We’ve gotten a few questions already about how we came up with our charts on the climate-warming impacts of travel choices. The charts compare the global warming impacts of cars, SUVs, vanpools, planes, buses, and trains. And since I couldn’t find a single, unified data source for all of that—at least, not one that I couldn’t poke some holes in—I compiled our figure from scratch, using a bunch of different sources.
So in case you’re the sort of person who gets excited about numbers and footnotes, there’s plenty—plenty, I say—of data wonkery to follow…
Find this article interesting? Support more research like this with a gift!
First off: the data for these charts represent mid-point estimates for both direct and indirect climate-warming emissions from travel. They do include an estimate for carbon dioxide emissions from extracting, transporting, and refining crude oil, as well as emissions of trace climate-warming gases in vehicle exhaust. The air travel emissions also include the effects of contrails, cirrus cloud seeding, and other wacky airplane stuff. But the estimates don’t include emissions from vehicle construction and maintenance, or any of the life-cycle emissions related to building roads and parking lots, insurance company operations, etc. Including these other life-cycle emissions would push up the totals a bit, but probably not enormously.
Second, the geekery:
- Petroleum fuel cycle emissions represent a mid-point estimate from fuel cycle models, including the GREET model, which is maintained by the Argonne National Lab. See table 14-1 of this pdf for upstream emissions estimates. In addition, our emissions estimates include trace gases such as NOx from tailpipes, which add roughly 5 percent of CO2 equivalents to total tailpipe emissions, per p. 4 of this EPA report. By our final count, burning a gallon of gasoline releases 25.8 pounds of CO2 equivalents into the atmosphere, including about 19.5 pounds of CO2 from the tailpipe itself. For diesel, it’s 27.8 pounds of CO2 per gallon total, including 22.3 pounds of CO2 directly from the tailpipe.
- Real-world passenger vehicle mpg is taken from a variety of sources. My Prius figure (46.4 mpg), is the average of values from here and here; the second link has real-world data for other efficient vehicles, too. If you care, Hummer H2 fuel efficiency (10.2 mpg), along with real-world mpg ratings for other efficiency duds, can be found here.
- More generally, passenger vehicle mpg for cars and light trucks is from this spreadsheet from the US Bureau of Transportation Statistics, which pegs light trucks at 16.2 mpg and passenger cars at 22.9 mpg, as of 2005.
- Aircraft emissions are tough to calculate in the abstract. Not only do estimates vary, there are also a lot of variables in play—how many seats are filled, how far and how high a plane travels, its route, the time of year and time of day, flight delays and reroutings, etc. This awesome analysis of online air travel emissions calculators suggests that Atmosfair.com is the best of the lot, so I averaged Atmosfair’s values for a short, medium, and long plane flight, coach class from Northwest cities. Emissions work out to just under a pound of CO2 equivalents per mile of point-to-point travel. But note that direct CO2-only emissions from air travel are actually quite a bit lower than the figure suggests—and usually are somewhat better than the per-passenger emissions from car travel. Much of the net impact of air travel aren’t from CO2, but from other gases and contrails. Herearesomegood sources for CO2-only emissions from air travel; but note that they disagree. This spreadsheet from GHGProtocol.org also lets you calculate your direct emissions from travel, including air travel.
- For rail and bus transit, as well as vanpools, fuel consumption and ridership figures are derived from tables 17 and 19 of the National Transit Database, a program of the US Department of Transportation. I’ve used King County, WA figures for buses, and national averages for rail transit (combining light, heavy, and commuter rail). On its face, bus transit doesn’t do all that well, since many buses around the US run almost empty. Late-night buses carrying only a few passengers are pretty much efficiency duds. But a full bus—like many of the rush hour buses in Cascadia—is quite efficient. So when it comes to the climate, filling seats is just as important as choosing the right kind of vehicle. And just for the curious—transit buses average 10.7 riders per vehicle mile traveled in King County, and trains average 25.9 riders per train car. Trains genuinely do attract more riders than buses, but much of their advantage stems from the fact that trains tend to be built in dense, transit-friendly neighborhoods, while buses often service more sparsely populated suburbs.
