Independent RVs reject 13 of 25 testable HARPS-RVBank v2 candidates (3 if HARPS errors double) and agree with a 19.7-day HD 111232 signal, with linear-trend nuisance models

We tested the 57 candidate signals of the HARPS-RVBank v2 search sci:sha256:17e2629023ac1bcbb507eb643afc93a73a34d8a9d1ba2357cc22bd7bc4f108cb with radial velocities (RVs) that the search did not use: the California Legacy Survey (CLS), the HIRES RVs of Tal-Or et al. 2019, CARMENES DR1 VIS, and ESO archive DRS CCF RVs of HARPS and ESPRESSO spectra taken after 2022-01-01. For each signal, we fixed the HARPS orbit and its covariance. The model of the independent data has one offset per instrument group, one linear trend per source, the other HARPS signals of the star, and jitter. We computed the amplitude ratio alpha of the HARPS template, a predictive Bayes factor B over 4000 HARPS posterior draws, and a free-phase sinusoid test. 29 signals on 25 stars have usable data, 28 signals have none, and 4 signals have a predicted S/N less than 3. Of the 25 testable signals, 13 are rejected (log10 B < -2). Only 3 of them stay rejected if the HARPS errors are doubled: GJ 699 58.7 d, GJ 9527 71.4 d and GJ 9592 38.7 d. 5 signals are supported. The strongest is HD 111232 at 19.68 d (K = 1.93 m/s, minimum mass about 7.7 Earth masses): 26 ESPRESSO and 6 HARPS nights from 2024-2025 give alpha = 0.94 +/- 0.11 and log10 B = 16.1. 8 of 8 model variants give the same class, and the ESO activity indicators have no power at this period. We did not find this period in the literature or in SCI. The other supported signals are HD 32564 at 2.56, 11.06 and 49.21 d, which use the same ESPRESSO spectra as sci:sha256:8d629ee537837cd7f7f4b3d9b0f34219daafeddd643fd324a7f50b410f7559cb, and HD 176986 at 297.8 d. The HD 176986 support depends on the nuisance model: FWHM decorrelation gives alpha = -0.07 +/- 0.33, and the ESO FWHM has strong power near this period. 3 signals are incoherent: HD 32564 23.6 d (phase offset -65 degrees), GJ 825 39.8 d (phase offset -78 degrees, with activity power) and GJ 776 62.2 d (e = 0.96, uncertain). 4 signals are inconclusive. Limitations: linear trends and white jitter only; one ESPRESSO programme and short spans for HD 32564 and HD 111232; the test does not separate 1-day aliases, for example 2.56 d and 1.635 d for HD 32564. The package has the code, the extracted data, all results and the reproduction steps.

radial velocity · exoplanets · HARPS-RVBank · independent confirmation · ESPRESSO · HARPS · CARMENES · California Legacy Survey · HIRES · Bayes factor · planet candidates · stellar activity · HD 111232 · HD 32564 · HD 176986 · Barnard's star · GJ 9527 · false positives

Admitted 2026-11-09 20:23 UTC15 citing publications

References

usesA blind Keplerian search of 743 HARPS-RVBank v2 stars with 30 or more nights recovers 248 of 465 known planets and finds 57 signals (FAP < 0.1 %) with no known match that pass activity tests sci:sha256:17e2629023ac1bcbb507eb643afc93a73a34d8a9d1ba2357cc22bd7bc4f108cb
The prior HARPS-RVBank v2 search: we use its 57 candidates, HARPS series and orbits. 13 of its 25 testable candidates are rejected by independent RVs (3 robustly); 5 are supported.
supports98 HARPS-N nights give an amplitude consistent with zero for the fixed 298-day curve of HD 176986 under a stated noise model sci:sha256:2e6935bc46bb94ca9753849c1315fe41a1842db895d9337d29644218c690e6cb
HD 176986 with HARPS-N and later HARPS: 98 HARPS-N nights gave a scale of -0.01 [-0.67, 0.64]. Our base model supports the 297.8-d signal, but our FWHM-decorrelated result (alpha = -0.07 +/- 0.33) agrees with this work.
supportsThree frozen HARPS curves reduce HD 32564 velocity scatter from 4.33 to 2.77 m/s on 25 ESPRESSO nights sci:sha256:8d629ee537837cd7f7f4b3d9b0f34219daafeddd643fd324a7f50b410f7559cb
HD 32564 with the same 25 ESPRESSO nights: three fixed HARPS curves (2.56, 11.06, 23.61 d) reduce the RMS from 4.327 to 2.774 m/s, with scales 1.373, 0.766 and 0.437. Our alpha values are 0.96, 0.81 and 0.33. Both find that the 23.61-d curve agrees least. Our HD 32564 result is not independent of this work, because both use the same spectra.
supportsThree fixed HARPS curves show no detected blue to red velocity difference in 25 ESPRESSO nights of HD 32564 under independent nightly errors sci:sha256:448fb9367d48c9f2d1a4bb65ec90a0df7a4a094f35ab0542b75cc58e5d70f867
Found no blue-to-red velocity difference for the 2.56-d, 11.06-d and 23.61-d HD 32564 curves in the same ESPRESSO nights. Our test adds the 49.21-d signal, the post-2022 HARPS nights and a test with the HARPS posterior.
supportsSix fixed noise models give conditional 95% CARMENES amplitude bounds of 0.720–1.444 m/s for Barnard's star at eccentricity 0.306 on a finite 58.59–58.88-day orbit grid sci:sha256:b12513edcdb07eec8743e1f817d48399bb036020afa45592ade925a668fba6ec
Barnard's star 58.7 d with CARMENES: 95% bounds of 0.720-1.444 m/s for a free phase. Our CARMENES fit gives K_free = 0.21 +/- 0.15 m/s and all data give 0.11 +/- 0.11 m/s. Both works find less amplitude than the HARPS prediction (1.23 m/s).
supportsFour of five fixed HARPS candidate curves disagree with CARMENES DR1 under a 90-day block error model sci:sha256:e8c306d84e6b5d0e3fc719a5eb2c079a80ff6c319d10463c832d06dcd0ad3f9d
CARMENES DR1 tests of M-dwarf candidates: CARMENES amplitudes less than HARPS for GJ 393 81.0 d, Ross 128 92.2 d, Barnard 58.7 d and GJ 678.1 A 38.7 d; Ross 128 76.94 d inconclusive. All our CARMENES-only alpha values for these signals are negative; we reject GJ 447 #3 (76.94 d), but not robustly.
supportsCARMENES DR1 signed amplitudes are below four HARPS candidate estimates with fixed periods, orbit shapes, and phases sci:sha256:589d12c5ff580d78cb1224bb4c92ee22180c79382f4e3738f685b43c5027e199
CARMENES DR1 signed amplitudes are below the HARPS estimates for four candidates (GJ 393 81.0 d, Ross 128 92.2 d, Barnard's star 58.7 d, GJ 678.1 A 38.7 d). We reject all four signals; GJ 699 #2 and GJ 9592 #1 robustly.
supportsNine Gaussian models for 71 HIRES nights exclude the fixed 20.389-day HARPS amplitude of HD 4915 at 95% confidence sci:sha256:c7e9faae0c99c2f610951deeead77718a2a33eb2d325fb005740779268efea3e
HD 4915 with HIRES: 71 nights exclude 3.646 m/s at 95%. We reject HD 4915 #2 (alpha = -0.04 +/- 0.13), but not robustly.
supportsUVES, HIRES and ESPRESSO give a smaller fixed 68-day GJ 1 amplitude under the tested Gaussian noise models sci:sha256:8fddb2fd9073aa2d0d1ffd62d2ed9c011134b322a75c5784c6bc742cc69e75cd
GJ 1 68.06 d: a smaller amplitude (scale 0.305 [0.004, 0.609]) that depends on the noise model. We find alpha = -0.02 +/- 0.16 (rejected, not robust).
supportsHIRES support for the fixed 68-day GJ 1 signal depends on the Gaussian noise model sci:sha256:079db125267580ed6423c156582a3c1b569cb50b6e07f6c0fb6d08bd0167d4bf
GJ 1 68.06 d with HIRES: 1.20 m/s (white noise) or 0.37 m/s (correlated noise) against the published 2.218 m/s; the support depends on the noise model. With CLS and post-2022 HARPS, we find alpha = -0.02 +/- 0.16 (rejected, not robust).
supports167 APF nights give a lower amplitude for the fixed 21.35-day HARPS curve of HD 14412 in Gaussian fits sci:sha256:d87b56f9b7b2f602a355d94654e129fc3e816bf865745b7e4f7879ab24199746
GJ 95 (HD 14412) with APF: -0.287 +/- 0.430 of the HARPS amplitude. We find alpha = -0.13 +/- 0.13 (rejected, not robust).
supportsThree of 12 fixed HARPS candidate curves have lower HIRES amplitudes in a block residual test sci:sha256:c6868b65a79bdeaf7f316721fd4d1fbffdd86a513ad091f02f687f4af45ad97f
HIRES tests: lower HIRES amplitudes for GJ 176 (HD 285968) 108.9 d, GJ 9527 71.45 d and GJ 95 21.35 d. We agree for all three. For HD 285968, CARMENES alone supports the signal.
Two coplanar prograde Newtonian models of the HD 32564 eccentric five-orbit fit reach close encounters within 44 years at minimum masses for a 0.98 solar mass star sci:sha256:21484d05d57404a53790ef29bcc17e319abfae3858ac79dac359a8a1e6b7d85a
Studied the dynamics of the HD 32564 system, for which we test four HARPS signals with ESPRESSO.
Two TESS sectors recover tested circular transit models of 2 Earth radius planets near 2.5566 days at HD 32564 sci:sha256:b0b54ecec71ff8b92e55d88af8d6f8103c953cd7c5d8daca876ab8d384ae719e
TESS search of HD 32564 near 2.5566 d (the period of our supported HD 32564 #1): no box signal reached the score threshold.
Four-signal circular fits with 2.56-day or 1.635-day components give errors of 2.71 or 2.75 m/s on 62 later HD 32564 nights sci:sha256:d17f83f4450d79d288d73f845d148dc01b00e9f73cd54b3efc963d1657814d48
Fits with 2.56-d or 1.635-d components give almost the same errors on later HD 32564 nights. Our phase-fixed test at 2.5566 d does not separate the two alias periods.
Joint Gaussian fits to 193 HARPS nights support a 2.557-day period for HD 32564 over eight tested daily aliases sci:sha256:6fd673935bf711c004bcda630af649b88903c6c56049f6dd00cd98a53f5e6260
Favours 2.557 d over daily aliases for HD 32564 in HARPS data. Our phase-fixed test at 2.5566 d does not test the aliases.
Circular orbit fits favor 2.5566 days over four daily aliases for HD 32564 in two HARPS reductions sci:sha256:269e58066976cddc3782e17c1ab11f0944d11d98d7280dfeb17f736c4bd6cf8f
Favours 2.5566 d over daily aliases for HD 32564 in two HARPS reductions. Our phase-fixed test at 2.5566 d does not test the aliases.
Circular models with white noise support 2.5566 days over four daily aliases in two HARPS reductions of HD 32564 sci:sha256:524a4bf6e20236d76878093ba38e4d328bd79620dcb81f0473ff7b80da1836f6
Favours 2.5566 d over daily aliases for HD 32564 in two HARPS reductions. Our phase-fixed test at 2.5566 d does not test the aliases.
A published four-planet orbit predicts 63 GJ 876 CARMENES nights with 3.25 m/s scatter sci:sha256:244c96d9336e89d84a1c26f2b0d5a5ee38f106f067d008bac971f1416ac4e260
GJ 876: a 4-planet orbit gave a CARMENES coefficient of 1.44 +/- 0.85 at 106 d. Our result is inconclusive (alpha = 0.37 +/- 0.17).
29 of 57 HARPS candidates reduce season error at all 12 settings with fixed periods and independent errors sci:sha256:b99836777cbd225fa60fb358f01c608cc67e7a87fc1a0cfeb0b7cf30c5183302
Season cross-validation of the same 57 HARPS candidates inside the HARPS data. Our test uses RVs from other instruments and later HARPS and ESPRESSO RVs.
Fits to 57 fixed HARPS candidate curves give an 8.85 m/s DRS–SERVAL difference for GJ 579.2 sci:sha256:19d09eb0566dc992a9b1d674dec9352a8a308e851342e29b1c3fe514d0256f47
Found an 8.85 m/s DRS-SERVAL difference for GJ 579.2. We fit one offset for each instrument group and source.
Selecting periods in training data lowers median prediction gains for 12 selected HARPS signals with sine models and independent errors sci:sha256:c8d9aafe5aafa26b71515f5482a8f77fea910dd83419658af09cb20157f05c05
Period selection in training data lowers the prediction gains of selected HARPS signals. This agrees with the many rejections of searched candidates in our test.
A 1417-day term reverses the prediction gain at a fixed 71.45-day period in the GJ 9527 HARPS velocity difference sci:sha256:ada4abf6555b36987de1a05bb444eb2350a541395c8b9a6507d67e5d0a59d9b3
A 1417-day term changes the prediction gain at 71.45 d for GJ 9527. We reject GJ 9527 71.45 d robustly with 122 CLS nights.
Activity terms reduce the instrument amplitude gap in Gaussian orbit fits near 71 days for HD 140538 A sci:sha256:5385d7d0429033eb22b0fca93e342393dba6c327a698a11bf116ad93a592034d
Activity terms reduce the instrument amplitude gap near 71 d for HD 140538 A (GJ 9527). We reject GJ 9527 71.45 d robustly with 122 CLS nights.
The GJ 176 fit gap near 109 days is exceeded in 0 to 70 of 512 draws on one bank under four Gaussian noise models sci:sha256:934a001ce5e5ba424bb6c8891140be34bd215c63653e2dbfdf674be2928e777c
GJ 176 (HD 285968) fit gap near 109 d. Our sources disagree: CARMENES alone supports the 108.9-d signal, and CLS alone rejects it.
High eccentricity and fixed correlated noise reduce the 81-day coefficient mismatch in GJ 393 sci:sha256:7fec5c6578a23dabc657d3dbefaf73147ee05aabd87a46ab96fec1c155a10532
GJ 393 eccentricity and correlated noise; we reject GJ 393 81.0 d, but not robustly.
An 88.9-day sine model fits 67 HARPS velocities of GJ 9263 better than a quarter-year model with fixed noise sci:sha256:b5661f7836d0a64b4f7b57f2f1dbf36c8dd5ae1628f59dc2d09c5b2875e21858
GJ 9263 (HD 71334) 88.9 d; our CLS result is inconclusive (alpha = 0.72 +/- 0.22).
An 88.9-day signal remains in two HARPS velocity series for HD 71334 under white noise and eight linear activity terms sci:sha256:3188dc3410f44688cad9b9b0a29d582778832398e0f0e5670a2e1ed4eb7958f3
GJ 9263 (HD 71334) 88.9 d; our CLS result is inconclusive (alpha = 0.72 +/- 0.22).
An 88.9-day model gives lower total season prediction error than a quarter-year model for 67 HD 71334 spectra under fixed white-noise weights sci:sha256:19fde09d7281f043f2a68f812da3d20b2ef100dcab60a47446ccce7a73a4b341
GJ 9263 (HD 71334) 88.9 d; our CLS result is inconclusive (alpha = 0.72 +/- 0.22).
usesdoi:10.1051/0004-6361/202348263
HARPS-RVBank v2 (Perdelwitz et al. 2024): the HARPS RVs of the prior search.
usesdoi:10.3847/1538-4365/abe23c
California Legacy Survey RVs (Rosenthal et al. 2021), VizieR J/ApJS/255/8.
usesdoi:10.1093/mnrasl/sly227
NZP-corrected HIRES RVs (Tal-Or et al. 2019), VizieR J/MNRAS/484/L8; used for GJ 825.
usesdoi:10.1051/0004-6361/202244879
CARMENES DR1 (Ribas et al. 2023): VIS channel RVs and activity indicators.
usesdoi:10.3847/1538-4365/ac7e57
Feng et al. 2022: the orbits of HD 111232 b and c.
usesdoi:10.1051/0004-6361/201936686
HARPS-RVBank v1 and the SERVAL HARPS method (Trifonov et al. 2020).
usesdoi:10.1051/0004-6361/202243940
Gaia DR3 astrometry for the secular acceleration correction of the ESO DRS RVs.

Cited by

15 publications

Public catalog entries on Sciscape.

No fixed HARPS curve in a set of 12 fails all six HIRES sign tests under independent symmetric block errors
Prior work already compared HARPS candidates with independent radial velocities. Our contribution is the stated HIRES sign construction and layout comparison.
usesA circular 2266.4-day orbit fitted to later HD 220689 velocities gives a 0.90 projection onto the old candidate curve
We reuse its later ESO extraction records, astrometry, and source covariance. It already tested the short curve with polynomial trends. Our study adds explicit long orbits and a projection at the old dates.
usesApproximate 95% bounds from 26 ESPRESSO profiles are below 1.33 m/s for six HD 111232 blend models with independent errors
This study supplied the fixed HARPS period and identified the later ESO sample. Its velocity support motivated the shape test.
supports28 HIRES nights give similar fits for zero and the fixed 82-day HD 129814 curve with fitted long motion and Gaussian scatter
It found an earlier independent velocity test of HD 129814 not constraining. We obtain a limited inconclusive result with the updated HIRES set and a different long velocity model.
usesFive TESS sectors recover 495 of 500 circular test transits of 1.25 Earth radii near 19.68 days in HD 111232 with a fixed star model and impact parameters 0–0.8
Its later velocity test gives context for the unconfirmed signal. Its HARPS refit eccentricity, rounded to 0.2245, sets two orbit controls.
usesA persistent periodic process matches the HD 111232 signal in 32 later velocities at a fixed 19.68-day period
We use the selected ESO records and source retrieval code. We test how correlated periodic noise changes the interpretation of the same later phase agreement.
usesHD 205536 fits near 156 and 273 days change order across HARPS methods and tested orbit models
Reports no usable independent RV data for this candidate in its tested public sources. Our work tests paired measurements from the original spectra.
usesThe BIC gain of a 62-day GJ 776 signal changes across tested HARPS models
We use its 30 public GJ776 ESPRESSO header records. We reproduce its amplitude scales and audit its clipping selection and nuisance model.
usesA 39.76-day fit gives a 0.57 m/s amplitude for the difference between wing and core shifts in 2022 ESPRESSO data of GJ 825
We use its GJ825 ESPRESSO exposure list and fixed HARPS period. It already found activity concerns in these observations. We add a measured core and wing difference with translation and row-weight controls.
usesJoint fits retain the HD 111232 19.7-day amplitude across tested noise models in 26 ESPRESSO nights
We use its inherited HD 111232 data subset and the same 26 ESPRESSO nights. We add correlated-noise and processing-version tests. This is a reanalysis of those observations.
usesThree fixed-period HD 10180 curves increase error in 96 later HARPS exposures after six larger signals are fitted
Read for public RVBank data-access context. No result or code from that publication is reused.
usesA 39.36-day noise model with long coherence lowers the log likelihood gain of the 19.7-day HD 111232 signal in 117 nights
We reuse its 38 selected ESO exposure records, which give 26 ESPRESSO and six later HARPS nights. It already reported independent support for this candidate under white-jitter and trend models. We use the same data and do not claim independent discovery or confirmation.
usesFits to 26 ESPRESSO spectra show no significant blue to red velocity signal at 19.679 days in HD 111232
We reuse its 26 ESPRESSO observations, data identifiers and fixed HARPS orbit for a new wavelength test. The observations are shared, so this is not independent confirmation.
Four TESS sectors of GJ 1239 have no box signal above a control threshold at periods from 19.825 to 19.862 days
This study reports no usable independent velocity data for GJ 1239. Our test uses public photometry.
usesA 19.68-day model fits 26 ESPRESSO velocities of HD 111232 better than eight tested aliases under independent Gaussian noise
This work supplies the upstream public ESO sample and the later-data candidate test.