Is Green U.S. Transit a whopping myth?

As part of my research into robotic cars, I've been studying the energy efficiency of transit. What I found shocked me, because it turns out that in the USA, our transit systems aren't green at all. Several of the modes, such as buses, as well as the light rail and subway systems of most towns, consume more energy per passenger-mile than cars do, when averaged out. The better cities and the better modes do beat the cars, but only by a little bit. And new generation efficient cars beat the transit almost every time, and electric scooters beat everything hands down.

I encourage you to read the more detailed essay I have prepared on whether green U.S. transit is a myth. I've been very surprised by what I've found. It includes links to the sources. To tease you, here's the chart I have calculated on the energy efficiency of the various modes. Read on, and show me how these numbers are wrong if you can!

I have added a follow-up post on the comparison between lots of small personal ultralight vehicles and larger shared transit vehicles.

Note: If you want to comment on the cyclist figure, there is different thread on the fossil fuel consumption in human food which details these numbers and invites comments.


That San Jose Light rail is at the top doesn't surprise me. Most of the time, I see 2 light rails, attached in tandem -- and both of these trains together carry about 2-3 passengers!. Perhaps San Jose light rail has a policy where each rider needs to have his/her own train car...

But I include it because it's here in the place that I and many of my friends and readers live. There are worse systems. Cleveland and Pittsburgh are even worse. Surprisingly, L.A. is pretty good, though it's not thought of as a transit city. But it's a big city, and big cities can get more ridership inherently.

Cleveland and Pittsburgh have ancient systems which were largely designed as 19th century property development schemes. It's no surprise that they're not efficient at people-moving; they weren't intended for it.

Not only that, but San Jose light rail actually DECREASES the commute efficiency for everyone else. They can override the lights and cause backups in some of the most congested parts of the city.

Further, if a car could have its own right of way, and stop all other traffic at intersections, I bet it would be pretty efficient, indeed.

for every other option you've only counted efficiency from motor output to distance travelled, so counting a larger portion for cyclists is IMO unreasonable. You could count source to motor as well for the others but that would get ugly fast. So lets just look at cyclist output per mile covered. I'm going to work in ISO units then convert to your wacky ones.

A mediocre cyclist can put out 150W and gets about 20kph from that, or 8J/km. Which is roughly 512 BTU/hr and 12.5mph according to google. So that's about 41 BTU/mile, not even slightly close to the number you came up with above. Interestingly though, it is slightly more than the electric car. Hmm.

Now, certainly if a fat person living in the US was to eat the usual US diet and then ride a bike, you could reasonably blow that number out by a significant chunk. But by rights you should only consider the marginal increase in food consumption, not the gross number. Of course, if you want to consider where the fuel comes from then the coal that powers the generator that feeds the grid that charges the car should be taken into account too. You could also factor in the cost of making the car and bike, but that would also not favour the electric toys. But countering that is the extended life expectancy of the cyclist so their lifetime recsource use goes up, favouring passive transport.

It's not the main thrust of this article, I covered it in this one and I stick by the number.

Biking one mile at 10mph (6 minutes) burns 38 more Calories than sitting for 6 minutes. 38 Calories of food from mixed agricultural sources consumes about 380 Calories of fossil fuel based on the 10 to 1 ratio I've seen in the literature. Thats 380,000 calories, which is 1507 BTU. This is a "fuel" (by which I mean petroleum, not food) to wheels analysis. As I note in the other post, for beef it is 4x worse, and a cyclist fed on beef is using 6,000 BTU of fossil fuel (or 1/20th of a gallon,) which is worse than the car average. You're talking about human fuel and human power output. Based on purely the calories in food, the human does about 140 BTU/mile based on my calculations. Almost as good as the electric bike does on electricity. Beats the electric bike because its electricity comes from burning fossil fuel, but the electric bike beats the human since the human's food is made by burning fossil fuel.

However, this is not the point of the essay, it's a tangent, and it's not about cycling being bad, it's about the fact that U.S. agriculture uses 1.1 gallons of fossil fuel per person per day to make our food and what that means. It doesn't even include processing and shipping the food! Of course a cyclist eating from their own garden does better, but this is about the average. Cycling is good exercise, but if you eat food from modern argiculture, it's not nearly as green as you think it is.

I assume that to go from about 50 BTU per mile to 1500 you've rated people at 25-30% efficient (reasonable) then multiplied by 10 for your fictional "only human food has an energy cost in its acquisition" example.

I think it would be useful to include embodied energy in your calculations too. So heavy rail drops a bit because over 50+ years the cost of a big steel rail and big steel railcar is pretty low, but over 5 years (if you're lucky) the electric car/scooter start to look a bit sad. electric cars also suffer the same problem as mass transit - they only really pay off if you do a lot of miles. So the taxi operator in Vancouver who drives his Prius fleet almost continuously wins big (and gets about 2-3 times the expected mileage out of a set of batteries), but some bunny who buys one "to be green" and drives 5000 miles/year would do us all a favour if they simply burnt the cash instead (and took taxis).

But then, that favours cyclists again. Even a carbon fibre and titanium bicycle looks pretty good next to equally exotic materials in the electric car but with 500x more of them. I do wonder if by "scooter" you mean the $100 things from Taiwan with 12V/7AH SLAs in them. Those are all recyclable, but unfortunately their life is short because they load the battery way too high. The more expensive ones work better but cost more (and weigh more, use more expensive materials).

Car manufacture is typically 120 million BTUs, which is about a 15% extra on the 150,000 mile lifespan of a typical car. So while this affects the numbers, it is only by a modest amount. (With some irony, the Prius is more affected, because that 120M BTUs is about 25% of what it will burn in fuel in that 150K mile life.)

Fuel refining and delivery adds about 22%, but applies to everything -- cars, power plants and the tractors that farm food for humans. I have a section on this in the article, I presume you missed it, so do I need to make it stand out more?

The scooter/electric trike/bike class I derive from figures for the various small-volume vehicles we see in this space. They show what's possible but it's not mass produced. Battery tech is rapidly improving in all areas, but for electric scooters/trikes, SLA is fine.

Ok, I'm regretting putting that bike figure in now because people are harping on it. It's a correct figure, but tangential to the main article.

I tied but I can't find your numbers in the literature. Of course, the literature I see uses ISO units so I could easily be doing the conversions wrong. But cycling is what I know and care about, and when I see people like yourself making misstatements it grates on me. At the very least you really should say "a US cyclist" to allow for the variations caused by different food sources. Other fuel sources we can leave out of it if you like, but that doesn't make them cost-equivalent. You talk about asian transit but not about asian food sources, for instance. One of the problems we all face is that many asians are starting to want to eat the way northerners do... and we can't afford that.

Connex who run the rail here do actually run smaller/shorter trains off peak, and we use light rail a lot. It's possible to do similar things with road transit - Christchurch NZ has been using minibuses off peak for some years now. But then they have also computerised everything so it's easy for the drivers (or potential passengers) to indicate full mini-buses and get more capacity quickly.

It's not so much a technical problem as a financial one. In NSW they at least have the "benefit" of an honest treasurer - he quite frankly opposes public transport and goes to considerable lengths to make sure it does not work. As a result PT in NSW is not much better than most of the US gets. But should we actually resource PT appropriately it can do much, much better (it's more than just money).

Trying to be clear right from the title, I hope, that this is about U.S. transportation figures. What numbers are you looking for literature cites on? I'll try to put in more. Calories burned? Or fuel burned in agriculture? I do not talk about food or cyclists in Asia at all, I just quote some numbers from an Australian study I found that compares the various regions. And indeed it notes that the US sucks. (Australia's better, and so is Canada, but not as good as WE or rich Asia.) The study shows you can do it better, but the US doesn't, and won't without a big seachange. But expensive gasoline might be that seachange, who knows? (Of course expensive fuel hits transit too, as long as it stays this inefficient, with only subsidies hiding that.)

And, if we could, can we move the discussion of the cyclist number to the specific blog post on cycling? I am getting tempted to remove the cyclist from the transit article to avoid distractions and would then probably want to purge or move these comments. (There is no easy comment move in drupal though.)

