High-Speed Rail and Cities
When preparing various maps proposing high-speed rail in Germany, I was told that it looks nice but it overfocuses on the largest cities and not about connecting the entirety of the country. I’ve seen such criticism elsewhere, asserting that high-speed rail is a tradeoff in which the thickest connections get fast trains but the long tail suffers, whereas the medium-speed system of Germany or Switzerland or Austria serves everyone. So with that in mind, let’s look at the actual population served by a Germany-wide high-speed rail program.
I made a proposal last year, but then made some additional tweaks, posted as part of a Europe-wide map. The most important tweak: the main east-west trunk line was extended to Dortmund, which trades off some additional tunneling in the Ruhr for both higher frequency on Berlin-Dortmund and fast, frequent Dortmund-Cologne and Dortmund-Düsseldorf service. To my later map I’ll add one proposal: moving the Hanover-Dortmund tracks so that trains can stop in Bielefeld. Otherwise, take the maps as given.
The question is, what population is served by those maps? The answer of course depends on what this exactly means. The sum total of the populations of the cities served – Berlin, Hamburg, Hanover, Dortmund, Würzburg, Erfurt, Mannheim, etc. – is 18.2 million, or 21% of Germany’s population. But is that the full story? This just includes central cities, where people in many nearby suburbs and satellite cities would travel to the rest of the country via the primary city center anyway.
For example, let’s go down this list of German cities. The largest city without a stop on my proposed network is Bonn. But Bonn is very close to Cologne and there are Stadtbahn subways connecting the two cities, in addition to regional rail lines; Bonn also has a shorter-range Stadtbahn to a suburban station at which some high-speed trains on the Cologne-Frankfurt line call. Is it really correct to say Bonn is unserved? Not really. So its population should be added to the 18.3 million.
Going down the list, the same can be said of Wiesbaden (near Frankfurt), Mönchengladbach (Düsseldorf) and likewise many Rhine-Ruhr cities, Halle (near Leipzig and potentially on some slower Berlin-Erfurt trains like today), Potsdam (near Berlin), Ludwigshafen (near Mannheim), and many others. Some cities remain unserved – the largest is Münster, like a few other northwestern cities not really near anything bigger or on any line – but overall this adds another 5.3 million. So we get 23.6 million, around 30% of Germany’s population.
But that list is just cities of 100,000 people or more, and there are smaller suburbs than that. These form counties (“Kreise”), which should be added as well when feasible, e.g. when they lie on S-Bahn systems of large cities or when they are right across from cities with stations, such as Neu-Ulm to Ulm. For example, the Kreise served by the Munich S-Bahn, excluding the city proper, have a total of 1.3 million people, and people in those suburbs would be connecting to the rest of Germany by train at Munich Hauptbahnhof anyway – a lower-intensity, higher-coverage network would do nothing for them.
Overall, these suburbs add another 18.7 million people. In Berlin I used this list of suburbs; elsewhere, I went by S-Bahn reach, or in a few cases used an entire region where available (Hanover, Göttingen, Fulda). There are a few quibbles on the margin in the gaps between the Frankfurt and Rhine-Neckar and in places that probably should count but aren’t on any big city S-Bahn like Frankfurt an der Oder, but it doesn’t change the big picture: a dedicated high-speed rail network would serve around half of Germany’s population pretty directly.
Very little of the remaining half would be genuinely bypassed the way Magdeburg and Brunswick were when Germany built the Berlin-Hanover line. Regensburg, for example, is and will remain peripheral under any rail plan, with regional connections to Ingolstadt and Nuremberg; high-speed rail serving those cities is the best way to connect it to destinations beyond Bavaria. Kiel, at the other end of the country, is and will always remain connected to Hamburg by regional rail. Münster, genuinely unserved, is not really bypassed, not with how close it is to Dortmund. And so on.
Such a plan cannot serve the entire country, but it can definitely then serve a majority of it. It mostly serves the largest metropolitan areas, but that’s fine – Germany is an urban country, around 40% of the country lives in metro areas of at least 1 million people (defined again mostly by S-Bahn reach, which is a conservative definition by American MSA or French aire urbaine or Japanese MMA standards) and much of the rest is either in metro areas somewhat below the cutoff or in exurbs served by regional trains but not the S-Bahn.
Pedestrian Observations 
I’m not sure what the benefit of serving Leipzig with the fastest trains rather than Halle is. Yes, Leipzig is the bigger city, but it has a terminus station which adds significantly to any train trip through it and unlike Munich or Berlin or Hamburg, it isn’t a place where most people will leave their train anyway. So going via Leipzig makes a shorter trip for people with the final destination Leipzig in exchange for longer trips for people with final destinations from ca. the midpoint between Halle and Leipzig west and all people who’ll stay in the train through the Leipzig area.
You might quibble that Leipzig has now built a tunnel, but it is basically saturated already with S-Bahn (where the through-running has a much higher benefit) and building that tunnel took over a century. In “organizations before concrete” there is ample reason to route trains thru Halle rather than Leipzig, as the Berlin Munich Sprinter already does…
The idea is that the fastest trains go Südkreuz-Erfurt nonstop. Not my choice, but Berlin-Erfurt in just under an hour is possible, the half of the line that exists doesn’t really point to Leipzig or Halle, and I think stopping at Halle breaks the hour connection (if it doesn’t then obviously trains should stop there, it’s bad enough that Magdeburg got skipped by VDE-4).
Well, the line was IIRC designed pre reunification
Yeah, the idea was to connect West Berlin with the rest of West Germany (still skipping Braunschweig, though), but then unification happened and the line was pitched as an equity project for the East but still was not redesigned to serve Magdeburg.
This seems wrong. Leipzig/Halle is clearly more important than Erfurt or Sudkruez, so if you were to have a local and express stopping pattern this would be irrelevant. It seems to me that you’d want four services passing through the area:
1. Berlin Hbf-Halle-Nuremberg-Munich Hbf
2. Berlin Hbf-Sudkruez-Halle-Erfurt-Bamberg-Nuremberg-Ingolstadt-Munich Hbf
3. Berlin Hbf-Halle-Frankfurt Hbf
4. Berlin Hbf-Sudkruez-Halle-Erfurt-Fulda-Frankfurt Hbf.
So no Leipzig service whatsoever?
