Debates about the merits of biofuels have been going on for at least a generation. My favorite clip from the early, oil-crisis era ethanol push was Nicholas Wade’s article, “Oil pinch stirs dreams of moonshine travel,” published by Science in June 1979. Save for one topic, the terms of the debate — the costs of producing biofuels, whether ethanol took more energy to make than it delivered, the extent to which it really helps energy security, the hope for cellulosic biofuels and the food-versus-fuel dilemma — were the same nearly forty years ago as they are today.
Global warming is the topic not on the table then that is so important now. The effect of biofuels on greenhouse gas (GHG) emissions is the focus of many recent studies. To compare fuels according to their GHG impact, policymakers have adopted a form of computer modeling known as lifecycle analysis (LCA). A new report from the U.S. Department of Agriculture (USDA) is the latest LCA study to claim significant GHG reductions from the use of corn-based ethanol, concluding that it has net GHG emissions 43 percent lower than those of petroleum gasoline. Those results are similar to the findings of lifecycle modeling from Argonne National Laboratory (ANL), on which this latest USDA study heavily relies.
My own work has long come to an opposite conclusion. It shows that the use of biofuels (both ethanol and biodiesel) makes GHG emissions worse that they would otherwise be. This finding is not based on computer modeling, but relies instead on field data to assess the real-world CO2 flows involved when substituting biofuel for fossil fuel.
A key issue is that LCA is a scientifically illegitimate way to evaluate the situation. Indeed, the original developer of the lifecycle methods embraced by policymakers now concludes that “Using attributional lifecycle assessment to estimate climate-change mitigation benefits misleads policy makers,” as stated by the title of a paper he co-authored. There are several serious problems with LCA but the most telling is its use of an accounting shortcut that zeros out the CO2 emitted when biofuels are burned.
Carbon is the backbone of any practical liquid fuel, whether a simple, two-carbon molecule such as ethanol or a more complex organic mixture such as gasoline, diesel and biodiesel. Burning carbon-based fuels releases CO2 into the atmosphere regardless of their origin. For the most common fuels, these direct exhaust emissions average 73 (±2) grams of CO2 per megajoule (gCO2/MJ) of lower heating value (see adjoining chart).
Like many studies sponsored by the U.S. Department of Energy (DOE), which underwrites ANL’s lifecycle model, the USDA report sidesteps this basic fact of chemistry. They claim that the carbon in a biofuel — called biogenic carbon, coming from contemporary biomass — need not be counted because it was recently absorbed from the air by crops, trees or other vegetation that can soon grow back. This logic is used to justify the omission of tailpipe CO2 from biofuel LCA tallies. As seen in the USDA chart below, the lifecycle results for gasoline are dominated by the large red bar for tailpipe emissions. Only a tiny sliver of red (for non-CO2 GHG emissions) is shown for the ethanol results, even though the actual amount of CO2 emitted when burning ethanol is nearly as large as the amount emitted when burning gasoline.
Although the reasoning behind this accounting shortcut might seem simple, it is unfortunately simplistic and incomplete. It does not reflect the fact that productive farmland removes carbon from the atmosphere regardless of how the harvest is used. The photosynthesis of growing crops does not absorb more CO2 from the air just because the corn is used for fuel instead of food. The LCA models used by USDA erroneously credit cropland CO2 uptake as a full offset of biogenic emissions, including the CO2 from fermentation as well as that from tailpipes, even though only a portion of that uptake can be properly credited as an offset. The correct principles for such analysis were spelled out by my Biofuel’s Carbon Balance paper in 2013, but were not addressed by the USDA study.
The extent to which carbon uptake offsets biogenic CO2 emissions is an empirical question, one that can be evaluated using field data. Evaluating a state-of-the-art corn ethanol facility and the farmland serving it, we found no significant difference in net GHG emissions compared to gasoline. This contradicts LCA modeling that claimed a 40 percent GHG reduction for the same ethanol. Even this result reflects a one-time gain in carbon uptake from crop rotation; as shown in the next figure, GHG emissions from corn ethanol would be worse (up to nearly 70 percent higher) than those from gasoline under more realistic assumptions about prior crop production. Moreover, that’s before considering the commodity market and land-use changes that inevitably push the net CO2 emissions due to biofuel production even higher.
It is fairly straightforward to analyze the data available since the Renewable Fuel Standard (RFS) was passed in 2005. A retrospective analysis based on nationwide crop production statistics reveals that the gains in CO2 uptake during crop growth were only enough to offset 37 percent of the biogenic CO2 emissions associated with biofuel use. Once one factors in other impacts, including those from the land-use changes that have been directly observed with satellite data and the indirect land-use changes projected globally, the implication is that corn ethanol makes matters worse as far as the climate is concerned.
Although USDA’s carbon accounting is wrong, the parts of the analysis that address GHG emissions on farms and at biorefineries are probably fine. Ethanol production efficiency has improved over the years and so new facilities built after the RFS was put in place use less energy than older ones. Thus, recent production is more efficient than it was a decade ago, and the USDA analysis reasonably projects that processing gains and other improvements will continue through 2022.
On the other hand, it is difficult to trust the sections of the USDA report that address the effects on agricultural (including livestock) sectors and land use in the United States and globally. Little of the computer modeling cited is scientifically verifiable and aspects of it are mathematically incoherent. Largely performed by ANL and other entities funded by bioenergy advocates at DOE and elsewhere, these elaborate calculations were done in response to critical studies by independent academics. But even granting USDA its disputable results about agricultural system optimizations and land use, the study’s carbon accounting errors still undermine the claims that corn ethanol reduces net GHG emissions.
In short, the latest USDA analysis mixes facts about gains in corn ethanol processing efficiency with the fictions that LCA is a legitimate way to analyze GHG impacts and that the CO2 released when biofuels are burned does not count. Those errors are egregious and fundamental. Once one faces the facts about carbon uptake, it is clear that ethanol use worsens rather than lessens GHG emissions.