Carbon Calculator: sources, references, and limitations
Here's what's behind the carbon footprint calculator.
All figures are for the U.S., and all tons are U.S. tons (short tons, i.e. 2000 lbs.) unless otherwise noted.
(1 short ton = 2000 lbs.)
Handy conversion factor: 1 g/km = 0.003545 lbs/mi.
Household energy use
- Average household electrical use, (920kWh/mo.), (Dept. of Energy, 2006, "Electric Sales, Revenue, and Average Price 2006", Table 5: U.S. Average Monthly Bill by Sector, Census Division, and State 2006)
- Pounds of CO2 per kWh (1.72) (EPA, Clean Energy, Calculations and References, Aug. 2008; note that the 2009 version now gives 1.30, but I haven't yet factored that into the calculator.) Formula given is 7.78 x 10-4 metric tons CO2 / kWh, and there are 2205 lbs. in a metric ton. I didn't use the following figures, but I'm including them here for comparison and for possible future reference. The Pennsylvania Dept. of Env. Protection (PPT) gives 2214, 1828, and 1425 lbs. of CO2e per MW of electricity produced by coal, oil, or gas respectively, with a weighted average of 2137, considering the mix of these fuels in the U.S. electricity diet. By "MW" I will assume they mean "megawatt-hour". 2137 lbs./MWh is 2.137 lbs./kWh, which is close to the EPA's figure.
- Tons of CO2 per household per year,
electricity component (7.1) (920
kWh/mo. x 12 mo. x 1.34 lbs./kWh ÷ 2000
lbs./ton) This is an oversimplification because
different parts of the country use a different
energy mix to produce electricity, some dirtier
than others. This is, of course, an average.
Natural Gas usage
- Natural gas used by U.S. residences (PDF) (4.4 million-million cf) (DOE, "Natural Gas Consumption by End Use",Natural Gas Delivered to Consumers by Sector 2007)
- Number of households using natural gas (PDF) (64 million) (Ibid.)
- Natural gas per household per year (68,750 cf) (4.4 million-million cf / 64 million households)
- Natural gas per household per year (688 therms) (68,750 cf x 1 therm / 100cf)
- Natural gas per person per year (14,667 cf) (4.4 million-million cf ÷ 300 million people)
- Natural gas per person per year (147 therms) (14,667 x 1 therm / 100cf)
of gas ($15.00/mcf) (DOE,
August 2008, Short-Term Energy Outlook)
I'm estimating the customer's gas
usage as [Price per month - $10 fixed charge] ÷
$15.00 = Mcf (thousand cubic feet) of gas. Then,
1 mcf x 1 therm / 100cf = 10 therms. WARNING: In
the industry, "Mcf" confusingly means thousand
cubic feet, not million cubic feet.
Natural Gas pollution
- Metric tons per therm (0.005) (EPA, Clean Energy, Calculations and References, Aug. 2008)
- (Short) tons per therm (0.0055) (0.005 metric tons/therm x 2205 lbs/metric ton x 1 short ton / 2000 lbs.)
- CO2 per household (3.8 tons) (688 therms * 0.0055 tons/therm)
- CO2 per person (0.8) (147
therms * 0.0055 tons/therm)
Or the hard way:
- Natural gas, pounds of CO2 per million BTU (117) (Dept. of Energy, 2007; Electric Power Annual: "Carbon Dioxide Uncontrolled Emission Factors"; also in Voluntary Reporting of Greenhouse Gases Program, "Fuel and Energy Source Codes and Emission Coefficients")
- BTU per therm (100,000) (EPA, 1997)
- Natural gas, pounds of CO2 per therm (11.7) (117 lbs / 1,000,000 BTU x 100,000 BTU / therm)
- Natural gas, tons of CO2 per therm (0.00585) (11.7 lbs. ÷ 2000 lbs./ton)
- Total fuel oil used by households annually (7.2B gallons) (Dept. of Energy, 2000, Fuel Oil Use in U.S. homes)
- Number of households using oil (8.1 million) (Dept. of Energy, 2007)
- Fuel oil per household (889 gallons) (7.2B gallons ÷ 8.1M households)
- CO2 from fuel oil (0.0121 tons/gallon) (EPA, Clean Energy, Calculations and References, Aug. 2008) Formula is 462.1 kg/42-gallon barrel, or 1017lbs./42 gallons, or 24.2 lbs./gallon, or 0.0121 tons/gallon
- CO2 per household (10.8 tons) (889 gallons x 24.2 lbs./gallon ÷ 2000 lbs./ton)
- CO2 per person (0.3 tons) (7.2B gallons ÷ 300M people x 24.2 lbs./gallon ÷ 2000 lbs./ton)
- Price of oil ($4.00/gallon) (DOE, August 2008, Short-Term Energy Outlook) I'm estimating the customer's oil usage as [Price per month - $10 fixed charge] ÷ $4.00 = gallons of oil. I assume the user enters the average winter bill and I multiply by only 4 months for winter instead of 12.
