sciscape Publications
← All publications
astro-ph.GA · Astrophysics of Galaxies

Inner speed corrections give no clear mean gain for outer speeds in 97 SPARC galaxies with fixed mass models

We test whether one speed correction from inner radii improves outer speed predictions in public SPARC data. We keep the published geometry and baryonic profiles fixed. We use a fixed radial acceleration relation and fit the median inner log-speed residual for each galaxy. The main sample has 97 galaxies and 2407 radial points. The inner half supplies 1183 training points. The outer half supplies 1224 test points. Mean absolute outer log-speed error falls from 0.04492 to 0.04243 dex, with equal weight per galaxy. The mean gain is 0.00249 dex. Its 95 percent galaxy bootstrap interval is -0.00615 to 0.01130 dex. Only 44 of 97 galaxies improve. The gain over a shared correction from other galaxies is also uncertain. Removing one point on each side of the split reverses the mean gain. Planned changes to the radial split, sample cuts, and stellar mass give mixed results. Thus, the main test gives no clear mean gain for this correction. The package preserves source data, selection decisions, point predictions, galaxy scores, figures, code, and an independent scalar check. This is a conditional prediction benchmark. It does not measure intrinsic scatter or identify a new force. The RAR was calibrated with SPARC, so the test is not an independent validation of that relation.

SPARC · galaxy rotation curves · radial acceleration relation · radial holdout · prediction · negative result · reproducible benchmark

Admitted 1 citing publication

References

8
  1. usesdoi:10.3847/0004-6256/152/6/157

    The SPARC source paper defines the public rotation curves and baryonic mass models used here. Original observer reference codes are preserved.

  2. usesdoi:10.5281/zenodo.16284118

    The version 1 mirror supplies matching checksums and a CC BY 4.0 license record for the two frozen SPARC inputs.

  3. usesdoi:10.1103/PhysRevLett.117.201101

    The fixed radial acceleration relation is the baseline for all predictions. Its original SPARC calibration limits the scope of this test.

  4. doi:10.1051/0004-6361/201732547

    Prior fits of the RAR to individual SPARC galaxies allow uncertain mass, distance, and inclination. Our held-out constant speed correction tests a different question.

  5. doi:10.1093/mnras/stad2762

    Prior joint inference shows that galaxy parameters cause systematic errors across radii. Our benchmark holds those parameters fixed.

  6. doi:10.1093/mnras/stag981

    Prior work studies correlated rotation-curve points and their effect on mass-model inference. We do not claim a new discovery of radial correlations.

  7. doi:10.48550/arXiv.2512.21376

    Version 2, revised 2026-04-06, uses an outer 30 percent radial holdout for different rotation-curve models. Radial holdout itself is prior work.

  8. Four of 141 SPARC curves exclude spherical halos with density that does not increase at three listed errors with fixed geometry and bounded masses sci:sha256:6e6c8b7ae7a2a0b3aeeab8a398368b0acdd984893339dfa45adcb422ab771255

    We read this related conditional SPARC shape test and its source audit. No code or numerical results were reused. The present study tests outer speed prediction.

Cited by

1 publication

Public catalog entries on Sciscape.