Variable Stars Observed With the Ast3-1 Telescope From Dome a of Antarctica


Dome A in the Antarctic plateau is likely one of the best astronomical observation sites on Earth. The first one of three Antarctic Survey Telescope (AST3-1), a 50/68 cm Schmidt-like equatorial-mount telescope, is the first trackable telescope of China operating in Antarctica and the biggest telescope located in Antarctic inland. AST3-1 obtained a large amount of data in 2012 and we processed the time-series parts. Here we present light curves of 29 variable stars identified from ten-day observations with AST3-1 in 2012, including 20 newly-discovered variable stars. 19 variables are eclipsing binaries and the others are pulsating stars. We present the properties of the 29 variable stars, including the classifications, periods and magnitude ranges in i band. For 18 eclipsing binaries, the phased light curves are provided with the orbital period values well determined.

techniques: photometric survey — stars: variables

Vol.0 (200x) No.0, 000–000

1 Introduction

Observations of variable stars play a crucial role in the understanding of stellar evolutions, as well as stellar formations and properties. Study of variable stars can make an important contribution in astrophysical researches such as the age of Universe, the distance of galaxies, the composition of the interstellar medium and the behavior of the expanding Universe. However, continuous observation is impossible in mid-latitude observatories because of the alternation of day and night. Luckily, there is a continent where night can last for almost six month: Antarctica.

Antarctica is a continent with the highest average altitude, the driest climate and the lowest average temperature. After tests lasting almost two decades by astronomers from different countries, Antarctica continent is shown to be the best place for all kinds of astronomical observations ([Storey 2005]). Apart from its stable atmosphere, Antarctica continent possess excellent conditions as follows,

  • Continuous observations: the latitude close to the South Pole is high enough for celestial objects to be observed almost continuously during the Antarctic nights ([Strassmeier et al. 2008]).

  • Clear skies: at Dome C in Antarctica, there are 75% to 90% days without clouds ([Rauer et al. 2008]).

  • Clean and dry atmosphere: the atmosphere in Antarctica is the cleanest on earth. In South Pole the amount of aerosol is one fifty of that in Hawaii ([Bodhaine 1995]).

Furthermore, the pollutions like light pollution don’t exist in Antarctica.

In order to take advantage of these remarkable observation conditions at Dome A, which is the highest place in the inland of Antarctica continent and considered as the best astronomical site in both optical/IR and sub-mm, the Chinese Small Telescope Array (CSTAR) was installed at Dome A in 2008 January. CSTAR undertook both site testing and science research tasks ([Wang et al. 2014]). As a following plan, the trio Antarctica Survey Telescopes (AST3) were designed for wide-field survey in multi-band from Dome A, mainly to search for supernovae and extrasolar planets.

2 Facilities and Observations

2.1 Facilities

The first of three AST3 telescope (AST3-1) is a 50/68 cm Schmidt-like equatorial-mount telescope, with the focal ratio of F/3.73, equipped with complete pointing and tracking components. AST3-1 was mounted with a STA1600FT CCD camera, with 10560 10560 pixels of 9 m, corresponding to 1 arcsec per pixel on the focal plane. Only half area was used to make exposure. The filter is SDSS i ([Wei et al. 2014]). The CCD chip is cooled with TEC to take advantage of the extremely low air temperature at Dome A in the winter, which is about -60 C on average ([Ma et al. 2012]).

The Frame Transfer mode of CCD is adopted to terminate the exposure rather than the mechanical shutter. The CCD is divided into two parts, the frame store regions (the top and bottom quarters with 10560 2640 pixels) and the image area (the central half 10560 5280 pixels), corresponding to a 2.93 deg 1.47 deg on the sky, which is the effective field of view of the AST3-1 ([Ma et al. 2012]).

Figure 1 shows a raw image of AST3-1taken in 2012 with 530012000 pixel, where one may see the image of stars and some bad pixels. There are sixteen readout amplifiers for the CCD chips to reduce the readout time. The area of one panel is 26401320 pixel. Among the panels there are 8 vertical overscan areas whose size is 5300180 pixel each and 1 horizontal overscan area whose size is 2012000 .