A number of your articles and entries allude to and talk around a topic I am trying to find out more about - the total energy used in the lifecycle of a vehicle (ideally, including raw materials, assembly, operation, maintenance, scrapping/recycle, etc). In this note, you stated "Car manufacture is typically 120 million BTUs" - have you found a reference which provides this information in any more detail?

One thorough (but not independently verified) document on the subject is "Dust to Dust" at This paper allows one to draw the suspicious conclusion that when normalized to the lifetime of the car, a Prius uses more energy per mile than a Hummer (it looks like this result is dominated by the expected usable lifetime of the vehicles and the complexity of the materials going into the vehicles). Obviously, a result like this is at odds with expectations and draws into question the veracity of the information (as many have already done).

However, the basic principle of what they are attempting is really the way we should be framing the energy conversation - what matters is not the immediate per-unit energy cost of an item (MPG, BTU/hour, etc), but the unit normalized end-to-end energy cost of the item. If it takes twice as much energy to manufacture a high efficiency solar panel as it produces in its lifetime, then it is not a good deal for the environment (it is a net loss). Until we ask the right question, we cannot hope to find the right answer.

The problem right now is you see a lot of different answers in the various lifecycle analysis papers. Plus you also see people wanting to incorporate different things, like the cost of roads both in manufacture and driving. That's very hard to do, as you would have to figure out when a road would not be built at all (or as large) rather than divide up its cost among all the vehicles. They also get multiple uses (streetcar tracks etc. often run down them.)

The Hummer vs. Prius paper is pretty famous, and pretty strongly disputed.

Well done on the blog post regarding energy per mile. As I understand it, and appreciate it, the main point is simply that the true nature of the petroleum use for human transportation is counterintuitive, or at least contraintuitive. In a society where 'moral choices' are affected so heavily by the currents of discussion in the society as whole,(e.g. mass media, Oprah, PTA, political ads), and this discussion is rarely informed by hard numbers and end to end analysis but rather by emotional appeals to vague concepts, the reminder of the true nature of these issues is very valuable. If only CNN would use 10% of its 24 hour coverage to discuss these subjects in analytical fashion, we could actually make progress in improving the entire world.
And thats the point of this article that I like so much. Issues like transportation need to be looked at from the entire world. Not just the whole surface of the globe, but conceptually including the soup to nuts analysis of everything affected, not just the microcosm of the green this or the green that.
Well done, Brad.

Everyone who starts to get serious about transportation rediscovers the importance of load factor and tare. Did you deliberately choose an inflammatory summary? The bland summary is that the statistics show that nationwide average transit uses 60% (per passenger) of the energy that is used by cars. Both figures can be improved substantially, but the easy targeting is different.

For low capacity methods like automobiles, you need to eliminate tare weight. That is why the scooters do so much better than autos. They have much less tare weight. There is a lot that can be done in this area.

For high capacity methods like transit, you need to improve load factor (reducing tare is good, but load factor is easier). The figures that you show look like the 2005 US numbers. Load factors have increased dramatically, so transit is probably doing better now. Carrying more passengers increases the energy cost by only a tiny amount. There is still a lot that can be done here.

Then, defeat the imbecile politicians and the transit union power. A mixed mode solution will almost always be best. You use a mix of heavy and light rail on the high density links, with personal vehicles for low density. If done right, this can make a huge difference. It does involve the mentally difficult task of accepting that there is no one true god. There will be a mix of choices made.

I would suggest pruning or re-organizing your lists to compare only genuinely competitive modes. Tricycles do not compete with jet aircraft, nor does heavy rail compete with city buses. Also, it is usually necessary to incorporate local terrain and weather issues. The transportation analysis for the northeast corridor is different than the analysis for Texas.

The analysis gets more complex if you consider traffic pattern modification. Work from home is one example of a substantial traffic pattern modification that can happen quickly. Relocation of housing, offices, and commerce happen more slowly. Traffic mix management happens as well. The oldest major traffic mix management project is the Alameda RR Corridor (which restructured automobile, truck, ship, and rail traffic for Long Beach). One that is just getting underway is the CREAT project around Chicago. These shift freight off trucks and onto rail. Rail is roughly 10x lower energy cost per mile for freight. Steel wheels on steel rails are incredibly efficient, but they have a very low route flexibility and a very high switching cost. These projects also re-organize transit and auto traffic to reduce the freight/passenger conflicts. It might not make much of an energy difference, but everyone involved is less upset by congestion.

The evolution of commerce also introduces changes. Mail and Internet ordering with home delivery is much more energy efficient than shopping trips. So there will be a transition toward this for non-perishable generic items. It will be much harder to manage for perishables and for items that require in person inspection or fitting. Complicating this some are the efficiencies from bulk purchase stores. Home delivery behaves much like transit. It has a high fixed initial cost, but very low per destination costs. If I can get a delivery route with an average of 100m between drop sites and fill the delivery truck, it's nearly impossible for any personal vehicle to match the delivery energy cost.

Where do you get the 60% number? The nationwide average of "all transit" (not counting airlines) is very similar to the car numbers. I believe it's about 3,300 btu/pm compared to 3,500 for cars. Only 6% better, which in my mind is meaningless, since the impression that most people have is that transit is "much greener."

The most recent stats I have don't show a "dramatic" increase in load factor. I'm using the DoE's most recent numbers. I have seen news reports of transit riderships up about 10% due to gas prices, but am waiting for consolidated numbers rather than anecdotal examples to make conclusions. Please point us to what you have seen.

As to what is best, I feel that's up for debate. The truth is even making transit twice as green by doubling ridership barely dents our transportation energy use. Americans don't tolerate transit's other problems -- long trips, unreliable times, unreliable transfers, no station near them, long waits, can't stay out late, etc. I think big change will have to come from making private vehicles much more efficient. Americans demand vehicles that take them from A to B without having to change buses at C.

The chart is not to show competition, it's to make people aware of energy numbers. But yes, subways (heavy rail) do compete with city buses -- in terms of what cities decide to build. I see subway/light rail/BRT debates going on.

Now as you'll see shortly, I predict a much bigger change than any of what you cite coming from a completely unexpected source.

My numbers are for differential effects. If a person switches from car to transit, the actual energy consumption by transit barely changes. It's less than 100 btu/pm. But, if that person takes a car pool from 2 people to one, there is no significant change in the energy consumption for the car either. Since the average occupancy of a car is 1.57 PAX that means about 60% net energy reduction unless I assume that the car to transit shift also changes the average occupancy to a significant degree.

For the longer term, you have to look at what load factors you can achieve. Here things get more complex. The roughly 30 per vehicle figure for transit means an average load factor of under 50%. If I can push that load factor up to the 80% level, which is still comfortable, the btu/pm drops to 2,000 btu/pm. Again, the 60% figure.

Getting better than 60% requires much more significant changes. Either the inherent vehicle efficiency must increase or the vehicles must become substantially lighter. Those are hard changes and will take time.

As for competition, I meant in the eyes of the user. I thought I made that clear. It's true that cities like Boston and Chicago spend money on aircraft, rail, road, and bicycle. All these modes compete for city capital spending. So in that respect airplanes and bicycles compete. But as a means of getting from point A to B, bicycles and airplanes do not compete.

Shifting to transit would dent transportation energy usage. I like dent as the analogy. Does a dent eliminate an automobile? No. A dent is something noticeable but not overwhelming. Doubling ridership implies adding capacity. If I use 80% load factor, those riders are 2,000 btu/pm. I might get a 10% reduction in fuel consumption. (Miles traveled changes, etc. make this a very fuzzy number.) That's a dent.

For really big effects you have to give up on efficiency alone and start supporting behavioral change. Behavioral change has already dropped vehicle miles per capita by about 10%. That is the current figure for April 2008 versus April 2005. When gasoline hit $2/gal VMT per capita started to drop. It was rather slow until more recent increases in cost. The drop in the year April 2007 to April 2008 was almost 6%. This is behavioral change, ranging from car pools to consolidating shopping trips to having children take the bus or walk to school.

The behavioral changes most often visible to the high tech audience are telecommuting. One extra day per week telecommuting from home is a 20% reduction in transportation energy use. It is much easier to ship electrons. They weigh only 0.1% of a proton or neutron. But a subtler and longer term shift is eliminating jobs. The transition from paper to electronic transactions saved the NYSE from imminent meltdown. It also eliminated tens of thousands of jobs. This did affect the commuting pattern for NYC. Work restructuring does have a large impact on transportation.

