Vegans vs. Hybrids

Time for another of the classic Fat Knowledge Vs. posts. Which has the bigger impact on the environment (specifically CO2 emissions): switching from a heavily meat based diet to a vegan diet or switching from a Hummer to a Prius?

I came across this report: Diet, Energy and Global Warming and thought it had everything I need as they were comparing diets with cars. But, when I looked into the report, I couldn't figure out how they came to their conclusions. And if I can't understand it, I don't want to post it on my blog.

Oh, but don't you worry loyal Fat Knowledge readers. I don't give up that easily. :)

I dug in a little deeper and tried to understand how they were attacking this problem (in my opinion they are using a very unintuitive method). Then I was further stymied by the fact they had mathematical errors in their calculations. So, I decided to ignore their analysis and instead take their raw numbers and do the calculations myself.

You can download my calculations in an .xls format or view it in Scribd.

So now, lets get back to the original question, which has a greater impact on reducing CO2 levels, switching from a heavily meat based diet to a vegan diet or switching from a Hummer to a Prius?

First, lets take a look at the impact of diet.

Diet	Tonnes of CO2 per Year
Avg American	2.19
Vegan	.19
Lacto Ovo	1.22
Mad Meat Eater	6.70
Avg American on Diet	1.44

All diets eat 3,774 calories a day, except the "Average American on a Diet" which eats 2/3s as much or 2,490. The 3,770 number is a bit misleading as it includes food that is wasted or thrown out, which accounts for somewhere between 14% and 50% of all food.

In the Average American Diet 5% of calories from from chicken, 11% milk, 1% eggs, 3.2% beef, 5.6% pork and the rest from plants. In the Vegan Diet all calories come from plant sources, and I use the .03 lbs CO2/100 Calories as an average value for all plant sources. The Lacto-Ovo is a vegetarian diet with 24% of the calories from milk, 4% from eggs and the rest from plants. The Mad Meat Eater is an estimate for an extreme meat based diet, with 20% of calories from beef, 20% from pork, 5% from chickens, 5% from eggs and the rest from plants. Take a look at the spreadsheet more details.

Now to compare this with cars, I will assume the average driver goes 12,000 miles a year and I will use estimates of 12.5 mpg for a Hummer (the comments on that link are hilarious), 25 mpg for a Camary and 50 mpg for a Prius. At 8.8 kg of CO2 emissions per gallon that works out to a yearly usage of 960 gallons and 8.5 tonnes of CO2 for the Hummer, 480 gallons and 4.2 tonnes of CO2 for the Camary, and 240 gallons and 2.1 tonnes of CO2 for the Prius.

Which has the greater impact, changing your car or your diet?

Going from a Mad Meat Eater diet to a Vegan diet saves 6.5 tonnes of CO2 a year while going from a Hummer to a Prius saves 6.4 tonnes. Given a margin of error on the values, I call that a tie.

If you switch from the Avg American Diet to being a vegan that saves 2 tonnes very similar to the 2.1 tonnes saved switching from a Camary to a Prius. Going from the Avg American Diet to a Lacto-Ovo vegetarian diet saves 1 ton a year, about 1/2 the savings of going from a Camary to a Prius.

In looking at their overall impact, the average American diet emits 2.19 tonnes which is almost the same as the 2.1 tonnes for the Prius. The Mad Meat Eater Diet emits 6.7 tonnes, halfway between a Camary and a Hummer. A vegan diet at .19 tonnes is less than 1/10 that of a Prius. In general, for the average American, the impact of their car will be greater than the impact of their diet.


CO2 equivalent emissions of various foods

Food	lbs CO2/100 Cal
chicken	0.37
milk	0.62
eggs	0.64
beef (grain fed)	3.04
pork	1.99
lamb	5.71
fish (avg value)	1.33
herring	0.06
tuna	1.05
salmon (farmed)	1.07
shrimp	6.79
veg (avg value)	0.03
corn	0.02
soy	0.01
apple	0.06
potatoes	0.05

These values take into account the methane emissions of animals as well as the fossil fuels needed to raise them. As you can see beef, lamb and shrimp have very high emissions, while all plant based foods are low, especially soy. In fact, if you were to trade 640 Calories of beef for soy, this would emit 22 lbs less of CO2, the same as using one gallon less of gasoline.

I was surprised by the fact that seafood emits so much CO2 per calorie. Shrimp in particular is high, more than any other food on the list and double that of beef. I have no idea why so much fossil fuel energy is needed to raise shrimp. Tuna is higher than milk, eggs or chicken. I guess it takes lots of energy to power the boats that capture the fish.

Assumptions and Caveats:
1) The fossil fuel usage for the various foods comes from Pimentel. This guy's estimates are always on the high end for fossil fuel usage. I first ran into his work when looking at ethanol. When compared to all others that have looked at the energy needs to create a gallon of ethanol, he is by far the highest. This makes me very leery about using his estimates here.

2) It is not clear to me how they came up with the average value for fish.

3) In the report they use "e values" of 1,2, and 4 for plants in one graph and doesn't specify the value in the other. This value determines the fossil fuel usage. I chose to go with the value in the middle of 2, which corresponds to .4 kCal of fossil fuel energy per kCal of food, and gives you a value of .03 lbs of CO2 per kCal. Soy is much lower and if more calories came from soy, the resulting CO2 impact of that diet would be lower than estimated here. In general all plant based foods use little CO2 emissions compared with animal products, so even if the estimate is off by a factor of two, it will have limited impact on the results.

4) The report estimates CO2 emissions from fossil fuels based on .07 tonnes CO2 per MMBtu. While I don't agree with the process they used to determine that number, it corresponds well with the value of .0703 for motor gasoline as determined by the EPA. Natural Gas emits .0528 and bituminious coal .0925. So, the more that fossil fuel usage is from natural gas, the lower the values for CO2 emissions should be. The more coal is used, the higher they should be. The .07 value works if natural gas, motor gasoline, and coal are used in roughly equal amounts, which seems to be a reasonable estimate to me, but could be off.

Mathematical Errors
The report incorrectly calculates weighted averages. First they do it in table 1 (page 28) on average mpg values by multiplying the mpg values by their weight rather than the inverse of the mpg values. The impact is minor, but shows they don't understand the underlying mathematics. Next they do the same thing in table 3 (page 30) where they are calculating the weighted mean of different diets. I don't know how significant the impact is, but it could be large. I chose instead to just attack the whole thing differently.