London Metro Rail

For London, one Crossrail isn’t enough

» There are another four years to go before Crossrail 1 opens, but consultation is advancing quickly on Crossrail 2. London is ready for more fast cross-town links.

As Paris begins construction on a massive new program of circumferential metro lines designed to serve inter-suburban travel, London has doubled down on its efforts to improve links within the center of the metropolitan area. The two approaches speak to the two regions’ perceived deficiencies: Paris with its inadequate transit system in the suburbs, London with a core that is difficult to traverse.

There’s one thing both cities deem essential, though: Much faster transit links to reduce travel times around each respective region. In London, that means growing support for additional new tunneled rail links designed to bring suburban commuters through the center city while speeding urban travelers.

Since the conclusion of the second World War, London’s Underground network has grown very slowly: The Victoria Line was added in 1968 and the Jubilee Line extended in 1979, but that’s about it. In some ways, that made sense: London region’s population peaked in 1951 at 8.1 million and declined precipitously until the 1980s. It only recouped it losses in 2011. But the region is now growing quickly, adding an estimated 100,000 or more people a year, reaching a projected 9.7 million 20 years from now. The number of commuters entering the city is expected to grow by 36% by 2031.

That growth has put incredible strain on the city’s transit network, with ridership growing by 40% in fifteen years. Through direct government grants, the support of the pseudo-public Network Rail, and the commitment of Transport for London, the local transit organizing body, the city has two major relief valves under construction. The Thameslink Programme, which will open for service in 2018, will improve the existing north-south rail link through the city by allowing for trains every two to three minutes; the Crossrail 1 project, also opening in 2018, will create a new, 21-km northwest-to-southeast subway corridor that is expected to increase overall transit capacity by 10% while significantly reducing east-west travel across the city center.

Those projects, which cost more than £21 billion ($36 billion) between them, will allow the system to accommodate new growth, but they won’t resolve London’s most significant transit bottleneck, the Victoria Line, which carries far more riders per mile than any other Underground Line. That’s where Crossrail 2 comes in.

Crossrail 2, as the following map shows, would extend from the southwest to the northeast of the city, connecting Victoria with Euston, St. Pancras, and King’s Cross Stations, roughly paralleling the alignment of the Victoria Line. The project will allow certain trains on the West Anglia Main Line to the north and the South Western Main Line to run through the city. The project was submitted to a public consultation process that ended last week that examined several options for line routings; a preferred route is expected to be selected this year, with construction beginning at the earliest in 2020 at a cost of £12 to 20 billion ($21 to 34 billion). Last year, a separate consultation for the route selected a “regional” option (allowing through-running commuter trains) over a “metro” option, which would have been an automated subway.

Like Crossrail 1, Crossrail 2 is expected to increase the transit capacity of central London by 10%, possible thanks to 10-car trains running every two minutes, allowing 45,000 passengers per hour per direction. As the following map illustrates, that capacity increase will be needed by the early 2030s if the project is not implemented. Major sections of the Victoria, Piccadilly, Northern, and District Lines are all expected to be crowded at more than four passengers per square meter at rush hour, enough to make much of London Underground a truly inhospitable environment.

The opening of the the high-speed rail line HS2, which will link London to Birmingham by 2026, makes the capacity bump provided by Crossrail 2 even more important because of the influx of passengers expected at HS2’s terminal, Euston Station.

The result of the new connection will not only produce less crowding on other lines, but it will significantly reduce journey times. To Tottenham Court Road, where Crossrail 1 will will meet Crossrail 2, the latter project will reduce travel times from Kingston in the southwest from 49 to 27 minutes and from Tottenham Hale in the northeast from 27 to 16 minutes.

There is little about Crossrail 2 that has been easy thus far, and certainly there is plenty more work to be done, particularly in assembling the project’s financing. The project has been studied since the 1970s (as the “Chelsea-Hackney Line”) and was considered as a serious alternative to the initial Crossrail project in the late 2000s. In other words, its necessity isn’t exactly a new idea.

Extensive support from business groups, including London First, however, is new. The organization has proposed funding the project, in part, with £3 billion in fare increases on all transit services, £2.4 billion in revenue from allowing denser development along the corridor, and £1.8 billion from expanded business taxes. In addition, the line — like Crossrail 1 — is expected to be operationally profitable and therefore able to raise some its capital funding by bonding on the back of future fares to the tune of an additional £3 billion.

