Indirect land use change from biofuels: exploring implications of producTION SYSTEM heterogeneity for results and mitigation strategies
Julie Witcover, Institute for Transportation Studies, University of California, Davis, Phone +1 530 792 7278, E-mail: Siwa Msangi, International Food Policy Research Institute, Phone +1 202 862 5663, Fax +1 467 4439, E-mail: Sonia Yeh, Institute for Transportation Studies, University of California, Davis, Phone +1 530 754 9000, Fax +1 530 752 6572, E-mail:
Overview
Recent studies have shown that increased biofuels production can cause expansion of cropland as farmers respond to price changes prompted by diverted feedstock. The phenomenon is referred to as market-mediated, or indirect, land use change (“iLUC”). More land converted to agriculture can result in significant GHG emissions that offset carbon sequestered in growing the biofuel (Fargione et al. 2008; Gibbs et al. 2008; Searchinger et al. 2008). Both US and California biofuels policy have incorporated iLUC emissions estimates to help assess whether or not particular biofuel pathways meet regulatory GHG emissions requirements (EPA 2010; CARB 2009). The iLUC analyses use a combination of models, but rely on economic equilibrium modeling to estimate the magnitude and location of land use change. There is, however, considerable uncertainty surrounding several key parameters within the economic equilibrium models and how they are applied to estimate iLUC given biofuel policy shocks. One factor still subject to uncertainty that has a large effect on iLUC results is yield, with its direct relation to how much additional agricultural land is needed to meet projected demand for agricultural commodities. Currently, models looking at international iLUC effects typically assign yield-related parameters to a geographical area – country or agroecological zone (AEZ) combination. The (average) parameters can cover a range of technologies and production system types, which each embody a different level of responsiveness to both market-driven and biophysical factors.
This study examines how iLUC model assumptions on production system homogeneity or heterogeneity could matter to model results. It also discusses how economic equilibrium models could evaluate or target policy efforts to reduce iLUC. More specifically, the study asks whether a slightly more refined characterization of agricultural production might allow for better use of information that we already possess about production response in alternative types of production systems in terms of yield growth (via extent of irrigation, or the intensity of other yield-enhancing inputs) vs. area expansion. Differentiating among production system responses could help identify areas at risk for agricultural expansion, as well as those promising for yield improvements. The exercise could also help pinpoint what additional information is needed to effectively model the heterogeneity that might matter most for area expansion, and how empirical efforts might best be directed to address this.
Using this approach, the study aims to illuminate how to identify production systems more likely to respond in a way that increases the chance of iLUC happening or offer a possibility of mitigating the iLUC response. While uncertainties and other differences within and across models still preclude definitive statements about where iLUC responses would occur under any given scenario, the study suggests how information derived from improved understanding of production responses could better focus policy actions addressing iLUC. Such policy-relevant considerations might include the question of how to target technology improvements effectively; what kind of trade policies matter for those commodities and production systems and their responsiveness to market-equilibrium shocks; and how to encourage production in areas less likely to elicit an iLUC response, among other potential issues.
Methods
The study consists of a conceptual framework for how heterogeneity in production systems could matter for iLUC, and illustrative experiments with a simulation model to explore the potential of this approach. It uses a global partial equilibrium model for trade in agricultural commodities (the International Model for Policy Analysis of Agricultural Commodities and Trade, or IMPACT), developed at the International Food Policy Research Institute. The conceptual framework suggests how an economic equilibrium model could be used to quantify ranges of likely production response, point to likely policy targets (e.g., regions or farming system types) for reducing iLUC, and provide an avenue for evaluating leakage associated with proposed targeting schemes.[1] The actual model exploration focuses on the distinction between irrigated and rainfed production systems, with discussion on additional disaggregation of rainfed systems into high input, low input, and subsistence categories as well as ways to evaluate policy using such a model.[2] Additional discussion touches on issues surrounding use of an economic equilibrium model to guide policy, including implications of uncertainties within and across models for iLUC and iLUC policy evaluation, as well as institutional and jurisdictional aspects of policies targeting iLUC emissions.
Results
Different production systems employing different technologies (and with different marginal cost curves) are expected to vary in their propensity to respond in terms of area or yield to commodity price changes such as those prompted by biofuels policy. We expect to show that measurable differences would be found in the location and magnitude of indirect land use change in a model that accounts for producer heterogeneity versus one that does not. Furthermore, we expect to illustrate the potential for economic equilibrium modeling to provide a useful policy targeting and evaluation tool for mitigation strategies, but with some caveats due to the inherent uncertainty in modeling of production response that are themselves mitigated by taking some account of producer heterogeneity.
Conclusions
Indirect land use change models have emerged as an important component of biofuels policy, but the magnitude of effects is still uncertain, because of lack of knowledge about current and future behavioral responses and how these responses differ across regions. Sensitivity analyses indicate that iLUC results are highly sensitive to assumptions about agricultural yield. While considerable uncertainty exists about yield growth and responsiveness to prices, average yields in many iLUC models are known to encompass diverse production systems. Separating out production systems in ways that distinguish between their propensity towards increasing yield vs. expanding area as a way of boosting production should better exploit existing knowledge to reduce uncertainty around this important dimension of iLUC response, and help inform policy options for reducing iLUC emissions.
References
CARB. 2009. California's Low Carbon Fuel
Standard: Final Statement of Reasons. California Air Resources Board. http://www.arb.ca.gov/regact/2009/lcfs09/lcfsfsor.pdf.
EPA. 2010. Renewable Fuel Standard Program (RFS2) Regulatory Impact Analysis. United States Environmental Protection Agency. http://www.epa.gov/otaq/fuels/renewablefuels/420410006.pdf.
Fargione, J., J. Hill, D. Tilman, S. Polasky, and P. Hawthorne. 2008. Land clearing and the biofuel carbon debt. Science 319, no. 5867: 1235.
Gibbs, H. K., M. Johnston, J. A. Foley, T. Holloway, C. Monfreda, N. Ramankutty, and D. Zaks. 2008. Carbon payback times for crop-based biofuel expansion in the tropics: the effects of changing yield and technology. Environmental Research Letters 3: 034001.
Searchinger, T., R. Heimlich, R. A. Houghton, F. Dong, A. Elobeid, J. Fabiosa, S. Tokgoz, D. Hayes, and T. H Yu. 2008. Use of US croplands for biofuels increases greenhouse gases through emissions from land-use change. Science 319, no. 5867: 1238.
[1]These could include efforts to direct production to geographical areas or production systems so as to minimize the iLUC effect (schemes that may emerge from the heterogeneity exercise), as well as broader efforts to improve overall supply efficiency (not specifically targeting the LUC caused by biofuels policy, but reducing it by boosting overall agricultural productivity) or reduce extensification pressure in carbon hotspot areas.
[2]For example, evaluation of efforts to reduce iLUC by improving overall supply efficiency or protect particular areas could be implemented in IMPACT by, respectively, selectively changing specific yield parameters or applying constraints on area response in certain regions.