Climatic Change

, Volume 144, Issue 2, pp 111–119 | Cite as

Commentary on “carbon balance effects of US biofuel production and use,” by DeCicco et al. (2016)

  • Robert D. De Kleine
  • Timothy J. Wallington
  • James E. Anderson
  • Hyung Chul Kim


In their recent publication “Carbon balance effects of U.S. biofuel production and use, DeCicco et al. present an empirical assessment of net CO2 emission effects over the period 2005–2013 after the US renewable fuel standard (RFS) came into existence and conclude that biofuels have resulted in a net increase in CO2 emissions over the period. The analysis presented by DeCicco et al. relies on three key assertions. First, that if biofuel carbon combustion emissions are not completely offset by additional net ecosystem production (NEP), then the biofuel should not receive full biogenic carbon credit. Second, that changes in agricultural NEP related to biofuel production can be accurately estimated from national-level agricultural production statistics. Third, that agricultural NEP is a pertinent measure of biofuel global warming impacts. We show that following the conventional definition of NEP the combustion of biofuel by definition leads to an exactly equal increase in NEP; therefore, the first assertion is not meaningful. Regarding the second assertion, we show that estimation of biofuel-related NEP changes from agricultural production statistics is not a robust methodology. Finally, we argue that agricultural NEP is an important parameter for estimating land-use change effects, but in isolation is an irrelevant GHG metric for current biofuels. We find that the conclusions above from DeCicco et al. are unfounded and do not invalidate the application of biogenic carbon offsets in life cycle assessments of biofuels currently used in national and international regulations.

1 Introduction

In their recent publication “Carbon balance effects of U.S. biofuel production and use, DeCicco et al. (2016) present an empirical assessment of net CO2 emission effects over the period 2005–2013 after the US renewable fuel standard (RFS) came into existence. DeCicco et al. come to three important conclusions. First, over the period 2005–2013 the “additional carbon uptake on cropland was enough to offset only 37% of the biofuel-related biogenic CO2 emissions.” Second, that this result “falsifies the assumption of the full offset made by LCA (life cycle assessments) and other greenhouse gas (GHG) accounting methods.” Third, “U.S. biofuel use to date has been associated with a net increase rather than a net decrease in CO2 emissions.” In this commentary, we argue that these conclusions are flawed and not supported by this analysis.

2 Calculating biofuel impact on NEP

It is useful to begin with definitions of biogenic CO2 emissions, carbon neutrality, and net ecosystem production. Biogenic carbon is the carbon contained in compounds produced by biological systems. Biogenic carbon is sometimes referred to as organic carbon. Biogenic CO2 emissions are CO2 emissions “related to the natural carbon cycle, including those resulting from harvest, combustion, digestion, fermentation, decomposition, and processing of biologically based materials” (U.S. EPA 2016). This definition excludes carbon emissions from non-biological sources, such as fossil fuel combustion, used as part of these processes. The term “carbon neutral” refers to a system or entity with net zero carbon emissions into the atmosphere; any carbon emissions are matched by an equivalent amount of carbon sequestration or carbon offsets. The argument that follows aims to address only the carbon neutrality of the biogenic carbon emissions and not the carbon neutrality of the full fuel life cycle.

Net ecosystem production (NEP) is defined as the difference between the amount of carbon fixed by photosynthesis, characterized as gross primary production (GPP) and the total ecosystem carbon respiration, Re (Woodwell and Whittaker 1968; IPCC 2000; Lovett et al. 2006).

$$ \mathrm{NEP}=\mathrm{GPP}-{R}_{\mathrm{e}} $$

The total ecosystem respiration is the sum of autotrophic (Ra) and heterotrophic respiration (Rh). In a traditional ecology context, only heterotrophs in the local food web are considered, but if the spatial boundaries grow to include both natural and human-managed ecosystems, respiration from humans and domesticated animals is also included. Biogenic carbon in agricultural products produced for food, or feed is returned to the atmosphere by heterotrophic respiration and is included in Rh.

$$ {R}_{\mathrm{e}}={R}_{\mathrm{a}}+{R}_{\mathrm{h}} $$
$$ \mathrm{NEP}=\mathrm{GPP}-{R}_{\mathrm{a}}-{R}_{\mathrm{h}} $$

Net primary production (NPP) is defined as the difference between GPP and autotrophic respiration, and hence, NEP is the difference between NPP and heterotrophic respiration

$$ \mathrm{NPP}=\mathrm{GPP}-{R}_{\mathrm{a}} $$
$$ \mathrm{NEP}=\mathrm{NPP}-{R}_{\mathrm{h}} $$

NEP represents the total amount of biogenic carbon in an ecosystem available for storage, export as organic carbon, or non-biological oxidation to CO2 by combustion or ultraviolet light and is expressed in terms of a mass of carbon per area per time (Lovett et al. 2006). As discussed by Lovett et al. (2006), NEP is perhaps best understood in the context of the organic carbon balance for an ecosystem. The change in organic carbon stored in an ecosystem, ΔCorg, is given by the following:

$$ {\Delta \mathrm{C}}_{\mathrm{org}}=\mathrm{NEP}+I-E-{\mathrm{Ox}}_{\mathrm{nb}} $$

where E is export of biogenic carbon, Oxnb is the non-biological oxidation of biogenic carbon by combustion or UV light, and I is import of biogenic carbon into the ecosystem. The change in carbon stored in an ecosystem is also often referred to as net ecosystem carbon balance (NECB) (Chapin et al. 2006), but this paper will continue to use the ΔCorg term to reduce jargon. Rearranging gives

$$ \mathrm{NEP}={\Delta \mathrm{C}}_{\mathrm{org}}+{\mathrm{Ox}}_{\mathrm{nb}}+E-I $$

Forests, grasslands, and croplands typically have positive NEP, while lakes, rivers, and cities typically have negative NEP; for worked examples, see Lovett et al. (2006). Agricultural harvests involving local or regional transfer of biogenic carbon are included in the term E for ecosystems from which they are exported, and in the term I for ecosystems into which they are imported. Biomass production for biofuels generally impacts the terms E, Oxnb, I, and possibly ΔCorg (e.g., if soil carbon is changed). Agricultural production of long-lasting fibers (e.g., construction wood) leads to storage of biogenic carbon and hence impacts ΔCorg, though possibly in a different region than its ecosystem of origin.

From the perspective of the global ecosystem, the terms E and I cancel and so

$$ \mathrm{NEP}={\Delta \mathrm{C}}_{\mathrm{org}}+{\mathrm{Ox}}_{\mathrm{nb}} $$

DeCicco and co-authors (DeCicco et al. 2016; DeCicco 2013) have challenged the treatment of biogenic CO2 emissions in life cycle assessments of biofuels. Typically, biogenic CO2 emissions from biofuel combustion are either not included in the assessments (US EPA), or they are accounted for, but offset by a matching credit (GREET 2016). The United Nations Framework Convention on Climate Change (UNFCCC) greenhouse gas (GHG) accounting method adopted for international GHG reporting (UNFCCC 2006) sets biogenic CO2 emissions from biofuel combustion to zero. The text in the UNFCCC report reads “Amounts of biomass used as fuel are included in the national energy consumption but the corresponding CO2 emissions are not included in the national total as it is assumed that the biomass is produced in a sustainable manner. If the biomass is harvested at an unsustainable rate, net CO2 emissions are accounted for as a loss of biomass stocks in the Land Use, Land-Use Change and Forestry sector.” Simply put, if the terrestrial carbon stock remains stable over time, then the biogenic carbon emissions are carbon neutral. This text is consistent with the expression for NEP in Eq. (8). This approach mirrors the physical reality that the biogenic carbon in the fuel was recently absorbed from the atmosphere and is completing a cycle back to the atmosphere. Biomass cultivation, harvesting, and processing to fuel typically involve fossil fuel inputs, and there may be additional carbon emissions associated with land-use changes; hence, although combustion of biofuels is carbon neutral, the full life cycle production of biofuels is typically not carbon neutral.

DeCicco and co-authors (DeCicco 2013; DeCicco et al. 2016) argue that the biogenic carbon captured from the atmosphere and converted into biofuel would have been otherwise captured by plants and that biofuels therefore should not necessarily receive full biogenic uptake “credit.” DeCicco and co-authors claim that NEP is the critical metric to consider and that if biofuel production does not increase NEP by at least as much as the embodied biogenic carbon, then the biofuel should not receive the full biogenic carbon credit. We argue that the logic applied and the argument developed by DeDicco and co-authors is false because it does not recognize the fact that biogenic carbon produced for food or animal feed in cropland ecosystems will be returned to the atmosphere via heterotrophic respiration in the non-cropland ecosystems to which the food/feed is exported. By definition (see Eq. 5), biogenic carbon which flows back to the atmosphere via heterotrophic respiration does not contribute to NEP. When using a holistic approach with both cropland and non-cropland ecosystems included in the analysis and considering the case with no change in carbon stocks (ΔCorg = 0) then Eq. (8) shows that the biogenic carbon that is combusted in biofuel and hence contributes to Oxnb leads to an exactly equal increase in NEP.

When discussing NEP, it is important to carefully specify the ecosystem being considered. DeCicco and co-authors use inconsistent spatial definitions in their analyses. In the definition of NEP in Eq. (7), harvesting and exporting biogenic carbon from croplands is accounted for in the term E. Harvesting and exporting biogenic carbon increases the NEP of the croplands and decreases the NEP of the non-cropland ecosystems into which the harvest is imported. To the extent that the harvested biogenic carbon is produced without changing the carbon stock in the cropland ecosystem (ΔCorg = 0) and that the harvested biogenic carbon is food or feed and returned to the atmosphere via heterotrophic respiration (Oxnb = 0), then increased NEP in the cropland ecosystem is exactly matched by decreased NEP in the ecosystems into which it is imported. If there is no change in the cropland carbon stocks (ΔCorg = 0) and we consider both cropland ecosystems from which the harvest is exported and the non-cropland ecosystems into which the harvest is imported, then harvest of biogenic carbon in the US (and elsewhere) for food and feed has no overall contribution to NEP. The impact of the harvest on NEP for the combined cropland and non-cropland ecosystem is only non-zero when the harvest is produced in a manner leading to decreased (ΔCorg < 0) or increased carbon stocks (ΔCorg > 0) in the cropland ecosystem, or when the biogenic carbon harvested is stored for long periods (ΔCorg > 0, e.g., wood for construction), or when the harvested biogenic carbon undergoes non-biological oxidation (Oxnb > 0).

It is straightforward to understand that biofuels produced from waste biomass such as forest residue or agricultural waste which would otherwise undergo decomposition (heterotrophic respiration) lead to a reduction in Rh and hence, by definition, lead to an equal increase in NEP because NEP = NPP − Rh (Eq. 5). It is also straightforward to understand that biofuels produced from additional biomass produced as a response to biofuel demand also leads to an equal increase in NEP. Less intuitively, but no less true, biofuels produced by either diverting biomass which would otherwise be used as food or feed, while perhaps not optimal from economic or moral perspectives of maximizing human consumption, also lead to an increase in NEP exactly equal to the biogenic carbon combusted.

We conclude that following the conventional definition of NEP (Lovett et al. 2006), the production of biofuel by definition leads to an exactly equal increase in net ecosystem production (NEP). The biogenic carbon offset applied in life cycle assessments of biofuels used in current national and international regulations (UNFCCC 2006) is consistent with the definition of NEP as discussed above. While biofuel production and use lead to an exactly equal increase in NEP and a 100% biogenic carbon offset is appropriate for biofuel combustion in life cycle assessments of biofuels, it is important to recognize that the production of biofuels does have important direct and indirect climate change impacts. Direct impacts include fossil-fuel-related emissions during biomass cultivation, harvesting, and processing. These direct effects are captured in conventional attributional life cycle assessments. Indirect effects on the agricultural sector are best understood through a consideration of economics in consequential life cycle assessments. The combination of attributional and consequential methodologies provides a comprehensive life cycle accounting of climate change impacts of biofuel production and use.

The importance of economics in biofuel production cannot be understated. Economic factors, such as the cost of inputs (including materials, fossil fuels, and land rents), cost of substitutes, and value of outputs, play a critical role in determining agricultural production. Production of biomass for first generation biofuels is embedded in a complex national and international market for agricultural goods whose purpose is to provide food, feed, fuel, and fiber. An increase in the demand for biomass X (e.g., corn) for biofuel production (e.g., corn ethanol) creates a price signal resulting in increased prices for X and likely decreased prices for coproducts (e.g., corn oil, distillers dried grains, and solubles). Figure 1 illustrates the high correlation between Iowa corn and ethanol prices especially starting in 2008 when ethanol’s share of total US corn production had grown to about 20%. These price signals causes producers to respond with intensification and extensification of X production and causes consumers to respond with decreased demand for X and increased demand for coproducts of X. As producers improve their efficiency and develop new technologies to increase X production and as X consumption decreases for other (non-biofuel) purposes, the price signal will fade and conceivably reverse over time. Agronomy efficiency improvements developed for X production may find application to production of other crops. Increased biofuel production could lead to decreased, or increased, production of non-biofuel agricultural goods. There are complex economics-mediated connections between food, feed, fuel, and fiber production. Consequential life cycle assessments typically use models of the cascading effects of increased biofuel production and associated coproducts on the interconnected global agricultural system. Consideration of these important economic effects is lacking in the DeCicco et al. analysis.
Fig. 1

Comparison of corn price to ethanol price 2005–2017 (AGMRC, 2017)

3 Estimating the terrestrial and atmospheric carbon exchanges associated with biofuels

DeCicco et al.’s second assertion is that NEP changes related to biofuel production on US cropland can be accurately estimated from agricultural production statistics. DeCicco et al. (2016) estimate the cumulative biogenic carbon emitted from combustion of corn ethanol and biodiesel in the US from 2005 through 2013 (132 TgC) and compare this with the estimated cumulative additional uptake of carbon on US cropland (49 TgC). DeCicco et al. (2016) calculate that for the years 2005–2013, there is a difference of 83 TgC (see Fig. 2) and that 37% (=49/132) of the biogenic tailpipe carbon emissions was offset by additional carbon uptake. As a result, the authors suggest that only 37% of the biogenic emissions should be offset in emissions analysis.
Fig. 2

Cumulative carbon emitted by US biofuel use compared to cumulative additional carbon uptake on cropland from 2005 to 2013 [Reproduced from DeCicco et al. 2016]

The assumption that NEP changes on US cropland related to biofuel production can be accurately estimated from agricultural production statistics is problematic. As shown in Fig. 2, the NEP metric used shows considerable year-over-year variability. Factors other than biofuel demand such as changes in the broader economy and overall demand (e.g., Great Recession in 2008–2010), increasing yields, and changing consumption patterns all play important roles in agricultural decisions. The correlation between biofuel mandates and acres of corn planted is somewhat low (R = 0.6) (Mueller 2016). Perhaps the most significant factor is weather variability. The dip in carbon uptake in 2012 is attributable to a major drought in the Midwest (Westcott and Jewison 2013). Ethanol policy is not responsible for the decline in this year. It is possible that the decline would have been worse in the absence of ethanol demand, but the study only considers year-over-year changes and does not attempt to account for weather effects. As a result, the analysis is highly dependent on the choice of the starting and ending years. As an illustration, consider a slight adjustment to the DeCicco et al. analysis by using a 2006 baseline instead of a 2005 baseline. If the method is robust, we should expect a similar result to the 37% carbon offset originally calculated. As shown in Fig. 3, the offset instead becomes 138%. By the same logic as the DeCicco et al. analysis, this would imply that biofuels’ biogenic emissions are net negative (before accounting for other emissions). The NEP data is so variable and confounded by other factors that a simple 1-year baseline adjustment refutes the main conclusion of the paper. The accounting approach used by DeCicco et al. (2016) is clearly not sufficiently robust to draw meaningful conclusions about biofuel-related NEP changes on croplands.
Fig. 3

Cumulative carbon emitted by US biofuel use compared to cumulative additional carbon uptake on cropland from 2006 to 2013

4 The relevance of agricultural NEP to biofuel life cycle GHG emissions

Finally, we contend that agricultural NEP is not an appropriate metric for determining biofuel biogenic carbon neutrality. To state the obvious, the relevant global warming mechanism for biofuels is the impact on atmospheric greenhouse gas concentrations—principally, carbon dioxide. Thus, biofuel climate impact studies need to ultimately evaluate changes to the net flow of greenhouse gas emissions into the atmosphere. The relevant question is whether biofuels increase or decrease the net amount of carbon emissions flowing into the atmosphere and by how much. The fate of agricultural crops is predominantly short-term consumption as either vehicle fuel or as food for humans or animals. Since in either case, the carbon embodied in agricultural harvests returns to the atmosphere on a similar, short timescale; there is no substantial change in net carbon emissions to the atmosphere when the feedstock is used for fuel instead of food.

To illustrate, we examine the relevant carbon flows from agriculture and fuel production as presented by DeCicco et al. 2016. DeCicco et al. considered the difference between the carbon flows in 2005 and 2013 to evaluate the impact of biofuels. As previously discussed, this type of historical comparison is problematic due to numerous confounding factors including changes in fuel demand and weather effects. Instead, we prefer to consider annual carbon flows and how counterfactual scenarios would alter these flows. Figure 4a presents the results for the year 2013 adapted from DeCicco et al. (2016).
Fig. 4

Material carbon flows (TgC) through the US agricultural-fuel system in a 2013 [Adapted from DeCicco et al. 2016] and in b a 2013 “no biofuel scenario”

The climate impact of this combined agriculture-petroleum system lies in the net carbon emissions to the atmosphere. On the input (left side), the 215 TgC uptake on cropland offsets the total output (right side) emissions of 715 TgC resulting in a net carbon addition of 500 TgC to the atmosphere. As a thought experiment, consider a scenario where biofuels were prohibited in 2013 and that the energy content of biofuel on a per carbon basis was the same as that of petroleum-derived fuel. Assuming the agricultural outputs and motor vehicle demand remained the same (i.e., no market responses to biofuel prohibition), the resulting flow is shown in Fig. 4b. Extraction from petroleum reserves would need to increase to 529 Tg (approximately 6% increase), and petroleum processing emissions would increase to 96 Tg to replace the 24 Tg of biofuel.

In this decoupled system, the carbon uptake on cropland remains the same (215 TgC) while the output emissions are greater, 744 TgC, because of the increased use of petroleum for motor vehicle fuel. At the end of a year’s time, this results in a net flow of 529 TgC (744–215 TgC) into the atmosphere. By comparing the actual and counterfactual scenarios, the conclusion from this simple exercise is that biofuel production reduces carbon flow to the atmosphere by 29 TgC. One could consider a similar counterfactual scenario where agricultural markets do respond to the elimination of biofuel, and crop yields decline. However, the net result would remain the same since the resulting biogenic emissions associated with the harvested biomass are completely offset by the initial uptake of biogenic carbon by the biomass. Assuming no changes to soil carbon and assuming the harvested biomass is consumed and emitted back into the atmosphere within a year, in this scenario, the net flow of 529 TgC is unaffected by the agricultural yields, whether it decreases, increases, or stays the same.

