Is Elevated Acceptable?


The Transport Politic has an excellent post on the debate over the plan to build Honolulu’s proposed light rail system elevated through downtown, as opposed to at the surface as a group of architects wants.

Everyone is prone to reduce the complexity of urbanism to a problem solvable by their own profession, and risks being dismissive of the expertise of other professions’ points of view.  (See here, for example.)  When a group of architects proposes that a major new transit investment should be made slower and more expensive to operate in order to foster a better streetscape, as is happening in Honolulu, one hopes that they have thought through the urbanist consequences of all the people who’ll be in cars instead of on transit because the transit is too slow, infrequent, and unreliable.  Let me clarify each of those words:

  • Slow, because a system running at street level has to interact with traffic or pedestrians.
  • Infrequent, because a system running at street level will require human drivers, while a fully grade-separated system can be driverless.  Human drivers imply a cost structure dominated by the cost of labor, which in turn means that high frequencies are just not affordable in the late evening, for example, as demand goes down.  Vancouver’s driverless SkyTrain runs extremely high frequencies to past midnight, because without drivers the incremental cost of more frequency is very low.  Note that poor evening transit service is bad for nightlife, which is something all urbanists care about.
  • Unreliable, because street operation, even in an excusive right-of-way, exposes the service to many random and unpredictable causes of delay.

I support the notion that locally-oriented transit systems (such as streetcars and local bus services) should be integrated into the streetscape, because those systems are by definition designed to serve short local trips where station access time is more important than travel time.  Honolulu’s proposed LRT is not such a line.  It is intended for very long trips, and for trips that go across downtown (Airport to Waikiki for example).  If you could find the money, you could follow Vancouver’s example and run the line underground through downtown, but running it on the surface would defeat too much of the purpose of the line.

Yes, Portland’s famous light rail system, MAX, is on-street through downtown.  Does it stimulate the downtown?  You bet it does.  Is it a time-competitive way of making a long trip from one side of the city to the other, say Beaverton to Gateway?  Well, that’s not its strong suit, as the long downtown Portland segment is very slow.  But the important difference with MAX is that it was never designed with an eye toward driverless operation.  It is mostly at-grade all through the system, and will always have drivers, while the Honolulu system is designed to be fully grade-separated, and hence driverless, everywhere.  It’s hard to overstate what a difference this will make to the frequencies the service will be able to offer.

UPDATE:  Welcome, readers from, where I find this post has been cited as support for an elevated extension of Vancouver’s SkyTrain to the University of British Columbia (UBC).  That’s more than I’m prepared to say.

I will say that I wish they’d call it the Broadway corridor (to emphasize the street it serves) rather than the UBC corridor (which refers only to the endpoint).  From a ridership standpoint, Broadway is by far the densest, most promising transit corridor in Vancouver that still lacks rapid transit.   Like Wilshire Blvd. in Los Angeles, it’s a place where buses in mixed traffic are doing everything such buses can do, and can’t run any more frequently because they would just pile up at the stops.  But I do not mean to imply a position on technology or profile (elevated vs surface vs underground), and am not sufficiently up to date to wade further into that debate at the moment.

27 Responses to Is Elevated Acceptable?

  1. Nathanael May 19, 2009 at 11:45 pm #

    Why has it never been considered that you could run the train line at ground level, and put overpasses or underpasses in for each of the roads it crosses (well, I’m sure some of the roads could be dead-ended)?
    Is the line planned to run directly *along* an important road, which couldn’t be shut or pedestrianized — is that the catch?

  2. Jarrett at May 19, 2009 at 11:54 pm #

    To retrofit rail on the surface in a built city, the expense of all those underpasses is probably prohibitive. The result is also not very poor urban form, because even with underpasses the line would become a major psychological barrier for pedestrians (and to some extent a practical one) creating an unavoidable sense of “wrong side of the tracks” about one side or the other. Current urban design is also very concerned with maintaining continuous activity at the ground plane, and avoiding underpasses or overpasses for pedestrians as much as possible. This is why skybridges such as those in Minneapolis are no longer considered best practice.

  3. Rail for the Valley June 18, 2009 at 8:41 am #

    Try this for site for some additional info.

