1: %\documentclass[12pt,preprint]{aastex}

2: %\documentclass[preprint]{aastex}

3: 

4: \documentclass{emulateapj}

5: 

6: %\usepackage{apjfonts}

7: %\usepackage[,longnamesfirst]{natbib}

8: 

9: \newcommand{\Teff}{\mbox{$T_{\rm eff}$}}

10: \newcommand{\Dnu}{\mbox{$\Delta \nu$}}

11: \newcommand{\acena}{\mbox{$\alpha$~Cen~A}}

12: \newcommand{\acenb}{\mbox{$\alpha$~Cen~B}}

13: \newcommand{\acen}{\mbox{$\alpha$~Cen}}

14: \newcommand{\tcet}{\mbox{$\tau$~Cet}}

15: \newcommand{\bhyi}{\mbox{$\beta$~Hyi}}

16: \newcommand{\dpav}{\mbox{$\delta$~Pav}}

17: \newcommand{\bvir}{\mbox{$\beta$~Vir}}

18: \newcommand{\nuind}{\mbox{$\nu$~Ind}}

19: \newcommand{\muara}{\mbox{$\mu$~Ara}}

20: \newcommand{\cms}{\mbox{cm\,s$^{-1}$}}

21: \newcommand{\eboo}{\mbox{$\eta$~Boo}}

22: \newcommand{\ms}{\mbox{m\,s$^{-1}$}}

23: \newcommand{\kms}{\mbox{km\,s$^{-1}$}}

24: \newcommand{\muHz}{\mbox{$\mu$Hz}}

25: %% \newcommand{\mynote}[1]{{\bf [Note: \it#1\bf]}}

26: \newcommand{\mynote}[1]{{\bf [#1]}}

27: \newcommand{\new}[1]{{\bf #1}}

28: \renewcommand{\new}[1]{\relax #1}

29: %\let\epsilon\varepsilon

30: \newcommand{\half}{{\textstyle\frac{1}{2}}}

31: \newcommand{\sixth}{{\textstyle\frac{1}{6}}}

32: \newcommand{\tenth}{{\textstyle\frac{1}{10}}}

33: 

34: \slugcomment{to appear in ApJ}

35: 

36: \shorttitle{Oscillations in Procyon. I. Observations}

37: \shortauthors{Arentoft et al.}

38: 

39: 

40: \begin{document}

41: 

42: \title{

43: A multi-site campaign to measure solar-like oscillations in Procyon.\\

44: I. Observations, Data Reduction and Slow Variations}

45: 

46: \author{

47: Torben Arentoft,\altaffilmark{1}

48: Hans~Kjeldsen,\altaffilmark{1}

49: Timothy~R.~Bedding,\altaffilmark{2}

50: \\

51: % HARPS

52: Micha\"el Bazot,\altaffilmark{1,3}

53: J{\o}rgen~Christensen-Dalsgaard,\altaffilmark{1}

54: Thomas~H.~Dall,\altaffilmark{4}

55: Christoffer Karoff,\altaffilmark{1}

56: \\ % CORALIE

57: Fabien~Carrier,\altaffilmark{5}

58: Patrick~Eggenberger,\altaffilmark{6}

59: Danuta Sosnowska,\altaffilmark{7}

60: \\ % McDonald

61: Robert A. Wittenmyer,\altaffilmark{8}

62: Michael Endl,\altaffilmark{8}

63: Travis S. Metcalfe,\altaffilmark{9}

64: \\ % Lick

65: Saskia Hekker,\altaffilmark{10,11}

66: Sabine Reffert,\altaffilmark{12}

67: \\ % UCLES

68: R.~Paul~Butler,\altaffilmark{13}

69: Hans~Bruntt,\altaffilmark{2}

70: L\'aszl\'o~L.~Kiss,\altaffilmark{2}

71: Simon~J.~O'Toole,\altaffilmark{14}

72: \\ % Okayama

73: Eiji Kambe,\altaffilmark{15}

74: Hiroyasu Ando,\altaffilmark{16}

75: Hideyuki Izumiura,\altaffilmark{15}

76: Bun'ei Sato,\altaffilmark{17}

77: \\ % Tautenburg

78: Michael Hartmann,\altaffilmark{18}

79: Artie Hatzes,\altaffilmark{18}

80: \\ % SOPHIE

81: Francois~Bouchy,\altaffilmark{19}

82: Benoit~Mosser,\altaffilmark{20}

83: Thierry~Appourchaux,\altaffilmark{21}

84: Caroline~Barban,\altaffilmark{20}

85: Gabrielle~Berthomieu,\altaffilmark{22} %?

86: Rafael~A.~Garcia,\altaffilmark{23}

87: Eric~Michel,\altaffilmark{20}

88: Janine~Provost,\altaffilmark{22} %?

89: Sylvaine~Turck-Chi\`eze,\altaffilmark{23}

90: \\ % EMILIE

91: Milena~Marti{\'c},\altaffilmark{24}

92: Jean-Claude~Lebrun,\altaffilmark{24}

93: Jerome~Schmitt,\altaffilmark{25}

94: Jean-Loup~Bertaux,\altaffilmark{24}

95: \\ % SARG

96: Alfio~Bonanno,\altaffilmark{26}

97: Serena~Benatti,\altaffilmark{27}

98: Riccardo~U.~Claudi,\altaffilmark{27}

99: Rosario~Cosentino,\altaffilmark{26}

100: Silvio~Leccia,\altaffilmark{28}

101: \\ % NOT/FIES

102: S{\o}ren~Frandsen,\altaffilmark{1}

103: Karsten~Brogaard,\altaffilmark{1}

104: Lars Glowienka,\altaffilmark{1}

105: Frank~Grundahl\altaffilmark{1} and

106: Eric~Stempels\altaffilmark{29}

107: }

108: 

109: \altaffiltext{1}

110: {Danish AsteroSeismology Centre (DASC), Department of Physics and

111: Astronomy, University of Aarhus, DK-8000 Aarhus C, Denmark;

112: toar@phys.au.dk, hans@phys.au.dk, jcd@phys.au.dk, karoff@phys.au.dk, srf@phys.au.dk,

113: kfb@phys.au.dk, f002769@phys.au.dk, fgj@phys.au.dk}

114: 

115: \altaffiltext{2}

116: {Institute of Astronomy, School of Physics, University of Sydney, NSW 2006, Australia;

117: bedding@physics.usyd.edu.au, bruntt@physics.usyd.edu.au,

118: laszlo@physics.usyd.edu.au}

119: 

120: \altaffiltext{3}

121: {Centro de Astrof{\'\i}sica da

122: Universidade do Porto, Rua das Estrelas, 4150-762 Porto, Portugal;

123: bazot@astro.up.pt}

124: 

125: \altaffiltext{4}

126: {Gemini Observatory, 670 N. A'ohoku Pl., Hilo, HI 96720, USA;

127: tdall@gemini.edu}

128: 

129: \altaffiltext{5}

130: {Instituut voor Sterrenkunde, Katholieke Universiteit Leuven,

131: Celestijnenlaan 200 B, 3001 Leuven, Belgium; fabien@ster.kuleuven.be}

132: 

133: \altaffiltext{6}

134: {Observatoire de Gen\`eve, Ch.~des Maillettes 51, CH-1290 Sauverny,

135: Switzerland; patrick.eggenberger@obs.unige.ch}

136: 

137: \altaffiltext{7}

138: {Laboratoire d'astrophysique, EPFL Observatoire CH-1290 Versoix;

139: danuta.sosnowska@epfl.ch}

140: 

141: \altaffiltext{8}

142: {McDonald Observatory, University of Texas at Austin, Austin, TX 78712,

143: USA; robw@astro.as.utexas.edu, mike@astro.as.utexas.edu}

144: 

145: \altaffiltext{9}

146: {High Altitude Observatory, National Centre for Atmospheric Research,

147: Boulder, CO 80307-3000 USA; travis@ucar.edu}

148: 

149: \altaffiltext{10} 

150: {Leiden Observatory, Leiden University, 2300 RA Leiden, The Netherlands}

151: 

152: \altaffiltext{11} 

153: {Royal Observatory of Belgium, 1180 Brussels, Belgium; saskia@oma.be}

154: 

155: \altaffiltext{12}

156: {ZAH-Landessternwarte, 69117 Heidelberg, Germany;

157: sreffert@lsw.uni-heidelberg.de}

158: 

159: \altaffiltext{13}

160: {Carnegie Institution of Washington, Department of Terrestrial Magnetism,

161: 5241 Broad Branch Road NW, Washington, DC 20015-1305; paul@dtm.ciw.edu}

162: 

163: \altaffiltext{14}

164: {Anglo-Australian Observatory, P.O.\,Box 296, Epping, NSW 1710, Australia;

165: otoole@aaoepp.aao.gov.au}

166: 

167: \altaffiltext{15}

168: {Okayama Astrophysical Observatory, National Astronomical Observatory of

169: Japan, National Institutes of Natural Sciences, 3037-5 Honjyo, Kamogata,

170: Asakuchi, Okayama 719-0232, Japan; kambe@oao.nao.ac.jp,

171: izumiura@oao.nao.ac.jp}

172: 

173: \altaffiltext{16}

174: {National Astronomical Observatory of Japan, National Institutes of Natural

175: Sciences, 2-21-1 Osawa, Mitaka, Tokyo 181-8588, Japan;

176: ando@optik.mtk.nao.ac.jp}

177: 