Driving: Tons of
carbon per household
- Pounds of CO2 from 1 gallon of gas, (20.4) (EPA, 2005; 19.4 for CO2 + 5.3% for equivalents of other gases)
- Number of U.S. households (99M) (DOE, 2001 data) The summary shows 107M total households, but that includes the 8 million who don't have a car, which I'm excluding.
- Gallons of gas used by households (113B) (DOE, 2001 data)
- Gallons per household per year (1,141) (113B ÷ 99M)
- Tons of carbon per household (11.6) (1,141 gallons x 20.4 lbs./gallon ÷ 2000 lbs./ton)
Driving: Miles traveled and gas mileage
- Residential miles traveled (2.3 trillion miles) (DOE, 2001 data)
- Miles per household per year (23,232) (2.3 trillion miles ÷ 99 million households)
- Miles per gallon (20.36) (23,232
miles ÷ 1,141 gallons)
Average fuel economy is reported higher, as 22.2 mpg (Dept. of Transportation, Table I-1, 2004), because actual performance is less than the standard set by the government, and the sticker that appears on the car. I include this note to let the reader know that I take this into account when figuring fuel use in the calculator. The calculator assumes that your actual fuel economy is 20.36 ÷ 22.2 = 91.7% of whatever figure you choose for your own car.
- Miles per month per household (1,936) (23,232 miles per household per year ÷ 12 months)
- Miles per month per person (1019) (1,936 miles per month ÷ 1.9 cars per household)
Driving: Tons of carbon per person
- Number of drivers (196 million) (Wikipedia, citing the DOT, 2003 data)
- Gallons per driver per year (591) (2.3 trillion miles ÷ 20.36mpg ÷ 196M people)
- Tons of carbon per driver (6.0) (591 gallons x 20.4 lbs./gallon ÷ 2000 lbs./ton)
Energy & carbon from food production
I found two sources for either the amount of energy required for each person's food, and the corresponding amount of CO2e, and these sources are almost spot-on in agreement with each other. I use the first figure.
University of Chicago
- Energy required to produce & distribute food per person per year (4x107 BTU) (Diet, Energy, and Global Warming (PDF), Gidon Eshel & Pamela Martin, University of Chicago, published in Earth Interactions, Vol. 10, paper no. 9, 2006, p. 3)
- Pounds of CO2 per million BTU (156) (Climate Trust, "2007 RFP Conversion Factors". Of course the figure is different for different fuels, but the range is actually fairly tight, 139 for LPG to 227 for the dirtiest coal available. The figure I picked is for automotive gasoline. Note that "Diet, Energy, and Global Warming" uses a rather peculiar conversion factor, by looking at the total energy used in the U.S. compared to the total GHG generated, according to DOE figures, as described on page 3. This works out to 154.35 lbs. per million BTU, as evidenced from their table where 6.8 x 107 BTU yields 4.76 tons of CO2 equivalents. I'm assuming they're using metric tons since it's a scientific paper.)