  4. CroMagnon July 21, 2009 at 3:24 pm #

    I’ve looked into this project a bit.
    For all intents and purposes this project is Heavy Rail, not Light Rail. The only thing “light” about it would be the platform length, about 180′ or so (expandable to 300′). Even still the capacity should be quite high if headways are 3 minutes.
    I think this project is great, because it diminishes the “nobody builds HRT anymore” reasoning. Too much is made of LRT in transit and the modern urban planning school of thought as a one size fits all, when BRT or HRT would be more appropriate technicall, economically, as well as agreeable to the neighborhoods through which it would travel.

  5. BobA August 11, 2009 at 12:54 pm #

    I use the Vancouver Skytrain frequently and like the elevated aspect of it — great views out while travelling. Some effort needs to be made to keep it visually acceptable – my recent trip to Chicago showed me what awful impact an old elevated system can be on the eyes and ears.
    The Portland system is nice but we need RAPID transit, not a bus on rails.
    The best mix to me is tunnel downtown and in significant built-up commercial areas and elevated the rest.

  6. Zweisystem August 20, 2009 at 1:03 pm #

    Just a note, Portland’s MAX has a faster speed than SkyTrain in Vancouver, but has about twice as many stops as SkyTrain so MAX has a lower commercial speed.
    Fact is, after being on the market for 30 years and four name changes from ICTS > ALRT > ALM > ART, only seven of these mini-metros have been sold. If Hawaiian politicos want SkyTrain, they also want to bankrupt the taxpayer. TransLink the transit operating authority is in a deep financial crisis and will be $450 million in the hole next year due to its mini-metro building program.
    Build with SkyTrain – Caveat Emptor – Let the buyer beware.
    For some interesting reading

  7. M1EK October 15, 2009 at 8:43 am #

    Honolulu’s line certainly is heavy-ish, and they often just call it “elevated rail”. One can justify it there, too, for some additional reasons:
    1. Views (give transit customers some of the great views users of the elevated freeways currently have).
    2. Cost (actually building this line elevated may be cheaper when you consider the high cost of displaced auto capacity at ground level and the limited options for supporting it).
    The big problem is it doesn’t GO to Waikiki; won’t for a long time. Perils of the political process – the tourists who would use this thing like gangbusters (Waikiki tourists arrive now without cars in droves) didn’t get to vote on the system; and local residents demanded service for less useful places first. Those tourists could have subsidized the system quite nicely but now won’t get the chance.

  8. zweisystem October 21, 2009 at 10:43 am #

    Actually the cost of elevated transit can be as much as 10 times more than at-grade LRT/streetcar. Cost for displaced auto traffic? That’s a new one; actually at-grade/on-street LRT creates passive traffic calming, which increases ridership on the transit system.
    Let’s be serious; if one invests billions of dollars on a elevated transit system, then one want to get as many people on to it as one can, to justify construction. Elevated transit systems are very poor in attracting the motorist. Vancouver’s SkyTrain, despite over 20 years in operation has failed to show a modal shift from car to transit.
    At-grade/on-street LRT has shown the desirable modal shift because it creates passive traffic calming, but with the bonus of increasing a road lane’s capacity from about 1,600 pphpd with cars to over 20,000 pphpd with trams.
    By greening or lawning the streetcar rights-of-ways creates linear parks, something that the ugly pillars for the elevated transit systems do not.
    But a final argument is that LRT, properly built can be operated without subsidy, which is a bonus for taxpayers!

  9. CroMagnon December 9, 2009 at 1:13 pm #

    ^Man, zweisystem, everything you claimed is either subjective or just factually wrong. I’m concerned and amazed at how many LRT advocates ignore physical and economic factors of time, space, energy, etc. External impacts of reduced capacity are usually ignored and dismissed as though they are negligible, i.e. “out to sea”.

