Commentary: Cellulosic ethanol worse than gas? Not so fast

This bailer, on display at an event near Emmetsburg, Iowa in 2010, is designed to harvest corn stover for biofuel use. (Photo by Joanna Schroeder via Creative Commons)

This bailer, on display at an event near Emmetsburg, Iowa in 2010, is designed to harvest corn stover for biofuel use. (Photo by Joanna Schroeder via Creative Commons)

Cross-posted from the Great Plains Institute with permission

By Brendan Jordan

A new academic paper on biofuels was released this week, and some media outlets have reached sweeping conclusions. “Fuels from corn waste not better than gas,” claimed the Associated Press. The Daily Caller one-upped them with “Study: Corn Ethanol is Nature’s Enemy.”

A careful reading of the paper doesn’t in any way support these conclusions.

In fact the academic study in question should be seen as reinforcing what many advocates for cellulosic biofuels have known all along: cellulosic biofuel production systems that systematically deplete soil organic carbon are unsustainable. The author does not argue that cellulosic biofuels are inherently unsustainable, or even that systems relying on crop residues are inherently unsustainable. This isn’t a general indictment of cellulosic biofuels, though it’s certainly being played that way in the media.

Simply put, there are good and bad ways to produce biofuels. This study points out the risks of unsustainable crop residue management. It also makes a few puzzling analytical decisions on the way, but we’ll get to that.

The study, “Biofuels from crop residue can reduce soil carbon and increase CO2 emissions” by lead author Dr. Adam Liska from University of Nebraska, appears in the April 2014 issue of Nature Climate Change. I know and like Dr. Liska, and respect his work. He has participated in several GPI projects and events.

The important thing to note about this paper is that it assumes bad agronomic practices – it assumes that crop residues are being removed in a way that depletes soil organic carbon (SOC). The paper is based around a simulation exercise, wherein the authors run a large model (“geospatial model and supercomputer simulations”) to estimate soil organic carbon loss across a wide geographic area due to residue removal. This is based on the assumption that if you simply leave the residue on the ground, it degrades slowly (around 20 percent is left after 10 years, mostly the lignin part). But if you remove it, the carbon is released quickly during the process of making biofuels, or by burning the biofuel in an engine.

Liska, et al., conclude that impact of SOC loss adds 50-70 grams of CO2 per megajoule (gCO2/MJ) to the lifecycle greenhouse gas (GHG) intensity of stover-based ethanol (around 30 gCO2e/MJ before SOC loss – or lower according to many studies). For context, the GHG intensity of gasoline is around 94 gCO2e/MJ, so this SOC loss is enough to push it well above the 60 percent GHG reduction threshold required by EPA for a fuel to be considered a cellulosic biofuel in the RFS.

There are many reasons not to take the high level conclusions at face value.  Others have pointed out that the authors use unrealistic assumptions for the level of stover removed by assuming that 75 percent is removed when typically only around 25 percent is actually removed. For example see USDA research leader Doug Karlen’s comments in this story from E&E’s ClimateWire. The paper fails to note that extensive research has been dedicated – with considerable success – to harvesting stover while avoiding soil carbon loss.

But now for the puzzling part – the authors invalidate the “big headline” conclusions in the discussion section of the study, but fail to follow through by changing their results.

First, the authors acknowledge that they do not calculate a co-product credit in their lifecycle emissions assessment (LCA) and thus include no benefit from burning lignin to provide process heat in the cellulosic conversion plant. However, every cellulosic plant I’m aware of does this. For LCA wonks – this is a bit like failing to account for DDGs in a corn ethanol LCA. The authors state that:

The lignin fraction of residue can also potentially be burned to produce electricity, offsetting coal generated electricity and saving emissions of up to 55g CO2e/MJ

Another problem – the authors fail to account for reductions in N2O emissions due to residue removal, even though they acknowledge that N2O emissions will be reduced. In the words of the authors:

As residue is a source of N2O emissions, residue removal would lower these emissions by 4.6 g CO2e/MJ, or 8% of SOC emissions.

N2O has 310 times the global warming potential of CO2 – agricultural soils are THE biggest source of N2O emissions. Several agronomists have told me that modest residue removal actually improves yields in the next year, and results in less need for added nitrogen, thus reducing N20 emissions. So if anything the beneficial impact on nitrogen emissions from residue removal is underestimated.

In review: you might add as much as 50-70g CO2e/MJ from soil emissions (assuming bad agronomic practices), but you gain back 55 from burning lignin for process heat, and another (at least) 4.6 from reduced N2O emission.  Doing the math, this research should result in a net impact of between negative 9.6 and 10.4 g CO2e/MJ.

Here’s a longer excerpt from the study’s discussion (emphasis mine):

Alternatively, development of other bioenergy systems, such as perennial grasses or forestry resources, may provide feedstocks that could have less negative impacts on SOC, GHG emissions, soil erosion, food security and biodiversity than from removal of corn residue. Soil CO2 emissions from residue removal, however, can be mitigated by a number of factors and management options. As residue is a source of N2O emissions, residue removal would lower these emissions by 4.6 g CO2e MJ-1, or 8% of SOC emissions. The lignin fraction of residue can also potentially be burned to produce electricity, offsetting coal generated electricity and saving emissions of up to 55g CO2 e MJ-1. Furthermore, use of improved soil and crop management practices, such as no-till cover crops, forage-based cropping systems, animal manure, compost, biochar and biofuel co-products, could replace the estimated SOC loss after residue removal. These management options require more research under different residue removal practices to ensure SOC stocks are maintained where crop residue is removed.

Final takeaway: The academic literature on cellulosic biofuel systems still overwhelmingly supports the conclusion that these systems offer deep GHG emissions reductions relative to gasoline. This paper is unlikely to alter the consensus view.

Brendan Jordan is Vice President of the Great Plains Institute, a Minneapolis-based policy organization. The Great Plains Institute is a member of RE-AMP, which also publishes Midwest Energy News.

3 thoughts on “Commentary: Cellulosic ethanol worse than gas? Not so fast

  1. Great analysis from the Great Plains, thanks Brendan. One question: are there any policy measures to encourage the positive agronomic practices that would lead to a better carbon footprint?

  2. Taken from actual story, “In a story April 20 about new research showing biofuels made with corn leftovers are worse for global warming than gasoline in the short term,”
    This is a very short sighted academic article with an axe to grind. Any bio-fuel research worth his salt understands that corn stover is hardly the only raw material to produce lignocellulosic ethanol. Asa research paper, the original academic paper should be considered a complete waste and the researcher who authored the paper is spinning his wheels..

  3. This analysis of soil carbon impacts clearly has a number of problems, included those stated above. I’m amazed at the conclusions based on the data and assumptions of the articles. The authors state that most organic carbon is found in the top 30 cm of soil, but cite no studies to support that, and I believe most studies of deep soil don’t support that conclusion. Our work here on Pacific Northwest forests shows this is not true of our soils, at least, as much as 80% of the soil carbon is found below 30 cm (to our 4 m maximum sampling depth), and most of the long-term impacts (in terms of Mg C/ha) of forest treatments on long-term soil carbon actually are found below 30 cm. Midwest Mollisols are famous for deep A horizons. I’ve seen many A horizons deeper than 30 cm. Increases in soil carbon have also been seen in switchgrass rotations, but only by 2.4 Mg C/ha in the 0-30 cm soil depths. Adding a consideration of soil depth to 120 cm increased that C sequestration by 16 Mg C/ha (Liebig et al. 2004). If these authors had only considered the soil as superficially as the study commented on above, the conclusions would have been very different.

    LIEBIG, M.A., H.A. JOHNSON, J.D. HANSON, AND A.B. FRANK. 2005. Soil carbon under switchgrass stands and cultivated cropland. Biomass Bioenergy 28:347–354.