astro-ph0604037/ms.tex
1: %\documentclass[manuscript]{aastex}
2: \documentclass[12pt,preprint]{aastex}
3: 
4: \shorttitle{Discovery of a new SPB: HD~163830 from MOST}
5: \shortauthors{Aerts et al.}
6: 
7: \begin{document}
8: 
9: \title{Discovery of the new slowly pulsating B star HD~163830  (B5II/III) from
10: MOST\altaffilmark{1} spacebased photometry}
11: 
12: \author{C.\ Aerts\altaffilmark{2,3}, 
13: P.\ De Cat\altaffilmark{4},
14: R.\ Kuschnig\altaffilmark{5},  
15: J.M.\ Matthews\altaffilmark{5}, 
16: D.B.\ Guenther\altaffilmark{6},  
17: A.F.J.\ Moffat\altaffilmark{7}, 
18: S.M.\ Rucinski\altaffilmark{8}, 
19: D.\ Sasselov\altaffilmark{9}, 
20: G.A.H.\ Walker\altaffilmark{10}, and 
21: W.W.\ Weiss\altaffilmark{11}}
22: 
23: \altaffiltext{1}{Based on data from the MOST satellite, a Canadian Space
24: Agency mission, jointly operated by Dynacon Inc., the University of Toronto
25: Institute for Aerospace Studies and the University of British Columbia, with the
26: assistance of the University of Vienna.}
27: 
28: \altaffiltext{2}{Institute of Astronomy, University of Leuven, Celestijnenlaan
29:   200 B, B-3001 Leuven, Belgium, email: conny@ster.kuleuven.be}
30: 
31: \altaffiltext{3}{Department of Astrophysics, Radboud University
32:   Nijmegen, P.O. Box  9010, 6500 GL Nijmegen, the Netherlands}
33: 
34: \altaffiltext{4}{Royal Observatory of Belgium, Ringlaan 3, B-1180 Brussels,
35:   Belgium} 
36: 
37: \altaffiltext{5}{Department of Physics and Astronomy, University of British
38: Columbia, 6224 Agricultural Road, Vancouver, BC V6T 1Z1, Canada}
39: 
40: \altaffiltext{6}{ Department of Astronomy and Physics, St. Mary's University,
41: Halifax, NS B3H 3C3, Canada}
42: 
43: \altaffiltext{7}{D\'epartement de Physique, Universit\'e de Montr\'eal,
44: C.P. 6128, Succursale Centre-Ville, Montr\'eal, QC H3C 3J7, Canada}
45: 
46: \altaffiltext{8}{David Dunlap Observatory, University of Toronto, P.O. Box 360,
47: Richmond Hill, ON L4C 4Y6, Canada}
48: 
49: \altaffiltext{9}{Harvard-Smithsonian Center for Astrophysics, 60 Garden Street,
50:   Cambridge, MA 02138, USA}
51: 
52: \altaffiltext{10}{1234 Hewlett Place, Victoria, BC  V8S 4P7, Canada}
53: 
54: \altaffiltext{11}{Institut f\"ur Astronomie, Universit\"at Wien,
55: T\"urkenschanzstrasse 17, A-1180 Wien, Austria}
56: 
57: 
58: \begin{abstract}
59: We report the discovery of a new slowly pulsating B (SPB) star, with the largest
60: number of detected frequencies to date by more than a factor of three, based on
61: 37 days of MOST (Microvariability \& Oscillations of STars) satellite guide star
62: photometry.  The star HD~163830 (V = 9.3, B5II/III) varies in twenty detected
63: frequencies in the range $0.035 - 1.06$ d$^{-1}$ ($0.4 - 12.3 ~ \mu$Hz) with
64: amplitudes from {0.7 to 7.6} mmag ($S/N$ from {4 to 43}).  Eighteen of these
65: frequencies are consistent with low-degree, high-order nonradial g-modes {of
66: seismic models of an evolved 4.5$M_{\odot}$ star. We are unable to identify one
67: unique model due to lack of mode identifications.}  The lowest two
68: frequencies may be associated with the rotation of HD~163830, but firm
69:   proof of this must await future spectroscopic data.
