AstroSat tells astrophotographers which satellites will cross a framed field of view during a planned exposure — when, where on the sensor, and whether they will actually leave a trail.
What it does
Give it your site, the patch of sky you are shooting, your optics and mount, your exposure window and sub length. It works through the catalogue of tracked objects and reports every pass that clips your field: the entry and exit times to a fraction of a second, the path across the sensor, and an estimate of how brightly each one registers. It then finds the cleanest stretch of the window - the longest run that fits the most consecutive subs untouched.
Crucially, it separates passes that are sunlit from those in Earth's shadow. Only a sunlit satellite leaves a streak; one in shadow crosses your field invisibly. The night-and-field view lets you scrub through the window and watch the passes sweep across the frame in real time.
How it works
AstroSat runs from public orbital data. Element sets - the general-perturbations catalogue plus supplemental operator ephemerides for the large constellations - are pulled from CelesTrak1. Each candidate satellite is propagated with the SGP4 model2 across your exposure window and tested against the framed field.
Two refinements keep the results honest. A small angular search margin absorbs the cross-track error inherent in public element sets, so a pass that grazes the edge of your field is not missed. And the day–night boundary uses a conical Earth-shadow model with a penumbra, so a satellite near the terminator is graded on how much sunlight still reaches it rather than being switched hard between lit and dark - the difference between a faint ghost of a trail and none at all.
Brightness, and what it can't tell you
Magnitudes use published photometry models3 for the constellations that have them (Starlink and OneWeb) and conservative standard magnitudes otherwise, each carried with an honest uncertainty - typically a few tenths of a magnitude of intrinsic scatter. What it deliberately does not do is predict specular flares: the brief mirror-glints off a satellite's flat surfaces depend on its exact orientation, which operators don't publish, so any "flare forecast" would be guesswork. Passes that can flare are flagged as such; their peak brightness is not promised.
Who makes this
AstroSat is built and maintained by Chris Beach, a software developer, electronics tinkerer, and longtime astronomy enthusiast based in Tasmania, under some of the darkest skies in the southern hemisphere. Chris is a board member at the International Space Association and the founder of Books.org. He made AstroSat free and browser-based on purpose: the forecasting runs on your own machine, and you don't need an account to use it.
Feedback
Spotted an error, or want a feature? Get in touch at hello@astrosat.org.
A forecast you share with the "Copy share link" button stores its inputs - including the observing site's coordinates - so the short link can restore them. Everything else stays on your device.
References
- Orbital element sets: CelesTrak (general-perturbations and supplemental GP), T. S. Kelso. celestrak.org
- SGP4/SDP4 propagation: satellite.js. github.com/shashwatak/satellite-js
- Constellation photometry after A. Mallama and others; phase-function fits drawn from published Starlink and OneWeb brightness studies (arXiv:2209.12060, arXiv:2012.05100).
- Deep-sky target catalogue: OpenNGC, by M. Verga, licensed CC-BY-SA-4.0. github.com/mattiaverga/OpenNGC