alternative, concrete, is at least as energy-intensive and carbon-intensive as rail). Perhaps we'll go back to brick roads, but then ride quality REALLY starts to tip the scales for rail.

So if we're really thinking in terms of rational long-term planning rather than emotive stuff, we want rail.

Yet you left this out. You are allowing buses the advantages from the existing installed system. Yet you also decry locking in bad geometry.

*But you haven't realized that building infrastructure for buses is locking in bad geometry*. Pittsburgh hampered itself for a generation by building the East Busway instead of a rail line, for a real-world example.


FYI, rail has had a nasty tendency to get gold-plated. We should note that this is not necessary.

Transit Planner Munich wrote:

Why just imitate the tram with a tram on tires and one instead of two tracks?

In the long run it is more expensive, cause tracks last much longer than busways, rubber tires are also not cheap."

The fact is that even a cheap branch line built to early 19th century standards gives service the equal or the better of a bus on a crumbling rural road. This is proven in many remote "only-reachable-by-rail" areas.

Technically, there's only one reason to use buses in low-density areas: because we already have the subsidized oil-based, car-centric asphalt infrastructure, paid for by "someone else".

Now we may simply have to accept that as a fact in many places. But it leads to an inexorable conclusion: buses are only useful as a supplement to the car. If you want a substitute for the car, you build rail.


"Imagine for a second, the very modern, sleek, low floor electric trams that roam europe. Now, instead of two tracks, make it one track and add tires (see Verona). It's now a guided bus! But unless you examine the tracks, you think it's a train. "

And it's more expensive to build and operate, and takes up more real estate!

Just because you CAN do something doesn't mean you SHOULD. I have a definite hostility to silliness like using non-standard track gauge (BART), nonstandard platform heights (oh, practically everyone), rubber tires for the sake of giving money to Michelin (Paris Metro), nonstandard power supply, nonstandard clearances, et cetera.... Guided busways sadly fall firmly in this "different from standard rail, but without enough benefits to justify it" category.


@Chris M, 40,000 on a single line in a "low-density New World city" actually is great - far exceeding that which any BRT system in any US city has accomplished.


Well, in most European (and most American cities either) you have the following situation:

There are usually just a few corridors with the dimension to have a seperate right of way.

In most other streets, if you like it or not, you cannot run a real BRT system.

So space is indeed the limiting factor.

On the other hand - to built a BRT system or a BRT system with tracks and overhead wire does not make a big difference in costs. But when you do that you can operate a light rail / tram system on it, with a larger capacity per vehicle.
To run a bus system with a headway of less than 3 minutes between the buses is tricky. Even if you have signal priority at traffic lights. And even with articulated buses and boarding on all doors. I don't know how they do it in Curitiba, but it is really a challenge.

But even light rail / trams have a limitation. They are good for distances of maybe 5 km. For longer distances a metro system will be better, or commuter rail. Cause you can reach average travel speeds that you will never have with a tram or bus system. But that is a different story and I just wait for a "why BRT is superior to commuter rail or metro trains" post on this blog.


Now here's something I'd like to know about dedicated-ROW BRT projects: when they talk about capital costs of BRT, what exactly is included in that? Obviously there's the bus roadway itself and the stations, and I'm sure no roads off the busway are included even if improvements are made for increased bus traffic. But what about the buses themselves? Or maintenance facilities for an expanded bus fleet? The cost cited for light rail lines generally includes vehicles and maintenance facilities, and often general streetscape improvements around the line too, but maybe BRT appears cheaper due to not including those costs.

Michael D

I think utility relocation is a big chunk of light rail capital costs. Is that included in BRT proposals? (Shouldn't it be?)


I think utility relocation is a big chunk of light rail capital costs. Is that included in BRT proposals? (Shouldn't it be?)

If the utility is relocated...

There are two reasons to do utility relocation:

* If the weight of the vehicles is damaging to the utility infrastructure itself.

