Ten years ago, Karen Daynard and I produced a detailed report for the Grain Farmers of Ontario on various issues relating to use of corn-based ethanol as a gasoline additive. Since then, while government policies in Canada, the United States and many other countries have continued to encourage this practice, there is also negativity in the media and other public statements about fuel ethanol. This column contains an overview of addition information published since our report was written, with a focus on greenhouse gas (GHG) emissions. It is targeted to those, especially in agriculture, who seek more information than contained in media reports but not to the detailed extent found in scientific papers.

The bottom line: The use of grain-based ethanol in Canada means about a 45% reduction (and perhaps at least 49%) in net greenhouse gas emissions compared to gasoline – and that’s including indirect effects on land usage in both Canada and internationally.

More detailed analysis follows, starting with a brief historical perspective.

Since the modern era of biofuels began about 40 years ago, the reasons for ethanol usage have changed. Initially ethanol was welcomed as an octane enhancer when lead was removed from gasoline.  In addition, because of its high oxygen content, ethanol addition to gasoline meant more complete fuel combustion and reduced ground-level concentrations of carbon monoxide caused by cold-engine operation in winter, especially in vulnerable cities like Denver and Calgary. Changes in automobile engines have largely eliminated the latter as an environmental concern, but the need for octane enhancement remains.

Since the Iranian conflict in 1979, there has been continuing major interest in the United States (not so much in Canada) in production of home-grown biofuels to reduce dependence on petroleum imports. This led to debate on the extent to which ethanol inclusion actually increased energy supply. Subsequent analyses – references here – showed that corn-based fuel-grade ethanol contains about 160% as much combustible energy as was used for its production a decade ago, with that percentage likely now in the vicinity of 200% in Canada.

Biofuels experienced a set back in about 2007-2008 when their production from grain and oilseed crops led to loud criticisms that this usage increased food shortages and raised food prices. Follow-up analyses showed that, while this effect was real, it was much smaller than proclaimed in initial analyses and media statements (more detail here). In the United States, analyses showed that any increase in food prices was more than counter-balanced by ethanol-induced reductions in transportation fuel prices.

However, this did dampen enthusiasm for biofuels to some extent, leading to increased interest in use of non-food crops for biofuel production. (This also meant introduction of some rather strange public policies such as encouraging the conversion of food crop land to the production of non-food crops for biofuel production.)

Ethanol policy now focused on reductions in greenhouse gas emissions – and indirect land use changes

Beginning about 2000, public policy on biofuels became driven primarily by climate change and the opportunity for fuel ethanol and biodiesel made from agricultural crops to reduce net greenhouse gas (GHG) emissions. This has been the main driver for post-2000 federal and provincial/state programs introduced to encourage biofuel usage in Canada, the United States and Europe – including the Clean Fuel Standard announced for Canada in December 2020 (more detail here).

A major influence on ethanol public policy was a high-profile paper published by Searchinger et al in Science in 2008. In it, the authors claimed that net GHG release with fuel ethanol usage was greater than that of the fossil-based fuels which ethanol replaced. This, they calculated, was because of the obligate clearing and conversion of forests and grasslands into food/feed grain production, to replace the grain now used for fuel production. Searchinger et al, using a combination of their own calculations and the FAPRU computer model developed by Iowa State University and University of Missouri, calculated that this ‘indirect land use change’ (ILUC) equated to an average of emission of 104 g CO2e (carbon dioxide equivalent) per MJ (million Joules) of combustible energy with fuel ethanol. This compares to a total of 92 g/MJ CO2e with gasoline production and combustion – and only 74 g/MJ CO2e for all other emissions associated with grain-based ethanol.

The report by Searchinger et al caused quite a stir and prompted the US Environmental Protection Agency (EPA) to release a major calculation of its own in 2010, pertinent to the US Renewable Fuel Standard (RFS) of 2007. EPA calculated that, while the ILUC factor was significant, it was about 70% less than the value published by Searchinger et al. EPA also calculated the GHG emission associated with all non-ILUC factors was about 45 g/MJ CO2e for ethanol from corn, versus the 74 g/MJ published by Searchinger et al. EPA projected, further, that with improvements in efficiencies expected by the year 2022, combustion of an average litre/gallon of grain-based ethanol in the United States would mean a 21% reduction in net GHG emissions compared to the same quantity of gasoline, adjusted for differences in caloric energy content. (Ethanol has a combustible energy content of 23.5 MJ/l – million Joules per litre – versus 33.5 MJ/l for gasoline.) The 21% exceeds the 20% minimum needed to qualify for inclusion under the RFS.

The United States Department of Agriculture (USDA) initiated its own review of the GHG balance with ethanol-blended fuels, first with the release in 2017 of a study done for USDA by Washington DC-based ICF, and then by a peer-reviewed paper published in 2019 in Biofuels. J. Lewandowsky of USDA was senior author of the latter. The two reports are similar though not identical; the later uses updated input data where available. As with the EPA report, the Lewandowsky et al/USDA-LCF analyses depend heavily on computer models available from several sources. The numbers shown below come from the 2019 paper.

Their analyses include calculations of net GHG emissions associated with corn-farming inputs including fuel and fertilizer, offsets from production/usage of byproducts of ethanol manufacture, associated effects on animal agriculture, ILUC both domestic and international, rice methane emissions (greater grain usage for ethanol means changes in rice production with its sizable methane emissions), fuel and feedstock transportation, fuel ethanol production, and emissions during fuel combustion (some methane and nitrous oxide emissions result from ethanol combustion in engines).

Table 1. Net greenhouse gas emissions, expressed as grams of CO2 equivalent per Mega-Joule of energy (g/MJ CO2e) in gasoline or ethanol, according to different sources/calculations.

Note: The categorization of emissions with ethanol production/combustion is taken from Searchinger et al (2008) with data presented in Lewandowsky et al, (2019) matched up as possible. The category, ‘Ethanol feedstock production and associated emissions,’ includes effects on rice methane production and livestock emissions. ‘Indirect land use change’ includes both domestic and international.

Detailed explanations of the respective differences among the calculations of Searchinger, EPA and Lewandowsky et al are beyond the scope of this column. Readers are referred to Lewandowsky et al for more explanation. Suffice it to say that the reasons include the use of data from different years, different assumptions, and updated/revised and different computer models for estimating the various quantities.

Biofuel policy must also recognize global differences as well as common goals

Interestingly, Dr. Searchinger addresses the same issue again in a chapter in Creating a sustainable food future, published by World Resources Institute (WRI) in 2019 (full report here; my critique on the full report here). Searchinger does not mention the analyses done by EPA (2010) and the USDA-ICF report in 2017. (The Lewandowsky et al paper of 2019 may not have been published in time for consideration for the WRI report.) His basic argument is that any significant increase in crop usage for biofuels must mean a corresponding reduction in land for food production – and hence, the need to convert forests and grasslands into new agricultural land for food production. He largely ignores a couple of key issues: 1) the corn being used to produce ethanol for fuel in North America comes in large part from yield increases – and to some degree from shifts within total grain crop hectares – rather than additions in new arable crop hectares and 2) policies that encourage developing countries to grow their own food rather depending on imports from grain producers in the developed world are to be encouraged.

The vulnerability of developed countries to large-scale food importation was made painfully apparent during the food price crisis of 2007-08. (See here for more detailed discussion and references.) The fact that sizable quantities of grain are being used to produce ethanol in the United States and Canada instead of depressing prices for farmers in developing countries should be viewed as a positive. This is not to argue that increased biofuel production is beneficial for countries where domestic grain supply is limited. The world is not a homogeneous supply-demand market place for agricultural commodities – even though that appears to be a basic assumption (and flaw, in my view) in the WRI analyses.

I will enlarge on the Canadian corn supply-demand situation as affected by ethanol later in this column.

If one uses the Lewandowsky et al (2019) calculations as being the most representative of current conditions, the usage of ethanol as a substitute for gasoline represented a 38% reduction in net GHG emissions compared to gasoline on a caloric basis, in years around 2015, with this projected to rise to 43% by 2022. Note that Lewandowsky et al estimate that the reduction could become as high as 71% (calculations not shown) by using best practices for reducing GHG emissions in both producing feedstocks and in ethanol refining.

To provide balance, readers are referred to this article published by a group named the International Council on Clean Transportation (ICCT) in Washington DC, which is critical of most of the calculations presented in USDA-LCF. Unlike the articles by Searchinger et al (2008) and Lewandowsky et al (2019), the ICCT report is not in a peer-reviewed publication.

What about Canada?

A final issue involves the extent to which GHG reductions caused by the substitution of ethanol for gasoline differ for Canada compared to the United States.

Jayasundara et al (2014) published a thorough analysis of energy and greenhouse gas emissions associated with corn production in Ontario using data from about 2006 to 2011. About 70% of Canadian corn is grown in Ontario. Unfortunately, it’s not that easy to compare most of their numbers with the values shown in Table 1. Firstly, data in Jayasundara et al are expressed in kg CO2e per tonne of grain, which necessitates a conversion to MJ of energy in ethanol – not that difficult with certain assumptions (eg. litre of ethanol per tonne) but a source of uncertainty. The Jayasundara et al calculations also include items such as energy used to make farm equipment, for seed production and for other operations not included in the analyses shown in Table 1.

In total, Jayasundara et al calculate that an average of 281 kg of CO2e was released per tonne of corn grown in Ontario from 2006 to 2011. If it is assumed that about 420 l of ethanol are produced per tonne of corn (see Table 8 in Lewandowsky et al for a reference), this equates to 670 g CO2e per lite or about 28 g CO2e per MJ (23.5 MJ/litre). This value is well above those shown in the second data line in Table 1.

However, analyses using data from Jayasundara et al, unlike the others, include no credit for byproducts produced in ethanol production, including the high-protein DDGS (‘distillers’ dry grains with solubles’) that represents about 30% of the dry weight of the initial grain feedstock. Calculation of the associated effect on net GHG emissions is not simple as it needs to include offsetting effects on the production and usage of other high-protein livestock feed sources and on livestock production per se.

If one assumes a value of 12 g CO2e/MJ of ethanol calculated by Lewandowsky et al for byproduct credits (a weighted average for US dry-mill and wet-mill ethanol plants), this means 28-12 = 16 g/MJ of CO2e for ethanol with byproduct credits. This is still above the 13 g/MJ calculated for USDA-LCF (2019) in Table 1, recognizing that the Jayasundara et al calculation includes more GHG-linked costs for ethanol production.

Another option is to use the comparisons with US data which are presented by Jayasundara et al. They refer to a study by Farrel et al (2006) showing average US GHG emissions per tonne of corn to range from 282 to 477 kg/t of CO2e. Another by Liska et al (2009) shows a range of 261 to 385 kg/t for 12 Midwestern Corn-belt states. Jayasundara et al state that the Liska et al values must be increased by 11% to account for operations included by Jayasundara et al but not by Liska et al. In total, these comparisons mean that the 281 kg/t of CO2e calculated for corn production in Ontario lies at or below the lower end of the range reported in US studies.

A decade or more ago, ethanol production in the US was considered to be less efficient than in Canada because of the extensive use of coal in US ethanol plants but not Canada. However, that difference has largely disappeared. All US dry-mill ethanol plants are now fired by natural gas as in Canada, and most US wet-mill plants are as well.

Somewhat more energy is used for grain corn drying at harvest in Canada than the US but the US uses much more fossil energy for irrigating corn (almost none for corn in Ontario).

An overall conclusion from the above is that fuel ethanol production/usage as a gasoline substitute in Ontario/Canada is likely to mean even somewhat larger reductions in MJ/l of fuel than the values shown for the United States in Table 1.

In the Clean Fuel Regulations published by the Government of Canada on December 18, 2020, a default value of 49 g CO2e per MJ is assumed for both Canadian and US-sourced ethanol. This is quite similar to the 52 g/MJ calculated by Lewandowsky et al for 2022 (Table 1). (Though beyond the scope of this article, there is a small but significant difference in how the US and Canada calculate the energy content of fuels; see this reference for an explanation of the difference between “lower” and “higher” heating values, with the US generally using “lower” and Canada “higher.”)  The value, 49g/MJ of CO2e, represents a 47-49% reduction compared to gasoline. A more thorough life-cycle analysis is presently being completed by the government, presumably for both Canadian and US-produced ethanol; completion date is unknown.

The conclusion from all of this: The use of grain-based ethanol in Canada means about a 45% reduction (and perhaps at least 49%) in net greenhouse gas emissions compared to gasoline – and that’s including indirect effects on land usage in both Canada and internationally.

—

Appendix: What has ethanol production meant for corn production in Canada?

Major production of fuel ethanol began in Canada in 1996 with the opening of a Commercial Alcohol, now Greenfield Ethanol, plant in Chatham Ontario. (A small fuel ethanol plant was started by Mohawk Oil at Minnedosa Manitoba in 1981 and another by Poundmaker Agventures at Lanigan Saskatchewan in 1990.) Further major expansion occurred with the opening of additional Greenfield Ethanol plants and other plants across Canada, such that production about tripled between 2005 and 2011. Exact data are difficult to secure since Statistics Canada only reports corn usage for all food and industrial processing for Canada (and not at all for Ontario where most corn is grown and processed), and this category also includes corn dry milling to produce food products plus ethanol for beverage alcohols and other consumer and industrial products.

Nevertheless, in the graph below taken from various reports from Statistics Canada, one can see (brown line) the rapid increase in corn usage for food and industrial use from 2006 to 2011 – essentially all of the increase being caused by increased fuel ethanol production.

By contrast, feed usage of grain corn (yellow line) peaked about 2002 to 2008, declined thereafter (likely in part because of the substitution by DDGS), and with recent recovery to near the levels of a decade and a half ago. Total domestic usage of corn (red line) has increased steadily from 1997 to the present, with this mostly exceeding domestic production (light blue line) for years before 2011 but about equivalent to production in years since.

One can see that the increased production has come from a combination of increased yields (dark blue) and planted hectares (green).  Interestingly, the 400,000 of increased hectares planted to grain corn in Canada from 1997 to 2020 matches up about exactly with a 400,000 ha reduction in perennial forage/hay production in Ontario over the same interval (data not shown) according to Statistics Canada.

Changes in number of hectares used corn/ethanol production for all of Canada are overshadowed by larger shifts in production of larger-hectarage crops in Western Canada over this period and a steady transition of Prairie farmland from summer fallowing to crop production. In fact, this shift from summer fallowing to cropping has resulted in reduced net GHG emissions from Canadian farm soils since 2005 according to data reported by Canada to the United Nations Framework Convention on Climate Change (more detail here).

In total, while some increase in land seeded to corn has occurred in Canada since 2006, one would be hard pressed to show that increased fuel ethanol production from corn has resulted in transition of land in forest and grassland into corn production.

Acknowledgment

Major thanks is extended to Mr. Don O’Connor, president of (S&T)2 Consultants of Delta, British Columbia,for his help in accessing some of the background material used for this article and for his critique of a draft copy. That notwithstanding, any errors or misinterpretations in the resulting product are solely the responsibility of the author.