  10. Zweisystem January 2, 2010 at 4:32 pm #

    The previous post is fundamentally flawed.
    If one wants to build a metro (ie grade separated transit), then one should have the ridership to sustain it, or about 400,000 passengers a day. Now LRT can be elevated like Manila, if ridership demands.
    If projected ridership is less, then at/grade/on-street is more cost effective. Modern LRT, by its very nature on operating on a reserved rights-of-way (a route for the exclusive use of a streetcar/tram) commercial speed and capacity is based on the number of stations or stops per route/km, size of station, and headways.
    Even the simplest of tram system can, in theory, accommodate 20,000 persons per hour per direction.
    Now transit capacity is adjusted for usage and very few transit systems run at capacity throughout the day, but enormous capacity can be accepted by LRT over short periods.
    Example Karlsruhe Germany: trams in pairs operate headways of 45 seconds during peak hours through the main street, gives a peak hour capacity of about 36,000 pphpd! (45 second headways = 80 trips per hour X 450 the capacity of two trams in MU) This is tremendous ‘lift’ from a simple tram system.
    This is why LRT has proven very good in moving large volumes of passengers affordable.
    The facts are on my side I’m afraid.

  11. Alon Levy January 2, 2010 at 5:49 pm #

    Where does the 400,000/day figure comes from? I’m asking because lower-ridership metros have been built for much lower per-passenger costs than US light rail systems. For example, Paris Metro Line 14, with 60,000,000 passengers per year, was built for about $7,000 per passenger in 2009 money, less than the cost of every US light rail system built to date.

  12. Zweisystem January 2, 2010 at 6:19 pm #

    The 400,000 thousand passengers a day comes from UBC Prof. Condon and as well used by other transit professionals here and overseas.
    I am very suspect of the figure of $7,000 per passenger and I am sure it is not the official cost. Of course metro’s in Europe are heavily subsidized; in France pay-roll taxes are used extensively to fund public transport.
    I would suspect that two different accounting methods were used to confuse the issue. And please remember, Paris is carrying on an extensive tram/LRT construction program, with over 100 km. to be built by 2015. Added to the mix, maintenance on subways is extremely expensive!
    As for subway construction, the costs can be very high.
    The fact be known, at-grade transit systems tend to attract more passengers than grade separated transit systems and is one of the reasons that modern LRT is the most popular mode of public transit in the world.

  13. Jarrett January 2, 2010 at 6:54 pm #

    Zwei, I’m sorry, but an abstract rule about a particular daily ridership that justifies metro is absolute nonsense. It depends entirely on how much the grade separation costs and over how much distance. The ridership outcomes, in turn, depend on the size of the total travel market and the travel time of the service compared to the alternatives. Surface systems have less access times but grade-separated systems can have a better in-vehicle travel time. Grade separated systems thus end up being faster except for very short trips, i.e. the local bus or streetcar market.

  14. Alon Levy January 2, 2010 at 8:34 pm #

    I am very suspect of the figure of $7,000 per passenger and I am sure it is not the official cost.

    The line cost €1.133 billion and gets 60,000,000 yearly boardings, which would translate to about 200,000 weekday boardings. That’s about $7,000-8,000 per daily boarding.
    Yes, Paris builds tramways. It also builds busways. How does that support any of your anti-subway claims?

  15. Zweisystem January 3, 2010 at 7:45 am #

    Jarrett, estimated daily ridership is a determining factor in choosing transit mode. What you speak of is 1980’s gibberish. It has been found in Europe that ‘at-grade’ transit systems are a lot better than grade-separated transit systems in attracting new customers and with LRT operating on a dedicated rights-of-way, can obtain commercial speeds approaching that of a metro.
    This accounts for the Renaissance of LRT in Europe in the ’90’s to date.
    Just a note, recent studies have shown that the over all ambiance of transit system, ease of use and no-transfers, trumps speed for attracting new customers to transit.
    The 400,000 customers a day is indeed a good marker for ridership for building a metro; any less for metro means higher operating subsidies from the taxpayer.
    Ron Levy, you did a naughty thing, you for got to add the €1 billion phase 1 cost of the METEOR into your calculations. Thus the cost of the 14.5 km. METEOR is €2.33 Euros or USD $3.34 billion. The cost is an estimate only and in 2003 Euros & dollars.
    Now 14.5 km is roughly 9 miles, so the cost of the METEOR is about USD $371 million/mile to build in 2003 dollars.
    Now if you financial calculations are questionable, then your claims that “Paris Metro Line 14, with 60,000,000 passengers per year, was built for about $7,000 per passenger in 2009 money, less than the cost of every US light rail system built to date”, is highly suspect. You have also forgotten to add in annual operating costs which are far higher for automatic metro systems than light rail, which also erodes your claim.
    I find it interesting that in France there has been a collapse of sales of the VAL metro system in favour of LRT, in spite of offers of heavy subsidies by the French government for the French built VAL system!
    I am not anti-subway, I just make sure that new subway systems do not bankrupt the taxpayer.