* If there are concerns about future utility work taking the road or rail infrastructure out of service.

Both things tend to affect rail more than bus. Railcars are heavier than busses. The point-source axle-load at a given point is lighter (more axles on trains than the two found on most 40' transit busses); but the linear load (weight over length) is higher. Many roadbeds need rebuilding, particularly for LRT-class vehicles, and if you do that, relocating the utilities often makes sense.

Secondly, the need to maintain utilities located under tracks often is more disruptive to rail service (which can only go where there are tracks) than to busways (as busses can detour surface streets during utility maintenance). There are construction techniques, highlighted at lightrailnow.org, to merit utility maintenance under tracks from a trench dug alongside, permitting the line to remain in service; but these might not be applicable in all situations. (These are civil engineering questions well outside my knowledge).

Jarrett at HumanTransit.org

Scotty's right about utility relocation.  Less of it is needed for bus projects because buses can physically detour.  Often utility work closes only one lane at a time, so buses can get around the construction but rail would be interrupted.

Michael D

Thanks - that makes sense. Are there ways to ensure that the transit agency will be able to secure temporary bus lanes for a detour?


Thanks - that makes sense. Are there ways to ensure that the transit agency will be able to secure temporary bus lanes for a detour?

Do you mean are there ways to guarantee an exclusive bus lane during construction--i.e. closing an auto lane temporarily so busses still have exclusivity when the busway is closed?

Depends. Some road authorities will permit this; others may not--and insist that the busses run in mixed traffic. If it is for a short stretch, the service degradation is probably tolerable.

One wrinkle: Some BRT lines in exclusive ROWs operate busses which are not street legal. (LAMCTA Orange Line?) These busses may have greater difficulty with construction in their ROW, but even then, there are more options (a slow zone, flaggers, running in the opposite lanes) for tires on concrete than for rail, which only can go where there are tracks.

With rail, unless you have lots of redundancy in the system (which few metropolitan rail systems do), the usual solution is to bus around the construction zone. Happens all the time in Portland, and TriMet is quite good at doing it, too...

Jarrett at HumanTransit.org

Anonymous, re capital costs of BRT, fleet can be a genuinely hard thing the figure, as it depends on how you're going to use the cost estimate. A closed BRT system will have a dedicated fleet whose cost should be part of the capital cost of the project, but an open BRT like Brisbane's is designed so that any old low-floor bus can use it. In that case, the fleet need is a result of the operating plan rather than the facility, and the operating plan can of course evolve in response to actual ridership.

So if you're talking about the capital cost as "sunk investment in a project that is exposed to loss if the project fails" it's logical to exclude some of the fleet, because in fact fleet will grow in response to actual ridership (very fast in Brisbane's case!) and isn't a sunk cost at the beginning.

Steve Lax

A point of interest: While not quite a BRT, the almost 4 mile long XBL (Exclusive Bus Lane) from the New Jersey Turnpike to the Lincoln Tunnel into NYC accommodates a bus every 10 seconds in the peak hour. On some days peak hour ridership has been measured at over 20,000 inbound passengers.

The value of the XBL is that it collects buses from various express and local bus routes and bypasses the vehicular backup of other vehicles trying to access the Lincoln Tunnel into midtown Manhattan in the AM peak period. In a number of cases, ridership via bus from suburban park-rides at train stations is higher than ridership by train. (Of course, Manhattan is such a dense market it is one of the few places that can support extensive bus and rail service from the same suburban areas.)

This does, however, show the power of BRT-type operations on a suburban to urban core corridor where a number of bus routes can feed into an exclusive bus lane and provide good quality single seat rides into the urban core from many neighborhoods. Note that this also reduces the needs for suburban park-rides as there are more options for people to walk to a bus stop with this model.

Alon Levy

The XBL is not a good example of BRT capability. It has no stations, which means the headway is limited only by a safe stopping distance. It also feeds into a multi-block station right at the Manhattan end. For a new BRT system, it would be prohibitively expensive to build such a station.

