What Does Subway Construction Cost?

An interesting article on the site Pedestrian Observations was recently linked on Twitter by Taras Grescoe (@grescoe), the author of Straphanger.  In Comparative Subway Construction Costs, Alon Levy attempts to bring together projects worldwide, adjust for currencies, inflation and other factors to derive comparable US$ values for subway construction.  The numbers are interesting in light of complaints about overpriced construction in Toronto.

The underlying problem, of course, is that no two projects are the same.  Varying proportions of underground construction, different soil/rock/water conditions, variations in station numbers/size/depth, not to mention rolling stock procurement and yard/maintenance facilities all affect the total cost and hence the cost/km.  Stir in political differences and the ease or difficulty with which projects are approved, and the number of variables is quite high.

All that said, Toronto’s costs are not wildly out of line.  This is not to say that they may not be excessive, but the cause could be our extended design and approval process and a preference for deep bore tunneling that drives up construction costs compared to shallower cut-and-cover.

The argument for LRT has always turned on the availability of a surface option where it is practical.  Whether we choose to exercise this in every possible case is another matter.

There are cases where underground construction is the only practical way to build a line, but that should not condemn Toronto to building only subways without looking at alternatives.

60 thoughts on “What Does Subway Construction Cost?

  1. I was at Davisville yard last Saturday and there appeared to be six 6 car trains, 5 TRs and an H6, in the yard. I believe it is going to be converted into a works yard but would there still not be room in it for some trains?

    Steve: I think the problem is more that if the number of trains is small, the overhead of using this as a location to base crews and dispatch trains becomes quite high. There is a similar problem with the reactivation of Keele/Vincent Yard.

    The TTC does not seem to follow the practice of most other agencies that have a yard at or near the end of each line. They had one at Six Points which they foolishly gave up and Wilson used to be at the end of the line. The TTC does not seem to make adequate provision for storage of equipment when it builds new lines. Metrolinx is putting storage yards at or near the end of each line as well as downtown as this makes for more economic use of trains. I hope they remember this with their LRT lines.

    All of these convoluted ideas about building connecting lines from Wilson to everywhere are a useless exercise in spaghetti bending. Though we could end up with a true PPP line, Pearson, Peterborough via Pickering. It has a certain ring to it. The line would be operated by the Shining Waters or whatever name they have for the proposed Peterborough line operating company.

    Like

  2. Steve: All of these convoluted ideas about building connecting lines from Wilson to everywhere are a useless exercise in spaghetti bending.

    Spaghetti bending / tossing for transit and transport planning is a fine example of using your noodle(s).

    Cheers, Moaz

    Steve: That’s not my quote, it’s Robert Wightman’s. I happen to agree with him. There’s nothing wrong with using our noodles, and having done so, saying, nope the idea isn’t workable. Just because I disagree with you doesn’t mean I am incapable of rational thought.

    Comments that imply “my idea is best” and dismiss critiques as lacking insight or some sort of vision are no substitute for analysis.

    Like

  3. Robert:

    Davisville currently stores around 80 subway cars (not including work cars), of which it dispatches 10 trainsets (60 cars) for the Y-U-S during the peak periods. Once the line is fully converted to TR operations, it’s been projected that it won’t be able to store more than 8-9 complete TR trains. It will definitely continue to store revenue trains, but clearly stuffing more trains into Wilson – with its disadvantages – is the only option in the short term.

    The big issue is the night maintenance window, which will be reduced accordingly, unless they move to an earlier closing time of the entire system, around 12:30 AM or so. Not likely to happen.

    Like

  4. “However, what you describe would not be elevated, certainly not through downtown. On one hand you are extolling elevateds, but then you propose an alignment that is mostly not.”

    I should have been clearer that, yes, the downtown portion would be elevated. Why is that categorically dismissed? Anyway, an excellent alternative routing would be to bring a DRL into downtown (elevated) above Eastern Avenue and through Bremner, linking to the Y-U-S at Union Station’s revamped south end and serving such major destinations as the ACC and the Skydome. Then on to points west. Elevated would perhaps suit that route better.

    The broad point being … let’s start thinking of ways to do things rather than thinking of all the reasons not to. Someone has to bring some vision and leadership into this and get a DRL built.

    Steve: There was a time when an elevated on the route you propose might have flown, but the waterfront is now a built-up area and you would be putting an elevated rapid transit line right through the middle of it. If you haven’t been to the south end of the Don River recently, you should take a look. There is a whole cluster of buildings going up along Front Street East (which aligns with Eastern Avenue west of the river). Similarly, Bremner is lined with condos out to the west all the way to the CNE.

    This is not a question of saying “no” or of lacking “vision and leadership”. Occasionally it helps to actually know what is going on in the area where one cavalierly draws lines on a map.

    Like

  5. All the ridicule of loops and interlining seems to miss the main point. I think everyone would agree that a Sheppard interline to York U would add more riders, and the interline to Union would take riders off the Yonge line. Maybe these numbers would not be huge, but if the Sheppard line is being extended, they are not huge expenses. What it would do is give the public a transit system that is better suited to their wants and desires. This would go a long way towards getting acceptance towards the entire Big Move plan and the necessary funding. I imagine we would think nothing about adding 10% to the cost of stations for aesthetics, but spending extra 10% to satisfy rider wishes seems out of the question.

    Like

  6. I’m curious why individuals keep bringing up the Kingston sub ROW as a way of reducing construction costs of a hypothetical DRL. There doesn’t appear to be enough room to set aside the extra 30-40 feet that a HRT ROW would require.

    From a pragmatic standpoint surely any room for a fourth set of tracks should be reserved for future GO-REX and not taken up just to save the TTC a few dollars in capital expenses.

    Steve: I agree. If this corridor is used for any part of the DRL, it would be underground, not at grade.

    Like

  7. Cheers, Moaz

    Steve:

    That’s not my quote, it’s Robert Wightman’s. I happen to agree with him. There’s nothing wrong with using our noodles, and having done so, saying, nope the idea isn’t workable. Just because I disagree with you doesn’t mean I am incapable of rational thought.

    Thanks, my mistake on that.

    Agree with the rest.

    Walter:

    I think everyone would agree that a Sheppard interline to York U would add more riders, and the interline to Union would take riders off the Yonge line. Maybe these numbers would not be huge, but if the Sheppard line is being extended, they are not huge expenses. What it would do is give the public a transit system that is better suited to their wants and desires.

    Steve: That Sheppard extension is a very big “if”. I can agree that a through service to York would be useful for some, but the problem is the cost of building the connection versus the likely ridership. As for a route to downtown, I am less sure. As you say …

    TTC tried interlining for 6 months and decided it didn’t work. They decided long ago that train interchanges would require customers to exit and go up or down to a separate platform … or take a ‘short’ distance walk to a separate platform.

    In contrast, Montreal decided to build paired cross-platform interchanges where possible. This happened around the same time that TTC was trying interlining.

    It would be interesting to understand why one worked and the other didn’t.

    Cheers, Moaz

    Steve: The TTC’s interlining was doomed because of how it was scheduled and operated. Rather than treating each service as an independent route, the TTC tied itself in knots trying to force trains to alternate so as not to “confuse passengers”. (That phrase, like “traffic congestion”, is a TTC catchphrase intended to preclude debate.) If that wasn’t bad enough, they had trains changing between routes so that it was essential for crewing for everything to be in exactly the right place and time. We all know how good the TTC is at that sort of thing, even without any external interference from plagues of locusts and other not-our-fault issues that might beset the subway.

    A three-way integration would be tricky enough to manage on close headways under ideal circumstances, but that’s not what obtained, especially due to the way the TTC insisted on operating the line.

    An across the platform transfer is much easier because the passengers do the work, not the trains. However, stations must be designed for that operation from day one, not as a retrofit.

    Like

  8. “Occasionally it helps to actually know what is going on in the area where one cavalierly draws lines on a map.”

    Been living in the downtown for about 14 years. But that’s about what I expected as your response. Yes, there are some new condos – all the more reason it would be a good alignment. There’s also a heckuvalot of empty lots and low-rise, smaller, insignificant, industrial buildings. Show me a better route.

    There is such a thing as making the perfect the enemy of the good. Fine, bring a DRL underground on the downtown side of the Don Valley and we’ll wait 20 years and pay billions more to see it run.

    But let’s just carry on discussing streetcar headways and interlining and pantographs or whatever the frig they’re called.

    Steve: You have still not addressed the issue of running an elevated subway line through the middle of a residential area, let alone how it would get through the core itself. Please consider this conversation closed.

    Like

  9. Timur Urakov says:
    June 13, 2013 at 3:35 pm

    “Davisville currently stores around 80 subway cars (not including work cars), of which it dispatches 10 trainsets (60 cars) for the Y-U-S during the peak periods. Once the line is fully converted to TR operations, it’s been projected that it won’t be able to store more than 8-9 complete TR trains. It will definitely continue to store revenue trains, but clearly stuffing more trains into Wilson – with its disadvantages – is the only option in the short term.”

    Thanks for the information. I couldn’t tell from the published info if they going to continue to run service trains or not. It would seem to be a waste to only use the yard for works equipment. Do you know if they are going to lengthen the maintenance building to hold more 6 car trains.

    Like

  10. rocinante says:

    “I should have been clearer that, yes, the downtown portion would be elevated. Why is that categorically dismissed? Anyway, an excellent alternative routing would be to bring a DRL into downtown (elevated) above Eastern Avenue and through Bremner, linking to the Y-U-S at Union Station’s revamped south end and serving such major destinations as the ACC and the Skydome. Then on to points west. Elevated would perhaps suit that route better.”

    The reason elevated is categorically dismissed is because most politicians want to get re-elected. Less than 50% of the population normally votes but if you tell them that you are going to put an elevated down their street I guarantee that the turn out will be much higher. If you were the dictator then you could build elevated lines everywhere but they are non starters in downtown Toronto.

    How do you plan to get from Eastern to Bremner? Eastern runs into Front Street which is North of Union Station and Bremner has its east end at the ACC south of Union Station. The TTC wants 1000 foot radius turns, 300 m, for its subway, rapid transit, cars. Show me where these curves will be. Granted if you go west of Union over Front to the North Bathurst Yard area there is room to curve but there are already high rises planned for most of those area.

    A transfer at Union is a non starter because there will not be platform capacity with all the passengers coming off GO. Do a little rational analysis of your plan before you take pot shots at someone else because their rational comments don’t suit your preconceived, and ill conceived, ideas.

    Like

  11. I am sorry to carry on about Sheppard interlining, but I think there is a big difference between it and the 1960’s interlining at St.George. I see the two big problems before were that stations were very close to each other (which led to the upper/lower Bay problems) and that each branch of the subway were equally important (leading to the alternating trains problems at the crossover). At Sheppard, the stations are far enough apart that there is plenty of room for connections. Also, the frequency on Sheppard is half as much so there could be a planned wait for trains to ensure the Sheppard trains join the Spadina line at the correct time.

    In terms of construction, I see Sheppard being on a bridge above the West Don, leaving only a bit over 2km to reach the Spadina line. This could be built by cut-and-cover to avoid the relative high cost of launch and extraction sites on such a short stretch. Again, I don’t think this is the most urgent project, but the acceptance of a somewhat more disruptive (and less expensive) construction method could be used to increase its priority.

    Like

  12. Robert:

    I haven’t heard of a plan to expand Davisville. There may be something in the 10-year capital budget papers, but I haven’t checked those. The only practical addition would be lengthening the carhouse on tracks 1 and 2, so that they can each hold a full 6-car trainset inside the building (normal procedure currently is to split a six car train, so that 2 married pairs are stored inside, and one pair outside the carhouse on both of those tracks, but that will no longer be possible with a 6 car fixed configuration trainset). That would add 2 trainsets – capacity-wise. Not very much.

    But in any event, the yard will continue to be used for service trains, but only for storage and cleaning. It cannot service the new TR trains as it does not have the necessary hoists. The minor repairs they are now doing in there on the T1 cars (motor replacements, compressors, driveshaft repairs etc.) cannot be done on the TR’s. The biggest change in the short term will be the consolidation of the work car fleet at Davisville to free up storage room at Greenwood.

    Steve: I have read the capital budget blue books, and they line up with what you describe. The more pressing problem for space is on the BD line with the shift of the T1 fleet off of Yonge (thanks to some not too bright fleet planning, but that’s another story) and the need to squeeze every inch of space out of the BD line to hold more trains than are needed for service there.

    The scheme for a twelve-trainset underground yard exists both in a Finch-Cummer and a Richmond Hill version. Where I think the TTC will get “bitten” is that they will need more capacity (ie shorter headways) on Yonge before they can possibly get the DRL built, but won’t have enough trainsets (or the storage to hold them) to actually operate that frequency. They keep talking about the benefits of ATC and closer headways but don’t address the fleet implications.

    Like

  13. Walter said:

    I see Sheppard being on a bridge above the West Don, leaving only a bit over 2km to reach the Spadina line.

    I sense another example of not checking a route before making a suggestion. On the north side of Sheppard in the Don Valley is a group of homes with the road access to them on the south side linking up with Sheppard just west of the existing bridge. When you factor in Earl Bales park to the south and the apartment buildings on both ends of the bridge on the south side and the ones to the west on the north side, the only bridge option for the Sheppard West extension would be to replace the existing road bridge over the West Don with a new one that goes straight across the valley with a subway deck beneath it. Since the only way to build it would involve closing Sheppard, you also have a steep political cost to go with the financial one.

    Steve: And, by the way, bridges are not cheap. Less than tunnels, to be sure, but not free. For both segments, east and west of the river, cut-and-cover is likely how they would be built given how short they are. Depending on where the stations were going to be, this may not endear the project to the neighbours who have the pain of construction with no benefit in service.

    Like

  14. I don’t know what the fascination is with interlining, especially Sheppard, with its 4 car trains with Spadina with it TRs. The existing stations would need to be finished to the full 6 car length, an unnecessary expense. Also Agincourt to downtown via Downsview is not very direct route. How can it be better than a DRL? I mean a Downtown Relief Line and not the Downsview Relief Line. Some people seem to believe that this line would be for the benefit of those bike riding latte swilling leftists in the downtown but the people it would really help are Ford’s little people who live in the outer 416 and need to get downtown to work.

    To propose that it could be used to run trains from Wilson to the Yonge line to run trains into and out of service is going to be very expensive and would require major upgrades at Wilson yard and at Sheppard Yonge station. It would probably cheaper to build a full yard near Sutton while providing a more useful service. We could get 416 from Rockwood to inaugurate service.

    To build a more useful transit service we need to complete more parts of the grid, not build more pretzel routes to go downtown via many circuitous routings. The most cost efficient way to build is LRT not HRT.

    Like

  15. It occurs to me that if we’re going to talk at length about hare-brained schemes, we should talk about all the details involved in interlining the Sheppard subway with the Sheppard LRT — i.e., converting the existing stations to LRT. Even if impractical, I’m sure it’s less infeasible than building subway lines all over the place. And hey, at least it wouldn’t involve lengthening the Sheppard line stations.

    Like

  16. Robert Wightman said:

    I don’t know what the fascination is with interlining, especially Sheppard, with its 4 car trains with Spadina with it TRs. The existing stations would need to be finished to the full 6 car length, an unnecessary expense. Also Agincourt to downtown via Downsview is not very direct route.

    I personally thought that the stations didn’t need to be finished up to the 6 car length as the doors of whichever carriages (presumably first and last) are not in station would be locked out. The driver would not be able to have eyes on the station but the guard would. And, when the train entered the Yonge line (bypassing Sheppard-Yonge) the guard could move to the last car and operate like all the other Y-U-S.

    And it would be Don Mills to Downsview via Union🙂

    In any case this is just fanciful thought and can only continue as long as it doesn’t get too crazy.

    Issac Morland said:

    It occurs to me that if we’re going to talk at length about hare-brained schemes, we should talk about all the details involved in interlining the Sheppard subway with the Sheppard LRT — i.e., converting the existing stations to LRT. Even if impractical, I’m sure it’s less infeasible than building subway lines all over the place. And hey, at least it wouldn’t involve lengthening the Sheppard line stations.

    I’m with you on that. A conversion of Sheppard to LRT is a great and feasible idea and may become viable after 2020 when people see that the Eglinton Crosstown LRT line does work (and the 3 car trains actually carry more people than Sheppard).

    The advantage of Sheppard as an LRT is that you can use the existing bridge over the West Don River which would save significant money. Being able to extend the line past Downsview (ideally through Downsview Park) to Jane or Weston Road would be great … of course it would be better with a Jane LRT.

    The big challenge is the LRT needing a larger tunnel than the subway, the presence of the 3rd rail, the high platforms, and of course all the money spent on the connecting track to the Yonge line.

    I know that at Don Mills the tracks would be raised to the high platform level. Does that mean that the existing ceiling height in the stations (or just Don Mills) is enough to accommodate an LRT? Or is there special work required?

    I hope we can continue to discuss the fanciful.

    Cheers, Moaz

    Steve: The question of LRT in the subway tunnels has been examined before. Aside from the need to shut down the line while it is reconfigured (and please let’s not start a comment thread on the many ways this might be accomplished — we have heard it all here before), there are some clearance problems notably in the box tunnel sections of the line (cut and cover) where the headroom of a round tunnel is not available.

    Like

  17. Moaz Yusuf Ahmad says:
    June 17, 2013 at 10:30 am

    “I personally thought that the stations didn’t need to be finished up to the 6 car length as the doors of whichever carriages (presumably first and last) are not in station would be locked out. The driver would not be able to have eyes on the station but the guard would. And, when the train entered the Yonge line (bypassing Sheppard-Yonge) the guard could move to the last car and operate like all the other Y-U-S”

    If one is operating TR trains then the only possible location from which to operate the doors is either the first or last car. Since they will hopefully be using One Person Operation there would be no guard and the operator would be in the front car. There is no way to get around the problem. Interlining Sheppard and Spadina is a non-starter, don’t even contemplate how to do something that has no rationale for existing.

    Like

  18. Steve:

    There are some clearance problems notably in the box tunnel sections of the line (cut and cover) where the headroom of a round tunnel is not available.

    Presumably there is no headroom issue at Don Mills Station? Since my last trip on Sheppard was a decade ago, could you refresh my memory on which sections are deep bore and which are cut-and-cover?

    Steve: All locations with special work (crossovers) are cut-and-cover. Also, the section east of Leslie Station.

    Robert Wightman said:

    If one is operating TR trains then the only possible location from which to operate the doors is either the first or last car. Since they will hopefully be using One Person Operation there would be no guard and the operator would be in the front car. There is no way to get around the problem. Interlining Sheppard and Spadina is a non-starter, don’t even contemplate how to do something that has no rationale for existing.

    Except that Sheppard uses T1s and will continue to do so for at least another decade (and will probably continue to do so after Yonge-University-Spadina switches to TRs).

    Steve: If you are going to interline, then you have to assume that any train on YUS might wind up on Sheppard, not the handful of T1s it has today. Moreover, adapting the T1 fleet for ATC is not going to be cheap. Finally, by the time any of this happens, the T1s will be nearing retirement and the TTC is probably going to have to buy a few “short” TR trains. These will not be able to interline for obvious reasons.

    Remote but still possible = interesting topic for discussion … and least to my mind🙂 Tomorrow is spaghetti night and I promise, no new crazy ideas.

    Cheers, Moaz

    PS. Steve, (if you are following the Gardiner East proposals), I’d like to know what you think of the Rem Koolhaas “Replace” proposal (which includes a Downtown Relief Line that serves a proposed new downtown east at the foot of Broadview). Gardinereast.com for more info. m

    Steve: I am least impressed by the Rem Koolhaas scheme of the whole lot. There is a common problem with some of the Gardiner proposals in that they are more suited to a complete expressway replacement rather than one stopping at Jarvis where the transit corridor would have to fend for itself. Also, the DRL south of the rail corridor is a non-starter as this would create a powerful disincentive for “relief” traffic that would not be delivered directly to the core. It is physically impossible for the DRL to serve the main business district and become a local subway line for the waterfront. I have an article in Torontoist that will appear later today (June 18).

    Like

  19. Steve:

    “If you are going to interline, then you have to assume that any train on YUS might wind up on Sheppard, not the handful of T1s it has today. Moreover, adapting the T1 fleet for ATC is not going to be cheap. Finally, by the time any of this happens, the T1s will be nearing retirement and the TTC is probably going to have to buy a few “short” TR trains. These will not be able to interline for obvious reasons.”

    At least the short TR train sets could be lengthened by the purchase or extra “B” and “C” cars. I will be truly amazed if the TTC manages to keep all its TR sets in proper [order]. Any place I have ridden these types of trains the numeric order is all screwed up as they take working cars from one dead train to place in another train to have one working and one dead train instead of two dead trains.

    Like

  20. Sorry if I’ve caused offence. None was intended. It just seemed to me that a thread about subway costs would be a reasonable one in which to suggest that els are much less expensive.

    1. Classic NIMBYism is the enemy of pretty well any plan. However, the citizenry at large might favour a leader who supported an alternative which could be built years sooner at a much lower expense.

    2.My suggestion is painted in broad strokes. There are thousands of impediments to be overcome in any plan. I do not purport to be any kind of expert with a fully-developed blueprint. I’m well aware that Eastern Ave. and Bremner aren’t on the same side of the tracks, which is partly why my original proposal was to use Richmond or Adelaide.

    Apparently, this topic has been declared closed. I hope you’ll be big enough to allow my response.

    Like

Comments are closed.