A Train Derailment, and The Need for More Downtown Train Stations in Toronto.

My pet project returns.

It is very uncommon, if not unheard of, for a major transit-oriented city to have a single city centre landing point for all regional and national rail services. I was going to cite New York with Penn station, but New York also has Grand Central, which as of a few years ago has service on the LIRR (while Penn Station access will bring MNRR trains to Penn).

London has a large number of rail terminals in its city centre, and with both Thameslink and the Elizabeth line, it also has a number of central through stations built underground that carry these services. A similar pattern can be found across numerous European cities, where at the very least suburban and regional trains tend to run through the city centre, usually in tunnels, often with just two tracks. This is the setup in Stockholm, Copenhagen, Paris, Berlin, Madrid, Milan, Oslo, and many more.

Toronto, on the other hand, has just Union station.

A packed Union station after the derailment.

The issues with this are many, but the main ones are capacity, coverage, and resiliency — something highlighted last week when a train derailing, combined with our “flexibility” oriented rail operations — where trains coming from any track are able to get to many tracks at the main train station, sort of like a plane being able to freely taxi around and airport — caused huge havoc across the entire Toronto regional rail network for days, as well as on the airport rail link. A single track having a problem in most of the cities I mentioned above would not have that effect. In some cases, it might shut down one line, or set of lines, or even all through traffic if it was in a critical location, but trains could fairly easily terminate elsewhere.

Resiliency

Unfortunately, the way things are set up in Toronto, our city centre station and its supporting infrastructure are an obvious single point of failure, at the centre of an interwoven and codependent network of trains that circulate throughout the region.

And while last week the single point of failure was in an important bit of track, the reality is it could be numerous other things: an impact to the station structure, or even some untold thing which renders platform space unusable.

You might expect that in such a brittle setup we would have supremely strong signalling technology to prevent a train derailing or going overspeed, but we do not. The entire GO network lacks positive train control, and instead of creating provably-safe electronic systems, we have created high-redundancy and high-cost human systems that still fail.

More or less, at every level, we have a system which is not resilient.

Capacity

Capacity is another huge issue. Both because its location, and the concentration of development, Union Station is superbly located with a huge number of jobs and destinations within a 15-minute walk of it.

Unfortunately though, the station itself only has so much capacity. The building feels large, but actually doesn’t have much space dedicated to passenger circulation, certainly compared to other major hubs in the developed world. The major expansion project that was undertaken in the 2010s was helpful, but it mostly added new amenities instead of improving the pure flow and function of the facility — this isn’t helped by dysfunctional policies that abound in the station’s operation, like numerous doors that need to be opened by passengers, and escalator banks that are not adjusted to best serve demand patterns throughout the day.

These issues will be mitigated somewhat with major future construction projects, which will widen platforms (creating more space at track level) and increase vertical circulation from platform to concourse level. However, these measures can only do so much. The concourses end up becoming the limiting factor, and there just isn’t that much real estate here. Crowding at Union is already severe at certain times of day, and even if this is made marginally better by additional space and operational changes, the reality is that every other factor is pushing us towards more congestion — significantly more development in the immediate area and waterfront (office towers have increasingly moved south to the area around Union Station — CIBC Square basically forms part of the station at this point), regional and urban population growth, massive increases to rail service, and mode shift.

These issues are already also becoming real problems for the subway and streetcar. These services already play the role of getting the vast majority of people arriving at Union, but not travelling to within a short walking distance (relieving these by improving, say, cycling options, would be a rational and cost-effective thing to do, however our government at the provincial level is instead driven by ideological considerations first on this issue). Unfortunately, as you can see most any day, the subway and streetcar platforms at Union are very crowded. There is a plan to expand the streetcar portion of the subway station with a series of straight platforms, meaning the current platform would function solely as a loop, and this should significantly increase capacity here (as well as improve circulation with a new direct connection into Union Station). In theory, we could also add one more subway platform for a Spanish solution setup of boarding for northbound University trains only from the island platform with exiting passengers using a new second side platform, and this would be made more practical with screen doors. However, these interventions only allow us to unlock the full subway and streetcar capacity through the station, which pales in comparison to that of the other rail services, which I would approximate as having the capacity of at least 7 directional subway routes, and up to 9 during peak periods, several times the ~2.25 available through the TTC station between the subway and streetcar (obviously cycling, buses and walking can absorb some number of passengers, but not the difference).

