high-speed rail versus highways

Now that the latest Ontario High-Speed Rail proposal has come out, we are getting the usual Serious Journalism about rail.  Which means some combination of intentional and unintentional concern trolling about rail, particularly around costs (call it “concern costing”).  A particular attribute of this kind of analysis is it always looks at rail in isolation as a standalone service, as if decades of subsidizing driving didn’t exist.

Here’s what a better comparison of Ontario high-speed rail versus highways that included more context might look like:

Measurement High-Speed Rail Highways
Est. deaths over 10 years 0 3000
Est. injuries over 10 years ~0 30,000
Est. expenditure over 10 years $20 billion? $26 billion
Est. direct revenue over 10 years ? $0

So just to be clear, spending $26 billion on roads to get zero revenue and kill 3000 people is a no brainer, but spending maybe $20 billion on rail to get some unknown millions in revenue and kill zero people needs Very Serious Analysis.

Estimated deaths over 10 years is from Ontario road deaths: Did drivers do better or worse in 2015? It applies to highways under OPP jurisdiction. Estimated injuries is just a direct 10x scaling of deaths.

Estimated highway expenditure is from Ontario budget 2017, at the bottom of Investing in Highway Infrastructure.

No one really knows yet how much the train will cost, or how much revenue it will bring in.

I’m not saying we shouldn’t examine the viability of High-Speed Rail. I’m just saying we should place it in context with our expenditures on highways. If you’re not asking the exact same questions about highway “viability”, then it’s not a level playing field.  If you’re not talking about the fact that you’re introducing HSR into an environment where driving has been massively subsidised on multiple fronts, including cost of gas, cost of highways and cost of parking, then it’s not a level playing field.

France iDTGV train and iDCAB taxi

France has not one but three high-speed train companies that share the same rails.  There’s the traditional TGV which you can book about 3 months in advance, the low-cost iDTGV which you can book up to 6 months in advance (but only runs on certain routes), and Ouigo, which you can book up to 9 months in advance but which departs from outside Paris and runs only along a single route.

(There’s also TGVpop, where it takes a certain number of votes before the train is confirmed to run and can be reserved.  Voting starts 2 weeks before the train would depart.)

Booking iDTGV

UPDATE 2017-03-17: SNCF has announced that it will discontinue iDTGV as a separate brand with separate booking.  See e.g. iDTGV brand to disappear.  ENDUPDATE

iDTGV can be booked from its own website or app, which has the advantage of offering seat selection, or through voyages-sncf, which I gather doesn’t.

It has two types of “zone”, basically a quiet one (iDZEN) and a family/louder one (iDZAP).  See http://www.idtgv.com/en/idservices/idzen-and-idzap  It also has two classes, but in my experience it’s not worth getting 1st class on French trains.

It sort-of works for booking outside of France (including a good English interface).  There are a number of issues though (and some outright bugs / website errors).

To make it work without a European credit card and French phone number:

  1. When creating an account or checking out, do not try to enter a cellphone number unless you have a French cellphone.  The form only accepts French numbers (of form e.g. 06 xx xx xx xx).  The cellphone number is optional, except…
  2. When going through the booking process, on the options page, do not select the iDCAB taxi option.  It will make the entry of a cellphone number at checkout mandatory (this is a bug in the system).
  3. If you have a non-European credit card, your first checkout will almost certainly fail with “rejected”.

iDTGV support says « les cartes étrangères, non Européenne sont bloquées par défaut sur notre site par mesure de sécurité » which translates roughly as “non-European cards are blocked by default on our site as a security measure”.

However, having done the initial transaction and gotten your card rejected, you can email or contact form them and ask them to unblock it.  Just send them the last four digits of your card, not your entire number.  Thanks to Seat 61 for the info about credit cards.

If that doesn’t work, try also calling your credit card company to see if they have blocked the card on their end too.  Sometimes transactions on European websites trigger North American credit card company blocks.

I had to both email iDTGV to get the card unblocked at the website end as well as phone my credit card company to get the card unblocked on the card end.  I had the credit card support stay on the line while I ran the transaction through, a practice I highly recommend as it saves you calling back if it’s still not working.