70: \end{abstract}
71: 
72: \keywords{stars: early-type; stars: individual (HD~163830); stars:
73: oscillations; stars: slowly pulsating B star (SPB)}
74: 
75: \section{Introduction}
76: 
77: Slowly pulsating B (SPB) stars (Waelkens 1991) show great promise for
78: asteroseismology.  Their g-mode pulsations should be sensitive probes of
79: their deep stellar interiors, since g-mode amplitudes remain large right
80: into the core (unlike p-modes whose amplitudes are significant only in
81: the outer layers).  But these stars also present serious observational
82: and theoretical challenges.  (1) Their pulsation periods are of order a
83: few days, making them difficult to disentangle in groundbased data due
84: to cycle/day aliasing.  (2) Their theoretical eigenspectra are so rich
85: (Dziembowski et al.\ 1993; Gautschy \& Saio 1993) that one cannot simply
86: find a unique match to a model by fitting the observed frequencies unless
87: there is prior information about the degrees and orders of the modes.
88: (3) Rotation is a serious complication for mode identification in these
89: stars because rotational splitting is large enough to cause multiplets of
90: adjacent radial orders to overlap.  The same great promise and obstacles
91: exist for other g-mode pulsators on the lower main sequence, like the
92: $\gamma$~Doradus variables.
93: 
94: Huge long-term groundbased observing campaigns (e.g., De Cat \& Aerts 2002 and
95: references therein for the SPBs, and Henry et al.\ 2005 and references therein
96: for the $\gamma$ Dor stars) had been successful in reliably discovering
97: multiperiodic variables. But the HIPPARCOS mission (Perryman et al.\ 1997)
98: demonstrated the power of spacebased observations to expand our understanding of
99: g-mode pulsators on the upper main sequence.  By mining the HIPPARCOS
100: photometric archive, Waelkens et al.\ (1998) and Handler (1999) discovered many
101: previously unknown SPB and $\gamma$ Dor stars, respectively. 
102: 
103: However, despite these efforts, the maximum number of well established
104: independent frequencies to be found in groundbased data of single SPB or
105: $\gamma$ Dor stars is five (see De Cat \& Aerts 2002; Poretti et al.\ 2002).
106: This paucity of observed frequencies compared to the densely populated
107: theoretical eigenspectra has made it difficult to even address issues like
108: rotational splitting.  Here, we report the detection of twenty frequencies in a
109: new SPB star discovered by the MOST space mission.  The WIRE satellite has
110: meanwhile also revealed numerous new SPB candidates with about five to fifteen 
111: frequencies (Br\"untt et al., in preparation).
112: 
113: 
114: \section{Observations and data reduction}
115: 
116: HD~163830 was one of 20 guide stars in the field for the MOST Primary
117: Science Target WR~103, observed in June 2005.  This field also contains
118: the guide star HD~163868 (B5Ve), in which the discovery by MOST of a
119: rich g-mode eigenspectrum made it the first in a new class of SPBe stars
120: (Walker et al.\ 2005).
121: The basic aspects of the guide star photometry in this field and the data
122: reduction procedure are already described by Walker et al.\ (2005).  A
123: constant star (at the level of precision of the measurements), HD~164388
124: (A2IV, V=7.9), was adopted as a comparison for the HD~163868 photometry,
125: and it is also used as a stable reference for HD~163830 in this paper.
126: 
127: HD~163830 was monitored for 36.6 days from 14 June to 21 July 2005, with
128: about 137,000 signal values in the original time series.  Data were
129: collected for only about 50 min of each 101-min orbit, since the target
130: field is outside the MOST {Continuous} Viewing Zone (CVZ) and was eclipsed
131: by the Earth for half of each orbit.  The final duty cycle of the time
132: series is about 38\%, after filtering of outliers and discarding data
133: contaminated by excess cosmic ray hits during satellite passages through
134: the South Atlantic Anomaly (SAA) in the Earth's magnetosphere.  For the
135: eventual frequency analysis, the data were binned into 2-min samples,
136: giving a total of 9890 points in the final light curve. The MOST
137: light curve of HD163830 is presented as dots in Fig.\,1. This plot
138: illustrates that the MOST data set far exceeds any previous photometric
139: data of known SPBs both in duty cycle and quality.