If these seem like huge sums, they are. But London transit proponents have successfully been able to make the case not only that the city’s residents rely on its transit system, but also that investing in a better transit system produces overwhelmingly positive benefits to the economy as a whole. Crossrail 2’s advocates note that, even with a £16 billion price, the project’s benefits to cost ratio is 4.1 to 1 when wider economic benefits, such as agglomeration, are considered. This is a message that American transit promoters, who are unable to effectively make the argument for new lines, should practice making, because while London’s a great town, there’s nothing particular about the benefits of fast transit there versus anywhere else.

Image at top: Crossrail station at Canary Wharf, almost complete, from Flickr user George Rex (cc); Crossrail 2 map from Transport for London; Crowding map from London First.

26 replies on “For London, one Crossrail isn’t enough”

Just one thing about the cost: it’s of course very high, but it’s also a long tunnel. The sum of straight line distances between the underground stations is 29 km, choosing the shortest option. The longer options make it about 34 km. Now, in PPP dollars, £12-20 billion is $18-30 billion. Being slightly pedantic and deflating to 2010 dollars makes it $17-29 billion (the cost figure is in 2012 pounds). Put another way, at the low end, Crossrail 2 is not much more expensive than several lines built around the world, including the RER A, which in today’s money cost about $450 million per km, in an era when real GDP per capita was half as high as it is today. Even at the high end, it’s only about as expensive as Crossrail 1 and a bit more than RER A scaled up to GDP per capita, still cheaper than easier projects built in New York.

Now, the benefits. As we see with Second Avenue Subway Phase 1, US cities do build things, even at extreme costs, when the benefits are commensurate. Not knowing London very well, I still think Crossrail 1 and 2 look like they can have very large benefits, each comparable in importance to the RER A. Like the RER and Grand Paris Express, they’re capable of sustaining high speeds through city center, effectively turning city center transparent to regional rail traffic. Ridership levels comparable to those of the RER A, a million riders per weekday, are plausible for each. The cost per rider would still be high, although that’s partly a function of the need to construct longer tunnels than for the RER A.

I don’t understand why such long tunnels are needed. Commuter rail lines currently terminate at core stations such as St Pancras and Waterloo. Why not just build a much shorter tunnel between such stations? Similarly, why is the RER E building an entirely new tunnel from Haussmann-Saint Lazare to La Defense, rather than just reusing the existing commuter line that connects the two?

Because the capacity crunch is on the approaches to those stations. Going all the way to Tottenham Hale avoids interaction with the Chingford line and keeps trains out of the way of Stanstead-bound fast trains for longer. On the other end, going all the way south to Wimbledon is actually not far enough, because the fast lines are already hitting capacity, and there’s talk of a tunnel for the fast lines running from Surbiton to past Wimbledon where the slows get out of the way and into the Crossrail 2 tunnel. The latter project is likely to be relatively cheap, though, since it doesn’t require any stations, just tunnels.

From Wimbleton to Victoria, there are a minimum of 4 tracks with what appears to be more space in the ROW for more. The northern approach is somewhat more constrained, but still a great deal of the tunneling seems to be unnecessary in order to add 2 more tracks. And given the time wasted by to completely loading/unloading in a train terminal, it is possible that capacity could be significantly increased without adding any tracks, if trains were to through-run in a tunnel between, say, Victoria and St Pancras. This system works great with the RER A/B/D, why shouldn’t it work in London?

Because the constraint is not at the Waterloo terminal. The constraint is partly that the four SWML tracks between Surbiton and Clapham Junction are full, between the inner-suburban slow trains, the outer-suburban semi-fasts and the fast trains from much further out. The same applies on the WAML: the terminal is actually not going to be as huge of a problem once Crossrail 1 takes a bunch of trains out of Liverpool St, but the mix of fast and slow trains is a bottleneck on the WAML.
The point of Crossrail 2 (and 1 for that matter) is really to relieve the Tube lines running parallel: the problem at Waterloo starts once the passengers get off and try to all cram onto the already-packed Tube lines.

But at least from aerial imagery, it appear much easier to add two surface tracks in parallel to the existing tracks, rather than build brand new tunnels.

Of course the main point is to relieve Tube lines, but all that matters for that is the station locations. The manner of construction between stations should be done in a cheaper above-ground way wherever possible.

Also, remember that the existing south-side terminals are all elevated. But the Crossrail tunnel has to cross below the Thames.

Once you figure in the tunnel approaches, you’re already quite some distance from the terminals, and even from the wide terminal approaches… getting into the congested main lines which anonyrat describes.

The northern route seems honestly a little odd, as I’d think they’d try to pick up the “spare branches” of the Lea Valley Lines rather than the mainline, but whatever.

