How to Simulate the Ancient Sky in Planetarium Software
A free planetarium program can show the sky over Chaco Canyon or Chichén Itzá on any night a thousand years ago. Here is how to set it up, what actually changed, and where the simulation needs a second check.

Key takeaways
- Stellarium, KStars and Cartes du Ciel are free and can show the sky for any place and date across thousands of years.
- Over 1,000 years the sky changes mainly through precession (about 14°), a slightly larger axial tilt and the drift of a few fast stars.
- Sunrise directions come out very accurately; clock times of events before about 700 BC, horizon refraction and the real skyline need a second check.
To simulate the ancient sky, open a planetarium program such as Stellarium, set the observer's latitude and longitude to the site, type the historical date in astronomical year numbering and load a panorama of the real horizon. The software then recomputes precession, the tilt of Earth's axis, planetary positions and star motions for that moment. For most questions about the Sun at Chaco Canyon or in the Yucatán, the result is accurate to a small fraction of a degree.
The same engines now power museum dome shows, classroom tools and phone apps that let visitors step back to a solstice morning at a real site. Building one of those starts small: a single site, a single horizon and a single event, which is the approach of teams that help launch a minimum viable product before a full app. This guide stays with the astronomy, so you can judge any simulated sky, whatever tool shows it.
If you want the underlying math first, start with how to calculate sunrise azimuth by hand. If you would rather click than install, our Chaco viewer runs in the browser.
Can planetarium software really show the sky as it was centuries ago?
Yes. A planetarium program does not store pictures of old skies; it calculates them. Given a date, a time and a place, it works out where Earth's axis pointed, how Earth was tilted, where the Sun, Moon and planets stood, and where each catalogued star had moved by then.
Those calculations rest on the same published theories professional astronomers use:
- Planets: long-term models that are checked against NASA's JPL Horizons ephemeris service.
- Earth's axis: modern IAU models of precession and nutation.
- Stars: catalogs that include each star's measured motion across the sky.
Researchers rely on this. In 2016 a team at the University of Texas at Arlington used Starry Night and a Digistar 5 dome system to recreate the sky over Lesbos in 570 BC and date the season of Sappho's "Midnight Poem" to mid-winter and early spring, when the Pleiades set around midnight, as reported by Phys.org. Archaeoastronomers use the same kind of simulation before they visit a site, to see what an observer could have seen.
Which planetarium programs work best for ancient skies?
For most readers, Stellarium is the best place to start: it is free, open source and has tools written specifically for archaeoastronomy. The alternatives are useful for cross-checks or for dome shows.
| Program | Cost | Strength for ancient skies | Limitation |
|---|---|---|---|
| Stellarium | Free, open source | Custom horizon panoramas, ArchaeoLines plugin, selectable Delta T models, realistic sky rendering | Star positions for a few fast stars far in the past are worth cross-checking |
| Cartes du Ciel | Free | Precise star charts; recommended by Stellarium developers as a reference for star positions at ancient dates | Charts rather than a realistic horizon view |
| KStars | Free, open source | Accurate sky for any location and date, good for Linux users | Fewer archaeoastronomy tools |
| Starry Night | Commercial | Polished interface, used in published research and in classrooms | Paid licence |
| Digistar and other dome systems | Professional | Full-dome shows of ancient skies for museum audiences | Only in planetarium theaters |
| JPL Horizons | Free web service | Reference positions for the Sun, Moon and planets | Numbers, not a picture of the sky |
Features change between releases. Check each program's current manual before you rely on a specific option.
How do you set up Stellarium for an ancient date and place?
Setting up an ancient sky in Stellarium takes six steps. The order matters, because the horizon and atmosphere settings change what you see at the moment of sunrise.
- Set the location. In the Location window, enter the site's latitude, longitude and elevation. Chaco Canyon sits near 36.06° N, 107.96° W at about 1,880 m; Chichén Itzá near 20.68° N, 88.57° W.
- Enter the date in astronomical years. Stellarium counts 1 BC as year 0, so a year N BC is entered as −(N−1). The year 570 BC becomes −569, and 1000 BC becomes −999. Years CE are entered as they are.
- Mind the calendar. Before October 1582, Stellarium shows dates in the Julian calendar. A "June 21" in the Julian calendar of 1100 CE is not the solstice, which then fell around June 14 Julian. Find the solstice by watching the Sun's declination peak instead of trusting the label.
- Load a real horizon. The default landscapes are generic. Stellarium accepts custom panoramas, and the user guide explains how to stitch a 360° photo or build a horizon line from survey data. At canyon sites this single step can move sunrise by one to three degrees.
- Turn on ArchaeoLines. This plugin draws lines for the solstices, equinoxes, the Moon's major and minor standstills and other characteristic declinations, so you can see at a glance whether a feature on the horizon matches one of them.
- Check the atmosphere settings. Set temperature and pressure close to local conditions, because refraction near the horizon depends on them. Leave Delta T on the default model unless you know why to change it.
What actually changes in the sky over 1,000 years?
Over a thousand years the night sky looks almost the same to the naked eye, but several slow motions add up. The largest is precession, the 26,000-year wobble of Earth's axis.
| Effect | Size over about 1,000 years | What it changes for an observer |
|---|---|---|
| Precession | About 14° along the ecliptic (roughly 1° every 72 years) | Which stars rise with the Sun on a given date, where the celestial pole lies, star rising directions |
| Change in axial tilt | About 0.13° (23.57° around 1000 CE, 23.44° today) | Solstice sunrise and sunset directions, by about 0.15° to 0.2° at Chaco and Chichén Itzá |
| Proper motion of stars | Usually tiny; Arcturus about 0.6° | Exact star rising points for a handful of nearby bright stars |
| Calendar drift | About 7 days between the Julian and Gregorian calendars by 1100 CE | Which calendar date the solstice falls on |
| Delta T | Roughly half an hour around 1000 CE, several hours by 500 BC | The clock time and the longitude band where an eclipse was visible |
Precession matters most for star alignments. A doorway that faced the rising of a bright star in 1100 CE may face empty sky today, which is why any claim about a star alignment has to be tested at the building's own date. For the Sun, precession mostly shifts the calendar date of the solstice, not its direction, so solar alignments are far more stable than stellar ones. The site's guide to solar alignments in ancient architecture explains why builders who watched the Sun left the most testable records.
Where does a simulated ancient sky stop being reliable?
A simulated sky is reliable for directions and unreliable for exact clock times far in the past. Four limits deserve a second check.
Delta T and the clock time of events
Earth's rotation slows irregularly, so the gap between uniform astronomical time and solar clock time, called Delta T, has to be reconstructed from historical records of eclipses. Stellarium's maintainers note that no trustworthy records exist much before about 720 BC, so the exact hour of an event before then is uncertain. That uncertainty barely moves a solstice sunrise direction, but it can decide whether a particular eclipse was visible from a particular place.
Star positions for fast-moving stars
Most catalogued stars move so slowly that their positions a thousand years ago are known very well. A few nearby bright stars move faster, and small catalog errors grow over long spans. For star alignments, Stellarium's developers suggest comparing against a second program such as Cartes du Ciel.
Refraction near the horizon
Software applies a standard refraction model, about 34 arcminutes at the horizon. Real morning air, cold and layered, can bend light more than that, which shifts the visible sunrise slightly. The effect is real but rarely larger than the Sun's own width.
The shape of the real skyline
A generic flat horizon is the most common source of error. Mesas, ridges and tree lines raise the skyline, delay sunrise and push its position along the horizon. A simulation without the site's real horizon cannot confirm or reject an alignment at a canyon site.
How do researchers, museums and teachers use simulated skies?
Simulated skies are a common first step in cultural astronomy research, a teaching tool in classrooms and a regular part of planetarium programs. Typical uses fall into four groups:
- Testing alignment claims. Researchers set the site, the building's likely date and a measured horizon, then check whether a solstice, equinox or lunar standstill lines up within the survey's error. Our guide to how archaeoastronomers test an alignment walks through that logic.
- Revisiting lunar standstills. The Moon's rising point swings to its extremes every 18.6 years. At Chimney Rock in Colorado, a Chacoan outlier, the major-standstill Moon rises between two rock pinnacles, and simulations let researchers and visitors preview that event years in advance.
- Studying Venus and the Maya calendar. The Dresden Codex tracks Venus through its 584-day cycle. Simulations show how the planet's appearances as morning and evening star line up with the dates in the tables, a theme covered in Maya astronomy explained.
- Teaching and museum shows. Planetariums use dome systems to show audiences the sky over ancient sites. The Adler Planetarium in Chicago has written about modeling a far older sky for a dinosaur exhibit, and about how small errors in star motions grow over millions of years.
Simulation shows what was in the sky. It does not tell us what a place meant. That knowledge belongs to published research and to Pueblo and Maya communities who keep their own traditions about the Sun, many of which are still practiced today.
What should a first sky viewer for a site include?
A first sky viewer for a site needs one location, one accurate horizon and one event that visitors care about. Everything else can wait. Museums, parks and educators who plan their own tool usually do best with a short list:
- A trusted solar algorithm. Published algorithms already give the Sun's position far more precisely than any field measurement, so there is no reason to write a new one.
- A measured horizon. A 360° panorama or a terrain profile of the site, checked against a survey.
- A date picker that handles history. Astronomical year numbering, Julian dates before 1582 and a clear note on calendar labels.
- Honest captions. State what the simulation shows and what is interpretation, and credit the research and the communities behind each claim.
- A way to compare then and now. Showing the same morning in 1100 CE and today is often the moment that makes the astronomy click for visitors.
Our own Chaco viewer and Yucatán viewer follow that pattern: pick a date and watch the Sun meet a real horizon. For dates worth watching this year, see the sky calendar.
Frequently asked questions
Which free software shows the night sky in the past?
Stellarium, KStars and Cartes du Ciel are free and can be set to dates thousands of years ago. Stellarium is the usual choice for archaeoastronomy because it adds custom horizon panoramas and the ArchaeoLines plugin.
How do I enter a BC date in Stellarium?
Stellarium uses astronomical year numbering, in which 1 BC is year 0. Enter a year N BC as -(N-1): 570 BC becomes -569 and 1000 BC becomes -999.
How accurate is a simulated ancient sky?
For solstice and equinox sunrise directions over the last few thousand years, very accurate. The weak points are the exact clock time of events far in the past (Delta T), star positions for fast-moving stars, refraction near the horizon and the shape of the real skyline.
Did the constellations look different 1,000 years ago?
Almost the same to the eye. Precession shifted the whole sky by about 14 degrees relative to the calendar and the pole, but individual stars barely moved against each other. Arcturus, one of the fastest bright stars, drifted only about 0.6 degrees.
Can I see the sunrise at Chaco Canyon as it was in 1100 CE?
Yes. Set the location to Chaco Canyon, the date to the solstice of 1100 CE and load a horizon panorama of the site. The Sun will rise about 0.2 degrees farther north than today, which is less than its own radius.