Earlier you wrote that you felt transit energy usage was 60% of car usage, which the numbers I have seen don't bear that out. (It is true in some cities of course, an definitely true in Europe/Asia, but low-usage cities bring the average back up.) You now seem to be suggesting we could get transit that took 60% of the energy of cars, and that's certainly possible, though whether the American public can be readily convinced is another story.

My conclusion was that since it needs something radical to get Americans out of their cars -- and I am not sure $5 gasoline is radical enough -- the right push is for more efficient private vehicles, and that push is already happening.

Indeed, since private vehicles are over 60% of all transportation energy usage, and urban transit is 1%, the truth is, if you're not focused on cars, you're not really serious about making things greener. (At least in transportation. Outside transport, the right targets are electricity generation, agriculture and a few industrial areas.)

A properly run transit is 2000 BTU/pm. Thats 60% of an automobile. That's all I've been saying there.

I'm not objecting to improving automobiles. I do object to picking single targets. You get your savings through system improvements and system change. Top down stalinist command and control is also not the effective way to make system change. You make system change through enabling individual decision makers to make lots of local decisions. I've also done system efficiency improvements. The big wins are usually from complete elimination of something, not from doing the same thing the same way but more efficiently. These are steps like email instead of Fedex. Already 10% of vehicle miles per capita have been eliminated by behavior change. That is a big change that took just a couple years.

Proper integrated planning of multi-modal transportation does require regional coordination, but this too need not be top down stalinist. It just needs a regional coordinating organization. It is the exception. Lots of systemic change can proceed without planning.

One simple step is presenting information to ease decision making. Instead of presenting automobile usage as miles per gallon do gallons per mile. Consider these four hypothetical cars:

Behemoth 10 mpg or 10.0 g/100mi
Better B 15 mpg or 6.7 g/100mi
Small car 30 mpg or 3.3 g/100mi
Super H 300 mpg or 0.3 g/100mi

There are people investing lots of money in modifying small cars to approach a Super H. I don't object to dreams or research. But if you look at the g/100mi you realize that the fuel savings from improving a behemoth to a better behemoth is greater. So the serious investment should go into the mundane improvement of behemoths and persuading people to replace behemoths. That has a much better return on investment. The advanced research should continue at a low level so that when the behemoths are getting 4 g/100mi, you have a research base for doing better. The step of persuading people to change has happened with persistent high gas prices.

The system problem is that today's small cars are not a functional substitute for the behemoths. If I occasionally need that large capacity, I have choices: buy a behemoth and use the extra gas all the time, buy two cars (small and behemoth) and save on the gas, or have an alternate way to meet the occasional large capacity need, e.g., rental. Financially, the buy behemoth is usually the best choice. From a system perspective, if I could provide an alternative occasional large capacity at a reasonable cost and availability, people would buy the small car. That's a system problem, not an automobile efficiency problem.

You agree that top down control doesn't make it happen, which is why I must dispute the assertion about a "properly run transit system." You can't just declare, "the solution is to make all transit systems properly run." First of all, I am not sure that all the bad numbers are a result of bad management. Transit load factors depend on a variety of things, some of them simply tied to the geography of the city, the demographics of the users, the history of the city and other factors. A vertical city like New York is going to do better than a rich sprawl like San Jose. Management can only do so much.

But the other point is that transit is such a tiny part of transportation energy that whatever we do to improve it, though not wasted, is only going to dent the big picture. The real change must come in cars/light trucks. Transit can take riders away from those, but that's about all it can do to change the big picture. But is it the most effective way to do that?

And yes, that's why I present the figures in BTU/pm, and not primarily MPG. High gas prices, which I support, do a lot to push people out of Hummers and SUVs.

BTW what's the cite of 10% drop in vehicle miles per capita? Not fedex, surely? Has it even reduced? I make far more use of delivery services in the ebay/online shop era than I did before.

And as for your last point, you will see in a week or two that this is exactly where I am going!

Yes, we need something radical to get people out of their cars .... by replacing them with electric cars that run on and off guideways, powered by the guideway, and by small batteries when off the guideway. The best design I've seen for that is the RUF. The cars form themselves into trains when appropriate, which probably gives them a better efficiency than even heavy rail.

"Unfortuantely, since transit is only 3% of total transportation energy use,"
What's the other 97%?

Also, have you got any figures on Caltrain? I don't ride the light rails much, but I Caltrain quite a bit.

It's in the DoE transportation factbook cited in table 2.6 and now that I look at it (I may have used another source for 3%) it's actually around 1%. 0.7% for buses, 0.3% for passenger rail. The rest is:

  • 63% for light vehicles (cars, light trucks etc.)
  • 16.7% for heavy trucks
  • 9% for planes
  • 3.1% for pipelines
  • 2.1% for rail freight

Transit is in the noise. Recreational boating uses a similar amount to transit!

No, I haven't found Caltrain numbers. I have numbers for BART and Muni, but Caltrain has too many counties. BART is an above average system, by the way, I suspect in part because it's half commuter-rail, half-subway. Commuter rail which has long distances between stops does better.

I'm trying to piece these bits from the essay together:
" is always the green move for any individual to take existing mass transit over a car."
"My first conclusion is that we would get more efficient by pushing small, fuel efficient vehicles instead of pushing transit, and at a lower cost."
"This issue is complex because we like transit for other reasons."
Because of the wheelchairs, children, and elderly, transit is still needed (at least until the "robocars" appear). If we agree that we still need transit, then what is family of 4 with no car (like mine) to do? Should we buy an electric car or continue to use transit? As you allude to in the article, one piece of the answer is in the incremental BTUs/passenger mile for the transportation options you've listed. The bar chart only shows the overall BTUs/mile though. Do you have a chart showing incremental BTUs/mile with some set of assumptions? It sounds like you've done a lot of research; I'd guess if you don't have such a chart already, you have nearly enough information to create one. Please do post it! Thanks.

Well, realize at the price of many transit systems, many cities could afford instead to have subsidized taxis at jitneys for qualified people. For the handicapped, that is in effect what many towns have, though many paratransit systems are just dreadful -- "we'll pick you up between 2 and 4."

Of course pure taxi subsidy wouldn't work -- everybody would take it, since unlike transit taxis go directly from A to B, and if you make an appointment, leave then. You would have to work out a subsidized taxi that truly only served those who would have been the off-peak transit riders. For peak times, transit makes all sorts of sense. Then it's green and reduces congestion, and if given its own right of way (which is very expensive but already done in most cities) it can be faster.

Jitneys may be an answer, since their stops and diversions and limited routes would appeal mostly to the current transit rider. And in fact there are countries where jitneys are very common, where they even out-compete the buses and "have" to be made illegal because they are doing so. What jitneys don't have our schedules. Many U.S. transit systems have schedules but effectively they are useless. In any case, most transit riders don't want transit with schedules. They love transit that goes so often you can just expect one to come in a short time. But in the end they really want to get to their destination on time, and not have to leave too early. The schedule is just a means to that end.

But in the long run, I point out we do want these other functions of transit, but what these numbers revealed to me is that they need to be judged on their own. The argument that we're saving energy by building transit systems seems to fall down, at least based on most recent figures. $4 and $5 gas might change things, who knows? But in the long run I think it means people buy more hybrid cars.

Some other useful things are carshare, computer coordinated carpooling (goloco) etc.

This is like the airline paradox they teach kids in stats class. Airline A has a higher ontime ratio than airline B. But they both serve the same routes, and B is on time more often for every single route. How can this be? Some of the routes tend to cause more delays, and Airline B flies those routes more often.

The forms of transit you're flagging as inefficient are used in places where cars might be stuck idling in traffic jams or looping around the block looking for parking. OTOH, intercity buses and trains seem to be more efficient than highway driving. (God knows every intercity bus I've ever been stuck taking was packed to the gills)..

Also--mass transit enables greater densities, as we don't have to use so much space for parking or roads. And that's where the real savings lies--not in per mile efficiency, but in decreasing the number of miles by allowing people to live much closer to where they work. If you eliminated mass transit, many cities would become simply unworkable, and homes and businesses would have to relocate to distant suburbs.

All that said, yeah, your chart does suggest that jitneys and other systems that allow people to share vehicles but maintain flexibility are a really good idea. And some individual mass transit systems look pretty dang suspect (I love Pittsburgh, but it wouldn't be the first time it screwed up its infrastructure...).

But I think there's pretty good reason to suspect mass transit saves more energy than efficiency per mile would suggest.

Car city mileage would compare with urban transit (light rail, city bus, heavy rail ) while car highway mileage would compare with commuter rail, suburban bus and rail etc. There is some validity to this point, but consider that newer car technology like the Prius and Insight have better city mileage than they have highway mileage, and these vehicles beat all the transit figures.

So the airline analogy is not as apt as you suggest when you consider future planning. We would do more to get greener by pushing efficient private vehicles than we have done with transit. Can we do better with transit? Theory and European practice say yes, but history in the USA says no. But expensive fuel might change it.

As I noted, transit has other uses besides (not) being green, including reducing congestion and parking need. The point of these numbers is to say those are the things to talk about, not about how transit is so good at saving energy.

Secondly, recall that my big shock is not that cars are greener than transit some of the time. It's that it's a close race. I think most people perceive transit as much greener, and the numbers just don't say that, no matter how many subtleties we may discuss. I want to see arguments why these figures are off by a factor of two.

Another way to look at it is to see that all transit averages to about 38 passenger-miles per gallon. While mpg is inverted from the best way to examine efficiency, it is more familiar to people. If you have a private mode that generates better than 38 pmpg, which hybrid cars and electric cars certainly do, it's going to do better anywhere it gets this efficiency.

Per passenger-mile? The real way to look at transportation costs is to use multiple variables. If you want to grind on some particular point, then you choose the variable that fits your agenda- the most common theme in U.S. energy discussions today. In other words, you aren't even wrong.

First, develop a procedure for individual states - then scale up to national scale. It is also worth noting that most consumed goods are now imported into the United States, and material transport doesn't even fit into the "passenger-mile" image - even though transport of goods (food, material, goods) around the United States accounts for the majority of the fossil fuel use.

Indeed, why not rework this into "fossil fuel emissions per passenger mile"? How do you decide to do the accounting? The point here is that this is just another example of bad use of statistics, period. You can draw no meaningful conclusions from the presented data.

Can you cite a source for your claim that freight is the big user? The DoE numbers I point to show light vehicles as 63% of transportation energy use and under 25% for freight. Check table 2.6 I think. Not that freight is not important, but it's nothing like what you say.

In other countries things are different... or perhaps the writers are just lying (I mean "using statistics") in a different way?

ANALYSIS of emissions from Australia's most popular motor vehicles and from public transport has shown the Hummer to be the most polluting, closely followed by the popular Ford Territory.

The research, by the Public Transport Users Association based on the Federal Government's Green Vehicle Guide, showed that diesel trains, such as those run by V/Line, were the least polluting form of transport per person per kilometre.

These trains emitted eight grams of carbon dioxide equivalents per person per kilometre, compared with 270 grams from a Hummer and 245 grams from a Ford Territory.


Comments: VLine is the medium-range train system in Victoria. Most Australian short-range heavy rail is electrified and in Victoria that means brown coal for fuel. Ditto for light rail. So diesel trains are used for interstate and intercity transit rather than intracity (although in NSW the electric system covers journeys within about 120km of Sydney CBD). Buses are mostly diesel here, although gas is increasingly popular. That's gas as the rest of the world uses the term... what do merkins call hydrocarbons that are gaseous at room temperature that are used for fuel?

It all depends on occupancy. Not on my chart is long distance rail like Amtrak, which is not viewed as transit per se, though it might be worthy of a slot. It is fairly efficient, for a number of reasons.

  • Trips are long, with few stops
  • Trains are long and well enough occupied. Long trains are the most efficient thing out there if you can fill them.
  • They run too infrequently in the USA to be considered transit, but that allows better occupancy.

Commuter rail, as shown, is next best. Is that the V/Line? I would like to see the numbers but 87 for Prius and 8 for V/Line is too far off the numbers I have seen unless it is regularly packed.

When considering electric vs. diesel/gasoline, it is *ludicrously* important to know how the electricity is produced.

IIRC, Australia produces almost all its electricity from coal. No wonder then that Australian diesel trains are better from a greenhouse gas perspective than electric trains.

In California or Washington state, which are almost entirely powered by hydropower, electric is far more carbon-efficient than gas.

Oh, by the way, diesel vs. gasoline: diesel is more energy-efficient and more carbon-efficient; it requires more energy (including the crude oil used) to refine a particular number of BTUs of gasoline than to refine the same number of BTUs of diesel, last I checked. Currently diesel is more expensive due to a shortage of diesel refining capacity; that should not last.

FYI, we call methane "natural gas" in the US.

While I don’t doubt the accuracy of your numbers, the issue is the myth or efficiency of public transport is largely dependent on ridership. As you've stated, fully loaded, the public transit system is very efficient. Asia’s and Europe’s systems are very efficient due to the ridership – ever been to China and Japan?

So really, the problem isn’t the transit systems but the people who refuse to use them. The efficiency calculation is largely cost of system/rider miles where the larger the number of riders, the greater the system efficiency.

It seems as though the essay’s ulterior motive is to trash the transit system and promote the idea of robocars.

No, that's not my motive. I've been a transit fan in the past, but the numbers are convincing me otherwise. However, my motives are not relevant. It's the numbers that matter, and only the numbers that should be debated.

That people in the USA don't fully utilize transit is not a counterargument, it is actually the point. You can say "it's the fault of the people for not riding transit" but that's a head-in-the-sand approach. People don't do what they should. They do what they want to do, what their wallet or tastes tell them to do. Rational transportation planning should not involve the idea that people will suddenly change their desires. Sometimes they do, but it's a tricky way to plan.

I believe it's a reasonable assertion you will do far more to cut transportation energy use by offering highly efficient private vehicles (robotic or otherwise) than by other means. At least for now, people in the USA don't follow the patterns of Asia and Europe, and you can't make them do so by fiat. (And I don't mean the small Italian car.)

"I believe it’s a reasonable assertion you will do far more to cut transportation energy use by offering highly efficient private vehicles (robotic or otherwise) than by other means."

I don't believe it's reasonable.

Here's why: everything we have seen in recent behavior and surveys indicates that Americans under the age of 30 really, really want better public transportation, they want it on rails, and they will ride it if it's there.

We are underbuilt for current demand as far as mass transit goes. We don't need to change people's tastes; they already changed; we just need to start satisfying the unsatisfied demand.

This demand is concentrated in particular locations, because people are self-sorting. Those locations also happen to be more suitable for rail transit than others, so that works out nicely.

If you're going to wave a magic wand and create a transit system which is fully loaded all the time, then I get to wave the same magic wand and fill every seat in a car with a rider.

Neither of us have that magic wand.

In the spirit of the essay, i.e. discussing the data, I think if a poster wants to call Americans to greater use of transit, or even dog them for their low usage rates, then that person should supply some numbers. I would suggest that they declare an assumption of their 'ideal' ridership numbers, and then run the numbers to post btu/person/mile numbers for buses etc. that can be compared to the essayist original post. He's done a fair amount of work putting together data, lets add to it and not detract.
Maybe I should get busy on the data myself, because I refuse to be led by the nose by a culture that insists that it already knows whats best, that I should take the bus, but is not willing to put up the hard numbers to prove that this is indeed the case. If I'm asked to sacrifice my convenience, I deserve to know for sure that it is worth it. Show me the wattage!

Actually, it's pretty clear that taking any vehicle already going your way is the efficient thing to do. The bus, train or car weighs so much that you won't add much load by getting on. So take the bus or train or form/join a carpool to be most energy efficient for any given trip.

The message here is instead for society, and transportation planning. How do we reduce energy usage and pollution in a way the public will embrace?

