Prospective dynamic average and marginal electricity emission factors for Germany until 2070: a methodological extension incorporating energy storage

Abstract

Background

Life Cycle Assessment (LCA) of electricity consumption requires temporally resolved emission factors. Two established metrics are the Average Emission Factor (AEF), allocating total system impacts to consumed electricity, and the Marginal Emission Factor (MEF), reflecting the response of marginal generation. Previous studies in regard to Germany (e.g. Seckinger & Radgen) provided hourly AEFs but focused mainly on the Global Warming Potential (GWP) and represented storage using constant annual discharge factors, neglecting round-trip losses and construction-related (upstream) emissions. This study addresses these limitations by proposing an extended electricity impact model that treats storage as an impact storage carrying time-varying charging impacts and allocates storage construction impacts to discharge cycles.

Methods

The reference model by Seckinger et al. and the new approach are applied to a consistent hourly dataset for Germany covering 2020–2070. Hourly AEFs and MEFs are calculated for GWP, AP, EP\(_{fw}\), ODP, POCP, ADPF, and ADPE. The new model accounts for storage round-trip efficiencies and consumption-based accounting of imports, exports, and charging. It also allocates construction impacts to discharge events using a cycle-based approach linked to the state-of-charge distribution. The results are compared with those achieved by existing studies, and a sensitivity analysis evaluates key assumptions on storage construction and power plant characterization factors.

Results

The new model yielded higher hourly emission factors than that from Seckinger & Radgen, particularly during periods of high storage activity. Differences increased over time with rising storage deployment and are primarily driven by construction-related impacts, while hourly attribution introduces additional temporal variability. Across most impact categories, the refined storage representation leaded to higher average and marginal emission factors in storage-rich scenarios. In contrast to GWP and several other categories, ADPE increased over time, indicating a shift in environmental burdens toward non-fossil resource use as renewable generation and storage expand.

Conclusions

Energy storage representation strongly influences time-dependent electricity emission factors, especially in future systems with high storage penetration. Although differences between modeling approaches are small in early years, explicitly accounting for time-varying charging impacts and storage construction emissions improves accuracy and avoids systematic underestimation of future electricity footprints. Extending dynamic LCA beyond operational emissions is therefore essential to capture emerging trade-offs in highly renewable electricity systems.