The existing Victoria terminal is already depressed not elevated.

Wimbledon-Victoria is about 7 miles, and St Pancras-Tottenham Hale is about 6 miles, and there is no way the distance need to go from tunnel to elevated is a significant fraction of those distances.

Victoria rail station platforms and concourse are at grade. The track approaches are depressed to an extent, though end up being elevated by the time you reach the river (due to the natural sloping away of the terrain down to the embankment), bridging over the A3212 and the Thames.

In any event, there are no tracks connecting the stations chosen: they can only be connected by using new routes. The goal behind the Crossrail routes isn’t just to allow through-running of trains across London but to increase interconnectivity. Wimbledon–Tooting Broadway links the southwest rail network and District line with the Northern line. And Tooting Broadway–Clapham Junction links the Northern line with the country’s busiest rail station, including Overground links to west London. The Chelsea options give increased service to an underserved part of the city.

So this plan gives new connections to the Overground at Dalston Junction/Hackney Central, and the Underground at Tooting Broadway. Every other connection would be the same under my plan.

But the Overground connection could be accomplished much more cheaply with a new transfer station north of Kings Cross. And a Tooting Broadway connection would be nice, but how many billions is it worth? Maybe better to build an Underground branch from Tooting Broadway to Wimbledon.

My uninformed guess is that Paris wants to connect the RER E to suburban lines that go due west of Paris. The line from Sainte-Lazare to La Defense only goes south of La Defense, toward Versailles, which is already served by the under-capacity RER C and the trains to Montparnasse. (There at least used to be long-term plans for an RER F line connecting Montparnasse with Sainte-Lazare.) In contrast, the plans to extend the RER E involve taking over an RER A branch and extending it further. At least in my experience the RER A is not at capacity, but it’s a lot closer to it than the other RER lines.

I do not know the situation in London. I suspect the answer is capacity, as Tom notes. In London, unlike in the US, commuter rail already runs frequently and is fare-integrated with the Underground, so there’s no way to relieve the Underground just by running better commuter service. (And honestly? Even in New York, the worst capacity bottlenecks on the subway can’t be fixed by commuter rail optimization, even though some of the most crowded lines – Lexington, Queens Boulevard, Fulton Street – closely parallel commuter rail. That’s why the city is building Second Avenue Subway to relieve the Lex, and why I think it should build a branch along Northern to feed 63rd Street and relieve QB.)

In that case, the RER E could connect to the RER A branch in La Defense (possibly in addition to the southbound branches). Seems like an easy way to avoid a lot of expensive tunneling.

“Commuter rail lines currently terminate at core stations such as St Pancras and Waterloo. Why not just build a much shorter tunnel between such stations?”
Because that wouldn’t add capacity.
The fundamental im to increase the number of people who can get into London, and hence why a new line is needed over such a long distance.
That said, some of its ridership will come from existing riders using this line instead their current line + the Victoria line.

Linking two termini by tunnels and thus allowing through services has been proven time and time again to add significant capacity.

Would it be possible to tunnel »between such [terminal] stations«? Tunnels need length to descend below ground and space to assemble the TBM and start it; I don’t think you could fit all this into Central London, let alone fit two ramps between two terminals.

I think Kings Cross–St Pancras might be a better cognate for Chatelet–Les Halles. It’s already sprawling with 5 tube lines and Thameslink rail (in addition to the terminating mainline and international services). The addition of Crossrail 2 (likely providing a direct pedestrian link to Euston) would cement that.

It’s worth noting that, although not branded as such, Thameslink will be fulfilling a similar role to Crossrail: when the full Thameslink Programme is complete in 2018, there’ll be 24tph through central London consisting of mostly 12-car train sets, stopping at Kings Cross–St Pancras, Farringdon (transfer to Crossrail 1), City, Blackfriars and London Bridge.

Good points. Thameslink was Crossrail 0 in a way. It’s all subjective, but Kings Cross-St Pancras seems more like Gare du Nord, if we’re continuing the Paris analogy. It is very interesting how similar (and effective) the systems will be. Even the London Overground and Docklands Light Rail are kind of analogous to Paris’ new outer Metro lines and all the tram lines they are building at the edges of their metro.

I really think the original Chelsea-Hackney line route would be better as another LU metro line instead of a crossrail heavy metro. There should be all the original inner-city interchanges finally connecting the LU metro and it should serve the city of London.