- Tons of CO2 per person per year from food production & distribution (3.1) (156 lbs. x 4 x 107 BTU ÷ 1,000,000 ÷ 2000)
- Tons of CO2 per household per year from food (8.1) (Food-Miles and the Relative Climate Impacts of Food Choices in the United States, Christopher L. Weber and H. Scott Matthews, Carnegie-Mellon University, published in Environmental Science and Technology, Vol. 42, No. 10, 2008, p. 3512)
- Number of people per household (2.67) (NY Times, 2007)
- Tons of CO2 per person per year from food production & distribution (3.0)
GHG from livestock
In addition to the GHG from the energy that goes into food production, livestock also produce methane from their waste and from bodily gases. It's not crystal-clear whether the Food-Miles paper includes these non-energy-related GHG emissions, but it doesn't appear to. The Diet, Energy, and Global Warming paper does tackle that separately:
- CO2e GHG produced by U.S. livestock annually (232.9 x 106 tons) (Diet, Energy, and Global Warming as cited above, p. 9; figure is for 2003 and based on 291 million people)
- Percentage of Americans who never eat meat (6.7%) (Harris Interactive survey commissioned by the Vegetarian Resource Group, 2006)
- Number of meat-eating Americans (271.5M) (291 million less 6.7%)
- CO2e GHG from livestock per meat-eating American per year (0.86) (232.9 x 106 tons ÷ 271.5M)
Energy for vegan vs. typical diets
Plant-based diets produce less GHG because meat requires a lot more energy to produce. The two sources below are nearly spot-on for the energy required for a vegan diet. I'll split the difference to get that result (2.2, 2.4=2.3).
For a typical (non-vegan) diet the sources disagree. For that I'm also splitting the difference between the two (3.2, 4.2 = 3.6), and then adding the first source's figure for GHG from livestock themselves (3.6+0.9 = 4.5).
I couldn't find good figures for tons per vegetarian, so I'm splitting the difference between vegan and typical (2.3, 4.5 = 3.4).
University of Chicago
- Extra carbon for typical American diet vs. plant diet, from production & distribution, annually (701 kg) (Diet, Energy, and Global Warming as cited above, p. 8)
- Same figure in tons (0.77 tons) (701kg x 2.2lbs/xg x 1 ton/2000 lbs)
- Total annual carbon for typical vs. vegan diet,
production & distribution (3.2 / 2.4 tons)
- 93.3%MT + 6.3%VT = 3.1 tons/person average
- (MT = tons of carbon per meat-eater; VT = tons of carbon per non-meat-eater; 93.3% & 6.7% are the percentage of meat-eaters and non-meat-eaters in the U.S. respectively; see sources above)
- MT = VT+0.77 (see above)
- VT = MT-0.77
- 93.3%MT + 6.3%(MT-0.77) = 3.1 tons
- 93.3%MT + 6.3%MT - (6.3%*0.77) = 3.1 tons
- 100%MT = 3.1+0.05 = 3.15 tons
- Total annual carbon for typical diet, production, distribution, and livestock effeccts (4.1) (3.2 from previous, plus 0.9 from "GHG from livestock", above)
- Fossil energy required to produce typical & vegan diets, daily (35,000 / 18,000 kcal) (Food, Energy and Society, David Pimental of Cornell University, p. 147)
- Kcal per gallon of gasoline (31,000) (How Stuff Works.com)
- Gallons of gas to produce typical & vegan diets, daily (1.13 & 0.58) (35,000/31,000 and 18,000/31,000)
- Gallons of gas to produce typical & vegan diets, yearly (413 / 212) (1.13 x 365.25 and 0.58 x 365.25)
- Pounds of CO2 from 1 gallon of gas, (20.4) (EPA, 2005; 19.4 for CO2 + 5.3% for equivalents of other gases)
- Tons of CO2e to produce typical & vegan diets, yearly (4.2 / 2.2) (413x20.4/2000 and 212x20.4/2000)
Their footnotes say that vegetarian and vegan diets use 42% and 72% less respectively compared to typical diets, and they cite the University of Chicago paper above, but it's not clear how they arrived at that conclusion. My calculations from that paper above show that vegan diets use 42% less energy, and I couldn't find anything in the paper which addressed the reduction from vegetarian diets specifically.
Comparison to flying
I used only the first item below in the calculator (or rather, in the Calculator Help, but I'm including the rest in case they prove useful in the future (to me or other researchers.)
- Food choices compared to flying. Per
capita flying as per above is 773,800 million
miles ÷ 300 million people = 2,579 miles/person.
