San Francisco-to-L.A. overnight train inches closer to reality
ktla.com
ktla.com
It's convenient and depending on your distance to an airport, only slightly longer, but much more convenient. I'm a big fan.
That said, having successful train systems in the US is hard. The US is a huge country with a lot of empty space. If it's a regional train in dense areas, it can make sense (like the NE regional). Maybe the traffic of LA <-> SF makes sense, and if it does and it can be done cost effectively, then it'll work out.
My assumption is that after building all the infrastructure into place, the cost of the tickets plus the additional travel time will not be worth it to most who are making this trip. I would love to be wrong, but I'm not counting on it.
I love the experience of taking a train, and I take them a good bit here in the North East. That said, they're not the best solution in the US currently, it's it's hard to see a world where they become the best for travel outside commuter rails (the MBTA commuter rail is also fantastic) and dense areas (again, like the NE regional).
Trains are not a catch all solution in the US. I hope for them to grow in popularity when they're appropriate, but I see a ton of online discourse (granted, it's likely from a vocal minority) about how trains are always better. They're not, and you can't compare US rail logistics to many regions to Europe.
If they get the beds and service right, it would be because it’s convenient and comfortable. Like, have a Michelin-class meal, good wine, reliable service that knows the regulars and comfortable beds and good shock absorbers and you’re trumping planes as the first-class option.
It doesn’t have to beat flying in all cases to be successful. Price always seemed like a weird niche for passenger rail to aim for—planes don’t have to maintain track!
The rewards on a Bay Area-Los Angeles flight are de minimis if you’re a frequent flier. I am one, and I was planning on taking the Coastal Starlight up next month until I learned—in this thread—that it’s kaput for the time being.
Well, they have to maintain airports, but there's a lot of federal funding that goes into that too. It's always interesting to see where the costs of infrastructure are borne. Rail seems to be one of the few where the carriers (and consumers) are almost entirely directly responsible for the infrastructure. (Admittedly, citation needed.)
I feel like even if we were to do it, folks would sabotage it for politics here. It's tiring.
China has the same problem... they simply made it a priority to build the infrastructure.
There is nothing stopping us from doing the same thing.
China has almost four times the population density of America’s (146 per square kilometre to 37).
The USA is a bit better, 25% live in the western half (mostly on the west coast though).
Well what's stopping us is:
1. prioritization of private property ownership over every other possible measure of societal health, people can drag out these projects fighting over the land in court for decades
2. government cronyism isn't even organized enough to stomp out the issues with #1 using eminent domain, because all the small fractious governments of states and local municipalities are working against each other to each win the cronyism battle. at least in more centralized governments, the cronysim is organized enough to get the project done while skimming 75% of the money off the top.
However achieving property rights for the land needed for the project is a much bigger issue in the US than in China, where I doubt they receive much dissidence from people affected due to their infrastructure improvements.
If you look at China's HSR map, they have exactly one line that goes into the western half of the country (from Lanzhou to Urumuqi). It only takes 9 hours, only a few stops in between.
No reason these couldn't be linked up and are more populous than many equivalent regions with HSR. It's like power grids, noone is arguing that one big one from SF to NY is the right solution.
The problem is that it’s not, at least from the standpoint of property rights covering that land. For example, the path from San Francisco to LA is relatively densely populated. And on top of that, the US has an intractable legal system for building infrastructure.
https://www.youtube.com/watch?v=wE5G1kTndI4
His video description: "Nothing that exists in North America fits the definition of true high-speed rail. If we actually ever started building a network, where would we start? And how would we know when we were done? This video details a methodology for answering all those questions, then counts up 56 city pairs we really should've had high-speed rail service between long, long ago."
That analysis makes several idealistic assumptions, e.g. frictionless international borders. It’s a good starting point, but his other videos analysing specific corridors seem more grounded.
As someone who did a lot of SF / LA travel for sales and marketing events the only way to get there early and be fresh for the presentations was to leave the night before on a flight ($100), stay at the airport hotel ($100-200) and then uber to the meeting / conference in the morning (always far away from LAX $50).
So $350 is not crazy if you have high quality accommodations and I'm sure it could be much cheaper making this a very compelling option for both business and leisure travel even if it's slow!
BNSF has zero interest in upgrading the track to handle higher speeds. That would not help operating ratio even if it let them move more freight and charge higher rents. For US railroads operating ratio is God.
The funny part is high-speed passenger service is usually very profitable. The trips are so short you don't need to turnover or time out crews. A lot of routes you don't even need to put them up in a hotel, they go out and come back. That keeps labor costs very low.
