2
5
10
20
50
0
5
10
15
20
25
∆
ln L
2
5
10
20
50
Period [days]
0
5
10
15
20
∆
ln L
2
5
10
20
50
Period [days]
0
10
20
30
∆
ln L
ln-Posterior
ln-Likelihood
HARPS+UVES pre-2016
HARPS PRD
0.1% FAP
1% FAP
ALL
10% FAP
a
b
c
Figure 1: Detection of a Doppler signal at 11.2 days. Detection periodograms of the 11.2 day
signal in the HARPS+UVES pre-2016 data (panel a), and using the HARPS Pale Red Dot cam-
paign only (panel b). Panel c contains the periodogram obtained after combining all datasets. Black
lines correspond to the
∆ ln L statistic, while the gray thick represent the logarithm of the Bayesian
posterior density (see text, arbitrary vertical offset applied to for visual comparison of the two statis-
tics). The horizontal solid, dashed and dotted lines represent a 10, 1, and 0.1 per cent false alarm
probability thresholds of the frequentist analysis, respectively.
each campaign and the photometry are detailed in the methods section. All time-series used in this
work in the online version of the paper as Source data.
The search and significance assessment of signals were performed using frequentist
14
and Bayesian
15
methods. Periodograms in Figure 1 represent the improvement of some reference statistic as a func-
tion of trial period, with the peaks representing the most probable new signals. The improvement in
the logarithm of the likelihood function
∆ ln L is the reference statistic used in the frequentist frame-
work, and its value is then used to assess the false-alarm probability (or FAP) of the detection.
14
A
FAP below 1% is considered suggestive of periodic variability, and anything below 0.1% is con-
sidered to be a significant detection. In the Bayesian framework, signals are first searched using a
specialized sampling method
16
that enables exploration of multiple local maxima of the posterior
density (the result of this process are the red lines in Figure 1), and significances are then assessed
by obtaining the ratios of evidences of models. If the evidence ratio between two models exceeds
some threshold (e.g.
B
1
/B
0
> 10
3
), then the model in the numerator (with one planet) is favoured
against the model in the denominator (no planet).
A well isolated peak at ∼11.2 days was recovered when analyzing all the night averages in the
pre-2016 datasets (Figure 1, panel a). Despite the significance of the signal, the analysis of pre-
2016 subsets produced slightly different periods depending on the noise assumptions and which
subsets were considered. Confirmation or refutation of this signal at 11.2 days was the main driver
for proposing the HARPS PRD campaign. The analysis of the HARPS PRD data revealed a single
significant signal at the same ∼
11.3 ± 0.1 day period (Figure 1, panel b), but period coincidence
3
0
2
4
6
8
10
Phase [days]
-4
-2
0
2
4
6
8
RV [m/s]
UVES
HARPS pre-2016
HARPS PRD
Figure 2: All datasets folded to the 11.2 days signal. Radial velocity measurements phase folded
at the 11.2 day period of the planet candidate for 16 years of observations. Although its nature is
unclear, a second signal at P∼ 200 days was fitted and subtracted from the data to produce this plot
and improve visualization. Circles correspond to HARPS PRD, triangles are HARPS pre-2016 and
squares are UVES. The black line represents the best Keplerian fit to this phase folded representation
of the data. Error bars correspond to formal 1-
σ uncertainties.
alone does not prove consistency with the pre-2016 data. Final confirmation is achieved when all the
sets were combined (Figure 1, panel c). In this case statistical significance of the signal at 11.2 days
increases dramatically (false-alarm probability
< 10
−7
, Bayesian evidence ratio
B
1
,0
> 10
6
). This
implies that not only the period, but also the amplitude and phase are consistent during the 16 years
of accumulated observations (see Figure 2). All analyses performed with and without correlated-
noise models produced consistent results. A second signal in the range of 60 to 500 days was also
detected, but its nature is still unclear due to stellar activity and inadequate sampling.
Stellar variability can cause spurious Doppler signals that mimic planetary candidates, especially
when combined with uneven sampling.
9, 17
To address this, the time-series of the photometry and
spectroscopic activity indices were also searched for signals. After removing occasional flares, all
four photometric time-series show the same clear modulation over
P ∼ 80 nights (panels b, c, d
and e in Figure 3), which is consistent with the previously reported photometric period of ∼83 d.
3
Spectroscopic activity indices were measured on all HARPS spectra, and their time-series were in-
vestigated as well. The width of the spectral lines (measured as the variance of the mean line, or
m
2
) follows a time-dependence almost identical to the light curves, a behaviour that has already
4