147,99 €
This book surveys our understanding of stars which change in brightness because they pulsate. Pulsating variable stars are keys to distance scales inside and beyond the Milky Way galaxy. They test our understanding not only of stellar pulsation theory but also of stellar structure and evolution theory. Moreover, Pulsating Stars are important probes of the formation and evolution of our own and neighboring galaxies. Our understanding of Pulsating Stars has greatly increased in recent years as large-scale surveys of Pulsating Stars in the Milky Way and other Local Group galaxies have provided a wealth of new observations and as space-based instruments have studied particular Pulsating Stars in unprecedented detail.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 982
Related Titles
Title Page
Copyright
Dedication
Preface
Chapter 1: Historical Overview
1.1 Discovery of the First Pulsating Variable Stars
1.2 The Recognition of Pulsation as a Cause of Variability
Chapter 2: Fundamentals of Stellar Variability Observations
2.1 Definitions
2.2 Photometric Bandpasses
2.3 Period Determination
2.4 Common Observational Techniques
2.5 Space-Based Versus Ground Observations
Chapter 3: Classification of Variable Stars
3.1 Regular, Semi-Regular, and Irregular Variables
3.2 Variability: Intrinsic and/or Extrinsic
3.3 Extrinsic Variables
3.4 Intrinsic Variables
Chapter 4: Stellar Structure and Evolution Theory
4.1 The Basic Equations of Stellar Structure and Evolution
4.2 The Evolution of Low-Mass Stars
4.3 The Evolution of Intermediate-Mass Stars
4.4 The Evolution of High-Mass Stars
Chapter 5: Stellar Pulsation Theory
5.1 Timescales
5.2 Ritter's (Period–Mean Density) Relation
5.3 Basic Equations of (Radial) Stellar Pulsation Theory
5.4 Linearization of the Stellar Pulsation Equations
5.5 Linear Adiabatic Oscillations: The LAWE
5.6 Eigenvalues and Eigenfunctions of the LAWE
5.7 Non-Adiabatic Theory: Conditions for Stability
5.8 The Linear Non-Adiabatic Wave Equation
5.9 Driving Mechanisms
5.10 Stability Conditions and Instability Strip Edges
5.11 Non-Radial Pulsations
5.12 Nonlinear Effects
Chapter 6: RR Lyrae Stars
6.1 RR Lyrae Stars as a Class of Pulsating Variable Star
6.2 RR Lyrae Stars as Standard Candles
6.3 Evolution of RR Lyrae Stars
6.4 Pulsation
6.5 The Blazhko Effect
6.6 RR Lyrae Stars in Globular Clusters
6.7 The Oosterhoff Groups
6.8 Period Changes
6.9 RR Lyrae Stars beyond the Milky Way
Chapter 7: Cepheid and Related Variable Stars
7.1 Classical Cepheids
7.2 Type II Cepheids
7.3 BL Boo Stars or Anomalous Cepheids
7.4 RV Tauri Stars
Chapter 8: Red Variable Stars
8.1 Convection and Pulsation
8.2 Mira and Related Long-Period Variables
8.3 Semi-Regular Variables
8.4 Irregular Variables
Chapter 9: Pulsating Stars Close to the Lower Main Sequence in the H-R Diagram
9.1 δ Scuti and SX Phoenicis Stars
9.2 γ Doradus Stars
9.3 roAp Stars
Chapter 10: Pulsating Stars Close to the Upper Main Sequence in the H-R Diagram
10.1
β
Cephei Stars
10.2 SPB (53 Per) Stars
Chapter 11: Pulsating Supergiant Stars
11.1 SPBsg Variables
11.2 PV Telescopii, V652 Herculis, and R CrB Stars
11.3 α Cygni, S Dor, and Wolf-Rayet Stars
Chapter 12: Hot Subdwarf Pulsators
12.1 EC 14026 (V361 Hya, sdBV, sdBV
p
, sdBV
r
) Variables
12.2 PG 1716+426 (V1093 Her, “Betsy,” sdBV
g
, sdBV
s
) Variables
12.3 sdOV (V499 Ser) Variables
12.4 He-sdBV Stars
Chapter 13: Pulsating Degenerate Stars
13.1 GW Vir Stars
13.2 DBV (V777 Her) Stars
13.3 DQV Stars
13.4 DAV Stars
13.5 ELM-HeV Stars
13.6 GW Librae Stars: Accreting WD Pulsators
13.7 Pulsations in Neutron Stars and Black Holes
Glossary
References
Index
End User License Agreement
ii
iv
v
xi
xii
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
263
264
265
266
267
268
269
270
271
272
273
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
Cover
Table of Contents
Preface
Begin Reading
Figure 1.1
Figure 1.2
Figure 1.3
Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4
Figure 2.5
Figure 2.6
Figure 3.1
Figure 3.2
Figure 4.1
Figure 4.2
Figure 4.3
Figure 4.4
Figure 4.5
Figure 5.1
Figure 5.2
Figure 5.3
Figure 5.4
Figure 5.5
Figure 5.6
Figure 5.7
Figure 5.8
Figure 5.9
Figure 5.10
Figure 5.11
Figure 5.12
Figure 5.13
Figure 5.14
Figure 5.15
Figure 5.16
Figure 5.17
Figure 5.18
Figure 5.19
Figure 5.20
Figure 5.21
Figure 5.22
Figure 5.23
Figure 5.24
Figure 5.25
Figure 5.26
Figure 5.27
Figure 5.28
Figure 5.29
Figure 5.30
Figure 5.31
Figure 5.32
Figure 5.33
Figure 5.34
Figure 5.35
Figure 6.1
Figure 6.2
Figure 6.3
Figure 6.4
Figure 6.5
Figure 6.6
Figure 6.7
Figure 6.8
Figure 6.9
Figure 6.10
Figure 6.11
Figure 6.12
Figure 6.13
Figure 6.14
Figure 6.15
Figure 6.16
Figure 7.1
Figure 7.2
Figure 7.3
Figure 7.4
Figure 7.5
Figure 7.6
Figure 7.7
Figure 7.8
Figure 7.9
Figure 7.10
Figure 7.11
Figure 8.1
Figure 8.2
Figure 8.3
Figure 8.4
Figure 8.5
Figure 8.6
Figure 8.7
Figure 8.8
Figure 8.9
Figure 8.10
Figure 9.1
Figure 9.2
Figure 9.3
Figure 9.4
Figure 9.6
Figure 9.5
Figure 9.7
Figure 9.8
Figure 10.1
Figure 10.2
Figure 10.3
Figure 10.4
Figure 11.1
Figure 11.2
Figure 11.3
Figure 11.4
Figure 11.5
Figure 12.1
Figure 12.2
Figure 12.3
Figure 12.4
Figure 12.5
Figure 12.6
Figure 12.7
Figure 13.1
Figure 13.2
Figure 13.3
Figure 13.4
Figure 13.5
Figure 13.6
Figure 13.7
Figure 13.8
Figure 13.9
Figure 13.10
Figure 13.11
Figure 13.12
Figure 13.13
Figure 13.14
Figure 13.15
Figure 13.16
Table 1.1
Table 2.1
Table 5.1
Table 5.2
Table 9.1
Table 10.1
Table 11.1
Table 12.1
Table 13.1
Shore, S.N.