Be aware that the iDTGV website will silently time-out if you’re idle for a while.

Other bugs:

  • In the Details of My Order section of Print My Tickets (the summary page), the links under “You Can Still” mostly don’t work (they just point to the summary page).  Use the links under My Booking (upper left) instead.
  • If you’re booking in English in the iDTGV app, check the dates very carefully.  it looks to me like there is an off-by-one error (e.g. you have to enter January 1 in order to get January 2), probably due to the fact that the French calendar week starts on Monday and the US/Canadian/UK calendar starts on Sunday.

UPDATE 2016-04-11: It looks like iDTGV has joined a number of other apps in now only being available in the French iOS Apps Store, rather than being available e.g. in the Canadian store as it was previously.  As the easiest way to display your tickets is in the iDTGV app, this makes it less convenient.  If you have already downloaded the iDTGV app, it will continue to work.  ENDUPDATE

Booking iDCAB

iDCAB is a taxi-like service for travel to and from the train station.  It’s available for quite a few of the major stations in France.
iDCAB 10 Euros
You can add iDCAB to your iDTGV booking (on the website) without any problems AFTER you’ve booked and paid for your main train trip.  However (if you have a one-way ticket at least) it appears to only offer the option for a taxi on the departure end of the trip, not the arrival.

In the iDCAB interface you can enter an international number (e.g. +1 xxx xxx xxxx) although it’s anyone’s guess whether it actually gets recorded correctly in the system.

iDCAB is not currently available as an option in the iPhone app.

You can also book iDCAB as a stand-alone service at http://idcab.sncf.com/ (thanks to Jérémie Croyère, CTO of iDCAB for this info – @cpasbanal on Twitter).

On the SNCF site you’re supposed to be able to submit your train booking reference to get it to autoload your train stations, but I couldn’t get it to work with an iDTGV booking reference.

Unlike the interface at iDTGV, the SNCF interface won’t give you options for how early you want to arrive or estimate travel time to the station; you have to choose your pickup time yourself.  It also won’t let you specify the number of passengers (however this doesn’t really matter as it’s a flat price regardless, up to max 4 passengers).  The price on SNCF is also higher than on iDTGV.  The SNCF interface will tell you what car service is picking you up though.

Baggage Restrictions

Note that iDTGV has tighter baggage restrictions than the TGV:

You can take 2 pieces of luggage for free (details below).

If you take more than two pieces of luggage, you will have to pay an extra charge of €35 per piece of luggage (on the iDTGV website in “My travel options”), within the limit of 2 extra pieces of luggage per person.

If you did not pay this extra charge when you booked, the payment will cost you €45 on board.

The free pieces of luggage per person cannot exceed:
– two pieces of hand luggage (suitcases, rucksacks, travel bags) per traveller; or
– one hand luggage and an object per traveller (a children’s pushchair, a wheelchair, a bicycle with its wheels removed and placed together in a special protective cover of 1.20 x 0.90m maximum, a surfboard placed in a protective cover of 1.20m x 0.90m maximum, a pair of skis, a monoski or a snowboard, a bag containing a ‘small-sized’ domestic animal; or
– a piece of hand luggage per traveller and a piece of luggage of 50cm x 50cm x 50cm ; or
– a piece of hand luggage per traveller and a musical instrument.

For more details, check out the full page at Seat 61: www.seat61.com/idtgv.htm

Primary and Secondary confusion about Ottawa LRT

2013-10-16: Found this in my drafts folder from 2012-06-29.  Published.

There are many elements of confusion about the Ottawa LRT project.  There has been poor communication and poor choice of language.  There are also legacy elements from the 2008 Transportation Master Plan.

First, the purpose of the system.  They are not just building “east-west rail”.  They are building a commuter rail system that will get you from Kanata (Terry Fox Station) to Trim Station (and/or Millennium Station).

That’s roughly 40km end-to-end.  That is a long stretch.  If you’re going to run commuter rail to get someone 40km every morning and evening, it must be

  • reliable service (always arrives when scheduled)
  • regular service (service departs frequently, ideally every few minutes at peak times)
  • fast service (the journey must take an hour or less, generally speaking, for commuters to tolerate it)

The service should also be high enough capacity to handle the current and anticipated load over its operating lifetime.  And it should be affordable.

There are various engineering ways to ensure you can achieve those goals.  In a modern commuter rail system, you ensure this type of service by

  • having continuous grade separation – this is a fancy way of saying “cars, pedestrians and cyclists never cross the tracks” – this means at all intersections the other traffic must either tunnel under or skyway over the transit corridor
  • having infrequent, widely-separated stops – the transit needs to have a high average speed, which means it needs to stop (with all the slowdown, boarding time, and startup from zero again that involves) as few times as possible
  • long buses or long train cars – this not only means high capacity for each transit unit but it importantly it also means many fewer passengers per driver (salaries are expensive)

Part of the problem is that, while the O’Brien east-west plan with a tunnel (to provide grade separation) downtown implies this service, it is a different service from what we have now, and while staff clearly understands this is what council has directed them to specify and procure, it appears that the councillors themselves and the vast majority of citizens don’t realise this is what we all agreed we’re building.  (The old north-south proposal was a mixed system that was only partially commuter rail.)

To make things worse council and staff use very confusing language to refer to this type of system.  Rather than calling it commuter rail or rapid commuter rail, they call it either LRT (“light rapid transit” or “light rail transit”) or PRIMARY.

As you can imagine, this causes confusion because people think primary means “the main route”.  So when staff says “route X should be the primary corridor” people think they mean “the main transit corridor”, they don’t realise it means “the commuter rail corridor”.

Now, there is a completely different goal that you can set.  You could set the goal of moving people within the city, and generating economic development continuously along the route.  If you want to do that, again there is a set of requirements:

  • convenient – it can’t be too far to a stop
  • integrated – the transit should be part of the streetscape, easily reachable from the sidewalk
  • moderate speed – the transit should be going at a speed that is safe and pleasant when experienced both by bystanders and those potentially crossing its route

This is the classic “clang, clang, clang went the trolley” type of street-level service most people are somewhat familiar with.  To meet these goals you:

  • run the transit on the street, with perhaps priority over other types of traffic but nevertheless sometimes intersecting with other kinds of traffic, including e.g. rails that people can just walk across
  • have stops close together, sometimes just a few blocks apart
  • (on the street and stops close together pretty much constrains the maximum speed)

This kind of service is what the city staff calls SECONDARY or SUPPLEMENTARY.  They should call it local service, or name it by a particular instance of the service (call it a tram).  Unfortunately the use of “secondary” makes people think it is less important, or lower priority, even though it is the service that people living in the downtown core would use the most.  They really just mean “slower local service that isn’t commuter service”.

(It’s not actually this simple as e.g. the Western LRT investigation calls it Primary and Supplementary, referring within Primary to LRT and BRT, and within Supplementary to “intensive bus” and “intensive rail” transit.)

To make things extra confusing, in Ottawa we have both these types of service, but we mix them together under the general heading of “bus” or “OC Transpo”.  And we mix the two types of service together in a single route (which you should never do, because as shown above, they have completely different requirements).

The closest to commuter service we have is the Transitway, our pseudo-BRT (Bus Rapid Transit) system.  (Pseudo because we don’t do pay-before-boarding, which slows the system down considerably.)  The dedicated parts of the Transitway are classic commuter design: grade separated, widely-spaced stops.

OC Transpo provides a map that is a representation of our current Transitway system.

The big problem is that as soon as the Transitway hits downtown, it turns into a local service with sidewalk-accessible frequent stops.  This completely screws up the end-to-end transit as you’ve inserted a slow, disruption-prone segement into the middle of a fast disruption-free route.  And that’s why as long as you’re building commuter rail, you have to do a tunnel.

We also have a local service in the non-Transitway buses.  People tend to mix them together as it is never clearly stated that there is a local system and a commuter system, both systems use the same buses and they interconnect.

The Parkway

The Parkway would be terrible for a secondary (local, tram) service.  There’s no one there.  There’s no shops by a sidewalk that people are going to go to.  There’s no existing corridor that will be uplifted by having frequent local transit.

The Parkway is ideal for a primary (commuter) service.  First of all, the commuter service ALREADY runs on the Parkway (the Transitway buses).  Replacing those buses with rail would mean less pollution, less noise, a lower number of transit “containers” and a predictable service (both in terms of it being fixed to rails and in terms of it keeping to its schedule).  A predictable service means you can cross the path safely (unlike the current life-in-your-hands situation now where it’s basically a highway).  A rail service means people can look out the big windows and actually see the river and the green.

The idea that single occupancy vehicles (eyes on the road!) and commuter buses is a better combination for highlighting and enjoying the Parkway than packed train cars gliding along is just bizarre.  The idea that a predictable train on fixed rails driven by a professional driver is more dangerous than thousands of amateur drivers careening around in cars is ridiculous (which is to say, if we don’t require a fence to protect us from the cars that kill people ALL THE TIME then we shouldn’t need a fence to protect us from the train that kills people approximately never).

That being said, the city has to completely change how the NCC thinks about the Parkway (it appears to imagine it is some 1950s Sunday drive scenic leisure route).  They need to understand how it is actually used (as a busy and dangerous car commuter highway) and appreciate how it could be better used (as a modern rail corridor).  Both data and storytelling will be required.  A Horizon 2067 that imagines the Parkway is best used in 55 years by cars and buses would be just bizarre.

The Parkwark (ORP) will never be used for the LRT.  NCC is completely opposed to it.


The Carling story is the inverse of the Parkway.  Carling would be terrible for primary (commuter, rapid rail) transit.  Either you do as you’re supposed to and grade separate the rail you put on Carling (at enormous expense) or you intersect the train with traffic, which introduces a fatal unreliability, unpredictability, and great danger into the middle of your route (not to mention which a wall of trains that will completely block your north-south traffic flow).

Carling is fantastic for secondary (local, tram) service.  You can have lots of stops.  People can start using the sidewalk and shopping along the whole length of the street.  Housing and shopping will be attracted to the route.  The economic development that Holmes wants is a strong possibility with a local route on Carling.

What about Route X?

If there is an existing local neighbourhood, running commuter rail through it is going to suck.  You can’t make it unsuck.  You can maybe make it less visible (e.g. using cut and cover) but you can’t make it fit well within an existing urban fabric.  By design it must be fast and separate.  If you put in stops that will just slow it down for everyone, end-to-end.  There is no way that you are going to be able to do this without making some local residents unhappy.  That’s the political reality.  That’s a job for the politicians to sort out.  What you must not do is try to turn it into a local transit segment, then you’re just reinventing the problems of our current Transitway (which slows almost to a standstill in peak hours when it hits the downtown core) in a new location.

Transportation in General

I already walked this out in a previous post.  First, it’s important to understand that TRANSIT is about moving people.  Not stuff.  People.  People are small.  The issue with cars is they make sense if you are moving a lot of stuff (which people almost always aren’t) but are nonsensical if you are moving a single person (which they almost always are).  Cars and small trucks are fine if you need to move stuff.  They make no sense for moving people in a dense environment.

In a dense environment within a city you:

  • walk if it’s a short distance
  • cycle if it’s a medium distance
  • tram (or bus) if it’s a long distance
  • subway if it’s a long distance and there are a lot of people to move

That’s it.  There are no other solutions for moving people around a city.

If you need to get from the city to a nearby village (or more usually in from the village to the city to work) you take commuter rail.

If you need to get from city to city you take high-speed rail.

That’s how you move people around.  The only time you need cars is if you need to move around places that aren’t served by rail.  In the North America we used to have (and in the Europe that still exists) all villages are served by rail.  The only places you need to use a car are to get around in rural areas.

High Speed Rail

The Government of Canada’s own study (or I should say, the latest of many studies) says that the Toronto-Ottawa-Montreal corridor can be done for $9 billion dollars and that the economic benefits mean the corridor would pay for itself.

There is however zero chance of a conservative government building high-speed rail.

Union Station

In a perfect world, Union Station would be used for the passenger rail service (and that service would be high-speed to Montreal and Toronto).

This is never going to happen.

In a reasonable world, Union Station would at least be used as a commuter rail station, since the commuter rail goes right under it.  This would connect with our history, and open our miniature version of Penn Station up for tourism.

Union Station will never be a train station again.  (It is bizarre beyond belief that the debate was about whether the commuter rail station should be to the east or the west of an actual existing station, but that’s Ottawa for you.)

In fact, given that Union Station needs millions in repairs, the most likely situation is that it will have to be demolished (after a decade or two more of total neglect).

Barry Padolsky’s firm has completed a heritage conservation plan for the building, but I don’t know where or if you can find it online.


There is zero funding for local (street-level, slower, many stops) rail in Ottawa.  We are building a commuter rail system that will do the centre (Tunney’s Pasture Station to Blair Station), then west (Lincoln Fields to Tunney’s Pasture Station) then east (Lincoln Fields Station to Trim Station and/or Millenium Station) then south.  That will cost a hijillion dollars and at Ottawa’s pace will take another three to five decades.

If we were France we would be building local trams in addition to commuter rail.  In fact we could just copy our 1948 street car system (which to a large extent created many Ottawa neighbourhoods including the Glebe with the B line and Westboro with the S line).  If you want trams you will have to do a ton of work to get the city to budget and plan for them.  This is currently not on the budget radar at all.  The Transportation Master Plan consultation and the Horizon 2067 plan are both key places you would insert tram planning and funding, if that’s what you want to have.

la Commission Mobilité 21 – French rail plans report

Summary of June 27, 2013 report from la Commission Mobilité 21, with a focus on how it will affect Rouen

76actu – LNPN, contournement Est de Rouen : où en est-on ?

The two major elements are bringing high-speed TGV to Rouen and building a new train station, Rouen-Rive-Gauche/Saint-Sever.  (The current station is Rouen-Rive-Droite; the original Saint-Sever train station was destroyed in WWII.)

The project is Ligne nouvelle Paris-Normandie (LNPN).  In 2012 there was discussion, followed by an April 5, 2012 decision (PDF) from RFF, the French rail infrastructure operator.  Following that decision, the Mobilité 21 commission looked at things further and produced its 2013 report.

Basically some activity is going to be delayed.  The new station and the TGV to Rouen may take until 2030.  The report proposed various scenarios, but the final decision hasn’t been made yet.  The article linked above says « Tout le suspense repose sur les choix de scénario que vont retenir Parlement et État dans les mois à venir. »

Other lines are already underway and not affected.  For example, by 2017, LGV-Est will extend to Vendenheim, just outside Strasbourg.  Also there will be TGV to Bordeaux by 2018.  The sub-project is LGV Tours – Bordeaux.  There’s supposed to be an entire site on the encompassing project, http://www.lgvsudeuropeatlantique.org/ but it’s down.

UPDATE 2013-07-04: Bordeaux high-speed rail project site is back up.  Réseau Ferré de France: Ligne á Grande Vitesse, Sud Europe Atlantique, LGV Tours > Bordeaux.  (RFF: LGV Bordeaux)  « Il s’échelonnera sur 5 années pour permettre une mise en service en 2017. »  The line will run at 320km/h.  « Une vitesse commerciale de 320 km/h »  There is also a more detailed project progress site from the contractor: http://www.lgv-sea-tours-bordeaux.fr/ You can read a PDF summary (press release) from June 2013 in French and in English.

high-speed rail in Canada

High-speed rail has been extensively studied in Canada.  The latest study There was a joint Federal-Ontario-Quebec investigation in 2011.  The links below are all to the same 2011 “EcoTrain” report, just distributed from the websites of the three different partners.

UPDATE 2017-05-19: In typical web fashion, all the links from 2013 were broken and gone with no trail to follow to find where they might have been relocated.  I have added new links where possible.

Available from Ontario MTO: English, French

Update – Ontario MTO English (from Internet Archive)

Available from Transports Québec: English, French (site also includes 1995 study of the same corridor)

Update – Transports Québec: French

Update – Transports Québec: Quebec-Ontario High Speed Rail Project, Preliminary Routing Assessment and Costing Study, Final Report, March 1995 (PDF)

Sort-of-kinda available (requires you to request full report via email) from Transport Canada: English press release, English “Executive Summary”, French PR, French Exec Summary

Update – Transport Canada: English press release, English “Executive Summary”, French Exec Summary (from Internet Archive)

There is also a Library of Parliament summary HillNote Number 2012-06E “High-Speed Rail in Canada” (also available in French “Le train à grande vitesse au Canada“)

and Wikipedia has a page of course: High-speed rail in Canada

There have also been separate studies of other corridors, including Montreal-New York and Montreal-Boston (PDF, via Transports Québec)

We are not understudied in this area.  But no one wants to spend any money.

Toronto to Montreal

The bottom line of the 2011 study: “Developing the section between Montreal-Ottawa-Toronto could cost between $9.1 billion for 200 km/h and $11 billion for 300 km/h” and “as a whole, the Montréal-Toronto segment of the project would provide a positive economic impact“.

Nine billion is not a lot of money for HSR connecting Toronto to Montreal, particularly considering you get a net positive economic impact (you get more back overall in the economy than you spend to build the infrastructure).

For the two speed options, which in the study are E300+ (electric, 300+km/h) and F200+ the specific dollar figures are $869 million net benefit and $817 million respectively, and that’s counting as “losses” the reduction in revenue to airlines and airports (report page S-21).


And that is with modal shares I consider ridiculous. At 300km/h, the modal share for business travel in 2031 would be 17%? Seriously? In 2031, 72% of business people are going to choose to drive rather than go 300km/h in first class on a train? (from page 58 of the EcoTrain report)


This is in a world where today, Amtrak’s “high-speed” train (which only goes an average 120km/h, with a top speed of 240km/h) and its much slower regular train service together get seventy-five percent of the Washington-New York modal share.

In other words, even with ridiculously low modal shares and even inexplicably counting the diversion of traffic from airlines as a “loss”, Toronto-Montreal HSR still has a net positive benefit.

But no one will build it.  All it would take is some outreach to external funding sources (Chinese government? Richard Branson?) and some political will.  We have neither.

This post inspired in part by Tyler Brule in the Financial TimesMaple leaves on the line (April 26, 2013).  Brule manages to say things like “Have neither the government nor the private sector ever thought about the economic benefits?” without mentioning any of the Canadian political context or any of the multiple studies (including repeated studies of the Quebec-Windsor corridor).

UPDATE 2013-05-02: I left a comment on Mr. Brule’s article.  Here is the text in full:

Mr. Brule makes excellent points about high-speed rail (HSR): convenient for getting from city centre to city centre and more convenient than flying for short distances.  However his question “Have neither the government nor the private sector ever thought about the economic benefits?” is somewhat puzzling in the Canadian context.  The Quebec-Windsor corridor has been extensively studied for HSR, with the most recent report released in 2011.  A copy of the report is available at e.g. Transports Québec: http://www.mtq.gouv.qc.ca/portal/page/portal/entreprises_en/transport_ferroviaire/thv_quebec_windsor

Key findings: the Toronto-Ottawa-Montreal route (the specific route that would have provided HSR for Mr. Brule’s trip to Ottawa) would cost $9.1 billion for 200km/h service and $11 billion for 300km/h service, and in both cases the investment produces an overall net positive economic benefit.  Other routes such as Montreal to New York have also been analyzed (available on the Transports Québec site).

What is needed is not more study of the economic benefits, instead what is missing is political will and private sector interest.  Perhaps Richard Branson or other transportation entrepreneurs are needed, in order to invest in this well-documented opportunity to improve Canadian inter-city travel options.