Uncertainty Implications of Hybrid Approach in LCA: Precision versus Accuracy

2019-03-07T00:00:00Z (GMT) by Jessica Perkins Sangwon Suh
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.