American High Speed Rail is a Terrible Idea

Let’s talk about this new push for high speed rail. Let me state this straight out: If you care about global warming, you should oppose plans to criss-cross the country with high speed rail.

There is a myth out there that passenger trains are somehow ‘green’, and use of them will lower CO2 output. The reality is very different anywhere where trains don’t run at 100% capacity…

Using data from the Transportation Energy Databook here:

Amtrak’s actual results are that it uses 1535 BTU per passenger mile. Commuter rail uses 1603.

Now let’s assume that we’re going to use Teslas instead of trains to move people around. A Tesla model 3 uses about 823 BTU per mile. That’s close to HALF of what it costs in energy to run a passenger train per passenger mile, even with just one person in the Tesla.

These numbers for the train are based on the average number of people that actually ride the train, not its theoretical efficiency at full capacity. These are real numbers. And they don’t help the case for high speed rail, because current passenger trains in the US are in dense corridors where they get higher utilization. There’s no way you are going to maintain those kinds of capacity factors in a nationwide system.

The fleet average of current passenger cars is 2840 btu per passenger mile, with an average of 1.5 people in the car. That’s at an average fleet fuel economy of 24.9 mph. So just by putting three people per car in the current fleet you would make it more efficient than trains.

But more to the point, the fleet average economy goes up constantly as older, less fuel efficient cars are retired. The fleet average in 2040 (long before the trains come online) is projected to be almost 40 mpg. If some of the electric car projections happen, it will be more like 50. But even at 40, that means the fleet average by 2040 will amount to 1775 btu/passenger mile with only 1.5 people per vehicle. That’s right about where current trains are, and would be a lot lower than a nation-spanning HSR.

And if we instead focused on going all-electric, then in a future of electric cars, passenger trains are energy-hogging relics. A Tesla with two people in it can travel cross country on ONE QUARTER of the energy of a passenger train. And for long distances there are options to make them even more energy efficient, such as ‘platooning’ or ad-hoc car trains run by the autopilots.
And when would these trains come online? Not for decades. California’s HSR boondoggle is a case in point: Construction was approved in 2010, with a projected cost of the system of $40 billion, and a promised completion date of 2028, 18 years after the start of the project. The cost has now ballooned to $100 billion, the route has been shortened and stops deleted, and it’s not projected to be fully running until 2033. Hands up, all who think either that budget number or deadline will be hit.

So, at least $100 billion and 20 years of development to build one 380 mile stretch of rail. Now do the math on what it would cost to build the many thousands of miles of HSR across the country.

And what happens before the trains are online, during this critical period where we have to reduce CO2 as much as possible? Well, you’re going to need thousands of miles of high-strength steel smelted, and you’re going to be burning huge amounts of energy on heavy machinery, grading, bedding, concrete, yada yada. The train system is a 100% CO2 emissions source for at least two decades, and more likely many decades.

High Speed Rail uses about 140 tons of steel per mile for the tracks. Smelting a single ton of steel causes the release of about two tons of CO2. So we are talking about 280 tons of CO2 emitted per track mile just for smelting. Then you have to mine the ore, refine it, transport it, work it, weld it, etc. And you need heavy machinery for all that. I don’t have a number for those additional tons of CO2, but it won’t be negligible. In fact, the smelting would only be a fraction of the total energy budget.

So let’s say we built a single HSR link along I-80 between New York and LA. That’s about 2800 miles. Just the smelting energy for the track needed will release 784,000 tons of CO2. And that’s a small part of the overall energy budget as you’ll have thousands of heavy machines, grading, concrete overpasses, etc. So you’ve now committed to decades of increased CO2 production before the first train ever runs.

Then by the time the trains come online, we’ll be driving electric cars and HSR will look like a giant energy sink - a relic of a previous century we can’t afford to run because of CO2 emissions. The same activists that want HSR now will probably be lobbying to prevent it coming online by then…

Another problem with rail is that it is ‘planned’. Putting a passenger rail link between two cities relies on those cities remaining the size that the planners planned for. Let’s say in 1970 you built an HSR between Detroit and somewhere else, basing its capacity on estimates using Detroit’s current population. Today, with Detroit being half its size, your trains would be running half empty and be even more expensive and less energy efficient. In a world of rapid change that’s increasing every day, Tying yourself to a fixed set of rail links between cities that were important 40 years ago is crazy. Much better to have networks of roads which can accommodate large-scale changes in usage patterns.