The Tapestry of Modern Astrophysics
2013
ISBN: 978-0-471-16816-4
Stepanov, A.V., Zaitsev, V.V., Nakariakov, V.M.
Coronal Seismology
Waves and Oscillations in Stellar Coronae
2012
Print ISBN: 978-3-527-40994-5 Also available in digital formats.
Foukal, P.V.
Solar Astrophysics
2nd Edition
2004
Print ISBN: 978-3-527-40374-5 Also available in digital formats.
Stahler, S.W., Palla, F.
The Formation of Stars
2004
Print ISBN: 978-3-527-40559-6 Also available in digital formats.
Márcio Catelan and Horace A. Smith
Authors
Prof. Márcio Catelan
Instituto de Astrofísica
Facultad de Física
Pontificia Universidad Católica de Chile
Santiago
Chile
and
Millennium Institute of Astrophysics
Santiago
Chile
Prof. Horace A. Smith
Dept. of Physics & Astronomy
Michigan State University
East Lansing, Michigan
United States
Cover
NASA, ESA, and the Hubble Heritage Team (STScI/AURA)-Hubble/Europe Collaboration
Acknowledgment: H. Bond (STScI and Penn State University)
ESA/Hubble & ESO
NASA, ESA, and the Hubble Heritage Team (STScI/AURA)
Acknowledgement: William Blair (Johns Hopkins University)
This image, obtained by NASA's Galaxy Evolution Explorer (GALEX) space mission, shows the comet-like tail left behind by the pulsating star Mira (o Ceti) as it loses mass and moves across the interstellar medium. (Image courtesy NASA.)
All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.
Library of Congress Card No.: applied for
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library.
Bibliographic information published by the Deutsche Nationalbibliothek
The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at <http://dnb.d-nb.de>.
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA,
Boschstr. 12, 69469 Weinheim, Germany
All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law.
Print ISBN: 978-3-527-40715-6
ePDF ISBN: 978-3-527-65521-2
ePub ISBN: 978-3-527-65520-5
Mobi ISBN: 978-3-527-65519-9
oBook ISBN: 978-3-527-65518-2
Cover Design Adam-Design, Weinheim, Germany
Printing and Binding Markono Print Media Pte Ltd, Singapore
To Carolina, Debbie, Gabriela, and Amanda, the driving mechanisms of our hearts' pulsations.
In this book, we introduce the readers to the history, observations, and theory of pulsating stars as they are known in the early twenty-first century. We are long past the time when a single book can recite all that is known about these stars. Our hope is to give the reader an overview of the different varieties of pulsating variable stars, what causes their pulsation, and how they fit into the general scheme of stellar evolution. The book grew in part from courses that we have taught over the years at our respective universities, and we thank the students and postdocs who have participated in those courses for helping us along the way.
Many researchers in this field are credited in our book, but there are many others, both living and dead, whose contributions to our understanding of pulsating stars must pass without explicit acknowledgement. We nonetheless appreciate our debt to those unnamed students of stellar variability.
Particular thanks for their assistance in the creation of this book is owed to authors who gave us permission to reproduce figures from their papers, and most especially those who have even volunteered updated versions of those figures, including J. Alonso-García, L. G. Althaus, A. H. Córsico, T. Credner, P. Degroote, L. Eyer, G. Hajdu, J. J. Hermes, C. S. Jeffery, and A. Mukadam, among many others. The original source of each borrowed figure is indicated within the figure caption. We also thank Acta Astronomica, AlltheSky.com, the American Association of Variable Star Observers, the American Astronomical Society, the American Institute of Physics, the American Physical Society, the Astronomical Society of the Pacific, Annual Reviews, Astronomy & Astrophysics, Astrophysics & Space Science, Cambridge University Press, the Carnegie Observatories, Communications in Asteroseismology, the European Physical Journal, the European Space Agency, the European Southern Observatory, the International Astronomical Union, the Institute of Physics (IOP), Katholieke Universiteit Leuven, Oxford University Press, the Royal Astronomical Society, the National Aeronautics and Space Administration, Nature, Springer, and the University of Chicago Press for granting us permission to reproduce figures from publications to which they own the copyright. We would also like to express our gratitude to our editor, Nina Stadthaus, for her patience, and to our families, for their unwavering support. Most especially, we thank M. E. Escobar and A. V. Sweigart for their careful proof-reading of the entire book.
We also wish to acknowledge the different agencies that have sponsored our research while we worked on this book. H.A.S. thanks the U.S. National Science Foundation for its important assistance to his research over many years. M.C. acknowledges the support of Chile's Ministry for the Economy, Development, and Tourism's Programa Iniciativa Científica Milenio through grant IC 120009, awarded to the Millennium Institute of Astrophysics; of Proyecto Basal PFB-06/2007; and of Fondecyt grants #1110326 and 1141141. M.C. is particularly grateful to the Simon Guggenheim Memorial Foundation, for awarding a Guggenheim Fellowship that allowed portions of this book to be written during his sabbatical year; to Pontificia Universidad Católica de Chile, for awarding him with sabbatical leave; and to those institutions that kindly hosted him on that occasion, including Michigan State University, Catholic University of America, NASA's Goddard Space Flight Center, Università di Bologna (Italy), and Universidade Federal do Rio Grande do Norte (Brazil).