Its your essay, you can define the purpose however you like, and I'll go along.
And well put about the efficiency of my own choice.
But my point still stands regarding the data required for my city to make the best overall choice for future planning. I've lived in 4 major cities, and Austin is the worst in relation to the convenience of transit. In any case, if I'm given a chance to vote on transit proposals, I want to decide based on something like your efficiency framework, not of a vague and pushy notion that buses are always better.
If we don't have numbers like btu/person/mile, or total cost, or grams_C02/marginal_$_GPD, or some such overarching metric, whats to prevent us from continuing to bark up the wrong tree in our pursuit to make the world a better place. Sometimes I feel like the only voter around who wants a full analysis. Thanks for bringing the DOE data to my attention.

Because estimates of ridership will vary wildly. And yes, those who want to build the transit systems tend to feel ridership will be high. There are some people who estimate better than others. Ridership will depend not just on the type of riders, but how well the transit is designed. In San Jose, the light rail largely runs down the media on freeways. Stations are not close to many homes and offices, if any. So ridership at those stations is quite poor, accounting for poor efficiency.

Worse, some ridership projections plan out decades, expecting that development will follow transit, with hi-rises at the stations, as are found in some cities with old transit. I would be wary of this, since I forecast robocars before two decades. I now believe transit plans must deliver today if they are to make sense.

So what type of metric would best help you evaluate if a transit plan in your area makes sense? What value should the threshold be set at? What scope (e.g. geographic area, governance zone, tax base) is most appropriate?

Did you include long distance electricity transmission loss in your plugin electric energy use numbers?

Studying charts like:, impacted my transportation choice(s).

The numbers are indeed based on 10,399 BTUs for a kw-h, which involves mostly generation losses, and then about 7-8% for transmission losses. In effect multiplying by 3. "Generation losses" are easy to calculate for coal and natural gas, which are about 70% of electrical power. They can also be calculated for nuclear (and are similar to coal) but with nukes the heat is not generating pollution directly. With hydro, again the comparison is less clear because we're not even using heat-based power. However, since 70% is fossil fuels, that's enough to derive useful numbers. What are the numbers in the grinning planet chart?

The US Transportation Energy Data Book 2004 has these figures:

Transport mode Efficiency per passenger
Rail (Commuter) 5.3 L/100 km (45 MPG)
Rail (Transit Light & Heavy) 5.6 L/100 km (42 MPG)
Rail (Intercity “Amtrak”) 5.6 L/100 km (42 MPG)
Cars 7.2 L/100 km (33 MPG)
Air 8.1 L/100 km (29 MPG)
Buses (Transit) 8.8 L/100 km (27 MPG)
Personal Trucks 8.9 L/100 km (27 MPG)

We have developed a mass-transit system that is really green. An inner-city UniBus can transport 50 passengers with a fuel efficiency of 100 MPG! It is even better if we supplement energy with solar-assist.

You can see more information on our web site:


Looks cool, but may miss the lesson learned from the Energy data book numbers (I used the 2006 book, but I did not think the 2004 book is that different.)

Namely, it's all about load. Transit is of course more efficient when full or nearly full. But it faces a problem. To be convenient, it must run frequently, including at low demand times. Unless you can use smaller, lighter vehicles at the low demand times, the efficiency of the entire system suffers, and this is what makes it score poorly in the national average.

So for a new system, the question is not simply can it be efficient, but will it actually be efficient based on the passenger loads. And alas, it's hard to answer that well until you actually deploy.

I don't know if these cable-suspended rails can work, but I will give them some points for having smaller single-unibus transit be possible off-peak. And the rails don't look nearly as eyesore generating as most other elevated trains and monorails. I would need to see more of the physics that keeps them so rigid.

"solar assist" doesn't truly change the efficiency of a transit system. You must work from grid power. If you put up solar arrays, great, you make the grid cleaner and more efficient. But it's a silly idea to suggest that if you put up solar panels just for your system that this makes it magically greener, because you still have to grid tie (solar panels are terrible if not grid tied so that every joule they make is used) I know people who power their electric cars from solar panels and try to say "my car doesn't pollute" even though they would cause even less pollution if they grid tied the panels and took car power from the grid, but would not feel as good about it.

Brad, I agree with your statement about load.

In fact, in making the comments below, I do not want to sound like I am disputing what you have said. In fact, we have recognized the inefficiencies of public transport that you have raised, and spent years developing our optimized system to deal with these inefficiencies.

The benefit of our system is that even if the UniBus was carrying 2 people (instead of the capacity of 50), the fuel economy is over 100MPG. This makes the per-person fuel economy 200MPG, which is still far higher than any other system.

The problem with rail-based or bus-based systems is that their capacity comes from the vehicle size. During off-peak times, they make it less frequent (as you have mentioned). In peak times, they operate say every 20 minutes; in off-peak times, they operate every hour. For our system, they operate every few minutes; in off-peak, they operate every 10 minutes. This is still more convenient than light-rail or buses during peak times.



Ah, so you claim your vehicle gets 100 vehicle miles per "gallon" then? That's very efficient -- 1250 BTUs per vehicle mile or just 120 watt-hours per vehicle mile at the DoE's generated power rate of 10400 BTUs to generate one kwh.

I have to say I am skeptical of a large vehicle getting 120 watt-hours per vehicle mile, can you point to the page where you calculate this number?

While this is very good, when empty it's still not as good as an electric scooter (250 BTUs/vehicle mile) but starts beating it with over 5 passengers average load. Still doesn't beat the ultralight electric vehicle on going directly from A to B though.

"Unless you can use smaller, lighter vehicles at the low demand times,"

A properly managed train system can do exactly that. As long as the low demand times can give a reasonable load for a single railcar, you put on one railcar at the lowest demand times, and more at higher demand times. One car at midnight, twelve cars at rush hour.

(This assumes the use of multiple-unit designs, and for that matter designs with fairly fast coupling and uncoupling. Locomotive-hauled designs will have energy usage somewhat less responsive to number of cars attached, though still rather responsive.)

Clearly San Jose is just being stupid if it's pulling two-car consists at all times; why not use one-car consists for the low-demand times? Seriously, why not?

In fact, a carefully designed bus system can also do this, with buses of several different sizes ranging from a minivan up to a full-sized bus, used at different times of day, but that's rather less efficient in terms of capital expenditure.

....actually, I just realized that you made an error of assumption in the essay. In a properly run rail system, those full rush hour trains are *not* running back empty: at midday, they're waiting near downtown to go back at evening rush hour; at night, they're waiting in the outskirts to go back at morning rush hour.

Bad scheduling and bad consist design can certainly render a rail system inefficient. It can do the same for a bus system very easily. Taken together, all this seems to indicate that perhaps we should spend more time and money on keeping our transit systems *good* and matching schedules and consists to demand, rather than just saying "Look, we got transit"?

Yes, there are rail systems that just keep the trains at one end after the rush. That requires a lot of extra rolling stock and engines, though, and not all systems do that. Ditto the bus systems. They could switch to vans during the low periods, but that requires a complete set of buses for rush hour and a complete set of vans for off-peak, which again they don't usually do.

As to what we should do, that is the hard question. Perhaps the main message of these figures is that one of the most important factors in "what should we do" is "will people use it?" You can build a beautiful transit system that's fabulously efficient if it gets good ridership, and it can still suck if you didn't estimate ridership properly -- and transit planners are notorious at estimating ridership properly. Transit planners who get paid if transit lines are approved may have a financial bias as well.

That's what has led me to the conclusion that taking steps to put much greener private vehicles on the road in place of the sucky ones we have today may be the best way to improve energy efficiency.

Every new transit system opening in the US in the last ten years has experienced significantly *higher* ridership than predicted, with the exception of Cleveland's Waterfront line.

I rest my case.

as I know it's not every system, since BART's SFO extension in right in my own town is way under projections.

So please give me a cite for this underestimation? Is it related to the brief transit peak that took place in 2008 during $4 gasoline?

Interesting article, but I'd be curious to see where the figures for average passenger load for rail transit come from. In my experience, 33 passengers per train would be a good bit lower than average on the late-night weekend trains that I frequently ride here in Philadelphia, and those are noticably desolate compared to similar trains in other cities that I'm familiar with. Trains at any other time would of course cary many times more, of course, which makes it hard for me to accept your figures at face value unless there is some other factor that's slipping my mind....

