Counterfactual safety modeling paper published in the proceedings of AIAA Aviation!

06 Jun 2026

Counterfactual Safety Modeling: Estimating the Impact of ATC Staffing and Scheduling Practices on Risky Events has been published in the conference proceedings of AIAA Aviation! A preprint of the paper is available here for free, and the official version is available on the AIAA website. I will be presenting the work on June 8th, and I hope to include the slides in my next post.

Here is the abstract:

Air traffic control (ATC) staffing-related factors, such as staffing levels, supervisor presence and controller fatigue, have been cited as contributing factors to recent aviation incidents and accidents. These events have brought into focus the need to develop new tools that can quantitatively assess the impacts of ATC staffing and scheduling practices on rates of risky encounters in the National Airspace System (NAS). This work addresses the question of how staffing impacts risk using a Bayesian causal modeling approach. As a proof of concept, we examine whether the presence of a dedicated supervisor in an ATC tower reduces the likelihood of runway incursions (RIs). Informed by operational, staffing and incident data, the model can identify operational conditions that have a higher likelihood of RIs, and make prescriptions for when and where supervisor presence can be most effective in reducing risk. These kinds of models can help the Federal Aviation Administration prioritize the most effective ATC technologies and staffing changes to increase safety and efficiency in the NAS.

Berk