No Leipzig thru-service. I’m sure there’s support for plenty of South/West Germany-Erfurt-Leipzig-Dresden service and Berlin-Leipzig direct service.
Yeah. It sucks, but Leipzig lost its ability to have ICE through-service even with a through-tunnel when DB listened to Thuringia’s tantrum and routed the Berlin-Munich line via Erfurt instead of going direct Leipzig-Nuremberg.
At least the trains are not going to go through the AfD-ville that is Gera.
Assuming that Germany had a real 320kph+ HSR system as Alon envisaged, it would be actually good for Erfurt-Leipzig-Dresden and Erfurt-Berlin being totally separate. For Berlin, their trains don’t need to waste time at Leipzig ever again. For DB, that means they don’t need to use their best trains on Berlin-Leipzig because there is no point going full speed between them anymore.
The difference is that Südkreuz is 130 km/h territory whereas serving Halle or Leipzig requires a detour around 300 km/h territory.
And also, local and express stopping patterns are kind of overrated and Germany shouldn’t use them. If the city isn’t worth serving every hour, ideally every half hour, it’s not worth serving; give it a timed RegionalExpress connection. The upshot is that the stopping pattern needs to be (Gesundbrunnen)-Berlin Hbf-Südkreuz-Erfurt-Nuremberg-Ingolstadt-Munich. If there’s room for a Halle stop in the timetable without messing up the knots then all trains should stop there; if there isn’t, none should, it can get an RE or Leipzig.
Seems to me like the HSR line from Berlin to Erfurt would have enough trains that it’s at show-up-and-go frequency which means you don’t need to worry about knots. If you move the stop from Erfurt to Halle, it wouldn’t really cascade down the line in Nuremberg and Ingolstadt. Also, there should be some sort of express service if only to facilitate 600 km+ international services where it really affects total travel time. Should Berlin-Paris trains really stop in Erfurt and Fulda?
Yes, quite silly of the Japanese to have Nozomi, Hikari and Kodama. Everybody not in the important places should just behave like good peons and take the local train to the biggest stations.
The Tokaido Shinkansen has 8 off-peak tph, we’re talking about lines that can get 2.
Then there’s no capacity problems with having a third one that toddles on and off the mainline.
The issue isn’t capacity, it’s frequency. TGV-style 6-trains-a-day patterns for trips that take an hour and a half aren’t what high ridership is made of.
And the train that toddles through 6 times a day is six trainloads of people not on the through trains. which would be less than six if they had to change trains. Just because you find complicated schedules offensive doesn’t mean they shouldn’t do it.
I mean the question is do those six trains come close to paying for themselves? Examples from the Tokaido Shinkansen are useless in Germany, since the former is arguably the best HSR market in the world. It’s like saying what works for Lebron James on the basketball court could work for me, and despite your snark as usual you don’t provide a ton of evidence that it will work in Germany, which seems IMO more like a mega-Switzerland than another Japan.
The honest answer though is DB should run a cost benefit analysis on your model and if it comes close to breaking it even it should try running it for a couple of years, and if it get patronage great keep it, if not cut it.
Shinkansen run on a isolated system. In Germany the trains would be running on the same tracks and using the same stations the slow local train would be using.
»I mean the question is do those six trains come close to paying for themselves?«
Depends on where they go. Maybe some states would chip in some money to have an IC or ICE connection. (like already all ICs to the Lower Saxonian north sea coast, on the Gäubahn and on Gera – Erfurt – Eisenach), where these ICs are integreated into the RE (RegionalExpress) takt.
Marginal costs or full costs?
»Leipzig/Halle is clearly more important than Erfurt or Sudkruez […]«
Which one, Leipzig, Halle or the stroke inbetween? 😉 Unless you’re zig-zagging, you can only stop at one of them, which reduces the catchment area and population. Stopping only at the ‘stroke’ (that would the airport which lost its last passenger flights this week) is not possible, because Leipzig Messe is missing the north↔west curve (and you lose your whole catchment).
Note that Leipzig already has ICE through-service every two hours, just on the east-west line, interlined with IC through-service every two hours east-northwest (to Hannover). They have timed connections to the north-south trains in Erfurt and Halle, respectively.
Either option would probably be fine. The metro area is pretty big and provided there are great regional rail connections I’m not sure it matters that much.
Which metro area, the ¼ million of Halle or the ½ million of Leipzig? If you want both, you have to stop in both.
Or offering a connection in the other city (½h travel time), Erfurt (¾h travel time) or Berlin (1h travel time). Which connection is better depends more on the destination than on the exact travel time.
This goes to confirm my frequent statement that Germans tend to not think in the category of “metro areas”
It doesn’t help that Leipzig and Halle are in two different states.
But they share an S-Bahn
»Yes, Leipzig is the bigger city, but it has a terminus station which adds significantly to any train trip through it […]«
Chainging direction at the terminus certainly doesn’t add much time over the stopping time needed for people to get off and on the train. The detour from Halle to Leipzig Messe, south to the Hbf., back north again to Messe and off to Berlin takes much more time.
Terminus stations also have complicated interdependent switches on the station throat both in and out. Those reduce capacity and increase the risk of delays.
Random request – could you add a category “Crayons” or “Proposals” to the blog with all the network proposals you have made – Europe, eastern US, and various cities?
Building a high speed network doesn’t mean that rail service to smaller towns will be abandoned. As long as the legacy and the high speed networks are well integrated, the people from Münster can benefit from the higher speeds too. The fastest train journey to Berlin is now ~3h30, going north to Rheine or Hamm and then connecting to a train going east. From your crayon, Dortmund-Berlin is 2h05, add 35 min Münster-Dortmund and a 20 min connection and it’s still 30 min faster than the current trip. And the frequency should be much higher too.
Yeah, there’s this worry that the TGV is why SNCF abandoned rural and exurban France… but British Rail abandoned local service in just about all of provincial England, and that’s even with a rather German focus on upgraded legacy rail and a high-frequency takt.
And there’s of course Brazil, which just abandoned all non-urban passenger rail.