The only other way to get bus headways so low is to have long stations, with multiple bays each, as in Bogota. This is less expensive, but still involves a large station footprint.

The BRT systems usually under discussion are not like the XBL or like Transmilenio, and have their headways determined by bus boarding time and signals, just like local buses and LRT.

As for the utility relocation issue: the best-planned New World LRT system, the C-Train, avoided this entirely. Calgary figured out in advance which corridors would be served and then built the roads and utilities on them to appropriate standards. It's impossible to do in an older, denser city, but in a rapidly-growing city, it's possible and advisable.

Matt Fisher

It would sound like David in Ottawa and I are saying the same thing re. BRT. Alas, it appears that people in Winnipeg were convinced that since BRT was good for Ottawa, it would be good for them, notwithstanding that Winnipeg is the city that is headquarters to bus manufacturer New Flyer Industries, a supplier of buses to transit authorities across North America (including OC Transpo).

Hamilton Transit History

I think that the biggest problem with BRT is not planning or technology related, but political.

When a bus route reaches the point that upgrading it to BRT makes sense, so much time is wasted with the debates, polls, studies, posturing, arguments, etc, that by the time you actually get around to building a BRT, demand on the route will have risen so high that a very short time will pass before it makes good sense to upgrade to an LRT. Because of the delays that naturally occur in democracy, You might as well skip over a BRT and go straight to the LRT

Steve Lax

Re: @Alon (8/28-11:33)

I did not mean to suggest that the XBL is a BRT, in part for the reason you cited (no stops). That is why I began with "While not quite a BRT". However, it is an excellent example of a busway that works in the extreme. Without it, the volume of people choosing bus to access midtown Manhattan would not and could not exist.

That there is a huge bus terminal on the Manhattan side is obviously important; but not all of the buses using the XBL go into the terminal.

Indeed, the XBL is so successful that many buses traveling from Staten Island and Central Jersey to lower Manhattan (Wall Street/World Trade Center Area) travel north to the XBL and then through the Lincoln Tunnel and south in Manhattan instead of traveling via the Holland Tunnel which is significantly shorter.

What the XBL does show is that people will choose to commute by bus when there are rail alternatives if the bus is, in their evaluations, the better option for the journey. And 20,000 people in an hour is a huge number voting for the bus.

Matt Fisher


"It's pointless, and I dont understand how some people can be so against buses (or rail) because of arbitrary criteria they decided upon. Like capacity."

So, I support rail based on "arbitrary criteria"? Well, I say BRT has its place, and will continue to in Ottawa even if the Transitway is converted to light rail. Were all the great rail systems of all time built on "arbitrary criteria"? Was Calgary's C-Train done the same way? Furthermore, I find capacity to be legitimate, not "emotional".

"Buses arent capacity limited, Ive mentioned it before, but you can attach various trailers to one engine, just like a train. There is no technical limitation. If an engine pulling THREE giant trailers is allowed to move at 60mph on an interstate, I think we can manage doing it on an exclusive ROW."

Oh sure, you CAN run a bus like a train. Like that'll ever happen. NOT! You CAN'T expect to couple two buses together, with or without an articulation. Not even Curitiba or Bogota, Bill "Better Rapid Transit" Vincent's favourite examples, does that! (Thanks to Michael Setty for suggesting the term to describe a major BRT proponent who claims BRT is better at fighting global warming than electric rail transit.)

"Another example: streetcar."

"Imagine for a second, the very modern, sleek, low floor electric trams that roam europe. Now, instead of two tracks, make it one track and add tires (see Verona). It's now a guided bus! But unless you examine the tracks, you think it's a train."

That's the so called "tram on tires" which is not a real tram. As the accidents in Nancy and Padua have shown us, it appears to be more of a lemon, but I won't totally rule it out. It does not literally constitute a train.

