Apr 23, 2026 04:29 AM
John Kraus is a longtime and widely respected rocket photographer. (They are a special breed.)
Currently John holds a NASA position created especially for him: Special Communications Assistant to the Administrator of NASA.
And he's been working on this own time creating the Space Jellyfish Predictor.
A Space Jellyfish is the often spectacular sky-show created when a rocket launches right after sunset or right before dawn. As the rocket ascends it leaves the Earth's shadow (we call it 'night') and passes into bright sunlight while the Earth surface below is still in darkness. This creates crazy illumination of the rocket's exhaust plume.
https://jellyfish.johnkrausphotos.com/
John describes the Space Jellyfish Predictor this way -
https://x.com/johnkrausphotos/status/202...3389072590
Currently John holds a NASA position created especially for him: Special Communications Assistant to the Administrator of NASA.
And he's been working on this own time creating the Space Jellyfish Predictor.
A Space Jellyfish is the often spectacular sky-show created when a rocket launches right after sunset or right before dawn. As the rocket ascends it leaves the Earth's shadow (we call it 'night') and passes into bright sunlight while the Earth surface below is still in darkness. This creates crazy illumination of the rocket's exhaust plume.
https://jellyfish.johnkrausphotos.com/
John describes the Space Jellyfish Predictor this way -
https://x.com/johnkrausphotos/status/202...3389072590
Quote:Pleased to share the v1 public beta of the Space Jellyfish Predictor, which outputs text predictions and heatmaps for upcoming rocket launches and their potential to produce a “space jellyfish.” This effect occurs when a launch is illuminated by the sun while an observer is in relative local darkness.
You can try it out here: https://jellyfish.johnkrausphotos.com — please report any issues and suggest features in this thread. I am also eager to see real-world photos from upcoming missions at distances close and far to help refine the models.
Especially during this early v1 public beta period, please treat this site as potentially helpful guidance, not a guarantee. In the event of issues, I may pull it down at any time.
How it works:
The model evaluates the geometry between the observer, rocket, and sun to estimate whether a launch plume is likely to be sunlit, how high it may appear above the horizon, and how strongly it may contrast against the sky. It identifies the strongest post-liftoff moment and converts it into a likelihood/prominence label, and does so in increments for two hours on either side of the current T-0 so you have insight into the evolving forecast.
This tool is focused on immediate post-launch ascent visibility and does not model delayed illumination from residual plumes. Nighttime launches can still be visible from great distances, and only sunlit ascents will produce relevant forecasts. As always, real-world conditions (especially weather and atmospheric quality) can affect what is actually visible.
The default location for each launch prediction is the launchpad, which is generally sufficient for nearby viewers. For distant observers, you can change your location to generate a location-specific text output for a given mission.
The site currently supports Starlink missions with strong fidelity across groups, ISS crew/resupply launches, and GTO/similar missions. Other Falcon 9 missions may use a generalized fallback ascent profile once their launch azimuth (direction) is determined, but these predictions will likely be a bit less reliable.
It was fun to work on this outside of the day job. I hope it is useful. If there is strong interest, I will evaluate the feasibility of adding additional vehicle types and launch sites in the weeks ahead.
![[Image: HGj37z8a0AETnVS?format=jpg&name=large]](https://pbs.twimg.com/media/HGj37z8a0AETnVS?format=jpg&name=large)