140: 
141: The exposure time of the star tracker was set to 1.5 sec.  As in the Direct
142: Imaging section of the MOST Science CCD (see Rowe et al.\ 2006), the Point
143: Spread Function (PSF) of the images has a Full Width Half Maximum (FWHM) of
144: about 2 pixels, where the focal plane scale is about 3 arcsec/pixel.  Only
145: pre-selected rasters of the CCD are read out, each 40x40 pixels in size, with a
146: guide star at the centre. The CCD parameters are: bias $\sim$550 ADU, gain
147: $\sim$20 electrons/ADU, and read noise $\sim$2.5 ADU/pixel RMS. A maximum of 20
148: such rasters can be chosen for a single guide field, due to downlink and onboard
149: software limitations. The brightness limits range for guide stars are from 6.5
150: mag to 10.5 mag.
151: 
152: Each raster/image is pre-processed automatically onboard. First, an
153: instantaneous background reading is determined by averaging the signals of the
154: first and last rows of the raster.  This subtracts much of the stray light
155: present in the image effectively, even if the scattered Earthshine changes
156: dramatically from exposure to exposure.  Then a fixed threshold value of 20 ADU
157: is added to the background; this number is subtracted from the signal from each
158: individual pixel. The threshold is applied to avoid introducting spurious
159: signals into the attitude control data processing.  As a consequence, the
160: contributions of the tails of the stellar PSF are not included in the
161: photometry.  The facts that no images are downloaded from the guide star data
162: and only the highest-signal pixels are included in the measurements combine to
163: limit the photometric precision to 2-2.5 times the photon noise limit.  However,
164: the long uninterrupted time series still provides low noise level results for
165: relatively faint guide stars such as HD~163830.
166: 
167: Finally, all pixels with a positive residual signal value (essentially the core
168: pixels of the PSF) are summed, representing a single intensity measurement per
169: exposure.  Residual background due to scattered Earthshine not subtracted by the
170: onboard procedure is filtered on the ground as described by Walker et al.\
171: (2005).
172: 
173: The MOST Startracker CCD must generally be operated at a sampling rate higher
174: than the Science CCD, in order to update the Attitude Control System. In the
175: case of the Primary Science Target WR~103, Science measurements were obtained at
176: 10-sec intervals, during which 5 guide star exposures were accumulated.  These 5
177: values were added and the cumulative signal was stored in the MOST Science Data
178: Stream, so the effective guide star sampling interval is also 10 sec.
179: 
180: 
181: \section{Frequency analysis}
182: 
183: We performed a Fourier time series analysis using Scargle's (1982) algorithm,
184: successively identifying the strongest peaks and removing them from the light
185: curve to search for additional frequencies in the residuals.  The spectral
186: window of these data is so clean due to the lack of long or periodic gaps in the
187: time series (see Fig.~1) that aliasing is not a concern.  The prewhitening was
188: done by 
189: performing a least squares fit to find the most likely value of the 
190: amplitude, phase and frequency, using
191: the height and position of the peak 
192: from the Scargle periodogram as starting value. The errors
193: derived in this way are formal errors which do not take into account the
194: correlation of the noise nor the dependence among the parameters. 
195: Hence they are an underestimation of the true errors
196: (Schwarzenberg-Czerny 1991).
197: 
198: From this process, we identified 21 frequencies with amplitude greater
199: than {3.9} S/N, where the S/N ratio was estimated as 
200: { the average peak amplitude in a
201: Scargle periodogram obtained after final prewhitening, }
202: in the range $0 - 2$ d$^{-1}$ where power was
203: detected (see Breger et al.\ 1993).   
