|Title||Robustness of life cycle assessment results : influence of data variation and modelling choices on results for beverage packaging materials|
|Author(s)||Harst-Wintraecken, E.J.M. van der|
|Source||Wageningen University. Promotor(en): Carolien Kroeze, co-promotor(en): Jose Potting. - Wageningen : Wageningen University - ISBN 9789462575097 - 217|
Environmental Systems Analysis
|Publication type||Dissertation, internally prepared|
|Keyword(s)||levenscyclusanalyse - onzekerheid - modelleren - gegevensanalyse - gegevens verzamelen - afvalverwerking - recycling - milieueffect - life cycle assessment - uncertainty - modeling - data analysis - data collection - waste treatment - recycling - environmental impact|
|Categories||Environmental Management (General)|
Life cycle assessment (LCA) is a well-established method to evaluate the potential environmental impacts of product and service systems throughout their life cycles. However, it can happen that LCAs for the same product have different and even conflicting outcomes. LCA results need to be robust and trustworthy if they are used in decision making. The aim of this thesis is to evaluate whether the use of multiple data sets and multiple modelling options can increase the robustness of LCA results.
The research starts with identifying reasons for differences in LCA results for the same product. The results of ten existing LCAs for disposable beverage cups are compared to each other as to examine the consistency and robustness of these results. The comparison of the LCAs shows no consistent best or worst cup material. And, the quantitative results for cups made from the same material vary across the LCAs. The evaluation of the methodological choices and the used data sources in each LCA made it possible to identify possible sources for discrepancies in the LCA results. Reasons for differences in results include the variation in the properties of the cups, production processes, waste treatment options, allocation options, choices in system boundaries, impact indicators, and potentially also the data sets that are used.
The thesis next describes a novel method to evaluate and include the influence of data sets and modelling choices on the LCA results. The method is applied in a case study of a disposable polystyrene (PS) beverage cup. The study purposely uses different data sets from various sources for processes with an influential contribution to the LCA results. The study includes two waste treatment options (incineration and recycling). The multiple data sets represent the variability among processes, and the waste treatments represent choices in the modelling of the life cycle of the PS cup.
This variability among the data sets for a similar process is presented as a spread in the results. This spread in the results for the PS cup is caused by differences in the amount and type of the used resources and energy, reported emissions, the origin of the production location, the time period of data collection, or choices in the value of recycled PS. The overlapping spread in the quantitative results for incineration and recycling prevents a decisive conclusion on the preferred waste treatment option for the PS cups.
Next, the method for the use of multiple data sets and modelling choices is applied in a comparative LCA of disposable beverage cups. Three cups are compared: a PS cup, a polylactic acid cup (PLA, a biobased plastic), and a cup made from biopaper (paper with a lining of biobased-plastic). The waste treatment options consist of incineration and recycling for all three cups, and additionally composting and anaerobic digestion for the PLA and biopaper cup.
The use of multiple data sets and modelling choices leads to a considerable spread in the LCA results of the cups. The results do not point to the most environmentally friendly cup material, and neither to a preferred waste treatment option. The results clearly identify composting, however, as the least preferred waste treatment option for the PLA and biopaper cups. The spread in the results makes the comparison of the results for the cups more complex, but the results provides more robust information for decision makers. The combined inclusion of the variability among data sets and the waste treatment options makes the results more trustworthy.
The thesis then dives deeper into the methodological modelling of recycling in LCA and describes and evaluates six widely used recycling modelling methods: three substitution methods, an allocation method, the recycled-content method, and the equal-share method. The main difference among the six methods lies in the assumption on where and how to apply credits for recycled material in the life cycle of the product.
These six methods are applied in two case studies: a disposable PS beverage cup and an aluminium beverage can. The results for the aluminium can clearly depend on the applied recycling modelling method, the recycling rate of the disposed cans, and the amount of recycled material used in the cans. The results for the PS cup additionally depend on the consideration of a drop in the quality of the recycled PS compared to the original PS, and the other waste treatments (landfilling and incineration) for the cups. Including several recycling modelling methods in the LCA incorporates the various underlying modelling philosophies of the methods, and thus makes the results more robust.
This thesis demonstrates the added value of including multiple data sets and multiple modelling choices in LCA. The use of multiple data sets is especially useful if general processes instead of specific processes are used in the representation of the product system. The use of multiple data sets increases the accuracy of the results, and is a supplemental tool next to statistical methods which increase the precision of the results. The simultaneous handling of variability among data sets and modelling choices is hardly performed in LCA. The method presented in this thesis fills this gap and provides a transparent tool to capture these uncertainties. The trade-off between an increase in the robustness of the results and the additional demand for resources (time, money, effort) should be assessed, and depends on the goal of the study and on the intended use of the results. This thesis shows that inclusion of the uncertainty in the LCA results provides the decision maker with valuable information. This thesis thus provides a useful method to increase the robustness of LCA results.