- Train travel poses a puzzle, since trains—particularly transit trains—often run on electricity. And electricity is slippery—when you turn on the lights, there’s no telling where the juice really came from. Which means that, in effect, the climate impacts of train travel depend on where you think the electricity comes from. Using national averages for the generation mix gives lots of credit for hydropower and other low-carbon generation. On the margins, though, much of the nation’s electricity comes from coal, which is much worse for the climate than “average” power. On the other hand, “peak” electricity—the marginal power produced at times of day when demand spikes—typically comes from natural gas, which is generally less carbon-intensive than the US average. I’ve used US averages here; but other, arguably equally valid assumptions would yield different estimates. Also, there’s one potential source of error: for lack of easily accessible data, I assume that the upstream and trace emissions from electricity—the energy used to extract and transport coal and gas, for example, or trace emissions from coal combustion—accrue at the same rate as the upstream and trace emissions for diesel fuel. It’s a wee bit sketchy, really, but it solved the problem of accounting for Amtrak, which runs on both diesel and electricity.
- Amtrak figures come from the “CO2 emissions from transport or mobile sources” spreadsheet at GHGProtocol.org. Note that these agree pretty closely with the figures used by the Transportation Energy Data Book, with intercity train travel assumed to be predominantly diesel.
- I chose a nominal value for walking and biking, as well as additional passengers. However, the calories from walking and biking have to come from somewhere — and the food system, from farms to stores to refrigerators and stoves, consumes lots of energy. So some people have suggested that, mile for mile, walking and biking aren’t quite as benign for the climate as is commonly thought. I’m not sure I agree, but I’ll leave this analytical dilemma as an exercise for the reader.
I think that’s all for now. Also, many, many thanks are due to former research intern Justin Brant, who did most of the early research for this project!!
Matt the Engineer
Great stuff, Clark. I’m a bit torn about counting contrails (which I see you left off for other reasons), because the effect ends the day you stop putting airplanes in the air. It’s not an additive effect like CO2 – every additional contrail adds to global warming as long as it’s in the sky whereas every pound of CO2 stays in our environment forever.On the other hand, contrails have a real effect and any extra warming could work to push us over some tipping point.
Great calculations, Clark. Very useful. A couple of related documents from the American Public Transit Association are 1. Public Transportation and Petroleum Savings in the U.S.: Reducing Dependence on Oil2. Public Transportation’s Contribution to Greenhouse Gas ReductionThe argument that walking is bad for the environment is bogus for a lot of reasons. Here’s one: A pedestrian walking at a slow stroll burns calories at the same rate (2 kcal/kg/hr) as the body of a car driver. In some cases it takes even more calories for a person to operate a vehicle than to walk, depending on the vehicle and the speed of the walker. Since most walking trips are completed in less time than most driving trips, on that basis alone walking trips take less energy. Then you start adding in gasoline consumption and the lifecycle energy costs of automobile infrastructure (highways compared to sidewalks), and there’s no contest.
Good points, Laurence. A simpler reason to ignore the arguments about the energy consumption associated with walking and cycling occurs to me.The calculations that suggest relatively high fossil energy inputs for human-powered transportation are really just pointing out that we’ve got an energy-intensive food system. The solution to this problem is to reduce the energy intensity of the food system. It’s not to exercise less. Indeed, we should almost all exercise much more than we do.Furthermore, I suspect there’s only a loose link between exercise and calorie intake. Just for myself, I do probably eat a little more when I bike or walk a lot on a particular day. But the main determinant of my daily calorie intake is probably an independent variable: whether there is ice cream in my freezer or not.Generalizing from this anecdote, I suspect that calorie availability—the so-called “food environment”–dwarfs exercise levels in determining calorie consumption. I’m also impressed by the evidence that our calorie consumption is directly proportional to how many people we are sharing a meal with: the more companions, the more we eat. We would not say, however, that people should all eat alone, to save fossil energy. (There’s evidence that family dinners are important to children’s development and education.)Anyway, I’m just guessing here. I haven’t studied the literature. But the argument that “walking burns fossil fuels” seems like sophistry to me.
Rail transport (and road transport) vs. air/sea. You really need to include a figure for trackway/roadway maintenance. One of the great boons of air/sea transport is that you don’t have to maintain your rights-of-way, just your port facilities. This difference can’t be ignored. Money has to be paid to staff to do this maintenance/new build, and that money will be spent on stuff which causes further emissions (eg air travel!). It’s the main reason that air travel is still competitive with road/rail in financial terms. Now if road/rail staff were paid in vouchers that could only be used for low-energy-intensity products (such as further rail travel) then we might save more energy and get close to the figures you now quote.- Stephen, rail fan and transport economics realist, England.