Coverage

Then we come to the final, and perhaps least critical issue — coverage.

Right now, regional trains have a single landing point in downtown Toronto, and the reality is that many major “downtown” locations like the hospitals, Queen West, UofT, The Eaton Centre, the sort of Bloor-Yonge and Bay offices, the AGO, and the ROM are not Union Station-adjacent.

To some extent we are solving this, and our plan is a bunch of new regional rail to rapid transit connections (along with new rapid transit that we can connect).

Calling this a plan is generous, it’s more like people eventually seem to be realizing that having two rail lines intersect and not connect is silly, but there doesn’t appear to be a lot of rational high-level planning going on (or at least there was not 10 years ago), which is reflected in the way things are planned.

If you look at the “GO RER” network as consisting of five radiating corridors — Lakeshore West, Kitchener, Barrie, Stouffville and Lakeshore East, which correspond with west, northwest, north, northeast, and east — then we have planned connections between all five corridors and rapid transit, three of those connections are integrated in-station, and two are fairly high-capacity.

For the Lakeshore West line, we have the new Exhibition hub station, a massive structure south of Liberty Village being built as part of the Ontario line.

Then for the Kitchener line we have Bloor-Dundas West, two existing stations that are currently being connected by a new (not particularly high-capacity, and so likely to need upgrade sooner or later) underground walkway. These stations already exist and lots connect between them, but a direct connection makes that connection faster and weather-protected, which will encourage more transfers. It’s worth noting that if the King-Liberty station ends up getting built, this will effectively act as a sort of northern annex for Exhibition station on the Lakeshore West line, as it is a short direct walk on a local street and a pedestrian bridge to connect (this would be very useful and would mean some connections between these services could happen here instead of Union, or optimistically Spadina).

Currently, the Barrie line has no station in central Toronto, with the first station after being Union being at Downsview Park near the city’s northern edge. A new station is underway at Caledonia to link to Line 5 (which already connects to the Kitchener and Stouffville lines, meaning it captures the northern three RER lines), but more importantly for our purposes, another new station is moving ahead at Bloor-Lansdowne, which, while, not in-station, will involve a very short walk down a multi-use path over Bloor street, and then turning onto a minor side street to access the subway.

Then, at the southeast of central Toronto, we have East Harbour station, which is similar in scope to Exhibition, but on the east end of the core, and which is also being built as part of the Ontario line to cement a new district near the waterfront. East Harbour’s design is less impressive than Exhibition despite some pretty incredible conceptual renders over the years, and like with Exhibition (and Bloor-Dundas West for what it’s worth) will bring together subway and eventually also streetcar services — which will provide additional onward trip options.

So basically, when you look at it, the five “GO RER” corridors all end up connection to an east-west rapid transit line across central Toronto, with the Lakeshore and Stouffville lines connecting to the Queen Street corridor and the Ontario line, and the Kitchener and Barrie lines connecting to the Bloor Street corridor and Line 2. There are also other “interceptor” stations further out, including Mt. Dennis on the Kitchener line, Caledonia and Downsview Park on the Barrie line, and Kennedy on the Stouffville line.

A real missed opportunity is that there isn’t a better, high-capacity connection at Danforth station (where the Stouffville line ought to stop, if only to connect with the Lakeshore line much further east of what is likely to be a very busy hub at East Harbour), there was even a news video highlighting congestion at this station recently, highlighting how little thought the powers that be have been, and have given it in the past decades.

