A major barrier to realising biofuels’ climate change mitigation potential is uncertainty concerning carbon emissions from indirect land use change (ILUC). Central to this uncertainty is the extent to which yields can respond dynamically to increased demand for agricultural commodities. This study examines the elasticity of soybean and corn yields in the USA for 1990–2017 using Bayesian network models to robustly quantify uncertainty. The central finding is that a single parameter value for yield elasticity does not adequately represent the effects of technology, policy and price pressures through time. The models demonstrate the limiting role of technological progress as well as farmers’ capital investment in response to system shocks. Results suggest evaluation of parameter uncertainty alone is unlikely to capture a full range of future ILUC scenarios and reiterate the need for ILUC studies to use probabilistic approaches as standard to robustly inform climate change mitigation policies.

实现生物燃料减缓气候变化潜力的主要障碍是间接土地利用变化（ILUC）产生的碳排放不确定性。不确定性的核心是单产对农业商品需求增长的动态响应程度。这项研究使用贝叶斯网络模型对不确定性进行了量化，研究了1990-2017年美国大豆和玉米单产的弹性。中心发现是，收益弹性的单个参数值不能充分代表技术，政策和价格压力随时间的影响。这些模型展示了技术进步以及农民对系统冲击做出的资本投资的有限作用。