An integrated approach to uncertainty assessment in LCA

The use of stochastic models and the presentation of ranges and confidence intervals enhances the decision support capabilities of an LCA study. However, an uncertainty analysis should not merely quantify the uncertainty in the output, but provide a mechanism to direct effort back into the LCA models to manage those uncertainties. This paper demonstrates the valuable assistance an uncertainty assessment can provide to an LCA study, including the selection of meaningful criteria against which to evaluate systems, directing further data collection and modeling effort, and systematically generating scenarios for comparison. The uncertainty assessment takes place according to a 3-layered framework, which is based on the recognition that different sources of uncertainty require different methods for their analysis and reduction.