This research assesses the effect of greenhouse gas (GHG) emission constraints imposed in biofuel importing countries on the export potential of biofuel producing countries. Several countries are promoting the introduction of biofuels on their energy matrix through ambitious biofuel mandates but also specify a certain level of GHG emission reduction that biofuels should fulfil. Biofuel producing countries focused on the international market should comply with this criterion in order to supply biofuels to those countries. Biofuel producers should then report the GHG emission saving (GES) of the biofuel they supply. A critical issue in this assessment is the inclusion of GHG emissions from land-use change (LUC) induced by the production of feedstock for biofuels. Focusing on the Argentinean case, this thesis analyses the soybean-based biodiesel export potential of Argentina to the European Union (EU), including the GES threshold imposed in the EU Renewable Energy Directive (RED). The thesis therefore focuses on estimating the biofuel GES based on the impact of soybean production on direct land-use changes (dLUC) at the country level. Key factors influencing this result include the policy framework regulating the biofuel supply chain, the evolution of prices and demand for soybean-based products and the feedstock production patterns. The thesis proposes a modeling approach to assess the effect of these factors on soybean-based biodiesel production and exports. The approach is based on a market analysis of soybean and of higher value-added products, a conceptual modeling framework and a simulation model. The market analysis serves as a background study to define the modeling foundations. The conceptual modeling framework specifies the main interaction among producers in the biodiesel supply chain and their link to international markets, land-use changes and GHG emissions. Simulations are then performed to assess how those key factors affect the Argentinean (AR) biodiesel export potential to the EU. To this end, a system dynamics simulation model is developed. The simulation model includes a life cycle assessment model used to estimate the biofuel GES. The research explicitly addresses the allocation of biodiesel production between two types of producers and two market destinations, provided that specific policies regulate the domestic biodiesel industry. Land supply for soybean production is estimated based on the evolution of demand for soybean, competing and higher value-added products. Dynamics in the international markets are addressed through a scenario-based approach to define a plausible scenario of the market evolution. Feedstock production patterns are accounted for by disaggregating soybean production in four different regions (Centre, South-East, North-East and North-West). In each region, the expansion of managed lands is modeled based on the current share of three soybean cultivation methods and seven unmanaged land types. The biofuel GES is finally compared with the EU-RED GES threshold to estimate the biofuel export potential under GHG emission constraints. Results indicate that the impact of biodiesel production on soybean land supply was small compared with the effect of soybean oil and meal exports. While biodiesel production affects mainly soybean oil exports, this effect is still marginal given the biodiesel production level and the economic value attached to soybean meal for the given scenario. Land supply for soybean production therefore seems to depend more on how Argentinean soybean meal exports affect the price of soybean in the international market. Despite the large share of Argentina in the soybean meal export market, this market is likely to be competitive. Biodiesel domestic policy instruments significantly affect the biodiesel export potential, especially when different domestic blending targets are applied. With respect to the national biodiesel mandate large firms are mainly export oriented while small and medium firms exclusively supply the domestic market. Moreover, export taxes seemed to significantly affect the biodiesel export potential through its direct effect on producer profits. Feedstock production patterns largely influence dLUC from soybean production. The supply of cropland for soybean cultivation differs among regions. Higher land productivities and the application of first-occupation no-tillage farming in the Central region led to higher net returns to land and lower land requirements. Soybean cultivation in the Central region leads mainly to displacement of other crops and pastures, given constraints in land availability. Cropland supply in other regions resulted in higher dLUC due to lower land productivities and the application of conventional tillage that lead to lower yields. In the South-East and North- East regions cropland expanded mainly into mixed land, grassland and shrubland. In the North-West region, cropland expansion into forests resulted in significant GHG emissions from dLUC. The allocation of dLUC from cropland expansion to biodiesel resulted in different biodiesel export potentials. Producers located in the C region seemed to be those with the highest potential for exporting biodiesel, given their higher profits and higher GES compared with other regions. Producers in the C region can supply biodiesel to the export market with a GES of 45% complying with the EU-RED GES threshold, at least until 2017. If no dLUC occurs, the GES for biodiesel produced in this region rises to 57%. Supply by other regions to the international market is constrained by the non compliance with the GES threshold. Perspectives for further research include additional simulations to assess the biofuel GES and the export potential under other market scenarios and policy contexts. The modeling framework may be extended to the individual producer level and may also be linked to a global approach to improve the modeling of market interactions in the world economy and the accounting of indirect land-use change. Finally, the extension to geographic information systems (GIS) can improve the representation of land heterogeneity and the induced land-use changes from soybean production.
EPFL2012