204: This is considered to be {a rather strict significance criterion for the
205: known sample of SPBs (De Cat \& Cuypers 2003).}
206: One of the identified
207: frequencies, 2.0018 d$^{-1}$ is a known artifact in MOST data arising due to
208: modulation of stray Earthshine during the satellite's Sun-synchronous dusk-dawn
209: orbit (see Reegen et al.\ 2006). 
210: { The detection of 20 independent well-separated frequencies 
211: within $[0,1.1]$\,d$^{-1}$ was possible thanks to the high S/N level of the
212: peaks and the sinusoidal waveform of the modes.}
213: {The satellite orbital frequency 
214: peak occurs at 14.2 d$^{-1}$, i.e.\
215: well outside the frequency range seen in the target.}
216: 
217: {Scargle periodograms at a few different stages of prewhitening (omitting
218: the 2.0018 d$^{-1}$ artifact) are shown in Fig.~\ref{fig2}.  The final fit to
219: the MOST light curve of HD~163830 (including the artifact) is defined by:
220: \begin{equation}
221: y_i = a + \sum_{j=1}^{21} b_j \sin [2\pi (f_j t_i + \phi_j)],
222: \label{vgl}
223: \end{equation}
224: where the values for the free parameters $a, b_j, f_j, \phi_j$ are given in
225: Table~\ref{table1}.  The solid curve in Fig.~\ref{fig1} corresponds to this
226: solution.  The residuals after subtraction of this fit have a standard deviation
227: of 4.20 mmag and are also provided in Fig.~\ref{fig1}. Note that most of the
228: peaks listed in Table\,1 have an amplitude well below this value. Indeed, the
229: S/N ratio in the Scargle periodogram of the residuals amounts to only 0.18 mmag
230: mainly thanks to the reduction factor $\sqrt{9890}$.}
231: 
232: There is a hint for additional low-amplitude quasi-periodic
233: variability in the residuals. All the candidate frequencies found in them have
234: an amplitude below {3.6} S/N, however, so it is dangerous to make further
235: identifications or to assign formal errors from continued prewhitening.  We
236: ended the frequency identification conservatively when the formal $S/N$ dropped
237: below {3.9}, having high confidence in the frequencies and amplitudes in
238: Table~\ref{table1}.
239: 
240: 
241: \section{Seismic interpretation}
242: 
243: One high-quality archival 7-colour Geneva measurement is available for
244: HD~163830. We used it to derive the effective temperature and gravity of the
245: star following the method by K\"unzli et al.\ (1997). This resulted in $T_{\rm
246: eff}= 13\,700\pm 500$ K and $\log g=3.79\pm 0.14$ dex which is compatible with
247: the spectral type and luminosity class of B5II/III assigned by Houk (1982). As
248: there is no parallax estimate available for HD~163830, we interpolated in
249: standard models computed with the stellar evolution code CL\'ES (Scuflaire 2005)
250: and found a mass of 4.5$\pm$ 0.4$M_\odot$.
251: 
252: To compare the detected oscillation frequencies of HD~163830 with those
253: predicted by theory in the approximation of a non-rotating B-type star, we
254: computed seismic models from the ZAMS until core exhaustion using CL\'ES for 
255: different stellar masses and for $X=0.71$, $Z=0.015$, a mixing-length parameter of
256: 1.75 times the pressure scale height and no core overshooting.  We plot these
257: models in a $(\log T_{\rm eff},\log g$) diagram in Fig~\ref{fig3}. The
258: observational error box of HD~163830 is also shown, as well as the SPB
259: instability strip for modes of $\ell=1,2,3$.  Subsequently, we computed the
260: frequencies of the zonal modes {of all the $M=4.5\,M_\odot$ models on the MS} for 
261: $\ell=1,2,3$ and tested the excitation of
262: these modes with the non-adiabatic code MAD (Dupret 2001). The outcome of these
263: instability computations is
264: presented in Fig.~\ref{fig4} for the $\ell=1,2$ modes. The spectrum for $\ell=3$
265: modes is even denser and is omitted here for brevity. We also show the observed
266: frequencies (except the stray light frequency and the lowest ones of {0.035} and
267: {0.079}~d$^{-1}$) and the upper limit of the effective temperature of
268: HD~163830. It can be seen that there are numerous $\ell=1,2$ modes excited in
269: these models and that they can explain well the observed frequencies of
270: HD~163830, particulary since we only show the zonal modes in Fig.~\ref{fig4}. We
271: thus interprete the variability of our target in terms of high-order low-degree
272: non-radial g-modes as expected in SPBs.  Given the expected large number of
273: excited modes, some of the residual power seen in Fig.\,\ref{fig1} is probably
274: due to g-modes we have not been able to adequately resolve from one another.