Finally, every dollar the government spends on a ‘solution’ that doesn’t fix anything is a dollar that could have been spent on something that did. For example, for the cost of California’s HSR you could build ten large nuclear plants (25 if you could do it for the price other countries do), which would help stabilize a grid full of renewables and save thousands of tons of CO2 every year.

There’s another way in which HSR could make global warming worse: If politicians give up on separate track because of all the right-of-way issues and piggyback the system on regular track, they’ll screw up rail freight. And the U.S. has the most efficient rail freight system in the world. So it’s entirely possible that after the politicians get done, the HSR won’t be fast, and more heavy freight will be moved to the roads as they had to do in Europe when passenger rail crowded out freight. And freight is what trains are really good at. Moving more freight to the roads will increase road wear and raise the energy cost of moving goods.

In a world of limited money and political capital, if you think global warming is a critical problem you are insane to support spending trillions on a fancy rail system that will cost huge energy to build, won’t be available for decades, and will likely be no better than, and possibly much worse than other travel alternatives by the time it is finished. Hell, given the pace of change it’s possible that by the time the rail lines are finished the demand for them will have shifted somewhere else…

Debate any of these points if you’d like.

There is no reason someone should take train from NY to LA, unless they are terribly afraid of flying or want to sightsee.

High-speed rail makes sense for densely populated regions like the Boston-NYC-Philadelphia-DC corridor or the Dallas-Houston-San Antonio triangle. For that, a shinkansen would be great.

So you are okay with extra CO2 emissions for decades to wind up with a system no better than alternatives and possibly worse?

I used NY to LA as an example. The same math applies to every route. California’s HSR will never make back the energy it will take to construct it. In fact, when it comes on line it may actually reduce transportation efficiency compared to the alternatives.

The same is true for every line. Lines that are running at constant full capacity might do better on a passenger mile basis, but they’d have to be about eight times better to beat a Tesla full of people.

I guess these electric cars will run on roads made of pixie farts or something.
(IOW your analysis leaves out things, to say the least.)

Pro tip: In order to say that high-speed rail is worse than flying, you need to compare to the numbers for flying.

Domestic airline operations use 2219 BTU per passenger mile, as of 2019.


Pro Tip: If you think I was talking about flying, it might be better to actually read what I wrote.

But if you’d like the numbers, intercontinental air travel averages 2.94 L/100km per passenger. A 2017 Toyota Prius consumes 4.2L/100km. So yes, a plane would be much more efficient than HSR, since a Prius is also more efficient than HSR…

If you are wondering why passenger trains are so fuel inefficient, it’s because of the amount of mass in the train compared to the mass of the passengers. An empty train costs almost as much to run as a full one, both financially and in terms of energy. One of the reasons why Light Rail has dismal energy statistics is that the run on time whether full or empty. At night, the LRT here runs at maybe 10% capacity, and probably less. Entire cars run empty. And yet, it’s still a huge vehicle that takes a lot of energy to start and stop.

Freight is a different matter. Rail freight is generally heavy, so the mass fraction of the train is lower. Heavy loads benefit more from steel wheels on steel rails, reducing rolling resistance compared to moving it on roads. Trains excel at moving heavy things. If you build passenger rail and it reduces the efficiency of the world’s best rail freight system, you can add another economic and environmental disaster to the list.

If you see HSR efficiency numbers from a supporter, pay very close attention to the capacity factor they are using. A common trick is to assume the train would run at capacity, but trains never do. Hell, I took the TGV in France a few years ago, and the four of us on the business trip were the only people within probably ten rows of us.

Of course, the TGV has the benefit of running on very inexpensive electricitry, because France powers it with Nuclear.

