1: \documentclass[12pt,preprint]{aastex} % for submission
2: %\documentclass[apjl]{emulateapj} % 2-col, single-spaced
3: %\usepackage{apjfonts,mathptmx}
4:
5: \newcommand{\kms}{km\,s$^{-1}$}
6: \newcommand{\ms}{m\,s$^{-1}$}
7: \newcommand{\hr}{HR\,3831}
8: \newcommand{\vb}{\mbox{\boldmath$V$}}
9:
10: \shorttitle{Indirect Imaging of Stellar Pulsations}
11: \shortauthors{Kochukhov}
12:
13: \begin{document}
14:
15: %\title{Indirect Imaging of Stellar Pulsations\footnotemark[1]}
16: \title{Indirect Imaging of Nonradial Pulsations in a Rapidly Oscillating Ap Star\footnotemark[1]}
17: \footnotetext[1]{Based on observations obtained at the European Southern Observatory, La Silla, Chile}
18:
19: \author{Oleg Kochukhov}
20: \affil{Institut f\"ur Astronomie, Universit\"at Wien, T\"urkenschanzstra{\ss}e 17, 1180 Wien, Austria}
21: \email{kochukhov@astro.univie.ac.at}
22:
23:
24: \begin{abstract}
25: Many types of stars show periodic variations of radius and brightness,
26: which are commonly referred to as `stellar pulsations'. Observed
27: pulsational characteristics are determined by fundamental stellar
28: parameters. Consequently, investigations of stellar pulsations provide
29: a unique opportunity to verify and refine our understanding of the
30: evolution and internal structure of stars. However, a key boundary
31: condition for this analysis -- precise information about the geometry
32: of pulsations in the outer stellar envelopes -- has been notoriously
33: difficult to secure. Here we demonstrate that it is possible to solve
34: this problem by constructing an `image' of the pulsation velocity field
35: from time series observations of stellar spectra. This technique is
36: applied to study the geometry of nonradial pulsations in a prototype
37: magnetic oscillating (roAp) star \hr. Our velocity map directly
38: demonstrates an alignment of pulsations with the axis of the global
39: magnetic field and reveals a significant magnetically induced
40: distortion of pulsations. This observation constitutes a long-sought
41: solution of the problem of the pulsation geometry of roAp stars and
42: enables very stringent tests of the recent theories of stellar
43: magneto-acoustic oscillations.
44: \end{abstract}
45:
46: \keywords{line: profiles --- stars: chemically peculiar ---
47: stars: imaging --- stars: individual (\hr) --- stars: oscillations}
48:
49: \section{Introduction}
50:
51: A classical solution \citep{uoa89} of hydrodynamic equations expresses
52: monoperiodic nonradial pulsational fluctuations of an idealized spherically
53: symmetric star in terms of a spherical harmonic function. Approximations of
54: this simple stellar pulsation theory are adequate for a significant fraction of
55: oscillating stars and, therefore, historically investigations of stellar
56: pulsation geometry were limited to assigning spherical harmonic `quantum
57: numbers' $\ell$ and $m$ to each observed pulsation mode. This problem is solved
58: by several varieties of the mode identification technique \citep*{apw92,w88},
59: which, with different degrees of success, are applied to photometric and
60: spectroscopic observations of pulsating stars. However, pulsation modes in real
61: stars are often significantly distorted by rapid stellar rotation \citep{ls90}
62: and global magnetic field \citep{dg85,bd02,sg04}. These effects result in a
63: considerably more complicated pulsation geometry, irreducible to a single
64: spherical harmonic function. Up to now no objective technique was available for
65: a comprehensive and reliable analysis of such distorted nonradial pulsations,
66: but it was understood that a detailed study of high-resolution stellar spectra
67: is the most promising approach to resolve these difficulties.
68:
69: In general, all inhomogeneities in the outer layers of stellar envelopes,
70: including pulsational disturbances, produce distortions in the stellar line
71: profiles. As a star rotates and starspots appear and disappear from the visible
72: hemisphere, spectral line shapes and strengths are modulated periodically.
73: Doppler imaging (DI, e.g., Vogt, Penrod, \& Hatzes \citeyear{vph87}) is a powerful
74: technique to decipher rich information about the physics and geometry of the
75: surface structures contained in the stellar spectrum variability. The Doppler
76: mapping method is based on the resolution of the stellar surface provided by
77: the effect of stellar rotation on the shapes of absorption spectral lines. If
78: the projected rotational velocity of a star is large enough for the shapes of
79: spectral lines to be dominated by the rotational Doppler broadening, there
80: exists a well-defined mapping relation between distortions in a line profile
81: and positions of the corresponding inhomogeneities on the stellar surface.
82: Thus, a single Doppler-broadened line profile represents a one-dimensional
83: projection of the visible stellar hemisphere, with all the structure blurred
84: along the lines of constant Doppler shift. Spectroscopic observations at
85: different rotation phases can be used to monitor the evolution of dips and
86: bumps in line profiles, thus obtaining a set of non-redundant projections of a
87: stellar surface map, and eventually to reconstruct a two-dimensional image from
88: a time series of line profile observations. This indirect stellar surface
89: imaging method was first applied to the problem of mapping chemical
90: inhomogeneities in early-type stars \citep{krw86} and was subsequently extended
91: to investigation of temperature spots \citep{vph87} and complex magnetic
92: topologies \citep{s89} in active late-type stars and global magnetic fields
93: \citep{pk02} in chemically peculiar stars. \citet{vp83} have emphasized that
94: rotational broadening of stellar spectra also aids in probing the horizontal
95: structure of nonradial pulsations. However, subsequent spectroscopic studies
96: of rapidly rotating oscillating stars \citep{gk88,kwm92,h98} used only
97: resolution of pulsation structures along stellar longitudes and restrictively
98: approximated velocity field with a single spherical harmonic.
99:
100: In a recent development Berdyugina, Telting, \& Korhonen (\citeyear{btk03}) presented a
101: new exploration of the capabilities of the DI technique in its application to mapping
102: of stellar nonradial pulsations. These authors employed a temperature DI code in
103: modeling pulsational line profile variations and assumed that pulsational fluctuations
104: at the stellar surface can be represented by a rotation of a fixed pattern. As
105: emphasized by \citet{k04}, the latter assumption restricts pulsational mapping to the reconstruction
106: of a spherical harmonic geometry, which is aligned with the stellar rotation axis and
107: is characterized by a unique non-zero azimuthal number $m$. This condition is not
108: fulfilled for zonal modes and is invalid for all types of oblique stellar pulsations,
109: hence the method of \citet{btk03} cannot be applied to study oscillations of roAp
110: stars.
111:
112: Here we present the first application of an alternative technique.
113: %developed by \citet{k04}.
114: In our approach the principle of Doppler imaging is extended to the
115: reconstruction of the time-dependent velocity field. The temporal and surface variation
116: of the velocity vector are represented with a superposition of the two constant
117: surface distributions:
118: \begin{displaymath}
119: \vb(t,\theta,\phi) = \vb^c(\theta,\phi) \cos(\omega t) +
120: \vb^s(\theta,\phi) \sin(\omega t),
121: \end{displaymath}
122: where $\omega$ is the pulsation frequency and
123: $\theta$, $\phi$ are usual spherical coordinates on the stellar surface. The $\vb^c$
124: and $\vb^s$ vector maps are recovered directly from the observed line profile
125: variability without imposing any specific global constraints on the pulsation geometry.
126: This is equivalent to mapping a two-dimensional surface distribution of the pulsation
127: amplitude and phase for each velocity component.
128:
129: The foundations of the pulsational Doppler mapping and description of
130: its computer implementation were presented by \citet{k04}. We refer the reader
131: to this paper for a detailed explanation of the technique and discussion of
132: the numerical simulations which were used to evaluate performance and intrinsic
133: limitations of the new surface mapping method.
134:
135: \section{Pulsational Doppler mapping of \hr}
136:
137: The pulsation Doppler inversion is applied to time-resolved
138: observations of the \ion{Nd}{3} $\lambda$~6145~\AA\ line in the spectrum of the
139: well-known roAp star \hr\ (HD\,83368). A total of 1860 spectra of this object
140: were obtained over the period of 11 nights using the Coud\'e Echelle
141: Spectrograph fiber-linked to the Cassegrain focus of the 3.6-m telescope at the
142: European Southern Observatory. A preliminary discovery report of the
143: pulsational variations in individual spectral lines in \hr\ using these
144: observational data was published by \citet{kr01}. Our observations evenly
145: sample the 2\fd851976 rotation period of the star \citep{kwr97} and are
146: characterized by the resolving power of $\lambda/\Delta\lambda=123,000$ and the
147: signal-to-noise ratio of 110--160 pixel$^{-1}$. A 70$^{\rm s}$ exposure time
148: was chosen to ensure an appropriate sampling of the 11\fm67
149: oscillation period. Fig.\,\ref{fig1} shows an example of the rapid line profile
150: variation of \hr\ and presents a comparison between the observed and
151: calculated spectra.
152:
153: Reconstruction of the pulsation velocity field of \hr\ took into account an
154: inhomogeneous surface distribution of neodymium which significantly distorts
155: the mean \ion{Nd}{3} line shapes. The maps
156: of chemical abundance and pulsational fluctuations were recovered
157: simultaneously in a self-consistent manner. The resulting Nd abundance map agrees
158: very well with the distribution derived by \citet{kdp04} using a different DI
159: code.
160:
161: Nonradial pulsations in roAp stars are characterized by very small
162: photometric amplitudes. The respective pulsational temperature variation
163: does not exceed $\sim$\,10~K \citep{mk98}. Such temperature changes do not have
164: a noticeable effect on the stellar line profiles and can be safely neglected in
165: our modeling.
166:
167: The formation region of the \ion{Nd}{3} $\lambda$~6145~\AA\ line is located
168: high in the atmospheres of roAp stars \citep{rpk02}. The horizontal motion due
169: to the high-overtone {\it p-}mode pulsation is expected to be either negligible
170: \citep{bd02} or considerably smaller \citep{sg04} than the radial motion at this
171: atmospheric height. Trial mapping of the full vector pulsation velocity
172: distribution demonstrated that, whereas including the horizontal pulsation
173: motion substantially increases the number of free parameters, it does not lead
174: to an appreciable improvement of the fit to observations. Taking this into
175: account we limited the pulsation DI reconstruction to the vertical velocity
176: component which is represented by the $V^c_r$ and $V^s_r$ velocity amplitudes.
177:
178: %\begin{figure}[t*]
179: %%\epsscale{0.5}
180: %\plotone{f1.eps}
181: %\caption{
182: %Observed variation of the \ion{Nd}{3} $\lambda$~6145~\AA\ line ({\it histogram}) is compared with the
183: %best fit synthetic spectra ({\it solid line}). The left panel shows the
184: %observed and synthetic spectra
185: %for 20 phases (out of 1860 used in the inversion) close to the
186: %maximum of pulsational fluctuations (rotation phases $\varphi=0.966$-0.973 according to the ephemeris of
187: %Kurtz et al. \citeyear{kwr97}). The right panel presents the difference between
188: %individual profiles and the average line shape. The spectra
189: %for consecutive pulsation phases are shifted in the vertical direction. The bars
190: %at the bottom of the panels show the vertical scale in units of the continuum
191: %flux.\label{fig1}}
192: %\end{figure}
193:
194: The Doppler maps of the pulsation velocity field of \hr\ revealed with our technique
195: are illustrated in Fig.\,\ref{fig2}. The $V^c_r$ map shows a clear oblique dipolar
196: pattern with the local pulsation amplitude reaching up to 4.0~\kms. On the other hand,
197: the amplitude of the stuctures in the second map, $V^s_r$, does not exceed a few hundred
198: \ms. Such a relation of the two maps is typical of nearly axisymmetric pulsations
199: \citep[see Sect 5.5 of][]{k04}.
200:
201: %\begin{figure*}[t*]
202: %%\epsscale{0.95}
203: %%\plotone{f2.eps}
204: %\plotone{f2_color.eps}
205: %\caption{
206: %Results of the first Doppler imaging reconstruction of the stellar
207: %pulsation velocity field.
208: %Rectangular panels show the $V^c_r$ ({\it left}) and $V^s_r$ ({\it right})
209: %vertical velocity amplitude maps.
210: %The contours of equal velocity are plotted over the greyscale images with a step of 1.0
211: %\kms. The horizontal dotted line shows the lowest visible latitude
212: %corresponding to the angle between the stellar axis of rotation and the
213: %observer's line of sight (`inclination angle', $i=68\degr$) adopted in the
214: %analysis of \hr. The plots below present spherical projection of the
215: %respective pulsation velocity maps. The star is shown at 4 different aspect
216: %angles corresponding to the rotation phases $\varphi=0.00$, 0.25, 0.50, and 0.75.
217: %The grid at the stellar surface is plotted with a 30\degr\ step in longitude and
218: %latitude.
219: %\label{fig2}}
220: %\end{figure*}
221:
222: \section{Discussion}
223:
224: Indirect imaging of the surface structure of nonradial oscillations has important
225: implications for stellar astrophysics. Results of the first Doppler inversion
226: of the pulsation velocity presented here can be regarded as an ultimate solution of the
227: long-standing problem of the pulsation geometry of roAp stars. Until now the
228: high-overtone {\it p-}mode oscillations in these unique objects were
229: investigated primarily with the high-speed photometric techniques. Analyses of
230: light variations of roAp stars \citep{k90} and, in particular, \hr\
231: \citep{kwr97}, demonstrated that the amplitude and phase of pulsations are
232: modulated with the stellar rotation. A coincidence of the times of maxima of pulsation
233: amplitude and magnetic field extrema was interpreted within the oblique
234: pulsator framework \citep{k82}. This phenomenological model postulated that
235: pulsations are represented by an axisymmetric dipolar ($\ell=1$) mode inclined
236: with respect to the stellar rotation axis and aligned with the axis of the
237: global magnetic field. Subsequent theoretical studies were carried out by
238: \citet{st93}, \citet{ts95} and, most recently, by \citet{sg04} assuming a
239: dominant role of the magnetic field. These calculations suggested that
240: important deviations of the mode geometry from the purely dipolar topology
241: might occur, although pulsation structure was still expected to remain almost
242: axisymmetric and aligned with the magnetic field. A totally different picture
243: has emerged from the theoretical investigation by \citet{bd02} who focused on
244: the essential role of the centrifugal force resulting from the stellar
245: rotation. According to their theory pulsations in roAp stars are described by a
246: superposition of all $\ell=1$ harmonic components and are neither axisymmetric, nor
247: necessarily aligned with the magnetic field. Limited information content of the
248: photometric observations did not allow astronomers to distinguish between these
249: competing approaches to the roAp pulsation structure.
250:
251: The surface velocity mapping technique is not limited to the spherical
252: harmonic parameterization of pulsational fluctuations and can be used
253: to explore the physics of nonradial oscillations in rotating magnetic stars.
254: In particular, it is able to supply a crucial observational constraint required
255: for testing the alternative theories of roAp pulsations. The pulsation map
256: (Fig.\,\ref{fig2}) reconstructed here for the prototype roAp star \hr\
257: displays an obvious correlation with the stellar magnetic field geometry.
258: \citet{kdp04} found that the dominant dipolar component of the global field
259: in \hr\ is inclined by about 90\degr\ relative to the stellar rotation axis
260: and is oriented at $\approx180\degr$ in longitude. At the same time, the
261: $V^c_r$ velocity map presented in Fig.\,\ref{fig2} indicates that pulsation
262: geometry of \hr\ is dominated by an axisymmetric structure. Furthermore,
263: it is clear that the two maxima of the pulsation amplitude coincide with
264: the magnetic poles. This is the first independent verification of the
265: alignment of pulsations and magnetic field in oscillating magnetized
266: stars.
267:
268: Our pulsational mapping of \hr\ suggests a dominant role of the
269: magnetic perturbation of the {\it p-}modes and considerably less important
270: influence of the stellar rotation. Indeed, the pulsation Doppler image
271: of the surface velocity field reveals no significant asymmetry which can be
272: attributed to the oblique $\ell=1$, $|m|=1$ components.
273: Therefore, we conclude that pulsations in \hr\ cannot be described
274: by the theory brought forward by \citet{bd02}, who attempted to fully take into
275: account the rotationally induced mode perturbations but limited their
276: calculations to a relatively weak magnetic field.
277:
278: Fig.\,\ref{fig3} illustrates an analysis of the dominant axisymmetric
279: pulsation structure reconstructed with our technique. This figure shows the
280: latitudinal dependence of the $V^c_r$ pulsation amplitude in the reference frame of
281: the stellar magnetic field. The inferred latitudinal trend is clearly
282: different from the one expected for a pure oblique $\ell=1$ pulsation. Axisymmetric
283: multipolar decomposition of the observed latitudinal dependence suggests
284: that oscillations in \hr\ can be roughly described with a superposition of
285: $\ell=1$ and $\ell=3$ harmonic components. The octupolar component has the
286: same sign as the dipolar contribution and an amplitude smaller by a factor
287: of two (A$_{\ell=3}$/A$_{\ell=1}$=$0.50\pm0.04$). In other words, we
288: discover that pulsations in \hr\ are strongly confined to the magnetic field
289: axis. This pulsation geometry is in remarkable agreement with the
290: predictions obtained in recent calculations assuming the dominant role of
291: the magnetic field \citep{sg04}. This concordance of the empirical model and theory
292: documents a significant breakthrough in our understanding of the physics of
293: {\it p-}mode oscillations in magnetic stars and possibly confirms an
294: important role of damping by the magnetic slow waves investigated by
295: \citet{sg04} in the context of excitation of the roAp-type pulsations.
296:
297: In addition to revealing the true pulsation geometry of roAp stars, our
298: precise velocity map recovered for \hr\ opens numerous possibilities for
299: asteroseismological determination of various stellar parameters. For
300: instance, \citet{ts95} argued that the presence of a large quadrupolar
301: component in the magnetic field geometry is expected to induce $\ell=0$ and
302: $\ell=2$ axisymmetric structures in the pulsation velocity distribution.
303: However, Fig.\,\ref{fig3} convincingly demonstrates that the even multipolar
304: components are very weak in \hr. Hence, the magnetic field of this star
305: must be basically dipolar -- a conclusion obtained independently from an
306: analysis of any classical magnetic field observables.
307:
308: \acknowledgements
309: This work was supported by the Lise Meitner fellowship granted by the Austrian
310: Science Fund (FWF, project No. M757-N02). We acknowledge extensive use of
311: the SIMBAD database, operated at CDS, Strasbourg, France.
312:
313: %\begin{figure}
314: %%\epsscale{0.7}
315: %\plotone{f3.eps}
316: %\caption{
317: %Analysis of the latitudinal dependence of the $V^c_r$ pulsation velocity
318: %amplitude. The symbols show average
319: %velocity amplitude as a function of the cosine of the angle $\theta$ counted from one
320: %of the pulsation poles. The solid line illustrates the fit with a superposition
321: %of axisymmetric $\ell=0$, 1, 2, and 3 spherical harmonic components. Contributions of
322: %individual multipolar components are shown by the dashed lines for $\ell=1$--3 and by
323: %the dotted line for $\ell=0$. \label{fig3}}
324: %\end{figure}
325:
326: \begin{thebibliography}{}
327: \bibitem[Aerts, De Pauw, \& Waelkens(1992)]{apw92}
328: Aerts, C., De Pauw, M., \& Waelkens, C. 1992, \aap, 266, 294
329: \bibitem[Berdyugina et al.(2003)]{btk03}
330: Berdyugina, S. V., Telting, J. H., \& Korhonen, H. 2003, \aap, 406, 273
331: \bibitem[Bigot \& Dziembowski(2002)]{bd02}
332: Bigot, L., \& Dziembowski, W. A. 2002, \aap, 391, 235
333: \bibitem[Dziembowski \& Goode(1985)]{dg85}
334: Dziembowski, W., \& Goode, P. R. 1985, \apj, 296, L27
335: \bibitem[Gies \& Kullavanijaya(1988)]{gk88}
336: Gies, D. R., \& Kullavanijaya, A. 1988, \apj, 326, 813
337: \bibitem[Hao(1998)]{h98}
338: Hao, J. 1998, \apj, 500, 440
339: \bibitem[Kennelly, Walker, \& Merryfield(1992)]{kwm92}
340: Kennelly, E. J., Walker, G. A. H., \& Merryfield, W. J. 1992, \apj, 400, L71
341: \bibitem[Khokhlova, Rice, \& Wehlau(1986)]{krw86}
342: Khokhlova, V. L., Rice, J. B., \& Wehlau, W. H. 1986, \apj, 307, 768
343: \bibitem[Kochukhov(2004)]{k04}
344: Kochukhov, O. 2004, \aap, 423, 613
345: \bibitem[Kochukhov et al.(2004)]{kdp04}
346: Kochukhov, O., Drake, N. A., Piskunov, N., \& de la Reza, M.
347: 2004, \aap, 424, 935
348: \bibitem[Kochukhov \& Ryabchikova(2001)]{kr01}
349: Kochukhov, O. \& Ryabchikova, T. 2001, \aap, 377, L22
350: \bibitem[Kurtz(1982)]{k82}
351: Kurtz, D. W. 1982, \mnras, 200, 807
352: \bibitem[Kurtz(1990)]{k90}
353: Kurtz, D. W. 1990, \araa, 28, 607
354: \bibitem[Kurtz et al.(1997)]{kwr97}
355: Kurtz, D. W., van Wyk, F., Roberts, G., Marang, F., Handler, G.,
356: Medupe, R., \& Kilkenny, D. 1997, \mnras, 287, 69
357: \bibitem[Lee \& Saio(1990)]{ls90}
358: Lee, U., \& Saio, H. 1990, \apj, 349, 570
359: \bibitem[Medupe \& Kurtz(1998)]{mk98}
360: Medupe, R., \& Kurtz, D. W. 1998, \mnras, 229, 371
361: \bibitem[Piskunov \& Kochukhov(2002)]{pk02}
362: Piskunov, N., \& Kochukhov, O. 2002, \aap, 381, 736
363: \bibitem[Ryabchikova et al.(2002)]{rpk02}
364: Ryabchikova, T. A., Piskunov, N. E., Kochukhov, O., Tsymbal, V., Mittermayer, P.,
365: \& Weiss, W. W. 2002, \aap, 384, 545
366: \bibitem[Saio \& Gautschy(2004)]{sg04}
367: Saio, H., \& Gautschy, A. 2004, \mnras, 350, 485
368: \bibitem[Shibahashi \& Aerts(2000)]{sa00}
369: Shibahashi, H., \& Aerts, C. 2000, \apj, 531, L143
370: \bibitem[Shibahashi \& Takata(1993)]{st93}
371: Shibahashi, H., \& Takata, M. 1993, \pasj, 45, 617
372: \bibitem[Semel(1989)]{s89}
373: Semel, M. 1989, \aap, 225, 456
374: \bibitem[Takata \& Shibahashi(1995)]{ts95}
375: Takata, M., \& Shibahashi, H. 1995, \pasj, 41, 219
376: \bibitem[Unno et al.(1989)]{uoa89}
377: Unno, W., Osaki, Y., Ando, H., Saio, H., \& Shibahashi, H.
378: 1989, Nonradial Oscillations of Stars, (2d ed., Tokyo: University of Tokyo Press)
379: \bibitem[Vogt \& Penrod(1983)]{vp83}
380: Vogt, S. S., \& Penrod, G. D. 1983, \apj, 275, 661
381: \bibitem[Vogt et al.(1987)]{vph87}
382: Vogt, S. S., Penrod, G. D., \& Hatzes, A. P. 1987, \apj, 321, 496
383: \bibitem[Watson(1988)]{w88}
384: Watson, R. D. 1988, \apss, 140, 255
385: \end{thebibliography}
386:
387: \clearpage
388:
389: \begin{figure}
390: \epsscale{0.5}
391: \plotone{f1.eps}
392: \caption{
393: Observed variation of the \ion{Nd}{3} $\lambda$~6145~\AA\ line ({\it histogram}) is compared with the
394: best fit synthetic spectra ({\it solid line}). The left panel shows the
395: observed and synthetic spectra
396: for 20 phases (out of 1860 used in the inversion) close to the
397: maximum of pulsational fluctuations (rotation phases $\varphi=0.966$-0.973 according to the ephemeris of
398: Kurtz et al. \citeyear{kwr97}). The right panel presents the difference between
399: individual profiles and the average line shape. The spectra
400: for consecutive pulsation phases are shifted in the vertical direction. The bars
401: at the bottom of the panels show the vertical scale in units of the continuum
402: flux.\label{fig1}}
403: \end{figure}
404:
405: \clearpage
406:
407: \begin{figure}
408: \epsscale{0.95}
409: \plotone{f2_color.eps}
410: \caption{
411: Results of the first Doppler imaging reconstruction of the stellar
412: pulsation velocity field.
413: Rectangular panels show the $V^c_r$ ({\it left}) and $V^s_r$ ({\it right})
414: vertical velocity amplitude maps.
415: The contours of equal velocity are plotted over the greyscale images with a step of 1.0
416: \kms. The horizontal dotted line shows the lowest visible latitude
417: corresponding to the angle between the stellar axis of rotation and the
418: observer's line of sight (`inclination angle', $i=68\degr$) adopted in the
419: analysis of \hr. The plots below present spherical projection of the
420: respective pulsation velocity maps. The star is shown at 4 different aspect
421: angles corresponding to the rotation phases $\varphi=0.00$, 0.25, 0.50, and 0.75.
422: The grid at the stellar surface is plotted with a 30\degr\ step in longitude and
423: latitude.
424: \label{fig2}}
425: \end{figure}
426:
427: \clearpage
428:
429: \begin{figure}
430: \epsscale{0.7}
431: \plotone{f3.eps}
432: \caption{
433: Analysis of the latitudinal dependence of the $V^c_r$ pulsation velocity
434: amplitude. The symbols show average
435: velocity amplitude as a function of the cosine of the angle $\theta$ counted from one
436: of the pulsation poles. The solid line illustrates the fit with a superposition
437: of axisymmetric $\ell=0$, 1, 2, and 3 spherical harmonic components. Contributions of
438: individual multipolar components are shown by the dashed lines for $\ell=1$--3 and by
439: the dotted line for $\ell=0$. \label{fig3}}
440: \end{figure}
441:
442: \end{document}
443:
444: