High-Speed Rail

Is Direct Service the Defining Element of Rail System Success?

» Eurostar has 85% market share on direct trips to and from London; trips requiring transfer, however, have less than 5% share.

Slightly less than fifteen years ago, Eurostar began offering services under the English Channel between the United Kingdom and continental Europe. The line has been met with unequivocal success. Railway Gazette reports today that Eurostar has grabbed an 85% share of the air-rail market on trips between London and Paris, Bruxelles, and Lille. The speed-up of the service over the years, from 2h56 between the French and UK capitals in 1994 to 2h15 today, has played a large role in the corridor’s high ridership. Eurostar’s popularity, like that of domestic high-speed services in France, Spain, and Japan, demonstrates the great value of investing in fast train systems.

Yet the Gazette article also notes that on routes where customers are required to transfer — such as between London and Koln, a link that can only be made via a connection to a Thalys train at Bruxelles — Eurostar market share is reduced to a pitiful 4%. What gives? Are people so disinclined to transfer trains that they’ll switch to an air trip if they can’t take a direct train? Is through service the only way to ensure high ridership?

There are obvious reasons for customers preference for direct trips. Once you’re settled in on the train, you have a cafe at your disposition, power outlets, sometimes wireless internet. You have far more space than on the airplane and the ability to walk around at will. In other words, a train trip, as long as it’s reasonably quick, can be incredibly productive for business people and far less stressful a travel experience for everyone else. The Gazette piece points out that on other European rail services, trains have a 50% air-rail market share on trips of up to even four hours — when there are direct connections.

But on trips requiring transfers between trains, the advantages of rail travel diminish significantly, because much of the productivity possible on the train comes as a result of the comfort it provides, not as a consequence of its speed. Once customers have to lug their bags from one train to another and walk through often-crowded stations, train travel looses many of its advantages. In the air, direct flights are easier to come by, and even with a transfer, your bags are handled for you, out of sight.

In some ways, of course, Eurostar is an exception that exasperates the problem. Because the United Kingdom is outside of the European free exchange zone, travelers must pass through customs and stations must be outfitted with immigration facilities if they are to provide direct service into London. Taking a train between France and Germany, on the other hand, requires no such bureaucracy.

Nonetheless, the disadvantages of transfers apply to other routes as well. If we are to assume that direct service is necessary to convince people to take the train, through-routing of trains should be prioritized over more frequent service. For customers using a potential future Midwest high-speed rail network to get between Milwaukee and Detroit, for example, four trains a day directly from Milwaukee to Detroit might attract double the patronage of eight trains a day between Milwaukee and Chicago and eight trains a day from Chicago to Detroit — even though the latter option would probably save a lot of people time. Eliminating transfers is incredibly important.

Jarrett makes a strong argument at Human Transit that terminal stations provide a better customer experience than through-stations, but the evidence presented above may demonstrate that the appeal of direct routes should dictate that virtually every high-speed station be through-running. France’s plans for new interconnections around Paris seems appealing from the perspective of increasing the number of direct trains, and a track connection between Eurostar services and the United Kingdom’s planned High-Speed 2 seems all but obligatory if British planners want to encourage rail travel to mainland Europe from anywhere other than London.

37 replies on “Is Direct Service the Defining Element of Rail System Success?”

This is an important article. Years ago, based on petty flimsy research, the rule of thumb was that you loose 50% of patronage with a transfer. Yet there remains a widespread assumption that changing trains is no problem. Supposedly, the longer the trip, the less riders mind. I think we need a lot more serious research on this issue. There is a Phd thesis or two here! It’s so easy for planners and managers to make assumptions.

Interesting article. I would think a line linking London – Brussels – Koln – Frankfurt in a single trip would be very successful considering the huge numbers of people living in that part of Germany. Much all of it is built up to 200 mph +/- standards already and the rest of that route soon will be.

The UK is outside the European free exchange zone? I had no idea. I remember taking a flight from Amsterdam to London about 6 years ago and I don’t remember going through customs, but I probably just forgot. Do you go through the customs process in the train like when you cross the Canadian border or do they do it in the station?

I think you need to take care when extrapolating from this particular example as there are a number of factors that depress the numbers, which don’t for other routes solely in the UK or solely on the continent.

The first is marketing. Eurostar almost only ever advertises the destinations it directly serves, so the common perception is that you can use it to get to Paris or Brussels only (there are additional direct services – ski trains, Avignon, Disney – but these are seasonal and advertised separately). I’ve only rarely seen mention that it can be used as a gateway to elsewhere in Europe.

Another is the difficulty of booking. When I tried to book London–Bordeaux last spring, it was impossible on the Eurostar website (have just tried it again, and it seems that it now does work). At that time, there were the options of RailEurope (either the UK or French site – each had its advantages). In the end, somewhat bizarrely, ended up being the cheapest and easiest way to book.

The Man in Seat Sixty-One website has pages of advice for the various possibilities of booking rail trips to continental destinations – something your average traveller isn’t going to bother researching if they can just deal with a low-cost airline’s website. I’d think there’s also a bias to just dealing with one company, one booking, rather than having to deal with multiple companies – some of which are foreign.

I think if RailTeam had been more heavily marketing the possibilities of where you could go by rail from the UK (see the map here), and if Eurostar had a functioning booking tool for single non-direct journeys and if the perception was created that it was a single albeit multi-leg journey, the numbers would have been higher.

People here are used to transferring, after all, even if it’s not always preferable. (Though sometimes it is, where it gives faster journey times or a better choice of trip times – much the same as it is with choosing flights, though when I’ve chosen indirect flights it’s usually been because it was cheaper). The difference is that for a trip solely in the UK you would see a list of possibilities that included direct (if there were any) and indirect routings in the same list. With Eurostar’s website, until recently, you couldn’t.

It’d be interesting to see what the numbers are for some of the routes in Germany, where you might need to transfer when travelling between certain city pairs. Again, I think it’s partly a perceptual thing – if you book, say, Frankfurt to Munich as one trip, the fact that you need to change in Mannheim or Stuttgart is less of an issue than if you need to book the legs separately. Another factor is confidence in the connections – it’s one thing if you’re going cross platform and it’s a five-minute wait, another if you need to traverse Paris on the metro during rush hour – and confidence in the frequency of service. If a missed connection means a half hour wait, it’s a chance to grab a cup of coffee. If it means a few hours’ wait, or if there’s a financial penalty to not being on a specific train, that’s a disaster.

The point you mention about customs and immigration could be a non-issue, if all the border formalities are handled at the UK end. Once you are moving through the Kent countryside, you’re effectively in France. As anyone could board/disembark in Lille, any travel beyond that would require the assumption that you’ve already been allowed in France, so no further checks would be needed in any of the other Schengen countries.

It’ll be interesting to see how the numbers play out once Deutsche Bahn and other operators start competing on UK to Europe services (there’s an article in the Guardian). Will it be competition on fares rather than through-routing that increases the numbers? Whenever I’ve looked, you’ve been able to get low-cost flights for as little as half the price of the train, so unless you’re on expenses, it’s a no-brainer. I’d pay a bit extra to travel by train but not that much. If they can get the prices down, we could really see a boom.

All of which is to say that though people will in most cases prefer a transferless trip, Eurostar’s numbers can’t really be used as a case study for how travel in the US market would play out. You’d really need to look at the numbers for journeys that can be booked in one go through DB or SNCF to get a clearer picture.

This argument would also point to unbalanced transfers as being preferable to balanced transfers. If you have a four hour trip with a transfer, a one-hour and three-hour leg is preferable to two two-hour legs, and a 1/2hr leg and 2 1/2 hr leg better still.

Of course, the last would as likely have been counted as no transfer at all, but rather just one more way the passenger got to the origin HSRail station and from the destination HSRail station to their final destination.

Maybe we don’t need massive through-running. Much can be done to make it easy to make connections. It’s usually NOT easy now.

“…unless we … made things easier for people to connect and treated it as a network … reassurance and help on … making connections fast so you don’t wait as long. … The biggest challenge, … so difficult, is the lack of standardisation … using technology that will sell fares and provide information.”

My one experience in a Brussels train station almost 10 years ago would not encourage more ridership. I was younger and healthier then, so I wasn’t looking for escalators or elevators, but I don’t recall any. Signs were in French and Flemish, none in English (or German, iirc). I asked for a downtown map, to walk to nearby historic buildings. They did not have one, not free, not for sale, not on the wall by the door with an arrow You Are Here. “No” is a word that translates all too well.

In those regards, I’m sure that every U.S. train station is worse than any in Belgium.

My suggestion for every Union Station, Penn Station, and others in the US: Signs in several languages, giving simple directions to get around. Lots of escalators and elevators. Free baggage carts. Assistance for elderly and disabled riders. Checked baggage with transfers to the correct train. Posted maps of the stations themselves, and at the exits, maps of the neighborhood.

Is that so difficult? Yes. It seems nearly impossible for bureaucracies running transport stations to consider the user’s point of view.

Case in point, AirTrain at JFK Airport operated by the Port Authority of NY. All signs are in English only. Destination signs are flush with the walls, rather than extending out to be readable by people coming onto the platform. And a dozen more examples of design failure that I notice every time I visit JFK.

I offer the AirTrain because it’s new. I’m sure the Manhattan bus station is free of any signage in French or German or Japanese.

Augmenting John W’s comment, Germany’s Deutsche Bahn makes transfers very easy…they plan trains so that transfers will only be across a platform and usually require no more than a 5-10 minute wait. This reduces a lot of the intimidation and inconvenience of a transfer.

The question is, does Eurostar perform poorly on London-Köln because there’s no direct connection, or is there no direct connection because Eurostar could not perform successfully?

It’s not just poor marketing. Eurostar isn’t a very fast service, because of the 160 km/h restriction through the Channel Tunnel and the indirect route taken in France. Brussels is close enough to London that Eurostar can do it in 2:00, but to get to Köln, it’d take about 4:10. Due to the indirect route taken it could probably only be marginally competitive with air.

The Detroit-Milwaukee example actually argues against the premise. Given Lake Michigan the trains will stop in Chicago. Even in its current cramped form Chicago Union Station can be used for through trains such that a Detroit to Milwaukee train could berth directly across a platform from a St. Louis train allowing quick and convenient passenger exchange. As a counter example, most Amtrak Wash-Boston trains nearly empty out at New York as few riders actually travel through NYC. As to ‘station experience’ Washington Union Station as well as NYP are both terminal and through. .

I think this is largely influenced by advertising, searching, and pricing. When one books a trip with Deutsche Bahn, one goes to a single webpage and simply searches for the starting station and destination, Deutsche Bahn handles the rest. At least for domestic trips, all transfers are mapped out reliably, and pricing is comparable to a direct train. When traveling between European systems, as in the example of going between London and Koeln, there is no central website on which to search and book tickets, and purchasing from multiple carriers tends to be much more expensive. If one had to switch airlines when booking multi-segment flights, including searching and paying for two separate transactions, with inadequate coordination between system schedules, multi-segment flights would be very unpopular indeed. Co-branding, harmonization of train schedules (including guaranteed transfers in which one train waits for another in the case of a short delay), and a more competitive pricing scheme would work wonders in making the London Koeln rail trip competitive with air travel. Some of this is taking place, e.g. Deutsche Bahn and SNCF’s coordination of their Stuttgart-Paris and Frankfurt-Paris routes, but I would like to see a schedule and joint pricing arrangement for all routes on all major rail networks in Western Europe.

I don’t think there are two types of stations. There are different types of platforms. Zürich Hauptbahnhof has a very nice “end station” upper level but then some trains run at a lower level and go through the station. There’s no reason (besides the billions it would take to change a station…) to assume that a station can’t do both.

It’s a matter of good design, planning, and a certain amount of clairvoyance. France. When the Parisian stations were built I doubt anyone was thinking about HSR and going all the way across the country. I think new stations and remodeling of old stations should think about how many platforms they need to go through, and how many they could make nicer for riders by having them be end platforms. For me the only failure is when a terminating train stops on a through platform, those are the cases when no one gets a better experience. Through lines through a major city will be tunneled so we have to accept that. Some places we can make terminating lines be at ground level to improve the experience and walk into the station. I’d be interested to here which stations do this the best.

And here’s the Paris issue.
Bordeaux -> Paris-Montparnasse: 3h03
Metro/RER across city
Paris-Est -> Strasbourg: 2h19

Bordeaux -> Strasbourg: 6h31 (runs twice a day now, I think, ran once when I was there).

The Extra time is for a few Paris suburb stops and a semi-circular route around the city. I suppose it’s lucky to even have the direct route, but in some ways like the NEC model better. 50 people start in DC/Bordeaux 450 get off at or before NYC/Paris where 450 get on and then these passengers, mostly new, some old get off in Boston/Strasbourg. It’s a hub system where you pair approximately even ridership on the sides. And, you can run some trains that don’t go through to make sure the ridership demand is met. But, because of the Paris station system you can’t do that.

David: Jarrett’s argument about different types of station isn’t about how many passengers get on and off, but about where the platforms are in relation to the station. At European end stations, you can walk directly from the platforms to the concourse. At through-stations, you need to climb or descend stairs, or cross the tracks on a bridge or in a tunnel.

wasn’t DB trying to buy a British tourist railroad so it could offer direct service to London? Wasn’t the main opponent SNCF because they feared lsot revenue or wanted riders to come through Paris?

John W pretty much nails it. But there are a few other things that detract from Eurostar:

1) The aeroplane/Home-Office-neo-fascist lobby in Britain forced them to microwave luggage, despite the fact that in 25 years of “war” the IRA never put a bomb on a train going through the Severn Tunnel: there are far easier ways to disrupt railways, mostly involving access to the right-of-way, not the stations or trains. This is a spectacular waste of time and money.
2) Transferring across London is a pain. Birmingham, Liverpool, and Manchester involve a ten-minute walk between stations along narrow pavements. Bristol and Cardiff involve some of the most infrequent Underground lines and playing guess the tube station at Paddington. There needs to be a plan to divert all long-distance services into Euston and shift its services to overgrown commuter villages to Paddington, Marylebone, and St Pancras.
3) You can’t just buy an off-peak ticket and ride virtually any Eurostar (except between Lille and Brussels). All-reserved seating means that every single connection has to be padded.

If the IRA did do something crazy to the Chunnel I think their days would be instantly numbered until all of their heads were on pikes along the highways of Great Britain.

I agree that Eurostar is a pathological case and generalisations should not be drawn. However, it should be pointed out that coming from Germany it is possible to buy a Köln-London ticket in a single transaction at any DB Reisezentrum (and in some cases on In fact there are special discounts for such journeys, such that they are often cheaper than just booking the Bruxelles-London segment on the same train directly with Eurostar. Coming from the UK end is a hopeless mess though. (St Pancras has four separate, incompatible ticket offices, three of which (like every other ticket office in the UK) are incapable of booking Eurostar tickets.)

Jarrett’s emphasis on the “experience” of terminal stations seems rather silly to me, reminiscent of the “Disneyland theory of transit” he criticises elsewhere. The purpose of transportation is not to “magnify a city’s importance”, but to move people, and in almost all situations a terminal station is far less practical and efficient for this purpose. Even when a terminal station does make operational sense, circumstances often mean the “experience” still involves climbing up and down stairs to cross the tracks (e.g. LA Union Station and the proposed SF Transbay Terminal). Political changes and urban growth mean that any new city-centre station to be built will almost certainly need to be accessed by tunnel, and so stairs and overpasses need to be dealt with anyway.

Other than frequency, I’d say it depends how ‘easy’ the transfer is, or better yet, how easy the transfer is psychologically.

I remember that the transfer from Boston’s Red Line subway to Mattapan Ashmont High Speed Line was about as good as it gets… the trolleys would enter the Ashmont station building, drop off literally 20 feet from the subway car doors, loop around over the subway tracks and pick up arriving passengers on the other side, again less than 20 feet from the subway car doors. And this was a multi-modal transfer too. Psychologically it was a very easy transfer. Unfortunately a few years ago they ruined this easy transfer when they rebuilt the station, not that its terrible now but its much less convenient. The MBTA used to have lots of these easy in-station transfers…Fields Corner, Harvard (still exists), Sullivan Square (the old terminal), Andrew, Fenway, Forest Hill (pre-1987) to name a few.

I’ve always thought these kinds of in-station multimodal transfers are needed more especially in new transit construction… the SMART commuter rail station at the Larkspur Ferry terminal should have a flyover allowing trains to pick up and drop off feet from the ferry boats (like Hoboken Terminal and the former Oakland Mole). It would also be great if the LAMTA Orange Line went into a very short tunnel and entered the North Hollywood Red Line station mezzanine like the Harvard station in Cambridge MA, preferably within the subway’s fare paid zone. Yes these are expensive but they break down the psychological barrier of transfering in general and especially when one thinks of needing to walk a long distance or cross a busy road, both of which are huge turn offs.

Its also nice to have large transfer terminals where it is not only easy to make multimodal transfers but also has services such as restaurants, bars, newsstands where one can spend time before catching their train/ferry/bus. Hoboken Terminal is one of the best examples that I can think of that still exists. I believe to some extent this is the case with SEPTA’s 69th Street Terminal, and was with the Pacific Electric building and MBTA Sullivan Square terminal decades ago and the Transbay Terminal before it smelt like piss.

I have two issues with your article.

1. The Thalys and the ICE3 high-speed trains operate between Bruxelles and Koln on classic tracks, meaning that they are not operating at high speed and are sharing tracks with other train services. The journey from Bruxelles to Koln takes 2 hours 15 minutes, not high speed for the distance. I think the lack of market share is that 5 hours from London to Koln (with a transfer) does not compete with the airlines time.

2. Customs when arriving in the UK or Frace on Eurostar are practically non-existent. There is a passport check before boarding the train in the UK or France. On arrival, there is no customs declarations to make, you simply walk off the train and into the country. There are custom officials watching, but there is not a formal check and no custom forms to fill out.

I think a lot of people have figured this out: it’s not direct SERVICE, it’s direct TICKETING with an easy change. A change at Brussels is confusing — but it shouldn’t need to be — and direct ticketing from the “important” end, the UK, is nonexistent.

In addition, travel time is key, and London-Koeln is over the “magic time”, though this will probably change as additional high-speed lines are put in.

David: There is no customs (because the UK is in the EU), but there is immigration before departure at both ends (because the UK is not in the Schengen Area). You therefore have to queue to have your passport checked (and if you are not from the EU, stamped). As a result of this and the moderate level of security theatre, passengers are told to “check-in” 30 minutes before scheduled departure time. How many other HSR routes could have 30 minutes cut from their effective journey time for zero capital cost?

“Because the United Kingdom is outside of the European free exchange zone, travelers must pass through customs”

The UK is part of the “European free exchange zone” if by that you mean, as you must, the European Union or the European Free Trade Association. The UK is not a part of the Schengen Area, which allows for passport free travel within its limits.

Not one person on this comment thread mentions Switzerland and its nation-wide timed connections policies. So much for pontification!

For the record, Switzerland has the second highest per capita rail ridership in the world, and relies very heavily on its national timed connections system.

Yes, Switzerland has high rail ridership. But it’s not just the timed connections – those are present in Britain, too. It’s also the way Swiss geography works: the country is too small for a large domestic air market, and the mountain crossings are such that most roads are poor with top speeds in the 40-50 km/h range. Only the most important passes have highway tunnels, but those also have rapid rail tunnels, eliminating the advantage of roads over rail.

Besides, it’s easier to time connections within one rail operator’s territory, as in Switzerland. Eurostar would have to time connections with DB, which is a more difficult feat.

Good post and thread.

But my post on end-stations vs through-stations should not be read as advocating or opposing requiring transfers. Through-stations are easier for through-trains, but trains can and often do flow through end-stations; they just have to reverse direction. German and Swiss trains do this all the time.

So the main effect of end-stations is that if you’re on a train continuing beyond the city, (a) the end-station is a little out of your way and (b) the train will have to reverse direction there. If you’re glued to your laptop, you’re not going to care about either of these things, just the total travel time and whether you have to change trains.

Second, you should probably separate out the Schengen issue in the UK from the main point. As the Bordeaux-Strasbourg example shows, the through train is not always faster, even if you chose to wait all day for it rather than take the next frequent train that puts you through the center of Paris.

So as Yonah states the question, I think there are too many separable issues here to offer an easy answer. In general the questions here are:

1. How is the handling of UK customs on trains weakening the train’s ability to compete with flying. (Pretty unrelated to the other issues.)

2. How do intercity rail customers make the following tradeoffs?:

a. Frequency vs single-seat ride. The itinerary requiring a change via a major city is ALWAYS going to be much more frequent than an itinerary that bypasses the city. So this depends a lot on the kind of traveller. Businesses travellers need flexibility about departure times, especially at the end of the day! So having their perfect itinerary only on a single train leaving at 17:47 may be way less helpful than having reliable trains every hour on the hour, even if they require a connection.

b. Travel time vs single-seat ride. On this I agree, a single seat ride with a slightly longer travel time often wins because I can spend the whole time productively, if the frequency is comparable. But it usually isn’t (see a).

These are specific enough that some surveys of current riders and nonriders could probably quantify them.

1. How is the handling of UK customs on trains weakening the train’s ability to compete with flying.

At the moment, it’s not – this is a plus not a liability. Check in time, which includes time to pass through customs, is 30 minutes (10 minutes for business travellers!). On arrival by rail, there are no formalities, giving it a further edge against air travel. The only thing comparable is London City Airport’s 20-30 minute check-in (and a presumably equally breezy arrivals procedure).

If direct trains were added to further European destinations, this could all become a lot more complicated. However, if the UK maintains just the two destination stations, passport control on arrival in the UK would likely be the simplest and cheapest solution. For outbound travel, the status quo works fine.

If regional UK cities are added into the mix (as was originally planned), whether on legacy tracks or as part of HS2, pre-boarding clearance is much less workable. A possible scenario would then be to go back to the standard procedure that was common in Europe pre-Schengen, when officious borders agents walked through the trains with their little boxes and stamps.

In the North American context, having both US and Canadian customs checks in Montreal and Vancouver would make sense for those routes as there’s only one stop on the Canadian side of the border. For Toronto and south-western Ontario (linking to Buffalo or Detroit and beyond), onboard checks would make the most sense in limiting delays and passenger inconvenience. But would it be acceptable to the US and Canadian governments?

David: Also, trains between Bruxelles and Köln do not use much legacy track. The HSL2 line between Bruxelles and Liege has been in use since 2002, and this summer ICE trains also began using the HSL3 line between Liege and Aachen (though for now Thalys trains still use the legacy track for this segment). Most of the Aachen to Köln line was upgraded for 250km/h running in 2002. Bruxelles-Köln journeys are unlikely to get much faster in the near future than the 1:57 taken by ICE trains, and for London-Köln far greater journey time reductions would be achieved by coordinating and improving connections in Bruxelles (if not through-routing).

Alon: “Besides, it’s easier to time connections within one rail operator’s territory, as in Switzerland. ”

Switzerland has multiple rail operators and they do time connections *with each other*.

“So the main effect of end-stations is that if you’re on a train continuing beyond the city, (a) the end-station is a little out of your way and (b) the train will have to reverse direction there.”

Through stations are preferred for *operational efficiency*. In some countries’ regulatory regimes, reversing direction creates a real delay.

More importantly, unless you have a complex set of flyovers like Sydney Central, end stations on a spur in the middle of a north-south line (for instance) cause northbound and southbound through trains to foul each others’ tracks. Even with the flyovers, you need twice as many tracks on the spur line to have the same capacity as a through station.

This is irrelevant for relatively-low-frequency lines, such as ones serving trains only every half hour in each direction, but once a segment starts serving trains every few minutes, it becomes very significant. This is why so many end stations have had through stations dug underneath them, even at enormous capital cost. If your station is a natural terminus, an end station is fine, but if the trains run through, an end station is only OK if there are relatively low frequencies.

“For Toronto and south-western Ontario (linking to Buffalo or Detroit and beyond), onboard checks would make the most sense in limiting delays and passenger inconvenience. But would it be acceptable to the US and Canadian governments?”

The US and Canadian governments have been forcing the trains to stop and doing onboard checks while stationary. And being very very slow about it.

….but more importantly, there are two Niagara Falls stations, one in the US, one in Canada, which are throwing distance apart. The train has no time in motion for the inspectors to work, assuming both stops are used. It would make the most sense to add a pedestrian crossing to the rail bridge, and simply have all the passengers walk across, with the border crossing there; but that would make TOO MUCH SENSE for the US Border Patrol, apparently.

Frankly, something similar is true at the Detroit-Windsor border, where the most appropriate station locations are *very close* either side of the river. It would be most appropriate to have a pedestrian crossing and timed connecting trains, though the bridge locations aren’t quite as perfect for it as they are in Niagara Falls.

Nathanael: Regarding Windsor and Detroit, by “very close” you mean two miles or more, hardly a reasonable distance to ask people to walk with their luggage (in the snow or rain!), and I don’t know how you could have a reasonable “timed connection” given such distance. If ever the article’s thesis of connections destroying ridership applied, this would be an example. Stations near the Ambassador Bridge would also be fairly difficult to reach from downtown Windsor/Detroit. In this case I think it makes much more sense for trains from Toronto to stop in Windsor near the tunnel portal, run through the tunnel, and terminate at the current Detroit Amtrak station, with customs/immigration in both directions conducted there. Passengers traveling onward to Chicago would still need to connect, but at least passengers for Detroit would not be inconvenienced. Ridership on routes like Toronto-Chicago seems pretty hopeless to me as long as the US and Canada insist on passport checks.

A pedestrian-based connection at Niagara Falls would be slightly less absurd, but still likely very damaging for through-ridership given the statistics in the article. Perhaps in addition to trains terminating either side of the Whirlpool Bridge, there could be NY-Toronto through trains that stopped at neither Niagara Falls station, allowing time for passport checks between Buffalo and St Catharines. Some sort of joint US-Canada station at Niagara would be even better, but would probably pose too many logistical and diplomatic difficulties. Again, the ideal long-term solution (for this and many other issues) would be a Schengen-like US-Canada immigration union.

Given that there’s already a railroad bridge, why go for pedestrian-based connections, which are proven ridership-killers, when you can have ridership-boosting cross-platform connections?

It’s not hard. Fence off a track with an island platform at one of the Niagara Falls stations, and at one of Detroit and Windsor. Designate that track to be equivalent to the other country’s territory. Have the connecting train stop at the track opposite the fenced off track, on the same platform, with immigration checks in between, and a train from the other country’s rail system waiting across the platform, ready to take in passengers who’ve cleared customs. This allows immigration and customs to take 20 minutes, as it does at border preclearance at Canadian airports, rather than 1:50, as it does on the Maple Leaf.

Another solution: run trains nonstop in either Canada or the US. For example, for Chicago-Toronto service, there could be two stopping patterns, one American train running all-local in the US but stopping only at Toronto and Windsor in Canada, and one Canadian train running all-local in Canada but stopping only at Detroit and Chicago in the US. The American train would stop at fenced-off platforms at Windsor and Toronto, the Canadian train would stop at fenced-off platforms at Detroit and Chicago. People who want to get from an intermediate point to an intermediate point can transfer to a domestic train, which should be timed to provide a transfer time of less than 30 minutes.

“Perhaps in addition to trains terminating either side of the Whirlpool Bridge, there could be NY-Toronto through trains that stopped at neither Niagara Falls station, allowing time for passport checks between Buffalo and St Catharines. Some sort of joint US-Canada station at Niagara would be even better, but would probably pose too many logistical and diplomatic difficulties. Again, the ideal long-term solution (for this and many other issues) would be a Schengen-like US-Canada immigration union.”

I agree with *ALL* of these suggestions. There should be plenty of time for passport control between Buffalo and St. Catharines, and anyone going to or from Niagara Falls can catch one of the other trains and doesn’t need to cross the border by rail. And there’s no reason whatsoever for the US and Canada to retain border control paranoia, but I’m afraid that’s going to be a long, hard process to change.

Regarding Detroit and Windsor, the current Amtrak station is not in a terribly great location — and it’s totally unsuitable for travel from Canada except as a terminus, and not much good as a terminus. And I’ve heard stories implying that the current Windsor station isn’t that great either. The Michigan Central station in Detroit isn’t in a terribly good location either. But I really don’t know how to get a good station location in *either* city without spending a fortune, which is not justified. The built-up environment seems to prevent the railroads from actually getting to downtown.

“Fence off a track with an island platform at one of the Niagara Falls stations,”
That might just work, but the Canadian station is quite constrained in its site, and the US area (for the planned station relocation — the current station is not interesting as a location) has challenging topography which would likely require a lot of earthmoving.

“Another solution: run trains nonstop in either Canada or the US. For example, for Chicago-Toronto service, there could be two stopping patterns, one American train running all-local in the US but stopping only at Toronto and Windsor in Canada, and one Canadian train running all-local in Canada but stopping only at Detroit and Chicago in the US. ”

I would expect that to be a ridership killer. There are people going from the US to “points in Canada” and people going from Canada to “points in the US”. Unless the transfers were very carefully arranged I don’t see how it could be made convenient.

I would note as a side point that if there *were* a pedestrian connection, it would be possible to use it to connect from *existing* VIA service to Niagara Falls to *existing* Amtrak Empire Service to Niagara Falls. This is unreasonable currently *solely* because there is no convenient way of crossing the border near either station.

Separate services for the US and Canada running nonstop in each other’s countries would surely be less of a ridership killer than forcing people to walk across a bridge on foot. The transfers could be arranged well, if Amtrak and VIA put their minds to it (this would require better schedule adherence, too). For example, if the nonstop-in-Canada Maple Leaf arrived in Niagara Falls, Ontario 20 minutes before the Niagara Falls-to-Toronto service departed, people could transfer relatively easily.

Some cooperation with customs authorities on the cross-border services is also needed. If you go back to the 1960s, CN’s service on the Grand Trunk from Chicago to Toronto involved both US and Canadian border officers boarding the train in Battle Creek and then proceeding to work their way down the train, interviewing everyone and checking IDs and passports. The result? No delay whatsoever at the border. US officials would leave the train in Port Huron and the Canadians in Sarnia, to await the next westbound service. Some similar degree of creativity, if not necessarily the same plan, is needed today in order to help with travel times in situations where you don’t have at least one endpoint next to the border.

A personal note to add to the commentary regarding train changes in Brussels.

I’m researching a trip from London to Amsterdam. I know that Thalys services recently started running over the new HSL Zuid line in the Netherlands cutting an hour off the old 3 hour trip. So I looked up Eurostar/Thalys trip combinations.

First of all, the lack of timing on connections in Brussels means that it’s actually faster to take the local intercity train rather than Thalys, canceling out all of the benefit of the high speed line. The best time provided was over 6 hours, while through-running could allow a trip in under 4.

Second, the pricing of the service varied widely between (£350!!!) and (€108).

I found a flight out of the centrally located London City Airport (22 mins. via DLR from the center of the financial district) for under £100. I love high speed trains, but sadly there’s no advantage to rail on this route as it stands.

Leave a Reply