2.2 Observations

AST3-1 was placed at Dome A by the 28 Chinese Antarctica Research Expedition (CHINARE) team in January of 2012 and carried out observations until May 8, 2012. In 2012, AST3-1 has in total observed 500 sky areas, corresponding to 2000 square degrees, and obtained more than 22000 images, including more than 3000 frames with exposure time of 60 seconds, and more than 4700 frames for LMC as well as 7000 images for the Galactic disk. The data collected with AST3-1 in 2012 were taken back to China by the 29 CHINARE team ([Ma et al. 2014]).

AST3-1 obtained a large amount of data in 2012 observation but the time-series data except LMC area is not much enough. This data was selected to process in order to find variable stars and it was list in table 1.

3 Data reduction

3.1 Pre-reduction

The overscan was used as bias. After subtracting and cutting down the overscan, the image size was changed from 530012000 to 528010560. We ignored the effect of the dark current due to its low values. On the other hand, Ma (2014) gave a new method to deduce the dark currents of the AST3-1 data .

As flat-field correction is a very necessary and critical step in data reduction, We selected 3763 images whose average ADU per pixel is larger than 5000 hence regarded as twilight images and combined them as a Super Flat, where the pixel values larger than twice of the background ones are replaced with the mean ADU value of the frame. Although the sky brightness gradient is minimum and stable, there is still a gradient of 1% across the AST3s 4.3 square degrees field-of-view ([Wei et al. 2014]). By combining numerous twilight images the gradient was averaged. After divided by the superflat, the images are ready for photometry. Figures 2 and 3 show the superflat and an image after the correction of bias and flat-field, respectively.

The data taken in 8 of April of 2012 show a good tracking capacity of AST3-1, as the images of stars on the frames kept stable during several hours observations. However, the frames collected after 12 of April show problems of tracking. Picture shift and rotation as well as star trails become common. As a result we programmed an IDL script to align stars in different frames and removed bad data. Some gaps in the light curves which will be mentioned later are the consequence of the bad data.

3.2 Photometry

For lists the datasets of AST3-1 collected in 2012, we used the command DAOFIND in IRAF to derive the coordinates of stars where the ADU values of the central pixel are larger than four times of the sky background ones to make sure a good photometric accuracy. The ADU values should be smaller than 30000, which is the limit of linearity of the CCD ([Ma et al. 2012]). For images whose exposure times are 60s each collected in 25th April, the number of stars satisfying this requirement is 12579. The result of flux calibration shows that the range of magnitude in i band is approximately from to . Aperture photometry is then used to derive light curves in order to detect the candidates of variables.

4 Light Curves

After light curves obtained, some corrections were made before further analysis. From the data we noted that the clock of AST3-1 has a serious error which leads to the recorded time being the same for a group of images taken by continuous collection. In order to correct the time error, we assumed that for one group of images the time separations are the same hence the times of the images can be deduced. This assumption is proved to be reasonable by the phase diagrams of all the eclipsing binaries.

Figure 4 shows the magnitude calibration result. With the reference stars the linear fit leads to , in which the related coefficient R² = 0.994, while means the differential magnitudes of eight check stars in i and means magnitudes in i from UCAC4 catalog ([Zacharias et al. 2013]).

Figure 5 shows the light curves of the star V23 observed in 30 of April of 2012, in which one can see that the light curves in even one night are discontinuous, due to bad data caused by either tracking components problems or bad pixels.

Figure 6 shows the diagram of the photometry accuracy versus the magnitude in i band when the aperture radius was taken as 4 pixels. From Figure 6, one may see that the photometry accuracy for most stars with magnitude in i band between and is between and . If the sigma for a certain star is larger than three times of the normal accuracy in its magnitude in Figure 6, in the meanwhile no CCD problems were found by eyes, the star is regarded to be a variable.

5 Results and Discussion

5.1 Variable stars

Table 2 and Table 3 summarize properties of the 29 variable stars we detected. 20 ones are newly discovered in Table 2 after searching in the AAVSO website and the other ones have already been known which are listed in Table 3. We present the information including name, coordinate in International Celestial Reference System (ICRS) from UCAC4 catalog ([Zacharias et al. 2013]), classification, period or time scale and magnitude range in i. Magnitude ranges are derived by using the polyfit to the extreme values. Three of the known variable stars were discovered by the Optical Gravitational Lensing Ex-periment (OGLE) ([Pietrukowicz et al. 2013]). Two were found by The All Sky Automated Survey(ASAS) ([Pojmanski 2002]).

