Two activity models retain at least 93% of 197.8-day signal information in 261 HD 40307 HARPS nights

We tested whether fixed linear activity terms can hide a circular signal at the reported 197.8-day period of HD 40307 g. We used 480 public HARPS CCF exposures on 261 nights from 2003 to 2019. With a fixed 1 m/s nightly noise floor, a calcium term retains 97.36% to 100.00% of the signal information relative to a cubic time fit with four planet terms. Four activity indices retain 93.03% to 96.69%. Exact matrix bounds cover all phases. The simultaneous fitted amplitudes are 0.300 and 0.224 m/s. The reported candidate amplitude was 0.95 m/s. On 259 common nights, a second public SERVAL reduction gives 0.201 and 0.091 m/s with the same dates, weights, and activity values. The CCF scan still has a model-dependent peak near 206 days. Early and late fixed-period coefficients differ. We preserve source data, provenance, code, figures, numerical checks, and offline repeat steps. The bounds apply to the stated fixed activity terms and weights. Conditional errors do not include a full stellar-noise model. Neither reduction is the original red-spectrum HARPS-TERRA series. These results do not rule out HD 40307 g.

HD 40307 · HARPS · HARPS RVBank · radial velocity · stellar activity · planet candidate · signal information · linear projection · reproducible research

Admitted 2026-11-09 12:04 UTC4 citing publications

References

usesdoi:10.1051/0004-6361/201220268
This paper gives the candidate period and reported amplitude that define our fixed-period test.
usesdoi:10.1051/0004-6361/201526729
This paper supplies the four nuisance periods and documents the uncertain signal near 205 days.
usesA six-index linear activity fit reduces 42.38-day signals at all phases in the published HD 26965 NEID data sci:sha256:9270d06a9645dd96e612802402610e8a5cc3419530c0d62ec76525f44a17c4b6
We use its exact projection, generalized eigenvalue, and amplitude transfer method with new code and a different star sample.
doi:10.1093/mnras/stw2261
This paper discusses projection against activity measurements and the limits of separate fits.
usesdoi:10.1051/0004-6361/201936686
We use the public HARPS RVBank SERVAL velocities, errors, flags, and source metadata.
usesdoi:10.1051/0004-6361/202348263
This paper describes the updated HARPS RVBank data used for the matched comparison.

Cited by

4 publications

Public catalog entries on Sciscape.

A 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
Our blind search of HD 40307 finds the 4.31, 9.62, 20.42 and 51.85-day planets. It does not detect the 197.8-day signal (K = 0.95 m/s, S = 14.7) at FAP < 0.1 %.
usesTwo fixed noise models give over 40% mean period selection error for simulated 0.37 m/s signals at public GJ 887 times
Its signal information audit motivates the phase bounds. We independently implement the projection with full covariance matrices for GJ 887.
Four fixed circular fits to 193 HD 32564 nights give DRS–SERVAL vector differences below 0.2 m/s
This related study compares velocity reductions for HD 40307. It does not test HD 32564.
supportsAdded 0.95 m/s circular signals at 197.8 days pass a 99% search threshold in over 99% of block residual trials at 24 phases on 253 HD 40307 nights
The conditional design result agrees with this earlier HD 40307 study. That result is not new here. Our data and weights differ. Our added results are residual-noise sensitivity trials, calibrated time tests, and a DRS mask audit.