Record-breaking temperatures¹ have induced governments to implement targets for reducing future greenhouse gas (GHG) emissions^2,3. Use of oil products contributes ∼35% of global GHG emissions⁴, and the oil industry itself consumes 3–4% of global primary energy. Because oil resources are becoming increasingly heterogeneous, requiring different extraction and processing methods, GHG studies should evaluate oil sources using detailed project-specific data⁵. Unfortunately, prior oil-sector GHG analysis has largely neglected the fact that the energy intensity of producing oil can change significantly over the life of a particular oil project. Here we use decades-long time-series data from twenty-five globally significant oil fields (>1 billion barrels ultimate recovery) to model GHG emissions from oil production as a function of time. We find that volumetric oil production declines with depletion, but this depletion is accompanied by significant growth—in some cases over tenfold—in per-MJ GHG emissions. Depletion requires increased energy expenditures in drilling, oil recovery, and oil processing. Using probabilistic simulation, we derive a relationship for estimating GHG increases over time, showing an expected doubling in average emissions over 25 years. These trends have implications for long-term emissions and climate modelling, as well as for climate policy.

创纪录的温度¹促使政府实施了减少未来温室气体（GHG）排放量的目标^2,3。成品油有助于使用〜全球温室气体排放量的35％⁴，和石油行业自身消耗全球一次能源的3-4％。由于石油资源变得越来越多样化，需要不同的开采和加工方法，因此，温室气体研究应使用针对具体项目的详细数据来评估石油来源⁵。不幸的是，先前的石油部门温室气体分析在很大程度上忽略了一个事实，即在特定石油项目的整个生命周期中，生产石油的能源强度会发生显着变化。在这里，我们使用来自全球25个重要油田（最终采收率超过10亿桶）数十年的时间序列数据，对石油生产中的温室气体排放随时间的变化进行建模。我们发现，随着采油量的减少，石油产量将下降，但这种消耗伴随着每MJ温室气体排放量的显着增长（在某些情况下超过十倍）。枯竭需要增加钻井，采油和石油加工中的能源消耗。使用概率模拟，我们推导了一种估算温室气体随时间增长的关系，表明在25年内平均排放量有望翻一番。