For the curious I once ran the numbers on China high-speed rail vs California's HSR project in terms of miles of HS track. A rough estimate was about 3x the cost to do the project here in the US. Given wages and most importantly property values in California that actually isn't out of line. China has the same vast distances across low density areas as the USA does so you can't blame it on Europe and Japan being higher density. Because they cover large distances on some routes they pay (all else equal) more than others do. China has simply invested more in their rail infrastructure because they see it as a priority.
FWIW The USA used to do those kinds of investments. It was government-backstops of investments that allowed the Transcontinental Railroad project to happen in the first place. Plus railroad land grants (free land if you build a railroad).
They also have much easier procurement. It's easy to build a railway line in China when local municipalities can rezone entire swathes of farm and urban land with minimal legal recourse.
The high speed line through the central valley makes more sense when you consider these limitations. If they're able to speed things up enough then it does become competitive with air travel - but this is definitely a case of "time = money". It's much easier for higher-paid individuals to just take a a short flight for day trips to LA/SF than sit in a train or traffic.
No. The unique trackage to be used between SFBA and LA is the "Coast Line" which is owned by Union Pacific.
FTA: "Dreamstar signed a Memorandum of Understanding (MOU) with Union Pacific Railroad, [...] that would allow Dreamstar to operate its trains on the tracks owned by Union Pacific."
Ctrl+F "BNSF" finds no matches in the article.
Coast Starlight uses both BNSF and UP track. Nevertheless the point stands: neither has any intention of making the track better for passenger service. Expect frequent delays and long trip times - far longer than driving.
I still hope that CA's HSR project will eventually complete the line and give us actual high-speed service. They seem to be focusing on building the middle segment because they can get it done quicker and easier, which turns the north SF and south LA extensions into a sunk cost problem: "well we already build HSR through the central valley, it makes no sense and is super embarassing not to finish it...". All told that's probably the best strategy.
There is only one HSR line that has been built in the western half of China, a line from Lanzhou to Urumuqi. It is pretty devoid of rails of any kind on that side. It was a huge technical achievement when they were able to build just a normal rail to Lhasa.
There are several regions of the US where HSR could be built with similar density to China which itself has lower density than some other countries yet spent the money to build HSR anyway.
There is a finite pool of money that can be spent on infrastructure.
Investing on enhancing and integrating the Capitol Corridor, ACE, Caltrains, BART, Muni, and VTA into a unified system has a better RoI than a 400-500mi rail line that will most likely have a train leaving on an hourly cadence.
It's a similar story for Los Angeles and San Diego.
Flights cost around $100-200 for a round trip and take around 3 hours even factoring in security and commute to and from the airport. It would be the same length of time commuting from a railway station to the other, the railway would take 2 additional hours, and the price would be roughly comparable.
> There is a finite pool of money that can be spent on infrastructure.
Exactly. The metric should be passenger-miles, nothing else. That is the most useful metric.
If spending $100B on SF to LA (400mi) impacts 1M passengers per year, then that's 400M passenger-miles, or 250 dollars per pax-mile.
But if spending $5B on (say) SF to SJ (50mi) impacts 1M passengers per year, which is 50M passenger-miles, then it is 100 dollars per pax-mile, a much better tradeoff.
San Jose to SF (Caltrain) has a yearly ridership of 5.5 Million.
Los Angeles to SF has a yearly ridership of 2.2 Million.
I don't think you've ever been to California, or at least haven't had to commute recently.
I live in California, chief. You know what they say about making ASSumptions?
And I was using round numbers to make the calculations easier and the point easier to understand.
They shouldn't though, it's absurd that they are that way in the US.
Airplanes are just the most cost-effective, reliable, and fastest method to get around for long-distance travel in the US. Train service is often so bad that people would rather drive their cars and sit in traffic for long periods of time, because train travel is still worse.
Make train travel cheaper, reliable, comfortable, and to places that people actually want to go, and people will take them. Of course some of these problems are much harder to fix than others.
> Make train travel cheaper, reliable, comfortable, and to places that people actually want to go, and people will take them. Of course some of these problems are much harder to fix than others.
How?
The closest city to the Bay Area is Sacramento which is 90mi/150km - which is the maximum distance that a commuter rail system can work between two metropolitan areas while still being cheaper than flying.
San Francisco to Los Angeles is 620-700km - so from Cologne to Prague.
Unlike that route, you'd need to be building through multiple mountain ranges in California AND there are very few cities in between to justify ridership.
There is no similar sized city in between the Bay Area and LA MSA.
