Sensitivity of life-cycle assessment outcomes to parameter uncertainty: Implications for material procurement decision-making

This paper aims to apply an analytical approach to propagate parameter uncertainty through Life-Cycle Assessment (LCA) outcomes, and identify equivalence intervals that are applicable during material procurement decision-making. LCA outputs are usually presented as point estimates without accompanying confidence intervals, or associated margins of error, due to methodological ambiguity in accounting for uncertainties, be they epistemic (due to inherent randomness) or aleatory (due to lack of knowledge). This reduces the reliability of using LCA for design selection, and of decision-making during material procurement. This paper aims to fill this gap using a methodology based on Taylor’s first-order approximation to examine LCA outcomes for asphalt mixtures. The method is applied to assess the impact of variation in input flows on the estimated output emissions by propagating the parameter uncertainties through the LCA. It allows for the reporting of LCA outputs using intervals rather than point estimates. For competing designs, the overlap in intervals is used to construct equivalence intervals within which the effective impact of the designs can be considered equivalent due to expected variability in input parameters. The method is applied to construct equivalence intervals for the Global Warming Potential (GWP) of three asphalt mixtures, each having different levels of Reclaimed Asphalt Pavement (RAP). Their role in communicating LCA outcomes for decision-making is illustrated. Foreground data collected from 40 North American asphalt plants were used to estimate uncertainty in the input parameters of electricity and energy use in asphalt production. The paper adds to existing literature on uncertainties in pavement LCAs by establishing the need to incorporate uncertainty when using LCA to support decision-making during material procurement. Life-Cycle Assessment (LCA) of products and processes typically communicate environmental impacts using point estimates of mid-point indicators (1). However, these values for environmental impacts implicitly include both aleatory (due to inherent randomness) and epistemic (due to lack of knowledge) uncertainties, even though they are not explicitly considered or communicated. This creates an ambiguity in deciding the relative ranking of different alternatives. Additionally, this lack of reliability restricts the widespread application of LCA outcomes in decision-making. There is limited understanding of ways to define margins of error in LCA estimates when developing estimates for specific pavement design alternatives. This is a critical gap when LCA is used in making comparisons between design alternatives and may eventually impact the procurement of construction materials. Even under the fairest of conditions knowing the significance of a difference between two options can be difficult. For instance, if the acceptable error margins for benchmarking GWP for two options of equivalent functionality are +/–X% and +/–Y% respectively, then for some overlap of the ranges that is a function of X% and Y%, the options should be considered to have an effectively equivalent impact. This is currently not part of the practice and is a deterrence to the meaningful use of LCA in decision-making. This paper presents, as a first step, a method to develop confidence intervals for reported mid-point indicators