"Peoples preference to trains over buses are simply because most people were brought up in a system where rail had much more money put into it than buses. 3 minute headways on rail, 30 minute headways on bus. Thats where the bias was born!"

Yeah, well, I like rail over bus DESPITE being brought up in Newfoundland and Labrador, where I was born in 1987. The year after, trains in the province stopped running (thank you very much, Brian Mulroney). My move to Ottawa in 1998, where I continue to live today, only furthers this, and we only had the Transitway at the time (and I was then blindly impressed by it).

@Transit Planner Munich:
I think Line 53 in Munich would make an excellent tram line sometime.

Matt Fisher

I almost forgot, you can't "attach various trailers to one engine, just like a train". There IS a serious technical limitation in practice. It's just not possible. I don't think OC Transpo would even want to do this on the Transitway, as much as they would want to use a "tram on tires" like in Nancy or Padua. That's likewise why almost all of the Montreal Metro is underground to begin with, with snow concerns here in Canada (except some train yards, but I never saw them on my travels while riding it).


Actually, all of the Montreal Metro is either underground or under a roof, even the train yards. Yes, that does sound like a pretty expensive price to pay just to be more like Paris.

As for the XBL, there are a number of things going on. First of all, the BRT system is not just the XBL, it's the XBL plus the four story bus terminal at the end of it. Station capacity can limit throughput pretty significantly, and it's only because of that terminal that they can put so many buses through the tunnel. Also, many of the buses serve areas of Jersey that do not have a one seat rail ride to Manhattan: you have to go via Hoboken and a transfer to the PATH, or via Secaucus. And the rail tunnel for the commuter trains is already literally at capacity in terms of trains and getting very near its total capacity. They're running 24 trains per hour through there with a good fraction being 12-car trains of double deckers with every seat filled and people riding in the vestibules. It's no wonder that some ridership spills over onto the bus lines.


Rail is more energy efficient than tires. And longer lasting than the pavement for a bus. So the extra investment for some tracks pays off. Also the life span for a tram / light rail vehicle is 30 to 40 years. For a bus 8 to 12 years. So a system that masks like a tram to attract ridership but runs on tires usually does not make sense. It costs more in the long run. So it would be a pure emotional reason to go for BRT if you are obsessed with rubber (what is fine if it makes you happy but should be lived out in your own house) instead of logically chosing a tram system.

@Hamilton Transit History:
When is the point reached to invest in BRT?
And when in light rail?

Dieter Ludwig, the former head of Karlsruhe transit, liked to say: Where you need to run an articulated bus every 10 minutes, you might as well build a tram. So why build BRT instead, just to save a relatively small part of investment... Also you will not invest in BRT either with buses running just every 20 minutes.

Probably there are two good reasons to invest in BRT:
- you have several bus lines with low ridership and low frequency coming together on a central section. And in that central section you have fortunately the space to build BRT. Then it makes sense to build BRT for this central corridor.
- you have a high demand on one or more lines but they go in some sections through narrow or steep streets and you think you will meet practical or political problems to build there rail. Then fine, BRT is better than no improvement.


Greetings from Berlin, where I am spending a week and have been riding the transit. I wrote the original comment, which Jarrett invited.

Berlin has a fantastic, integrated transit system, which includes quality bus service. I don't know whether you'd call it BRT, or not, but it includes many features that are often cited as BRT. It is a proof-of-payment system, which allows all-door-boarding, and most riders have passes, including day passes. Many stops have electronic signs showing when the next buses will arrive. There are many exclusive bus lanes. There are automated bus stop announcements. But stops are generally about every 400 meters, I believe. And there is a frequent service network which doesn't differentiate trams and buses - all designated by M.

Having experienced it for the past 4 days, it also demonstrates some of the drawbacks:
- tour buses seem to be allowed to park in the exclusive busways. in multiple locations they were parked with blinkers flashing, and the buses had to maneuver into GP lanes to get around them
- at several stops alighting end entering passengers took 2 traffic light signal cycles on a bus I was riding. the following bus (5 minute headway) passed us. in the opposite direction, 4 buses of the same route number were queued at the same traffic light - the first two were overcrowded and the last two were quite empty. Perfect example of bus bunching and headways not maintained. This route is operated with double-deckers, which exacerbates the stop dwell times and bus bunching.
- there were two route diversions due to construction, including missed stops. The flexibility of the bus to go around construction is a double-edged sword. Because it is easy/cheap to re-route, that gets done, instead of finding a way to keep the normal route open which is what be done for tram/LRT - maybe auto traffic would be re-routed but tram tracks can be maintained during construction.

I wasn't aware of this article until today, but my experience maintains my basic point. For lower capacity, bus is the only feasible mode. But once capacity increases, it is difficult to provide the same service quality, capacity and reliability as can be provided by rail - perhaps in part for what Jarrett calls cultural reasons, which make it too easy to re-route, delay, block, and bunch the buses.

Damien Haas

Having watched and talked to Jarret in Canberra, I don't think he is bus biased. i do think that as a consultant he will work for bus biased transport agencies. In that case, what do you do ?

The report he prepared for Canberra was heavily bus-focussed but that is what the local government want.

I, as a public transport lobbyist take his findings and recommendations and look at what they can tell me about future transport needs and technologies.

I truly want Canberra's bus ridership to reach 16% of all trips taken - I cant see a better way to convince people that they will need light rail when they are trying to jam 200 people onto a bus that only comes once an hour...

Integrated light rail and bus will work, so its not an either/or argument.

The routes he identified can easily be used in future for light rail. Just because the current government want BRT, doesn't mean a future government will. Consultants consult, governments respond to electoral pressure.


Berlin has a really nice transit system.
But also some partly home made problems:
- it has a still only fragmental priority at traffic lights for buses, that
- gets even more sabotaged by the requirement to board on the front door and show your ticket, despite proof of payment system. That makes the amount of time a bus has to stop uncalculable and also a proper prioritization difficult
- most bus lanes are on the side of the street instead of the center. That makes it easier to place bus stops but also makes the bus lanes basically useless. The right lane is bound to be blocked by illegally stopping cars, delivery services and trucks etc. who all have a good excuse but block the bus. And: on several bus lanes taxis, bicycles and even freight trucks are allowed to use them. That again causes chaos and bad signal priority and delays.

Steve Lax

@anonymouse (8/29 - 18:05)

As I mentioned in my response to Alon Levy, not all of the buses use the bus terminal; though it is helpful. Of course, the rail line has a rail terminal (Penn Station); so maybe the issue is not bus or rail but the availability of a terminal that provides an enclosed waiting area with auxiliary facilities (rest rooms, food, etc.) and sheltered connections to local subways that really encourage people to use public transit.

Would some users of the buses into NYC switch to rail if rail offered a one seat ride? Absolutely. New Jersey Transit saw that when they opened up their Midtown Direct service. However, where bus remains competitive time wise for the total trip (home to transit to job taking frequency into consideration (average/median wait including any transfer penalties) bus retains a large share of the market.

Of course, the job base in Manhattan is so dense that both modes can be supported.

The total travel time is not only a factor in the New York market. In the Austin, Texas area commuters who commuted into Austin on two express bus routes from the Leander area are upset that Capital Metro wants to discontinue them now that the commuter rail is up and running. The reason they are upset: the train increases the time of their commute.

Alon Levy

Steve, you're right that the trains have a terminal, too. But the train terminal is underused. Out of 21 tracks, it needs about 6 for good operations - 2 for through-service, and 4 for terminating extra LIRR service from the east. The other 15 tracks are just for work rules, agency turf battles, and steam-era railroad traditions.


@ Anton Levy

I have to agree with you. Lets look at what the XBL does in peak hour to the city and surrounding roads:


I think I'd rather have a train carrying so many people. There are no prizes or awards for carrying people in buses every 10 seconds.