For SNCF it’s really a shame that the network works so badly for trips not involving Paris. It’s understandable that provincials get frustrated.
Yeah, and some of it is just small-scale bullshit, like building Lyon-Saint-Exupéry instead of extending Paris-Part-Dieu trains to Marseille. SNCF just has genuinely rotten operations, like hating the idea of seat turnover on grounds that boil down to the French elite’s hate for the French people (“French people aren’t as disciplined as Germans about getting up when someone else has the seat,” say people who have never seen how Germans cut in line at Lidl).
And this hatred of seat turnover carries over to their rolling stock choices. You like to point out how unusual it is for HS trains to still have power cars (only Alstom does it) and presumably this means worse acceleration. I like to complain about the low doors/seats ratio, it takes a long time to embark/disembark. All of this makes for an increased stop penalty. Though even then, do you really need to stop 15 minutes in Liège?
Germans cannot for the life of them form an orderly queue when waiting for something… This is particularly odd since such queues were a daily occurrence in the GDR…
What’s this seat turnover issue? To me turnover is when a lot of people get off the train and thus irrevocably vacate the seats. Are you saying people leave their seats temporarily and come back and find others sitting there who won’t move? In the US people leave something on the seat like a jacket or book to show they’re coming back, and other people respect it. Does that not happen in France?
And on Amtrak there’s a kind of boarding pass above the seat showing your destination.
On a plane you wouldn’t sell the same seat twice on a single trip.
SNCF management is very beholden to airline ideas.
DB says “If the seat is empty, the next person can get on” – on a Munich-Hamburg train the same seat may be occupied by three different people. One from Munich to Nuremberg, one from Nuremberg to Hannover and one from Hannover to Hamburg…
I worry there’s divergence between the worst English regional service got in the 70s-80s and France’s current position. Even quite promising French routes (I took Lille to Calais-Ville a couple of years ago) have erratic, low-frequency scheduling in a way that no comparable English regional rail (which is now almost all on hourly takts) does. I do think this is attributable to the focus on TGV and TGV-style service (big trains, wacky timetables) on the TER.
That’s not the TGV, though – that’s France. You see some adoption of takt schedules, e.g. in PACA, but it’s not systematic, and France is really bad at learning from other countries (as is Britain, to be honest). Likewise, you see adoption of transport associations in the largest regions, i.e. Ile-de-France and TCL, but it’s slow, just as how giving the metropolitan counties in England the ability to coordinate transport is taking forever and I don’t think either Boris or the prime minister is interested in continuing Cameron’s process to that end.
Another french particularity is that trains feed the station, not the city–people shall make their downtown by bus or whatever.
If the rail station is a kilometre or more away from the town centre but the rail line passes a few hundert metres by the centre, never shall a station be built there!
Are there any french equivalents to Schwerin Mitte, Koblenz Stadtmitte, Döbeln Zentrum, Weida Altstadt, …?
I think this almost gives too much credit to England’s incredibly dysfunctional system of local transport planning which doesn’t really seem to be learning from anywhere, let alone from abroad (with the exception of Manchester and London). Metropolitan counties’ transport systems are significantly less integrated than they were forty years ago.
What can these local transport planners actually do? Let’s take railroads for example. National transport planners, their buttocks sitting in London, will naturally put London’s interest first. Rail operators are happy to oblige because London gives the best profits. Local trains become second-class citizens which nobody in power cares.
British local transport planners do lack many powers, but they also fail to use what powers they do have. It is true that until a short while ago, local transport bodies outside London only operated trams, bus infrastructure (stations, stops etc), and in some places maintained rail stations. The most many places could do in terms of actually setting bus and train routes was nudging through the subsidy regime, though a couple of places (e.g. Brighton, Nottingham) managed to retain bus services in-house and these have been successful.
But even with these minimal powers, British local transport planning has on the whole been poor. This is proven by the existence of a few good examples like Nottingham, which did build a decent tram network, integrated (and simple) ticketing, a comprehensive and legible bus network, a workplace parking levy etc. As a result it has the second-highest bus usage in the UK despite being something like the 10th largest city.
The City Deals for more devolved powers were happening and fit with their rebalancing Britain narrative that is now in play. But Covid and Brexit consume nearly all government time now. We shall what happens next year and whether Johnson survives or the Conservatives cull him for Rishi (someone who can make a decision).
But Covid hastened the end of franchising and towards management contracts and is likely to prove the death knell of the season ticket, if home working takes off. (rail companies have been wanting to provide tickets designed for people who commute 3 or 4 days a week for ages, but annual season ticket has dominated for so long that No one in the DFT was in a hurry to change it. Before COVID, various metropolitan areas were grouping towards a bus management contract system like London but we shall see if the current crises proves a catalyst for further change.
My understanding was that city deals were more about money and (re)organisation than powers. I struggle to think of what statutory powers were granted that some level of local government didn’t already have, although I know there were a couple of minor ones.
I believe both Manchester and Newcastle now have powers to start franchising bus routes in their area. But both are approaching it very slowly. I think the first phase of Manchester is supposed to start in North Manchester. I believe the roll out will be more than a decade.
Wacky scheduling is the way all railroads operated in the bad old days. Takt timetables are an innovation that not all countries have adopted. No TGVs are to blame for the extremely weak Lisbon-Evora service (it’s about 130 km, and there are only 4 daily trains).
Yeah of course – but takts are less relevant on long-distance, high-capacity, few-connection routes like TGV (and small, cheap, nimble trains are actively harmful for longer-distance services) and I think this mindset has at least contributed to France’s poor uptake of best practice for regional rail. For the record, this isn’t the only factor; I think France’s low population density, for example, has also contributed to its poor regional rail.
Why did they do that in the bad old days?
Lack of travel demand? Lack of competition by private car?
Lack of competition and low expectations are key. Transit didn’t need to be that good to have a high number of passengers. Passengers could be treated more like cargo, moved when it was convenient for the railroad. Of course this is a bit of a caricature but you see the same effect in poor metropolises today where people do 2h bus commutes out of poverty, not because they love the service.
Perhaps centralism really does play a role? Switzerland and Germany are famously decentralized and at least historically, the Netherlands were, too.
