CO2 emissions reduction by switching conference online: Uncertainty analysis of global air travel

Yinan Li, Xiaonan Wang, Jinyue Yan, Mingkun Jiang, Dan Fang, Jin Yang, Haoran Zhang2021 · DOI

Abstract

Ideally, primary data collection is recommended for every life cycle assessment (LCA) study. However, due to limited availability or accessibility to first-hand data, related sources of secondary data can be a good alternative in practice. In this work, the uncertainty of using secondary data from the Ecoinvent Life Cycle Inventory (LCI) database is illustrated with an LCA case study on global air travel. Inside the database, both parameters' basic uncertainty from measurements and additional uncertainty from data quality criteria are considered with the pedigree approach. The effect of updated pedigree matrix coefficients is also evaluated. Furthermore, the sensitivity with respect to the choice of system boundary is studied with a hotspot analysis for air travel. Outside the database, the uncertainty associated with mapping real world processes to those available in the database is analyzed. In particular, the influence of flight specific parameters, e.g. plane type and occupancy level, is assessed by comparing the International Civil Aviation Organizations (ICAO) carbon emissions calculator with database calculations. The results show that emissions calculated by ICAO generally lie on the lower end of confidence intervals provided by uncertainty analysis of the database, especially for very long-haul flights. Finally, for the LCA case study on air travel, a two-step method combining the advantages of both the ICAO calculator and the Ecoinvent database is proposed.

Introduction

The shift toward virtual conferences during the COVID-19 pandemic highlighted the potential for significant carbon savings. However, quantifying these savings accurately requires Life Cycle Assessments (LCA) that often rely on secondary data. While primary data—direct measurements from specific activities—is ideal, researchers frequently depend on databases like Ecoinvent due to the proprietary nature of airline fuel logs. This paper investigates the uncertainty inherent in these secondary datasets and how it impacts our understanding of the carbon footprint of global air travel.

Methodology

The study employed a matrix-based LCA approach using the Ecoinvent v3.7 "cut-off" system model. To quantify uncertainty, the authors used the pedigree approach, which combines basic measurement uncertainty with factors reflecting data quality (reliability, completeness, etc.). Monte Carlo simulations were conducted to determine the distribution of potential outcomes.

Crucially, the researchers performed a hotspot analysis to trace emissions through the supply chain and benchmarked Ecoinvent’s results against the International Civil Aviation Organization (ICAO) carbon emissions calculator. The ICAO tool provides a real-world benchmark by incorporating flight-specific parameters like aircraft type and occupancy levels, which are often generalized in broader LCI databases.

Results and Discussion

Uncertainty and Emission Factors

The analysis revealed that Ecoinvent tends to provide higher emission estimates than the ICAO calculator. This discrepancy arises because Ecoinvent uses generalized fuel consumption rates (approx. 0.0267 kg/personkm for long-haul) that are significantly higher than the ICAO’s specific estimates (approx. 0.0136 kg/personkm).

CO2 intensity and uncertainty by flight distance

As shown in Figure 1, the uncertainty ranges provided by the database are substantial. Interestingly, deterministic values often fall below the medians of Monte Carlo simulations due to the skewed nature of lognormal distributions, suggesting that standard "average" reports might actually underestimate the likely carbon impact if the underlying distribution is not considered.

Supply Chain Hotspots

A recursive calculation traced the majority of air travel's carbon impact to direct emissions during flight, which account for 86.4% of the total. Upstream fuel refining is the next largest contributor. In contrast, the "infrastructure" components—including aircraft manufacturing and airport construction—account for less than 2% combined.

Hotspot analysis of air travel emissions

Figure 2 visualizes this concentration, where the large central node represents direct emissions, dwarfing the contributions from airport and aircraft infrastructure.

Conclusion

The study identifies a clear need for higher specificity in aviation LCAs. The authors propose a two-step hybrid method: use the ICAO calculator to obtain accurate, flight-specific fuel consumption data, and then apply Ecoinvent’s LCI data to calculate the emissions associated with that fuel. This approach leverages the strengths of both tools—flight-specific accuracy and comprehensive supply-chain coverage—to provide a more robust assessment of air travel's environmental impact. Future work should focus on refining pedigree uncertainty factors to ensure confidence intervals remain practically useful for decision-making.

Summary updated

Abstract

Ideally, primary data collection is recommended for every life cycle assessment (LCA) study. However, due to limited availability or accessibility to first-hand data, related sources of secondary data can be a good alternative in practice. In this work, the uncertainty of using secondary data from the Ecoinvent Life Cycle Inventory (LCI) database is illustrated with an LCA case study on global air travel. Inside the database, both parameters' basic uncertainty from measurements and additional uncertainty from data quality criteria are considered with the pedigree approach. The effect of updated pedigree matrix coefficients is also evaluated. Furthermore, the sensitivity with respect to the choice of system boundary is studied with a hotspot analysis for air travel. Outside the database, the uncertainty associated with mapping real world processes to those available in the database is analyzed. In particular, the influence of flight specific parameters, e.g. plane type and occupancy level, is assessed by comparing the International Civil Aviation Organizations (ICAO) carbon emissions calculator with database calculations. The results show that emissions calculated by ICAO generally lie on the lower end of confidence intervals provided by uncertainty analysis of the database, especially for very long-haul flights. Finally, for the LCA case study on air travel, a two-step method combining the advantages of both the ICAO calculator and the Ecoinvent database is proposed.
BibTeX
@misc{li2021co,
  author = {Yinan Li and Xiaonan Wang and Jinyue Yan and Mingkun Jiang and Dan Fang and Jin Yang and Haoran Zhang},
  title = {{CO2 emissions reduction by switching conference online: Uncertainty analysis of global air travel}},
  year = {2021},
  doi = {10.46855/energy-proceedings-7227},
}
Figure