Planetary Frames and Mission Frame Hand-offs
Planetary frames, the Moon’s two meridians, rotating frames, and the frame hand-offs of a real mission
Ocean of Storms, 19 November 1969
Four months earlier, Apollo 11 had landed about six kilometres beyond its target. Nobody minded: Armstrong was flying over a boulder field looking for anywhere flat, and the goal was to land at all.
Apollo 12 was given a harder job. Two and a half years before, an uncrewed probe called Surveyor 3 had landed in the Ocean of Storms. Pete Conrad and Alan Bean were to set down close enough to walk over to it.
That required knowing where Surveyor 3 was in a lunar coordinate system, knowing where the lunar module was in the same system, and knowing how that system was oriented in space at the moment of descent — on a world whose gravity field was lumpy in ways nobody had fully mapped. The navigators refined the landing target in flight using tracking from Earth.
Intrepid touched down about 160 metres from Surveyor 3. Conrad and Bean walked over and cut pieces off it to bring home. It was the first pinpoint landing on another world, and it was, at bottom, a frame problem solved correctly.
Watch first · presented by Shoaib Aziz
Whoopee! Man, that may have been a small one for Neil, but that's a long one for me.
Every major body in the solar system has a body-fixed frame defined by the IAU Working Group on Cartographic Coordinates and Rotational Elements (WGCCRE), which updates the definitions every few years. The recipe is the same for all of them, and it is just three angles:
α₀, δ₀ — the right ascension and declination of the body's north pole, in the ICRF. They change slowly (the body's own precession).
W — the angle of the prime meridian, measured along the body's equator from the point where that equator crosses the ICRF equator. It increases by the rotation rate every day.
From ICRF to body-fixed is then three elementary rotations — the same machinery as Earth, with fewer moving parts:
R₃(90° + α₀) swings the node of the body’s equator onto x; R₁(90° − δ₀) tips the ICRF pole onto the body’s pole; R₃(W) spins to the prime meridian. d is days of TDB since J2000.
| Body | Pole α₀, δ₀ (ICRF) | Ẇ (°/day) | Sidereal rotation | Prime meridian defined by |
|---|---|---|---|---|
| Sun | 286.13°, 63.87° | 14.1844 | 25.38 d | Carrington’s 1853 convention |
| Mercury | 281.01°, 61.42° | 6.1385 | 58.65 d | Crater Hun Kal placed at 20° W |
| Venus | 272.76°, 67.16° | −1.4814 | 243.0 d, retrograde | Central peak of crater Ariadne |
| Moon | 269.99°, 66.54° | 13.1764 | 27.32 d | Mean direction to Earth (Mean-Earth frame) |
| Mars | 317.27°, 54.43° | 350.8920 | 24 h 37 m 22.7 s | Airy-0 crater, now pinned to the Viking 1 lander at 47.95137° W |
| Jupiter | 268.06°, 64.50° | 870.536 | 9 h 55 m 30 s | System III: the magnetic field’s radio period |
| Uranus | 257.31°, −15.18° | −501.16 | 17 h 14 m, retrograde | Convention; poorly constrained |
💡Which pole is north?
For the planets and their major moons, the IAU defines the north pole as the one on the north side of the solar system's invariable plane — roughly, the same side as Earth's north. W then increases if the body spins counter-clockwise seen from that pole, and decreases if it spins the other way.
That is why Venus and Uranus have negative W rates. Venus turns slowly backwards. Uranus is tipped 98° and its IAU north pole sits at declination −15°: measured from that pole, it rotates retrograde.
For dwarf planets, asteroids and comets — many of which tumble or have no clean relationship to the invariable plane — the WGCCRE instead uses the positive pole given by the right-hand rule of the spin. Read the definition before trusting the word "north".
💡Planetocentric versus planetographic
The Earth lesson's geodetic/geocentric split reappears everywhere, with a twist:
Planetocentric latitude is measured from the body's centre, and longitude increases east, 0–360°. The modern standard for Mars and the Moon.
Planetographic latitude is measured from the normal to the reference ellipsoid (like geodetic latitude), and longitude increases in the direction opposite to the rotation — so on Mars, it increases west. Older Mars maps (Viking, Mariner 9) use it.
Jezero crater, where Perseverance landed, is at 18.44° N, 77.45° E planetocentric — and 18.65° N, 282.55° W planetographic. Same rock. Hand the wrong pair to a landing system and you are on the other side of the planet.
The Moon has a problem no other body has to the same degree: two perfectly good body-fixed frames, in daily use, that disagree by up to 875 metres on the surface.
Principal Axis (PA) — axes along the Moon's principal moments of inertia. This is the frame in which the Moon's rotation and its gravity field are naturally written, so lunar ephemerides (DE421, DE440) and gravity models such as GRAIL's are produced in it.
Mean Earth / polar axis (ME) — x-axis along the mean direction to Earth, z along the mean rotation pole. Because the Moon keeps one face to Earth, this is the natural frame for maps. LRO's topography and imagery, and almost every lunar map, use it.