Yeah, decentralisation means that more decision-makers are aware of the need to facilitate transfers. So that probably helps the adoption of takt scheduling.
For relatively long distance travel hourly trains are probably OK, with two-hourly being less good but just about liveable with. The main reasons for travelling by car in the UK is poor transport or car hire options at the destination and poor connections between services. For example all the trains on the west coast main line don’t stop at Stafford and perhaps Tamworth even though that would let them all connect with the cross country trains from Southampton and the South West
The French services can be much better if they just adopted the Japanese paradigm: also with HSR as backbone, but with timed transfers to regional trains and slower intercity trains that go *perpendicularly* with the HSR.
Japan is geographically dissimilar from France. Plus, France has legacy rail lines at 1435 mm, Japan at 1067 mm
Perhaps these criticisms are correct. I genuinely don’t know.
But one does need to consider that with almost identical populations England is only 24% of the area of France. Even that doesn’t really reveal another big difference, that the vast majority of England’s population are along a few linear routes (and the two main Scottish popn centres are close to the end of the two main north-south routes). Of course that is the logic behind HS2 and why it should have been built decades ago, as a single line (more or less) it links up a high fraction of the entire population.
They should never have destroyed the great central mainline
The accusation is often that there is a connection between “senseless prestige projects” on the one hand and the abandonment of branch lines “for economic reasons” on the other.
In my opinion those people have it exactly backwards. A railway with a growing and dynamic high speed sector will have an easier time getting the money and the political majorities for keeping branch lines open or even modernizing them…
Improving part of a network makes all of it stronger, not just the part that was upgraded!
By now suburban voters know that an S-Bahn tunnel benefits them a lot (maybe even more than city residents) even though all the digging is happening downtown. People in peripheral cities should understand that same dynamic with high speed rail.
A lot of the opposition to Stuttgart 21 comes from parts of Baden Württemberg that aren’t Stuttgart…
Didn’t the Greens win in Stuttgart on an anti-S21 platform? The referendum on canceling S21 went down statewide, inc. in Stuttgart, but I think in Stuttgart it lost by smaller margins than elsewhere…
https://de.m.wikipedia.org/wiki/Volksabstimmung_zu_Stuttgart_21 Stuttgart did vote against S21 more than the state average, but the only places that voted in majority against S21 have nothing to do with it…
Have you thought of making similar HSR-crayons for other continents such as South America?
Colombia is perhaps the country with the most ridiculously short but busy domestic air routes in LatAm. But it also has formidable mountains in the way.
Argentina, Paraguay and a large chunk of Brazil are all definetly flat enough for HSR
Yes, but apart from Buenos Aires-Montevideo, I don’t think international HSR would make sense in South America, the cities are too far apart. But some national networks could still make sense.
There is certainly a case for a continental freight network which would probably entail a lot of tunneling anyway.
To my knowledge there are already sleeper buses in south America, so maybe sending sleeper trains along such lines would pencil out…
If the route already exists for freight anyway, sure. But don’t expect huge volumes of passengers.
What’s the route for Buenos Aires Montivedo? A Rio de la Plata tunnel? Or jus going around?
With some quick and dirty crayoning, going around is 320 km, vs 210 km as the crow flies. It’s 220 km with a 50 km Buenos Aires-Colonia del Sacramento tunnel. I don’t think tunneling 50 km under an estuary is worth it just to save 100 km of track (25 min at HSR speeds?).
Also the 30 km or so of urban approach in the “going around” case would be free, because that could be shared with the line to Córdoba.
The ferry from BA to Montevideo can do it in 2h15, so I’m not sure if building an entirely new railway line going the long way around just to save at best 30-45 minutes on the journey time really stacks up.
I had no idea ferries could be that fast. But the time saving from HSR would be more like 50 min to 1h.
The ferries are quite fast between Montevideo and Buenos Aires and with a distance of 106 nautical-miles(196km), the 2:15 trip time is 87km/h average speed. This service is done on the LNG fueled “HSC Francisco” with a top speed of 58 knots(107km/h) and a cruise speed of 51 knots(94km/h) carrying 1024 passenger similar capacity to a 16 car HSR trainset but HSR has ~2X – 2.5X the average speed.
The ferry also has the advantage of low upfront capital cost. I can’t find a cost for the HSC Francisco, but ferries in the 100m length class range between $80 – $100 million compares to the cost of building a HSR corridor. However operating costs are huge which contributes to the high ticket prices of ~$650 average compared to similar distances on HSR where in Japan the Kodama service on the Tokaido Shinkansen does the 239km trip between Tokyo and Hamamatsu in 1:52 for ~$80 at an average speed of 128km/h and in China where the 219km trip from Beijing to Cangzhou is done in 58min(225km/h average speed) at a cost of $14, $25, and $45 for 2nd class, 1st class, and business class respectively.
Moving on water at relatively high speeds is very energy intensive. The Francisco is propelled by a pair of 22MW gas turbines for a total of 44MW in order to achieve its cruise speed meanwhile most 350km/h capable 16 car trainsets achieve that operating speed with 20MW.
It seems the operator of the Buenos Aires to Montevideo route, Buquebus is focusing on luxury as their new vessel currently under construction emphasizes that instead of the speed emphasis of the Francisco. It will cost $130 million, have room for 2,100 passengers, 3000m2 of retail floor space onboard the vessel, and a cruise speed of ~40knot(74km/h). Which suggest a longer trip time of 3:00 compared to the 2:15 of the Francisco.
Building HSR to serve the two metro areas will be a huge financial endeavor, but the infrastructure would not only save time, but provide better service to more of the metro areas especially if they have multiple stops within each city and more service can be provided instead of the current 2 times daily service from Buquebus and 3 times daily service from their competitor Colonia Express. Also it will avoid the frequent cancellations due to bad wild weather.
A direct tunnel could establish an Øresund-like binational metro area…
But isn’t that ferry just across to Colonia ie. directly opposite Buenos Aires? Then one has to take a bus for the 180km to Montevideo. At least that was the option when I did it several decades ago. I took the slower night ferry (leaves after midnight) from Colonia to Buenos Aires so I could do Montevideo as a day trip. Quite a long day trip because quite a lot of time was in the travel. I think most would choose to use a train if it was an option.
