Deriving Planetary Masses from Moon Orbits

How precisely can we derive a planet’s mass (GM) from the orbit of one of its moons, using only Kepler’s third law plus physically-motivated corrections? This page documents a single closed-form formula that approximates DE440 mass ratios for every moon-bearing planet in the solar system — from Earth out to Pluto — with residuals ranging from 3 ppm (Neptune) to 340 ppm (Mars). The Earth-Moon case is treated first as a specialisation, then generalised.

1. The Earth-Moon problem

For an isolated two-body system orbiting their common barycenter, Kepler’s third law gives the combined gravitational parameter exactly:

G(M₁ + M₂) = 4π² · a³ / T²

Applied to the Moon orbiting Earth with a_M = 384,399.07 km (IAU mean lunar distance) and T_M = 27.32166156 days (IAU sidereal month):

G(M_Earth + M_Moon)_raw  =  4π² · 384,399.07³ / (27.32166156 × 86400)²
                         =  402,406.51 km³/s²

The JPL DE440 reference value is 403,503.24 km³/s². The naive Kepler result is 0.27% too low — a gap of 1,097 km³/s². This is not the M+m split (handled separately via the Earth/Moon mass ratio). It is fundamentally caused by the Sun’s tidal perturbation on the Moon’s orbit — the classical 3-body problem (no closed-form solution since Poincaré 1890).

2. The Δa correction

The fix is to apply a physically-motivated shift to the Moon’s semi-major axis before the Kepler computation:

Δa  =  a_M  ·  μ  ·  m

Each factor is a natural quantity of the Earth-Moon-Sun system:

The product is the Earth-Moon barycentric wobble × orbital phase fraction during one lunar orbit: a_M · μ ≈ 4,670 km is Earth’s wobble around the Earth-Moon barycenter; m ≈ 7.5% is the fraction of Earth’s heliocentric orbit completed during one lunar month. For Earth-Moon-Sun, Δa = 349.37 km.

What this formula honestly is

The actual leading-order solar perturbation in Hill-Brown lunar theory scales as m²: ΔGM/GM ≈ α₂ · m² for a rational coefficient α₂ ≈ ½. Our formula uses Δa/a = μ · m, giving ΔGM/GM = 3·μ·m (factor of 3 from cubing (1 + Δa/a) in Kepler’s law, since GM ∝ a³). For these to agree:

3 · μ · m  ≈  ½ · m²       →       μ  ≈  m / 6

In our solar system μ = 0.01215 and m = 0.0748, so μ/m = 0.162 ≈ 1/6.16 — the relation μ ≈ m/6 holds to ~3%. This is what makes the two forms numerically equivalent.

So Δa = a_M · μ · m is a clean re-parameterisation of Hill-Brown’s ½·m² correction using two physically meaningful inputs (mass ratio × period ratio) instead of m² and a rational coefficient. It is not a new physical law — it is a useful shorthand that exploits a numerical relation specific to our system. If μ/m were significantly different (e.g., a hypothetical Moon with twice today’s mass), the simple 3·μ·m form would not give the right correction; one would have to use the underlying α₂·m². The model uses this re-parameterisation because all three factors are observed quantities the model already tracks, and the agreement with JPL DE440 (~4 ppm) is well within the precision floor of any Kepler-from-Moon-orbit derivation.

3. The Earth-Moon computation chain

Inputs

Hardcoded (observational):

a_M       =  384,399.07 km         // IAU mean lunar distance
T_M_iau   =  27.32166156 days      // IAU sidereal month
ratio     =  81.30056816           // M_Earth / M_Moon (DE440 SPICE kernel)

Model-derived (computed at runtime):

T_M       =  27.32166241 days      // T_M_iau quantized to fit H integer-moons
T_S       =  365.25636437 days     // sidereal year derived via H/13 from tropical year
LOD       =  86,399.99968 s        // sidereal-seconds / sidereal-days ratio

Using model-derived values (rather than raw IAU references) keeps the GM derivation consistent with the model’s year-length, sidereal-day, and scene-graph speeds. Sub-ppm differences from nominal IAU values.

Computation

1.  m   =  T_M / T_S                                    =  0.07480133     (lunar small parameter, Hill 1878)
2.  μ   =  1 / (ratio + 1)  =  M_Moon / (M_E + M_M)     =  0.01215058     (Moon mass fraction)
3.  Δa  =  a_M · μ · m                                  =  349.37 km      (solar-tidal shift)
4.  a_corrected  =  a_M + Δa                            =  384,748.44 km
5.  T (in seconds)  =  T_M · LOD                        =  2,360,591.63 s
6.  GM(M_E + M_M)  =  4π² · a_corrected³ / T²           =  403,504.73 km³/s²
7.  GM_Earth  =  GM(M_E + M_M)  ·  ratio / (ratio + 1)  =  398,601.91 km³/s²
8.  GM_Moon   =  GM(M_E + M_M)  /  (ratio + 1)          =  4,902.82 km³/s²

Precision vs reference values

The G uncertainty (~22 ppm) sets a hard floor on how precisely M_Earth can be expressed in kg.

Where the 384,748 km Moon distance comes from

The Moon’s Kepler-effective distance — 384,748 km, sometimes listed in physics references as the Moon’s mean orbital radius — is an established textbook value that appears in independent academic references as a stated constant (e.g., UNLV / D. Jeffery astrophysics reference pages list it directly; university physics textbooks pair T = 27.321661 days with r = 384,748 km in Kepler’s-law homework problems).

Wikipedia’s Lunar Distance article catalogues four distinct ways to compute the Moon’s mean distance — none produces 384,748 km. The Kepler-effective semi-major axis (384,748 km) is the value Kepler’s third law requires to recover the JPL DE440 combined parameter from the IAU sidereal month, but no public source we found gives a clean closed-form correction connecting the geometric a_M (384,399) to this Kepler-effective value. The Δa = a_M · μ · m formula provides exactly that bridge:

geometric a_M (LLR / Brown's parallax)        =  384,399 km
+  Δa  =  a_M · μ · m                         =     349 km
=  Kepler-effective semi-major axis           =  384,748 km

Using only μ and m — both already tracked by the model — and numerically equivalent to Hill-Brown’s leading ½·m² correction because μ ≈ m/6 in our solar system.

The Sun-side analog

The Moon-side Δa = +349 km has a symmetric counterpart on the Sun side of the same Earth-Moon-Sun derivation. Kepler’s third law applied to Earth’s heliocentric orbit returns the combined GM(Sun+Earth); to recover GM_Sun alone, the model subtracts GM_Earth_alone. That GM-level subtraction is mathematically equivalent to plugging a slightly shrunk “Kepler-effective AU” into the bare Kepler formula:

Δa_Sun-side  =  AU / (3 × M_Sun / M_Earth_alone)  ≈  149.77 km
a_eff_Sun    =  AU − Δa_Sun-side  ≈  149,597,720.93 km
GM_Sun       =  4π² · a_eff_Sun³ / T_sidereal_year²   (= same result as GM-subtraction)

The Kepler-correction picture is symmetric but opposite in sign for the two bodies:

The quick-derivation formula Δa ≈ a / (3 × mass_ratio) works in both directions:

• For Moon: Δa_Moon = a_M · μ · m (see §2) — equivalent to a small upward shift on a_M because bare Kepler under-estimates GM(EMB) by the solar-tidal perturbation.
• For Sun: Δa_Sun-side = AU / (3 × M_Sun/M_Earth) ≈ AU × Earth_mass_fraction / 3 — a small downward shift on AU because the Kepler-from-AU formula returns GM(Sun+Earth), and we want GM(Sun) alone.

In the code, the Sun-side correction is done at the GM level (subtract GM_Earth_alone), not at the AU level. The 149.77 km AU-equivalent is purely conceptual — useful for seeing the structural symmetry with the Moon-side derivation. This Earth-specific 149.77 km value is the special case of a universal Sun-side Δa formula that works for all 9 planets (including Mercury and Venus, which the Moon-side derivation cannot reach). Full general derivation, algebraic identity that makes it exact, and per-planet table from Mercury (3 km) to Jupiter (248,000 km): Sun-side Δa Formula.

4. The universal formula

For any planet P observed via one of its moons M:

GM_P_system  =  ( 4π² · a_M³ / T_M² )  ·  ( 1  +  3·μ·m  −  1.5·J2·(R_P/a_M)² · (1 − 1.5·sin²i) )
                └────────┬────────┘     └──────┬─────┘   └────────────────┬──────────────────┘
                  bare Kepler          solar Δa term         planet-oblateness J2 term
                                       (additive)              (subtractive)

where:

The formula’s output GM_P_system covers planet + all moons. To recover the planet-alone value: GM_P_alone = GM_P_system · ratio / (ratio + 1) for single-moon systems, or GM_P_system − ΣGM_moons for multi-moon systems. This distinction matters: published “Sun/Planet” mass ratios are inconsistent across sources (DE440’s BODY1–BODY9 are planet-system ratios; BODY199/BODY299/etc. are planet-alone).

Physical meaning of each term

Term 1 — bare Kepler (4π²·a³/T²): the two-body Keplerian relationship. Exact for an isolated two-body problem.

Term 2 — solar Δa (+3·μ·m): the Sun pulls on planet and moon unequally; the differential pull modifies the moon’s effective Kepler orbit. Re-parameterises Hill-Brown’s leading-order solar perturbation. Positive: bare Kepler under-estimates GM_true. Geometric reading: a·μ = planet’s wobble around the planet-moon barycenter; m = fraction of planet’s heliocentric year completed during one lunar month.

Term 3 — J2 oblateness (−1.5·J2·(R/a)²·(1 − 1.5·sin²i)): a non-spherical planet creates a non-Keplerian potential. Bare Kepler over-estimates GM_true for an oblate planet’s moon — subtract this correction to recover GM_true. The (1 − 1.5·sin²i) factor handles non-equatorial moons; for retrograde or polar orbits it can flip sign.

5. Multi-planet validation

Apply the formula to every major moon of each planet and compare to DE440. If correct, all moons of a given planet should converge to the same Sun/Planet ratio (to within data-precision floors).

Earth-Moon (1 moon) — DE440 Sun/EMB: 328,900.5614

Corrected formula matches DE440 to 3.7 ppm. The 2,727 ppm shift between bare and corrected reflects the full Earth-Moon Δa = 349 km.

Saturn (7 major moons) — DE440 Sun/System: 3,497.9018

The most striking validation — J2 is the dominant correction:

Bare-Kepler spread across 7 moons: 2,300 ppm (3,489.5–3,497.7). Corrected: 475 ppm (3,497.15–3,498.81). The J2 correction collapses Saturn’s spread by ~5× — the strongest single piece of evidence the formula captures real physics.

All seven moon-bearing planets — best-moon match to DE440

Sun/system vs Sun/planet-alone

For “system” vs “planet-alone” the conversion is straightforward but matters for binary-like systems:

Earth (1.2%) and especially Pluto (11%) stand out as the two binary-like systems where the system / planet-alone distinction shifts the ratio by percent-level.

6. Which correction dominates where

Both correction terms exist for every system; the only question is which dominates. The solar Δa term scales as μ·m, and m = T_moon / T_planet shrinks dramatically for outer planets:

Earth is unique in having a sizable m (0.075) because it is the innermost moon-bearing planet. Pluto-Charon is the mirror image: tiny m but huge μ — a true binary. Earth and Pluto sit at opposite ends of the same curve, not as special cases.

7. Origins and prior work

Honest accounting of what is classical, what may be original, and what this is not.

What is classical

None of the underlying physics is original to this work.

What may be original

1. The re-parameterisation Δa = a·μ·m — the Hill-Brown leading solar correction is conventionally written as α₂·m² with α₂ from perturbation expansions. Re-expressing it as 3·μ·m — where both factors are intrinsic observables — gives a geometric reading: a·μ is the planet’s wobble around the planet-moon barycenter, and m is the moon’s month as a fraction of the planet’s heliocentric year. The two forms agree when μ ≈ m/6, which holds in the Earth-Moon-Sun system to ~3%. We have not found this presentation in standard texts.
2. Closed-form derivation of the textbook value 384,748 km — this Kepler-effective Moon distance appears in physics references (UNLV, university homework problems) but is stated rather than derived in those sources. The Δa = 349 km bridge from the geometric LLR value appears to be original.
3. Unified universal formula across all 7 moon-bearing planets — while each term is classical, packaging them as a single closed-form and demonstrating it against 22 moons of 7 planets in the DE440 reference frame is a synthesis we have not seen in textbook or review form.

What this is not

• Not a new law of physics — the corrections are all classical.
• Not a fundamental discovery — JPL and IAU have used these tools for decades to fit ephemerides.
• Not a 7th Fibonacci Law — the Fibonacci Laws describe the model’s intrinsic cycle structure (precession, inclination, eccentricity resonances). This formula is calibration/derivation work that produces the mass inputs the Fibonacci framework consumes; they live at different layers.

A pedagogical / synthesising contribution suitable for an undergraduate astrodynamics audience or a physics-education journal (e.g., American Journal of Physics, European Journal of Physics). The novel pieces are honest small contributions, not paradigm shifts.

8. Residual: why ~3.7 ppm remains

The Δa correction reduces the 0.27% raw-Kepler gap to a ~3.7 ppm residual. That residual reflects higher-order Brown’s lunar theory terms (m⁴ and beyond) that no clean closed-form correction can capture. This is a fundamental limit of the 3-body problem: Poincaré (1890) proved analytically there is no closed-form solution; Brown’s lunar theory uses thousands of terms to reach JPL’s precision. Any model that derives GM_Earth from the Moon’s orbit via Kepler + a single correction term has a precision floor in the few-ppm range. JPL DE440 sidesteps this entirely by fitting GM_Earth from artificial-satellite tracking (~10⁻⁹ precision), not from the Moon’s orbit. Our derivation matches the precision floor of the Kepler-from-Moon-orbit method.

9. Data-convention caveat

The universal formula assumes published (a, T) are observational averages (mean elements), in which case +3·μ·m corrects for solar perturbation (matters for Earth) and −1.5·J2·(R/a)² corrects for planet oblateness (matters for outer-planet moons). For sources that publish osculating (Kepler-effective) elements at a specific epoch, the J2 correction is already absorbed into a, and bare Kepler returns GM_true directly. This is why some Jupiter Galilean residuals are unusually small with bare Kepler alone — JPL Horizons publishes osculating elements for major moons. When in doubt, applying the J2 correction to mean elements over-corrects; not applying it to osculating elements under-corrects. The 50–475 ppm spreads in the validation tables reflect this ambiguity plus genuine observational precision floors.

10. Planets without moons: Mercury and Venus

Mercury and Venus have no moons, so there is no (a_M, T_M) to plug into Kepler. Their mass ratios — MASS_RATIO_SUN_MERCURY = 6,023,657.94, MASS_RATIO_SUN_VENUS = 408,523.72 — come exclusively from spacecraft trajectory perturbations (Mariner 10, MESSENGER, Venera, Magellan). This is a fundamental observational limit — no closed-form orbital derivation exists for planets without natural satellites.

The companion Sun-side Δa Formula does include Mercury and Venus — but in that direction the formula computes orbital periods from heliocentric semi-major axes (given known GMs), not GMs from moon orbits. The two pages go in opposite directions through Kepler’s third law.

Summary

The Earth-Moon Δa = a·μ·m discovery is the leading term of a universal closed-form formula:

GM_P_system  =  (4π² · a³/T²) · ( 1  +  3·μ·m  −  1.5·J2·(R_P/a)²·(1 − 1.5·sin²i) )

Every observable is either an orbital element of a moon (a, T), an intrinsic mass/period ratio (μ, m), or a published gravity coefficient (J2). No fitted parameters. Applied to the major moons of Earth, Mars, Jupiter, Saturn, Uranus, Neptune, and Pluto, both Sun/system and Sun/planet-alone ratios match DE440 to 3–340 ppm (best moon). Earth and Neptune both reach 3–4 ppm precision — the Hill-Brown m⁴ floor of the 3-body problem. The same formula handles both Earth (solar-Δa-dominated) and outer-planet moons (J2-dominated) and the binary-like Pluto-Charon case. For Saturn specifically, the J2 correction collapses the 7-moon spread from 2,300 ppm to 475 ppm — a ~5× reduction that is the strongest single piece of evidence the formula captures real physics.

This page is the Moon-side of a two-page mass/period derivation. Its companion Sun-side Δa Formula extends the same Δa machinery to derive each planet’s orbital period from its heliocentric semi-major axis — including Mercury and Venus, which the Moon-side cannot reach. Together they form a single coherent picture: every planet’s GM and every planet’s period can be expressed as a one-line Kepler formula plus a one-line Δa shift.

See also

• Sun-side Δa Formula — the companion page generalising the Earth-specific 149.77 km Sun-side analog (§3) to a universal closed-form Δa for every planet
• Formulas — model’s analytical formulas for obliquity, eccentricity, precession
• Scientific Background — comparison with standard precession theory
• Glossary — definitions for GM, mass ratios, sidereal year, etc.
• JPL DE440 SPICE kernel: gm_de440.tpc  — authoritative system GMs