This impacts the incentives to build in the US.
Metropolitan Areas in the Western United States are VERY far from each other. And this isn't because of sprawl.
Take a look at a satellite map of the Western US sometime.
Stops (and Population) on Tokaido Shinkansen between Tokyo and Kyoto: Yokohama: 3mm+ Odawara: 200k+ Hamamatsu: 700k+ Shizuoka: 600k+ Nagoya: 2mm+
train 1:30hs, €109
flight 1:00hs (+ check in time and baggage pick up), ~$300
driving 3:30hs
https://www.google.com/maps/dir/Brussels,+Belgium/Paris,+Fra...
San Francisco - Sacramento (~85 miles):
train: 2:30hs, $33
driving: 1:30hs
Furthermore, the Capitol Corridor doesn't pick you up from SF. You need to commute to the Amtrak station in Richmond.
Bay Area to Sacramento is the farthest distance that can be financial feasible in the Western United States.
A better use of money would be better integrating all of the Bay Area and San Joaquin Valley into a single high frequency system (think ACE on steroids).
It is not so much generically “very intermittent” as optimized for Sacramento to East Bay commuting, with reduced traffic on the extended connection to San Jose, and in the counter-commute direction and off-commute times.
I've literally had to wait for Capitol Corridor trains for an hour because my Bart ride was delayed.
This is why money is better spent integrating the systems at the subregional level instead.
I'm baffled that there are large American airports that have no connectivity to public transport at all. Just improving that would already be significant. That doesn't mean that rail isn't worthwhile to pursue: SF to LA makes sense (mountain ranges are not necessarily relevant, that's the reason it goes the "long way" through the central valley, lack of density in between isn't as relevant because HSR has very infrequent stops in between destinations, having a handful of stops in the central valley would be enough to cater to those low population centers that don't wish to drive the full way), but if when you arrive to LA you have to rent a car to get around, that reduces the benefit or rail.
These are easy. They even both take Clipper.
> money is better spent integrating the systems at the subregional level
Totally agree. Even just making San Francisco better traversed by public transit than car would be a blessing.
Timing is an issue. I'd say 40% of the time I ever took Caltrain, I missed the Bart connection at Millbrae, adding either 25 minutes to the commute, or adding an additional $15-20 dollars.
Also pricing isn't integrated. While they all take Clipper, funds are not coalesced together. I cannot use clipper cash to pay for a caltrain ride and vice versa.
Clipper is like a crappy Oyster knockoff.
You absolutely can use cash on your clipper card to pay for a one way caltrain ride.
Likewise Bart and Muni are integrated in SF. My muni monthly pass includes bart rides in the city.
The Golden Gate Bridge (San Francisco to Marin) does, in fact, have sidewalks.
https://www.goldengate.org/bridge/visiting-the-bridge/bikes-...
It’s true there is no pedestrian/bike connection from San Francisco to Oakland, and that seems unlikely to change (the main improvement planned there involves electric ferry service from SF to Treasure Island connecting to the East Span of the Bay Bridge.
https://mtc.ca.gov/planning/transportation/bicycle-pedestria...
> You need a car or to find a bus that will take you to a rail line.
Or you can take a train that takes you to the other rail line (in the East Bay, but then there’s not much rail to connect to in Marin anyway); BART is a thing.
Rail systems work in Europe because European cities have decent local transit/walkability. Rail systems are challenging in the US because the US invested heavily in destroying local transit and walkability for a mixture of reasons (auto industry influence, deliberate attempts at de facto racial segregation, especially in the wake of pressures against de jure segregation, and some other reasons), so once you get somewhere by rail you can’t get anywhere else, unless the train station wastes a whole lot of space – and airports already do, but its less of a burden for them because airports need to be pretty sprawling to start with — having giant parking lots for round trip passengers and rental cars and for access to serve as giant road transport (bus, taxi, rideshare) hubs.
Which is why you can’t just build long-range passenger rail in the US, you have to also build or significantly upgrade local transit networks at the termini and intermediate stops and also deal with the local development patterns in all those places (you don’t have to do it all at once, but if you don’t, you have to accept that you aren’t getting all the benefits until you do.)
And you are going to be fought the whole way by forces that recognize the connections and oppose any part of it, for any of the reasons the status quo exists as it does.
When I travel from Brussels to Berlin (similar to SF to LA), I can pass through Cologne (pop 3.3m), Frankfurt (pop 770k), Hannover (pop 500k), and Hamburg (pop 5.5m).