The 2h15 trip time is for the direct trip from Buenos Aires to Montevideo. I stated the distance as 196km in my earlier post, but it’s actually 225km so it looks like they operate at a higher cruise speed to achieve that time @ 100km/h average. The route to Colonia is 65km long and is served by the other slower ferries in the fleet. Colonia receives 3 daily trips with a trip time of 1:15 a 52km/h average speed.
The Colonia route is more popular because it’s cheaper, about $260 plus $12 for the bus from Colonia to Montevideo. That’s a total of $272 compared to $650 for the direct ferry. The total trip from Buenos Aires via the Colonia ferry and bus combo takes 4 – 5hrs depending on the traffic. So roughly double the time for less than half the price.
Right. That wasn’t an option back when I did it, or I’d have taken it. It was worse because they force you to take a bus to Colonia much earlier than necessary and I remember hours of waiting around at the ferry terminal.
That new ferry must be a catamaran or even a hydroplane to manage such speeds. And generally susceptible to cancellations in bad weather–it’s really open ocean there?
It’s a wave piercing Trimaran (3 hulls) with 44MW of brute force. For reference an Airbus A320 uses a max of 12MW, a Boeing 777 uses 41MW, an A380 56MW and those are things that cruise at 850km/h instead of 100km/h.
So, Greta Thunberg won’t be taking it anytime soon …
South America has about 3 times the area as Europe (minus Russia), and 4 times the area of the EU (including Turkey), but a slightly smaller population, and less than half the GDP adjusted for purchasing power. Cross-border travel is somewhat restricted, unlike in Europe. So the most reasonable places to start HSR networks would be in the larger countries which have several cities >1 million population within reach of the largest city or cities, and relatively developed economies. Brazil, Argentina, Colombia.
Argentina has an obvious choice from Buenos Aires (~15 million people) to Cordoba (~1.4 million) via Rosario (~1.3 million), total distance ~710km (440 miles) over flat land: https://en.wikipedia.org/wiki/Buenos_Aires–Rosario–Córdoba_high-speed_railway
Other proposed extensions to Mendoza in the west and Mar del Plata to the southeast would be harder to justify from ridership, but might work due to network effects? Unfortunately Argentina has been in fiscal crises for years.
Brazil has the largest cities, most near the east Coast, though unfortunately not all are in a straight line and there are many hills.
São Paulo – Rio de Janeiro is a slam dunk: 403 kilometres (250 mi) between 2 urban populations of >11 or >12 million people each. The initial segment was planned to include Campinas (>3 million), close to Sao Paulo. https://en.wikipedia.org/wiki/Rio–São_Paulo_high-speed_rail – Further lines to Curitiba (3 million people), Belo Horizonte (5 million) look feasible, and a line connecting to Uberlandia and Ribeirao Preto is possible if ridership to Sao Paulo is high enough. There’s also a proposal to connect Brasilia to Goiania which is marginal since it would be difficult to connect to the other cities, and the two cities only are moderately large.
Colombia looks like it might have some options to connect Bogota to nearby cities, but the mountains in the way are significant: Bogota to Medellín (second largest city) is ~400 km but across 2 ranges of the Andes, with few cities in between. A route to Ibagué and Calí faces the same issues.
Venezuela could have a feasible route from Maracaibo to Caracas via Barquisimeto and Valencia, but the country has political and economic crises.
Chile has large distances between cities with rather small populations and some mountains even in the central valley south of Santiago, but there are plans for a short route (120 km) to Valparaíso: https://www.infrapppworld.com/project/santiago-valparaiso-high-speed-rail-ppp-project – but not to Concepción, 500 km south.
The other South American countries do not have strong enough economies or clearly good routes within the country, at first glance.
Further north, Mexico has a good potential route from Guadalajara to the capital, but the terrain is difficult. The first 210 km from Mexico City to Querétaro was being considered: https://en.wikipedia.org/wiki/High-speed_rail_in_Mexico
A pretty good first order approximation of potential is the list of busiest air routes in the world. Colombia and Brazil are immediately obvious from that, but as mentioned, the mountains are quite the obstacle…
In the Rio-São Paulo route the mountains aren’t that bad. The existing road follows the valley most of the way. The one tricky part is the Serra das Araras close to Rio, where the road drops some 500 m of elevation in 13 km full of twists and turns. This is I think the only place where you’d need some sort of tunnel or viaduct for the HS line.
Tunnels are good! They’re obviously expensive, but they also provide the best speed up! As you point out the Dutra is pretty straight between Rio in São Paulo except for the Climb to Piraí so the difficulty would be on the approaches. A more interesting question IMO is how do you build the low speed transport network to compliment this. Resende, Barra Mansa, and Barra do Piraí aren probably not getting HSR stops, but good bus network coordination, or low speed rail build out focused on Volta Redonda, would probably compliment the HSR network nicely.
Mar del Plata is not as big as Córdoba and Rosario, but it has the advantage that there’s an extremely straight existing ROW (there’s literally 125 km of straight track between Mar del Plata and Maipú), which should make the upgrade pretty cheap.
On the opposite side, Quito-Guayaquil is a good candidate in terms of distance (~300 km) and population (3M + 3M), but Quito’s extreme altitude is very challenging: there’s only 80 km between Santo Domingo (550 m above sea level) and Quito (2850 m), so there’d be no possibility for sharing the line with freight.
In Mexico, it could make sense to build in stages, since there’s a legacy network: first from DF to Querétaro as planned, then to León (also serving the intermediate medium-sized cities, and finally the branches to Guadalajara and San Luis Potosí. Perhaps you could then have sleeper trains from the capital to Monterrey…
Google maps says it’s 900 road kilometers from Mexico City to Monterray. Depart at midnight you’d get there at 4. And then the train sits around for 20 hours not earning any money.
Mexico City-Monterrey is 900km on the current road, slightly shorter on a new ROW. Viable for HSR.
Guayaquil-Quito needs a ~15km tunnel just west of Quito, but possibly that is the only significant obstacle (eyeballing it).