  16. Zweisystem January 3, 2010 at 8:38 am #

    The trouble with dealing with the subway lobby is the the many conflicting websites, with conflicting numbers. And confusion reigns – which is what the metro lobby wants.
    The previous post was based on info that I have now found to be incorrect.
    differs from
    Dear me who/what to believe!
    The METEOR Line original cost was well over $1 billion Euros, not including vehicles and signaling. A .49 km. (about one third of a mile) extension was €133 million or about USD $190 million or about USD $570 million a mile to build.
    “487m long, is costing €133 million and is due to open in December 2003.”
    The METEOR Line is about 9 km. long (5.6 miles) and at an estimated cost of over $3.5 billion!
    The METEOR travels through the heart of Paris and connects many important transit routes, thus generating it’s high ridership numbers.
    The massive costs of the METEOR Line, precludes any comparisons with American LRT systems.
    Apologies for earlier mistakes, but the more I study the issues about the METEOR, the more outrageous the comments comparing the METEOR with LRT become.
    I still stand by the comment, “Now if you financial calculations are questionable, then your claims that Paris Metro Line 14, with 60,000,000 passengers per year, was built for about $7,000 per passenger in 2009 money (which it is), less than the cost of every US light rail system built to date, is highly suspect.”

  17. Alon Levy January 3, 2010 at 9:21 am #

    Who the fuck is Ron Levy?
    And you’re double-counting the cost of Line 14. The link gives the cost as “€1 billion, Phase 2 €133 million.”

  18. PRT Strategies January 3, 2010 at 10:35 am #

    A better idea for elevated transit in Honolulu is at

  19. zweisystem January 3, 2010 at 1:16 pm #

    What ever, I corrected my initial error, but you fail to indicated that the phase 2 was €133 million for 487 metres (1/3 of a mile) of line!
    The €1 billion was an estimate only as per article and certainly did not include vehicles or signaling, nor does it include debt servicing.
    Thus your observation that “Paris Metro Line 14, with 60,000,000 passengers per year, was built for about $7,000 per passenger in 2009 money, less than the cost of every US light rail system built to date.”; is also wrong.
    You also fail to take into account that most of the METEOR’s ridership comes from other metro/railway, which again makes a farce of your cost estimates.
    Vancouver’s SkyTrain metro system which total cost to date is about $5 billion and carries over 70 million passengers a year is subsidized by over $230 million annually.
    You can build a metro in Honolulu, but if it doesn’t carry large volumes of customers (over 400,000 a day) be prepared for higher taxes and subsidies.
    I also failed to mention that French public transit is highly subsidized, through additional taxes and surcharges, again not factored in by your accounting. There is good reason why subways are only built in high traffic demand locations in large conurbations. Subways are very, very expensive to build and operate.

  20. Alon Levy January 3, 2010 at 3:52 pm #

    I don’t really care about Vancouver. It’s no more an example of elevated failure than San Jose is of light rail failure. Sometimes, the planners don’t know what they’re doing.
    The cost per passenger numbers I used are in line with those used to compare light rail systems. Nobody tries to subtract the passenger volumes of buses diverted to LRT from the passenger numbers, except maybe Randall O’Toole when he pretends to like bus transit. As for operating subsidies, give me a break; it’s not as if any American LRT system is operating in the black, either.
    Rolling stock isn’t part of the deal, either… but for what it’s worth, there are 120 cars; I can’t find the cost estimates, but elsewhere in the world they’d be $1-2 million per car. It’s not a budget buster; the numbers still end up a bit lower than for the lowest cost-per-rider LRT system in the US.

  21. M1EK January 4, 2010 at 7:50 am #

    @zweisystem, the cost of lost auto capacity is a big deal with the locals – and part of the political process here. Honolulu’s traffic is a lot more like Manhattan’s than Portland’s; and any additional non-trivial capacity lost would have doomed the rail plan to a loss at the ballot box.
    Additionally, again, Honolulu is one of the few places in this country with such high density and no existing rail transit – justifying a fairly high-cost investment based on anticipated high ridership compared to LRT cities.