Alon Levy


Sigh. It's not as if the name isn't routinely spelled correctly at the bottom of every comment that I write.




Alon, you are assuming that every single train, Amtrak included, can get in and out of Penn Station in under 2 minutes, and that it's safe to have a full reverse-commute load already on the platform when a full peak direction train arrives and lets out its passengers, or that the detraining process will happen at all quickly when this is the case. In the real world you do need more than two platforms. And work rules or no, you'll find that NJT's electrics will have some difficulty working without overhead wire, and likewise for LIRR's electrics and third rail, though that doesn't preclude through service to the New Haven Line (which does occasionally run, btw).

Alon Levy

For terminating trains, 15 tph is routine in the real world - see for example multiple lines in Tokyo. The Chuo Rapid Line terminates 26 tph on two tracks at Tokyo, but it's unique. Neither of these performance levels is unusual in New York: the WTC PATH loop is almost as fast, and the 42nd Street Shuttle turns trains around faster. They're just unusual on commuter rail.

The NJT electrics don't ever need to go anywhere except the New Haven Line. But sending them to Connecticut would free dual-mode EMUs. It would take a trivial amount of reconfiguration to change those EMUs from Metro-North to LIRR third rail.

The basic issue is that the trains coming from Jersey to New York can continue onward without clogging Manhattan and the buses by and large can't.


The limiting factor in Penn Station is not necessarily how fast you can move the trains through but how quickly you can move the passengers off the trains and off the platforms. I bet the Chuo Line trains have more doors and the platforms have much better vertical circulation to get passengers out of there as quickly as possible. Dwell time is generally the limiting factor in any transit system.

As for dual-modes, I assume you mean the M-2/4/6/8 series? The 2/4/6's can't run on 25kV, while the M-8s can't run on 25Hz, as that would require a noticeably heavier transformer. So there won't be any Metro North MUs going west of NY Penn for the foreseeable future. On the other hand, if the third rail were extended through Secaucus to Kearny Junction, the M-8s could run through to the M&E lines. Except they wouldn't be able to stop at most stations, since those only have low platforms. But really, what for? Are there that many people commuting through NY? Is it worth the risk of propagating delays across the network? And the agency turf battles you speak about are not all that simple to resolve, given that there are two states involved which don't exactly have a history of friendly cooperation.

Alon Levy

Yes, the Chuo Line has unusually good vertical circulation. The other lines in Tokyo don't; they have the same number of staircases/escalators leading out of platform level as Penn's NJT platforms (2), and fewer than Penn's LIRR platforms (5). The trains have more doors, but they're so much more crowded that there are more people per door.

And yes, there are that many people commuting through NY. Fewer than commuting to NY, but there's a market of a couple tens of thousands to be served at zero extra infrastructure.


"zero extra infrastructure"... aside from replacing all the trains and/or rebuilding quite a few platforms. Unless you're just talking about a limited through service of running a few NJT trains to/from Stamford, in which case I agree that it's quite doable, especially in the off peak, because in the PM peak, an NJT train coming from Stamford would either need a long stand time at Penn to buffer delays (especially ones at Shell, though that's gotten less bad) or else risk losing its slot through the tunnel or delaying everyone else. You won't get away with just two through platforms at Penn, or even four. And if you look at the busiest lines in Tokyo, I bet most of them have their own mostly-dedicated tracks where other lines don't interfere with them too much.

Steve Lax

@alon and anonymouse re NY Penn Station -

By coincidence, I went into NYC Penn Station from New Jersey this past Saturday at midday. The train consist was multi-level cars. Because of my interest in this sort of thing, I timed the emptying of the train at almost exactly four minutes with no interfering activity on the across platform track. And the train was not completely full; though there were the "which way should I go" occasional users slowing progress.)