Brazil has two obvious good lines in the Southeast. One going N-S in São Paulo between Uberlandia and Santos, and the other going E-W between Rio and Curitiba. After that you extend to Belo Horizonte. You can use Alon’s Metcalfe’s Law blog post to do some rough calculations, and you find that’s a pretty good network. There’s maybe one or two more extensions north (Campos, Vitoria) or south (Joinville, Balneário Camboriú, Florianópolis), but these are more marginal.
My one quibble is that it’s quite a long way to Uberlandia, even if you start from Campinas. A branch from the Rio-São Paulo mainline going to Belo Horizonte via Juiz de Fora and Ouro Preto would actually be much shorter, and it would make sense to build it before any extension beyond Campinas. BH is bigger than the sum of all the small/medium cities after Campinas.
That’s definitely true. My initial thoughts for the continued northward extension were 1.there’s probably a bonus to in state travel like there is in country travel. 2. It’s easy to build and the state of São Paulo could probably build the whole thing itself until Riberão Preto.
On the old abandoned railway system, Rio-Belo Horizonte was the mainline (and originally intended to go all the way up to Belém) and Rio-São Paulo was actually the branch line!
Brazil’s population being mostly concentrated in coastal cities does mean that there’s a *tonne* of cities along a single North-South line. Starting from the Rio-São Paulo route, Metcalfe’s law gives us 0.3*(22^0.8)*(13^0.8), or 27.68 mi pax/year; already higher than any individual city-pair in the US. However, we should also take into account intermediary cities, since there’s a pretty high population density along the entire corredor: Assuming regional lines feeding into Volta Redonda and São José dos Campos, we get the following passenger figures:
Rio de Janeiro-Volta Redonda: 0.3*(13^0.8)*(1^0.8) = 2.33 mi pax/year
Rio de Janeiro-SJC: 0.3*(13^0.8)*(3^0.8) = 5.62 mi pax/year
São Paulo-Volta Redonda: 0.3*(22^0.8)*(1^0.8) = 3.56 mi pax/year
São Paulo-SJC: 0.3*(22^0.8)*(3^0.8) = 8.57 mi pax/year
For a total of 47.76 mi pax/year between Rio and São Paulo alone. Going south, we go through Sorocaba (2.1), then Curitiba (3.6), Joinville (1.4), Florianópolis (1.2) and Porto Alegre (4.3). Beyond this the only real option is to cross the border into Uruguay and head straight to Montevideo (1.9) and even to Buenos Aires (16).
Using the formula, on a ~2684 km line from Rio de Janeiro to Buenos Aires, the annual ridership is projected at 57.31 million; if the line stops at Porto Alegre (1661 km from Rio), the annual ridership is projected at 44.63 million. Similar numbers should be found going north from Rio up towards Recife and Natal.
*Forgot to take into account passengers on the Rio-São Paulo route for the Rio-BA/Porto Alegre route; annual ridership from Rio to Buenos Aires would actually be projected at 105.07 million, and from Rio to Porto Alegre would be 92.39 million
It’s poor when compared to the rest of the line, sure, but Porto Alegre is still a pretty strong node and generates noticeable ridership even considering its distance. Buenos Aires is a very long way away though, the only reason I left it in at all is because it’s such a huge and important city that it still manages to break over 2 million yearly passengers to Porto Alegre despite being over 1000 km away.
This is why going actually beyond Curitiba is a pretty poor investment all things considered. Your route is 2684 km in length, and it produces 92.39 million riders. Going Curitiba -> Rio and Santos -> Franca gets 136 million riders. The line from Santos to Franca is 398 km. Rio to Curitiba is around 900 km. So for half the distance you get 50% more riders.
Ugh, I just saw now that the spamfilter got overly aggressive with a previous version of this comment with a link to a full list of city pairs in Brazil. Here is the link:
https://pastebin.pl/view/a35ba894
Colombia, despite the cost, seems like it could be worth it. The tunneling would get crazy but a basic T system connecting Medellin, Cali, and Bogota would have enormous ridership/km. I’d like to see a closer look at it. As a starter, just Medellin-Cali would get decent ridership without the tricky infrastructure of connecting to Bogota.
Gauge breaks are a way bigger problem in South America at border crossings than they are on any other continent. . .
Compatible border crossings:
Argentina || Bolivia: both Metre Gauge (1000 mm)
Argentina || Chile (2 of 3 crossings): both Indian Gauge (1676 mm) — Third out-of-service crossing built to Metre Gauge to accommodate rack railway segment, connects via transfer to Indian Gauge on each side.
Argentina || Paraguay: both Standard Gauge (1435 mm)
Argentina || Uruguay: both Standard Gauge (1435 mm)
Brazil || Bolivia: both Metre Gauge (1000 mm)
Chile || Bolivia: both Standard Gauge (1435 mm)
Chile || Peru: both Standard Gauge (1435 mm) — out-of-service.
Incompatible border crossings:
Argentina: Standard Gauge (1435 mm) || Brazil: Metre Gauge (1000 mm)
Brazil: Metre Gauge (1000 mm) || Uruguay: Standard Gauge (1435 mm)
Columbia: 3 Foot Gauge (914 mm) || Venezuela: Standard Gauge (1435 mm) — out-of-service
Peru: Standard Gauge (1435 mm) || Bolivia: Metre Gauge (1000 mm)
All specific country pairs not listed have never had any border-crossing rail links to begin with, but the gauge-break spaghetti gets even more complicated from here when tallying up differences in the yet-to-be-proposed crossings needed to infill a high-functioning intercity/international crayon map. Some countries, like Argentina, aren’t even internally unigauged nearly well enough to string together major intercity connections between the prospective border crossings. So overall a total crapshoot compared to Europe or the Indian Subcontinent where breaks-in-gauge have certain economies of scale and logic to them by clumping en masse to certain borders (e.g. ex- Eastern Bloc borders being the big Standard vs. Russian dividing line, France-Spain + Pyrenees being the Standard vs. Iberian division, newly Standard-ized India vs. Russian clumping at the ex- Soviet influenced borders, and so on). Even Africa by stroke of luck tends to clump its gauge breaks a lot more intuitively (i.e. along larger-scale geography + ex-colonial contours) than South America, unfortunately.