  22. Adrian January 6, 2010 at 12:06 am #

    I’m not sure where does this $4 billion dollar figure come from. Unless we are planning to build 200 metre platforms and Dubai-elaborate stations, there is no possible way that Broadway Transit Corridor will cost $5 billion.
    Professor Condon’s plan included generalized figures that will not work in the Broadway Corridor. Much of the figures are based on other LRT/Tram systems within North America, where either the figures may exclude the cost of vehicles, common in many American Transit Agencies, or have its own Right of Way (ROW).
    Unfortunately, the Broadway Corridor does not have a parallel rail nearby or have its own ROW. This, coupled with the fact that Broadway is critically limited in space, are factors that must be taken into consideration. In the 1999 Technical Study conducted by the City of Vancouver and independent from TransLink and/or the Provincial Government, the LRT will reduce traffic lanes down to two, or even one lane in each direction depending on the area. And while there are many other parallel routes in Vancouver around Broadway that can take in more traffic, the traffic situation is still bound to become more chaotic. This is also a problem for pedestrians and cyclists. Within morning peek periods, cyclists can use the HOV lanes. When the Broadway LRT is built, it will be a safety problem for cyclists who wish to use the Broadway corridor. There will be no room to expand sidewalks in Broadway, a problem especially around Cambie and the South Granville area. These are more are the many critical problems that myself and the other people behind the UBC SkyTrain group has pointed out countless times, but have not yet been addressed.
    As for the high cost of Line 14 or METEOR Line, many of the stations are buried quite deep underground along with an immersed tunnel. The ALRT option, firstly, has the flexibility of not being fully tunneled, especially between the portion of UBC and Sasamat, and will not be deep underground. Based on the 1999 Technical Study as cited earlier, possible methods of construction will allow platforms of SkyTrain (ALRT) stations to be connected to the station street-level entrance with only a short incline ramp, similar to the one at Broadway-City Hall Stn for the Canada Line. This allows for improved accessibility, quite similar to a LRT platform. I also like to point out that if stations are built with multiple entrances or one is at the south side of Broadway, there is no need to cross the street to reach the platforms if it was built ALRT unlike with LRT.
    Another problem with the LRT plan is that there is no room for an Operations and Maintenance Centre (OMC), another fact that the I have pointed out earlier. Where are you going to build in? In Pacific Spirit Park? Within the residences of the Westside? There is no space. With the ALRT plan, the OMC does not have to be located within the Broadway Corridor and can be combined with an expanded Edmonds Rail Yards.
    From what I know, the Expo and Millennium Lines have already broken even. The Millennium Line, unfortunately, took quite a while for that to happen. But the reason is because the original Millennium Line plan did not come into fruition. Instead, we only ended up with Phase 1. The original portion included plans of extending the network to Coquitlam (i.e. Evergreen Line) and to Cambie/Arbutus/South Granville (i.e. “UBC” Line). With the Canada Line, ridership is growing and has is approaching TransLink’s projected ridership, ones that you claimed were “pixie dust.” Now, you state that the Canada Line only exists to attract customers to casinos? Ridership is ridership. And the Canada Line has done more than just attract customers to casinos. Richmond Centre and Oakridge, the anchor shopping destinations, have reported a dramatic rise the number of shoppers, ones who would otherwise use cars to travel to their malls. The airport portion of the Canada Line has also attracted many people. Even at 5:00 in the morning, there are quite a few passengers with luggage bound for the airport. Even trains bound for Waterfront coming from the airport are just as full as the trains coming out from Richmond. Initial ridership on buses peeked at a maximum of 40 000 passengers (adding the ridership of the routes directly serve) and has now rose to an average of 92 852 passegers. That’s more than double, showing the success of light metros in Vancouver. This success will further be shown in the Millennium Line Extension, where there are many developed, dense, and attractive areas. If the Millennium Line extension opened today, I wouldn’t be surprised if the ridership reached 200 000 passengers, a little more than double that of initial ridership on all the added bus routes serving today. And since you brought up the issue about subsidizing for transit, I’ll be surprised if we aren’t subsidizing for your TramTrain plan of linking Vancouver with Chilliwack.
    In your plan, you also brought up the issue transferring, and how transfers decrease ridership. Yet with the Millennium Line extension, there will be an elimination of a transfer that would otherwise exist with a LRT on Broadway. Quite hypocritical here.
    As M1Ek stated, capacity is a huge problem. The “subway lobbyists” never doubted the potential capacity for LRT. However, most LRT systems in the world have a limited frequency that cannot be compared to the frequency of an automated system. If we want to achieve a higher capacity LRT, that requires more room which isn’t available on Broadway. If the LRT on Broadway is built with the ALRT standards, then yes, the LRT will have a capacity that rivals that of ALRT.
    Infrastructure is expensive but is important. There is a place for LRT, there is a place for buses, there is a place for Streetcars, and there is a place for Subways. You have categorized many who have different opinions as “subway lobbyists” and “anti-LRT” when that’s not even the case. Personally, I strongly support LRT on routes such as Victoria/Commercial Dr, Marine Dr Hastings Corridor, 41st Avenue, Fraser Hwy, Interurban, and 200th just to name a few. I believe that those corridors are designed or can be easily reconfigured for LRT, with fewer pedestrians crossings or space constraints. It’s time to put an end to this name calling and begin debating like an adult. If you are to use scholarly sources from Professor Condon, I suggest you to begin debating like a scholar then.

