Sovereign Star Registry — Deep-Time Astrometry Ledger

1. Audit Context

• Target Subsystem: moira.stars._propagate_icrs_vector
• Objective: Ensure stellar proper motion, parallax, and radial velocity calculations remain stable and mathematically rigorous across ±10,000-year epochs.
• Core Threat: Unbounded non-linear drift and coordinate distortion in high-proper-motion luminaries caused by tangential plane approximations.

2. Findings & Diagnostics

• Original State: The proper motion was computed using a first-order linear approximation on the tangent plane.
• The Divergence: At an 8,000-year propagation horizon, fast-moving stars exhibited catastrophic angular displacement. For example, Kapteyn's Star diverged by approximately 26.7 degrees relative to the exact IAU SOFA/ERFA truth vectors.
• Identified Causes:
  1. The projection completely ignored 3D space motion and perspective secular acceleration.
  2. The star's motion along the line of sight (radial velocity) was excluded from the calculations.

3. Implemented Interventions

• Record Structure: Expanded the _SovereignStarRecord data model and parser in moira/stars.py to securely store and coerce radial_velocity_km_s from the registry dataset.
• Vectorized Space Motion: Rewrote _propagate_icrs_vector to fully support Cartesian 3D space-motion.
  • Converted angular proper motion into exact tangential velocities ($v_T = d_{pc} \cdot \mu$).
  • Aligned radial velocity precisely along the vector line of sight using explicit exact IAU coefficients ($1.022712 \times 10^{-6}$ pc/yr per km/s).
  • The actual coordinate translation is now resolved purely in 3-dimensional Cartesian space, normalizing the unit vectors after time progression, prior to returning to spherical coordinates.

4. Validation Results

• Test Regimen: tests/unit/test_deep_time_astrometry.py
• Sweep Parameters: Iterating from Julian year J−8000 to J+8000 in 100-year increments.
• Comparison Oracle: ERFA's position and proper motion reference formulas (erfa.starpm).
• Outcome:
  • The catastrophic 26-degree drift has been eliminated.
  • Angular divergence for all tested stars (Barnard's Star, Kapteyn's Star, 61 Cygni A, and Groombridge 1830) strictly conforms to bounds $< 0.05$ degrees.
  • This negligible remaining variance appropriately isolates the relativistic aberration iterations present in ERFA, which are deliberately absent from Moira's physical substrate translation.
• Status: Stable. The astronomical substrate is sovereign and precise across the tested deep-time boundaries.