Announcing StarFall: Near Real‑Time Bolide Detection and Alert System

StarFall is now publicly released! This open‑source project, developed by the Space Dynamics Laboratory with support from NASA’s Planetary Defense Coordination Office, provides a near real‑time bolide detection and alert system.

What is StarFall?

StarFall ingests Geostationary Lightning Mapper (GLM) L2 netCDF data from NOAA’s GOES satellites. Although the GLM sensors were originally built to detect lightning, they also capture the distinctive signatures of bolides.

Key Features

• Continuous near‑real‑time alerts – The GLM Trigger Generator (GLM TG) continuously pulls GLM data from NOAA’s G‑Cloud, creating real‑time alerts for bolide events.
• AI/ML‑driven false‑positive reduction – Integrated with the ROCKET model (McKinney et al. 2024/2025; Dempster et al. 2020), the system dramatically cuts daily false‑positive alerts to a manageable level.
• Comprehensive parameter estimates – For each trigger, GLM TG computes:
  • UTC time of the event
  • Location of peak intensity
  • Total radiated energy (using Lockheed Martin continuum calibration tables)
• Stereo‑event velocity estimates – When sufficient GLM data are available for stereo observations, the system provides rough velocity estimates, including speed and directional components.
• Web‑based UI – An interactive, web‑based interface lets users explore the raw GLM data, view the trigger formation process, and visualize the derived parameters.

Why it matters

Bolide detection is a critical component of planetary defense. By leveraging existing GLM data streams and advanced AI/ML techniques, StarFall offers a lightweight solution that can be deployed continuously with few false‑positive alerts.

If you are interested in trying out StarFall or contributing to its development, visit the repository on GitHub:

https://github.com/Space-Dynamics-Laboratory/StarFall

Feel free to open an issue, submit a pull request, or leave a comment below if you have questions or suggestions. Happy hunting!