Biofuels provide an alternative liquid fuel to petroleum for transportation. Their projected 2008 production, at about 8.0-8.5 billion gallons, will represent about 5 percent of U.S. gasoline market volume. Over the last year, much has been written about whether this increasing volume is good or bad based on whether it may also be causing a reduction in food production, an increase in food prices, and a net increase in greenhouse gas emissions from new land coming into production or simply from petrocarbon energy needed to grow the feedstock. In this article, we will examine this issue and propose a policy framework that could encourage the development of biofuels in a way that produces a net benefit to food production while reducing greenhouse gases. The framework has three elements: (1) fully account for all land-use impacts of biofuels, (2) focus incentives towards biofuels with higher productivity and greater co-benefits, and (3) move towards an integrated food and fuel land use policy.

Background

Transportation is responsible for one-third of U.S. greenhouse gas (GHG) emissions. To address climate change, the U.S. needs a combination of: driving fewer miles, efficient vehicles and fuels that emit less GHGs. Over the past five years, E2 has written about the opportunity for alternatives to displace petroleum fuels to both reduce greenhouse gases and provide greater energy security:

• Hydrogen - Is it the Answer for Clean Cars? (August 2003)
• What Comes After Oil? (June 2004)
• Can Biofuels Replace Oil? (November 2004)
• Alternative Fuels Bill Passes California Assembly (June 2005)
• What will it take to eliminate foreign oil? (February 2006)
• Will Electricity Challenge Ethanol? (June 2006)
• Developing Low Carbon Ethanol for California (July 2006)
• The Future of Transportation (October 2007)

In these articles, we examined the role of alternatives to petroleum fuels including: hydrogen, biofuels, algae and electricity. We discussed how fuels could be developed to scale in a manner that is cost-effective and also low in greenhouse gas emissions.

The Problem

Our articles did not consider the "indirect effects" of using land for biofuels. For example, if we replace 60,000 acres of alfalfa (used for animal feed) with sugar cane to produce 100 million gallons of fuel, what happens to the alfalfa market? Does someone else grow additional alfalfa and does that happen in a manner that causes a net increase in greenhouses gases? It may also be possible for a modern cane-to-ethanol process to produce fuel and high-protein animal feed bi-products that can replace alfalfa. Sugarcane might be grown in a way that improves the annual soil and biomass accumulation to an extent greater than what the original alfalfa crop could provide. We need to calculate the negative and positive impacts of these indirect effects.

In a January 2008 article in Science magazine, Timothy Searchinger and colleagues explored the question of indirect land-use impact on stored soil carbon. The chart below (courtesy of NRDC) illustrates a scenario for corn ethanol:



The article calculates that the complete annual GHG emissions of growing the corn and producing the ethanol are 2.7Mg CO2e/acre (million grams of CO2 equivalent per acre) and the gasoline avoided saves 3.4 Mg CO2e/acre for a net benefit. However, there is still demand for that lost acre of corn which causes 0.85 new acres to come into production. (The protein left over from making corn ethanol is sold as animal feed and this is why only 0.85 acres are replaced and not the whole acre.) If the lost acre of corn is replaced by corn produced from clearing new land (based on certain assumptions for a standard acre of land), 4.7 Mg CO2e/acre of previously stored carbon is released, resulting in an overall increase in GHG emissions of 3.3 Mg CO2e/acre!

The direct GHG emissions from producing corn ethanol are very dependent on the specific refineries, but this can be reasonably measured. The State of California has published a list of GHG emissions for a variety of feedstocks and refinery techniques (see State Alternative Fuels Plan, page 32) without considering any indirect effects.

The Searchinger study examines whether the lost corn is actually replaced, the type and location of new land put into production, and what that land’s GHG characteristics were before being used for corn. The study is modeled on data from specific types of land that were converted for agriculture during the 1990’s. Assumptions about how the conversion happens and the type of agricultural practices used after the conversion are based on historic trends. Since land in undisturbed ecosystems is very good at storing carbon, disrupting that land releases carbon and creates a "carbon debt" which may take many years to recover through biofuels displacing gasoline (see "Land Clearing and Biofuel Carbon Debt" by Fargione, et. al., for details).

Increasing demand for corn ethanol has had an impact on the price of corn, but ethanol is not the only factor increasing food prices. Food price increases are also due to increases in the prices of fertilizers (derived from natural gas) and diesel fuels used to work the farm and deliver the products. Price increases are not limited to corn. Rice, for example, has increased in price dramatically and no rice is used in the production of biofuels. Global food demand is increasing, as developing economies can afford better diets, while production has lagged. Food price inflation is a complex issue on which multiple opinions exist. Biofuels alone do not explain rising food prices. (See "The Effects of Ethanol on Texas Food & Feed," Texas A&M University, for one analysis.)

Biofuels may also be helping to lower the cost of petroleum. A June 6 Merrill Lynch report, "Biofuels driving global oil supply growth," says: "On a global scale, biofuels are now the single largest contributor to world oil supply growth. We estimate that retail gasoline prices would be $21/bbl higher, on average, without the incremental biofuel supply."

Why Single Out Biofuels?

A major policy goal for encouraging biofuels is to reduce GHG emissions compared to fossil fuels. We must account for the total impact of biofuels - direct and indirect. We also must protect against shifts in land use that harm food production. There are several ways to reduce the indirect land-use GHG emissions from biofuels: (1) reduce/eliminate the land requirements, (2) increase the co-benefits of additional protein or carbon sequestration, and (3) displace more gasoline by increasing the yield per acre. Policies need to prioritize and reward biofuels that can do all three and discourage those that can’t.

There are about 440 million acres of land in crop rotation and cropland pasture, representing about 20 percent of all land in the U.S. That land and our food consumption are not optimized for either the healthiest diet or the lowest GHG emissions. Fifty percent of corn in the U.S. is used for animal feed and high-fructose corn syrup. (In 2007: 46 percent feed/residual, 25 percent ethanol, 19 percent export, 4 percent corn syrup, 6 percent other.) Cattle are fed corn - which they do not eat naturally - instead of grasses. The result is both an increase in GHG emissions and, as extensively documented in "The Omnivore’s Dilemma" by Michael Pollan, a less healthy human diet. Could we encourage a change in agricultural practices that reduces demand on land, reduces GHG emissions and produces a healthier diet? The answer is "yes" but to date, agriculture has been "off limits" as a topic of discussion. A well-designed land-use policy must ultimately balance food production, fiber production, diet and fuels. Until that policy exists, the burden of avoiding GHG increases from land-use changes will rest with biofuels.

We also need to think globally. In addition to reducing U.S. GHG emissions, we should develop technologies that can help developing countries independently meet their food and fuel needs. For example, Mozambique alone has 95 million acres of total arable land with available water, of which only about 12 million acres are in use. The government has recently completed an extensive land mapping process to make sure land use is properly allocated between food, fuel and environment, and is beginning to approve food and fuel development projects against that plan in order to meet local and export demand. These projects will provide foreign capital, jobs, food production, fuel production and general tax funds to improve the lives of people in one of the poorest nations on earth whose land and water are increasingly valuable national assets that can benefit this emerging economy. Production on much of that land can be done in a manner that not only does not release GHGs but also sequesters additional carbon (see "Biofuels, climate change and industrial development").

Corn is not the answer

We do not believe corn ethanol is the long-term solution for petroleum displacement. It has created a significant ethanol market - approaching $20 billion in 2008 (assuming $2.50 average wholesale price/gallon) - and now is the time to use policies to guide the market forward with performance standards, independent of specific technologies, which have the following benefits:

• Net GHG Reductions - significant life-cycle reduction in GHGs compared to petroleum
• Scale - capable of scaling to displace a significant portion of gasoline and diesel demand
• Water - does not create new demands on fresh water
• Environmental - neutral or positive benefits for air/water quality & toxics
• Economics - cheaper than gasoline and diesel at scale; possible benefits to small landholders
• Land - minimal additional land required or use of marginal land

We believe solutions that aim toward these benefits will also have a neutral to net benefit to global food supply because they will not compete for land and they create competition for fossil fuels resulting in a more competitive energy market. These six benefits will need to be pooled to produce a net benefit.

Existing Policy Mechanisms

There are three primary types of policy structures currently in use for alternative fuels:

• Distribution Requirements - A requirement for distributors to use a certain volume of alternative fuels. The federal government has established a mandate for a volume of ethanol to be purchased - typically by blending the ethanol with gasoline. The requirement includes specific environmental safeguards.
• Performance Standards - A requirement for distributors to reduce the lifecycle GHG emissions from their fuels. The California Low Carbon Fuel Standard (LCFS) is expected to require a 10-percent reduction in the carbon intensity of fuels sold in California by 2020. The standard will include specific environmental safeguards.
• Subsidy - The federal government currently pays $0.45/gallon as a subsidy to the blenders of corn ethanol, $1.01/gallon for cellulosic ethanol, $0.60/gallon for renewably derived alcohol fuels and up to $1/gallon for biodiesel.