178: \altaffiltext{17}

179: {Global Edge Institute, Tokyo Institute of Technology

180:  2-12-1-S6-6, Ookayama, Meguro-ku, Tokyo 152-8550, Japan;

181:  sato.b.aa@m.titech.ac.jp}

182: 

183: \altaffiltext{18} 

184: {Th\"uringer Landessternwarte Tautenburg, Sternwarte 5, 07778 Tautenburg,

185: Germany; michael@tls-tautenburg.de, artie@tls-tautenburg.de}

186: 

187: \altaffiltext{19}

188: {Institut d'Astrophysique de Paris, UMR7095, Universit{\'e} Pierre \& Marie

189: Curie, 98$^{bis}$ Bd Arago, 75014 Paris, France;

190: bouchy@iap.fr}

191: 

192: \altaffiltext{20}

193: {Observatoire de Paris, LESIA, UMR 8109, F-92195 Meudon, France;

194: benoit.mosser@obspm.fr, caroline.barban@obspm.fr, eric.michel@obspm.fr}

195: 

196: \altaffiltext{21}

197: {Institut d'Astrophysique Spatiale, Universit{\'e} Paris XI-CNRS,

198:   B{\^a}timent 121, 91405 Orsay cedex, France; Thierry.Appourchaux@ias.u-psud.fr} 

199: 

200: \altaffiltext{22}

201: {Laboratoire Cassiop{\'e}e, UMR CNRS 6202, Observatoire de la C{\^o}te

202:   d'Azur, BP 4229, 06304 Nice cedex 4, France;

203:   Gabrielle.Berthomieu@obs-nice.fr, Janine.Provost@obs-nice.fr}

204: 

205: \altaffiltext{23} 

206: {DAPNIA/DSM/Service d'Astrophysique, CEA/Saclay, 91191 Gif-sur-Yvette

207: Cedex, France; rafael.garcia@cea.fr, sylvaine.turck-chieze@cea.fr}

208: 

209: \altaffiltext{24}

210: {Service d'A{\'e}ronomie du CNRS, BP 3, 91371 Verri{\`e}res le Buisson,

211: France; milena.martic@aerov.jussieu.fr,

212: jean-claude.lebrun@aerov.jussieu.fr,  jean-loup.bertaux@aerov.jussieu.fr}

213: 

214: \altaffiltext{25}

215: {Observatoire de Haute Provence, 04870 St Michel l'Observatoire, France;

216: jerome.schmitt@oamp.fr}

217: 

218: \altaffiltext{26}

219: {INAF - Osservatorio Astrofisico di Catania, via S. Sofia 78, 95123

220: Catania, Italy; abo@oact.inaf.it, rco@ct.astro.it}

221: 

222: \altaffiltext{27}

223: {INAF - Astronomical Observatory of Padua, Vicolo Osservatorio 5, 35122

224: Padova, Italy; serena.benatti@oapd.inaf.it, riccardo.claudi@oapd.inaf.it}

225: 

226: \altaffiltext{28}

227: {INAF - Astronomical Observatory of Capodimonte, Salita Moiariello 16,

228: 80131 Napoli, Italy; leccia@na.astro.it}

229: 

230: \altaffiltext{29}

231: {School of Physics \& Astronomy, University of St.\ Andrews, North Haugh,

232: St.\ Andrews KY16 9SS, Scotland; Eric.Stempels@st-andrews.ac.uk}

233: 

234: \begin{abstract} 

235: %% Version: \today.  

236: We have carried out a multi-site campaign to measure

237: oscillations in the F5 star Procyon~A.  We obtained high-precision velocity

238: observations over more than three weeks with eleven telescopes, with almost

239: continuous coverage for the central ten days.  This represents the most

240: extensive campaign so far organized on any solar-type oscillator.  We

241: describe in detail the methods we used for processing and combining the

242: data.  These involved calculating weights for the velocity time series from

243: the measurement uncertainties and adjusting them in order to minimize the

244: noise level of the combined data.  The time series of velocities for

245: Procyon shows the clear signature of oscillations, with a plateau of excess

246: power that is centred at 0.9\,mHz and is broader than has been seen for

247: other stars.  The mean amplitude of the radial modes is

248: $38.1\pm1.3$\,\cms{} (2.0 times solar), which is consistent with previous

249: detections from the ground and by the WIRE spacecraft, and also with the

250: upper limit set by the MOST spacecraft.  The variation of the amplitude

251: during the observing campaign allows us to estimate the mode lifetime to be

252: $1.5_{-0.8}^{+1.9}$\,d.  We also find a slow variation in the radial

253: velocity of Procyon, with good agreement between different telescopes.

254: These variations are remarkably similar to those seen in the Sun, and we

255: interpret them as being due to rotational modulation from active regions on the

256: stellar surface.  The variations appear to have a period of about 10 days,

257: which presumably equals the stellar rotation period or, perhaps, half of

258: it.  The amount of power in these slow variations indicates that the

259: fractional area of Procyon covered by active regions is slightly higher than for

260: the Sun.

261: \end{abstract}

262: 

263: \keywords{stars: individual (Procyon~A) --- stars:~oscillations}

264: 

265: \section{Introduction}

266: 

267: Measuring solar-like oscillations in main-sequence and subgiant stars

268: requires high-precision observations -- either with spectroscopy or

269: photometry -- combined with coverage that is as continuous as possible.

270: Most of the results have come from high-precision Doppler measurements

271: using ground-based spectrographs, while measurements from spacecraft have

272: also been reported (see \citealt{B+K2007c} and \citealt{AChDC2008} for

273: recent summaries).

274: 

275: Procyon has long been a favourite target for oscillation searches.  At

276: least eight separate velocity studies have reported an excess in the power

277: spectrum, beginning with that by \citet{BGN91}, which was the first report

278: of a solar-like power excess in another star.  For the most recent

279: examples, see \citet{MLA2004}, \citet{ECB2004}, \citet{BMM2004} and

280: \citet{LKB2007}.  These studies agreed on the location of the excess power

281: (around 0.5--1.5\,mHz) but they disagreed on the individual oscillation

282: frequencies.  However, a consensus has emerged that the large separation

283: (the frequency separation between consecutive overtone modes of a given

284: angular degree) is about 55\,\muHz. Evidence for this value was first

285: given by \citet{MMM98} and the first clear detection was made by

286: \citet{MSL99}.

287: 

288: Controversy was generated when photometric observations obtained with the

289: MOST satellite failed to reveal evidence for oscillations

290: \citep{MKG2004,GKR2007,BAB2008}.  However, \citet{BKB2005} argued that the

291: MOST non-detection was consistent with the ground-based data.  Meanwhile,

292: \citet{R+RC2005} suggested that the signature of oscillations is indeed

293: present in the MOST data at a low level (see also \citealt{Mar08}).  Using

294: space-based photometry with the WIRE satellite, \citet{BKB2005b} extracted

295: parameters for the stellar granulation and found evidence for an excess due

296: to oscillations.

297: 

298: All published velocity observations of Procyon have been made from a single

299: site, with the exception of two-site observations by \citet{MLA2004}.  Here

300: we describe a multi-site campaign on Procyon carried out in 2007 January,

301: which was the most extensive velocity campaign so far organized on any

302: solar-type oscillator.  The only other comparable effort to measure

303: oscillations in this type of star was the multi-site photometric campaign

304: of the open cluster M67 \citep{GBK93}.

305: 

306: 

307: \section{Velocity Observations}

308: 

309: We observed Procyon from 2006 December 28 until 2007 January 23, using a

310: total of eleven telescopes at eight observatories.  These are listed in

311: Table~\ref{tab.sites}, ordered westward by longitude.  Note that the FIES

312: spectrograph on the Nordic Optical Telescope was still being commissioned

313: during the observations and the velocity precision is therefore somewhat

314: lower than for the other telescopes.

315: 

316: The team members from each telescope were responsible for producing a

317: velocity time series from the observations, together with estimates of

318: uncertainties.  In six of the spectrographs, the stellar light was passed

319: through an iodine absorption cell to provide a stable wavelength reference.

320: In four others, wavelength calibration was achieved by recording the

321: spectrum from a thorium-argon emission lamp alongside the stellar spectrum,

322: while with EMILIE, exposures of the stellar spectrum were alternated with

323: those of a white-light source passing through an iodine cell.  Details of

324: the methods used with each spectrograph are given in the references listed

325: in Table~\ref{tab.sites} and details of the observations are given in

326: Table~\ref{tab.obs}.

327: 

328: The velocity time series of Procyon is shown in Fig.~\ref{fig.series}{\em

329: a\/}, using a different color for each telescope.  Differences between

330: telescopes in the absolute zero point of velocity are not significant, and

331: so all the curves have been shifted into alignment by subtracting a

332: constant offset.  This was done by setting the velocities from each

333: telescope to have zero mean, excepting Lick and SARG, for which better

334: alignment was achieved by setting the means to $-6$\,\ms{} and $-5$\,\ms,

335: respectively.  Note that EMILIE is not shown in Fig.~\ref{fig.series}{\em

336: a\/} because those data were referenced to a different value on each night

337: and so the night-to-night variations are not measurable (this does not

338: affect their usefulness for oscillations studies, however).  FIES is also

339: not shown in Fig.~\ref{fig.series}{\em a\/} because it has much greater

340: scatter than the rest.