Clearly, these carbon flows are not representative of the complete carbon impacts associated with biofuel production. Impacts associated with energy and material inputs associated with growing, harvesting, and processing biofuels are not captured in these flows, but they are represented in life cycle assessment studies of biofuels. Furthermore, prominent life cycle assessment models, such as the GREET model, account for additional carbon impacts resulting from induced land-use change. These models more comprehensively describe the net impact of biofuels.

Again, this is not to say that increasing agricultural NEP is an undesirable goal. Increasing agricultural NEP (e.g., intensification) is a mechanism that can reduce competition between harvested biomass for food and fuel, thus limiting land-use change necessary to offset displaced food resources (extensification), but this is already understood. The impact of biofuels on food availability, price, and import/exports, and overall impact on social sustainability (i.e., human wellbeing) is an important consideration that deserves continued attention and study. However, beyond land-use change considerations (which are included in consequential life cycle assessment research), it is not pertinent to climate. Increases (or decreases) in the NEP of agricultural systems, in itself, do not significantly alter the carbon emissions to the atmosphere.

DeCicco et al. (2016) claim that their results falsify the fundamental assumption about carbon flows associated with biofuels made in traditional life cycle assessments. In contrast, we believe that a close examination of their data and assumptions in fact falsifies the analytical approach used by DeCicco et al. (2016).


  1. AGMRC (2017) Monthly comparison—energy prices vs. grain prices. Agricultural Marketing Resource Center. Accessed March 15, 2017
  2. Chapin FS, Woodwell GM, Randerson JT, et al. (2006) Reconciling Carbon-cycle Concepts, Terminology, and Methods. Ecosystems 9: 1041–1050.Google Scholar
  3. DeCicco JM (2013) Biofuel’s carbon balance: doubts, certainties and implications. Clim Chang 121(4):801–814CrossRefGoogle Scholar
  4. DeCicco JM, Liu DY, Heo J et al (2016) Carbon balance effects of U.S. biofuel production and use. Clim Chang 138:667–680CrossRefGoogle Scholar
  5. GREET (2016) GREET model, the greenhouse gases, regulated emissions, and energy use in transportation model,
  6. IPCC (2000) In: Watson RT, Noble IR, Bolin B, Ravindranath NH, Verardo DJ, Dokken DJ (eds) Land use, land-use change, and forestry. Cambridge University Press, UK, p 375 Google Scholar
  7. Lovett GM, Cole JJ, Pace ML (2006) Is net ecosystem production equal to ecosystem carbon accumulation? Ecosystems 9:152–155CrossRefGoogle Scholar
  8. Mueller S (2016) Review of John DeCicco et al. article titled “Carbon balance effects of U.S. biofuel production and use.” In: Energy Resources Center; The University of Illinois at Chicago. Accessed 10 Sep 2016
  9. UNFCCC (2006) United Nations Framework Convention on Climate Change, updated UNFCCC reporting guidelines on annual inventories following incorporation of the provisions of decision 14/CP.11, report FCCC/SBSTA/2006/9Google Scholar
  10. Westcott P, Jewison M (2013) Weather effects on expected corn and soybean yields. United States Department of Agriculture, Washington, DCGoogle Scholar
  11. Woodwell GM, Whittaker RH (1968) Primary production in terrestrial ecosystems. Am Zool 8:19–30CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media B.V. 2017

Authors and Affiliations

  • Robert D. De Kleine
    • 1
  • Timothy J. Wallington
    • 1
  • James E. Anderson
    • 1
  • Hyung Chul Kim
    • 1
  1. 1.Research & Advanced EngineeringFord Motor CompanyDearbornUSA

Personalised recommendations