Great post, but you failed to include the pollution spewed out daily by ocean going freighters. They have been deemed as the single most polluting source of all our ways to transport anything. I would love to see a follow up blog concerning this topic. thanks 🙂
Stephen – It seems fair to include emissions directly tied to maintenance of infrastructure in emissions totals. However, the idea that these totals should further include emissions related to the spending habits of those working in these sectors is absurd. All sectors have related staff with likely broadly similar spending habits, and it’s not like we want to penalize employment in general. If the voucher system you suggest made sense, it would have to apply to everybody, not just rail workers, but undoubtedly there are better ways to make prices reflect GHG -related environmental damage.
I love the data and agree heartily with the choice of atmosfair as reflecting the best science on air travel ghg emissions.However the chart itself has a large anti-car/pro-flying bias. It compares apples to oranges and the result is very misleading. The chart seems to show flying as somewhere in the middle of driving options. It is possible for flying to be lower ghg than driving the same distance…but it is very unlikely and very uncommon. The chart shows flying averages but not car averages. The flying bar is average for commercial airliner at average capacity. It should compare this to average car at average capacity going same distance. What would that look like? The average car is around 20mpg and the average capacity for everyday car use is 1.6 to 2.2 (us govt stats). However average occupancy for long car trips is even higher. Choosing 2 as average is conservative. So if you want to accurately compare “airplane” to “car”, a striped “car” bar is needed and it would be somewhere in the Prius range or lower. In otherwords, “car” is far more efficient than “plane”. Your chart doesn’t give this impression at all. I suggest a new chart comparing averages to each other like this.A second bias is the use of outlier data for cars but not for planes. You top your chart with worst-case car scenario. The SUV with one person is far below average mpg and far below average occupancy. To be accurate you should include a similar outlier bar for air travel. It’s easy to do. I went to atmosfair just now and checked emissions for seattle to dallas flight. Flying “business” class increases emissions to 140% of “average”. Flying on a Gulfstream jet cranks it up to almost 4 times as much as “average”. And that is with standard occupancy. You then need to also adjust for similar low-occupancy. (note: Private jets are booming and are fastest growing sources of airline emissions. Congress has held hearings on this very topic. It’s a reasonable comparison to solo SUV considering the number of people flying half-way across nation in private jets vs those driving alone in SUV each day.) How about adding some outlier bars for plane travel to balance your car ones? I recommend three:1) HALF-FULL PLANE: a 50% of average-occupancy of average-plane. That would be double the current “airplane” bar and would compare to your “car (solo)” which is also ~50% of average-occupacy of average-car over same distance.2) HALF-FULL GULFSTREAM: a 50% of average-occupancy of Gulfstream jet. That would be about 7.5 times the length of your current “airplane” bar and would compare to your “SUV (solo)” which is ~50% of average-occupancy of low-mileage car over same distance.3) ALMOST FULL HIGH-MILEAGE PLANE: a higher-than-average-occupancy, high-mileage plane. Atmosfair showed the lowest-ghg Boeing plane for this flight to be Boeing B737-400 with 92% of average emissions. I’ve read that average occupancy is around 80% so i’d ballpark a 20% occupancy increase and decrease per-passenger emissions a similar amount. That would put this bar in solo Prius territory.If your goal is to provide accurate comparisons to allow people to easily see the relative costs of each…then you need equivalent bars for flying and driving. Both average and outlier…so people can compare similar things.
I got curious so i made a quick chart of different modes of travel based on average occupancy (as discussed in my previous comment). This allows real world, like-to-like comparisons of the various modes.I posted this chart on my website: http://www.stonebreakerdesigns.com/reports/modes-of-transport-ghg.jpgIt clearly shows plane travel is much worse on average than car travel in real world use. Average commercial air trip is 40% more ghg per passenger mile than average car trip. Average business class air trip is 40% worse than average SUV trip. Average Gulfstream private jet trip is…well, off the charts. I think it is important to have this real-world occupancy chart first. Apples to apples.Next would be to throw in equivalent high/low scenarios in both directions. For example low occupancy and high occupancy for each mode. If i get time to do that i’ll post it in the same image referenced above.Data in my chart is based on data in your chart plus occupancy stats from US Dept of Energy and aircraft emission scenarios from atmosfair.com.
