Relativistic Lunar Reference Frame

• Relativistic Lunar Reference Frame is a Moon-centered space-time system that integrates relativistic corrections to enable meter-level positioning and nanosecond-level timing.
• It differentiates between an inertial lunar celestial system (LCRS/TCL) for orbit determination and a body-fixed realization (ILRF) for surface mapping and navigation.
• The framework underpins enhanced lunar navigation and scientific measurements via precise time-transfer models validated through Lunar Laser Ranging and advanced ephemerides.

A relativistic lunar reference frame is the Moon-centered space-time reference architecture required when lunar and cislunar positioning, navigation, and timing must be performed at meter-level accuracy, where timing errors must be only a few nanoseconds and relativistic effects can no longer be treated as secondary corrections. In current work, the problem is treated in two linked parts: a Moon-centered relativistic reference system, usually the Lunar Celestial Reference System (LCRS) with coordinate time TCL, and a body-fixed realization needed for cartography, navigation, and timing services. Recent IAU-based analyses place this architecture directly beside the terrestrial BCRS/GCRS framework and argue that future lunar infrastructures such as LunaNET and Moonlight require a common coordinate system and common reference timescale for interoperable PNT, communications, and science (Defraigne et al., 4 Nov 2025, Sośnica et al., 17 Oct 2025).

1. Relativistic hierarchy and the lunar-centered system

The modern lunar problem is formulated inside the same relativistic hierarchy already used for Earth and the solar system. IAU Resolution B1.3 (2000) defines the Barycentric Celestial Reference System (BCRS) with coordinate time TCB, and the Geocentric Celestial Reference System (GCRS) with coordinate time TCG. IAU 2024 Resolution II recommends constructing a Luni-centric Celestial Reference System (LCRS), using the same methodology as for the GCRS, with associated coordinate time TCL (Lunar Coordinate Time), and explicitly recommends that the unit of TCL be consistent with the SI second. IAU 2024 Resolution III recognizes the practical need for an internationally recognized lunar reference time and calls for work on relating possible lunar time realizations to other timescales, especially UTC (Defraigne et al., 4 Nov 2025).

In the explicit IAU-style lunar construction, the LCRS is the lunar analog of the GCRS. Its origin is the lunar center of mass, the Moon’s own gravitational field is excluded from the external potential in the same way the Earth’s field is excluded from the GCRS definition, and the axes are assumed non-rotating with respect to the BCRS axes. The spatial coordinate is the selenocentric radius vector,

$\mathbf z = \mathbf x - \mathbf x_L.$

This literature is clear that the primary development is the temporal component of the lunar relativistic reference system; a complete operational spatial lunar frame is left to broader IAU and geodetic work (Kopeikin et al., 2024).

A further 2026 synthesis places the Moon inside a broader operational chain consisting of BCRS/TCB, GCRS/TCG, terrestrial operational scales TT, TAI, UTC, the lunar LCRS/TCL, and in one proposed implementation a lunar operational scale LTC steered to TCL. In that treatment, cislunar operations do not receive a separate full metric reference system; they traverse the BCRS/GCRS/LCRS hierarchy through explicit transformations (Jin et al., 1 Jul 2026).

2. Celestial versus body-fixed lunar frames

The relativistic lunar frame in the narrow sense is not identical to the lunar body-fixed frame used for maps, surface coordinates, or landed assets. Current work separates the Moon-centered celestial frame from a body-fixed realization. In the ESA Moonlight review, the intended architecture is an inertial or kinematically non-rotating LCRS for dynamics and clocks, and a Moon body-fixed LRS analogous to Earth’s ITRS for surface geodesy and cartography. Within existing lunar practice, two body-fixed systems are emphasized: the Principal Axis (PA) system and the Mean Earth / Rotation Axis (ME) system. The PA system is the natural frame for rotational dynamics, whereas ME remains historically important for cartography and archival products. The static offset between PA and ME corresponds to roughly 860 m on the surface (Fienga et al., 2024).