This connection would allow Stouffville line passengers to save time by transferring to Line 2 further to the west, and also properly connect the Lakeshore East line to the subway system east of the core, also making it so that all but the Lakeshore line are connected to the Bloor Street corridor (if we had King-Liberty, then all but the Barrie line would be linked to the Queen Street corridor).

NOT THE RIGHT STATIONS!

Now, all of these very useful stations are what seems to come to mind for people when I say “We need more downtown train stations!”, which is both funny and a bit troubling: funny because somebody — who seems like they must live in Calgary or something, considers East Harbour or Dundas West to be “downtown”, but troubling because what I’m trying to get across here is stations serving the “central business district”, which is a term we just don’t use much in this country.

So let me be explicit …

When I say we need more downtown train stations, I explicitly see this as two new stations (it could be more, but at least two seems like the bare minimum) roughly one kilometre east and west of Union station — likely around Spadina Avenue and Sherbourne Street.

These stations play a very different role from the ones I just talked about. The stations I just talked about largely solve the problem of “centre city distribution”: instead of having everyone go to Union and ride Line 1 back north, many trips in the future can involve folks taking GO trains getting off early and taking a different subway line to their final destination. For example, someone from the Lakeshore West line taking the Ontario line to the Eaton Centre, someone from the Kitchener line taking Line 2 to the ROM, or someone on the Stouffville line taking the Ontario line to Queen West. These stations are very useful for these connections, and they can also be useful fallbacks when Union has a problem (they should be set up to act as termini, which they currently are not) — trains can drop passengers off here, and they can use the subway network, and some streetcars to get to their final destinations.

The issue is, this is only sort of half of a reasonable solution to central city distribution. That’s because Union is such a key hub right in the centre of downtown, that a lot of density is clustered around the density which is clustered around Union, meaning that there is a lot of density one kilometre east and west of the main station — the huge The Well development for example. There is no subway super close to The Well, but a large station there for regional trains would more or less provide the same access. In fact, passenger riding Line 2 from the west could get onto a Kitchener GO train (filling the seat of passengers getting off to go to UofT for example) to ride here.

Such a station would also just be useful because as anyone travelling will tell you, backtracking is psychologically annoying. For many people, riding GO trains into Union every day they need to watch with frustration as they roll past their final destination, only to turn around on foot at Union and head back.

For example, for people working at the offices at 320 Front, or at The Well, or at CBC, getting off a GO train approaching Toronto from the west and then walking to their final destination on foot would both likely be faster (plus these places don’t have Path access anyways so that doesn’t make a difference) and less crowded. You can imagine similar situations coming in from the east.

While these stations would not have last-mile subway connections (Spadina has streetcars, which if I can do anything about it will be much better), people often forget that pedestrian infrastructure is actually very high-capacity, and that a lot of the places you’d be serving are already places people walk to from Union — for example the SkyDome.

These particular locations need not be set in stone, but seem like the best bets to me, because of the good connections to major north-south streets and the proximity to Union. If Union isn’t operable for some reason you can easily walk to that general area from Sherbourne or Spadina in 10 minutes, it’s a much longer walk from Exhibition or East Harbour, much less Bloor-Dundas West.

These stations are also importantly located on the Union Station rail corridor where all services on the network run on the same corridor; this lets you build platforms connected by concourses that let passengers change between lines on regional trips — say, Markham to Oakville, or Brampton to Scarborough. Passengers can do this at Union, but changing trains right before the main station means you’re more likely to get a seat as people start standing up to disembark, but also that these stations can be purpose built for huge transfer capacity with big concourses and wide platforms, as well as natural light, which is sorely lacking at Union.

That the need for these stations needs to be explained is super weird, and given how unusual this suggestion is in Toronto, despite something like this being widespread in basically every city with a competent regional rail system, I’ve been giving thought to why that is.