341: 

342: In Fig.~\ref{fig.series}{\em a\/} we see variations in the radial velocity

343: of Procyon on timescales of days.  The good agreement between the different

344: telescopes indicates that these slow variations have a stellar origin,

345: although the imperfect match in overlapping sections shows that there is

346: also a contribution from instrumental drifts.  Figure~\ref{fig.series}{\em

347: b\/} shows a close-up of the central part of the campaign, during which the

348: coverage was above 90\%.  The solid curve shows the velocities after

349: smoothing, to better reveal the slow variations, which are discussed

350: in~\S\ref{sec.slow-var} below.

351: 

352: While interesting in their own right, the slow variations in the velocity

353: series significantly affect our ability to detect oscillations, due to

354: spectral leakage of power from the low-frequency part of the spectrum to

355: the oscillation region at higher frequencies.  Before merging the data from

356: the individual telescopes, we therefore removed the low-frequency

357: variations from the velocity series.  This was done for each telescope by

358: removing all the power below 280$\mu$Hz, a value that was chosen so as to

359: effectively remove the slow variations without affecting the oscillation

360: signal.  This filtering was done by subtracting a smoothed version of the

361: time series that contained all the power below that cut-off frequency.

362: 

363: In Fig.\ref{fig.series}{\em c\/} we show a close-up of a segment during

364: which three spectrographs were observing simultaneously.  The stellar

365: oscillations are clearly visible, with typical periods of about 15 minutes,

366: and there is good agreement between the different telescopes.  Note that

367: these data have been filtered to remove the slow variations.

368: 

369: \section{Optimizing the Weights}     \label{sec.weights}

370: 

371: The procedures for extracting velocities for each telescope also produced

372: estimates of the uncertainties,~$\sigma_i$.  In our analysis, we used these

373: uncertainties to calculate noise-optimized weights in the usual way, namely

374: $w_i=1/\sigma_i^2$.  If weights are not used when calculating the power

375: spectrum, a few bad data points can dominate and increase the noise floor

376: significantly.

377: 

378: We now describe the process we used to adjust these weights, which aims to

379: minimize the noise level in the final power spectrum.  The procedure

380: involves identifying and revising those uncertainties that were too

381: optimistic, and at the same time rescaling the uncertainties to be in

382: agreement with the actual noise levels in the data.  These methods have

383: already been described in previous papers \citep{BBK2004,BKA2007,LKB2007},

384: but the present analysis differs slightly from those descriptions and we

385: will therefore describe them in some detail.  One difference is that the

386: analysis had to be tailored to the individual time series because of the

387: large range of Nyquist frequencies (see Table~\ref{tab.obs}).

388: 

389: To illustrate the process, we show in Figs.~\ref{fig.show.harps}

390: and~\ref{fig.show.emilie} segments of data at different stages in the

391: process for two telescopes (HARPS and EMILIE).  The top panels

392: (Figs.~\ref{fig.show.harps}{\em a\/} and~\ref{fig.show.emilie}{\em a\/})

393: show the velocities for a single night, with the slow variations removed.

394: The remaining panels show the uncertainties at different stages in the

395: analysis.

396: 

397: It is important to stress that we are not adjusting the velocities, only

398: the uncertainties.  Of course, those adjustments still affect the power

399: spectrum of the velocities (which is, after all, why we are making the

400: adjustments) and so it is important to ensure that they do not distort the

401: oscillation signal and that the final weights reflect as accurately as

402: possible the actual noise properties of the series.

403: 

404: \subsection{Scaling the uncertainties}

405: 

406: We have scaled the uncertainties so that they agree with the noise level in

407: the corresponding amplitude spectrum, $\sigma_{\rm amp}$, as measured at

408: high frequencies.  This was done for each night and each telescope by

409: multiplying the uncertainties, $\sigma_i$, by a constant so that they

410: satisfied equation~(3) of \citet{BBK2004}:

411: \begin{equation}

412:       \sigma_{\rm amp}^2 \sum_{i=1}^{N} \sigma_i^{-2}  = \pi. 

413:         \label{eq.condition}

414: \end{equation}

415: This scaling was repeated after each step in the process described below.

416: Figures~\ref{fig.show.harps}{\em b\/} and~\ref{fig.show.emilie}{\em b\/}

417: show the uncertainties after scaling, and before any further adjustments.

418: 

419: \subsection{Filtering the uncertainties}     \label{sec.init-chk}

420: 

421: It is clear that the uncertainties in some parts of the time series show

422: variations that correlate with the oscillations of Procyon.  The clearest

423: example is HARPS, as shown by comparing the top two panels of

424: Fig.~\ref{fig.show.harps}, but the effect is also visible for other

425: telescopes (e.g., Fig~\ref{fig.show.emilie}).  To remove this structure in

426: the uncertainties, we have bandpass-filtered each of the uncertainty time

427: series to remove all power in the frequency range 280--2200$\mu$Hz.  This

428: removed fluctuations in the weights on the timescale of the stellar

429: oscillations, while retaining information on longer timescales (such as

430: poorer conditions at the beginnings and ends of nights) and on shorter

431: timescales (such as individual bad data points).  This process resulted in

432: slightly lower noise levels in the final power spectra for some of the

433: individual telescopes, reflecting the fact that the uncertainties, like any

434: measurement, contain noise that is reduced by bandpass filtering.

435: Figures~\ref{fig.show.harps}{\em c\/} and~\ref{fig.show.emilie}{\em c\/}

436: show the uncertainties after filtering.

437: 

438: We also noticed a few data points ($\sim$10) with unrealistically {\em

439: low\/} uncertainties.  These points would be given too high a weight in the

440: analysis and the uncertainties were therefore reset to the mean uncertainty

441: for that telescope night.

442: 

443: \subsection{Down-weighting bad data points}

444: 

445: Down-weighting of bad data points was done following the method described

446: by \citet{BBK2004}, with one difference that is discussed below.  The first

447: step was to make a high-pass-filtered version of the velocity time series

448: in which both the slow variations and the stellar oscillations were

449: removed.  This gave us a series of residual velocities, $r_i$, in which we

450: could identify data points that needed to be down-weighted, without being

451: affected by spectral leakage from the oscillations.  The frequency limit of

452: this high-pass filter varied from telescope to telescope, depending on the

453: Nyquist frequency of the data.

454: 

455: We compared the velocity residuals, $r_i$, with the corresponding

456: uncertainty estimates, $\sigma_i$.  Bad data points are those for which the

457: ratio $|r_i/\sigma_i|$ is large, i.e., where the residual velocity deviates

458: from zero by more than expected from the uncertainty estimate.

459: \citet{BBK2004}, who analyzed data similar to ours, found that the fraction

460: of good data points was essentially unity up to $|r_i/\sigma_i|=2$ and then

461: dropped off quickly for larger values of $|r_i/\sigma_i|$ -- see Figure~3

462: of \citet{BBK2004}.  They therefore introduced the factor $f$, which is the

463: fraction of good data points as a function of $|r_i/\sigma_i|$ and which

464: they obtained as the ratio between the distribution of data points in a

465: cumulative histogram of $|r_i/\sigma_i|$ and a best-fit Gaussian

466: distribution.

467: 

468: %% The weight of each data point was then decreased according to its value of

469: %% $|r_i/\sigma_i|$, so that the weights of data points with values of

470: %% $|r_i/\sigma_i|$ up to about 2 were unchanged, while points with higher

471: %% values of $|r_i/\sigma_i|$ were adjusted by dividing the corresponding

472: %% $\sigma_i$ with $\sqrt f$ (the weights are calculated as

473: %% $w_i=1/\sigma_i^2$).

474: 

475: We used a slightly different approach.  With the knowledge that points with

476: large values of $|r_i/\sigma_i|$ are bad, we introduced an analytical

477: function

478: \begin{equation}

479:   f(x_i)=\frac{1}{1+\left(\displaystyle{x_i\over x_0} \right)^{10}},\quad x_i=|r_i/\sigma_i|,

480:   \label{eq.fx}

481: \end{equation}

482: which has shape very similar to the fraction $f$ as a function of

483: $|r_i/\sigma_i|$ in \citet{BBK2004}.  The adjustable parameter $x_0$

484: controls the amount of down-weighting; it sets the value of

485: $|r_i/\sigma_i|$ for which the weights are multiplied by 0.5, and so it

486: determines how bad a data point should be before it is down-weighted.  The

487: optimum choice for $x_0$ was found through iteration, as described below.

488: Once this was done, we used $f(x_i)$ to adjust the weights by dividing

489: $\sigma_i$ by $\sqrt{f(x_i)}$, as in \citet{BBK2004}.

490: 

491: %% At this point we re-scaled the nightly weights to satisfy

492: %% Eq.~\ref{eq.condition}, however this time we used  weights both

493: %% in the calculation of the high-pass filtered time series and for

494: %% calculating the amplitude spectrum of the high-pass filtered velocities to

495: %% obtain $\sigma_{\rm amp}$.  Although the scaling constants were now close

496: %% to unity, some amount of re-scaling was required because both $\sigma_{\rm

497: %% amp}$ and $\sum_{i=1}^{N} \sigma_i^{-2}$ in Eq.~\ref{eq.condition} had been

498: %% changed because of the use of the  weights.  New values of

499: %% $|r_i/\sigma_i|$ were calculated and the individual uncertainties

500: %% $\sigma_i$ were adjusted once again using Eq.~\ref{eq.fx}.  We then

501: %% re-scaled the uncertainties using Eq.~\ref{eq.condition} again, merged the

502: %% individual nights of data from the spectrograph under investigation, and

503: %% made a final re-scaling of the {\it combined} time series using

504: %% Eq.~\ref{eq.condition}.

