Uncertainty Implications of Hybrid Approach in LCA: Precision versus Accuracy

The hybrid approach in Life Cycle Assessment (LCA) that uses both input-output and process data has been discussed in the context of mitigating truncation error and burdens of data collection. However, the implication of introducing input-output data on the overall uncertainty of an LCA result has been debated. In this study, we selected an existing process LCA, performed a Monte Carlo simulation after hybridizing each truncated flow at a time, and analyzed the dispersion and position of the distribution in the results. The results showed that hybridization effectively moved the mean of the life cycle greenhouse gas (GHG) emissions 38% higher while maintaining the standard deviation within the 0.62–0.78 range (relative standard deviation, 3–4%). We identified key activities contributing to the overall uncertainty and simulated the potential effect of collecting higher quality supplier-specific data for those activities on the overall uncertainty. The results showed that replacing as few as 10 of the largest uncertainty contributors with high precision supplier-specific data substantially narrowed the distribution. Our results suggest that hybridizing truncated inputs improves accuracy of LCA results without compromising their precision, and prioritizing supplier-specific data collection can further enhance precision in a cost-effective manner.生命周期评估（Life Cycle Assessment, LCA）中同时采用投入产出（input-output）数据与过程（process）数据的混合方法，此前已被用于缓解截断误差（truncation error）与数据收集负担的相关讨论中。然而，引入投入产出数据对LCA结果整体不确定性的影响仍存在争议。本研究选取了一项已有的过程型LCA，针对每一处截断流（truncated flow）依次开展混合操作后进行蒙特卡洛模拟（Monte Carlo simulation），并分析了结果中分布的离散程度与分布位置。结果显示，混合方法可使生命周期温室气体（Greenhouse Gas, GHG）排放的均值提升38%，同时将标准差维持在0.62~0.78区间内（相对标准差为3%~4%）。本研究识别了对整体不确定性具有显著贡献的关键活动，并模拟了为这些活动收集更高质量供应商特定（supplier-specific）数据对整体不确定性的潜在影响。结果表明，仅需将前10个最大的不确定性贡献源替换为高精度供应商特定数据，即可大幅收窄结果分布。本研究结果提示，对截断输入进行混合可在不损失精度的前提下提升LCA结果的准确性，而优先收集供应商特定数据还能以更具成本效益的方式进一步提升精度。