Lark et al. (2022) recently published “Environmental Outcomes of the US Renewable Fuel Standard” and addressed domestic land use change (LUC) of corn ethanol and associated greenhouse gas (GHG) emissions that are potentially caused by the U.S. Renewable Fuel Standard (RFS), as introduced in the 2005 Energy Policy Act and in the 2007 Energy Independence and Security Act (EISA). To do so, they considered the corn ethanol volume changes and LUC between 2008 and 2016.
In their assessment, Lark et al. assumed a business-as-usual (BAU) scenario (representing the goals of RFS1 for ethanol volume, as adopted in the 2005 Energy Policy Act by Congress, between 2008 and 2016), a new scenario (representing the goals of RFS2 for ethanol volume, as adopted in the 2007 EISA by Congress, between 2008 and 2016) to determine domestic LUC due to the RFS2. With no integrative modeling exercise, the authors simply calculated the average of the annual differences between the goals of RFS1 and RFS2 (5.5 billion gallons [Bgal]) and considered that volume of ethanol as the average annual contribution of RFS2 to new ethanol consumption between 2008 and 2016. Instead of using an integrated, coherent framework, as is the case with equilibrium models, in which changes in crop prices and associated LUCs at the intensive and extensive margin are determined simultaneously, Lark et al. applied a few loosely connected empirical methods to examine the impact of the RFS2 on three crops (corn, soybeans, and wheat). They estimated the short-term increases in commodity prices between 2008 and 2016 induced by the RFS2 and estimated that the prices of corn, soybeans, and wheat would increase by 30%, 20%, and 20%, respectively, due to an increase in the annual consumption of ethanol by 5.5 Bgal.
In the next step, Lark et al. used the Cropland Data Layer (CDL) from the United States Department of Agriculture (USDA) in combination with some other information on returns on cropland to estimate the probabilities of land transitions between cropland, pasture land, and Conservation Reserve Program (CRP) land. Using their projected increases in the prices of corn, soybeans, and wheat in combination with the estimated land transition functions, Lark et al. calculated that the area of corn plantation, adjusted for distiller’s dried grains (DDG), would increase by 2.8 million hectares (Mha) due to the RFS2, and that would lead to an increase in cropland area by 2.1 Mha. They showed that an overwhelming share of land conversion due to the RFS2 would be conversion of CRP land to active cropland. Lark et al. assigned a set of significantly large land use emissions factors to the CRP land conversion and likely double counted the N2O emissions in adding their LUC emissions to the rest of life-cycle analysis (LCA) emissions of corn-based ethanol, leading to the conclusion that the GHG emissions (commonly called carbon intensity) of ethanol are at least 24% higher than those of gasoline.
After a detailed technical review of the modeling practices and data used by Lark et al., we conclude that the results and conclusions provided by the authors are based on several questionable assumptions and a simple modeling approach that has resulted in overestimation of the GHG emissions of corn ethanol. In what follows, we present the general findings of our review.