Waste not want not: Understanding the greenhouse gas implications of diverting waste and residual materials to biofuel production
If we use livestock feed for biofuels, what will the cows eat?
Around one-third of all biodiesel produced in the U.S. is made from waste feedstocks like used cooking oil, animal fats, and inedible corn oil. But few waste feedstocks are really wastes; none of these materials was thrown in a landfill before the advent of the biodiesel market. The U.S. Environmental Protection Agency (EPA) recently acknowledged this fact and has begun counting emissions from displacing distillers sorghum oil used in biodiesel from livestock feed in response to ICCT comments. When corn and sorghum are used to make ethanol, large quantities of distillers grains and solubles (DGS) are left over from the process and are typically fed to livestock. Increasingly, ethanol producers are pressing corn and sorghum oil out of the DGS before they are mixed into livestock feed, and selling the oil to biodiesel companies. As a result, farmers must make up the lost calories in their feed. The question is, with what?
This matters because different potential replacement feed ingredients have different climate impacts, and this will change our understanding of the net greenhouse gas (GHG) effect of using these waste oils for biodiesel. For example, EPA estimates the GHG emissions from producing additional palm oil for livestock feed would be around 25% higher than if sorghum grain were used. The difference across feedstocks is much larger according to modeling done for the California Air Resources Board and the European Commission.
To better determine the climate effect of using waste oils for biodiesel, I looked at historical data in the U.S. to parse out what ingredients have replaced inedible corn oil in livestock feed. I focus on corn oil in livestock feed in particular because there is enough data available to accurately track changes in its use over time, and because corn oil may be a fair proxy for other waste fats and oils used in livestock feed.
The chart below shows the amount of inedible corn oil included in DGS fed to U.S. livestock over time in purple. Inedible corn oil in livestock feed increased from 2000 to 2011 as U.S. corn ethanol production ramped up, increasing the amount of DGS produced. From 2011 onward, corn oil has been increasingly pressed from DGS for biodiesel production, reducing the amount fed to livestock.
What feed ingredients were displaced as inedible corn oil ramped up from 2000-2011 and what took its place as corn oil in feed dropped after 2011? Common sense suggests that corn oil in livestock feed might be replaced by other oils, since low-fat livestock feed is metabolized less efficiently in most types of livestock, apart from dairy cattle. However, virtually no virgin oils—soy, palm, or otherwise—are used in livestock feed in the U.S. Instead, according to U.S. Department of Agriculture data, whole oilseeds, rather than refined vegetable oils, tend to be crushed into livestock feed. Soybean and cottonseed are most commonly added, in addition to very large quantities of soymeal to boost protein content. Indeed, the graph does show whole soybeans and cottonseed following an inverse pattern to the use of inedible corn oil in livestock feed, declining from 2000 to around 2011 and increasing thereafter, and regression analysis confirms an inverse correlation. However, if the oil in soybeans or cottonseed substitutes for corn oil, we would expect to see an increase in the use of meal concomitant with the reduction in oilseeds in feed over 2000-2011, with the reverse patterns from 2012-2017, because any changes in oils in feed shouldn’t change total protein needs. We don’t see these patterns in the data, suggesting that the use of whole oilseeds in livestock feed is driven by other factors, such as total cotton or soy demand.
The other obvious possibility is that the calories in corn oil are substituted with the largest source of calories in feed overall: corn grain. While corn oil and corn grain, which is mostly starch, don’t have exactly the same effect on livestock metabolism, EPA argued in its final rule on distillers sorghum oil that the loss in oil from DGS with biodiesel production is probably not large enough for farmers to deliberately add more oil to feed. If true, farmers would simply use more primary feed ingredients overall, and the vast majority of this would be corn grain. There is a strong and significant inverse correlation between corn grain and corn oil in livestock feed. The overall changes in the use of corn, shown in brown on the right-side axis of the chart, are much larger than the changes in corn oil in feed, so corn oil cannot be the only factor affecting total corn use. Overall changes in corn are also likely driven by changes in total livestock feed consumption, which dropped following the U.S. recession in 2008-2009, as well as the use of other grains, including wheat and oats, in feed. Total corn use in feed still responds significantly to the amount of corn oil in feed when controlling for these other variables, so it appears that the relationship between corn and corn oil is real. This finding supports EPA’s assumption that additional corn grain replaces distillers sorghum oil in livestock feed.
While the amount of corn oil that has gone missing from any particular batch of livestock feed may be small, on aggregate the effects of waste-based biodiesel production on U.S. agriculture are substantial. In 2017, 13% of total U.S. biodiesel was produced from corn oil, and 35% was produced from all waste oils and fats. Using EPA’s assumed substitution ratio of corn to sorghum oil, the diversion of corn oil alone from livestock feed to biodiesel would have resulted in around 500 thousand tonnes CO2e emissions from producing more corn in 2017. That level of emissions is worth accounting for.