275: {We stress that we did not perform a matching between the observed
276:   frequencies and theoretical ones of the models, because we lack information on
277:   the $m$-values of the modes and on the stellar rotation frequency. 
278: Fig.~\ref{fig4} only shows that the observed spectrum is compatible with
279: theoretical models in the appropriate area of the HR diagram.} 
280: 
281: 
282: The frequencies $f_{11}=0.035$~d$^{-1}$ and $f_{20}=0.079$~d$^{-1}\approx
283: 2f_{11}$ are clearly of a different character, being an order of magnitude
284: smaller than the other frequencies we have identified.  A possible explanation
285: for these frequencies is that they are associated with the star's rotational
286: frequency $\Omega$. Unfortunately, the corresponding period of {28.6} days is only
287: slightly less than the total time coverage of the MOST photometry of HD~163830,
288: so it is not safe to say whether this variation even repeats periodically. Such
289: low frequencies have also been seen in other SPBs (De Cat \& Aerts 2002) where
290: they are difficult to interpret in terms of pulsation.
291: 
292: Some of the adjacent frequencies of HD~163830 listed in Table\,\ref{table1} show
293: spacings in the range $0.03 - 0.04$ d$^{-1}$, and are compatible with rotational
294: splitting, since high-order g-modes produce multiplets with spacings $m[1 -
295: 1/{\ell}(\ell + 1)]\Omega$.  However, without prior mode identifications of
296: (${\ell},m$) it is impossible to decide whether these are multiplet components
297: or isolated modes.  De Cat et al.\ (2005) have shown that the oscillation
298: frequencies in the corotating frame of the star can differ substantially from
299: the same frequencies seen in the observer's frame, even in the case of slow
300: rotation and low-degree modes.  If $\Omega = 0.035$ d$^{-1}$, then one would
301: expect adjacent frequencies for $\ell = 1$ to be spaced by only 0.0175~d$^{-1}$;
302: for $\ell = 2$ by 0.029~d$^{-1}$.  Yet, as can be seen in Fig.\,\ref{fig4}, even
303: independent eigenfrequencies of adjacent radial order and the same degree in a
304: nonrotating model show separations comparable to this.  For these reasons, it is
305: premature to try to find a more specific match to {one particular} 
306: seismic model until there is
307: independent information about the rotation rate of the star or constraints on
308: the mode values, both of which will require spectroscopic follow up.
309: 
310: 
311: 
312: \section{Summary}
313: 
314: We discovered twenty intrinsic frequencies in the 37-day MOST light curve of the
315: little studied B5II/III star HD~163830. We interpreted eighteen of these
316: frequencies in terms of high-order non-radial g-modes in an SPB, thus adding
317: HD~163830 to this class of oscillators. It becomes at once the SPB with the
318: largest number of detected oscillation frequencies. We provided a seismic model
319: that is compatible with the observed frequency spectrum. This model is only one
320: of numerous other possibilities. In order to fine-tune our seismic analysis, we
321: need empirical mode identification, either from multicolour photometry or from
322: line-profile variations.
