Influence of emission-factor uncertainty and farm-characteristic variability in LCA estimates of environmental impacts of French dairy farms

Abstract

Life Cycle Assessment (LCA) is a useful framework for environmental assessment; however, the reliability of LCA results suffers from many sources of uncertainty and variability. Now that systematic uncertainty analysis in LCA is recommended, it can be useful to revisit past LCA studies to see whether inclusion of uncertainty (or additional types of it) changes interpretation of their results. In this study, we added uncertainty in 67 emission factors (EFs) to the variability in farm characteristics of 47 French dairy farms analyzed in a previous LCA study (van der Werf et al., 2009). We propagated uncertainty in EFs with Monte-Carlo simulation to estimate contributions of uncertainty and variability to uncertainty in potential climate change, acidification, and eutrophication impacts. For individual farms, uncertainty in emission factors added uncertainty to the farm's formerly deterministic impacts (coefficients of variation of 2–7% for climate change, 4–11% for acidification, and 2–46% for eutrophication). By farm type (conventional vs. organic), the addition of uncertainty in EFs increased uncertainty in impacts. Although uncertainty in emission factors contributed less to impact uncertainty than variability in farm characteristics did, it did add enough to potentially change decisions about whether differences in certain impacts between farm types were significant, depending upon the significance level and functional unit chosen. Variance-based sensitivity analysis identified emission source categories whose uncertainty contributed most to the uncertainty in impacts: manure deposited in pasture for climate change, cattle housing and manure storage for acidification, and leachate for eutrophication. Although larger uncertainties in potential impacts decrease apparent differences between the systems or scenarios studied, considering more than one type of uncertainty provides decision makers with a more complete and realistic assessment of the state of knowledge. Based on the degree of uncertainty in impacts, they can decide which location on impact intervals (e.g., mean, lower limit, upper limit) is best suited for decisions in a given system. Future studies should explore additional methods to combine multiple sources of uncertainty in LCA and express their relative influences on potential impacts.