Understanding\nUncertainty\nin Market-Mediated Responses\nto US Oilseed Biodiesel Demand: Sensitivity of ILUC Emission Estimates\nto GLOBIOM Parametric Uncertainty

The life cycle greenhouse\ngas (GHG) emissions of biofuels\ndepend\non uncertain estimates of induced land use change (ILUC) and subsequent\nemissions from carbon stock changes. Demand for oilseed-based biofuels\nis associated with particularly complex market and supply chain dynamics,\nwhich must be considered. Using the global partial equilibrium model\nGLOBIOM, this study explores the uncertainty in market-mediated impacts\nand ILUC-related emissions from increasing demand for soybean biodiesel\nin the United States in the period 2020–2050. A one-at-a-time\n(OAT) analysis and a Monte Carlo (MC) analysis are performed to assess\nthe sensitivity of modeled ILUC-GHG emissions intensities (gCO<sub>2</sub>e/MJ) to varying key economic and biophysical model parameters.\nAdditionally, the influence of the approach on the simulation of future\nILUC effects is explored using two alternative ILUC-GHG metrics: a\ncomparative-static approach for 2030 and a recursive-dynamic approach\nusing model outputs through 2050. We find that projected ILUC-GHG\nvalues largely vary based on which vegetable oils replace diverted\nsoybean oil, market responses to coproducts, and the carbon content\nof land converted for agricultural use. These are all, in turn, subject\nto decision uncertainty through the choice of the modeling approach\nand the time horizon considered for each ILUC-GHG metric. Given the\nlonger simulation period, ILUC-GHG emission uncertainty ranges increase\nunder the recursive-dynamic approach (42.4 ± 25.9 gCO<sub>2</sub>e/MJ) compared to the comparative-static approach (40.8 ± 20.5\ngCO<sub>2</sub>e/MJ). The combination of MC analysis with other techniques\nsuch as Bayesian Additive Regression Trees (BART) is powerful for\nunderstanding model behavior and clarifying the sensitivity of market\nresponses, ILUC, and associated GHG emissions to specific model parameters\nwhen simulated with global economic models. The BART reveals that\nbiophysical parameters generate more linear ILUC-GHG responses to\nchanges in assumed parameter values while changes in economic parameters\nlead to more nonlinear ILUC-GHG results as multiple effects at the\ninterplay of food, feed, and fuel uses overlap. The choice of the\nrecursive-dynamic metric allows capturing the longer-term evolution\nof ILUC while generating additional uncertainties derived from the\nbaseline definition.