Add to that the fact that so many Narrow Gauge lines of varies intra-Narrow incompatibilities still malinger on this continent on a number of key spanning routes for drawing up functional crayon-intercity networks and you’ve also got very divergent physical differences in the gauges themselves complicating the scope of what Unigauging efforts need to be mounted. For unigauging, Indian vs. Standard isn’t that big a physical difference so several of these countries are following India’s lead and studying a purge of 1676 mm lines in favor of 1435 mm. India’s changeover sort of dead-ends 1676 on the global rolling stock + railway engineering supply chains by virtue of sunsetting the world’s largest 1676 network, so 1435 gets flat-out easier to long-term support even in some countries (like Argentina) where Indian mileage substantially outnumbers Standard mileage. But it’s a much larger physical difference Metre vs. Standard/Indian, and Metre is a closer unigauge reach for uprating the enormous hodgepodge of unstandardized Narrow Gauges in largely poorer and geographically-challenged regions. So even parsing between a choice of Metre vs. Standard unigauging to clump the breaks on a crayon map to wider, more supportable scale like Euroland and the Subcontinent faces manifold difficulties and complexities in South America not found to anywhere close that degree anywhere else in the world.
Brazil has de facto adopted the 1600 mm gauge as the standard, the only 1435 mm gauge lines are a couple metro and light rail lines. 1000 mm is technically more common throughout the country, but for the most part 1000 mm gauge railways tend to be older, little-used freight lines that don’t really have any relevance to a modern passenger rail system, and I suspect the same is true of Argentina with 1000 mm/1676 mm gauge – And considering the lack of population density and general inhospitality of border crossings elsewhere, most breaks of gauge would be between the Brazilian and Argentinian railway systems, and therefore between 1600 mm and 1676 mm. Paraguay also has no railway system and the crossing into Argentina is just a tourist trap calling itself an international train. Breaks of gauge in general are also a lot less of a problem than they used to be, with modern equipment you can change a train’s gauge within minutes
“Breaks of gauge in general are also a lot less of a problem than they used to be, with modern equipment you can change a train’s gauge within minutes”
Not a blanket statement, because entirely dependent on **how big** a gauge break you’re covering. Metre-to-Indian is functionally a bridge way too far for integrated equipment. 40% difference is punishing cost/complexity premium vs. a 4.6% difference in Brazil 1600 mm vs. Argentinian 1676 mm and 6.6% difference in world-most-common 1435 mm/Standard vs. 1520 mm/Russian gauge breaks. It’s far too big a leap to justify the kludges.
Completely different universe from the largest-scale bulk gauge breaks we find between Western/Eastern Europe and the northwestern border of the Indian Subcontinent. While there are many Standard-to-Russian gauge-changing thru freights at those big breaks, even at one-digit percentage point differences it’s still enough of a premium to sharply thin the herd on the passenger side and force transfers at the breaks on majority of international trains. 1668 mm/Iberian vs. 1435 mm/Standard at the Pyrenees is an even bigger gulf at 14% difference, with only one Standard Gauge-only border-crosser running thru the French-Spanish border on the Madrid-Leon HSR line. Spain unilaterally swallowed the premiums for interoperability by bringing Standard gauge inside of Iberian territory and installing all the gauge-change stations at the junctions between the new line and any legacy Iberian lines…at level of above-and-beyond investment that upstart countries are simply not going to have the in-house means to swing for new service starts. And that’s largely been a passenger-only benefit (albeit substantial) to them; the Pyrenees break is still a major freight bottleneck imposed on the Iberian Penninsula economies.
But again…in South America we’re talking of a crayon-draw service universe here because there isn’t much of a starting map to work with. The extant border crossings are largely non-useful, and enough from-scratch corridors have to be created to facilitate intercity and international travel in the first place that it’s not reliable supposition that the resulting crayon draws are going to contour safely away from problematic gauge breaks. You’re going to encounter lengths of those “little-used” metre-gauge lines that need to be folded into segments of the new networks in between segments of virgin ROW, and in some cases where mountain geography comes into play at fashioning links the extra-problematic backwater Narrow Gauge lines have to be dealt with. It’s not a Europe or India analogue…the breaks don’t clump to enormous-scale geographic or geopolitical borders like the non-problematic Standard vs. Russian or semi-problematic Standard vs. Iberian megaborders. It’s a born hodgepodge with breaks distributed by chance across the map. Virtually all South American intercity crayon networks except for the ones that self-distort to diminishing returns in avoidance of gauge breaks are going to come face to face with multiple problem breaks as base cost of doing business. Breaks that are way more individual/situational in nature than the megastructures that give Standard-Russian or Standard-Iberian interops some measure of alright-to-excellent overriding scale. And breaks that in many of these situational cases are physically way wider (in the case of Metre vs. Standard/Indian) than applied interops can feasibly bridge, requiring tough binary decisions to be made on which lines get changed to what gauge at what cascading effects to their networks.
There’s going to have to be some brutally-tough decisions made on unigauging or “semi”-gauging to deliver a continental canvas that can take the crayon-draw networks. And/or the crayon-draw networks will need to rationalize their route choices in a very different way than we approach on other continents. Or likely some combination of both. The interoperability considerations don’t have an analogue with the other hemispheres of rail-on-Earth and can’t easily be pounded flat to act the same way for the same apportionment of resources. South America probably doesn’t need its own wholly separate rulebook on how to crayon-network, but it does need some sort of corollary twist on “Build to ___ Costs” to properly account for how large-by-share the gauge break complexity looms over there vs. anywhere else in the world.
IMO the gauge breaks aren’t that big of an issue, because outside of Buenos Aires-Montevideo, I can’t think of a single South American border that has two larger cities near enough to it to justify HSR in the short term to medium term (now-30 years in the future.)
