For many years, governments and researchers have debated whether corn and wheat ethanol reduce greenhouse gas emissions when counting the emissions from land use change needed to replace the food, and the debates have relied on different global agriculture and land use models. Ultimately, the U.S. Environmental Protection Agency and the California Air Resources Board found that these ethanols did generate modestly fewer emissions than gasoline, relying on different models. The European Commission also commissioned a model that found lower emissions. Meanwhile, many supporters of biofuels argue that because all these models differ in their results, impacts on both greenhouse gases and food are too uncertain to reflect in policies.  

In this paper, we analyzed the model results carefully and found that they estimate lower emissions for ethanol because they estimate that from 20-50% of the calories in food are not replaced. Physically, the emissions result from reduced respiration of carbon dioxide (and wastes) by people and livestock. The models attribute these "savings" to the biofuels.  And if these estimates were wrong, the models would estimate higher greenhouse gas emissions.  The paper therefore highlights that much of the debate between models is which adverse effects predominate not whether these ethanols generate adverse effects.

Those without subscriptions to the journal can gain free access to the paper through these links.

Summary: http://www.sciencemag.org/cgi/content/summary/347/6229/1420?ijkey=3HDvP4VYvLiK2&keytype=ref&siteid=sci

Reprint: http://www.sciencemag.org/cgi/rapidpdf/347/6229/1420?ijkey=3HDvP4VYvLiK2&keytype=ref&siteid=sci

Full Text: http://www.sciencemag.org/cgi/content/full/347/6229/1420?ijkey=3HDvP4VYvLiK2&keytype=ref&siteid=sci