Considerations public policy options for reducing ghg emissions from travel:1) Currently, in real world use, flying averages 140% of ghg emissions of driving same distance. 2) At maximum occupancy, average airplane has triple the emissions per-passenger-mile of average car. 3) Today’s auto technology provides much larger possible emissions reductions per vehicle than plane technology does. For example, the best Boeing plane listed in atmosfair for seattle to dallas flight has six times per-passenger-mile emissions when full than a full Prius does. 4) Increasing occupancy of transit dramatically increases the ghg gap between driving and air travel. Driving has much greater potential to respond to ghg reduction policies.My charts of average occupancy and maximum occupancy can be seen at http://www.stonebreakerdesigns.com/reports/modes-of-transport-ghg.jpg
FROM YOUR CHART PAGE: “The best strategy for reducing your impact: walk, bike, or fill up a seat that’s already going your way!”This is certainly true for walking, biking and *land* transport choices. However i’m concerned it leads to increased emissions when applied to air travel. The main excuse i hear from family and friends for why they still fly when they know the emissions are far higher than other choices is: the plane is already going there, I’m just filling an empty seat.Here is a similar quote from a recent WSJ article on climate change impacts of flying: “Chances are you are just taking a seat, not adding flights to the schedule” ( http://online.wsj.com/article/SB120061177398199015.html )Using this immediate “fill a seat” justification for flying emissions will never lead to decreased demand in the worst of all multi-passenger emission modes: flying. In fact the same article says domestic fliers increased 14% last year.Perhaps you could re-phrasing your recommendation to: “The best strategy for reducing your impact: walk, bike, or fill up a land-transport seat that’s already going your way!”
The only “1.6 average vehicle occupancy” figure that I can find comes with the caveat that it’s total people traveling divided by vehicles; that is, it includes people who aren’t even in cars to begin with. Actual field surveys in the U.S. pretty consistently find AVOs in the 1.2 range, and the empty HOV-2 lanes surrounding many U.S. cities (so under-used that several cities now permit paying solo drivers) attest to that.
Here are some occupancy of vehicles sources i’ve used.US DOE: http://www1.eere.energy.gov/vehiclesandfuels/facts/2003/fcvt_fotw257.htmlPickup: 1.46Auto: 1.57SUV: 1.76Van: 2.22EU-based report on aviation impacts: http://www.climnet.org/EUenergy/aviation/2006-06_aviation_clearing_the_air_myths_reality.pdf“The occupancy rate of cars (and lorries) at distances competing with aircraft (i.e. long hauls) is higher than the average occupancy rate of 1.6.””Per passenger kilometre, aviation is three to ten times worse in terms of climate impact than cars on competing distances.”This report is extensively footnoted with primary research studies.
Payton’s points about empty HOV-2 lanes and low-occupancy for city driving are perhaps valid but are not relevant to comparing driving vs. flying impacts. To compare flying to driving you need to consider the equivalent trip. For medium and long haul trips, stats show vehicle occupancy as higher than around-town driving.That is one of my fundamental problems with the chart above. It is comparing apples and oranges: average real-world occupancy for medium and long-haul flying but not average occupancy for medium and long-haul driving. Furthermore it shows low-occupancy emissions for driving but not for flying. It shows low-efficiency driving options but not low-efficiency flying options. The result is several layers of bias in favour of flying vs. driving.
Thanks for your data sources, Clark. I used them to create the chart I was hoping for: showing per passenger emissions for many vehicle types AND many levels of occupancy. People can use this to plan low-carbon ways to get from here to there.This chart was just published in the latest issue of the Watershed Sentinel, a BC environmental magazine. Readers can view it online at http://www.saxifrages.org/eco/show1h0s/Climate_Emissions_by_Type_of_VehicleThanks again to Sightline for getting interesting research out into the public via your Daily Score blog.
People often forget to include scooters, mopeds, and motorcycles in the transportation mix. Modern scooters (150cc) get about 80 mpg and mopeds get over 100mpg. Although they have a reputation for being dirty, I think that comes from the old two-stroke mopeds and scooter.I should disclose that I have an interest in promoting scooters, as I co-own a scooter dealership (Scooter Gallery) in Seattle. But the main reason I’m involved in that business is because I believe scooters and mopeds offer a more efficient alternative to single-occupancy cars for people who aren’t able to or don’t want to walk, bike, carpool, or use transit for every trip.
When comparing walking & cycling with SOV travel, the CO2 expelled as part of normal respiration will not be appreciably more than than the motorist expels as they sit in their car–motorists have to breath too! So, the main difference in contribution to CO2 per pedestrian or bicycle mile is in the energy embedded in the production of the bicycle as motorists will also generally be wearing shoes!
What I want to know is the # of ounces of CO2 per average commuted bicycle mile. This figure belongs on the back of my bike commuting jacket!