In order to deduce the orbital periods of eclipsing binaries from the light curves, we analyzed the shapes of light curves to estimate the rough values of periods, then apply a ten-order polyfit to the light curves to optimize the periods hence obtain the phased light curves with the least-square fit. Figure 7 shows the phased light curves of the 18 eclipsing binaries among the 29 variable stars listed in Table 2 and Table 3. The orbital period values are listed in the 5 column of Table 2 and  3. Figure 8 shows the light curves of the other 11 variable stars, which can not be combined to phase ones.

As light curves of 11 variable stars are not easy to be classified, we made a rough classification depending on their colour index and time scales, as well as the shapes of the light curves. We obtained their magnitude information from UCAC4 catalog, as listed in table 4. Because of the variability of the luminosity, the effective temperatures determined with the colour index are not accurate. We take three stars, as examples.

  • : one can know from UCAC4 catalog that the magnitude of V01 is and , so the colour index is , which means that this star is F0 type with ([Bessell et al. 1998]). In the meanwhile, it is apparent that the increase is faster than decrease in light curve. Combining its time scale, it is possible for V01 to be an classical cepheid([Strohmeier 1972]).

  • : the magnitude is and , so the colour index is . The effective tempareture is hence ([Bessell et al. 1998]). This star is close to G0 type and likely to be a Cepheide II ([Strohmeier 1972]) with the time scale of 1 day and the amplitude of . Nevertheless, according to the shape of the light curves, we are not sure so it is classified as ’unknown’ in Table 2.

  • : the magnitude is and , so the colour index is . One can deduce the effective tempareture as ([Bessell et al. 1998]). A star between F0 and G0 with the time scale of 0.6 days and the amplitude of could be a Scuti star([Strohmeier 1972]). However, [Sahade & Berón Dàvila (1963)] thought it as an eclipsing binary and gave the period of 1.676868 days. This inconsistency may be caused by the incomplete light curves in phase.

5.2 Discussion

Figure 9 plots distributions of variable fraction versus the magnitude in i, period and amplitude from the data of AST3-1 in 2012. We present a discussion and the photometric accuracy which are compared with the result from CSTAR ([Wang et al. 2015]).

  • Distribution of magnitude: From Figure 9, one finds that the mean magnitude is , and a peak appears at the magnitude larger than , which is significantly deeper than CSTAR with mean magnitude of .

  • Distribution of period: A jump is seen at 1 day. Since the observation data are shorter than one week, tt is then difficult to detect long-period variable stars and most detected periods are shorter than 2 days. By contrast, CSTAR can find the period up to 88.39days. AST3-1 also has such an ability with long-time observations.

  • Distribution of amplitude: Almost 60% stars have small amplitude smaller than , which is the same as that in CSTAR’s result. The distribution proved that AST3-1 has the capacity to find stars with a slight varibility.

  • Photometric accuracy: As shown in Figure 6, the photometric accuracies remain at the level of for stars of to . As the contrast, the photometric accuracy of CSTAR rises from at to at . This is believedto the larger aperture of AST3-1 than that of CSTAR.

In short, AST3-1 shows significant improvements in both deeper limit magnitudes and higher photometric accuracy.

6 Conclusions

Data observed with AST3-1 in 2012 are analyzed and 29 variable stars a are detected mong which 20 ones are newly discovered. For the stars with magnitudes in i band in the range from to , the typical accuracy of photometry is . The magnitude calibration leads to a good result with the coefficient of 1.03. Properties of all the 29 variable stars are presented, including the classifications, periods and magnitude ranges in i band. For 18 eclipsing binaries, the phased light curves are presented with the orbital period values well determined. Compared with the results from CSTAR, AST3-1 showed significant improvements in both limit magnitude and photometric accuracy. The data observed by AST3-1 in 2012 proved that observations by AST3-1 in Dome A can provide a unprecedented oppportunaties to the detection and study of variable stars.

7 Acknowledgements

The authors thank the support from the Scientific Research and Entrepreneurship Plan of Beijing College Students. JNF acknowledges the support from the Joint Fund of Astronomy of National Natural Science Foundation of China (NSFC) and Chinese Academy of Sciences through the Grant U1231202, and the support from the National Basic Research Program of China (973 Program 2014CB845700 and 2013CB834900).

Figure 1: A raw image of AST3-1. Note the 16 readout areas and its overscan areas.
Figure 2: The superflat
Figure 3: An image after correction of bias and flat field
Figure 4: Magnitude calibration. The linear fit shows a good magnitude calibration result.
Figure 5: Light curves of V23 in i in 30 of April of 2012.
Figure 6: Photometric accuracy versus magnitude in i band when the aperture radius is 4 pixels.
Figure 7: Phased light curves of 17 eclipsing binaries among the 29 variable stars, listed in Tables 2 and  3
Figure 8: Light curves of 12 vriable stars whose phased diagrams can not be made with certain values of periods.
Figure 9: The distributions of variable fraction versus the magnitude, period and amplitude. As to the magnitude distribution, one can see that the deeper the magnitude is, the larger the number of variable stars is. As for the period, one can see that the stars whith period shorter than 2 days occupy nearly 90%. Note that the observation durations are shorter than one week for the data we processed. In terms of amplitudes, there is a dramatic drop in
Date of observation Name of field of view (ICRS) (ICRS) Number
, and Apr. HD 143414 and WR star 240.9560 -62.6887 673
and Apr. HIP75377 231.0471 -61.6771 618
to Apr. transit 163.9000 -61.5000 1647
Table 1: Datasets we have already processed

(ICRS) (ICRS) Type P/TS(d) m range

10:52:32.93 -61:45:54.9 Cep 3.0 13.63 - 13.76
V02 10:52:59.15 -60:32:54.1 EB 0.997 13.68/no data - no data/13.88
V03 10:53:33.00 -61:29:16.1 EA 3.426 13.32/no data - 14.84/no data
V04 10:53:55.55 -61:28:13.8 EA 1.827 13.52/13.52 - 13.85/no data
V05 10:55:29.21 -62:04:08.3 EB 0.312 13.48/13.50 - 13.74/13.70
V06 10:56:41.07 -62:26:21.5 EA 1.997 13.50/13.50 - 13.54/13.60
V07 10:57:00.54 -61:20:35.8 unknown 0.5 13.74 - 13.83
V08 10:57:10.97 -61:58:34.1 unknown 1.0 13.15 - 13.25
V09 10:57:54.18 -61:31:45.0 EB 1.997 13.63/13.64 - 13.70/13.68
V10 10:58:13.81 -60:49:28.5 unknown 0.2 13.72 - 13.92
V11 10:58:52.16 -61:57:31.9 EA 1.997 13.68/13.68 - 13.84/13.83
V12 15:21:57.46 -62:09:15.7 DSct 0.3 12.41 - 12.50
V13 15:21:57.65 -61:30:28.7 EB 0.356 12.48/12.49 - 12.55/12.54
V14 15:22:16.35 -60:55:30.4 DSct 0.5 9.06/9.07 - 9.18/9.10
(HD 136194)
V15 15:22:25.31 -61:21:06.7 EB 0.203 12.08/12.08 - 12.11/12.10
V16 15:22:55.53 -61:16:38.2 unknown 0.4 12.42 - 12.55
V17 16:02:33.40 -62:37:41.5 EB 0.850 10.80 - 0.88
(TYC 9040-2606-1)
V18 16:04:44.70 -62:37:22.4 DSct 0.3 11.63 - 11.68
(TYC 9040-1073-1)
V19 16:05:03.11 -62:05:33.7 EA 0.550 10.44/10.46 - 10.60/10.60
V20 16:05:18.75 -62:07:24.6 EA 1.099 9.75 - 9.93
(HD 143660)
Table 2: Variable stars which are discovered by us. We list the name, coordinate, type, period or timescale (P/TS) in days and magnitude range in i (m range). If the phase is completed for an eclipsing binary, we give magnitude range in this format; primary/secondary max - primary/secondary min, otherwise we just give the range in: max - min. For the type, we indicate Eclipsing Binary(EB,EA,EC or EW), classical cepheid ( Cep) and Scuti stars(DSct).

(ICRS) (ICRS) Type P or TS m range

10:52:05.33 -61:20:31.7 EA 1.981 12.20/12.21 - 13.93/12.85
(OGLE-GD-ECL-04022) EA 1.977267
V22 10:53:03.30 -61:38:21.8 EB 1.582 13.27/13.27 - 13.71/13.50
(OGLE-GD-ECL-04249) EC 1.57794
V23 10:53:45.48 -61:15:35.0 EC 0.417 13.56/13.60 - 14.00/13.98
(OGLE-GD-ECL-04451) EC 0.41731592
V24 10:55:23.07 -60:50:51.5 EB 1.205 14.02/no data - 14.36/14.17
(OGLE-GD-ECL-04845) EC 1.20580
V25 10:56:23.41 -61:14:45.4 DSct 0.6 13.32 - 13.45
(V* IN Car) EA 1.676868
V26 15:59:05.37 -62:40:36.2 EB 0.623 12.07/no data - 12.37/12.20
(ASAS J155905-6240.6) EC 0.62332
V27 15:59:06.40 -63:17:49.4 unknown 1.1 10.07 - 10.67
(V* V336 TrA ([N/A 2005])) EW 0.266768
V28 16:06:32.63 -62:38:55.5 EA 1.634 9.86/9.86 - 9.99/no data
(ASAS J160633-6238.9 ) EA 1.6343
V29 16:08:24.83 -61:40:08.3 EB 1.0 13.66 - 13.72
(V* AD TrA ([Hoffleit 1931])) EB 1.2604

Table 3: Variable stars which are discovered by us. We list the name, coordinate, type, period or timescale (P/TS) in days and magnitude range in i (m range). If the phase is completed for an eclipsing binary, we give magnitude range in this format; primary/secondary max - primary/secondary min, otherwise we just give the range in: max - min. For the type, we indicate Eclipsing Binary(EB,EA,EC or EW), classical cepheid ( Cep) and Scuti stars(DSct).
Name B(mag) V(mag) J(mag) K(mag)
V01 13.269 13.068 0.201 7030K
V07 12.096 11.296 0.800 3850K
V08 13.580 12.999 12.872 12.756 0.581 0.116 6000K
V10 13.021 12.194 11.491 0.703 4150K
V12 13.607 12.923 11.494 11.060 0.684 0.434 5750K
V14 10.215 9.743 8.649 8.189 0.472 0.460 6500K
V16 13.742 13.069 11.453 11.045 0.673 0.408 5660K
V18 12.397 11.897 10.864 10.612 0.500 0.252 6250K
V25 14.192 13.748 13.154 12.690 0.444 0.464 6500K
V27 11.511 10.696 8.897 8.301 0.815 0.596 5240K
V29 13.338 12.889 11.999 11.624 0.449 0.375 6400K
Table 4: Information of variable stars from the literature whose types can not be deduced clearly.


  1. Send offprint request to:


  1. 2005, Information Bulletin on Variable Stars, 5600, 1
  2. Ashley, M. C. B., Burton, M. G., Calisse, P. G., Phillips, A., & Storey, J. W. V. 2005, Highlights of Astronomy, 13, 932
  3. Bessell, M. S., Castelli, F., & Plez, B. 1998, \aap, 333, 231
  4. Bodhaine, B. A. 1995, \jgr, 100, 8967
  5. Hoffleit, D. 1931, Harvard College Observatory Bulletin, 884, 10
  6. Li, X., & Wang, D. 2013, IAU Symposium, 288, 329
  7. Ma, B., Shang, Z., Wang, L., et al. 2012, \procspie, 8446, 84466R
  8. Ma, B., Shang, Z., Hu, Y., et al. 2014, \procspie, 9154, 91541T
  9. Mosser, B., & Aristidi, E. 2007, \pasp, 119, 127
  10. Pietrukowicz, P., Mróz, P., Soszyński, I., et al. 2013, \actaa, 63, 115
  11. Pojmanski, G. 2002, \actaa, 52, 397
  12. Rauer, H., Fruth, T., & Erikson, A. 2008, EAS Publications Series, 33, 41
  13. Sahade, J., & Berón Dàvila, F. 1963, Annales d’Astrophysique, 26, 153
  14. Storey J. W. V., 2005, in: Antarctic Science, 17(4), 555-560.
  15. Strassmeier, K. G., Briguglio, R., Granzer, T., et al. 2008, \aap, 490, 287
  16. Strohmeier, W. 1972, Oxford, New York, Pergamon Press [1972] [1st ed.],
  17. Wei, P., Shang, Z., Ma, B., et al. 2014, \procspie, 9149, 91492H
  18. Wang, S., Zhang, H., Zhou, J.-L., et al. 2014, \apjs, 211, 26
  19. Wang, S., Zhang, H., Zhou, X., et al. 2015, \apjs, 218, 20
  20. Zacharias, N., Finch, C. T., Girard, T. M., et al. 2013, \aj, 145, 44
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description