Santiago and East Lansing
March 2014
Márcio CatelanHorace A. Smith
Although cataclysmic variable stars of the nova or supernova type had been seen since antiquity (Stephenson & Green, 2002), by the sixteenth century stars were generally regarded as fixed and unchanging in both position and brightness (Hoskins, 1982). The outburst of a bright supernova in Cassiopeia in 1572 (Tycho's supernova) startled the astronomical community and reawakened interest in apparently new stars. Almost 14 years later, in 1596, David Fabricius (1564–1617) observed what he thought was yet another new star, this time in the constellation Cetus.
Fabricius's new star was only of the third magnitude, far less brilliant than Tycho's supernova, but nonetheless easily visible to the unaided eye. First seen in August, by October the star had faded below naked-eye visibility. However, Fabricius's star had not forever vanished. A few years later, it was recorded by Johann Bayer, who named the star omicron (o) Ceti and placed it on his 1603 star charts, although Bayer appears to have been unaware that he had rediscovered Fabricius's nova. In 1609, it was Fabricius himself who was surprised to see the new star make a reappearance. While o Ceti certainly seemed unlike Tycho's new star of 1572, Fabricius did not suspect that he had discovered a star that was not a nova but one that instead showed periodic changes in brightness (Hoskins, 1982).
In 1638, Johannes Holwarda (1618–1651) made yet another independent discovery of o Ceti. Like Fabricius, Holwarda watched the star fade from view, only to see it subsequently reappear. Now other astronomers began to pay increased attention to the variable, but nonetheless the periodic nature of its variability still eluded recognition.
Johannes Hevelius (1611–1687) carried out a detailed study of o Ceti, and, in 1662, he published Historiola Mirae, naming the star Mira (meaning the wonderful in Latin), a name which has ever since been applied to it. Ismael Bullialdus (1605–1694) made the next advance. He noticed that the peak brightness of Mira occurred about a month earlier each year, finally discovering the cyclic nature of its brightness changes. In his 1667 Ad astronomos monita duo he determined the period to be 333 days, about 1 day longer than the current determination of the mean period. Mira thus became the first variable star whose period was determined, and the archetype of the class of long-period variable stars now known as Mira variables. It would nonetheless be a long while before it was established that Mira's brightness changes had anything to do with pulsation (Hoskins, 1982).
Mira itself can become as bright as second or third magnitude at maximum light, making it easily visible to the naked eye (Figure 1.1). Other long-period variables, although not as bright as Mira, also have peak magnitudes that place them within the bounds of naked-eye visibility. That begs the question of whether Mira or other long-period variables might have been discovered before the time of Fabricius. Hoffleit (1997) summarized the evidence for pre-1596 observations of Mira. A number of possible pre-1596 observations have been suggested, going back to ancient times. Unfortunately, for none of the proposed pre-discovery observations is the historical evidence ironclad. The case is perhaps strongest for a 1592 “guest star” recorded in Asian records. However, Stephenson & Green (2002) concluded that even that object was unlikely to have actually been Mira.
Figure 1.1 Mira near maximum (a) and minimum (b) light. The image (a) was taken from Römerstein, Germany, on July 17, 2004, whereas the image (b) was obtained on October 30, 2005, from Austria. Several of the nearby stars in Cetus are identified by their Bayer designations, and the Cetus constellation itself is also drawn. The bright object on the upper left corner in (b) is the planet Mars. The July 2004 image presents a view similar to that seen by David Fabricius at his discovery of that long-period variable star in August, 1596. (Images courtesy Till Credner, AlltheSky.com.)
The number of confirmed periodic variable stars increased only slowly following the recognition of the periodicity of Mira. For instance, in 1686 Gottfried Kirch discovered the variability of χ Cygni, a Mira variable with a period of 408 days (Sterken, Broens, & Koen, 1999). R Hydrae, a 384-day-period Mira variable, was found by Maraldi in 1704. A third Mira variable, 312-day-period R Leonis, was found by Koch in 1782. The long-period variability of these stars, and of Mira itself, would eventually be attributed to pulsation. However, little was known of shorter-period variable stars when the British team of Edward Pigott and John Goodricke took up the study of variable stars in earnest late in the eighteenth century.
Pigott and Goodricke soon confirmed the variability of Algol1 (β Persei), which had been originally established by Geminiano Montanari, and determined its period. Goodricke, in a 1783 report to the Royal Society, suggested that an eclipse might be responsible for the periodic dimming of Algol's light. Pigott discovered the variability of η Aquilae in 1784, and soon thereafter Goodricke identified the variables β Lyrae and δ Cephei. While β Lyrae proved to be an eclipsing variable star, η Aquilae and δ Cephei were the first representatives of the important Cepheid class of pulsating stars. The circumstance that Goodricke was deaf (and possibly mute) did not slow the progress of the pair, but Goodricke's untimely death at the age of 21 ended the collaboration in 1786 (Hoskins, 1982; French, 2012).
A list of the earliest discovered variable stars and their discovery dates is shown in Table 1.1, from which cataclysmic variables such as novae or supernovae have been excluded. The first column of this table identifies the variable star by name, while the second indicates into which category the variable star was eventually classified. The discovery date of the variable star is given in the third column, followed by the name of the discoverer. In some cases, there are questions as to what should be called the actual discovery date: the date at which variability was first suspected or the date at which the variability was clearly established. The period of the variable, if it is periodic, is listed in the fifth column, while the brightest visual magnitude and amplitude of the variable are given in columns six and seven. Finally, column eight lists the spectral type of the variable. Note that all of the variable stars in this table reached naked-eye visibility, with the possible exception of S Serpentis, which at its peak brightness is just fainter than the usual magnitude limit for the naked eye.