A few years ago, when oil sold for $40/barrel, alternative fuels were dependent on the mandate and the subsidy. With oil at $125/barrel, the U.S. market is not necessarily dependent on either. This weakens the environmental protections because biofuels producers can profitably sell their fuel even if it doesn’t qualify under the federal requirements. This problem can be addressed by imposing performance standards (such as the LCFS or a carbon cap) on all fuels and by limiting subsidies only to alternative fuels with low GHG emissions and the other benefits outlined above.

Biofuels Land-Use Policy Proposal

The direct GHG emissions from the growing, processing and use of biofuels is relatively well understood. It is the indirect land-use impacts that must be calculated.

In June 2008, a group of scientists wrote the California Air Resource Board recommending that indirect land-use not be included in the LCFS for the next five years due to lack of understanding and scientific consensus on the issue. This was countered by a group of scientists from the University of California at Berkeley and Davis who argue that while there is uncertainty, the expected impact is not "zero." The evidence continues to show that emissions from indirect land-use impact can be large, and a policy that treats those as zero may allow bad solutions to come to market that will be harder to correct later on. We agree that it would be a mistake to treat indirect land-use emissions as zero. At the same time, the difficulty of modeling indirect land-use emissions is very dependent on the source of the biomass and the agricultural product they displace.

We propose placing biomass sources into "bins" based on their indirect land-use impacts and the type of product they displace. The "bin" determines the indirect land-use GHG penalty they are assessed and is based on the best available modeling of indirect land-use GHG emissions. We also propose a credit that can be applied against the indirect land-use GHG penalty for biofuels with significant co-benefits and yields.

Biofuels can be manufactured from a variety of organic materials. These fall into five primary bins:

1. Agriculture and Forestry Residues (rice straw, wheat straw, corn stover, managed forest thinnings, etc.)
2. Waste Feedstocks (waste juice, fermentation solids, manure, municipal waste etc.)
3. Winter Rotation Crops (grown for ground cover in certain climates)
4. Primary Energy Crops (energy cane, switchgrass, etc.) that might also have food co-products or are grown on land not currently suitable for food
5. Food Crops (corn, soy, etc.)

The first two bins (residues and waste) use no new agricultural land and would have no indirect land-use GHG penalty. In some parts of the country, winter rotation crops can be managed in a way that has minimum to no additional land use and opportunities are likely to expand as more research is done on this approach. Primary energy crops have land-use impacts. This category needs to be further divided based on the type of existing land used. For example, crops such as cotton may be going out of production anyway and therefore the land use associated with the energy crop grown in its stead may have a small penalty, or none at all. Current food crops have a large indirect land-use penalty and that will continue unless productivity improvements can outpace demand for food and fuel resulting in no increase in total land use.

According to the April 2005 study "Biomass as Feedstock for a Bioenergy and Bioproducts Industry" by the U.S. Departments of Agriculture and Energy, up to 1.3 billion dry tons of biomass could be made available annually. Excluding scenarios of land-use change and productivity gains, we could get up to 325 million dry tons of biomass without any impacts on existing agriculture or forestry. At a conversion efficiency ratio of 80 gallons per ton, this would be 25 billion gallons of fuel, or up to 15 percent of the 180 billion gallons currently consumed annually (the actual amount is dependent on whether ethanol, renewable diesel, or other fuels are produced; and the energy penalty associated with ethanol).

Productivity and Co-Benefits

The productivity of current biofuels includes corn (up to 450 gallons/acre) and Brazilian sugar cane (around 800 gallons/acre - assuming 85 tons/hectare and 85 liters/ton).

Emerging "second generation" processes can extract the fermentable sugar locked inside plant cells (cellulosic processes) or reduce them to energetic gases. These processes enable many more energy crops and much higher value per acre. Eight to 20 dry tons per acre can be produced, depending on geography and advanced genetics, and at a conversion of 80-100 gallons per dry ton, total yields can range from 640 to 2,000 gallons per acre.

Instead of converting biomass into ethanol, new companies such as Amyris and Solazyme may be able to economically convert sugars into renewable diesel, gasoline, jet fuel or the equivalent of crude oil that is then processed in a standard refinery. Converting biomass to hydrocarbons rather than to ethanol eliminates the energy penalty associated with ethanol and increases the GHG benefit by displacing more fossil fuels.

It is unclear how much dedicated energy cropland will be needed. While electricity, algae, hydrogen and biomass derived from waste materials and cover crops may play a significant role, at this point we don’t know if they will be enough and we should not discourage investment. We need to encourage innovation that focuses on very high productivity, measured in terms of tons of biomass/acre and gallons of fuel/ton of biomass, and plants that can produce both fuel and protein. For example, soybeans currently produce 0.4-0.65 tons/acre/year of protein. It may be possible for certain grasses to produce significant biomass for fuel and also 0.4-1.2 tons/acre/year of protein.

