Quantifying Policy-Induced Cropland Dynamics: A Probabilistic and Spatial Analysis of RFS-Driven Expansion and Abandonment on Marginal Lands in the U.S. Corn Belt

Rapid biofuel expansion has significantly reshaped agricultural land use in the United States, raising concerns about the conversion and long-term sustainability of marginal croplands. Understanding how policy incentives influence these land-use changes remains a key challenge in sustainable land management. This study aims to quantify the effects of the Renewable Fuel Standard on cropland expansion and subsequent abandonment in the U.S. Midwest using a probabilistic and spatially explicit framework. The analysis integrates geospatial datasets from USDA, USGS, gridMET, and the U.S. Energy Information Administration, combining indicators of soil productivity, slope, precipitation, temperature, and market accessibility. Bayesian logistic regression models were developed to estimate pre-policy baseline probabilities of corn cultivation and to generate counterfactual scenarios—hypothetical conditions representing land-use patterns in the absence of policy incentives. Results show that over one-quarter of marginal land cultivated in 2016 would likely not have been planted without biopower policy-related incentives, indicating that policy-driven expansion extended into less suitable areas. A second-stage analysis identified regions where such lands were later abandoned, revealing the role of climatic and economic constraints in shaping long-term sustainability. These findings demonstrate the effectiveness of integrating probabilistic modelling with high-resolution spatial data to evaluate causal policy effects and quantify counterfactual impacts—that is, the measurable differences between observed and simulated land-use outcomes.