| Life cycle assessment (LCA), defined as a systematic approach to evaluate environmental burdens associated with a product, process or activity covering its whole life cycle (i.e. from raw material acquisition to disposal) has been increasingly used by firms and government agencies to try to estimate the environmental impact of packaging options. Among the many LCA studies published, a great number of them are comparative studies to identify the environmentally better option, but whose results end up often being challenged by the losing parties.; This dissertation elaborates that conflictive results are a sign of LCA limits, which in turn can be traced back to the uncertainties inherent to the LCA approach. In order to fully represent these ideas, this work offers a two-step approach. First a critical overview of the state of the LCA method, its disparate applications and impact on topics that are part of the packaging field is offered. Second, the effect of some types of uncertainty (i.e. inventory data and scenario) in the outcome of a packaging based LCA featuring three different packaging materials is analyzed for a hypothetical drink product using published average process values. The three packaging systems are based on PET bottles, PLA bottles and an aluminum cans, along with their whole set of distribution packaging and transportation services and end of life alternatives which were evaluated with regards to four environmental burdens: energy use, water use, global warming potential (GWP) and ozone depletion potential (ODP).; The stochastic simulation for the simultaneous assessment of uncertainty due to LCI data, through Monte Carlo simulation, and scenarios relevant to packaging, through a non-parametric procedure, has demonstrated that, in this particular comparison. LCI parameter uncertainty seems to have a dominant effect in the outcome of the LCA results considered. Furthermore, by including the aforementioned noise, or considering temporal or spatial information or allocation procedures, the overlapping of uncertainty intervals (at 95% confidence) indicates that, despite the differences in the primary raw materials, the systems under study can be similar to one another, at least with regard to the environmental indicators of this work.; The dissertation also demonstrates that when the knowledge of the systems under study is limited to average values of the different operations, or to conservative uncertainty estimations, a clear option is unlikely to surface. On the other hand, the dissertation shows that when more information is known about how, where and when the system operates (i.e. location and time of operations, type of technology, raw material duality) more conclusive results can he obtained within an impact indicator, but it also shows that this gain cannot be fully exploited unless a priority-based approach across indicators is used for performing the impact assessment study. |