Well, you could designate some existing highways as autonomous car only roads. That would take out a lot of the brainpower needed for the cars. It would still need upgrades like lining it with sensors and transmission stations but I wouldn’t imagine it would approach the cost of a rail line.

One great way to make a very small fortune is to start by investing a very large fortune in U.S. passenger rail service.

The only two areas where passenger rail in the U.S. makes any sense at all is in the northeast and in southern California. The northeast is probably the best case (DC-Baltimore-Boston-etc) and even then it doesn’t make enough money to take track preference away from freight. Southern California has similar traffic and population density issues, though in the long run I sincerely doubt that California is going to end up making a profit off of it.

Rail service makes sense in Europe, where the population density is high enough that you aren’t paying for miles and miles of tracks between stations and where folks actually tend to use public transportation. In Europe, if you take a train to someplace you can then take a bus or a taxi to get to your final destination. In the U.S., public transportation has withered away to almost nothing due to a lack of paying customers.

Interesting. My numbers are different, but I think I see why. I was using ‘fuel per seat’, which assumes full capacity. Your numbers are probably the average across the fleet, which is a fairer way to look at it since that’s how my other numbers work.

On the other hand, we have no idea what the capacity utilization of a nationwide rail system would be, but undoubtedly the average would be lower than the current rail which serves dense corridors only. So I’d say it’s likely that HSR will be much worse than the cars of 2040, but maybe about the same as air travel.

For reference in case someone uses full-capacity numbers for the efficiency of HSR, at full capacity an Airbus A330-900 burns 2.48L/100 km/Passenger… All current large jet aircraft burn between 2.4 and 3.5 L/100 km per passenger, with the difference mainly being things like how many first class and business class seats there are.

As a side note, if you care about global warming, real low-hanging fruit can be had by getting rid of first class and business class seating. First class seats have up to SIX times the carbon footprint per passenger kilometer as does economy seating. All for a little more leg room and a better meal. People who care about global warming should be ashamed to fly first class.

I want to hear more about the “push”, since I’m not sure anyone is seriously proposing intercontinental Shinkansen. Who is? What prompted this post?

It works great in Japan, where density is very high. I can see a limited use in, as stated above, a few limited dense areas but am unsure who the proposed customers are.

Environmental issues are one important consideration to weigh. Not always the first thing the energy industry brings up, but times are apparently changing.

I assume currently it would be this:

Ahh, upgrades! I guess we can just ignore what it cost to get all the roads in the first place while we focus on the cost of new rail lines.
(And I daresay after my stay in Pennsylvania, where the state flower is a ‘ROAD WORK AHEAD’ sign, that the cost of upkeep isn’t trivial compared to the upkeep cost of rail.)

Also this:


And about a dozen stories in the same vein. Biden’s story isn’t specifically hsr, but the same numbers apply to the kind of trains he wants.

Well, not exactly but do you want to also cost out all the existing rail lines that might be reutilized for a high speed system? For instance, the California system will share rail line with current local systems. Do you think all the cost for those was part of the press release?

I don’t care, but you can’t compare them if you don’t include, or exclude, the ‘same’ costs.
Someone wants to include the ‘cost’ of making new rails but doesn’t mention where the metal all these new electric cars need is coming from.

As far as I can see, Sam didn’t propose any road upgrades, that was me. His argument was based on improvement in automobile tech.

Feel free to add what you think I missed. As for the electric cars, is it your belief that the existence of an HSR link will stop people from buying cars? Those electric cars will exist anyway, they’ll just be in the garage or parked at the train station.

But if you’d like to quantify some things I missed with real numbers instead of handwaving, be my guest.

I’m hand-waving?

“the glossing over of details” hmm…

I clicked on the link you posted in the OP. There were so many tables in the data section that I gave up looking for the specific numbers you cited. Could you post a more specific link?

I don’t think that’s true at all. Here’s the Amtrak route map from 2018 and it runs multiple routes from coast to coast. Faster trains may well increase ridership. I believe that already happened when the Acela service was introduced on the Boston-NYC-Washington corridor.