Yes, and it’s already happening.  Here’s a little Christmas gift for high-speed rail (HSR) fans.  I just pulled this Guardian UK article from August (via The Infrastructurist) about how the demand for HSR in Germany is really starting to crowd out the airline market.  For an option that was once better than taking the car, but not as good as flying, rail is becoming the predominant mode in the intercity travel market.  Many short-haul train trips are now faster than comparable plane trips.  According to Pierre-Stephane Austi, CEO of Rail Europe, trains are, on average, a quicker ride than planes for distances up to 1500km (932 miles– for comparison, roughly the distance from Seattle to Fresno) when factoring in check-in, security, and recommended buffer time for air travelers.

It’s a phenomenon that’s beginning to be realized across Europe, where 90% of travelers between London and Paris are now taking the Eurostar over any airline.  For a train to beat a car in the States is quite a feat, but winning over air travel is pretty unthinkable, at least for now.  From the Guardian:

The journey considered to be the nearest modern equivalent to magic carpet rides is the Cologne-Frankfurt route, which used to take over two hours but has been cut to just over an hour. Taking the car is hardly an alternative, when even whizzing in your BMW on the speed limit-free autobahn would take twice as long as the train. Berlin to Hamburg by rail now takes about 90 minutes, whereas a few years ago a flight would have taken at least two hours, taking check-in time into account.

Okay, Secretary LaHood and America, let’s get on it!

[UPDATE: This isn’t exclusive to just Europe.  The China Post reports that a Wuhan-Guangzhou HSR line will hit the airline industry hard there (H/T: Gordon Werner).]

65 Replies to “Can HSR Really Compete with Air Travel?”

  1. Oh how I wish we could have this in the States. I know it’s unlikely, but oh how I dream of it. I love the idea of not having to pad a 1 hour plane ride with 2 hours of security and waiting.

    Calgon, take me away.

    1. Yes – I love arriving at the train station just minutes before departure. And the best thing is I don’t have to get undressed to board which benefits the people boarding around me.

      Love the Calgon comment.

    2. It’s happening… construction of high-speed rail in California starts in 2011 and there’s an outside chance it could begin operation before our Capitol Hill light rail station. It will almost certainly begin operation before light rail service to the Eastside.

      We should really get cracking on building Washington and Oregon’s connection to California’s high-speed rail. If we can work together to build light rail from Vancouver (WA) to Portland (OR), we can work together to build high-speed rail through our states, too.

      1. I don’t even like the fact that California HSR is separated from the Northwest corridor. We should be treating this as a West Coast corridor.

      2. Sherwin,

        No, we shouldn’t. There’s nowhere between Sacramento and Portland big enough to support HSR and there are the twenty mile wide Siskiyou Mountains and three steep east-west divides each about four to ten miles thick at the base to cross in southern Oregon. The construction costs would be stupendous. You can’t even escape the tunneling costs by going through Klamath Falls and sacrificing the modest accumulations of folks around Medford and Roseburg, because the east side of Willamette Pass is too steep to traverse in one pass. It would require looping, not exactly the highest speed of alignments.

        Cal HSR blinked at the costs of tunneling under the Grapevine. You think Oregon and Washington can deal with the Siskiyous?

        Also, remember who is making this claim that trains are superior than planes up to 1000 miles: the president of Rail Europe. If his quote were from Herb Kelleher about how much more cost effective Southwest is between Dallas and Houston than building the Texas Triangle, would you be jumping up and down? He’s talking his book, just like any other CEO.

        Finally, remember that the primary reason that HS trains emit so little CO2 in Europe is that the ones in France are all powered by nuclear plants. Here in America where we generate 55% of our electricity from coal, the figures for trains are much closer to planes than they are there.

      3. “Finally, remember that the primary reason that HS trains emit so little CO2 in Europe is”

        Actually a lot of other things are more important than the power generation method:
        (1) the sheer number of people per train. This beats the pants off of small airplanes — but can only be achieved on *busy* routes.
        (2) The overhead of taking off and landing airplanes. This is why even large airplanes are less efficient than trains for short hauls.

        Unfortunately Portland-Sacramento pretty much lacks all of these advantages. It’s long enough that the ‘fixed overhead’ of taking off and landing is spread over a longer plane trip, it’s got enough volume that larger planes can be used, but it doesn’t have enough volume to run giant 20-car double-decker trains.