Crops with either significant biomass productivity (over 12 dry tons/acre producing 1,000 gallons/acre) and/or significant protein productivity may be able to overcome the indirect land-use GHG penalty.

Innovation Credits

The goal of the innovation credit is to give credit that can be applied against the indirect land-use penalty for crops and cropping systems that are making real steps in improving biomass productivity and/or significant protein productivity, and have potential to make larger steps. As this is largely an innovation credit, the credit would not required to be exactly proportional to the current benefits.

In our example above, the indirect land-use penalty is 4.7Mg CO2e. If a future biofuel had twice the productivity per acre of corn it would displace twice the gasoline, overcome the penalty and be GHG neutral (assuming the other factors stayed the same).

To encourage the development and deployment of highly efficient biofuels ideally with co-benefits, indirect land-use credits could be awarded. For example, initially the amount could be 4Mg Co2E/acre and that amount could decline with time. This creates an incentive to develop advanced biofuels and crops and also allows time for our understanding of indirect land-use modeling to improve and for an integrated food/fuel land use policy to be development.

Ranking Biofuel Feedstocks

Based on what we currently know, certain biofuels are clearly bad or good, or are in an ambiguous position.

Known to be Bad

Biofuels produced from deforested virgin land is clearly bad from at least a GHG emissions perspective. The common examples of this practice include palm oil produced on former forest land in Indonesia and soy from Brazil produced on former rain forest. To deal with these cases, a fuel could only participate in a mandate, performance standard or incentive if the company involved does not own, harvest or trade any virgin biomass whose source is feedstock planted on land that was detimbered at any time after the policy was first proposed, unless an independent third party has verified that the biofuel crop is improving the land and its carbon sequestration value.

Known to be Good

Biomass produced from agricultural residues and waste products can be individually inspected and demonstrated not to have any indirect land-use effect.

Ambiguous - but likely

In the "ambiguous but likely to be good" category is biomass produced by winter cover crops and biomass from forests and logging residue. Winter cover crops in areas where there is sufficient precipitation are used to protect and replenish the soil between growing seasons. The crops selected for this purpose have not been optimized to also produce biomass that could be used for biofuels. In one model (see Khosla: "Where will biofuels and biomass feedstocks come from?"), assuming cover crops were used in 50 percent (159M acres) of crop land and they supplied an average of four dry-tons/acre and the crops were converted to fuel at 80 gallons/dry-ton, 50 billion gallons of fuel could be produced by 2030 (or six times this year’s corn ethanol production). While many would consider this scenario to be optimistic, even one quarter of this result would be significant. It may be possible to integrate biomass-optimized cover crops into current agronomic practices in a way that has minimal impact on the ongoing land use in geographies where there is sufficient precipitation to support year-round biomass growth.

Ambiguous

Biofuels that replace other crops or are on new land would require overcoming the indirect land-use GHG penalty by having significant productivity increases and/or other co-benefits. They would be aided by a declining indirect land-use GHG credit.

Summary

First-generation biofuels have established a significant market for alternative fuels, but as demand has grown and the price of oil has increased rapidly, it is apparent that current policy is inadequate to protect against possible adverse effects of biofuels and does not go far enough to ensure that biofuels are done well. We recommend policies that focus on achieving benefits in

• Net GHG Reductions - significant life-cycle reduction in GHGs compared to petroleum
• Scale - capable of scaling to displace a significant portion of gasoline and diesel demand
• Water - does not create new demands on fresh water
• Environmental - neutral or positive benefits for air/water quality & toxics
• Economics - cheaper than gasoline and diesel at scale; possible benefits to small landholders
• Land - minimal additional land required or use of marginal land

Our framework has three elements: (1) fully account for all land-use impacts of biofuels, (2) focus incentives towards biofuels with higher productivity and greater co-benefits, and (3) move towards an integrated food and fuel land use policy. Ultimately, significant emissions from agriculture, forestry and land use need to be part of a total cap on emissions. If that happens, indirect land-use emissions from biofuels production will no longer need to be a special case and the world will be able to find - and benefit from - better solutions.

Acknowledgements

E2 members Paul Zorner and Anna Halpern-Lande, and NRDC staff Nathanael Green and Dan Lashof, contributed to this article. We would like to express our gratitude to the following people who provided critical feedback: Tony Bernhardt, John Dawson, Allen Dusault, Will Gardenswartz, Larry Gross, Dan Kammen, Richard Plevin and Dan Sperling.