When I travel from the Bay Area to Greater Los Angeles, I can only pass through Fresno (pop 700k) and Bakersfield (pop 500k)
This is the issue. There isn't enough population between metros in the Western United States to support the financial burden of managing such a system.
If you go by air, first you need to leave the city to go to the airport, hand off your luggage to a stranger with the promise it will be reunited with you, go through dense security, wait an hour for your plane to arrive, funnel through a single door, sit in a tiny cramped seat, then wait for the whole plane to exit through a single door, go find your luggage, then find ground transportation to the city center.
Any trip that involves air travel instantly adds 3 hours to the process. More if you are planning to go to the airport during rush hour.
I show up 30 minutes before boarding at my small American airport. Add 20 minutes for travel to the airport and 45 for getting out and you’re sitting comfortably at 90 minutes + flight time door to door to anywhere in the world.
In Europe, I agree, it was closed to 90 minutes pre and post flight plus travel time to and from the airport.
I show up to JFK about 45 minutes before boarding in the early weekday AM, or, if I have the misfortune of having to check a bag, about 75 minutes before. (Same for SFO, PHX and IAD.) American airports are far more efficient, for frequent fliers, than European ones are for Europeans.
For good measure. They get it right. If you're rich in Europe, take the train.
[0] American Airlines at my small airport, for two early-morning flights to different destinations, had all checked baggage going through one agent whose intervention was required for every. single. bag. even if checked in with a kiosk. Of course there was no way we could get the bags on, so we were going to miss the flight. But we went through TSA (both Global Entry/Precheck) with obviously too-big bags and liquids, and they said "no way that's a carry-on", and we said we were going to have the gate agent check it for us, and they said good luck.
We made the flight.
Bullshit.
I went to school in Boston. To get a one way like that I'd have to book weeks in advanced and it would be in the middle of the night.
I'm on the Amtrak website right now and if I book a one way trip from South Station (boston) to Penn Station (NYC) for today, it is $250 (tbf, it would be a last minute booking). If I book one for weeks in advance, it's still at $40-100.
May 7, 7:50pm-12:15am
$20
[X] Weeks in advance
[ ] Middle of the nightIt's cheaper, and I take it because it's a better experience in a lot of ways, but it's also regularly not full. I've had entire train cars to myself before.
I commute on along the North East Corridor, and if you want a train that'll get you into NY by around 9am or leaves around 5-6pm, you're definitely not getting that for $15 on a weekday.
This isn’t the high speed train which is what I thought at first reading the headline. That’d require new tracks.
This is just an overnight Amtrack that’ll inevitably get delayed like Amtrack always is and you’ll wake up not in San Francisco but in the middle of the Gilroy garlic festival.
If they strike an agreement with the track owner, I don’t see why that would be. Amtrak failed because nobody enforces the part of the law that gives passenger trains primacy.
That actually makes some sense. The railroad plans a scheduled hole in the flow of freight trains that they can put a passenger train in. (Passenger trains are faster than many freight trains, so it takes some juggling to make it work.)
If the passenger train is late, the planned spot to put it in isn't there. Now they have to create a spot for it on the fly. That isn't always possible (or possible at an acceptable cost) to the freight railroad.
https://www.smartcitiesdive.com/news/california-storms-sever...
I know this was meant as a humorous example of where one might end up because of a delay, but I can think of a million worse places to be stuck at than Gilroy. Better that than somewhere like Hanford, King City, Barstow, or even Bakersfield.
Oh no, the festival is permanently cancelled [1].
[1] https://www.sfgate.com/food/article/Gilroy-Garlic-Festival-n...
I'm not trying to make a pun, but it's times like this I feel the world is losing its flavor. So many neat things that were around when I was a kid have been cancelled because I guess no one wants to bother or because one bad thing happened.
The renderings and anecdotal knowledge about that rail route make me think this is not going to be an electrically powered train. What fuel would it use?
https://www.nytimes.com/2024/04/04/climate/trains-planes-car...
It does.
> this would be a low-priority to electrify because it’s not a major shipping route.
US rail electrification is more focused on passenger routes than shipping routes; an electric freight engine for mainline use has been introduced (just last year), but its not intended as the sole power source for a train, but rather to improve efficiency by being used in a consist with diesel engines, recharging with regenerative braking, reducing overall fuel use.
On the other hand, I took an overnight train from Wuhan to Beijing. It was crowded, the cars not so modern, and I didn't sleep all that well. The high speed rail train between those two cities was a much shorter and better experience.
Also my experience with trains are they are way more casual of a travel experience. Show up a few minutes before and walk on. Airplanes are always an ordeal. Like You need to add security line buffer time. Or wait for for bags
Less threat of a firey death in a Boeing plane to be topical too lol
I definitely prefer the experience of rail travel - boarding and riding though.
But you start and end your train ride very close to your destination and you aren’t locked in a shitty small seat. Every single bit of the trip is better, with the possible exception of the travel time. Even this might be better if you measure door to door.
overnight on land will happen on self driving cars before it happens on rails
What infrastructure? This new service would run on existing rails and terminate at existing stations. The only thing new would be the trains themselves. (The proposed later addition of a drive-on/drive-off car carrier service would involve some additional infrastructure.)
Furthermore, planes and driving have always been an option in the US.
Sofia to Bucharest is approx 400km.
San Francisco to Los Angeles is 600km.
Sofia to Bucharest is largely flat.
San Francisco to Los Angeles requires building rail lines in multiple mountain ranges that rival the Carpathians.
Imagine having to build a 600km high speed railway line through 3 mountain ranges the size and height of the Carpathians, oh and almost nobody lives in between the two metros.
It does not make financial sense.
San Francisco to Los Angeles is approx 630-700km.
Sofia to Bucharest is largely flat.
San Francisco to Los Angeles requires building rail lines in multiple mountain ranges that rival the Carpathians.
Imagine having to build a 600km high speed railway line through 3 mountain ranges the size and height of the Carpathians, oh and almost nobody lives in between the two metros.
It does not make financial sense.
Somehow Europe builds those things even across national borders. To be fair, there's not really an alternative if one wants to connect southern and northern Europe.
Absolutely, and those were multibillion dollar projects that could be justified to help connect 4-5 different heavily populated regions - Vienna, Innsbruck, Veneto, and Milan.
In between the Bay Area and Los Angeles Metro there are only 2 metropolitan areas (Fresno, Bakersfield), and they have a combined population less than 20% of Veneto.
That's it.
If the californian high speed railway would be the right thing to build, it'd have been built 60 years ago. Maybe it will be viable one day, but apparently it has not been so far.
Yes, you need some enormous tunnels. That's more than possible, ask the Swiss or Norwegians. And yes, it's an earthquake zone, but it's still functionally a political problem.
I'm not saying it's in any way practical reality because there's always a million reasons why It Just Can't Be Done. But that those million reasons are human choices, not physical laws of nature (other than the mountains, but it's 2024, not 1824, that can be done). It's just a chronic lack of intestinal fortitude.
It can be done, but it is very expensive. A 8 mi mountain railway tunnel like Wienerwald cost around $400m to build. That same amount could literally be spent to expand the Altamount Corridor Express railway (ACE) into downtown San Jose, thus having an impact in a closely integrated metropolitan area of 10m residents.
Why spend that (and much more) money integrating two cities that aren't even that closely tied from a commuting standpoiint.
Only 2.2 million people travel between Greater LA and the Bay Area annually, but 5.5 million travel between San Jose and San Francisco annually.
Clearly there is some demand for LA-SF connectivity, but it's hard to justify.
Yes, some of it is for dumb reasons, but there's also plenty of reasons that aren't just "we used to be a real country". It is, in fact, not 1824 any more. We can't just offer thousands of Chinese immigrants $2 a week and have half of them die while putting a hole in the Rocky Mountains. Yes, that's "a human choice", but the physical dimensions of the problem define the very real tradeoffs that have to be made. The "physical laws of nature" don't stop us from doing all sorts of things we'd probably regret, many of which we do regret now.
To go back to my original point, those reasons are a lot more compelling when you're spending some billions of dollars that connect to another country's billions of dollars of trains, and you get some regular amount of cross-traffic. Sure, we could build more, but we're not going to build a Deutsche Bahn amount of rail within Arizona. They're about the same geographic size, and Germany has 10x the population. Likely same with Nevada and Oregon. So what do you propose we connect to that has anywhere near the same network effect of cross-travel?
That said, trains are a nice way of seeing the country. Because the rail system is mainly for cargo, passenger service in most places is not good for time sensitive travel. It is great for sightseeing. This route being overnight is disappointing as it takes away the sightseeing but maybe they've found some available slots during the night between cargo loads.
Anyone who has a chance to travel between Portland and Vancouver by train should take it. You get views of the coast not available by road. Just be sure to get a seat on the side of the train that faces the ocean or you're going to have to enjoy seeing lots and lots of trees.
Hopefully when the cars and the line are rundown in 20 years regular people will be able to afford this as a way of travel