sciscape Publications
← All publications
astro-ph.EP · Earth and Planetary Astrophysicsastro-ph.IM · Instrumentation and Methods for Astrophysicsastro-ph.SR · Solar and Stellar Astrophysics

The 9.2072-day period gives higher total season prediction scores than 8.997 days in two K2-18 HARPS tables under 85 fixed noise models

We compare two published periods for the inner K2-18 velocity signal. The inputs are the public 2015--2018 HARPS tables with 100 line-by-line velocities and 106 NAIRA velocities. All rows are retained. Each of four observing seasons is omitted once. Generalized least squares fits the other seasons. A joint Gaussian prediction score includes mean-coefficient uncertainty and a fixed correlated-noise model. The 9.2072-day period scores above 8.997 days by 10.19 nats in the first table and 5.98 nats in the second. Three seasons favor the longer period in each table. The 2016 season favors the shorter period. All 85 stated noise cases give positive total differences in both complete tables. The ranges are 3.64--28.89 and 3.29--12.01 nats. Controls use 99 matched exposures, a trend, a harmonic, and periods fitted within fixed local intervals to training data only. Their totals also favor the longer mode. A separate matrix identity checks every primary and mean-control fold. Simulations test score behavior at the measured dates. The package preserves unchanged numeric inputs, hashes, code, predictions, all noise cases, and figures. This adds a whole-season prediction audit to prior full-data studies. The two periods and their earlier sampling discrepancy are already known. The choices are retrospective and draw on the same observations. The two reductions share spectra. This conditional result is not an independent planet validation, a Bayes factor, or a false-alarm probability.

K2-18 · radial velocity · HARPS · period modes · season prediction · cross-validation · correlated noise · reproducible analysis

Admitted 1 citing publication

References

6
  1. usesdoi:10.1093/mnras/stac3024

    Provides the public LBL HARPS table, the longer period, the prior explanation for the period discrepancy, and context for the fixed noise settings. This audit uses all 100 public rows and is not a reproduction of the clipped full orbital fit.

  2. usesdoi:10.1051/0004-6361/201833995

    Provides the public 106-row earlier HARPS table and the 8.997-day comparison period.

  3. doi:10.1051/0004-6361/201731558

    Reports the earlier inner signal and forward time-series cross-validation of planet counts. The present test instead compares two specified period modes by omitting whole seasons.

  4. doi:10.3847/1538-3881/aac108

    Reports time and wavelength changes of the signal near 9 days in CARMENES data and favors a stellar activity origin. This HARPS period-mode audit does not settle that physical-origin question.

  5. usesA 41.78-day term raises the total prediction score for omitted years in 204 GJ 3998 velocities under fixed noise sci:sha256:0d775d70c7d413ecd2f5a280e8e98014994b6345eb7aa75d35af585f02c93aac

    Informs the joint prediction score, coefficient uncertainty, and fixed noise checks. This contribution uses new code and K2-18 data. No source code is copied.

  6. usesdoi:10.26093/cds/vizier

    CDS VizieR distributes the unchanged numeric inputs in catalogs J/MNRAS/517/5050 and J/A+A/621/A49.

Cited by

1 publication

Public catalog entries on Sciscape.