There’s only 3.5 million people in all of Uruguay. They can fly to Argentina
The ”brutally-tough decisions” to me at least just look like no-brainers – re-gauge the metric lines to broad gauge where applicable, build new freight bypasses where applicable and accept the break-of-gauge between Brazil and Argentina. Most of the narrow-gauge lines, and *especially* the non-metric-gauge ones, are in such bad shape (i.e. single-tracked, unsignalled, 10-20 km/h speed limit) that they’re only really useful for the right-of-way itself, and in some cases not even that. The only real *problem* that I can see wrt changing gauges is Uruguay, since it uses 1435 mm gauge despite being surrounded by two huge countries which have different standards – And even then, none of its rail lines are suited for HSR (Though regional rail crayons from Rio Grande do Sul and Entre Rios could probably take a hit). It’s worth noting that urban rail systems in Brazil, Argentina and Chile for the most part use broad gauge, especially the larger systems like São Paulo and Buenos Aires, which makes it a lot more worth it to avoid narrow and even standard gauge.
why does it matter what the subway or the trolley cars do? are they going to shipping grain or livestock to market on subway cars?
2 million people connected to 15 million in Buenos Aires by rail within 300 km there isn’t really a distance penalty is at least plausible in most places I’d think, despite your insistence otherwise.
It matters what urban rail does because a lot of it (including São Paulo’s CPTM, Rio’s Supervia and Porto Alegre’s Metro) is operated on old long-distance rail lines – The CPTM and formerly the Supervia even have/had cargo trains running on them at night! Ignoring them would be like the TGV ignoring the Transilien
Berlin has a similar number of inhabitants
Another example of fantastic shortsightedness is Australia. Despite not building their first railway before the UK gauge unification legislation of the mid-19th century, no state shares a gauge with any neighbouring state. For 5 of the mainland states none shares a gauge with their three neighbours, and, remarkably, South Australia doesn’t share a gauge with any of its 5 neighbours!
Wiki:
(That excludes 4,000 km of 610m track that serves the sugar-cane industry in Queensland.)
.png)
Unification of gauges was official policy since 1901 (federation) but all capitals were only linked by standard gauge 94 years later.
Why are you saying that India is “standard-izing” or “ sunsetting the world’s largest 1676 network”? Project Unigauge has been a program to convert all narrow gauge railways in India to 1676mm, not to convert 1676mm to standard gauge. As of March of this year the Indian railway network was 93% broad gauge and conversion to 1676mm continues on many routes.
https://en.m.wikipedia.org/wiki/Project_Unigauge
You often say “assume Nordic costs”. Would “Nordic costs” let a Stockholm-Helsinki tunnel (via Åland) pencil out?
Probably not? These are not big cities. Finland is only now moving on a Helsinki-Turku line that’s only 10% in tunnel.
But the Nordic countries enjoy strong links and high incomes. Between Helsinki and Tallinn alone, ferries carry six million pax every year. And the Stockholm-Helsinki route is also very popular. If you let the train stop in Mariehamn, you can even finagle in duty free booze…
Stockholm is the busiest air route out of Helsinki at 1.4 mio pax (2019) which is however lower than some Norwegian domestic routes…
Oh, they have strong links, but at the end of the day, it’s a very long underwater tunnel for two metropolitan areas that combine to have slightly more people than Berlin.
You only need to get to Åland by tunnel. The archipelago sea has islands and skerries close enough to each other that it’s basically an elevated railway line more than a bridge…
And I think the NIMBYs can be placated with Mariehamn service, which is where 40% of Ålanders live…
Interestingly the busiest domestic connection from Arlanda airport is to Luleå which seems right on the edge of hsr range…
The problem is that Luleå is a village. Stockholm-Luleå HSR would get excellent modal split, destroying the 1 million/year air market and inducing more travel, but at the end of the day, 2-3 million/year isn’t what successful HSR looks like. Contrast this with Stockholm-Gothenburg, which is 900,000/year by air but also has a lot of driving and legacy rail, so HSR would get somewhat more traffic and could be a success with a Y-line to Malmö and Copenhagen and with Swedish construction costs.
Where would you locate the neck of the Y. Jonkoping, Falkoping, or Nassjo?
On a related note it’s wild that Malmö has no urban rail but Lund is opening a tram. What’s your guess? The Lund tram being extended to Malmö or “Øresundmetro”?
Also, while Luleå is probably not strong enough to justify a line, what will Sweden do after it has built a line to Gothenburg? There’s little left after that that doesn’t involve crossing big bodies of water
Malmö has Citytunneln, but I don’t know if it’s used as a real S-Bahn or not.
After the Stockholm-Gothenburg-Malmö Y, I guess the next link will be to Oslo. The Greens promised to study this option in the 2014 election.
Will Norway ever get off its asses and build hsr?
And what about a direct tunnel or bridge from Rügen to Southern Sweden? With such a connection, even Stockholm-Berlin becomes competitive and it is an attractive freight route…
They seem to be considering the possibility of extending Rail Baltica with a ~50 km tunnel between Tallinn and Helsinki. It would cost a lot of money (current estimate is 9-13 billion €) and the population served isn’t huge, but as international connections go it’s strong (low language barrier) and there’s also a strategic interest in allowing rail freight to bypass Russia…
Six million ferry passengers each year are nothing to scoff at… Plus it’d be one more step in making the Baltic an EU lake
How many of those 6 million passengers are just part of the booze fueled love boat experience making the round trip on the boat without actually crossing? If rail service is available, theres a high likelihood that those customers will not port over. But rail rail connection from Helsinki to the EU will definitely benefit from through traffic.
I’m not sure booze on a boat from EU member Finland to EU member Estonia is cheaper than booze inside EU member Estonia. The cheap booze on the other routes is due to technically leaving the EU VAT area by serving Åland…
Also, a reliable all weather crossing between cities in countries with similar languages will induce commuting, just look at Øresund…
Finland is not only EU member but also a nordic country, and alcohol tends to be really expensive there.
But you can always divide your train into a drinking and a non-drinking section, like the old division into smoking and non-smoking.
Or you run dedicated party trains, like a party tram.
Yes, but before the EU put the kibosh on it, you could get duty free sales simply by sailing out a few nm. Nite you have to actually drop anchor and allow people to board/alight at a place outside the EU customs territory. So in the case of the Baltic, that’s Åland (or in theory Russia, but that complicates immigration and ID checks needlessly). Estonia of course has lower taxes on booze than does Finland. I don’t know which tax regime applies for a boat between Estonia and Finland (or a train for that matter). Flag country? Port of departure? Does it change halfway through the trip?