Figures are sourced from the U.S. Government Dept. of Energy Transportation Energy Fact Book. It is linked to from the article. And a few other government sources. Figures for Tesla, Electric Scooter and Cycle are from other sources.

They are national averages. Some cities do much better, some do worse. I would guess that since the cities tend to buy the same equipment (other than diesel vs. electric) the load factor is what makes the difference. The charts I got indicate Pittsburgh light rail is one of the very worst in the USA, though. Heavy rail (commuter and subway) is presumably better based on your figures.

Nice presentation, and very interesting figures.

A thought: rather than present performance based on the average passenger, could you show the different transportation modes in terms of the BTUs/passenger-mile for the marginal passenger (additional BTUs required per additional passenger mile)? I'm guessing that the public transit options will come in close to zero on this basis. This is the real answer to the question "What should I do as an individual to be green" though of course it doesn't answer the question of what we should do on a systemic basis.

For all the modes. Cars also won't add too much more energy to add a passenger. The only difference will be how efficient, overall, the engines of the various vehicles are at the particular load. A small scooter might see its weight load increase 50% adding a passenger, but it was so efficient to begin with it still might do a better job. The simple message, "take a vehicle that's already going your way if you can" gets it down.

I have used public transit a lot in the DC area, and the DC metro carries over a million people per workday. Making a simple calculation of BTUs per car vs. metro train is not helpful, because it does not factor in the dramatic increase in BTUs (due to traffic gridlock) that would take place if those million+ riders switched to cars every day. It takes a hell of a lot more BTUs to get from DC to the suburbs in stop-and-go traffic than in free-flowing traffic conditions, but free-flowing traffic conditions -- already rare -- would be nonexistent if those million+ riders switched to cars.
As more people switch from cars to public transit, if traffic goes down as a result, then it lowers the btu/mile number for the remaining car commuters who are traveling in less traffic.

The DOE's number is a national average, combining the stop and go with the smooth sailing. However, it is a valid point (noted to some degree but expanded on) that our current number is based on our current congestion.

Current thinking is that traffic expands to fill the roads. Build more roads, get more traffic, build fewer, get less. People adapt. There is a level beyond which they won't take a road if it's too gnarly. They will divert to other roads, or live elsewhere, or take transit.

yeah but the trains run on clean green coal, nuke or hydro...bring it on!

Well, we haven't seen a lot of that yet. If you have to burn something, natural gas is going to be better but of course they both emit a lot of CO2. Work's being done on how to sequester that but it's a long haul. However, I do believe we can have a greener electric grid, and when we do, it makes electric transportation (cars or transit) greener.

"However, I do believe we can have a greener electric grid."

*That* is actually the key, #1, way to become greener. Practically everything can be converted to run on electricity at this point, and the trends are naturally pushing in that direction, but it's all for naught if we don't have a renewable electric grid.

There is no alternative fuel which can currently plausibly be generated renewably in a reasonably efficient fashion, other than electricity.

Thank you for posting the numbers. Too much writing about transit is based on fantasy rather than fact. A good example is the proposed high speed rail for California. The planners promise a system which will be profitable and they promise to deliver same for a fraction of what it will really cost --- despite the fact that no public transit system in the world operates without substantial, ongoing government subsidies.

The only truly, inherently efficient mass transit system is a variation on the old electric trolley system. The trolleys trundle down the street from corner to corner, dropping off and picking up according to a set schedule, and trolleys are added or removed from the lines according to the time of day. When you exit the system you should never be more than 1 city block from your destination and it should never take more than ~15 minutes to cross from one side of the system to the other - there probably must be express trolleys incorporated into the design. The electricity comes from clean electric generators, solar or nuclear, and regenerative braking returns power to the system at each stop. Net efficiency would be at least 50% better than ICE vehicles and these would create negligible pollution.

The only reason most cities don't have these is that in the 1940s and 1950s the Big Three automakers bribed - er., convinced the cities to install bus lines that were noisy, polluting, and had horrible schedules. Then when riders got angry at the cities they sold them cars.

Someone mentioned the deleterious effect of mass transit on regular traffic. I presume there is also a benefit, but it it's hard to quantify.

More importantly, when calculating carbon costs of any transit system, don't we need to include the carbon footprint of all the drivers/engineers/brakemen/ticket-takers? They are taking their (well deserved) salaries and using it to consume lots of energy as well--for their own cooling/heating/transportation/entertainment/manufacturing.

In other words, they could have different jobs in which they produced value for their carbon emissions. They do not hove those jobs, so that value-per-tonne-carbon has been lost as a society.

"This issue is complex because we like transit for other reasons."

It is also complex because there are reasons to not like it for other reasons. Actually, not like it is perhaps an overstatement. More like it will never be feasible in some settings. In rural areas your major premise that transit fails because ridership is never close to 100% will always be the case. Not because rural people are not green but because I suspect that below a certain population density you come to a point that you take the mass out of mass transit.

I have lived in Hong Kong two years, and the Philippines four.
You have not taken into account population density. The MTR
(Mass Transit Rail) in HK is the ONLY unsubsidized (in fact,
VERY profitable, and run by a private company) mass transit
in the world. This is due primarily to the fact that HK has a
population density of 14,000 people/sq mile (the highest on
earth). Most of Asia averages 300-700 people/sq mile, much
higher in urban areas. The US averages 30-70 people/sq mile,
almost nowhere (even downtown Manhatten) high enough to make
Mass Transit even break even, let alone profitable. This is
why it works in Asia (where they still prefer individual means
of transport, such as scooters, over MT). The MT in Atlanta
(Moving AXXXXXXs Rapidly Through Atlanta) is a total (and
typical) joke from a financial viewpoint. It can never be anywhere
near profitable, or even usable to most of the city's 5M plus
inhabitants, although we ALL have to subsidize it. When it comes
into a suburban area, crime rates soar and property values
plummet. We fight it tooth and nail, yet we are forced to
support it. For a long time, a former welfare brood mare with
NO prior business experience was running it (she had a budget
of about US$1B a year, and produced essentially no benefits).

And other areas of low population density. I hope that's clear in the article, it's all about ridership and whether a transit system can get enough of it to be efficient and profitable. I've ridden the transit in HK, and it is heavily used and cheap and rapid, and the roads are few enough that cars are slow.

For cases like Hong Kong, my future predictions involve a mix of robotic personal vehicles and robotic transit. A multi-modal solution. This also may happen in other large cities at rush hour with minimal transit outside of that. Once transit drops in frequency below a certain amount it encounters a death-spiral -- people don't want to use it because they will wait and can't readily predict the wait. In HK, you have one street car visbily following the next.

Do you think the highways you use are free? They cost the goverment money too. A LOT OF MONEY.

Your ignorance is pretty amazing, but typical.

Don't you have a primordial ooze to slither back into?

There are some other factors that I think need to be considered when attempting to deterine the overall energy efficiency of mass transit. The first is that mass transit routes are necessarily what may be called average routes: they are not the most direct route for most of the passengers. In other words, some fraction of the users of mass transit have to travel more miles to get from point A to point B than they would if they used their personal vehicles. The other factor is that mass transit operates on its own schedule and users of mass transit have to adjust their schedules to that of the transit system. This means that you have "wait for the bus" and while you're waiting you're not really working or producing anything. If you have to transfer from one mass transit system to another (train to subway, bus route to bus route) more time is spent waiting. Both of these factors involve small amounts per instance and tremendously large numbers of instances. It is the same sort of thing with allowing "right turn on red" which saves maybe ten seconds or so per person in idling time but a very large amount of time (and fuel) in total.

I am awfully glad you did this research and wrote this article. I, for one, have no intention of ever using mass transit, except for air travel. The reason for has nothing to do with "green" or btu per passenger mile. When I was in high school in Detroit in the mid 1970's, I rode a city bus, not a school bus, to school. I observed first hand then that the majority of users of mass transit were people with no respect for others and obviously with no morality. I was robbed at knife and gunpoint enough times to cause me to A)Start illegally carrying a concealed weapon, and B)Quit school to avoid getting on the bus. This many years later I have not forgotten those lessons and I now have a license to carry concealed and my own motor vehicle. I know for certain that I face a much lower risk of violent confrontation in my vehicle than I would on mass transit and that if I do encounter a criminal, the chance for turning the criminal into the "victim" has expanded exponentially. I also now reside in Florida, whose new laws make it quite legal for me to shoot the perpetrator son of a bitch in the head. So... I will take my SUV to the dance and screw green!