The two differ because the Moon is not a perfect triaxial body: its axis of minimum inertia does not, on average, point exactly at Earth. DE421 puts the offset at rotations of about 68″, 79″ and 0.3″ — which, at 1,737 km from the centre, is up to 875 m.
🛰Artemis lunar frames and lunar time
In 2024 a NASA working group studied the ME/PA problem for Artemis. Both frames are accurate to about a metre, but an 875 m disagreement is a large fraction of a crewed landing's navigation budget. The recommendation: support both, with documented transformations — and use Mean Earth as the standard for surface operations and for exchanging location data.
The same year, the US Office of Science and Technology Policy directed NASA to establish Coordinated Lunar Time (LTC), because clocks on the Moon run faster than clocks on Earth — by roughly 56–59 microseconds per day, depending on exactly how the comparison is defined. In August 2024 the IAU General Assembly adopted Resolution II, establishing a Lunar Celestial Reference System (LCRS) and Lunar Coordinate Time (TCL), built the same way as the GCRS and its time scale. The Moon is getting the full apparatus of this unit — origin, axes and clock — for the first time.
Between the planets, two more frames do most of the work.
Ecliptic J2000 (SPICE's
ECLIPJ2000): the ICRF axes tipped by the J2000 obliquity, ε₀ = 23.439291°, about the shared x-axis. The planets orbit close to this plane, so interplanetary trajectories are drawn and discussed in it.Barycentric ICRF: the BCRS of lesson 1 — the ICRF axes, origin at the solar system's centre of mass. Planetary ephemerides are computed here, and a cruise trajectory is integrated here.
One rotation about x — the equinox direction is shared by both frames, which is the whole reason it was chosen as the x-axis.
💡Frames that rotate with two bodies
Near the Moon, or near the Sun–Earth Lagrange points, the most useful frame is not inertial at all. The synodic (rotating) frame of the circular restricted three-body problem turns with the two primaries, so that Earth and Moon (or Sun and Earth) stand still on the x-axis.
In that frame, the five Lagrange points are fixed places, and orbits that look like tangled spirographs in inertial space become simple closed loops. JWST, Gaia and SOHO live near Sun–Earth L1 and L2, about 1.5 million km from Earth. Artemis I's Orion flew a distant retrograde orbit around the Moon, and NASA's Gateway is planned for a near-rectilinear halo orbit. Both are designed and flown in Earth–Moon rotating coordinates.
The price is that the frame is not inertial: the equations carry Coriolis and centrifugal terms, energy is not conserved, and the conserved quantity is instead the Jacobi constant. Converting a state in or out needs the transport theorem one more time: .
Loading interactive…
💡How a real mission hands coordinates along
Mission design is a relay of frames. A Mars lander, end to end:
Launch — the launch vehicle delivers a hyperbolic escape described by C3 (twice the escape energy), and the direction of the outgoing asymptote as a declination (DLA) and right ascension (RLA), conventionally referred to EME2000. A pad at 28.5° N cannot easily reach some DLAs, which is a frame fact with a launch-window consequence.
Cruise — the trajectory is integrated in barycentric ICRF against the DE ephemerides, and plotted in ecliptic J2000.
Approach — the aim point is a pair of numbers B·T, B·R in the B-plane: the plane through the target's centre perpendicular to the incoming asymptote. T̂ is defined parallel to a reference plane — the ecliptic, or the planet's equator — and the mission must say which, or the aim point is rotated.
Entry and landing — the state is converted to the planet's IAU body-fixed frame, and the landing ellipse is drawn in planetocentric latitude and east longitude.
NASA's SPICE toolkit exists to make these hand-offs mechanical: ephemerides in SPK kernels, body orientation in PCK kernels, frame definitions in FK kernels, spacecraft attitude in CK kernels, leap seconds in an LSK. When someone says "load the kernels", they mean: load the frames.
⚠A day on Mars is not one number
Mars spins once against the stars every 24 h 37 m 22.7 s — that is what Ẇ gives you. But a Mars sol, sunrise to sunrise, is 24 h 39 m 35 s, for the same reason Earth's solar day is longer than its sidereal day. Mission operations schedule by sols; the orientation model runs on the sidereal rate. Mixing them puts the prime meridian about 0.54° per day in the wrong place — 32 km on the equator after a single sol.
↓But there is a problem
You now have the whole tree: quasars to ICRF, ICRF to J2000 and back, precession and nutation, Earth rotation and polar motion, the crust and its plates, the spacecraft's own axes, other worlds' poles and meridians, and frames that turn with the Moon.
Every single conversion in this unit asked the same question first: when? TT for precession, UT1 for rotation, TDB for ephemerides, an epoch for every station velocity. We have been treating "the instant" as a number you simply have.
It is not. Kepler could draw an orbit perfectly and still not tell you where Mars would be next Tuesday — and the equation he wrote to try has no solution you can write down.