Friday, November 11, 2016

A simple comparison of ABC to LCA results for corn ethanol

Our recent paper [1] provides an example that compares annual basis carbon (ABC) accounting results to typical lifecycle analysis (LCA) results for corn ethanol. As noted in the paper's discussion section, we used the same process GHG emissions as used in a standard LCA but then adjusted for the fact that biofuel combustion is not fully carbon neutral, as LCA assumes. The implication is then that corn ethanol is 27% more carbon intensive than gasoline instead of 44% less carbon intensive as was claimed by the LCA previously published by Wang et al [2]. This post describes how we made that comparison.

But first a caveat about any such comparisons. As our paper explains, ABC accounting is not the same as LCA and does not generate a "carbon intensity" value as does LCA. In reality, the net CO2 emissions impact of a biofuel varies from year to year, and so LCA-generated carbon intensity values are scientifically meaningless, politically popular though they may be. ABC accounting generates results based on an annual tracking of CO2 flows. Therefore, unlike LCA, ABC results correctly reflect the real-world dynamics of carbon uptake and emissions in the biosphere, including the rate of CO2 absorption on the farmland from which biofuel feedstocks are sourced.

With this caveat in mind, the adjoining table uses our ABC findings to correct the Wang et al LCA results that were generated with the DOE-backed GREET model. That model makes the erroneous assumption that biofuel-related biogenic emissions are completely (i.e., 100%) offset by CO2 uptake during feedstock growth.

The first line of the table gives the "well-to-wheels" carbon intensity of gasoline, which is the baseline to which the ethanol LCA results are compared. The second line gives Wang et al's core result for corn ethanol, as shown in Figure 5 of their paper. This value, 76 gCO2e/MJ, is 19% lower than the 94 gCO2e/MJ carbon intensity value for gasoline.

That core result includes Wang et al's modeled land-use change (LUC) impact, amounting to 9 gCO2e/MJ, but does not include the credit of 14 gCO2e/MJ for distiller's grains with solubles (DGS) supplied as a co-product of corn ethanol. Subtracting those two values leaves what we label the net ALCA (attributional lifecycle analysis) result of 53 gCO2e/MJ, which is the basis for Wang et al's main finding that corn ethanol is 44% less carbon intensive than petroleum gasoline, as listed in Table 7 of their paper.

To make the comparison to ABC accounting while using the same process-related GHG emissions as estimated by Wang et al's GREET modeling, we need to adjust their 53 gCO2e/MJ result for our finding was that there was only enough carbon uptake on farmland to offset 37% of the biogenic CO2 emissions over the 2005-2013 period we evaluated. We therefore add in a portion of the biogenic emissions that are otherwise fully credited in the GREET model. Using GREET's internal data table, the CO2 directly emitted from the combustion of ethanol is 71 gCO2e/MJ. Total biogenic emissions are 1.5 times as much when counting the CO2 from fermentation, bringing total biogenic emissions for ethanol to 107 gCO2e/MJ. If 37% of that, i.e., 39 gCO2e/MJ, is offset by carbon uptake during crop growth, that leaves 67 gCO2e/MJ of net biogenic emissions entering the atmosphere. Adding that value to the 53 gCO2e/MJ ALCA estimate for process emissions implies 120 gCO2e/MJ, which is 28% higher than the carbon intensity of gasoline [3].


[1] DeCicco, J.M., et al. 2016. Carbon balance effects of U.S. biofuel production and use. Climatic Change, published online 25 August 2016.  http://dx.doi.org/10.1007/s10584-016-1764-4   

[2] Wang, M.Q., et al. 2012. Well-to-wheels energy use and greenhouse gas emissions of ethanol from corn, sugarcane and cellulosic biomass for US use. Environ Res Lett 7: 045905. http://iopscience.iop.org/article/10.1088/1748-9326/7/4/045905  

[3] The value of 28% is slightly more than the 27% reported in our paper, likely due to a rounding issue between the pencil-and-paper calculation used for the paper and the spreadsheet calculation just developed for this blog post. 

No comments:

Post a Comment