What’s the real impact of indirect land-use change?

Alternative fuels Life-cycle analyses

If shifting crops from food to biofuels may have unintended effects on land use and climate change, should policymakers consider those impacts when designing fuels policies? David Zilberman’s recent editorial in Global Change Biology Bioenergy titled “Indirect Land-Use Change: Much Ado About (Almost Nothing)” questioned the practice of including emissions attributable to indirect land-use change (ILUC) in biofuel policy design. Zilberman, a professor in Berkeley’s Agricultural and Resource Economics Department, argued that efforts to include the indirect, market-mediated GHG emissions from cropland expansion in biofuels policies, such as California’s Low-Carbon Fuels Standard (LCFS) and the Renewable Fuels Standard (RFS), have been “flawed and misleading,” and that quantifying those emissions, which are in his view small and uncertain, was a wasted effort and has perhaps even stifled investment and growth in the biofuels industry. Is the use of ILUC in biofuel policies just a case of fighting an imaginary enemy? Not so fast.

In the context of fuels policy, ILUC is typically quantified in terms of an emission factor similar to other components of a given fuel’s production emissions, such that the climate impact of land-use change is presented in grams of carbon-equivalents released per unit of fuel produced. While ILUC factors vary according to feedstock, underlying econometric models, and scenario assumptions, for many feedstocks they are large enough to fundamentally change our understanding of the climate benefits of biofuels.

Take, for example, a recent study conducted on behalf of the European Commission using the GLOBIOM economic model to assess the ILUC effects of biofuels consumed in response to European policies. While the direct emissions from commonly consumed biodiesel feedstocks such as soybean oil and canola oil were significantly smaller than those from palm oil, their overall emissions were significantly higher after accounting for ILUC. In fact, the well-to-wheel emissions from soy biodiesel exceeded those of petroleum diesel when the ILUC effects were considered, cancelling out the climate benefits of switching to biodiesel in the first place. If acknowledging the ILUC for a given fuel can make or break its GHG savings, shouldn’t it be a key factor in designing biofuel policy?

One of the main reasons GLOBIOM, as well as an earlier study using the MIRAGE model, report high ILUC emissions for all types of biofuel from vegetable oil feedstocks is that these oils are closely linked to palm oil production. In fact, the GLOBIOM study points out that the direct emissions from rainforest destruction and peatland drainage would more than offset the emissions reductions from displacing petroleum and thus disqualify palm oil as a sustainable fuel. Other fuels are also affected by the indirect impacts of palm oil emissions: In our previous research we’ve demonstrated a link between increases in soy oil price and rising palm oil imports to the U.S., indicating that increasing demand for U.S. biodiesel feedstocks is driving palm oil expansion. This linkage in vegetable oil markets is reflected in the high ILUC estimates for biodiesel pathways using the GLOBIOM and MIRAGE models.

A key pillar of Zilberman’s argument is that estimates of ILUC emissions have decreased in recent years as our understanding of the effect has grown and models have been refined. However, the GLOBIOM results stand in stark contrast to that trend. While it is true that the very first estimates of ILUC from Tim Searchinger in 2008 exceeded a whopping 100 gCO2e per MJ of corn ethanol, it’s wrong to conclude that the recent decline in estimated corn ILUC towards 20 gCO2e per MJ reflects a decline in all ILUC estimates. Much of that decline appears evident only in results produced from a single econometric model, GTAP, which are then compared to Searchinger’s very early study. But if other models, such as GLOBIOM and MIRAGE, and additional food-based feedstocks, such as soy, are considered, a very different picture emerges. Adding more data points by using ILUC results from other models, which make different land use and soil carbon assumptions, as well as expanding our scope to assess more types of food-based biofuel feedstocks, suggests that in fact the trend of a universal decrease in ILUC intensity is overstated.

Acknowledging the impact of ILUC on biofuel carbon intensity doesn’t mislead policymakers and waste their effort; rather, it shines a light on a key effect of biofuel policies. Emissions attributable to ILUC can’t be ignored—they constitute a sizeable share of the life-cycle emissions for some biofuels. Therefore, it is critical that those emissions are factored into the carbon intensity calculations underpinning biofuels policies in order to ensure that those policies support only the fuels that produce the most substantial GHG savings. This methodological approach can have profound effects: high ILUC values for some first-generation biofuels mean that they offer minor GHG reductions, if any, relative to petroleum, and could even have higher emissions when fossil fuel rebound effects are taken into account. Rather than stifling the biofuels industry as a whole, ILUC factors encourage both the development of innovative technologies to reduce direct emissions and the transition to second-generation feedstocks that have smaller indirect effects.