Smith and Searchinger (2012) argue that existing life cycle assessments (LCAs) pertaining to biofuels seriously overestimate carbon absorption on the part of bio­energy crops and do not take sufficient account of GHG emissions resulting from the cultivation and harvesting of these crops.

The type of land, i. e. unfertilized grassland, forest land or traditional cropland, used for biomass feedstock is an important determinant of GHGs emitted from the soil (EPA 2006). Preparing fallow or underutilized land for agricultural production usually requires clearing off the majority of the animal and plant species. This can destabilize the soil by releasing significant amounts of stored carbon (EPA 2006). Some studies conclude that conversion of native land such as forest, grassland and abandoned land for biofuel crops leads to carbon debts[15] ranging from one to sev­eral 100 years (Fargione et al. 2008; Gibbs et al. 2008; Fritsche 2008). For exam­ple, Fargione et al. (2008) estimated the carbon debt of producing palm oil on forest land (releasing 3452 tCO2/ha) to be approximately 423 years. Table 5 below provides an overview of the estimated payback periods for a range of biofuels.

In contrast, biofuel crops grown on traditional croplands are less threatening to the environment since they have less embedded soil organic carbon (SOC) (Englund et al. 2011). However, intensive biofuel cultivation, especially if using annual crops, could lead to a substantial release of SOC. This is due to frequent disturbance to the soil (i. e. via tillage), which exposes protected organic matter and increases the rate of mineral decomposition, thereby resulting in lower SOC storage (Grandy and Robertson 2007).

Aside from tillage, farming and irrigation practices could also affect the net carbon balance of biofuels. Mechanized farming or the use of fossil-fuel-powered machinery for soil preparation, sowing, planting, weeding and harvesting activi­ties releases GHGs. Likewise, water for irrigation of biofuel crops is often sourced from rivers, lakes, canals, dams and groundwaters. While this reduces water avail­ability for other uses, it also leads to soil salinization when the irrigation process is poorly managed (Englund et al. 2011). These impacts can be mitigated where rain harvesting systems such as terraces, bunds and small dams are available.

Another issue related to harvesting is mono-cropping, or the planting of only a single species or cultivar. While harvesting a particular biofuel crop on a large-scale over several years makes the process more economical, it can also increase the environmental footprint. Repetitive harvesting of a single variety of crop results in a

lack of biodiversity and a decline in soil fertility. To control pests and maintain yields in such environments, more chemical input and fertilizers are generally applied (Englund et al. 2011), which can lead to serious ecological impacts (more in Sect. 5 in A Comparison Between Ethanol and Biodiesel Production: The Brazilian and European Experiences). However, as Dale et al. (2010) report, such impacts can be minimized by adopting sustainable land management practices.[16]

Studies of LCAs have shown that GHG emissions can vary substantially between biofuels, but are mostly lower than those associated with conventional fossil fuels. Through a meta-analysis of LCA literature, Davis et al. (2008) found that the results range between -89 MgCO2 per hectare per year for corn-based biofuel (Farrell et al. 2006) to 9.6 MgCO2 per hectare per year[17] for biofuel pro­duced from switchgrass (Searchinger et al. 2008). Results also varied between authors for biofuels produced from the same crop. For example, Shapouri et al. (2002) found that corn ethanol reduces CO2 emissions by 1.2 Mg per hectare per year, while Delucchi (2006) determined that it increased CO2 emissions by 5.14 Mg per hectare over the same period. Some studies reported the results in terms of change in GHG emissions compared to fossil fuels. The variation in this case was once again large and ranged between -114 % for switchgrass (Adler et al. 2007) to 93 % for corn (Searchinger et al. 2008). LCAs, however, have often overlooked the impacts of LUC on overall GHG emissions. When Bailis and Baka (2010) compared biodiesel from Jatropha in Brazil with conventional biodiesel without considering LUC, they noted a 55 % reduction in GHGs. In contrast, when they included LUC, the net emissions were estimated to increase by 59 %.

Despite providing a cradle-to-grave assessment, LCAs therefore reach varying conclusions on any biofuel depending on the methodological approach adopted. While using an LCA should ideally be an ongoing process for handling and prior­itizing information as new data comes to hand, it is worth noting the “seven grand challenges” that McKone et al. (2011) identified for undertaking a comprehensive LCA of biofuels. These are

• Understanding farmers, feedstock options and practices.

• Predicting biofuel production technologies and practices.

• Characterizing tailpipe emissions and their health consequences.

• Incorporating spatial heterogeneity in inventories and assessments.

• Accounting for time in impact assessments.

• Assessing transitions as well as end states.

• Confronting uncertainty and variability.

A proper understanding of these issues will have profound implications with respect to what feedstocks should be used for biofuel production, together with what lands are most suitable for environmentally sustainable feedstock produc­tion. Any conclusion reached from an LCA must consequently be tempered by the knowledge that the same assessment could provide a different result at another point in time.