Table 1.1 Some of the first variable stars discovered.a),b),c)
Star
Type
Year
Discoverer
P
(d)
V
max
A
V
SpT
o
Ceti
Mira
1596
Fabricius
331.96
2.0
8.1
M5e-M9e
P Cygni
S Dor
1600
Blaeu
—
3.0
3.0
B1Iapeq
β
Persei
Algol
1667
Montanari
2.8673043
2.12
1.27
B8V
η Carinae
d
S Dor
1677
Halley
—
−0.8
8.7
pec(e)
χ Cygni
Mira
1686
Kirch
408.05
3.3
10.9
S6,2e-S10,4e(MSe)
R Hydrae
Mira
1704
Maraldi
388.87
3.5
7.4
M6e-M9eS(Tc)
R Leonis
Mira
1782
Koch
309.95
4.4
6.9
M6e-M8IIIe-M9.5e
μ Cephei
SRc
1782
W. Herschel
730
3.43
1.67
M2eIa
β Lyrae
β Lyr
1784
Goodricke
12.913834
3.25
1.11
B8II-IIIep
δ Cephei
Cepheid
1784
Goodricke
5.366341
3.48
0.89
F5Ib-G1Ib
η Aquilae
e
Cepheid
1784
Pigott
7.176641
3.48
0.91
F6Ib-G4Ib
ι Bootis
W UMa
1785
W. Herschel
0.2678159
5.8
0.6
G2V+G2V
R Coronae Borealis
R CrB
1795
Pigott
—
5.71
9.09
C0,0(F8pep)
R Scuti
RV Tau
1795
Pigott
146.5
4.2
4.4
G0Iae-K2p(M3)Ibe
α Herculis
SRc
1795
W. Herschel
—
2.74
1.26
M5Ib-II
R Virginis
Mira
1809
Harding
145.63
6.1
6.0
M3.5IIIe-M8.5e
R Aquarii
Mira, Z And
1810
Harding
390
5.2
7.2
M5e-M8.5e+pec
Aurigae
Algol
1821
Fritsch
9892
2.92
0.91
A8Ia-F2epIa+BV
R Serpentis
Mira
1826
Harding
356.41
5.16
9.24
M5IIIe-M9e
S Serpentis
Mira
1828
Harding
371.84
7.0
7.1
M5e-M6e
R Cancri
Mira
1829
Schwerd
361.60
6.07
5.73
M6e-M9e
α Orionis
SRc
1836
J. Herschel
2335
0.0
1.3
M1-M2Ia-Ibe
a The properties of these stars, as given in the last four columns, are taken from the online edition of the General Catalogue of Variable Stars (Kholopov et al., 1998).
b Excluding novae and supernovae; see, e.g., Zsoldos (1994) for a historical explanation.
c R Delphini is incorrectly shown in Table 3 of Petit (1987) with a discovery date of 1751. The star's variability was actually discovered in 1851 by M. Hencke.
d Enlightening reviews of historical observations of this star are provided by de Vaucouleurs (1952) and Smith & Frew (2011).
e η Aquilae will also be found in early writings as η Antinoi, after the (now defunct) constellation of Antinous. Antinous was the name of Roman emperor Hadrian's favorite and posthumously deified lover, turned into a constellation by Imperial decree.
The perceptive reader will have noticed, from Table 1.1, that the first variable stars to be discovered have names that begin either with a Greek letter or with an R. The Greek letters follow the original nomenclature proposed by the German lawyer and amateur astronomer Johann Bayer (1603) in his groundbreaking atlas entitled Uranometria.2 In Bayer's atlas the stars were assigned letters of the Greek alphabet, then, when those were exhausted, letters of the Roman alphabet, first in lower case, then starting anew in upper case, until the letter Q was reached – the last one to be used in Bayer's atlas. In all cases, the letter is followed by the Latin genitive form of the constellation's name. Hence β Persei, for instance, corresponds to Algol, the second brightest star in the constellation Perseus. It is a common misperception that the Bayer letters are always ordered by apparent magnitude, with the brightest star in a constellation assigned the Greek letter α, the second brightest β, and so on, but this is an oversimplification. In Bayer's days precise stellar photometry was not available. Bayer usually ordered his letters according to the traditional magnitude groups (magnitude 1 including the brightest stars and magnitude 6 those just visible to the naked eye). However, he did not order the stars by apparent brightness within each magnitude. Thus, although α is usually the brightest star within a constellation, that is not invariably so (Swerdlow, 1986).
As several of these stars were eventually shown to be variable, their original Bayer name was retained. However, the discovery of fainter variable stars that were not included in Bayer's original catalog required a new convention, or at least an extension to the Bayer scheme. This was provided by the Director of the Bonn Observatory, Friedrich Wilhelm August Argelander, when he produced the Uranometria Nova (Argelander, 1843), followed by the monumental Bonn Observatory Survey (the famous Bonner Durchmusterung, or BD).
Argelander reasoned that it would be only natural to continue the lettering sequence from Bayer, and thus named the first variable star detected in a constellation that did not have a Bayer symbol with the letter R, again followed by the Latin genitive of the constellation's name. As a matter of fact, an explanation that R stood for a continuation of the Bayer symbols, rather than, for instance, the first letter of “Rot” (“Red” in German, or “Rouge” in French, or “Rojo” in Spanish), as had often been speculated – not an unreasonable hypothesis, since so many of the stars in Table 1.1 are very red – appears to have been given by Argelander only a few years later, in his article on R Virginis published in the Astronomische Nachrichten (Argelander, 1855). Thus, for instance, R Leonis was the first variable star without a Bayer letter to be cataloged in the constellation Leo. One natural drawback of Argelander's proposed scheme was that, from R to Z – the available letters of the alphabet for variable star classification – there were only nine available “slots,” and thus no more than nine new variable stars could be classified in this way. Argelander, however, never thought this could pose a serious problem (Townley, 1915), since in those days stellar variability was still considered a rare phenomenon – and indeed, a mere 18 variable stars in total were included in the table of variable stars compiled by Argelander.3
Naturally, the number of reported variable stars quickly increased over the 18 originally reported in the BD, and by 1912 the number of known variables was already around 4000 (Cannon, 1912). Ironically, it was partly Argelander's invention of the “step method” of visual estimation of the magnitudes of variable stars (see Hearnshaw, 1996) that was responsible for the increased number of variability reports around the world, as quantitative estimates of stellar magnitudes became more widespread among amateur astronomers in particular. Naturally, the other key development that helped make the number of reported variable stars skyrocket was the development of the photographic plate. One way or another, the fatal limit of nine variable stars per constellation was quickly surpassed, and therefore an extension to Argelander's scheme became necessary.
To achieve this goal, Hartwig suggested, at a meeting of the Astronomische Gesellschaft in 1881, that double letters – still in the range between R and Z – be used (Townley, 1915). Thus, after the variable star Z in a given constellation, to the next variable star the letters RR would be assigned, followed by RS, RT, and so on, until RZ – followed by SS, ST, …, SZ, then TT, TU, …, TZ, and so on, and ending with ZZ. This gives an additional 45 slots, thus bringing the grand total to 54. Thus, RR Lyrae is the tenth variable star (without a Bayer letter) in the constellation Lyra, whereas ZZ Ceti is the fifty-fourth such star in Cetus. This scheme seems to have first been adopted by Chandler (1888) in his variable star catalog, where RR Virginis can be found.
Soon enough – more specifically, with the discovery of ZZ Cygni in 1907 – even 54 slots became insufficient, and so astronomers eventually resorted to the remaining letters of the alphabet, following a suggestion by the astronomer Ristenpart to the Astronomische Gesellschaft. In order to avoid confusion with the Bayer letters, however, two such letters were always used. From A to Q, the only letter that was not utilized was J, in order to avoid confusion with I.4 At the time – and again reflecting an amazing shortsightedness of the astronomical community in the late 1800s and early 1900s – it was widely believed that even the letters QZ would likely never be reached in any constellation (Schweitzer & Vialle, 2000).
Including these letters, all combinations starting with AA, AB, …, AZ, BB, BC, …, BZ, until ZZ become available (again except for those involving the letter J). Since there are 25 combinations involving the letter A, 24 involving letter B, 23 involving letter C, and so on and so forth, down to a single one for letter Z (i.e., ZZ), we find, based on the expression
with , that a total of 325 additional “slots” can be obtained in this way – which, on top of the original nine single-letter ones, gives a grand total of 334. This is the maximum afforded by the expanded Argelander scheme.
Even before 334 variables were first found in a constellation – which finally happened in 1929, with the discovery of QZ Sagittarii – some authors realized that this too would be insufficient, and therefore alternative, more intuitive naming schemes were suggested. In particular, Nijland (1914) and Townley (1915) suggested, following Chambers (1865), that it would be much more reasonable to simply adopt a “V” followed by the order of discovery of the variable – and this is the approach that has been followed until the present, starting with variable 335 in any given constellation.5 The first star to be classified in this way was then V335 Sagittarii. The next several constellations to “overflow” the original 334 variable star name slots were, in succession, Ophiuchus (1929), Cygnus (1933), Centaurus and Scorpius (1936), and Aquila (1937). According to the latest available edition of the General Catalogue of Variable Stars (GCVS; Kholopov et al., 1998), as presented by Kazarovets et al. (2009), the constellation with the highest number of cataloged variables is Sagittarius, with 5579 stars, followed by Ophiuchus, with 2671. On the other hand, many constellations are still far from exceeding 334 classified variables; for instance, in the constellation Caelum (the Chisel), the list extends only as far as SX Cae (i.e., no more than 24 stars), whereas in Equuleus (the Little Horse) the list reaches TX Equ (31 stars). This reflects the fact that constellations located close to the Galactic plane tend to contain many more variable stars than those at higher Galactic latitudes.
To avoid confusion, variable stars are usually given a provisional name upon classification, before entering the GCVS – at which point its naming is considered definitive. For a description of how provisional names are usually chosen, the reader is referred to Schweitzer & Vialle (2000).
With the advent of extensive variability surveys, such as EROS, MACHO, and OGLE6 – to name only the first gravitational microlensing ones – the naming of variable stars has started to become once again quite chaotic, and the previously described naming scheme is often not followed by the teams in charge of these projects. This is the reason why recently discovered variable stars are nowadays often classified into such complicated-looking, “telephone number-like” classes as “EC 14026” (after star EC 14026-2647, from the Edinburgh-Cape Survey), or “PG 1605+072” (after the star in the Palomar-Green Survey). As pointed out by Kurtz (2002b), the MACHO project alone has discovered tens of thousands of variable stars, “and they have chosen to use their own naming convention, and even changed conventions part way through the project giving two naming schemes.” The ongoing Vista Variables in the Vía Láctea (VVV) project (Minniti et al., 2010; Catelan et al., 2011 2013) alone is expected to discover of order one million variable stars in the Galactic bulge direction. With Gaia and the LSST, variable stars will be discovered by the hundreds of millions, and perhaps billions. Therefore, Kurtz's question – How will we name them? – is very pertinent. After all, as he points out, it may not be very practical “to work on a star called ES 220418.23+190754.62, and refer to it always like that;” his advice is accordingly that astronomers “continue with tradition,” and thus number the stars in each constellation in its order of discovery – even if these numbers will unavoidably become very large at some future point or another. Indeed, the present authors tend to agree with this view: while it may not be particularly pleasant to work with a star called (say) V12345678 Sgr, it may be even worse to have to deal with one called ES 220418.23+190754.62 on a daily basis. An important drawback, on the other hand, is to establish the actual order of discovery, with an unavoidable delay between discovery and the establishment of an official “V-number” designation.
By the early twentieth century, numerous examples of stars that we should today class as RR Lyrae variables, Mira variables, and Cepheids had been discovered. These are all types of variable stars now long recognized as pulsating stars, but at that time, the cause of their variability was still unknown. As early as 1667, Bullialdus had suggested that the brightness changes of Mira might be caused by rotation, with one side of the star being more luminous than the other. In 1783 Goodricke added the idea that the periodic dimming of Algol might be caused by an eclipse. The realization that stars might vary because of pulsation was longer in coming. In fact, even the eclipse model for Algol remained in doubt until Vogel's (1890) spectroscopic observations confirmed it (Clerke, 1903).
By the start of the twentieth century, variable stars of the δ Cephei type posed a particular problem for astronomers trying to understand the nature of their light changes. Spectroscopic binary stars had been discovered not long before: as we have already noted, Vogel (1890) used the Doppler shift in its spectroscopic lines to determine the radial velocity of the brightest star in Algol throughout the light cycle. The changes in radial velocity confirmed the eclipsing binary explanation of Algol's brightness changes. A few years later, Bélopolski (1895) observed the spectrum of δ Cephei, finding that its radial velocity also changed during its light cycle. Because periodic radial velocity variations were, at that time, known only for binary stars, where the changes in velocity indicated orbital motion, it was believed that Cepheids, too, were binary stars. The light curves of Cepheids, however, did not at all resemble the light curves of eclipsing binaries such as Algol (Figure 1.2). They often had asymmetric light curves, with a steep rise to maximum light followed by a slower decline (Figure 1.3). A further clue came when Schwarzschild (1900) found that the Cepheid η Aquilae changed in color as well as brightness during its light cycle.
Figure 1.2 Joel Stebbins's 1910 photoelectric light curve of the eclipsing variable Algol, showing the near symmetry of the primary and secondary eclipses. (From Stebbins (1910).)
Figure 1.3 Stebbin's 1908 light curve of the variable δ Cephei (with brighter magnitudes toward the top). Unlike that of Algol, the light curve of δ Cephei shows a steeper rise to maximum light followed by a more gradual decline in brightness. (From Stebbins (1908).)
Despite the very different light curves of Algol and Cepheid stars, the radial velocity argument in favor of binarity seemed persuasive to many (Brunt, 1913; Furness, 1915), as noted in Gautschy's (2003) historical review. It followed that the cyclic brightness changes of Cepheids must in some way be caused by the existence of a binary star system, even if the exact mechanism remained obscure. Perhaps the companion star somehow caused an eruption on the primary star that became stronger as the two stars approached. However, the absence of a satisfactory binary star mechanism to explain the variations of the Cepheids caused some to contemplate alternative explanations. Plummer (1913), noting the difficulty in explaining Cepheid variability under the binary hypothesis, raised the alternative of radial pulsation, but he did not explore the possibility in depth. It remained for Harlow Shapley (1914) to put forward a strong argument that pulsation underlay the variations of Cepheid-type stars.
Shapley (1914) marshaled several arguments against the binary hypothesis. His strongest argument employed the recent discovery of the existence of giant and dwarf stars. Shapley argued that the Cepheids were giant stars, in fact so large that their radii exceeded the calculated sizes of the orbits in the supposed binary systems: “Interpreted as spectroscopic binaries these giant stars move in orbits whose apparent radii average less than one tenth the radii of the stars themselves.” Could such binary systems actually exist, if the Cepheids were so large as to engulf the supposed companion stars? It seemed unlikely. That was an argument from which the binary hypothesis never recovered – although attempts at resurrecting it did pop up occasionally (Jeans, 1925), and in fact, somewhat surprisingly, as late as 1943 a theory of Cepheid variability was advanced in which members of a binary system moved within a common atmosphere (Hoyle & Lyttleton, 1943).
Shapley advocated radial pulsation as an alternative to the binary hypothesis, although he offered no detailed model as to how stellar pulsation might be established and maintained in Cepheid variables. At the time of Shapley's publication, theorists such as Ritter (1879) and Emden (1907) had considered some aspects of the physics of pulsating stars, but the processes that might drive stellar pulsation were unknown. Within a few years, Eddington (1918a 1919a) would significantly advance the discussion of the physical processes that would be needed to keep a star pulsating. Nonetheless, three more decades would elapse before Zhevakin (1953) and Cox & Whitney (1958) began to elucidate the specific driving mechanism behind the pulsations of Cepheid variable stars (see Chapter 5).
1
The name comes from “al-Ghul,” arabic for “the ghoul.” It is often stated that this name was chosen to reflect the awe, and perhaps fear, that its variability would imply (to be contrasted with the name “Mira” that was given for
o
Ceti). However, and as discussed by Davis (1957), the name al-Ghul does not necessarily imply knowledge of the variability of Algol. Before the Arabs, β Persei was associated with the head of the Gorgon Medusa (as in Ptolemy). Davis says that the Ghul to the Arabs was a female demon and sorceress, and may have been adopted as a sort of monster similar to Medusa, rather than an indication of variability. On the other hand, Jetsu
et al
. (2013) have recently raised the intriguing possibility that the star's variability may have been known in ancient times to the Egyptians.
2
http://www.lindahall.org/services/digital/ebooks/bayer/index.shtml
3
Of course, when the first variable stars were discovered, Argelander's naming scheme had not yet been created. Therefore, other names were initially given to those stars, based on the naming conventions of other catalogs (unless the star was already present in Bayer's catalog, in which case – as we have seen – its Bayer letter was usually retained). Other such catalogs included, in particular, John Flamsteed's, which was edited and published (unauthorized) by Edmund Halley in 1712. Interestingly, when Flamsteed eventually published his catalog himself, he did not include the famous “ Flamsteed numbers” for designating stars (Bakich, 1995). Hence, for instance, R Leo was initially called 420 Mayer Leonis or 68 (Bode) Leonis (Zsoldos, 1994, and references therein).
4
Recall that in many languages “I” and “J” have the same pronunciation – it is no coincidence that the acronym “INRI” in the Christian cross stands for “Jesus of Nazareth, King [Rex, in Latin] of the Jews.”
5
As a matter of fact, both Nijland (1914) and Townley (1915) envisioned that this system should be used to rename the 334 stars from AA to ZZ in every constellation, as well as those with Bayer letters – thus, for instance, Algol/β Persei would be renamed V1 Persei, SS Cygni = V20 Cygni, SY Andromedae = V25 Andromedae, and so on. This went a bit beyond what traditionalist astronomers would have preferred, and so these stars have kept their original designations.
6
For a list of the abbreviations and acronyms commonly employed in this book and/or which are widely used in the relevant literature, the reader is referred to the Glossary at the end of this volume.
At this point, it is worth defining a few of the common terms and usages of variable star astronomy. It is assumed that the reader is already familiar with the basics of astronomy, but might not be familiar with variable star astronomy in particular.
The word “variability,” as it appears in the term stellar variability, is in fact shorthand for “time variability.” Therefore, we cannot overemphasize the fact that proper timekeeping is of crucial importance in stellar variability studies. As pointed out by Aerts, Christensen-Dalsgaard, & Kurtz (2010a), in stellar variability studies, and in asteroseismology in particular, “observations must be made with times known to high precision, and for long data sets to high accuracy.”
Historical astronomical observations, often dating back to hundreds of years BC, can still provide precious information for present-day astronomers; a very interesting recent example is provided by Stephenson & Morrison's (2005) analysis of historical eclipse reports, dating as far back in time as 700 BC. For this reason, it is of strong interest to establish a dating scale extending as far back in time as possible. Just such a scale is provided by the so-called Julian Proleptic Calendar.
The Julian Calendar was officially implemented in 46 BC by the Roman ruler Gaius Julius Caesar. While it has been replaced for civil use by the Gregorian Calendar, implemented by Pope Gregory XIII in 1582, it is still of great importance in astronomy, especially after the work of the historian and philologist Joseph Justus Scaliger (1629) – a man who was once called the most learned man in Europe (Grafton, 1993, p. 744).
Realizing the need to have a sufficiently long baseline so that all historical events are encompassed by a single calendar without the a priori need for “negative dates,” Scaliger proposed the notion of Julian Day Numbers, by setting a “zero point” for the Julian Calendar dating as far back as noon (GMT) of 1 January 4713 BC (or 01/01/−4712).1 The Julian Date is the interval of time elapsed since that point in time, in units of days; it is a real number, and is decimalized accordingly. The Julian Day Number, in turn, is the integer part of the Julian Date. Analytical expressions that allow the computation of Julian Dates for given dates in Julian and Gregorian calendars are provided by, for instance, Seidelmann & Urban (2010).2
As pointed out by Scaliger, the Julian Calendar possesses a “characteristic period” or cycle of 7980 years, each of which lasts exactly 365.25 days. The 7980 year Julian cycle comes about as follows. First, the calendar contains a “solar (S) cycle” with period 28 years, corresponding to the time required for the days of the week to repeat themselves exactly in the calendar – in other words, one finds that every 28 years 1 January will fall on a Monday, and thus only 28 consecutive calendars are needed. Second, there is also a “lunar (L) cycle” (also known as Metonic3cycle, or cycle of Golden Numbers) of duration 19 years, which corresponds to the time needed for the phase of the Moon to be the same on a given date; in other words, there is a 19-year gap between a year that begins with full Moon and the next. Therefore, years will pass between two New Year's days that fall on the same day of the week and with the same phase of the Moon.
Since 532 years is a relatively short period of time (in historical terms), Scaliger took into account a third cycle, namely the Roman “cycle of indiction” (I), corresponding to the 15-year interval between consecutive tax censuses, originally established by the Roman emperor Diocletian (Gaius Aurelius Valerius Diocletianus) in 285 AD. It is important to note that the Roman cycle of indiction was still widely used in Europe in the Middle Ages, and in fact remained in use by the Holy Roman Empire until 1806, when the Empire was dissolved during the Napoleonic Wars – and so it is natural that Scaliger would have invoked it in his work. Thus the Julian Calendar period of years.
The year 4713 BC as starting point of Scaliger's cycle can be understood as follows. According to the sixth-century monk Dionysius Exiguus, the year of Jesus Christ's birth was characterized by year 9 (out of 28) of the Solar cycle (S = 9), year 1 (out of 19) in the lunar cycle (L = 1), and year 3 (out of 15) of the indiction cycle (I = 3). (The exact reasons why Exiguus assigned this particular combination to the event remain obscure.) He then verified that this particular combination (S, L, I) = (9, 1, 3) took place 4712 years after (S, L, I) = (1, 1, 1); hence if we assign “year 0” to Christ's birth, the combination (S, L, I) = (1, 1, 1) corresponds to -4712, or 4713 BC.
The main drawback affecting Julian Dates is the fact that at present they are very large numbers. As an example, at the time of writing of this paragraph, 19:17:2.40 UT (or 15:17:2.40 Chilean Standard Time) on 7 April 2010, the Julian Date is 2,455,294.3035. Moreover, as we have seen, the Julian Day starts at noon rather than at midnight. For these reasons, one also defines a Modified Julian Date (MJD), which is the same as Julian Date, except for a different “zero point” – namely, midnight (GMT), 17 November 1858. Since the latter corresponds to Julian Date 2,400,000.5, it follows that the MJD can be easily computed from the Julian Date by subtracting 2,400,000.5 from the latter.
Another important definition is that of Heliocentric Julian Date. Since the Earth revolves around the Sun, and since light travels at a finite speed, observations of a given object taken at different positions in the Earth's orbit are not equivalent, and so a correction for the Earth's orbit around the Sun is required. If not properly accounted for, this can lead to a spurious signal in a periodogram, at a period of around 1 year (and corresponding aliases). An equation providing such a correction was derived by Binnendijk (1960), and reads (as quoted by Landolt & Blondeau, 1972)
where R is the radius vector of the Earth, is the celestial longitude of the Sun, and are the star's right ascension and declination, respectively, and is the obliquity of the ecliptic (IAU recommended value: , for epoch 2000).
In fact, for ultra-high-precision asteroseismology, even this heliocentric correction may not be sufficient, and a correction for the barycenter of the solar system – that is, accounting for the influence of Jupiter and the other planets – may instead be required, thus giving a Barycentric Julian Date (BJD). Indeed, and as pointed out by Aerts et al. (2010a), an improper BJD correction was one of the reasons why the first claimed detection of an extrasolar planet around a neutron star (Bailes, Lyne, & Shemar, 1991), with an orbital period of around 6 months, had soon to be retracted (Lyne & Bailes, 1992): the detected 6-month signal was actually due to the Earth's orbit around the Solar System's center of mass.
As a matter of fact, when one obtains images at a given telescope, the time recorded on the resulting fits files is not the Julian Date, but typically universal time (UT). There are different UT definitions. UT0, UT1, and UT2 essentially provide measures of the rotation period of the Earth, with UT1 constituting a correction over UT0 that takes the Earth's polar motion into account, whereas UT2 provides a correction over UT1 that takes the annual seasonal variation in the Earth's rotation speed into account. Finally, coordinated UT, or UTC, is an atomic system in the sense that it differs from International Atomic Time (TAI) – which is defined as a function of atomic transitions between hyperfine levels of the ground state of Cesium 133, as measured by different laboratories throughout the world – by an integer number of seconds. UTC is kept to within 0.9 s of UT1 by adding leap seconds as may be necessary. Such leap seconds must obviously be carefully taken into account when performing asteroseismology.
Given the fact that the Earth's rotation is slowing down, and doing so in a fairly complex way, timescales based on the Earth's rotation are not uniform, which again can be extremely problematic for asteroseismology if not properly accounted for. Thus surfaced the concept of ephemeris time (ET). As a matter of fact, ET and TAI differ by a constant, which is equal to 32.184 s (Equation 6.4 in McCarthy & Seidelmann, 2009). Other definitions exist, but it is not our purpose to delve into the many intricacies and subtleties of different timescale definitions. For practical purposes, it is always possible to convert from one time system to another, using suitable analytical formulae and computer programs that are freely available in the literature. The interested reader will find an authoritative account in the recent monograph by McCarthy & Seidelmann (2009).4
Before closing, we would like to point out that while light's finite speed may make high-precision measurements of time-dependent astronomical phenomena difficult (see the preceding discussion of the Lyne & Bailes 1992 results), it also presents astronomers with a unique window into self-gravitating stellar systems, perhaps most notably planetary systems. This is because light's finite travel time leads to its taking longer to reach us when an emitting star is farther away from us in its orbit than when it is closer – and this can sometimes be detected in high-precision experiments. As a matter of fact, this light-travel time effect provided the means for the first positive detection of an extrasolar planet (orbiting a neutron star), as the detected anomalies in the pulsar's period could only be accounted for by the presence of a planetary-mass companion (Wolszczan & Frail, 1992). Likewise, the same method has been used to discover giant planets around the pulsating blue subdwarf star V391 Peg (Silvotti et al., 2007), the pulsating white dwarf GD 66 (Mullally et al., 2009), the eclipsing polar DP Leo (Qian et al., 2010a), and the hibernating cataclysmic variable QS Vir (Qian et al., 2010b).
A light curve, per se, is simply a plot of the apparent magnitude of a variable star versus time. As noted, it is often convenient to use Julian Date for the time coordinate. If a variable star has a period, and that period is known, it becomes useful to plot a phased light curve. The phase of an observation can be calculated from the relation (Hoffmeister, Richter, & Wenzel, 1985)
where is the phase, t is the time of an observation, is some adopted reference time, is the period, and is the integer part of , sometimes termed the epoch. is computed as follows:
where and denote the “floor” and “ceiling” of x, respectively.
Thus, only the portion of after the decimal point is retained. The phased light curve essentially folds all of the observations of a variable star to show how the variable changes brightness during a single cycle. Often, is chosen to be a time of maximum light when one is dealing with a Cepheid or RR Lyrae variable. For an eclipsing variable star is more likely to be a time of minimum light. An excellent introduction to working with light curves is Grant Foster's (2010) book Analyzing Light Curves. The phased light curve of AT And, based on the observations plotted in Figure 2.1, is shown in Figure 2.2.
Figure 2.1 The light curve of the variable star AT And, based on observations obtained during the Northern Sky Variability Survey (Woźniak et al., 2004). Magnitude is plotted versus Modified Heliocentric Julian Date minus 50 000.
Figure 2.2 The light curve of AT And shown in Figure 2.1 has been phased with a period of 0.6169 days.
In this book, we refer to observations of variable stars in a variety of photometric systems. It is perhaps useful to include a table of the properties of a few of the more commonly encountered photometric bandpasses (Table 2.1). Because there are slightly different versions of the J, H, and (especially) K near-infrared bandpasses in use, we list those used by the 2MASS project (Skrutskie et al., 2006). Complications in defining and using these and other photometric systems are described in Bessell (2005).
Table 2.1 Photometric bandpassesa
Name
Effective wavelength (nm)
Bandwidth (nm)
Johnson
U
366
65
Johnson
B
436
89
Johnson
V
545
84
Cousins
R
641
158
Cousins
I
798
154
Strömgren
u
352
31
Strömgren
v
410
17
Strömgren
b
469
19
Strömgren
y
548
23
Strömgren w
489
15
Strömgren n
486
3
Sloan
u
350
64
Sloan
g
463
131
Sloan
r
614
116
Sloan
i
747
127
Sloan
z
893
118
J
(2MASS)
1236
162
H
(2MASS)
1646
251
K
(2MASS)
2160
262
a Effective wavelengths and bandwiths from the following sources: Strömgren system, Bessell (2005); Sloan system, Doi et al. (2010); 2MASS system, Cohen, Wheaton, & Megeath (2003) and Blanton & Roweis (2007).
Time-series analysis for pulsating stars is one of the chief aspects in the analysis of photometric and spectroscopic observations. The search for periods, or in some cases the search for multiple periods, in the light curve or the radial velocity curve of a variable star is one of the most important steps in the analysis of observational data, and a variety of period determination methods specifically aimed at astronomical time-series data, which are often unequally spaced in time, have been developed for this purpose. Those interested in the period determination techniques may wish to consider the discussions in Lafler & Kinman (1965), Lomb (1976), Ferraz-Mello (1981), Dworetsky (1983), Horne & Baliunas (1986), Foster (1995, 1996, 2010), Percy, Ralli, & Sen (1993), Reimann (1994), Scargle (1982), Schwarzenberg-Czerny (1989), Stellingwerf (1978b, 2011), Templeton (2004), and Graham et al. (2013). Those papers cover some of the many different approaches to the mechanics of determining periods of photometric or spectroscopic observations. In many methods, a number indicating some measure of goodness of fit for the light curve is calculated for many trial periods, and the probable period is sought within maxima or minima of the calculated quantity, depending on the method. Some methods rely on the calculation of peaks in the Fourier spectrum of a dataset. Others minimize the scatter when the observations are binned according to the phases calculated using the trial periods. Still others minimize the length of a “string” that connects adjacent observations when they have been phased according to the trial periods. The derived curve showing the selected goodness-of-fit indicator versus trial period is frequently called a periodogram or power spectrum, especially when Fourier methods are used. When a period is selected as the most likely one, the periodic signal (plus its harmonics) can be subtracted out from the original signal, and the power spectrum recomputed – a technique called prewhitening, which is widely used in the case of multi-periodic stars.
The approaches that can be adopted are many, and the details of period determination methods, well covered in these other papers, are beyond the scope of this book.5 Nonetheless, it is worth reminding the reader of a few of the difficulties that can arise in determining periods from observational datasets.
First, there are limits to the periods that can be found using any particular set of observations. If the observations were obtained over a 1-week interval, they will be of little use in identifying a 10-year period. In general, the time interval spanned by the observations will need to be about two or three times the length of the longest period that can be reliably determined. There is a limit, too, to the accuracy to which a period can be determined. If n data points are sampled with a spacing in time between observations, then the frequency of the variable star can be determined with a resolution , which is the inverse of the time span covered by the observations. Transformed into period, , this becomes
The Nyquist frequency (Nyquist, 1928) is often used to represent the highest frequency that can be detected with a given set of observations. If the magnitude of a variable star is sampled at a rate of n observations per day, then the corresponding Nyquist frequency would be cycles per day. It is sometimes possible to do better than this, identifying frequencies that are higher (periods that are shorter) than the Nyquist frequency. In that case, however, one runs the risk that the period determined might be an alias