323: 
324: 
325: Our results illustrate the gain of nearly continuous photometry compared to
326: ground data in disentangling the frequency spectra of the g-modes in
327: SPBs. Similar results are to be expected for $\gamma$~Doradus stars.  We can
328: therefore anticipate numerous new results from space-based observatories that
329: are capable to gather long continuous strings of data from space, such as MOST,
330: WIRE, COROT and Kepler.
331: 
332: 
333: \acknowledgments
334: 
335: CA and PDC are indebted to Richard Scuflaire, Marc-Antoine Dupret and Mario
336: Ausseloos for providing their software codes CLES, MAD and SCAN. CA is supported
337: by the Research Council of the K.U.Leuven under grant GOA/2003/04. JMM, DBG,
338: AFJM, SR and GAHW acknowledge funding from the Natural Sciences \& Engineering
339: Research Council (NSERC) Canada.  RK's work is supported by the Canadian Space
340: Agency.  WWW received funding from the Austrian
341: Forschungsfo\"rderungsgesellschaft (FFG-ALR) and the Science Fonds (FWF-P17580).
342: 
343: \begin{thebibliography}{}
344: 
345: \bibitem[]{} {Breger, M., Stich, J., Garrido, R., et al.\ 1993, A\&A 271, 482}
346: 
347: \bibitem[]{} De Cat, P., Aerts, C.\ 2002, A\&A 393, 965
348: 	
349: \bibitem[]{} De Cat, P., Briquet, M., Daszy\'nska-Daszkiewicz, J., Dupret,
350: M. A., de Ridder, J., Scuflaire, R., Aerts, C.\ 2005, A\&A 432, 1013
351: 
352: \bibitem[]{} De Cat, P., Cuypers, J.\ 2003, ASPC 292, 377
353: 
354: \bibitem[]{} Dupret, M.-A., 2001, A\&A 366, 166
355: 	
356: %\bibitem[]{} Dupret, M.-A., Grigahc\`ene, A., Garrido, R., Gabriel, M.,
357: %Scuflaire, R.\ 2005, A\&A 435, 927
358: 
359: \bibitem[]{} Dziembowski, W.A., Moskalik, P., Pamyatnykh, A. A.\ 1993, MNRAS
360: 265, 588
361: 	
362: \bibitem[]{} Gautschy, A., Saio, H.\ 1993, MNRAS 262, 213
363: 
364: \bibitem[]{} Handler, G.\ 1999, MNRAS 309, L19
365: 
366: \bibitem[]{} Henry, G.W., Fekel, F.C., Henry, S.M.\  2005, Astron.\ J.
367:  129, 2815
368: 
369: \bibitem[]{} Houk, N.\ 1982, Michigan Spectral Survey, Univ. Michigan, 3 
370: 
371: \bibitem[]{} K\"unzli M., North P., Kurucz R.L., Nicolet B., 1997, 
372:    A\&AS 122, 51
373: 
374: %\bibitem[]{} Montgomery, M.H., O'Donohue, D.\ 1999, Delta Scuti Newsletter 13,
375: %  28 
376: 
377: \bibitem[]{} Perryman, M.A.C., et al.\ 1997, The Hipparcos and Tycho Catalogues
378: (ESA SP-1200, Noordwijk: ESA)
379: 
380: \bibitem[]{} Poretti, E., Koen, C., Bossi, M., Rodr\'{\i}guez, E., Mart\'{\i}n,
381: S., Krisciunas, K., Akan, M. C., Crowe, R., Wilcox, M., Ibanoglu, C., Evren, S.\
382: 2002, A\&A 384, 513
383: 
384: \bibitem[]{} Reegen, P., Kallinger, T., Frast, D., et al., 2006, MNRAS,
385: in press
386: 
387: \bibitem[]{} Rowe, R., Matthews, J.M., Kuschnig, R., Seager, S., Sasselov, D.,
388: Guenther, D.B., Moffat, A.F.J., Rucinski, S.M., Walker, G.A.H., Weiss,
389: W.W.\ 2006, ApJ, submitted
390: 
391: \bibitem[]{} Scargle, J.D.\ 1982, ApJ 263, 835
392: 
393: \bibitem[]{} {Schwarzenberg-Czerny, A.\ 1991, MNRAS 253, 198}
394: 
395: \bibitem[]{} Scuflaire, R.\ 2005, CL\'ES User Manual, Version 18, Universit\'e
396:   de Li\`ege, Belgium
397: 
398: \bibitem[]{} Waelkens, C.\ 1991, A\&A 246, 453
399: 
400: \bibitem[]{} Waelkens, C., Aerts, C., Kestens, E., Grenon, M., Eyer, L.\ 1998,
