Life cycle impact assessment (LCIA) transforms inventories of environmental flows to environmental impacts in life cycle assessment (LCA) studies. Many impact assessment methods have been developed for impact categories such as global warming, acidification, and ecotoxicity. These impact assessment methods provide different characterization factor values and impact units for the same impact category. LCA studies often report results for one or at most a few impact assessment methods and overlook the uncertainty caused by selecting and using different methods. In this study, we systematically evaluated uncertainties that result from considering different LCIA methods; cases from the US LCI and the ecoinvent database were used to demonstrate the uncertainties. Results showed that the discrepancies of the total impact results were caused by differences in 1) the total emission values included in the inventory; 2) the coverages of substances in methods, and 3) the differences in the characterization factor values of the substances. Case studies showed large uncertainties for all impact categories except for global warming. For most of the categories studied, the maximum values were 10,000 times larger than the minimum values. These results should be taken into consideration for impact assessments that involve categories with large uncertainties. The information provided by this paper can help researchers and software developers to understand the uncertainty caused by choosing different LCIA methods and the importance of communicating such information to LCA practitioners. The results can also help decision-makers to build guidance that is related to build and choose LCIA methods for various topics.

生命周期影响评估 (LCIA) 将环境流量清单转化为生命周期评估 (LCA) 研究中的环境影响。已经针对全球变暖、酸化和生态毒性等影响类别开发了许多影响评估方法。这些影响评估方法为同一影响类别提供了不同的特征因子值和影响单位。LCA 研究通常报告一种或至多几种影响评估方法的结果，而忽略了因选择和使用不同方法而导致的不确定性。在本研究中，我们系统地评估了考虑不同 LCIA 方法所导致的不确定性；来自美国 LCI 和ecoinvent 数据库的案例被用来证明不确定性。结果表明，总影响结果的差异是由以下方面的差异造成的：1）列入清单的总排放值；2) 物质在方法中的覆盖范围，以及 3) 物质特征因子值的差异。案例研究表明，除全球变暖外，所有影响类别都存在很大的不确定性。对于大多数研究的类别，最大值比最小值大 10,000 倍。对于涉及不确定性较大的类别的影响评估，应考虑这些结果。本文提供的信息可以帮助研究人员和软件开发人员了解选择不同 LCIA 方法所带来的不确定性以及将此类信息传达给 LCA 从业者的重要性。