October 4, 2019
One of the more interesting sessions at LCA XIX focused on uncertainties associated with Life Cycle Assessment (LCAs) on fossil fuels and their impact on downstream products – fuels and petrochemical-based materials like plastics.
Although analysts have traditionally used a single number for fossil fuel GHG emissions, it’s clear that no two quantities of crude oil or diesel will have exactly the same GHG profile, for example. Crude oil has many different properties, and the methods used to extract and refine it can vary significantly. Therefore, our LCA results inevitably have great variability due to the data (or lack thereof) used to model the GHG emissions. Our certainty is only as strong as our least-certain link.
How can we improve the situation and reduce these uncertainties for fossil fuels? The obvious first step is to acquire better data about crude oil characteristics and extractions, petroleum movements, refineries, and final product slates.
Our challenge, as LCA practitioners, is to be able to communicate both uncertain results and our need for better data in flow, model and parameter uncertainties in the tools we use. We need to recognize what level of certainty is required for our study, understand the uncertainty, standardize the uncertainty and carefully communicate this uncertainty to decision-makers in a way that provides them with trustworthy information for important decisions.
There are many possible next steps, both near- and long-term. Enthusiastic discussion at the session focused on a couple of things, including the need for an agreed-upon glossary of terms. Some consensus best-practice guidance would be a big help, too: Do we use a scale, put a quantitative value on uncertainty? A checklist? Standardize uncertainty factors for certain products (eg. 5% for fossil fuels and 10% for natural gas, and 50% for ethylene glycol)? Use a contribution analysis?
More work needs to be done, but the conversation at LCA XIX was a good start.
About the Authors:


PHOTO: Creative Commons: "Image" by jmrgmbl is licensed under CC PDM 1.0