Scrutinizing the logic on biofuels

This post picks up a thread based on comments that Prof. Robert Brown and Prof. Bruce Dale made on my "Don't pitch low-carbon fuel ..." post. Both Robert and Bruce take strong exception to my analysis. They invoke the commonly made assumption that substituting a biofuel for a fossil fuel reduces net CO₂ emissions because biofuel use recycles carbon while fossil fuel use does not.

This widely used view of the biofuel lifecycle does not tell
the whole story.  [Image credit: www.EthanolRFA.org] 

Their argument is based on the following logic:

(1) Fossil fuels send old carbon on a one-way trip to the atmosphere, thereby increasing the amount of CO₂ in the atmosphere.

(2) Biofuels use carbon recently taken from the air that is then released back to the air, resulting in no net change in the amount of CO₂ in the atmosphere.

(3) Therefore, substituting a biofuel for a fossil fuel reduces the rate of CO₂ buildup in the atmosphere.

Although this basic analysis neglects processing emissions, we can leave those aside for the purpose of this discussion. They are not what's at the heart of the disagreement, and in any case processing emissions do get tracked by lifecycle models, e.g., as used in the RFS and LCFS. 

There is a fundamental problem with the logic of steps (1)-( 3) as outlined above. At issue is the use of inconsistent system boundaries. The fossil fuel system of step (1) is physically different from the biofuel system of step (2) because the fossil fuel system does not include ecologically productive land. So the common logic compares a system of fossil fuel production and use, which lacks productive land, to a distinct system of biofuel production and use that does include productive land. In other words, it involves an "apples to oranges" comparison, leading to an incorrect conclusion.

The crux of the issue is the need for consistent treatment of land within the system boundary. Note that we're not even talking about land-use change at this point, we're just addressing the carbon flows on existing productive land. In this context, "productive" means land that has some level of positive net ecosystem production (NEP), i.e., an area of the biosphere that, on balance over the course of a year, removes more CO₂ from the air than gets released from that same area through plant respiration or decomposition. Any harvested cropland has strongly positive NEP, the vast majority of which is the carbon embodied in the harvest.

However, farmland removes CO₂ from the air regardless of the disposition of its harvest. Therefore, shifting the use of the harvest (say corn or soybeans) from a prior use (as feed or food) to an energy use (such as ethanol or biodiesel) does not itself change the rate at which the land is removing CO₂ from the air. 

Now, there will be displacement effects, that is, the need to make up the food or feed use displaced for fuel use by growing replacement crops somewhere else, or through a reduction in food consumption. Displacements can be partly offset with coproducts, such as DDGS used for feed. The displacement effects will change net CO₂ releases in other locations, but estimating those changes requires complex economic modeling as done by EPA and others who compute lifecycle GHG estimates for fuels. Such displacement effects can also lead to land-use change, which is even more difficult to quantify and serves to further increase the uncertainties regarding net impacts.

But let's return to the core carbon cycle logic, this time using a consistent system boundary that always includes productive land and represents a single physical system that incorporates both the biofuel supply chain and the fossil fuel supply chain. Without getting into either processing emissions or displacement effects, does substituting a biofuel for the fossil fuel reduce CO₂ emissions, and if so, where?

Well, there is no change in how much CO₂ is removed on the land; it's the same regardless of the where the harvest ends up. At the same time, there's no significant change in CO₂ released from fuel combustion, because similar fuels release essentially the same amount of CO₂ per unit of useful energy. So, to first order, the substitution of the biofuel for the fossil fuel results in no net change in the amount of CO₂ flow to or from the atmosphere. Yes, some fossil carbon stays in the ground, but a net reduction can only occur if more CO₂ gets removed from the atmosphere somewhere else. But that does not happen if land that is already in production is used to grow the feedstock, as is the situation for the vast majority of biofuel produced today.

Are there circumstances where biofuel use can reduce CO₂ emissions? Yes, of course, and what's required is a demonstrable increase in the net rate of CO₂ uptake where the feedstock is sourced. But that's not happening anywhere at commercially significant scale.

This analysis is fully described in my Biofuels' Carbon Balance paper, which provides a formal stock-and-flow analysis of a coupled biofuel and fossil fuel system. Ecologically, it reduces to the condition: 

      d(NEP)/dt > 0

That is to say, the threshold condition for a biofuel to have a CO₂ mitigation benefit is that the production of its feedstock must raise net ecosystem production. The failure to correctly address this condition is a fatal error at the heart of the lifecycle models used by many bioenergy researchers and policy makers. 

Editor's note: see also the newer post, When do biofuels really balance carbon?