401: A\&A 330, 215
402: 
403: \bibitem[]{} 
404: Walker, G.A.H, Kuschnig, R., Matthews, J.M., Cameron, C., Saio, H., Lee,
405: U., Kambe, E., Masuda, S., Guenther, D.B., Moffat, A.F.J., Rucinski,
406: S.M., Sasselov, D., Weiss, W.W.\ 2005, ApJ 635, L77
407: 
408: 
409: \end{thebibliography}
410: 
411: \clearpage
412: 
413: \begin{figure}
414: \begin{center}
415: \rotatebox{270}{\resizebox{12cm}{!}{\includegraphics{f1.ps}}}
416: \end{center}
417: \caption[]{Top: the observed light curve of HD~163830 (dots) is compared with
418: the final fit given in Eq.\,(1) (full line). Bottom: the residuals after
419: subtraction of the final fit.}
420: \label{fig1}
421: \end{figure}
422: 
423: \clearpage
424: 
425: \begin{figure}
426: \begin{center}
427: \rotatebox{270}{\resizebox{10cm}{!}{\includegraphics{f2.ps}}}
428: \end{center}
429: \caption[]{Scargle periodogram of the MOST data of HD~163830 (upper panel),
430:   after prewhitening with the 15 dominant frequencies (middle panel) and after
431:   prewhitening with the 21 frequencies listed in Table\,1 (lower panel). The
432:   dotted vertical lines indicate the frequency values listed in Table\,1 from
433:   least-squares fitting. The dashed horizontal lines indicate the 3.6\,S/N
434:   level.
435: Note the different scale of the y-axes.
436: The stray light frequency at 2.0039 d$^{-1}$ is omitted
437:   for clarity.}
438: \label{fig2}
439: \end{figure}
440: 
441: \clearpage
442: 
443: \begin{figure}
444: \begin{center}
445: \rotatebox{270}{\resizebox{6cm}{!}{\includegraphics{f3.ps}}}
446: \end{center}
447: \caption[]{Evolutionary models (dashed and dotted lines) in steps of
448:   0.1\,$M_\odot$ computed with CL\'ES from the ZAMS until core exhaustion. The
449:   masses for the models in dashed lines are given in solar values. The models
450:   whose frequency spectrum is given in Fig.~4 are indicated as a full black
451:   line. The observational error box of HD~163830 is represented by the
452:   rectangle. The instability strip computed from MAD is indicated by thick
453:   lines and represents the region in which modes of $\ell=1,2,3$ are excited. }
454: \label{fig3}
455: \end{figure}
456: 
457: \clearpage
458: 
459: \begin{figure}
460: \begin{center}
461: \rotatebox{270}{\resizebox{10.25cm}{!}{\includegraphics{f4a.ps}}}
462: \rotatebox{270}{\resizebox{10.25cm}{!}{\includegraphics{f4b.ps}}}
463: \end{center}
464: \caption[]{Predicted frequencies for $\ell=1$ (top) and $\ell=2$ (bottom) modes
465:   for main-sequence stellar models with parameters 4.5$M_\odot$, $X=0.71$,
466:   $Z=0.015$, mixing length $\alpha=1.75$ pressure scale heights and no core
467:   overshooting. Excited modes are indicated with full red symbols, stable modes
468:   with open black symbols (colors only in the online version of the paper). The
469:   observed frequencies of HD~163830 are indicated as full horizontal
470:   lines. HD~163830 is situated to the right of the vertical line.}
471: \label{fig4}
472: \end{figure}
473: 
474: \clearpage
475: 
476: \begin{deluxetable}{cccccc}
477: \tablecaption{Light curve solution of HD~163830, according to the
478: 21 terms given in Eq.\,(\protect\ref{vgl}) besides $a=921.56(5)$ mmag.  
