Towards supply chain specific life cycle assessment: challenges and solutions

Greenhouse gas emissions and other environmental impacts of products and organizations can vary significantly according to individual supply chains. To increase supply chain specificity and thereby account for more sustainable sourcing of energy, materials, and products, life cycle assessment (LCA) is moving from average secondary life cycle inventory (LCI) data towards supply chain specific primary data. This dissertation identifies challenges and provides solutions regarding this change. It analyzes at a general level how supply chain specificity can be measured and how chain of custody (CoC) modelling can be used to model supply chain specific data. An in-depth analysis is performed regarding challenges of double counting for electricity accounting. Industry initiatives, such as Pathfinder/Partnership for Carbon Transparency, Catena-X, and Together for Sustainability, are driving the push for supply chain specific primary data to enhance product carbon footprint (PCF) specificity. These initiatives advocate for sharing primary data along the supply chain and request a primary data share (PDS) alongside the PCF to measure supply chain specificity. Whereas the definitions for PDS calculations with exclusively positive PCF contributions are consistent, PDS definitions need further refinement regarding negative emissions and system boundaries for PDS calculations. Beyond this, further specification is needed regarding multi-output processes and CoC models, which enable a separation of physical and accounted products. CoC models are designed to preserve specified characteristics, such as biogenic and recycled content, or PCF, throughout supply chains. However, these models vary significantly in design and requirements. The mass balance-credit method and the book and claim model enable a separation of physical and accounted products. Thus, they do not follow the LCA approach of describing physical systems as specified in ISO 14040. Further, the mass balance-credit method does not follow the allocation hierarchy of ISO 14044. Conversely, the market-based method for electricity accounting, which is analogous to the book and claim model, is permitted and partially required by standards like the GHG Protocol and ISO 14067. This sets a precedent for a broader application of CoC models that separate physical and accounted products. Major challenges include defining consistent allocation rules, setting and enforcing requirements, and avoiding double counting of energy, materials, and products with specific characteristics. Due to the importance of electricity-related emissions and the acceptance of the market-based method, this dissertation examines electricity accounting methods in LCA and associated double counting challenges in detail. The location-based method uses regional average emission factors, III while the market-based method relies on energy attribute certificates (EACs) to verify exclusive claims on electricity from specific sources, such as renewable energy sources (RES). Using both electricity accounting methods in parallel leads to double counting of electricity from specific sources, such as RES. Avoiding double counting demands consistent method application throughout the whole LCA, which is relatively straightforward for the location-based method but more challenging for the market-based method. LCAs typically use average secondary LCI data that include location-based electricity mixes. For consistent market-based method application, electricity-related impacts in the inventory system must also be market-based. In this dissertation, the Python-based software package Brightway 2 is used to replace European location-based electricity consumption in all 21,238 processes of the ecoinvent database with market-based residual electricity mixes. This implementation leads to average changes in life cycle impact assessment results for climate change of 1%, 10%, 20%, and 50% for 2,958, 733, 387, and 107 processes, respectively. Significant changes are observed for processes with large differences between location-based and residual electricity mix, such as in Icelandic and Norwegian processes. This database represents an approach to avoid double counting. Further research is needed to avoid regional distortions and extend the database modification beyond electricity. European energy attribute certificates, such as Guarantees of Origin (GOs) can be traded in the European residual mix area. In countries with high shares of electricity from RES, interest in GOs is lower, leading to a trade imbalance. In 2020, Iceland and Norway had GO trade deficits amounting to 74.2% and 62.4%, respectively, of their gross final electricity production from RES. Whereas GOs play a central role in market-based electricity accounting, they have no function regarding country specific RES targets. A hypothetical inclusion of GO trade in the RES share calculations results in 11 European countries missing their 2020 RES targets for the electricity sector. Their total GO trade deficit adds up to 212 TWh, equivalent to 31% of all issued GOs from RES. An alignment between the product, organizational, and country accountings could improve the validity and effectiveness for both governmental and corporate assessments and targets. The identified challenges and solutions regarding double counting of electricity from specific sources are also relevant for CoC modelling in general. Accounting for energy, materials, and products with specific characteristics on an individual level – while they are part of average secondary datasets – leads to double counting. To avoid double counting and thereby inconsistencies in environmental impact calculations, consistent rules regarding supply chain specific data usage and CoC modelling are necessary. This dissertation serves as a basis for future developments and the corresponding stakeholder processes for consensus finding.