For example, in London ,Crossrail has Bond Street,Tottenham Court Road, and Farringdon, in Berlin the S-Bahn has Hbf., Friedrichstraße and Alexanderplatz, and in Paris RER A has Charles de Gaulle-Etoile, Auber, and Chatelet.

Notice stations are evenly spaced even right through the centre of London where there is not a single “downtown” station. By David Arthur, CC BY-SA 2.0 ca, https://commons.wikimedia.org/w/index.php?curid=328648

“Central” Paris is roughly between Charles de Gaulle -Etoile and Nation.

Locals generally consider “central” Berlin to be within the ringbahn, at the very least it has a number of S-Bahn stations. By Arbalete - Own work, CC BY-SA 4.0, https://commons.wikimedia.org/w/index.php?curid=6532703

Central Copenhagen is more or less where all the S-Tog lines stop from Osterport to Copenhagen Central.

I think in part, people’s psychogeography hasn’t really updated to both recognize that there is a crap-ton of density east and west along the Union Station rail corridor, but also that these places — Spadina and Sherbourne — are actually decently far away. People sometimes act like this is close to Union for more stations, but at one kilometre, it’s more than twice the distance east and west of Union as the next stations north of Union on the subway (which is what we are very roughly trying to emulate with GO — thus 1 km centre city station spacing and not 400 metres). We are also blessed with through-running as a given which means the obviously missing thing is a little less obvious. For Paris and London it’s “the trains don’t run to the actual city centre!”, in Toronto it has to be “the trains don’t run to these places adjacent to the city centre where the main train terminals would probably be in most European cities!”. There’s also the issue of speed: the Union Station rail corridor operates at a snail’s pace, and so the idea of adding stops seems torturous.

So let me paint you a picture. What we should hope for is something like these European cities. GO trains don’t go particularly slower between Spadina and Union than they do between Bloor and Spadina, and we are making progress on this by simplifying the track work and resignalling (slowly) the Union Station rail corridor. With three stations here, we ultimately should hope for a subway-like experience: does the train go at regular speed to King before crawling to Union? Well, there is a super tight curve, but not really. People sort of accept a slow crawl into Union because they mostly all get off at Union right now, but that will be less and less the case as more trains through-run and carry passengers through the city. Part of why passengers don’t ride through right now is that most people are obviously going to the Central Business District, and there are not other stations you could ride to beyond Union that would get you closer to your destination.

Part of the reason people on the GO rail-side seem to have a mental block on this is because of the series of slip switches east of Union, and the huge flyunder to the west, both of which would be done away with when such stations get built. We have an extremely conservative and non-adaptable and change-averse railway culture, and so even the suggestion that something else for passengers might be used to take space for operational convenience (really operational convenience designed for a railway decades in the past) seems absurd. While we have “passenger” trains in Canada, the fundamental attitude is not passenger-centric — this is just human freight.

This is all so frustrating, because to back up the huge value these stations would provide in terms of coverage, capacity, resiliency, encouraging slightly less centralization, encouraging more TOD, improving speed by spreading boarding and alighting load and the like, they should also be trivially-easy to build. Sure they are a lot more work for GO than some beetfield station in the greenbelt, but compared to the caverns we are currently digging out under skyscrapers, this is by far the easiest conceivable way to add more rapid transit to the core of the city — by moving some tracks around, filling in an underpass, and building some above-ground structures.

I get that Toronto is not exactly a curious place, particularly when it comes to transportation infrastructure, but surely at some point when we reckon with the fact that 5 billion-per-kilometre for more downtown subway trackage (if we stay on our current trajectory!) is not tenable. At this point, maybe someone will have listened to that weird dude from Vancouver who made sure to put this idea on the shelf to build new downtown rapid transit infrastructure, directly adjacent to tons of density and destinations, and crucially above-ground, and get us started on something we should have built 20 years ago (and would have if we were Australia).