I think your "crime" or "respect" problem had nothing to do with mass transit.

Your problem was being in Detroit. Heck, people get carjacked in Detroit.

Thank you for the posting Brad. Not surprise about your findings but still think we are comparing apples to bananas. The obsessions to compare transit to autos is getting old and in practical terms is of little consequence. I am with you 100% on the robo-cars. I personally believe the future of transportation will be something like an ipod-car that is called on demand. This will be the typical public transit but it will never come to reality if we keep on insisting that public transit in US cities should look like those in European cities.
I don't believe in density. In fact the best green technology is the internet and over time there will be more and more people working remotely using the internet and computers. Both passenger travel and freight transportation will be changed due to technology and (if we could only open our brains to the future of technology) transportation corridors will be much different than what we envision them now.
Regardless, good thoughts on your research. Reminds me that public transit such as buses use up over 50% of the pavement's life. Thus in your calculations you should attribute more than 50% of cost of pavement to buses and heavy vehicles, maybe 80 to 90%. Then you can incorporate the cost of pavement and related to per vehicle trip and extrapolate that to per mile of rider.
One factor that is never addressed in this type of comparisons is the factor of convenience. The reason why people driver cars is because it is much more convenient and if we equate convenience to dollar value the people use cars because using transit is too expensive; in most cases it is impossible to do with transit what we do with car, i.e. go where we want to go, whenever we want to.

Brad, superb article. Since moving to a larger metropolitan area (coincidentally San Jose) and having to deal with the higher gas prices, I recently opted for an alternative to driving everyday that I find works EXTREMELY well, the Electric Bicycle ( Based upon my commute distance, gas mileage, cost of the bike, cost of recharging it, etc; I figured I'd break even in about a year (assuming a modest 50% usage).

From Randal O'Toole CATO Institute
‘Rails won’t Save America’ 6 pages
Does Rail Transit Save Energy or Reduce Greenhouse Gas Emissions?

I really like the Cato report. And Brad's too. Thanks!

And thanks for making the point that once the rail is running, that's a good system to use because there are no additional marginal emissions. However, from a planning perspective, this is very provocative that adding buses on sparse lines can be very bad for emissions.

When looking at 'green'ness, it's vital to note that California (for instance) gets something close to 80% of its electricity from renewables (mostly hydro), while Texas gets something more like 20% of its electricity from renewables.

This means that an electric train is *MUCH* greener in California than in Texas. A huge, huge difference. Apples to bananas.

This is going to make cities like San Jose look much better on a greenhouse gas basis. And in fact this difference outweighs pretty much every other number.

Anyway, those extremely low average passenger numbers simply make it clear that running an unpopular and inefficient 'mass' transit system is stupid. Mass transit is for where there's a large *mass* to transport. If the roads aren't congested, you don't need or want mass transit, from an efficiency and emissions point of view. And if the roads are congested, buses and streetcars will be very slow, and therefore unpopular and inefficient, unless they have their own right-of-ways. This leads to the conclusion that the main point of such a system, from an efficiency point of view, is to speed commuters past congested areas. Most US bus systems have zero dedicated right-of-way, and are therefore close to entirely wasted.

The Pittsburgh "LRT" is actually a piecemeal conversion of a very old set of streetcar lines. There is no doubt that it's better than shutting down the streetcar lines and driving every single streetcar user into buses or cars, since that would add a very nasty amount of congestion to Pittsburgh streets; however, the line is not on a particularly rational route for a 'core rail' line (it exists largely because of streetcar tunnels which made bus conversion difficult during the 'transit holocaust'). In other words, it *goes the wrong places*. Pittsburgh's transit agency is currently attempting to extend it so it actually connects to more than one place people want to go.

Incidentally, it's sad to see lunatics claiming that mass transit increases crime and reduces property values; this has been proven untrue a gazillion times.

If you're curious, the energy costs of maintenance and construction of asphalt roadway are generally higher per mile than those of steel railways, largely due to how fast asphalt wears out and the fact that it's basically pure oil. Although it varies depending on any number of things such as the use of concrete rail ties (much more energy-intensive than wood), different maintenance standards (a road allowed to go to pot will be more energy-efficient in construction and maintenance than a railway inspected every week, though it might well make up for it in efficiency of the vehicles using it), the steel production method for rails (recycled obviously better), whether the asphalt is hot-rolled or not, etc.

For another minor point, the degree to which rail vehicles have lost energy-efficiency over the period when cars have gained energy-efficiency is largely a US phenomenon owing to very poor regulatory choices. The Federal Railroad Administration really likes tank-like vehicles; the trend in the rest of the world is lighter, lighter, lighter.

As I note, there is considerable variation in local situations, both in where power comes from and how well it is used. The west has more hydro but still has a fair bit of coal, nuke and gas. Whether nuke is green or not is a question I won't resolve here. But the short answer is that yes, if the electricity is green, then electric vehicles are a win over fossil fuel vehicles. Especially electric cars and scooters.

There are lots of issues with this analysis. First of which is saying that all cars carry 1.5 people. Hahahaha. Come on, it's more like 1.15 people per car if that. The DOE is on crack if it thinks 1.5 is normal. How many people do you know that carpool to work or drive with other people that much. That is 3 people riding in every two cars. Next time you're on the freeway look around and see how much that happens. I also find it hard to believe that trolley buses are higher energy use than diesels.

Second, is comparing cars to transit without talking more about land use even realistic if we're going to talk about being green? Cars have created a land use pattern that results in a lot of wasted energy. New Yorkers use 1/3rd of the energy that the rest of America uses. This is because of the land use patterns from the transit network there, not just the transit itself versus cars.

Using transit also means that you're not going to drive as much because you probably live in a more walkable/bikable environment. In Portland, Metro did a study in 1994 that showed compact development and good transit would push a use of 9 vehicle miles per day. In exurban Portland that number increased to 21 vehicles miles today. So even if we let your 1.5 number in cars stand, that still creates a situation where people are driving 12 more miles per day per capita. So lets be honest about what being green means in terms of transportation and not forget all the connections that occur. Looking at cars versus transit in energy alone is short sighted and misses a lot of issues that come with auto-centricity that seems like an illness in this country and is a huge WASTE of energy.

Then there is the issue of using BTU between all modes, specifically electric ones.

I will presume the DOE used fairly scientific methods to get their figures. However, even if they made a major error, and it's much lower, like 1.2 that raises the average car to about 4500, and it's worse but largely in the same range. I suspect the DOE's number is actually correct, but appears skewed to us because it includes the whole range of car travel, from highway road trips to solo commutes.

However, it remains true that the new generations of more efficient cars, such as the Prius and Insight, deliver better efficiency that many of the transit systems in the country, even when driven solo, and electric cars do even better, even solo, and electric scooters do vastly better, even solo.

Land use patterns are important but are not readily amenable to change in the short term. What people drive can change rapidly in the short term -- people replace their car every 5 years or so. They replace their streets and parking lots every 100 years, to make a wild guess.

As for their energy methodology, as long as 50% of U.S. electricity comes from coal plants, I think it's not an unreasonable number. Note that whatever the number is, I use the same number for electric cars and scooters as I do for electric trains, and they are the real answer, and they are better than transit with no argument needed about conversion factors.

You're (trying to) use per-passenger-mile figures. Is the DOE car occupancy averaged over the correct base? If it's the total number of people in cars at time X divided by the total number of cars on the road at time X, it's the wrong number for your purposes; if it's the average number of people in the car per trip, it's the wrong number for your purposes; if it's the average number per mile driven, it's *STILL* the wrong number for your purposes.

Check your math very carefully.