479: The reference epoch for the phases
480: $\phi_j$ is the time of first observation (HJD\,2453536.33498108). 
481: {The quoted formal errors underestimate the true variance because 
482: they were derived from a least-squares fit
483:   ignoring the correlation of the noise and the dependence among 
484: the parameters.}
485: Frequency
486: $f_{13}$ is due to the detection of stray light. }
487: \tablehead{\colhead{ } & 
488: \colhead{$f_j$ (d$^{-1}$)} & \colhead{$f_j$ ($\mu$Hz)} &
489: \colhead{$b_j$ (mmag)} & \colhead{$\phi_j$} & \colhead{S/N}} \startdata
490: $f_{ 1}$ &  0.6744(5) &  7.806(5)  &  7.6(2)  & 0.125(8)  & 41.4 \\ 
491: $f_{ 2}$ &  0.621(2)  &  7.19(2)   &  5.5(5)  & 0.34(3)   & 30.0 \\
492: $f_{ 3}$ &  0.647(4)  &  7.48(4)   &  2.8(4)  & 0.46(5)   & 15.3 \\ 
493: $f_{ 4}$ &  0.756(1)  &  8.75(1)   &  2.38(9) & 0.37(3)   & 13.0 \\
494: $f_{ 5}$ &  0.8007(9) &  9.27(1)   &  2.2(1)  & 0.15(2)   & 12.0 \\
495: $f_{ 6}$ &  0.412(1)  &  4.77(2)   &  1.64(8) & 0.50(2)   &  8.9 \\
496: $f_{ 7}$ &  0.949(1)  & 10.980(7)  &  1.40(6) & 0.45(2)   &  7.6 \\
497: $f_{ 8}$ &  0.576(2)  &  6.67(3)   &  1.4(6)  & 0.17(4)   &  7.6 \\
498: $f_{ 9}$ &  1.0546(7) & 12.206(9)  &  1.37(6) & 0.53(2)   &  7.5 \\
499: $f_{10}$ &  0.521(2)  &  6.03(2)   &  1.31(7) & 0.53(3)   &  7.1 \\
500: $f_{11}$ &  0.0350(8) &  0.405(9)  &  1.18(7) & 0.92(2)   &  6.4 \\
501: $f_{12}$ &  0.448(2)  &  5.18(3)   &  1.17(7) & 0.44(4)   &  6.4 \\
502: $f_{13}$ &  2.0018(8) & 23.17(1)   &  1.11(6) & 0.21(2)   &  6.0 \\
503: $f_{14}$ &  0.718(4)  &  8.31(5)   &  0.9(1)  & 0.61(8)   &  4.9 \\
504: $f_{15}$ &  0.257(2)  &  2.98(3)   &  0.9(1)  & 0.25(5)   &  4.9 \\
505: $f_{16}$ &  0.832(2)  &  9.63(3)   &  0.87(8) & 0.49(5)   &  4.5 \\
506: $f_{17}$ &  0.285(3)  &  3.30(3)   &  0.83(9) & 0.20(5)   &  4.5 \\
507: $f_{18}$ &  0.494(3)  &  5.72(3)   &  0.80(7) & 0.25(4)   &  4.4 \\
508: $f_{19}$ &  0.343(1)  &  3.97(2)   &  0.77(6) & 0.49(3)   &  4.2 \\
509: $f_{20}$ &  0.079(2)  &  0.91(2)   &  0.72(6) & 0.71(3)   &  3.9 \\
510: $f_{21}$ &  0.921(2)  & 10.66(2)   &  0.72(7) & 0.66(4)   &  3.9 \\
511: \enddata\label{table1}
512: \end{deluxetable}
513: \end{document}
514: