Above: The Canada Line at Marine Dr. Station. Featured photo by Larry Chen.
There’s been a lack of clarity when it comes to the big numbers that define the planning of transit systems in Canada. It’s particularly evident when transit technology becomes a matter of discussion.
Of course, millions of dollars are at stake. So there’s no doubt that when the cost estimate for a major project is higher by so much as a few million dollars, it’s the kind of thing that sends transit advocates scrambling to get attention and some people in the media practically screaming.
So I decided to take all the recent and upcoming Light Rail projects in Canada, research their costs and alignment details, and put them in a table for proper comparison. I put the data in a Google spreadsheet:
All projects were included regardless of technology. Alignment was divided by percentage and split into/measured in 7 categories: on-street, above-grade (i.e. elevated), below-grade (i.e. tunnel, open cut), disused R.O.W. (i.e. railway R.O.W., other empty lands), bored tunnel (the most expensive kind of tunnelling), shared-lane (on-street in mixed traffic like a streetcar), and the total at-grade percentage.
Since the transit planning complaints here in Vancouver always seem to be directed at grade-separation, I decided to focus on seeing if there was a cost trend regarding the amount of grade separation for the line.
Same data as above, but sorted by amount of grade-separation
What I found is that there is a trend that occurs when the chart data is pinpointed on a graph and assessed by percentage, but it’s very inconsistent and the projects are all over the map:
Several projects end up below the average and several end up above it. As an example, there’s a difference in the four projects on this chart closest to the 100% mark. The highest mark is for the proposed Scarborough extension of Toronto’s Bloor-Danforth subway line, which will be fully underground. The lowest mark is from the estimate for a SkyTrain Expo Line extension in Surrey, which will be fully grade-separated but built in an elevated guideway as opposed to a tunnel.
Despite the use of grade-separation, many of the highest-cost projects are not fully grade-separated and feature many at-grade segments that can limit potential. Even projects with only about 20% grade-separation can come close to or even breach $200 million per km.
In order to account for the differences associated with much more expensive below-grade (tunnelled) segments, I took the data and assessed it by percentage below-grade and found a much steeper and more consistent trend-line:
The amount of systems at the 100% mark has decreased from 4 to 3, and the trend-line now hits the middle of these three dots. The middle dot, closest to the line, is the current ongoing extension of Toronto’s Yonge-University Spadina subway line. The lowest dot is the cost estimate for the ‘Broadway Subway’ (the Millennium Line’s proposed extension down Broadway), which is below the trend-line but is built around a medium-capacity system unlike Toronto’s fully-fledged, high-capacity subway.
Still, there are some differences to account for in terms of alignment. At the 45-50% mark there are two projects that deviate both from the trend-line and from each other.
The higher of these two marks, at $279 million per km, is the Eglinton Crosstown LRT being built in Toronto. The Crosstown was planned as an on-street LRT system, but the central portion will be placed in a 10km dual underground bored tunnel, which spans more than half of the final construction. The lower of these two marks is actually our SkyTrain system’s Canada Line. The Canada Line is a fully grade-separated light metro and a slightly higher total percentage of it is below grade. However, only a much smaller portion of this is expensive bored tunnel – the rest was done as less expensive cut-and-cover. Therefore, it manages to be less expensive despite the full grade-separation.
To account for that difference I created one more plot excluding everything but projects with bored tunnel segments. The plot line managed to stay the almost same, and the relationship between high capital costs and tunnels is thus made clear:
Since only 13% of the Canada Line was built in a bored tunnel, it is now to the left of where it was in the last chart and sitting very close to the trend-line (the Eglinton Crosstown is also closer to the trend-line). Meanwhile, our Evergreen Line SkyTrain extension, which encountered challenging soils with its single tunnel bore, is right on the trend-line when set amongst the other systems.
Canada can’t be compared to Europe
The Tyee has probably been one of the most prominent to sound the cost-comparison alarm when they published a 2012 article titled, “Why Is TransLink’s Price for Light Rail Triple What Other Cities Pay?”
This article surmised that our Light Rail cost estimates are triple what they should be, based on cost estimates being about one-third as much in European and American cities. (And it was, of course, brought up as a way of hurling tomatoes at the idea of a Broadway Subway line – which is still a great idea for a number of reasons).
Interestingly, of all the American cities that could’ve been chosen in the comparison, it was Minneapolis and its Hiawatha Blue Line. This comparison is invalid as over 80% of the line is placed in either disused R.O.W. or tunnel, with only 20% of it being on-street. All of the other examples are from cities in Europe.
Regardless of whether you believe these numbers or not, the reality is that transit projects and their costs are more complicated than being able to be broken down into a simple cost-per-km value that can apply nationwide, across nations, or across transit projects. There are differences in labour laws, work schedule expectations, material costs, acquisition costs, logistics costs, varying land values, differences in local terrain and differences in economy. All of these need to be accounted for and thus it can’t be assumed that a transit project that cost a certain amount in Europe (or any other country, really) could be replicated in Canada for a similar cost.
Here in Vancouver, for example, any big rapid transit projects are likely to cost more than anywhere else in Canada simply because the higher cost of land would likely significantly raise the costs of project elements such as the operations & maintenance centre (OMC).
Despite this, at the end of the day, both the Broadway Subway and the LRT proposals were consistent with the trendlines across Canadian rapid transit systems.
To further address the point raised by The Tyee, I compiled one more chart between the predominantly on-street LRT systems:
From the wide spectrum in cost of what would otherwise be similar at-grade, on-street LRTs, it may appear that The Tyee would have a point. Even this can be explained, however. The two lowest-cost systems on this chart are Kitchener-Waterloo’s ION rapid transit and the proposed Victoria LRT system. They also happen to have the highest percentages (44% and 31% respectively) on a disused right-of-way (i.e. beside a railway), which is the least expensive place to build any transit because there’s no utility removal, property acquisition or street-scaping work adding to the cost.
In the middle are the Mississauga and Hamilton systems, which are slightly lower than the big-city systems in Greater Vancouver and Greater Toronto (they are also among the 3 systems with occasional mixed-traffic rights-of-way), which seems just right to me. The Mississauga system (Hurontario LRT), in particular, is being built on a wide roadway that in most places still has significant allocations on either side where the roadway can be expanded if necessary (in other words, there’s almost no property acquisition).
The cost for a Broadway LRT system is certainly on the high-end of the spectrum. This makes sense as a Broadway system would need to offer the highest capacity of all of these systems and would face street-scaping challenges with the need to stay within property lines (though this won’t stop property acquisitions from being necessary at station locations). There’s also the uncertainty around an OMC, which would have likely had to be built underground and/or expensively due to the lack of lands along Broadway and high land costs in Vancouver.
In the end, the amount of bored tunnel has a somewhat linear relation with project costs – but grade-separation altogether does not. This doesn’t mean we should avoid building systems with bored tunnel segments from end-to-end (at the end of the day, whether to go that far or not should come down to detailed evaluations of each corridor and transportation needs), but what I do hope to achieve with this article is to facilitate an improvement in the discussion of rapid transit projects (Especially capital costs, since it seems to be the only thing people want to talk about when thinking of rapid transit projects – I, of course, completely disagree).
It’s time to stop thinking that we can build paradise if we replicate the results of other countries, at the costs those other countries experience – it’s impossible. Let’s build transit systems that are adapted to the way our cities work, so that we are sure to be rewarded with positive outcomes.