        And as previously noted the construction costs would be stupendous. If California and Pacific Northwest HSR *both* get built, both do extremely well, and the last plane service on the West Coast is a busy Sacramento-Portland “connector” between the two rail services, *then* maybe we would want to consider the massive base tunnels and bridges following the route of I-5 from Eugene to Medford to Redding.

      4. Actually, there would still be airplane service along the west coast, mostly either regional jets flying to places like Bend or Boise or Yakima that lack any train service, or to places like Spokane that have poor service. There would still be 737/A320 type single aisle airplanes connecting the major cities (SEA, LAX, SFO) for those who have to get there in a few hours, for a connecting flight overseas or a business meeting.

      5. Nathanael,

        People per train doesn’t necessarily depend on a route being busy. Obviously, if one runs lots of trains empty, one is not reducing emissions, but the Builder is sold out between Chicago and Minneapolis and Seattle and Spokane nearly every day, even though it goes but once a day. I expect that especially on the flat between Chicago and Minneapolis it gets very low emissions per passenger mile.

        You’re right that people per train plays a big part in overall rail efficiency and also that frequent take-off and landing operations do drive up the emissions of aircraft. Southwest and Hawaiian have higher emissions per passenger mile than do the long-distance operators.

        But you are wrong about the CO2 emissions for HSR trains being potentially equivalent in the US, at least with our current and foreseeable power system. The primary factor in any electric usage emissions is the source of the electricity, considered over the entire fuel and construction cycle. Here in the Northwest we forget that in places like Ohio and South Carolina 95% of every kilowatt-hour used comes from coal plants. You certainly don’t want transportation completely changing to electric propulsion in that environment. Emissions would actually rise, compared to a fleet of efficient internal combustion powered high-occupancy vehicles (i.e. buses)

        There’s a place for major trunks to do so as long as it results in a reduction of SOV and short hop air miles. Frankly, coal powered HSR trains would be worse CO2 per person emitters than a minivan with a five member family.

        In Europe the combination of France’s extensive HSR network and the fact that it gets 82% of its electricity from nuclear plants significantly reduces HSR emissions per passenger mile for the entire European system. While that’s a very good thing, it will not be matched in the US. What is so wrong with me pointing that out? Do we all have to sing in the “Rail Everywhere” choir here?

        And finally, I was questioning the I believe ridiculous claim that trains are competitive with planes at distances up to 1000 miles. In the first place, there are no through runs in Europe of 1000 miles that don’t cross at least some noticeable mountain range, reducing the 180 mph speed. Even if a there were a Thalys run through to a TGV Marseilles via the Paris Est bypass (which would be close to 1000 miles) allowing only the minimum time of five hours 35 minutes to be required, a direct flight from Amsterdam to Marseilles would take about four hours end to end, including the security frou-frou at Schiphol. That leaves 45 minutes at each end of the trip to get to and from the airport. At the absolute best it’s a wash, and the truth is most train trips include intermediate stops and a distributor at each end too.

      6. One step at a time.

        Switzerland didn’t start off by building the Gotthard Base Tunnel, they did a gobload of smaller upgrades first. Russia didn’t build the Trans-Siberian Railway until they’d already built a huge railway system in the western 6 time zones.

        Once there’s a high-speed Seattle-Portland(-Eugene) run and a high speed Sacramento-Los Angeles run, you can start thinking about whether Portland-Sacramento makes sense. With the current state of rail at either end, it certainly *doesn’t* make sense.

  2. Nice post. But more to the point, hasn’t Acela already captured 60+ percent of its market?

    1. Indeed, and the Delta and US Airways shuttle flights have been down-gauged to regional jet aircraft on some legs.

    2. Yes, Acela is the closest thing we have to HSR, but it’s only limited to the Northeast Corridor. But even still, it’s not as fast as air travel.

      1. Acela is as fast as air travel city-center to city-center. Throw in the TSA Goons’ Kubaki Theater and Acela is a no-brainer.

    3. Yes, and by European standards, Acela is a joke. Imagine what a few billion to straighten out the curves and replace the catenary would do…

      1. Ryan,

        The curves are an issue on the Shore Line north of New York. The Pennsy “Standard Railroad of the World” south of NYC would be pretty much ready for 150+ with the catenary replacement. It’s already engineered with very broad curves.

        Except for the freight trains and the 90 mph Clockers and NJT Arrows.

        Until some way to get NS trains into Baltimore without running on the NEC south of Perryville is provided the route can’t be upgraded to true HSR. But 135 mph seems to be doing very well already. It’s really not that far between New York and Washington.

  3. Washington state, more than most states, has spent quite a bit improving rail service like the Amtrak Cascades. Right now a trip to Portland takes about three hours, still more than driving or flying, but it has improved over the past and is starting to be competitive with the other two modes. The problem is that Oregon has not contributed anywhere near as much to its portion of the Cascades network, nor has British Columbia. I doubt we will see high speed rail outside of the Vancouver BC – Seattle – Portland corridor, unless Oregon gets lots of federal money.

    Unfortunately, it will take a lot of money before a trip from Seattle to Fresno or more likely Oakland/San Francisco via train is competitive with flying. Right now the Coast Starlight takes about 23 hours to get to Oakland, provided there are no delays. An airplane takes about two hours flying time, plus a few hours going through security, etc. So the delta is 19 hours, and changing that will take lots of money. I don’t think any of the high speed rail lines in Europe have the terrain challenges like between here and Oakland, except the tunnel under the Channel.

    I think it is better for Washington state to spend money improving service between Seattle- Portland and Seattle- Vancouver. We should do much as can be done by improving the rail lines within our borders. When we hit the point of dimishing returns because our neighbors are not funding improvements as much as we are, we should turn to improving rail service east-west.

    1. The Swiss especially have been very good at tunneling under mountains and through mountain ranges, but at vast expense. The US and Canada should focus on higher speed rather than HSR in segments like Eugene-Redding, Sacramento-Reno, Calgary-Edmonton etc, and build out HSR over the next 50 years in relatively flat areas where there is less environmental and monetary cost.

      As Erik G says below, though, we’ll have to have much greater adherence to operating regulations and much stricter management and equipment maintenance if we are going to avoid mayhem and chaos on rail lines. The old American way of “Why not pretty good?” just doesn’t work in the air (hello Delta pilots) or on the rails (hello Link operator) in the 21st Century.

      1. Lloyd,

        Calgary to Edmonton is pretty flat. There are two parallel rail lines (CPR direct and CNR through Drumheller) that are pretty much crow-flies straight between the towns on their respective routes.

        If there is adequate ridership between them, the topography would be no issue.

        The other segments you mention are right, though.

    2. If you want higher speed rail connecting the West Coast, the target to aim for is probably 8 hours from Portland to Sacramento, enabling an easy overnight ride between the Northwest Corridor and the Bay Area. And that sort of speed is much more realistic than full-on HSR in a fairly sparsely populated corridor.

      1. You bring up population density, which is another reason why HSR along the whole west coast is unlikely. Our population centers- San Diego, Los Angeles, San Francisco/Oakland, Portland, Seattle, are too far away from each other for all to be linked by HSR. Europe and the Northeast Corridor have HSR, but their population densities support it. Your idea of a higher speed rail to link the Northwest with CA is much more realistic.

      2. HSR makes sense in California — LA and SF are less than 3 hours apart by HSR, there are a number of smaller but non-trivial metros along the way (Bakersfield, Fresno, etc.), and there are regional centers that will happen to be along the route, too (Anaheim, San Jose, etc.)

        It just doesn’t make sense to connect California to the Northwest.

      3. So does anyone have any sense of what would be required for 8 hours Portland to Sacramento?
        Preferably via Eugene, Medford, and Redding (higher population than the “eastern” route).

      4. Currently, the Coast Starlight runs through Eugene, then heads over the mountains to Klamath Falls, then heads south to Dunsmuir and Redding. Any HSR built along that route will miss Medford. Improving that route hits two of the three cities you mention.

        As for improving the other route, I do know that an Oregon DOT study of the rail line south from Eugene to Medford is that it is too slow- top speed is only 25 mph or so due to track quality and terrain. Persumably it would take a lot more $$ to get that route up to HSR standards than the one the Coast Starlight currently follows at about 65 mpr or so. Oregon currently is spending a small amount of money improving the Portland – Eugene route, so any new funding would have to come from the federal government.

        In any case, these cities are small and the terrain rugged. It is probably better to shoot for higher speeds and better on time service than HSR in this area, at least for the next 10 years or so.

      5. Well said, Dan, but the timeline for “higher speeds and better on-time service” is more like forever. Honestly, one can’t even do HSR via Willamette Pass without a twelve to fifteen mile tunnel. The UP passes through the stretch between McCredie Springs and the overpass bridge at the base of the climb to the summit three times to gain 1000 feet of elevation. The steep valley around Lookout Point would also be very difficult to build through for HSR, so it would probably have to have a tunnel, too.

        Barring a huge revolution in hard rock tunneling, which is still a pretty darn Newtonian activity, there will never be an HSR link south of Eugene.

        It might be worthwhile someday to spiff up the old O&C line as far as Medford for Cascades service, but it would not be cheap. State sponsored buses can probably do the job much more cheaply and with more frequent service. There are already many examples of Oregon-sponsored Ambus schedules sticking out from the Cascades stem. They go as far east as Ontario and west to Brookings, Coos Bay, Newport and Astoria.

      6. Oregon has made smart use of buses because much of the state in not in a position to be served by rail. You can read the Oregon 2001 Rail plan at:

        The document details each and every rail line in the state, and on page 98 discusses each potential route for intercity rail. Most rail lines in Oregon are not very well suited for intercity rail without a lot of money. So buses that link to rail stations make a lot of sense.

      7. One possibility would be run a train between Eugene and Redding via the line through Roseburg, Medford, Grants Pass, and Ashland. The track would have to be upgraded to class 4 or 5 from the class 1 or 2 track it is currently. A short collapsed tunnel would have to be repaired as well. While true HSR speeds couldn’t be obtained with Talgo gear speeds could be better than with say Superliner or Amfleet cars. If high-power locomotives (like the Talgo diesel units) were used the steep grades are less of a problem.

        The big advantage of switching passenger service to the line in the I-5 corridor is the UP line through Klamath Falls tends to be rather congested with freight and really can’t handle increased passenger frequencies. The line through Medford is rather lightly used so freight conflicts aren’t really a problem.

  4. “People get to work on time”???

    Who the hell is going to use HSR to commute to work??

    I love the dream, but based on the American culture of property rights, the lack of respect for regulations, and low tolerance for strict management, I do not think that the USA can achieve a train that runs more than the 150 mph it does today. Still that would nice to see in more places, but there will be many, many, many trespasser fatalities.

    Kudos to M. Stanley for not mentioning Kitty once in his spiel!

    1. In Japan commute via Shinkansen is fairly common. It’s more expensive than a local train of course, but you can get a nicer place 200km from the office. It’s not all that different from car commuters here, except that you don’t need all the parking downtown or at the suburbs.

    2. Supposedly some people use HSR in France to commute. The smaller cities 30-60 minutes via TGV from Paris are much cheaper than anything the same travel time via the Paris Metro, RER, or commuter lines.

  5. would love to see this happen in america but it won’t happen in the next 50 years (if at all).

  6. Aaaahhhh…wouldn’t it be nice to have high-speed rail in this region. Then, people could live in Bellingham and just have a ‘normal’ commute length of 45-60 minutes, which is how long it takes to drive just to Everett sometimes in this traffic. Or, maybe you want to live in sale tax-free Oregon but work in high-tech Seattle/Bellevue. Isn’t that a kind of freedom of choice that Republicans always scream about? Me, I would love to know I could get up to Vancouver or down to Portland in about an hour or so without having to worry about snow, air travel delays, stripping for security…

  7. There was a trend to build train stations in the outskirts of a city, instead of downtown. They were abandoning the downtown terminals and selling off the right-of-way. Should a HSR have to use a station in the outskirts, the trip to the true destination (downtown usually) could be like a trip to and from the airport.

    1. We had that problem in the Northeast (Penn Central abandoned an awful lot of well-located stations), but only in a few cities (Syracuse) did we actually lose the downtown *ROW*. Luckily the center of activity in those cities has since moved (tin the case of Syracuse, towards the rail ROW). In the West, you still have downtown station locations practically everywhere. In the Midwest, the stations are often gone but the ROW is still there, and similarly in most of the South.

      In Europe, HSR almost always goes fairly close to downtown in the *big* cities — but those big cities usually have connecting metro or S-Bahn services to downtown anyway. Outskirts stations were used in the TGV in France on smaller cities, and it has been a problem. Germany avoided this.

  8. Unfortunately, by the time we build rail, cars will have the technology to be auto piloted and able to run at high speed on corridor fastways.

    Rail is 19th century technology. Self/Auto/Guided personal transportation is always the future…

    1. Blue Swan, did you stumble upon your Dad’s 50s Era Popular Science collection in the attic while home for the holidays? ;)


        Cars are approaching ‘auto’ pilot mode
        With humans known to be unreliable sorts, vehicles get helping ‘hands’

        But a new breed of prototype automobile can drive without the help of unreliable humans, and major car companies are paying attention.

        In 2007, the federal government’s Defense Advanced Research Projects Agency sponsored the Urban Challenge. Corporate-sponsored teams from all over the nation retrofitted regular cars with sensors and artificial intelligence, transforming them into fully autonomous ground vehicles. Eleven finalists unleashed their driverless cars on a peaceful mock city where they proved capable of obeying traffic signals while merging, passing and parking. ”

      2. Unfortunately cars can’t see cats crossing the road. Or if they can, they stop for leaves falling from trees.

        This is why non-grade-separated transportation requires a human (or at least animal — horses are actually better at this than humans) controlling it: the pattern-recognition problem. Generalized visual pattern-recognition is a hard problem which computers are not going to be able to solve for decades if *ever*.

        These “automated cars” might work on fully-grade-separated superhighways, but if you’re going to do that, it’s cheaper to build and run a train instead. Plus the train can run faster (thanks to having tracks to keep it where it’s supposed to be) — the fastest speed for safe road vehicles is much less than that for railed vehicles, for simple mechanical reasons.

    2. Cars are 19th century technology, and older than railways. Gasoline cars especially, but even electric cars.

      ‘autopiloted’ cars have been claimed to be possible since the 50s. They still aren’t possible.

      Autopiloted *trains* on the other hand operate in revenue service *right now* — visit Docklands Light Rail in London or the SkyTrain in Vancouver, BC.

  9. Quoting a serious UK transportation sustainability advocate, John Whitelegg, writing in the same Guardian newspaper noted in the thread starter by Sherwin Lee, quoted at and in Vancouver, BC by sustainability advocate Gordon Price at

    [High Speed Rail] is promoted as something that can sort out nasty carbon producing
    aircraft on domestic routes. It has done this on the Paris-Lyon and Madrid-Seville lines, but this ability to trash a single air route should not be interpreted as something than can dent the growth of air travel. Germany has one of the largest HSR systems in the world yet has seen an explosion in internal air travel. HSR does not reduce the fuel consumption of domestic aviation or reduce annual carbon emissions from aircraft. And it produces twice as much CO2 per passenger kilometre as a non high speed train. If we are serious about reducing our carbon emissions by 80% by 2050, we should not move towards higher speed, more carbon intensive forms of transport and a policy of increasing the mass of travel. …

    1. An interesting assortment of straw men vanquished. If you can’t dazzle them with brilliance, read Whitelegg’s quote for an excellent example of baffling them with bullshit.

      To sort this out a little, California chose HSR because it is the most cost-effective way of meeting future transportation needs, and the only way to meet those needs with acceptable emissions. The airport association endorsed the choice because there won’t be enough airport capacity to meet air travel needs. Looking to the future, you can run trains on solar or wind power- but not airplanes.

      Whitelegg’s main trick here is to switch straw men about every half sentence, trying to produce false impressions by the close association of unlike claims. He’s very good at it, but a lay reader should be able to spot the spurious comparisons and counter them quickly.

    2. Hilariously, despite the many strawmen he uses to make “true but misleading” statements, Whitelegg manages to be simply wrong about Germany: internal air travel is collapsing. (External-to-Germany air travel is making up the difference, because air travel was previously limited by slots at airports.) Perhaps he’s using out-of-date numbers?

  10. And if high-speed rail couldn’t compete with air travel on December 24, 2009, it certainly can on December 26. Ugh.

  11. Why do many people believe that pre-boarding security processes on future high speed rail in the USA will be less rigorous than the security measures applied to air travel?

      1. No, but you can blow them up. That’s still several hundred people you could kill at once. Not all air travel security is just to prevent planes from flying into buildings, but also to prevent them from being blown up, as witnessed to by the event on the Delta flight the other day.

      2. You don’t need to be on the train to blow it up. Watching any old WWII movie will show you lots of ways to put explosives on the track and then wait for to come by.

    1. Pre-boarding security processes on existing high speed rail in every country are almost nonexistent. The only exception is Eurostar because of the Channel Tunnel, and even there the security check takes much less time than at airports; Eurostar requires you to arrive at the station 30 minutes before departure, compared with an hour or more for flying.

      An explosion on a train doesn’t kill everyone on board, unlike with aircraft – trains can be built to be able to derail without deaths. So terrorists and saboteurs typically target tracks, not trains.

      1. That reminded me of taking Amtrak Cascades from Vancouver, B.C. to Seattle. I arrived 30 minutes before departure and we went through the baggage x-ray, dog sniffing, and passport check really quick.

      2. Yes, the Cascades do security much like Eurostar, I think. International HSR in North America is probably going to converge to the same model elsewhere. Right now the way they do it on Amtrak’s New York-Toronto and New York-Montreal routes is that the train stops at the border for an hour for customs and immigration.

        But tellingly, such controls do not exist on the domestic train routes in either the US or Canada.

  12. HSR that could realistically compete with air travel over 1500 km would be astronomically expensive, and will most likely not happen within any of our lifetimes. I would be more inclined to believe that shorter distance HSR would be much more likely.

  13. Austi’s claim about trains winning up to 1,500 km is factually incorrect. In Japan, the breakeven point is at Tokyo-Hiroshima, a distance of 820 km. On the 1,070-km Tokyo-Fukuoka route, most people fly; trains only run the full route because of the intermediate city pairs it includes. Even allowing for the slightly lower speed of Japanese high-speed rail, there is no chance today’s trains can beat planes at distances much higher than 1,000 km.

  14. I don’t know whether the following Letter to the Editor will get printed in The Seattle Times, but one can hope so.

    Editor, The Seattle Times:

    Time for North America to emulate China’s High-Speed Rail

    The story about China announcing the fastest rail line in the world (at over 217 mph) and plans to build and complete at least 42 more lines stretching over 8400 miles by 2012 stands in stark contrast to our nation’s misguided automobile-centric transportation policies, as well as the Treasury’s financial deficit situation that limits options.

    A U.S. economic stimulus package for high-speed rail allocates $8 billion to “study” the issue. While the I-5 corridor between Vancouver, B.C. and Eugene OR is one of ten potential high-speed rail corridors that will be studied, the first U.S. line will probably be built between San Diego and San Francisco, where California voters approved a $9.95 billion bond issue last fall.

    Meanwhile, China is slated to spend over $300 billion through 2020. While I certainly have major issues with China’s poor records with worker safety, due process, political freedom, and environmental degradation, at least it is embarking on an ambitious program that promises millions of new jobs, a cleaner, greener future and society-wide benefits for the common citizen.

    Our unwillingness to embrace, support and build public transportation infrastructure will increasingly put our nation in competitiveness disadvantage compared to other developed and developing economies.

    Leo N. Egashira

    1. Time for North America to emulate China’s High-Speed Rail

      Yes, it’s amazing what you can do in a Communist country when the government controls ALL of the money. I hear they build a mean hydroelectric dam, too.

      Emulate THAT? No thanks.

      1. We also have those massive hydro dams in Eastern Washington. I think we come close to a one-party state and Frank Chopp holds too much power, just like in China!

      2. Oran,

        I didn’t know that Frank was a member of the Central Committee in Beijing, too. (“just like in China!”) This is a great advantage for Washington in its quest for trade with the PRC!


      3. If I’m not mistaken, relative to GDP the biggest HSR program in the world is in Spain. Spain is planning to build 10,000 km of HSR routes by 2020, just like China.

        You don’t need to be an authoritarian state to build things well.

Comments are closed.