Given the design of the NJ Transit multi-level cars, most people from both the lower and upper levels do not use the end doors; but rather the first doors they see at the mid-level stairs. If the stream of passengers closer to that door (on my train emptying it was the lower level) were the only stream to use the door and the other stream (upper level in this case), were to use the end doors, emptying would be faster. The other observed reason for delays was the large number of people with luggage, most likely coming from the Newark Airport Station.

Alon Levy

Anonymouse, in Tokyo the standard procedure is that a two-track line used to full capacity will expand to four tracks at the terminus, or occasionally six. A two-track through-line will have two platform tracks, with four only in very few cases (Yamanote at Ikebukuro, Saikyo/Shonan-Shinjuku at Shinjuku), whose crowding makes Penn Station look like a quaint branch line.

Trains aren't an infrastructure cost. Rebuilding them to a certain standard (or, better yet, making sure the new trains are capable of through-running) isn't expensive. Even upgrading platforms is quite cheap; in Germany they build regional rail single-track stations with mid-level platforms for about $200,000. Multiply that by ten and the total would still be a rounding error by the standards of the cost overruns on ESA and ARC.

Steve, four minutes is terrible for through-trains, and could be remedied only with higher-throughput rolling stock in the future. But for terminating trains, it's normal. Normally, trains are timed to turn in 5-10 minutes; even 10 minutes allows turning about 10 tph on two tracks, which reduces Penn's requirement from 21 tracks to about 12 with zero through-running. With through-running, with 4-minute dwells (=12 tph on through-track), make that 8 tracks.

The obvious answer to the question "Why don't they do it, then?" is that Penn has 21 tracks; there's little gain in fast turnarounds. The real crime begins when people internalize a certain turnaround rate, and then can't cope with higher traffic, building deep-level caverns.

Anyway, this is way, way off-topic. Going back to this thread's original issue, what I wanted to say is that trains running through central cities can achieve high capacity with relatively little station footprint, and diesel buses can't. Obviously, if there's yard space for either trains or buses, it should be used. Port Authority allows a certain bus throughput, and it's good that the XBL takes advantage of it (and there are certainly efficiency gains to be made there, too: for one, buses could have three doors). But for a city that wants to build a greenfield rapid transit connection to its downtown, rail's ability to do so with less station footprint is a major advantage.


For a hypothetical highly-used two track rail line coming into a hypothetical center city, you still need four or six platforms, it's just that you can have them all at one station or have three different stations, and in fact that is exactly how subways work. And likewise you can do the same with buses, as is in fact done with the Brooklyn, Queens, Staten Island, and Bronx express buses, without having a huge central terminal.

As for Penn, I agree that they could make better use of their platform capacity and that they have some capacity to spare (even more so after ESA), which makes the ARC deep-cavern project so ridiculous.

Alon Levy

The outer borough express buses don't begin to approach the capacity of the XBL.

And many of the four- and six-track terminals used in Tokyo are in fact the only CBD stations on their respective lines, or almost the only CBD stations; some are not even in the CBD, but offer the only/best CBD connections. It's not like with subways, where the center of the CBD might host 3 stations on one line. (However, many of the near-CBD and edge-of-CBD stations - Shimbashi on Tokaido Main, Kanda and Ochanomizu on Chuo, Kita-Senju on Joban and Isesaki - have much better subway connections than Newark and Jamaica.)


If you have one CBD station, then you need 4 or 6 tracks. If you have two or three CBD stations, you can get away with only two tracks throughout. But you still have 4 or 6 platforms in total either way.

Alon Levy

For through-stations? Not really... plenty of CBDs get by with two tracks - not just in Tokyo, but also in Paris. The RER A only makes about one stop in La Defense, and one in the proper CBD. And it runs 30 tph, on trains with terrible passenger circulations.

The comments to this entry are closed.

the firm

Jarrett is now in ...

Related Posts Plugin for WordPress, Blogger...
Related Posts Plugin for WordPress, Blogger...