I'm not a brilliant person, but I do know that government oversite comes at a price: less money or less freedom. I live 20 miles from a city. I prefer to be right where I am. Mass transit is not "friendly" out here. It takes a lot of time to ride the bus when it stops every few miles. I've been trying to keep up on solutions to the oil crisis. John Hofmeister, former CEO of Shell (on Glenn Beck last week) said gas is only 20% efficient and ethanol is less. He proposes hydrogen cars. They're clean energy, emitting only water and some heat. But, he failed to mention that hydrogen is produced by coal, gas or oil, so we're back to square one. Electric cars, hmmm, they're okay tooling around town, but they don't go far before they need to be charged again. Hybrids are a possibility, the prototypes can get over 100 miles per gallon when you calculate the cost of electricity, but I don't want to "plug in" my car. I don't want battery powered cars ruining my environment. Plus, they are painfully expensive. I just want a gas powered car that can get 100 miles to the gallon. If we could do that, we would reduce our CO2's by at least half. Is that asking too much? (A rhetorical question.) There's an international competition to build a 100-mile-per-gallon car. The Automotive X Prize will have an award of at least $10 million to the team that builds the car. I'm keeping my fingers crossed that somewhere someone has the ingenuity and brilliance to create a car that is capable of rocking our world!

So, until that happens, government will want us to do our part to make sacrifices and spend hours on a train or bus, move closer to a city, limit our vacations to the closest beach, and pony up more for the cost of fuel. No more going to visit grandma down in Florida (a 1,500 mile trip). Maybe we should just start riding horses.

Medical injury costs and value-of-life paments make bicycles and scooters non-starters:
Bicycles are about 180 times as dangerous as cars:

Per 100 million miles traveled automobiles incur 1.41 fatalities and bicycles 250. [ and }Farming -- all that extra food the cyclist needs -- has one of the highest fatality rates of all occupations.

Injured and dead folks inur large medical costs and loose work time.
The average roadway fatality costs about $980K. Average non-fatal bicycle accident costs $220K to one million dollars.[]

Also Most deaths on major roads. Fifty-seven percent of bicycle deaths in 1999 occurred on major roads, and 37 percent occurred on local roads. (6) The yearly economic costs also include $61 billion in lost workplace productivity; $20.2 billion in lost household productivity; $59 billion in property damage; $32.6 billion in medical costs; and $25.6 billion in travel delay costs.(source = ? Insurance institute link broken)

I am not surprised at your blog as I as I have previously made similar calculations.

Further, I recommend that every educated mass-transit fan read their local transit authority annual report -- it's an eye opener. Then they should visit the large, classy transit administration building.

I agree that existing scooters have too high an accident rate. I am projecting a world of advanced safety systems (and eventually robotic driving) that make ultralight vehicles, notably small electric trikes, both safe and efficient.

The guy that said that bicycles are 180 times as dangerous as automobiles is way too high. I did my own computations for motorcycles and came up with a of between 20 to 30 times more dangerous than driving a car.

Common sense would suggest bicycles are safer than motorcycles simply because there is a lot less kinetic energy for the passenger to dissipate in an accident. This reference suggests that that the bicycle fatality rate per passenger mile is 3.4 to 11.5 higher than cars, which seems plausible.

An electric scooter would probably fall between the bicycle and motorcycle number, since they would tend to be operated at an intermediate speed. Any sort of enclosed vehicle with safety equipment and the number would drop, perhaps below even the bicycle numbers.

BTW, I just did a calculation on the number of pedestrians killed per year in the US and came up with a walking fatality rate of 8.5 per 100 million miles or 6 times higher than driving. This suggests that it might not be impossible to come up with an ULV that is safer than walking.

Bike safety is highly incumbent on not getting hit by larger vehicles, as in cars, trucks, busses, etc.

I spend many hours a year on a bike and look forward to a RoboCar world. First they'll weight less so if you get hit by one it will hurt less. Second they will be much better at avoiding me. And finally won't intentionally buzz, honk or hit the brakes after swerving in front of me.

Green doesn't mean energy efficient. If you're concerned about climate change, it means carbon efficient. If you're concerned about peak oil, it means fossil fuel efficient.
You're using a metric that's meaningless.

I'd like to see your chart redrawn to take into account carbon efficiency, not energy efficiency, because, as some other commenters have mentioned, the greenhouse gas emissions of an electric vehicle, such as a train or trolley bus, depend upon how the electricity is produced.

I'm living in a city with electric rail and electric trolley buses. 70% of the power comes from hydro, geothermal and wind. So for this city, public transport starts looking like a low-carbon way of helping people move around. Even in the US, 30% of electricity generation isn't fossil fuel.

I do cite the national averages: 50% coal, which is not just greenhouse gas but a whole bunch of other bad stuff. 20% natural gas (fairly clean other than CO2). 20% nuke -- no greenhouse gas but you have to judge for yourself what you think about its environmental consequences. 10% hydro -- no greenhouse gass, but it destroys valleys and fish and other animal habitats.

It's complex, and there is no one number. This balance does indeed shift around the country. I think actual energy used is a good number to use for national analysis, and it's the right number to use in general systems design, because you need to design systems for all the cities.

But yes, if we move to renewable electricity, the electric transport (both cars and trains and trolleys) will score better.

The fact is that we all demand an awful lot of mobility. We want to live far from our jobs; we want the freedom to roam quickly and freely over a vast physical area during the course of our days and weeks; we don't want to walk any further than absolutely necessary, which means parking very close to our destination at all times.
Walking is an amazingly simple and efficient way to get from one place to another -- and it doesn't take as long as you'd think.

But people right now want sprawl, and even if they stopped wanting it you can't make the vast amounts of it go away within a century, I fear. So while you can lament over sprawl (or rather, the desire of people to buy larger homes on lower traffic streets for their kids even if it means they must drive everywhere) that lament doesn't address the problem.

But there is a problem even in non-sprawl. Even in the urbanized areas, most trips (and certainly most non-rush-hour trips) are not to or from downtown, they are between different, non-downtown parts of the city. Traditional transit designs don't solve that well at all. PRT might, and robocars could.

I've used these types of data sets quite a bit for environmental research, and can tell you from a quick look that there is something very wrong with your numbers.

There are many pitfalls in interpreting these numbers. In addition, the numbers don't always give the full picture. E.g., there are huge differences in the environmental impacts of electric and gasoline energy.

Before you publish, even on the internet, you might want to validate your findings with somebody who knows what they are doing. In fact, many such studies are available for you to compare, and they show pretty different results.

They come from the Department of Energy, and they do factor in the difference between electrical generation and gasoline burning. They are not carbon numbers, so if your experience is in carbon numbers that may account for your instinct, but instead of just a "you're wrong" you need to be specific about which methodology is wrong. There is already analysis about the issues with the numbers in the article you may have skipped.

Another big caveat on these numbers is that they were probably measured at different vehicle speeds, and in some cases that has an apples-and-oranges effect. The numbers *presumably* reflect average speeds for each vehicle, and we all know from our experience that a bicycle or scooter won't take us someplace as fast as a car on a freeway; so on the one hand it's perfectly legitimate to present them like this. However, we do need to keep speed in mind when making general statements that one form of transit is more efficient than the other. Here are some thoughts how speed affects these numbers (in no particular order):
* All the cars and buses and motorcycles are probably measured at 55 or 65 MPH; local trains probably less; electric scooters and bicycles certainly were measured at much lower speeds. At 65 MPH, most of the car or motorcycle motor's energy is expended overcoming aerodynamic drag; so they were going at the same speed as the scooters, they would look much better.
* Cars and scooters and bikes and buses all have rubber tires which have significant rolling resistance. There's a huge difference in resistance between racing-style bikes and ordinary ones; I wonder if DOE averaged the mix of bicycles actually in use well at all.
* Rolling resistance may predominate over aerodynamic drag at low speeds; without knowing the measurement speed it's not clear which factor is responsible for a difference.
* At a high enough speed for aerodynamic drag to predominate, a car does far worse than scooters or motorcycles or bicycles simply because its frontal area is much greater. Excellent streamlining can give a factor of 5 or more (Ca=.2 or less) better than a box with the same frontal area. Skin effect drag dependent on vehicle length is 100x less I think.
* Aerodynamics (plus friction in the motor and gears -- I know nothing about that) predominates for trains at much lower speeds, because rail rolling resistance is trivial compared with pneumatic tires. A long train amortizes frontal-area drag over many cars; assuming gaps between cars are filled well, the actual energy difference between pulling one or two or six cars is probably much less than you'd think; conductor pay and crime risks for a mostly-empty train are probably more significant cost factors.


Add new comment