  23. Adrian January 6, 2010 at 12:07 am #

    I like to correct my earlier post, first paragraph, last sentence to $4 million rather than $5 million.

  24. M1EK January 6, 2010 at 8:07 am #

    I wasn’t saying anything about Vancouver; you may or may not have the same issues as Honolulu; I honestly don’t know; but I do know that the standard surface LRT plan implemented successfully by many cities in the US is far too LITTLE for Honolulu.

  25. :) January 6, 2010 at 5:44 pm #

    Thank you Adrian!

  26. MB February 19, 2010 at 4:10 pm #

    I’m with Adrian, who deserves much credit for pointing out the obvious to anyone who has lived near or frequented Broadway: the existing pedestrian crossings are a crucial consideration in planning rapid transit in this corridor.
    Every single intersection in Central Broadway is signalized (Main Street to Arbutus Street). Moreover, 90+% of all cross streets on Broadway west of Main are signalized. Vancouver has developed its Greenway Program with several north-south crossings over Broadway. The sheer density of crossings make this is one of the most unique arterials in Western Canada.
    Thousands of pedestrians, bicyclists and service vehicles will be severely impacted by a dedicated (fenced) median for surface light rail. In essence, it will sever all crossings except at major arterials.
    Without a dedicated median, light rail becomes a milk run and merely replicates the existing B-Line express bus at great expense.
    Rergarding cost, I have spent 28 years in urban design and have managed many construction projects. I’d have to ask anyone who claims they know light rail can be built on this particular corridor from Commercial Drive to UBC for less than 1.8 billion dollars if they were distracted by swine flu when they did their calculations.
    Lastly, no one has addressed safety. LRT kills people at surface crossings. Period. For Zwei (and whatever his other monikers are) to say SkyTrain kills people too, is overtly callous because he’s including suicides. There are no crossing anywhere on SkyTrain. He’s done this before in desparation to post-justify his arguments.
    This doesn’t mean LRT can’t be implemented safely elsewhere (and Adrian pointed out a few of many different routes), and without a major improvement to human-scaled urbanism, but given the number of crossings on Broadway, you’ll either sever the cross links or put too many people at risk.
    The choice to me on Broadway is narrowed to two options: extend the Millennium Line in twin bored tunnels to UBC (do it much better than the Canada Line, but that’s another story), or making improvements to the B-Line without severing the existing crossings. Surface light rail is either a waste of money or a dangerous choice for Broadway.
    Lastly, I refuse to be characterized as being a ‘SkyTrain Lobbyist’. Such creatures exist only in the mind of one overly myopic light rail afficionado from the burbs who has created a jargon-ladden soap opera on several blogs, and who lives nowhere near Broadway or the inner city, and thus does not understand them.
    I am, though, a strong proponent of making our cities more resilient, sustainable and beautiful.

  27. tybuilding February 24, 2010 at 4:59 pm #

    I agree underground is the way to go, extend the skytrain from the milleneum line. Hopefully Honoloulu makes the right choice. The highways are congested there!