Ethanol & Greenhouse Gas Emissions – Reconsidering the University of Nebraska Study
The debate about the environmental impact of ethanol rages on. Last month, the most recent study on the greenhouse gas (GHG) emissions associated with ethanol use was published by researchers from the University of Nebraska (Liska et al.). That analysis used the most recent data available on individual facility operations and emissions, observed corn yields, nitrogen fertilizer emissions profiles, and co-product use; all of which prove important because of improved energy efficiencies associated with ethanol production over the past several years. The authors found that the total life-cycle GHG emissions from the most common type of ethanol processing facility in operation today are 48-59 percent lower than gasoline, one of the highest savings reported in the literature. Even without subtracting-out the GHG emissions associated with ethanol co-products (which accounted for 19-38 percent of total system emissions), ethanol would still present GHG advantages relative to gasoline. The ethanol lobby went wild.
This may be the best study on the subject, but it is not the final word. There are three fundamental problems with the analysis.
First, the study examines only a subset of corn production operations and ethanol processing facilities; dry mill ethanol processors fired by natural gas in six corn-belt states. Together, those facilities accounted for 23 percent of US ethanol production in 2006. While this approach makes the study stronger because the authors are not forced to rely as heavily on estimates and aggregated analysis, the down-side is that the study ignores a large number of older, less efficient ethanol processing facilities and thus cannot be used to assess the GHG balance of the ethanol industry as a whole. While the findings may well point to where the industry will be in the future as older, less efficient facilities lose market share and are upgraded or retired, the bankruptcies that are shutting down many newer facilities at present caution against certainty on this point.
Second, estimates regarding emissions are still relied on to some degree, and one of those estimates in particular – the estimate pertaining to the release of nitrous oxide (N2O) from fertilizer use in corn production – is problematic. While the study comports with convention in that it relies on emission estimates offered by the Intergovernmental Panel on Climate Change, a recent study finds that the IPCC estimates as they pertain to N2O release from fertilizer does not comport with the observed data (Crutzen et al., 2007). That study finds that N2O emissions from fertilizers used in biofuels production are 3-5 times greater than assumed by the IPCC and that, if we plug those higher emissions into the ethanol life-cycle models, “the outcome is that the production of commonly used biofuels, such as biodiesel from rapeseed and bioethanol from corn (maize), can contribute as much or more to global warming by N2O emissions than cooling by fossil fuel savings.” Given that the lead author of the study – Paul Crutzen – is a Nobel laureate chemist who has specialized in fields related to atmospheric science, his findings cannot be lightly dismissed.
Third, the study acknowledges the importance of the impact that ethanol production has on crop prices and, thus, on global land-use patterns, but it does not account for the GHG emissions associated with those changes. Those emissions are substantial and no life-cycle analysis of ethanol can credibly ignore them.
A worldwide agricultural model constructed by Searchinger et al. (2008) finds that the increases in crop prices that follow from the increased demand for ethanol will induce a global change in the pattern of land use. Those land use changes produce a surge in GHG emissions that is only dissipated by conventional life-cycle emissions savings many decades hence. Although Searchinger et al. modeled ethanol production increases that were beyond those mandated in existing law, “the emissions from land-use change per unit of ethanol would be similar regardless of the ethanol increase analyzed.”
While critics of Searchinger et al. are right to point out that the agricultural model employed in the study was crude, that much is unknown about the factors that influence global land use decisions, that improved yields are reducing the amount of land necessary to meet global crop demands, and that any land additions to crop production do not need to come from forests or other robust carbon sequestration sinks, none of those observations is sufficient to reject the basic insight forwarded in that study. If ethanol demand increases corn and other crop prices beyond where they otherwise would have been, profit incentives will induce investors to increase crop production beyond where production would otherwise have been. If that increased production comes in part from land use changes relative to the baseline, then significant volumes of GHG will likely be released and those emissions threaten to swamp the GHG savings found elsewhere in the life-cycle analysis. Even if the upward pressure on crop prices that are a consequence of ethanol consumption is more than offset by downward price pressures following from other factors, crop acreage retirement will not be as large as might otherwise have been the case and terrestrial sequestration will be lower as a consequence. Every link in that chain of logic is unassailable.
This is but one of the many impacts that ethanol might have on hundreds of industrial sectors worldwide. Searchinger et al. is ultimately unsatisfying because it is only a crude and partial consideration of those impacts, many of which might indirectly affect global land use patterns. For instance, if ethanol consumption reduces the demand for – and thus the price of – crude oil in global markets, how much of those “booked” reductions in oil consumption will be offset by increased demand induced elsewhere by the lower global crude oil prices that follow (known as a “rebound effect” in economics)? How might that increase in global demand for crude oil in response to lower price affect all sort of GHG emissions vectors? None of these sorts of questions are asked in ethanol GHG life-cycle analyses but they are clearly crucial to the analysis.
To summarize, the narrow, conventional consideration of the GHG emissions associated with ethanol from Liska et al. suggest that it reduces climate change harms relative to gasoline. If the IPCC has underestimated N2O emissions from fertilizer – as appears to be the case – then ethanol probably is at best a wash with regards to GHG emissions. Even if that’s not the case, consideration of secondary and tertiary emissions impacts strongly suggests that most if not all of all advertised GHG gains are lost in the changes in land use patterns that follow from increases in ethanol production relative to the baseline. Other changes in anthropogenic emissions – positive and negative – would almost certainly follow as well, but existing models do not bother to search for them and thus we do not know enough to say much beyond this with confidence.
Based on what we know now, it would be hard to make a compelling case that ethanol is preferable to gasoline with regards to total greenhouse gas emissions – and last month’s study out of the University of Nebraska does not change that.