Carbon from that is 0.47 lbs./mile x 2,579 miles
÷ 2000 lbs./ton = 0.6 tons of carbon. Amount
saved by going vegetarian as per previous note
is 4.5 tons - 2.3 tons = 2.2 tons.
CO2 from beef
- CO2 from beef (36.4 lbs per lb.) (Bill Miller; this is the most frequently-cited statistic in the media)
- CO2 from beef (32-54 lbs. per lb.) ("The Cheesburger Footprint"; 3.6 to 6.1 kg per 1/4-lb. of beef)
- CO2 from beef (21 lbs./lb.) (NY Times, "Rethinking the meat guzzler", Jan. 27, 2008; given as 2.2 lbs. per 100-watt light bulb for 20 days; 100 watts x 24 hours x 20 days = 48 kWh; 48 kWh ÷ 2.2 kg = 21.8 lbs./lb.)
- CO2 from beef (16.5-18.7 lbs./lb.) (Animal Science Journal, V78 #4, July 2007, pp 424-434, "Evaluating environmental impacts of the Japanese beef cow&endash;calf system by the life cycle assessment method"; given in the abstract as 4550 kg per animal; 4550 kg carbon. x 2.2 lbs/kg ÷ 536 lbs. animal as per note further down note=18.7 lbs; New Scientist says this same study gives a figure of 36.4 kg. of carbon per kg. of beef. They note that the study did not include the impact of managing farm infrastructure and transporting the meat. It's hard to say how applicable the study of Japanese beef is to U.S. beef production; here is criticism & response)
per beef animal (536 lbs.) (Gourmet
Sleuth, 569 lbs. less kidney, fat, suet, and
- Amount of global warming attributable to livestock (18%) (Food and Agriculture Organization of the United Nations)
- Agriculture's share of climate change gases in U.S. (PDF) (6%) (EPA, "Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2004", Chapter 6)
- CO2 for lamb, pork, chicken (PDF) (probably erroneously reported as 7.4, 6.35, 4.57 kg/kg) (I'm listing this source to warn other researchers who might come across those numbers here or here. The DEFRA PDF they cite and which I link to appears to have been misquoted. First of all, the units are given as "g 100 year CO2 equiv. per kg. of animal meat production" -- what the hell is "g 100 year"? Second, the table says that the animal weight is at the farm gate, which presumably means the whole animal, unprocessed.)
- Per capita meat consumption in U.S. (124kg total; 44 kg beef, 31 kg pork, 48 kg chicken, 1 kg other; 23 kg fish not included in total ) (American Journal of Clinical Nutrition, 2002, "Sustainability of meat-based and plant-based diets and the environment")
- Savings by switching from typical American diet to vegetarian (PDF) (1.5 tons/yr.) (Gidon Eshel & Pamela Martin, University of Chicago, published in Earth Interactions, Vol. 10, paper no. 9, 2006)
- Climate change and agriculture article (Wikipedia)
This calculator gives results 1.9 times higher than most other calculators, because we include the effect of radiative forcing, which means that plane emissions are more damaging than ground-level emissions for various reasons related to the planes being up in the sky. Tufts University's Climate Change Initiative has a good explanation of this on pages 26-27 of their Dec. 2006 PDF report, "Voluntary offsets for air travel carbon emissions," excerpted in brief:"The IPCC has estimated total radiative forcing of air travel to be 1-5 times larger in the stratosphere than in the troposphere and calculated the average for full radiative forcing to be a factor of approximately 2.7 (IPCC, 1999.) Therefore to estimate the impact of an airplane trip a multiplier should be used on the CO2 emissions from jet fuel to account for full radiative forcing. ...
"Although more research is needed to fully understand the chemical processes in the stratosphere, the research used by the IPCC is robust. We therefore recommend using a calculator that includes a multiplier for the increased radiative forcing in its carbon calculations."
Note that Tufts' 2.7x figure comes from an old 1999 IPCC report. We used that in an older version of this calculator, but now we're using a more recent figure of 1.9x. (Oxford University report (PDF), p. 7; Carbon Independent)
Note we don't know what percentage of passengers flying on U.S. airlines are actually foreign citizens, whose use shouldn't be included in our estimate of the impact of the average American. Similarly, many Americans fly on foreign airlines (even when leaving the U.S.), and their travel isn't included in the stats. But we have to use some figures, so we simply use the number of air miles traveled on U.S. airlines, hoping that our citizens flying foreign airlines and foreign citizens flying our airlines is close enough to a wash.
- Miles flown by a typical American per year (0). Most Americans do not fly in any given year. (Only 43%; Gallup, 2007) I'm therefore considering that the carbon output from flying for a typical American is zero.
- Number of people per household (2.67) (NY Times, 2007)
- Total passenger air miles traveled: (848 billion) (Bureau of Transportation Statistics, June 2007-May 2008)
- Portion of air miles that are personal (non-business) (70%) (Washington Times, 2008; article says that 30% of airlines' revenue is from business customers)
- Air miles per household (5,548) (848B miles ÷ 107M households x 70%)
of CO2 per mile of air travel per
passenger (0.39-0.64) (CarbonNeutral.com,
2007) I used the figures from An Inconvenient
Truth and Firm Green, both of which got their
numbers from the GHG
which bills itself as "the most widely used
international accounting tool for government and
business leaders to understand, quantify, and
manage greenhouse gas emissions." Those figures
- 0.64 lbs./mile for short flights (<727 miles) (Note that the BBC gives 170-240 g/km for short-haul flights, which works out to 0.60-0.85 lbs./mile.)
- 0.44 lbs./mile for medium flights (728-2575 miles)
- 0.39 lbs./mile for long flights (>2575 miles)
- Radiative forcing factor (1.9) (Oxford University report (PDF), p. 7; Carbon Independent)
- Hours to Miles conversion. I use "hours" in the calculator because people will likely have a better idea of how many hours they fly vs. how many miles. To convert hours to miles, first I made a table of air miles between various cities, using 15 city pairs ranging from 276 miles to 6440 miles. Then I fired up Orbitz and found the flying time between those cities. I then tinkered with formulas until I came up with one that seemed to fit most cities reasonably well: 200 miles for the first hour of flight, and 525 miles for each additional hour. I figure each flight is the average of the flight miles divided by the number of flights.
- Airplane passenger MPG (43). U.S. aircraft used 19,704 gallons of fuel and traveled 848 billion passenger miles in June 2007 to May 2008, which works out to 43 pMPG. Note that David Lawyer gets 45.6 pMPG for 2005 using the DOE's Transportation Energy Data Book.
- Bus passenger MPG (125). (From David Lawyer. Greyhound says its fleet gets 184 pMPG but figures supplied by industry should always be viewed with skepticism.)
- Train passenger MPG (45). (From David Lawyer.)
- Science Daily, 2008
Figures I didn't use, but which I'm saving in case I need them for the future
- U.S. population, (300M) (Census Bureau, 2006)
- Total U.S. car miles traveled, (1,700,000M), (Bureau of Transportation Statistics, 2004) This differs significantly (35%) from the 2.3 trillion miles given by the DOE.
Results are useful
Before I discuss the limitations which might make you think the calculator is worthless, let me give you the good news: I think the ratios are valid. Most calculators give different results from each other, but the ratios between those figures are fairly consistent. That means you can compare different parts of your lifestyle (say, driving vs. eating), and compare your total use with the average American's or the average world citizen. It's not the raw number of tons that really matter, it's how those numbers stack up against the other numbers. In other words, if you see that one number is half the size of another, or 3x the size of some other number, you can be pretty confident that that ratio is correct. And you can enter different numbers into the calculator to get an idea of what percent they might reduce your impact.
So with that out of the way, let's talk about some limitations and challenges in calculating the data.
Do we count everyone, or just those who use a certain kind of energy?
The biggest problem is where I try to show you the amounts used by the average person or household. Let's take air travel. How do we figure how much the average American fliies? Our first thought might be to take the total number of passenger air miles flown divided by the population of the U.S., multiplied by the percentage of airline trips that are personal, as opposed to business (70%). That would give us 1,979 miles per person. But you know what? Fewer than half of Americans fly in any given year. So for the typical American, the amount of air travel per year is zero. But if I list average air travel per person as zero, then I'm failing to show you how wasteful Americans are as a whole when it comes to energy use.
Okay, so let's say I use that average figure of 1,979 miles per person. Now when you're entering your own air travel figures, you see that the "average" American flies 1,979 miles, and you feel all happy because you didn't fly at all, so you think you're doing better than most people. But actually, you're not, because most people don't fly. So you're actually in the same camp as most people, not better, although the numbers in the calculator lead you to believe otherwise.
(There's no good solution for this, so my choice is that when calculating for an individual I show the average as 0 miles, and when calculating for a household, I divide total miles flown by the number of households -- though the typical family flies less than this, and those who do fly, fly more. And there are yet more problems with figuring air travel besides the ones just discussed, which are listed in the air travel sources section above.)
Another problem is that passenger planes typically haul cargo and mail (besides passenger baggage). How do we account for that? (In my case, lacking good information for what percentage of hauled weight is cargo vs. passengers, I don't even try.)
In a similar vein, how the hell do we figure the average amount of fuel (heating) oil used by an average household? Most households don't use fuel oil, so we have all the same problems as the above. What most calculators seem to have done is to consider only the households that actually use oil. Thus the average usage shown is a lot more than the average usage of all American households. If we included that number in the total usage for the average household then we'd be overstating how much the typical household uses. So some calculators, including mine, show you the amount of oil used by oil-using households when you're entering your own figures, but the total for the average household is calculated independently from the line items.
Perssonal consumption vs. consumption from infrastructure
Then there's the question of what to count. On one extreme is Nature.org, which tries to incorporate the energy needed to build and maintain roadways when they're figuring your driving footprint. I'm not sure that's really fair, since the roadway still exists even if you stop driving. And if you start riding a bike, you'll still use some roadways.
Most calculators don't take Nature's route, and instead look at your personal consumption only. However, many estimates of per-capita consumption are done by dividing a whole country's carbon output by the number of citizens, making each person partially responsible for all energy consumed in the country, no matter how or where it happened.
Households vs. individuals
Another problem is how to accurately count household vs. individual use. In fact, it looks to me like CarbonCounter.org ran headfirst into that problem, as their Sources section mixes results, listing some things per household and others per individual. For my calculator, to calculate individual use I first took the household use and divided by around 2 -- not the actual 2.67 people per household, because when you have more people in a house, you have more economies of scale. For example, both a 3-person household and a 1-person household have a refrigerator, that will use a similar amount of electricity. So I consider a small "penalty" for living alone.
What's a ton?
Finally, a "ton" means different things depending on whom you talk to. In the U.S. a ton is 2000 lbs., but in the U.K. it's 2240 lbs. And then there are metric tons which are 2205 lbs. In researching my sources I tried to make sure I was using U.S. tons consistently, and converting other types of tons into short tons (U.S. tons) when comparing my figures with other calculators', but possibly I missed the conversion in a couple of cases. Fortunately, the difference between various kinds of tons isn't all that great.
When we look at different calculators, we see that they come up with different results for the same stuff. There's a fairly wide range, but my results are in the middle, leading me to believe that they're fairly reliable, since the extremes on either end are less likely to be accurate. My total for individual use is 16.6 tons (vs. others at 7.5, 8.0, 21, 22, 22.5, 23.8, and 27), and my household total is 38.2 tons (vs. others at 27.7 and 66.8). I have tables comparing my results to other calculators at the bottom of the calculator page. Of coursse, for transparency, I clearly detail all my sources and calculations (above) so you can see how they were done. And as I mentioned at the top of this section, even though the numbers differ across calculators, the ratios appear fairly consistent, so I think you can successfully compare different parts of your lifestyle (e.g. driving vs. eating), or your total use to the average total use.
My ratio for pollution from home energy vs. car energy is right in line with what Google calculated. They said a 10% reduction in energy use in 6 households is like taking one car off the road. Therefore 60% of home energy carbon should equal 1 car's worth of carbon. My calculator shows 9.5 tons of carbon from electricity for the average home, and 6 tons for the average car. So 60% of 9.5 is 5.7 tons, very close to my 6 tons figure.
This page last modified June 2013