For the next heavy rapid-transit link; I would suggest re-boring the Waterloo & city line shuttle tunnel, and then linking it with the Northern and city line at Moorgate to give some of those services terminating at Waterloo direct access to the city and to link with services to Welwyn garden city, etc.
Then there could be investigations into another tunnel linking Waterloo with Euston. Or maybe Liverpool Street with Cannon St…

I’m in total agreement with your suggestion to rebore the Waterloo and City tunnel in order to replace the current LU service with a Crossrail-like service. As part of this work, I would suggest a secondary tunnel from Waterloo East into the reconstructed tunnel to connect the South Eastern Main Line to the South Western Main Line, via the old Waterloo Eurostar station. This would give commuters from Kent access to train services west of Waterloo without having to transfer to the Underground.

That would be a neat trick, given that Waterloo East’s platforms are about 20 feet above the ground, and to connect to the former Eurostar lines would involve a near 90-degree turn, ploughing straight through the main concourse of Waterloo station.

Any re-use of the Waterloo & City line tunnel is likewise completely impractical – it’s perpendicular to and some 40 or 50 feet below the terminating mainline tracks. And it departs north-west from Waterloo station and then does a sharp bend to the north-east under the Thames. There’s some useful background info in Wikipedia’s article on Waterloo & City.

The Waterloo and City line tunnel doesn’t do any sharp bends at all.

It begins close to the old Eurostar terminal and then does a turn not very sharp close to bank.
There have been proposal and from what I understand even actual plans to connect it to the Northern and city line with its larger loading gauge (as a tube line) for decades. So why not make it a crossrail service?

There’s a sharp curve almost immediately it exits Waterloo station. From the Wikipedia article I referenced:

The route starts from a point half-way between Lower Marsh and Aubyn Street; leaving towards the north west, the line turns in a 339-foot (103 m) curve towards the north east. The curve is constructed by cut-and-cover, and the twin tubes start immediately after it, under Stamford Street, turning north-north-east to pass under the River Thames, converging with Blackfriars Bridge on the north bank. The line turns east there, under Queen Victoria Street, to the station adjacent to the Mansion House, running for part of the way under the District line. The sharpest curves other than those at Waterloo are 603-foot (184 m) radius.

Plans to connect various lines have occurred with some frequency for over a century, including the one you suggest. However, the morass of lines around Bank seems to it make impractical to extend it further north (and at its current depth, the W&C line would intersect the Central line tunnels).

At the southern end, you might be able to straighten the curve so that you could align the tunnel with the old Eurostar tracks, though there’s the small matter of bringing it up to the same height as the elevated mainline tracks – not sure if you’d have to distance to do that without blocking a lot of busy streets (or starting the tunnel miles out from Waterloo station).

There’s plenty of discussion on extending the W&C line at one or both ends on London Reconnections, but the upshot seems to be that it would be cheaper to build an entirely new line.

If they’re going to build an RER-style “heavy metro” line as I propose like they are with Crossrail and with what they’re upgrading Thameslink too; the Waterloo and city line stations would need to be completely dug-out and enlarged. the would be no reason to continue with the current alignments at the tunnel’s two ends, and the existing station could become part of greater station space.

I’ll have to take your word for it that the Waterloo and city line would intersect the Circle line at bank. But I still think that with the tunnel-route re-bored, new platforms at a new depth could be rebored within that ~90 meters underground. Especially as I understand that the Northern & City line terminus at Moorgate is rather shallow.

Likewise; a new station at Waterloo would need to be cut-out, and would need to be very large, which would use the space of the now defunct Eurostar Terminus. I appreciate that the station would be ~100m under the surface, and that would mean some distance before surfacing to the mainline.

I appreciate that it would be expensive, but I think it would be worth it. Because so many commuters whose suburban trains terminate at Waterloo then transfer to the Waterloo & city shuttle to get to the city and many then walk or bus some distance north to Moorgate & the surrounding areas in the tower hamlets. And conversely it would give commuters on the northern & city line the extra stop to bank station.
Then there’s the other probably not-as-large benefits of allowing south-west Londoners to get to north London & beyond and vice-versa. And what’s the price tag on using the Chelsea-hackney alignment for a heavy metro? I really can’t imagine it being cheaper. The Waterloo and city line alignment is ideal for the next Crossrail to link two termini, and the leaky old tunnel can justify being replaced with something more modern. The Chelsea-Hackney line alignment seems more ideal as its original purpose as a metro line across boroughs of London, finally bringing a real tube to Chelsea and Hackney and providing the missing pice linking tube stations in central london. I really don’t think nearly as many of the commuters from South West London would want to go to Tottenham Court Road via vChelsea and the Victoria terminus as would want to go to Bank station in the city and Moorgate station in Tower Hamlets.

Leave a Reply