The first combined operational realization is the International Lunar Reference Frame (ILRF). ILRF is defined as the Principal Axis (PA) system, attached to the surface and co-rotating with the Moon, with its origin in the lunar center of mass (lunocenter). Its orientation is realized by combined Euler angles from INPOP21a, DE430, and EPM2021, and the transformation from the quasi-inertial lunar celestial frame to the body-fixed PA frame is written

$$\mathrm{PA} = R_z(\psi)\,R_x(\theta)\,R_z(\varphi)\,\mathrm{LCRF}.$$

The combined ILRF is characterized by the mean error of 17.6 cm for 2010–2030, where 15.3 cm comes from the origin and 8.6 cm from the orientation realization. The dominant weakness is the origin, because the lunocenter is inferred indirectly from a sparse near-side retroreflector geometry, with a reported correlation coefficient

$r = -0.97$

between the $X$-component of the lunocenter in PA and scale (Sośnica et al., 17 Oct 2025).

This distinction between celestial and body-fixed realizations is operationally decisive. The relativistic LCRS/TCL provides the Moon-centered coordinate system and coordinate time for orbit determination and timing, while the PA-based ILRF provides the body-fixed realization required for positioning, cartography, and surface operations.

3. Coordinate time, proper time, and lunar clock behavior

A central distinction in all relativistic lunar-frame work is the separation between coordinate time and proper time. TCL is the coordinate time of the LCRS. A real clock measures proper time along its own worldline. For a clock $l$ on or around the Moon, a standard first post-Newtonian expression is

$$\frac{d\tau_l - d\mathrm{TCL}}{d\mathrm{TCL}} = -\frac{1}{2c^2}\left[v_l^2 + 2W(\mathrm{TCL},\mathbf{X})\right],$$

with

$$W(\mathrm{TCL},\mathbf{X}) = U_L(\mathrm{TCL},\mathbf{X}) + U_{\text{tidal}}(\mathrm{TCL},\mathbf{X}) + U_{\text{inertial}}(\mathrm{TCL},\mathbf{X}).$$

Here $U_L$ is the lunar Newtonian potential, $U_{\text{tidal}}$ is mainly due to Earth and Sun, and $U_{\text{inertial}}$ is the inertial potential due to the non-geodesic acceleration of the Moon’s center of mass. The inertial term is said to be negligible, but the Moon differs from Earth because the 1:1 spin–orbit resonance makes one component of the Earth-raised lunar tide permanent, contributing at the level of

$$\mathrm{PA} = R_z(\psi)\,R_x(\theta)\,R_z(\varphi)\,\mathrm{LCRF}.$$0

in relative frequency, depending on location on the lunar surface (Defraigne et al., 4 Nov 2025).

Using GRAIL gravity and LOLA topography, recent estimates give a mean relative frequency difference of about

$$\mathrm{PA} = R_z(\psi)\,R_x(\theta)\,R_z(\varphi)\,\mathrm{LCRF}.$$1

between a clock at rest on the lunar surface and TCL, corresponding to about

$$\mathrm{PA} = R_z(\psi)\,R_x(\theta)\,R_z(\varphi)\,\mathrm{LCRF}.$$2

The variation over the surface due to topography is within about

$$\mathrm{PA} = R_z(\psi)\,R_x(\theta)\,R_z(\varphi)\,\mathrm{LCRF}.$$3

This means there is no single simple “surface clock rate” independent of location. Surface proper time differs from TCL in a way that depends on gravity, topography, and position (Defraigne et al., 4 Nov 2025).

The Earth–Moon comparison is likewise not reducible to a universal scalar offset. One major conclusion is that there is no universal $$\mathrm{PA} = R_z(\psi)\,R_x(\theta)\,R_z(\varphi)\,\mathrm{LCRF}.$$4 difference. The comparison depends on the worldline of the Moon clock, the worldline of the Earth clock, the coordinate system used in the time-transfer calculation, and the light-time model. For casual coordination with Earth, a predictable secular correction may be sufficient at the microsecond level, but for precise PNT and science the exact relation between a lunar clock and UTC must be computed through a relativistic time-transfer model (Defraigne et al., 4 Nov 2025).

An explicit TCL–TCG program gives further scale. At the Moon’s center, TCL runs slower than TCG by about $$\mathrm{PA} = R_z(\psi)\,R_x(\theta)\,R_z(\varphi)\,\mathrm{LCRF}.$$5. The long-term mean proper-time difference between a clock on the lunar selenoid and one on the Earth geoid is

$$\mathrm{PA} = R_z(\psi)\,R_x(\theta)\,R_z(\varphi)\,\mathrm{LCRF}.$$6

The dominant periodic TCL–TCG term is monthly, with amplitude about $$\mathrm{PA} = R_z(\psi)\,R_x(\theta)\,R_z(\varphi)\,\mathrm{LCRF}.$$7–$$\mathrm{PA} = R_z(\psi)\,R_x(\theta)\,R_z(\varphi)\,\mathrm{LCRF}.$$8 and period 27.55 days. For surface clocks, the local term

$$\mathrm{PA} = R_z(\psi)\,R_x(\theta)\,R_z(\varphi)\,\mathrm{LCRF}.$$9

produces a constant offset up to 19.8 ns and monthly location-dependent variations at the 1.1 ns and 2.3 ns level (Kopeikin et al., 2024).

4. Choice of the lunar reference timescale

Once the LCRS and TCL are accepted, a separate question remains: which practical coordinate-time realization should be adopted as the operational lunar reference? A generic scaled lunar time is written

$r = -0.97$0

and three options have been analyzed in detail (Defraigne et al., 4 Nov 2025).

Option	Definition	Main consequence
1	$r = -0.97$1	use TCL directly
2	$r = -0.97$2	match average proper time on a chosen lunar equipotential
3	$r = -0.97$3	remove secular drift relative to TT/UTC

Option 2 is explicitly analogous to TT, which is a scaled version of TCG with

$r = -0.97$4

and Option 3 is conceptually parallel to TDB, defined with

$r = -0.97$5

The key objection to lunar scaling is that, in the IAU relativistic framework, scaling a coordinate time requires scaling the associated spatial coordinates and gravitational parameters. By analogy with TT and TDB scaling,

$r = -0.97$6

and similarly for a lunar scaled time $r = -0.97$7, corresponding lunar scaling of $r = -0.97$8 and $r = -0.97$9 would be required. The reported Earth–Moon distance impact is about 1 cm for Option 2 and about 30 cm for Option 3. This is operationally undesirable in high-accuracy orbit determination, laser ranging, and navigation (Defraigne et al., 4 Nov 2025).

The steering burden also differs sharply. If TCL is used directly, a perfect clock on the lunar surface differs in rate from TCL by about

$X$0

so a surface clock realizing TCL would need a frequency steering correction of about

$X$1

If the surface-matched scaled time is used, the residual can still be up to about

$X$2

If the no-secular-drift-relative-to-TT option is used, a perfect lunar surface clock differs from the reference by about

$X$3

equivalent to about

$X$4

Recent trade-off analyses therefore conclude that TCL itself should be used as the practical lunar reference time (Defraigne et al., 4 Nov 2025).

Other papers represent alternative positions within the same problem. One derivation introduces a scaled lunar surface time TL with

$X$5

together with

$X$6

and obtains a secular TL–TT drift of $X$7, plus periodic terms and Moon-centered spatial scale and Lorentz-contraction corrections (Turyshev et al., 2024). A 2026 time-scale paper, by contrast, argues that any scaling of local coordinate times like TCL is unreasonable and provides TCB $X$8 TCL time ephemerides for all major bodies using INPOP19a, represented as Chebyshev polynomials (Klioner, 17 Apr 2026).

5. Dynamical realization through LLR, ephemerides, and lunar infrastructure

In practice, a relativistic lunar reference frame is realized through observations and dynamical models. The most important current observable is Lunar Laser Ranging (LLR). Each LLR observation is the round-trip travel time of a laser pulse transmitted from an Earth observatory, reflected by a lunar retroreflector, and received back on Earth. The lunar frame is therefore realized through a coupled model of Earth–Moon dynamics, Earth orientation, lunar orientation, station coordinates, reflector coordinates, and relativistic light-time (Williams et al., 2012).

One operational formulation states that “the lunar ephemeris and the positions of the lunar retroreflectors thus form a most precise lunar reference frame.” In that work, reflector coordinates are determined in a selenocentric coordinate system based on the principal axes (PA) of the Moon’s figure, and LLR can determine two out of three daily lunar orientation parameters, $X$9 and $l$0, but is not sensitive to the third parameter $l$1, the rotation around the Earth-pointing lunar $l$2-axis (Pavlov, 2019).

LLR also supplies the relativistic and dynamical validation needed by any lunar frame. Classical equivalence-principle analyses give

$l$3

from 17,580 ranges, with modern 2005–2011 data fit at 1.8 cm weighted RMS (Williams et al., 2012). More recent high-precision LLR analyses using APOLLO and infrared OCA data increased the dataset to 27,485 NP, improved reflector and synodic coverage, and reduced parameter correlations affecting frame-like quantities, lunar libration, and relativistic parameter estimates (Biskupek et al., 2020).

Future infrastructures extend this realization beyond LLR alone. The ESA Moonlight review treats LCRS/TCL as the inertial timing and dynamical backbone of a lunar radio-navigation system, while also assessing improvements from added lunar retroreflectors and orbiter altimetry (Fienga et al., 2024). The NovaMoon concept goes further by co-locating a lunar laser retroreflector, a VLBI transmitter, a receiver for navigation signals compatible with LunaNet standards, high-stability atomic clocks, and direct-to-Earth radio links. It is designed to improve the origin, orientation, and scale of the lunar reference frame and to provide the first long-duration physical realisation of a lunar time reference (Molli et al., 9 Feb 2026).

6. High-precision extensions, competing formulations, and open issues

A newer line of work pushes the lunar relativistic frame well beyond meter-level PNT. One high-precision framework introduces a fully post-Newtonian LCRS metric retaining all contributions above a fractional threshold of

$l$4

and timing terms above

$l$5

by expanding the lunar gravity field to spherical-harmonic degree $l$6, including Love-number variations, and carrying external tidal and inertial multipoles to the octupole. In that treatment, harmonics through $l$7 and tides through $l$8 are required for sub-picosecond synchronization and centimeter-level navigation in cislunar space (Turyshev, 29 Jul 2025).

A parallel 2026 implementation paper supplies a unified 1PN documentation chain from metric and proper time to null geodesics, light-time, and two-way range-rate for Earth, Moon, and Mars. It treats the Moon as a selenocentric body-centric CRS embedded in the BCRS, uses TCL as the coordinate time of the LCRS, and introduces a proposed operational lunar scale LTC steered to TCL. In that framework, lunar selenoid–geoid rates are reported as about $l$9–$$\frac{d\tau_l - d\mathrm{TCL}}{d\mathrm{TCL}} = -\frac{1}{2c^2}\left[v_l^2 + 2W(\mathrm{TCL},\mathbf{X})\right],$$0, and the authors stress that multi-CRS consistency relies on documented transformation chains rather than a single master clock (Jin et al., 1 Jul 2026).

Not all published approaches accept the same conceptual basis. One 2025 paper contests recent TT-to-LTC and TCG-to-TCL derivations, argues for an intermediate Earth–Moon barycentric coordinate system with coordinate time TCC, and proposes a stronger locality principle together with a Frenet-frame interpretation of non-rotation. It explicitly criticizes defining Earth-centered orientation using distant celestial bodies and presents an alternative hierarchical transformation chain through BCRS $$\frac{d\tau_l - d\mathrm{TCL}}{d\mathrm{TCL}} = -\frac{1}{2c^2}\left[v_l^2 + 2W(\mathrm{TCL},\mathbf{X})\right],$$1 EMCRS $$\frac{d\tau_l - d\mathrm{TCL}}{d\mathrm{TCL}} = -\frac{1}{2c^2}\left[v_l^2 + 2W(\mathrm{TCL},\mathbf{X})\right],$$2 GCRS/LCRS (Liu et al., 21 Jul 2025). This contrasts with the IAU-based literature, where the LCRS/TCL pair is constructed directly as the lunar analogue of the GCRS/TCG (Kopeikin et al., 2024).

Several open issues therefore remain. The body-fixed lunar realization is still limited by retroreflector geometry and long-term libration modeling. The exact operational lunar timescale is not yet standardized internationally, and names such as TL and LTC are used differently across papers. A universal scalar “Moon time minus UTC” formula is explicitly rejected. The most consistent trend in recent IAU-based work is instead to use unscaled TCL as the lunar reference coordinate time, realize the body-fixed frame through a PA-based ILRF, and connect Earth and Moon clocks through explicit relativistic time-transfer models rather than through a single universal lunar–terrestrial offset (Defraigne et al., 4 Nov 2025).