505: 

506: %% In Fig.~\ref{fig.fx} we show the weight correction factor $f$ as a function

507: %% of $r/\sigma$ for our 5698 HARPS data points, along with the analytical 

508: %% function in Eq.~\ref{eq.fx} with $x_0 = 3.1$ which was found to be optimal 

509: %% for HARPS.  The uncertainty of each data point have been divided by the

510: %% corresponding $\sqrt{f}$, however for the vast majority of points, 

511: %% $f$ is unity and only a small fraction of the data points were 

512: %% significantly down-weighted.  

513: 

514: The noise level used for optimizing $x_0$ was measured in a frequency band

515: near 2\,mHz in a weighted amplitude spectrum, between the oscillations and

516: the high-frequency part of the spectrum used for determining the

517: $|r_i/\sigma_i|$ values.  The exact position of the band was chosen for

518: each spectrograph separately, because of the differences in Nyquist

519: frequencies.  For each trial value of $x_0$, the noise level was determined

520: from a time series in which all power had been removed at both the low- and

521: high-frequency side of the frequency band, i.e., from a bandpass-filtered

522: time series containing information in the specific frequency band only.

523: This was done because applying the weights changes the spectral window

524: function and thus the amount of spectral leakage into the frequency band

525: where we determine the noise: if we did not filter out the oscillations and

526: the high-frequency part of the spectrum, we would optimize for a

527: combination of low noise {\em and\/} minimum amount of spectral leakage

528: (from both the low- and high frequency side of the passband).  In other

529: words, the spectral window function would influence our choice of $x_0$,

530: which is not optimal for obtaining the lowest possible noise level.

531: 

532: The procedure described above was repeated for a range of $x_0$ values and

533: we chose the one that resulted in the lowest noise in the power spectrum.

534: %% The aim was to down-weight bad data points without reducing too much the

535: %% effective number of data points, which increases the noise level and

536: %% affects the spectral window.  

537: Depending on the telescope, and hence the noise properties of the time

538: series, the optimal values of $x_0$ ranged from 1.7 to 4.3.

539: Figures~\ref{fig.show.harps}{\em d\/} and~\ref{fig.show.emilie}{\em d\/}

540: show the final uncertainties for HARPS and EMILIE.  

541: 

542: This completes our description of the process used to adjust the

543: uncertainties.  The results from calculating weighted power spectra using

544: these uncertainties are presented in~\S\ref{sec.power.spectra}.  First,

545: however, we discuss the slow variations in the velocity of Procyon that are

546: present in our data.

547: 

548: \section{Results} 

549: 

550: \subsection{Slow variations in stellar velocity}     \label{sec.slow-var}

551: 

552: The slow variations in the radial velocity of Procyon seen in

553: Fig.~\ref{fig.series} are remarkably similar to those seen in the Sun.

554: This can be seen from Fig.~\ref{fig.golf}, which shows a typical time

555: series of solar velocity measurements made with the GOLF instrument on the

556: SOHO spacecraft \citep{UGR2000,GTB2005}.  Longer series of GOLF data show

557: variations with a period of about 13\,d arising from active regions

558: crossing the solar disk (Fig.~11 in \citealt{GTB2005}; see also

559: \citealt{TGK2008}).  \new{This 13-day periodicity in solar velocities was

560: first observed by \citet{CIMc82}, who attributed it to rapid rotation of

561: the core, but it was subsequently shown to be due to surface rotation of

562: active regions \citep{Du+S83,A+M83,E+G83}.}  Similarly, we attribute the

563: slow variations in the radial velocity of Procyon to the appearance and

564: disappearance of active regions and their rotation across the stellar disk.

565: This explanation was also invoked by \citet{MBC2005} to explain much larger

566: velocity variations measured with HARPS for the COROT target HR~2530 (=

567: HD~49933; spectral type F5\,V).

568: 

569: The slow variations in Procyon appear to have a period of $P_{\rm slow} =

570: 10.3\pm0.5$\,d, as measured from the highest peak in the power spectrum.

571: \new{This agrees with an apparent periodicity of about 10\,d in two-site

572: observations of Procyon obtained over 20 nights by \citet[see their

573: Fig.~12]{KAS2008}.}  Identifying this as the stellar rotation period and

574: using a radius of 2.05\,$R_\sun$ \citep{KTM2004} implies a surface

575: rotational speed at the equator of $v = 10 \pm 0.5$\,\kms.  The measured

576: value of $P_{\rm slow}$ period might also correspond to half the rotation

577: period \new{\citep{Cla2003}}, in which case the rotational speed would be

578: half the value given above.  Combining with the spectroscopic value of $v

579: \sin i = 3.16 \pm 0.50$\,\kms{} \citep{APAL2002} gives an inclination angle

580: of $i=18 \pm 3^{\circ}$ (if $P_{\rm rot} = P_{\rm slow}$) and $i=39 \pm

581: 7^{\circ}$ (if $P_{\rm rot} = 2 P_{\rm slow}$).  The inclination of the

582: binary orbit is $31.1 \pm 0.6^{\circ}$ \citep{GWL2000} and so, if we

583: require that the rotation axis of Procyon is aligned with the orbital

584: rotation axis, it may be that $P_{\rm rot} = 2 P_{\rm slow}$.

585: 

586: \subsection{Power spectra and stellar activity} \label{sec.power.spectra}

587: 

588: The weighted power spectrum for each telescope and for the combined time

589: series, based on the uncertainties discussed in~\S\ref{sec.weights}, are

590: shown in Fig.~\ref{fig.all.power}.  The noise levels, as measured at high

591: frequencies in the amplitude spectrum ($\sigma_{\rm amp}$), are given in

592: column~8 of Table~\ref{tab.obs}.  The final column of that table gives the

593: mean noise level per minute of observing time, with a spread that reflects

594: a combination of factors, including telescope aperture, observing duty

595: cycle, spectrograph design, as well as atmospheric conditions such as

596: seeing.

597: 

598: The power spectrum of the combined time series is shown again in

599: Fig.~\ref{fig.power}, both with and without the use of weights.  As is

600: well-established, using weights reduces the noise level significantly, at

601: the cost of increasing the sidelobes in the spectral window (because the

602: best data segments are given more weight -- see insets).  When weights are

603: used, the noise level above 3\,mHz is 1.9\,\cms{} in amplitude, but this

604: does include some degree of spectral leakage from the oscillations.  If we

605: high-pass filter the spectrum up to 3\,mHz, the noise level drops to

606: 1.5\,\cms{} in amplitude.  Note that without the use of weights, the noise

607: level is higher by more than a factor of two.

608: 

609: Looking again at Fig.~\ref{fig.power}, we see that the use of weights

610: appears to have increased the amplitude of the oscillations.  In fact, this

611: indicates the finite lifetime of the oscillation modes: in

612: Fig.~\ref{fig.power}{\em b} the HARPS data are given the highest weight,

613: and so the effective duration of the observations is decreased (and the

614: sidelobes in the spectral window become much stronger).  Our estimate of

615: the mode lifetime is given in~\S\ref{sec.lifetime}.

616: 

617: It is also useful to convert to power density, which is independent of the

618: observing window and therefore allows us to compare noise levels.  This is

619: done by multiplying the power spectrum by the effective length of the

620: observing run, which we calculate as the reciprocal of the area integrated

621: under the spectral window (in power).  The values for the different

622: telescopes are given in column~3 of Table~\ref{tab.obs}.  In

623: Fig.~\ref{fig.pds.harps} we show the power density spectrum on a

624: logarithmic scale for the HARPS data, which has the lowest noise per minute

625: of observing time.  We see three components: (i)~the oscillations (about

626: 300--1100\,\muHz); (ii)~white noise at high frequencies; and (iii)~a

627: sloping background of power at low frequencies (stellar granulation and

628: activity, and presumably also some instrumental drift).

629: Figure~\ref{fig.pds.amp} compares the power density spectra for the

630: different telescopes.  

631: They show a similar oscillation signal and similar background from stellar

632: noise at lower frequencies (below about 250\,\muHz), with different levels

633: of white noise at higher frequencies (above about 2000\,\muHz), reflecting

634: the different levels of photon noise.

635: 

636: In Fig.~\ref{fig.pds.harps}, the lower two dashed lines indicate the

637: background level in the Sun, as measured from the GOLF data during solar

638: minimum and maximum, respectively.  The upper dashed line is the solar

639: maximum line shifted to match the power density of Procyon, which required

640: multiplying by a factor of 40.  We can use this scaling factor to estimate

641: the fraction of Procyon's surface that is covered by active regions,

642: relative to the Sun, as follows.  The low-frequency part of the velocity

643: power-density spectrum from the Sun falls off as frequency squared

644: \citep{Har85,PRCJ99}, and we see the same behaviour for Procyon.  Hence, in

645: both cases we have

646: \begin{equation} 

647:   PD(\nu) \propto \nu^{-2}.

648: \end{equation} 

649: Let $T$ be the typical time for an active region to be visible on the

650: surface (which may depend on both rotation and the typical lifetime of

651: active regions).  The amplitude of the velocity signal at frequency $\nu_0

652: = 1/T$ will be proportional to the fractional area covered by active

653: regions, $da/a$, and to the projected rotational velocity, $v\sin i$.  The

654: power density at $\nu_0$ is therefore:

655: \begin{equation} 

656:   PD(\nu_0) = (da/a)^2 (v\sin i)^2.

657: \end{equation} 

658: Combining these gives

659: \begin{equation} 

660:   PD(\nu) = \left(\frac{da}{a}\right)^2 \left(\frac{v \sin i}{T}\right)^2 \nu^{-2}.

661:   \label{eq.spots}

662: \end{equation} 

663: We will assume that $T$ is proportional to the rotation period, which we

664: take to be $10.3$\,d (or twice that value) for Procyon and 25.4\,d for the

665: Sun.  The measured values for $v \sin i$ are 3.2\,\kms{} for Procyon

666: \citep{APAL2002} and 2.0\,\kms{} for the Sun.  Combining these values with

667: our measurement of the power densities indicates that the area covered by

668: active regions on Procyon is about 1.6 times the solar maximum value (or

669: twice that value).  No detection of a magnetic field in Procyon has been

670: reported, and published upper limits imply that the average field cannot be

671: more than a few times solar (see Table 3 in \citealt{KHV2007}), which

672: appears to be consistent with our results.

673: 

674: 

675: \subsection{Oscillation amplitude and mode lifetime} \label{sec.lifetime}

676: 

677: To measure the amplitude of oscillations in Procyon, we have used the

678: method described by \citet{KBA2008}.  In brief, this involves the following

679: steps: (i)~heavily smoothing the power spectrum (by convolving with a

680: Gaussian having a full width at half maximum of $4\Dnu$, where $\Dnu$

681: is the large frequency separation), to produce a single hump of excess

682: power that is insensitive to the fact that the oscillation spectrum has

683: discrete peaks; (ii)~converting to power density

684: (see~\S\ref{sec.power.spectra}); (iii)~fitting and subtracting the

685: background noise; and (iv)~multiplying by $\Dnu/4.09$ and taking the square

686: root, in order to convert to amplitude per radial oscillation mode.

687: Note that 4.09 is the effective number of modes per order for

688: full-disk velocities observations, normalized to the amplitudes of radial

689: ($l=0$) modes -- see \citet{KBA2008} for details.

690: 

691: We applied this method to each of the telescopes separately, and the result

692: is shown in Fig.~\ref{fig.amp}.  There are significant

693: differences between the different curves, which we attribute to intrinsic

694: variations in the star arising from the stochastic nature of the excitation

695: and damping.  To investigate this further, we also applied the method to

696: the combined time series, after first subdividing it into ten 2-day

697: subsets.  Figure~\ref{fig.amp.segments} shows these amplitude curves and

698: also their average.

699: 

700: The amplitude curve of Procyon has a broad plateau, rather than the single

701: peak that has been seen for other stars.

702: \new{Figure~\ref{fig.amp.stars} shows the smoothed amplitude curve for Procyon

703: compared to the Sun and other stars.  It is an updated version of Fig.~8 of

704: \citet{KBA2008}, where the following stars have been added: 

705: $\mu$~Ara \citep{BBS2005}, 

706: HD~49933 \citep{MBC2005}, 

707: $\mu$~Her \citep{BBC2008}, 

708: $\gamma$~Pav \citep{MDM2008}, 

709: $\tau$~Cet \citep{TKB2008}.

710: }

711: 

712: The plateau for Procyon is centred at 900\,\muHz{} and is about 500\,\muHz{} wide, with

713: a mean amplitude across that range of $38.1\pm1.3$\,\cms.  This is our

714: estimate for the amplitude of the radial ($l=0$) modes in Procyon.

715: Comparing with the corresponding measurement for the Sun ($18.7\pm

716: 0.7$\,\cms; \citealt{KBA2008}) implies that the velocity oscillations in

717: Procyon are $2.04 \pm 0.10$ times solar.  In both Procyon and the Sun, the

718: modes with $l=1$ are higher by a factor of 1.35 (see Table~1 of

719: \citealt{KBA2008}).

720: 

721: The corresponding intensity amplitude, after accounting for the higher

722: effective temperature of Procyon (see Eq.~5 in \citealt{K+B95}), is 1.60

723: times solar.  This implies an amplitude at 500\,nm of 6.8\,ppm for $l=0$

724: and 8.5\,ppm for $l=1$ (see Table~1 of \citealt{KBA2008}).  These

725: amplitudes are completely consistent with the detection of oscillations in

726: Procyon by WIRE \citep{BKB2005b} and with the upper limit set by MOST

727: \citep{MKG2004,BKB2005}.

728: 

729: The standard deviation of the ten segments in Fig.~\ref{fig.amp.segments}

730: is $\sigma_A/A = 10.4\% \pm 2.3\%$, which reflects the finite lifetime of

731: the modes.  We can use equation~(3) from \citet{KBA2008} to

732: estimate the mode lifetime, but we must account for the much greater width

733: of the oscillation envelope in Procyon.  

734: % \mynote{How did we do that?}

735: Note that this equation was established empirically and we have confirmed

736: it analytically using the work of \citet{T+A94}.  We estimate the mode

737: lifetime to be $\tau = 1.5_{-0.8}^{+1.9}$\,d.  This equals, within rather

738: large uncertainties, the solar value of 2.9\,d \citep[e.g.,][]{CEI97}.

739: 

740: 

741: \section{Conclusions}

742: 

743: We have presented multi-site velocity observations of Procyon that we

744: obtained with eleven telescopes over more than three weeks.  Combining data

745: that spans a range of precisions and sampling rates presents a significant

746: challenge.  When calculating the power spectrum, it is important to use

747: weights that are based on the measurement uncertainties, otherwise the

748: result is dominated by the noisiest data.  We have described in detail our

749: methods for adjusting the weights in order to minimize the noise level in

750: the final power spectrum.

751: 

752: Our velocity measurements show the clear signature of oscillations.  The

753: power spectrum shows an excess in a plateau that is centred at 0.9\,mHz and

754: is broader than has been seen for other solar-type stars.  The mean

755: amplitude of the radial modes is $38.1\pm1.3$\,\cms{} ($2.04 \pm 0.10$

756: times solar), which is consistent with previous detections from the ground

757: and by the WIRE spacecraft, and also with the upper limit set by the MOST

758: spacecraft.  The variation of the amplitude during the observing campaign

759: allowed us to estimate the mode lifetime to be $1.5_{-0.8}^{+1.9}$\,d.

760: 

761: We also found a slow variation in the radial velocity of Procyon, with good

762: agreement between different telescopes.  These variations are remarkably

763: similar to those seen in the Sun, and we interpret them as being due to

764: rotational modulation from active regions on the stellar surface.  The variations

765: appear to have a period of about 10 days, which presumably equals the

766: stellar rotation period or, perhaps, half of it.  The amount of power in

767: these slow variations indicates that the fractional area of Procyon

768: covered by active regions is slightly higher than for the Sun.

769: 

770: The excellent coverage of the observations and the high signal-to-noise

771: should allow us to produce a good set of oscillation frequencies for

772: Procyon.  This analysis will be presented in subsequent papers.

773: 

774: \acknowledgments

775: 

776: This work was supported financially by 

777: %

778: the Danish Natural Science Research Council,

779: %

780: the Australian Research Council, 

781: %

782: the Swiss National Science Foundation,

783: %

784: %

785: NSF grant AST-9988087 (RPB) and by SUN Microsystems.

786: %

787: We thank Hugh Jones, Chris Tinney and the other members of the

788: Anglo-Australian Planet Search for agreeing to a time swap that allowed our

789: AAT observations to be scheduled.

790: %

791: MM is grateful to Prof.\ N. Kameswara Rao, G. Pandey and S. Sriram for

792: their participation in campaign with VBT Echelle Spectrometer, which was

793: used for the first time for Doppler spectroscopy observations.

794: 

795: %% \input{bibfiles-simple}

796: %% \bibliographystyle{natbib/mynatbib}

797: 

798: %% \input{paper-procyon-torben.bbl}

799: 

800: \begin{thebibliography}{52}

801: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

802: \expandafter\ifx\csname url\endcsname\relax

803:   \def\url#1{{\tt #1}}\fi

804: 

805: \bibitem[{Aerts} et~al.(2008){Aerts}, {Christensen-Dalsgaard}, {Cunha}, \&

806:   {Kurtz}]{AChDC2008}

807: {Aerts}, C., {Christensen-Dalsgaard}, J., {Cunha}, M., \& {Kurtz}, D.~W., 2008,

808:   Sol. Phys.

809: \newblock in press ({\tt arXiv:0803.3527}).

810: 

811: \bibitem[{Allende Prieto} et~al.(2002){Allende Prieto}, {Asplund}, {L{\'o}pez},

812:   \& {Lambert}]{APAL2002}

813: {Allende Prieto}, C., {Asplund}, M., {L{\'o}pez}, R.~J.~G., \& {Lambert},

814:   D.~L., 2002, ApJ, 567, 544.

815: 

816: \bibitem[{Andersen} \& {Maltby}(1983){Andersen}, \& {Maltby}]{A+M83}

817: {Andersen}, B.~N., \& {Maltby}, P., 1983, Nat, 302, 808.

818: 

819: \bibitem[{Baudin} et~al.(2008){Baudin}, {Appourchaux}, {Boumier}, {Kuschnig},

820:   {Leibacher}, \& {Matthews}]{BAB2008}

821: {Baudin}, F., {Appourchaux}, T., {Boumier}, P., {Kuschnig}, R., {Leibacher},

822:   J.~W., \& {Matthews}, J.~M., 2008, A\&A, 478, 461.

823: 

824: \bibitem[{Bedding} \& {Kjeldsen}(2007){Bedding}, \& {Kjeldsen}]{B+K2007c}

825: {Bedding}, T.~R., \& {Kjeldsen}, H., 2007, In Stancliffe, R.~J., Houdek, G.,

826:   Martin, R.~G., \& Tout, C.~A., editors, {\em Unsolved Problems in Stellar

827:   Physics: A Conference in Honour of Douglas Gough}, volume 948, page 117.

828:   American Institute of Physics.

829: 

830: \bibitem[{Bedding} et~al.(2007){Bedding}, {Kjeldsen}, Arentoft,

831:   et~al.]{BKA2007}

832: {Bedding}, T.~R., {Kjeldsen}, H., Arentoft, T., et~al., 2007, ApJ, 663, 1315.

833: 

834: \bibitem[{Bedding} et~al.(2005){Bedding}, {Kjeldsen}, {Bouchy},

835:   et~al.]{BKB2005}

836: {Bedding}, T.~R., {Kjeldsen}, H., {Bouchy}, F., et~al., 2005, A\&A, 432, L43.

837: 

838: \bibitem[{Bonanno} et~al.(2008){Bonanno}, {Benatti}, {Claudi}, {Desidera},

839:   {Gratton}, {Leccia}, \& {Patern{\`o}}]{BBC2008}

840: {Bonanno}, A., {Benatti}, S., {Claudi}, R., {Desidera}, S., {Gratton}, R.,

841:   {Leccia}, S., \& {Patern{\`o}}, L., 2008, ApJ, 676, 1248.

842: 

843: \bibitem[{Bouchy} et~al.(2005){Bouchy}, {Bazot}, {Santos}, {Vauclair}, \&

844:   {Sosnowska}]{BBS2005}

845: {Bouchy}, F., {Bazot}, M., {Santos}, N.~C., {Vauclair}, S., \& {Sosnowska}, D.,

846:   2005, A\&A, 440, 609.

847: 

848: \bibitem[Bouchy \& {Carrier}(2002)Bouchy, \& {Carrier}]{B+C2002}

849: Bouchy, F., \& {Carrier}, F., 2002, A\&A, 390, 205.

850: 

851: \bibitem[Bouchy et~al.(2004)Bouchy, Maeder, Mayor, M{\'e}gevand, Pepe, \&

852:   Sosnowska]{BMM2004}

853: Bouchy, F., Maeder, A., Mayor, M., M{\'e}gevand, D., Pepe, F., \& Sosnowska,

854:   D., 2004, Nat, 432, 7015.

855: 

856: \bibitem[Bouchy et~al.(2002)Bouchy, {Schmitt}, {Bertaux}, \& {Connes}]{BSB2002}

857: Bouchy, F., {Schmitt}, J., {Bertaux}, J.-L., \& {Connes}, P., 2002, In Aerts,

858:   C., Bedding, T.~R., \& Christensen-Dalsgaard, J., editors, {\em IAU Colloqium

859:   185: Radial and Nonradial Pulsations as Probes of Stellar Physics}, volume

860:   259, page 472. ASP Conf. Ser.

861: 

862: \bibitem[Brown et~al.(1991)Brown, Gilliland, Noyes, \& Ramsey]{BGN91}

863: Brown, T.~M., Gilliland, R.~L., Noyes, R.~W., \& Ramsey, L.~W., 1991, ApJ, 368,

864:   599.

865: 

866: \bibitem[{Bruntt} et~al.(2005){Bruntt}, {Kjeldsen}, {Buzasi}, \&

867:   {Bedding}]{BKB2005b}

868: {Bruntt}, H., {Kjeldsen}, H., {Buzasi}, D.~L., \& {Bedding}, T.~R., 2005, ApJ,

869:   633, 440.

870: 

871: \bibitem[Butler et~al.(2004)Butler, Bedding, Kjeldsen, et~al.]{BBK2004}

872: Butler, R.~P., Bedding, T.~R., Kjeldsen, H., et~al., 2004, ApJ, 600, L75.

873: 

874: \bibitem[Butler et~al.(1996)Butler, Marcy, Williams, McCarthy, Dosanjh, \&

875:   Vogt]{BMW96}

876: Butler, R.~P., Marcy, G.~W., Williams, E., McCarthy, C., Dosanjh, P., \& Vogt,

877:   S.~S., 1996, PASP, 108, 500.

878: 

879: \bibitem[Chaplin et~al.(1997)Chaplin, {Elsworth}, {Isaak}, {McLeod}, {Miller},

880:   \& {New}]{CEI97}

881: Chaplin, W.~J., {Elsworth}, Y., {Isaak}, G.~R., {McLeod}, C.~P., {Miller},

882:   B.~A., \& {New}, R., 1997, MNRAS, 288, 623.

883: 

884: \bibitem[{Clarke}(2003)]{Cla2003}

885: {Clarke}, D., 2003, A\&A, 407, 1029.

886: 

887: \bibitem[Claudi et~al.(2005)Claudi, {Bonanno}, {Leccia}, {Ventura}, {Desidera},

888:   {Gratton}, {Cosentino}, {Paterno}, \& {Endl}]{CBL2005}

889: Claudi, R.~U., {Bonanno}, A., {Leccia}, S., {Ventura}, R., {Desidera}, S.,

890:   {Gratton}, R., {Cosentino}, R., {Paterno}, L., \& {Endl}, M., 2005, A\&A,

891:   429, L17.

892: 

893: \bibitem[{Claverie} et~al.(1982){Claverie}, {Isaak}, {McLeod}, {van der Raay},

894:   {Palle}, \& {Roca Cortes}]{CIMc82}

895: {Claverie}, A., {Isaak}, G.~R., {McLeod}, C.~P., {van der Raay}, H.~B.,

896:   {Palle}, P.~L., \& {Roca Cortes}, T., 1982, Nat, 299, 704.

897: 

898: \bibitem[{Durrant} \& {Schr\"oter}(1983){Durrant}, \& {Schr\"oter}]{Du+S83}

899: {Durrant}, C.~J., \& {Schr\"oter}, E.~H., 1983, Nat, 301, 589.

900: 

901: \bibitem[{Edmunds} \& {Gough}(1983){Edmunds}, \& {Gough}]{E+G83}

902: {Edmunds}, M.~G., \& {Gough}, D.~O., 1983, Nat, 302, 810.

903: 

904: \bibitem[Eggenberger et~al.(2004)Eggenberger, {Carrier}, {Bouchy}, \&

905:   {Blecha}]{ECB2004}

906: Eggenberger, P., {Carrier}, F., {Bouchy}, F., \& {Blecha}, A., 2004, A\&A, 422,

907:   247.

908: 

909: \bibitem[{Endl} et~al.(2005){Endl}, {Cochran}, {Hatzes}, \&

910:   {Wittenmyer}]{ECH2005}

911: {Endl}, M., {Cochran}, W.~D., {Hatzes}, A.~P., \& {Wittenmyer}, R.~A., 2005, In

912:   {\em Revista Mexicana de Astronomia y Astrofisica Conference Series},

913:   volume~23, page~64.

914: 

915: \bibitem[{Frandsen} \& {Lindberg}(2000){Frandsen}, \& {Lindberg}]{F+L2000}

916: {Frandsen}, S., \& {Lindberg}, B., 2000, In Teixeira, T., \& Bedding, T.~R.,

917:   editors, {\em The Third MONS Workshop: Science Preparation and Target

918:   Selection}, page 163. Aarhus: Aarhus Universitet.

919: 

920: \bibitem[Garc{\'{\i}}a et~al.(2005)Garc{\'{\i}}a, {Turck-Chi{\`e}ze},

921:   {Boumier}, {Robillot}, {Bertello}, et~al.]{GTB2005}

922: Garc{\'{\i}}a, R.~A., {Turck-Chi{\`e}ze}, S., {Boumier}, P., {Robillot}, J.~M.,

923:   {Bertello}, L., et~al., 2005, A\&A, 442, 385.

924: 

925: \bibitem[Gilliland et~al.(1993)Gilliland, Brown, Kjeldsen, McCarthy, Peri,

926:   et~al.]{GBK93}

927: Gilliland, R.~L., Brown, T.~M., Kjeldsen, H., McCarthy, J.~K., Peri, M.~L.,

928:   et~al., 1993, AJ, 106, 2441.

929: 

930: \bibitem[{Girard} et~al.(2000){Girard}, {Wu}, {Lee}, {Dyson}, {van Altena},

931:   et~al.]{GWL2000}

932: {Girard}, T.~M., {Wu}, H., {Lee}, J.~T., {Dyson}, S.~E., {van Altena}, W.~F.,

933:   et~al., 2000, AJ, 119, 2428.

934: 

935: \bibitem[{Guenther} et~al.(2007){Guenther}, {Kallinger}, {Reegen}, {Weiss},

936:   {Matthews}, {Kuschnig}, {Moffat}, {Rucinski}, {Sasselov}, \&

937:   {Walker}]{GKR2007}

938: {Guenther}, D.~B., {Kallinger}, T., {Reegen}, P., {Weiss}, W.~W., {Matthews},

939:   J.~M., {Kuschnig}, R., {Moffat}, A.~F.~J., {Rucinski}, S.~M., {Sasselov}, D.,

940:   \& {Walker}, G.~A.~H., 2007, Commun. Asteroseismology, 151, 5.

941: 

942: \bibitem[Harvey(1985)]{Har85}

943: Harvey, J., 1985, In Rolfe, E., \& Battrick, B., editors, {\em Future Missions

944:   in Solar, Heliospheric \& Space Plasma Physics}, ESA SP-235, page 199.

945: 

946: \bibitem[{Hatzes} et~al.(2003){Hatzes}, {Guenther}, {K{\"u}rster}, \&

947:   {McArthur}]{HGK2003}

948: {Hatzes}, A.~P., {Guenther}, E., {K{\"u}rster}, M., \& {McArthur}, B., 2003, In

949:   {Fridlund}, M., {Henning}, T., \& {Lacoste}, H., editors, {\em Earths:

950:   DARWIN/TPF and the Search for Extrasolar Terrestrial Planets}, volume 539,

951:   page 441.

952: 

953: \bibitem[{Kambe} et~al.(2008){Kambe}, {Ando}, {Sato}, et~al.]{KAS2008}

954: {Kambe}, E., {Ando}, H., {Sato}, B., et~al., 2008, PASJ, 60, 45.

955: 

956: \bibitem[Kervella et~al.(2004)Kervella, {Th{\' e}venin}, {Morel}, {Berthomieu},

957:   {Bord{\' e}}, \& {Provost}]{KTM2004}

958: Kervella, P., {Th{\' e}venin}, F., {Morel}, P., {Berthomieu}, G., {Bord{\' e}},

959:   P., \& {Provost}, J., 2004, A\&A, 413, 251.

960: 

961: \bibitem[{Kim} et~al.(2007){Kim}, {Han}, {Valyavin}, {Plachinda}, {Jang},

962:   et~al.]{KHV2007}

963: {Kim}, K.-M., {Han}, I., {Valyavin}, G.~G., {Plachinda}, S., {Jang}, J.~G.,

964:   et~al., 2007, PASP, 119, 1052.

965: 

966: \bibitem[Kjeldsen \& Bedding(1995)Kjeldsen, \& Bedding]{K+B95}

967: Kjeldsen, H., \& Bedding, T.~R., 1995, A\&A, 293, 87.

968: 

969: \bibitem[Kjeldsen et~al.(2008)Kjeldsen, {Bedding}, Arentoft, et~al.]{KBA2008}

970: Kjeldsen, H., {Bedding}, T.~R., Arentoft, T., et~al., 2008, ApJ.

971: \newblock in press, arXiv:0804.1182.

972: 

973: \bibitem[{Leccia} et~al.(2007){Leccia}, {Kjeldsen}, {Bonanno}, {Claudi},

974:   {Ventura}, \& {Patern\`o}]{LKB2007}

975: {Leccia}, S., {Kjeldsen}, H., {Bonanno}, A., {Claudi}, R.~U., {Ventura}, R., \&

976:   {Patern\`o}, L., 2007, A\&A, 464, 1059.

977: 

978: \bibitem[{Marchenko}(2008)]{Mar08}

979: {Marchenko}, S.~V., 2008, A\&A, 479, 845.

980: 

981: \bibitem[Marti{\'c} et~al.(2004)Marti{\'c}, {Lebrun}, {Appourchaux}, \&

982:   {Korzennik}]{MLA2004}

983: Marti{\'c}, M., {Lebrun}, J.-C., {Appourchaux}, T., \& {Korzennik}, S.~G.,

984:   2004, A\&A, 418, 295.

985: 

986: \bibitem[Marti{\'c} et~al.(1999)Marti{\'c}, Schmitt, Lebrun, Barban, Connes,

987:   Bouchy, Michel, Baglin, Appourchaux, \& Bertaux]{MSL99}

988: Marti{\'c}, M., Schmitt, J., Lebrun, J.-C., Barban, C., Connes, P., Bouchy, F.,

989:   Michel, E., Baglin, A., Appourchaux, T., \& Bertaux, J.-L., 1999, A\&A, 351,

990:   993.

991: 

992: \bibitem[Matthews et~al.(2004)Matthews, Kuschnig, Guenther, et~al.]{MKG2004}

993: Matthews, J.~M., Kuschnig, R., Guenther, D.~B., et~al., 2004, Nat, 430, 51.

994: \newblock {Erratum}: 430, 921.

995: 

996: \bibitem[{Mosser} et~al.(2005){Mosser}, {Bouchy}, {Catala}, et~al.]{MBC2005}

997: {Mosser}, B., {Bouchy}, F., {Catala}, C., et~al., 2005, A\&A, 431, L13.

998: 

999: \bibitem[{Mosser} et~al.(2008{\natexlab{a}}){Mosser}, {Bouchy}, Marti{\'c},

1000:   {Appourchaux}, Barban, et~al.]{MBM2008}

1001: {Mosser}, B., {Bouchy}, F., Marti{\'c}, M., {Appourchaux}, T., Barban, C.,

1002:   et~al., 2008{\natexlab{a}}, A\&A, 478, 197.

1003: 

1004: \bibitem[{Mosser} et~al.(2008{\natexlab{b}}){Mosser}, {Deheuvels}, {Michel},

1005:   {Thevenin}, {Dupret}, {Samadi}, {Barban}, \& {Goupil}]{MDM2008}

1006: {Mosser}, B., {Deheuvels}, S., {Michel}, E., {Thevenin}, F., {Dupret}, M.~A.,

1007:   {Samadi}, R., {Barban}, C., \& {Goupil}, M.~J., 2008{\natexlab{b}}, A\&A.

1008: \newblock in press ({\tt 2008arXiv0804.3119}).

1009: 

1010: \bibitem[{Mosser} et~al.(1998){Mosser}, {Maillard}, {M\'ekarnia}, \&

1011:   {Gay}]{MMM98}

1012: {Mosser}, B., {Maillard}, J.~P., {M\'ekarnia}, D., \& {Gay}, J., 1998, A\&A,

1013:   340, 457.

1014: 

1015: \bibitem[{Pall{\' e}} et~al.(1999){Pall{\' e}}, {Roca Cort{\' e}s}, {Jim{\'

1016:   e}nez}, {Golf}, \& {VIRGO Teams}]{PRCJ99}

1017: {Pall{\' e}}, P.~L., {Roca Cort{\' e}s}, T., {Jim{\' e}nez}, A., {Golf}, \&

1018:   {VIRGO Teams}, 1999, In Gim\'enez, A., Guinan, E.~F., \& Montesinos, B.,

1019:   editors, {\em {Proc. Workshop on Stellar Structure theory and Tests of

1020:   Convective Energy Transport}}, volume 173, page 297. ASP Conf. Ser.

1021: 

1022: \bibitem[{R{\'e}gulo} \& {Roca Cort{\'e}s}(2005){R{\'e}gulo}, \& {Roca

1023:   Cort{\'e}s}]{R+RC2005}

1024: {R{\'e}gulo}, C., \& {Roca Cort{\'e}s}, T., 2005, A\&A, 444, L5.

1025: 

1026: \bibitem[{Rupprecht} et~al.(2004){Rupprecht}, {Pepe}, {Mayor}, et~al.]{RPM2004}

1027: {Rupprecht}, G., {Pepe}, F., {Mayor}, M., et~al., 2004, In Moorwood, A. F.~M.,

1028:   \& Masanori, I., editors, {\em Ground-based Instrumentation for Astronomy},

1029:   volume 5492 of {\em Proc. SPIE}, page 148.

1030: 

1031: \bibitem[Teixeira et~al.(2008)Teixeira, Kjeldsen, Bedding, et~al.]{TKB2008}

1032: Teixeira, T., Kjeldsen, H., Bedding, T.~R., et~al., 2008, A\&A.

1033: \newblock to be submitted.

1034: 

1035: \bibitem[Tingley et~al.(2008)Tingley, Grundahl, \& Kjeldsen]{TGK2008}

1036: Tingley, B., Grundahl, F., \& Kjeldsen, K., 2008, MNRAS.

1037: \newblock to be submitted.

1038: 

1039: \bibitem[{Toutain} \& {Appourchaux}(1994){Toutain}, \& {Appourchaux}]{T+A94}

1040: {Toutain}, T., \& {Appourchaux}, T., 1994, A\&A, 289, 649.

1041: 

1042: \bibitem[Ulrich et~al.(2000)Ulrich, Garc{\'{\i}}a, Robillot, {Turck-Chi{\`

1043:   e}ze}, {Bertello}, {Charra}, {Dzitko}, {Gabriel}, \& {Roca Cort{\'

1044:   e}s}]{UGR2000}

1045: Ulrich, R.~K., Garc{\'{\i}}a, R.~A., Robillot, J.-M., {Turck-Chi{\` e}ze}, S.,

1046:   {Bertello}, L., {Charra}, J., {Dzitko}, H., {Gabriel}, A.~H., \& {Roca

1047:   Cort{\' e}s}, T., 2000, A\&A, 364, 799.

1048: 

1049: \end{thebibliography}

1050: 

1051: \clearpage

1052: 

1053: \input{tab1}

1054: \input{tab2}

1055: 

1056: \clearpage

1057: 

1058: \begin{figure*}

1059: \epsscale{0.8}

1060: %% \plotone{Figures/slow_timeseries_all.eps}

1061: %% \plotone{Figures/slow_timeseries_zoom.eps}

1062: %% \plotone{Figures/overlap.eps}

1063: %% \plotone{Figures/sigma_timeseries_b.eps}

1064: \plotone{f1a_low.eps}

1065: \plotone{f1b_low.eps}

1066: \plotone{f1c.eps}

1067: \plotone{f1d_low.eps}

1068: \caption[]{\label{fig.series} Velocity measurements of Procyon, color-coded

1069: as follows: 

1070: %

1071: HARPS         = red;

1072: CORALIE       = brown;

1073: McDonald      = gray;

1074: Lick          = cyan;

1075: UCLES         = blue;

1076: Okayama       = green;

1077: Tautenburg    = black;

1078: SOPHIE        = dark green;

1079: SARG          = dark blue;

1080: FIES          = magenta;

1081: EMILIE        = yellow.

1082: 

1083: ({\em a})~The full time series, before any removal of slow trends (EMILIE

1084: and FIES are not shown).

1085: 

1086: ({\em b})~Close-up of the central ten days (FIES not shown).  

1087: 

1088: ({\em c})~Close-up of a five-hour segment during which three spectrographs

1089: observed simultaneously: HARPS (red circles), SOPHIE (dark green squares) and

1090: SARG (dark blue triangles).  All three series have been high-pass filtered

1091: to remove slow trends and the SOPHIE and SARG data have been smoothed

1092: slightly (using a boxcar with a width of three data points).

1093: 

1094: ({\em d})~The time series of the final noise-optimized uncertainties,

1095: showing all 11 telescopes.

1096: }

1097: \end{figure*}

1098: 

1099: 

1100: \begin{figure*}

1101: \epsscale{0.9}

1102: %% \plotone{Figures/show_sigma.eps}

1103: \plotone{f2.eps}

1104: \caption[]{\label{fig.show.harps} Steps in the adjustments of weights,

1105: illustrated using HARPS data from a single night. 

1106: ({\em a})~The velocities, with the slow variations removed.

1107: ({\em b})~The uncertainties, after scaling to satisfy

1108: Eq.~\ref{eq.condition} but before any further adjustments.

1109: ({\em c})~The uncertainties after filtering to remove power on the

1110: timescale of the oscillations.

1111: ({\em d})~The final uncertainties, after adjusting to optimize the noise

1112: (see text).

1113: }

1114: \end{figure*}

1115: 

1116: \begin{figure*}

1117: \epsscale{0.9}

1118: %% \plotone{Figures/show_sigma_emilie.eps}

1119: \plotone{f3.eps}

1120: \caption[]{\label{fig.show.emilie} Same as Fig.~\ref{fig.show.harps}, but

1121:   for a single night from EMILIE.  

1122: }

1123: \end{figure*}

1124: 

1125: \begin{figure*}

1126: \epsscale{0.9}

1127: %% \plotone{Figures/golf_timeseries.eps}

1128: \plotone{f4.eps}

1129: \caption[]{\label{fig.golf} Time series of velocity measurements of the Sun

1130: obtained over 21 days with the GOLF instrument on the SOHO spacecraft.  }

1131: \end{figure*}

1132: 

1133: \begin{figure*}

1134: \epsscale{1.0}

1135: %% \plotone{Figures/allspectra.eps}

1136: \plotone{f5.eps}

1137: %(query-replace "1.25e-06" "1.25e-03" nil nil nil)

1138: \caption[]{\label{fig.all.power} Power spectra for all 11 telescopes,

1139: together with that of the combined series.  Note that the vertical scale is

1140: not the same for all panels.}

1141: \end{figure*}

1142: 

1143: \begin{figure*}

1144: \epsscale{0.8}

1145: %% \plotone{Figures/wnow.eps}

1146: \plotone{f6.eps}

1147: %(query-replace "8e-09" "8e-03" nil nil nil)

1148: \caption[]{\label{fig.power} Final power spectrum based on the

1149: noise-optimized weights (lower panel), and also without applying

1150: the weights (upper panel).  The inset shows the spectral window.   }

1151: \end{figure*}

1152: 

1153: \begin{figure*}

1154: \epsscale{0.8}

1155: %% \plotone{Figures/gausssmooth_log_white.eps}

1156: \plotone{f7.eps}

1157: \caption[]{\label{fig.pds.harps} Power density spectrum of Procyon from the

1158: HARPS data, and the same after smoothing.  The lower two dashed lines show

1159: the solar activity level at minimum and maximum, and the upper line is the

1160: solar maximum activity shifted upwards by a factor of 40.  }

1161: \end{figure*}

1162: 

1163: \begin{figure*}

1164: \epsscale{0.8}

1165: %% \plotone{Figures/stellarnoise_pds_nykvist_w.eps}

1166: \plotone{f8.eps}

1167: \caption[]{\label{fig.pds.amp} Smoothed power density spectra (see

1168: Fig~\ref{fig.pds.harps}) for the nine telescopes shown in

1169: Fig.~\ref{fig.series}{\em a\/}, showing a similar background from stellar

1170: noise at low frequencies and different levels of white noise at high

1171: frequencies.  The color coding is the same as in Fig.~\ref{fig.series}. }

1172: \end{figure*}

1173: 

1174: \begin{figure*}

1175: \epsscale{0.8}

1176: %% \plotone{Figures/amplitudes.eps}

1177: \plotone{f9.eps}

1178: \caption[]{\label{fig.amp} Smoothed amplitude curves for Procyon for ten

1179:   telescopes, using the same color coding as Fig.~\ref{fig.series}. }

1180: \end{figure*}

1181: 

1182: \begin{figure*}

1183: \epsscale{0.8}

1184: %% \plotone{Figures/amp10-2d.eps}

1185: \plotone{f10.eps}

1186: \caption[]{\label{fig.amp.segments} Smoothed amplitude curves for Procyon

1187:   from ten 2-day segments of the combined time series (thin lines),

1188:   together with their mean (thick line).  }

1189: \end{figure*}

1190: 

1191: \begin{figure*}

1192: \epsscale{0.8}

1193: %%\plotone{Figures/amp12-log.epsi}

1194: \plotone{f11.eps}

1195: \caption[]{\label{fig.amp.stars} Smoothed amplitude curves for oscillations

1196:   in Procyon and other stars.

1197:   }

1198: \end{figure*}

1199: 

1200: 

1201: \end{document}

1202: 

1203: cp Figures/amp12-log.epsi revised/f11.eps

1204: 

1205: Torben Arentoft, Hans Kjeldsen, Timothy R. Bedding, Michael Bazot,

1206: Joergen Christensen-Dalsgaard, Thomas H. Dall, Christoffer Karoff, Fabien

1207: Carrier, Patrick Eggenberger, Danuta Sosnowska, Robert A. Wittenmyer,

1208: Michael Endl, Travis S. Metcalfe, Saskia Hekker, Sabine Reffert, R. Paul

1209: Butler, Hans Bruntt, Laszlo L. Kiss, Simon J. O'Toole, Eiji Kambe,

1210: Hiroyasu Ando, Hideyuki Izumiura, Bun'ei Sato, Michael Hartmann, Artie

1211: Hatzes, Francois Bouchy, Benoit Mosser, Thierry Appourchaux, Caroline

1212: Barban, Gabrielle Berthomieu, Rafael A. Garcia, Eric Michel, Janine

1213: Provost, Sylvaine Turck-Chieze, Milena Martic, Jean-Claude Lebrun,

1214: Jerome Schmitt, Jean-Loup Bertaux, Alfio Bonanno, Serena Benatti, Riccardo

1215: U. Claudi, Rosario Cosentino, Silvio Leccia, Soeren Frandsen, Karsten

1216: Brogaard, Lars Glowienka, Frank Grundahl and Eric Stempels

1217: 

1218: 

1219: We have carried out a multi-site campaign to measure oscillations in the F5

1220: star Procyon A.  We obtained high-precision velocity observations over more

1221: than three weeks with eleven telescopes, with almost continuous coverage

1222: for the central ten days.  This represents the most extensive campaign so

1223: far organized on any solar-type oscillator.  We describe in detail the

1224: methods we used for processing and combining the data.  These involved

1225: calculating weights for the velocity time series from the measurement

1226: uncertainties and adjusting them in order to minimize the noise level of

1227: the combined data.  The time series of velocities for Procyon shows the

1228: clear signature of oscillations, with a plateau of excess power that is

1229: centred at 0.9 mHz and is broader than has been seen for other stars.  The

1230: mean amplitude of the radial modes is 38.1 +/- 1.3 cm/s (2.0 times

1231: solar), which is consistent with previous detections from the ground and by

1232: the WIRE spacecraft, and also with the upper limit set by the MOST

1233: spacecraft.  The variation of the amplitude during the observing campaign

1234: allows us to estimate the mode lifetime to be 1.5 d (+1.9/-0.8 d).  We

1235: also find a slow variation in the radial velocity of Procyon, with good

1236: agreement between different telescopes.  These variations are remarkably

1237: similar to those seen in the Sun, and we interpret them as being due to

1238: rotational modulation from active regions on the stellar surface.  The

1239: variations appear to have a period of about 10 days, which presumably

1240: equals the stellar rotation period or, perhaps, half of it.  The amount of

1241: power in these slow variations indicates that the fractional area of

1242: Procyon covered by active regions is slightly higher than for the Sun.

1243: