Democratizing Aviation Emissions Estimation: Development of an Open-Source, Data-Driven Methodology

Andy Eskenazi, Landon Butler, Arnav P. Joshi, Megan Smirti RyersonarXiv 2022 · arXiv

Introduction

Accurately estimating aviation emissions is a critical yet challenging task. Existing models often oscillate between broad, "top-down" heuristics and highly accurate but proprietary "bottom-up" software. This lack of transparency prevents community advocates and urban planners from independently verifying airline sustainability claims. Researchers from the University of Pennsylvania have developed a new, open-source methodology that leverages public datasets to provide granular, flight-level emissions estimates for the U.S. domestic aviation system.

Methodology and Data Integration

The methodology centers on a multi-step pipeline that matches individual aircraft to their specific engine types. By utilizing Jaccard Similarity to link FAA tail numbers with the ICAO Aircraft Engine Emissions Databank, the researchers identified engine models for nearly the entire U.S. commercial fleet.

The data integration pipeline linking BTS flight schedules, FAA registries, and ICAO/BADA emission databases.

The model segments each flight into two distinct phases:

  1. Landing and Take-off (LTO): Activities below 3,000 feet, where engine thrust is standardized by ICAO.
  2. Climb, Cruise, and Descent (CCD): Activities above 3,000 feet, where emissions vary by altitude and distance.

The tool tracks four key pollutants—CO2, NOx, CO, and HC—and converts them into CO2 equivalents (CO2e) to provide a comprehensive view of total climate impact.

Key Insights: The NOx Surprise and Short-Haul Penalty

A primary finding of the study is the "NOx Surprise": when converted to CO2e, Nitrogen Oxides often have a greater climate impact than CO2 itself across both LTO and CCD cycles. This suggests that carbon-only metrics significantly underestimate the industry's environmental footprint.

Additionally, the research highlights a "short-haul penalty." Because the fuel-intensive LTO cycle represents a larger portion of total flight time for shorter journeys, these flights have a much higher emissions-per-distance ratio compared to long-haul travel.

CO2e emissions per capacity-mile vs. flight distance, demonstrating the increased environmental intensity of short-haul flights compared to standard estimates.

Results and Economic Impact

Analysis of 1.6 million flights from Q3 2021 revealed that low-cost carriers like Frontier (0.097 kg CO2/seat-mile) are significantly more efficient than regional carriers like Endeavor (0.243 kg CO2/seat-mile), primarily due to fleet age and seating density. Geographically, major hubs such as Atlanta (ATL), Dallas (DFW), and Los Angeles (LAX) were identified as the highest sources of LTO emissions. The researchers estimate the associated local health and environmental costs at these top airports range from $50 million to $200 million USD per year.

Conclusion

By open-sourcing the code and relying on public data, this methodology moves aviation emissions modeling into the public domain. While future work is needed to incorporate real-time weather and high-altitude radiative forcing, this tool provides a transparent foundation for holding the aviation industry accountable and designing more equitable air transit systems.

Summary updated

BibTeX
@misc{eskenazi2022democratizing,
  author = {Andy Eskenazi and Landon Butler and Arnav P. Joshi and Megan Smirti Ryerson},
  title = {{Democratizing Aviation Emissions Estimation: Development of an Open-Source, Data-Driven Methodology}},
  year = {2022},
  eprint = {2202.11208},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2202.11208},
}
Figure