Hubble Source Catalog (HSC) Use Case #2- Time-variable Phenomena

Searching for variable stars in dwarf irregular galaxy IC 1613

Reference:- The ACS LCID project.
                      II. Faint variable stars in the isolated dwarf irregular galaxy IC 1613.
                      Bernard et al. (2010) ApJ, 712, 1259

Figure 1 (from Bernard et al. 2010 reproduced by permission of the AAS). Shown is the Color-magnitude diagrams of IC 1613 for the ACS (left) and WFPC2 (right) fields. Confirmed and candidate variables stars were plotted, as labeled in the inset panel. The different size error bars in each panel shows the mean photometric error bars at given magnitudes. To show the location of the old main-sequence turnoff, an isochrone from the BaSTI library (Z = 0.0003, 13 Gyr; Pietrinferni et al. 2004) is overlaid in the right panel.

• STEP 3- Download the Bernard et al. (2010) data set from CDS.
  
  Note that Bernard et al. (2010) used DAOPHOT/ALLFRAME (Stetson 1994) to obtain the instrumental photometry of the stars on the individual, non-drizzled images, so they have better sampling over the light curve. In total they used 24 detections for each filter. The HSC have a binned phase covering of 10 points in F475W and 11 in F814W, since the HSC use the visit-based combined drizzled images. 

• STEP 4- Use the downloaded HSC Summary table:
  

• Calculate the "composite V magnitude" Vmag as ~ 0.5*(A_F475W+A_F814W).
  
• Calculate the Vmag "Variability Index (VI)" [i.e.  VI= sqrt(A_F475Wsigma**2 +A_F814Wsigma**2)]
• Plot the V mag [AB] vs the VI [AB] mag. A large fraction of the stars have a VI  <0.025 mag, which will be needed further for selecting non-variable stars. 
• Plot the F475W-F814W in [AB] vs the VI [AB mag].
  

        
Of particular interest are the objects that lies between (A_F475W-A_F814W) ~ -0.9 and 0.7 and have VI > 0.1. These are our 210 variable candidates.

• To select variable and non-variable stars we required that a star be detected in all 21 frames to be checked for variability. We also excluded all stars in crowded region by rejecting candidates with a companion contributing more than 50% of total light within a two pixel radius.
• A large fraction of the stars have a VI [AB] mag <0.025, to select non-variable stars we required that VI [AB] mag <0.01.
  
• Transform the magnitudes from the HSC in the AB mag system to the VEGA mag system using the Bernard et al. 2009 transformation equation. This allow us to make a direct comparison with Bernard et al. (2010) .

• STEP 5- Plot the Color Magnitude Diagram.

Shown above is the Color-magnitude diagrams of IC 1613 for the ACS data, where for the ordinate axis (F475W + F814W)/2 ~V filter combination is used, so that the horizontal branch (HB) appears approximately horizontal. HSC candidate variables are plotted in blue. Bernard et al. (2010) confirmed and candidate variables stars are labeled in the inset panel. The vertical gray dashed lines roughly delimit the instability strip. The magnitudes have been transformed to the VEGAMAG system. Variable star candidates were selected near the instability strip with Variability Index (VI) >=0.1. In total we selected 210 candidate variable stars from the HSC, compared to 259 in the Bernard et al. (2010) paper.

So what is the instability strip?
The instability strip is a narrow, almost vertical region in the HR diagram which contains many different types of variable stars (RR Lyrae, Cepheid variable, W Virginis and ZZ Ceti stars all reside in the instability strip). Most stars more massive than the Sun enter the instability strip and become variable at least once after they have left the main sequence. It is within this region that they suffer instabilities that cause them to pulsate in size and vary in luminosity.

• STEP 6- Download the HSC data using the HSC Detailed Search Form.
  

We now need to get the time-resolved data using the HSC Detailed Search Form and the "File Upload Feature". Upload the local saved comma separated values for the candidate variable stars file as follows:-

    Local File Name= File for the 210 candidates identified in Step 4.
    Column Delimiter= ,
    RA/Target/Data ID Column Number= 2
    Dec Column Number= 3
    Resolver= Don't Resolve
    Radius arcmin= 0.018
   
    Output Columns= add Filter & Wavelength
    Output Format= File: < your choice from the 8 options >
    Maximum Records per Target= 50

The form will look like this.

Save the resulting file.

• Make Aperture Correction (as in step 2 above).
• If you plan to compare your result to others reported in the literature you may need to convert to the BVI system (e.g. using the Bernard et al. 2009 transformation equation).  Note that the Bernard et al. (2009) equations are limited to the variable stars confined in the instability strip, as their intermediate color and narrow temperature range minimized the difficulties related to extreme temperatures.
• Convert  the observed start and end time to Julian Date (JD) or Modified JD (MJD), via e.g. (WCSTools, python PyAstronomy, pyephem, idl, topcat, etc.).
  
• Estimate the midMJD  (or midJD) 
  
• midMJD=0.5*(startMJD+endMJD)

• STEP 7- Visual inspection (optional)
  

1. Produce finding charts for selected variables in IC 1613. We used the Hubble Legacy Archive website and search around the position (RA & Dec) of each variable star with a search Radius: 0.01 degrees, then select from the Inventory "level  4" to get the color image cutout with window size:10 arcsec (i.e. 200x200 ACS WFC pixels). The finding chart are shown below.
2. The phased light curves for the candidate variable stars are also included shown below (produced using topcat). To get the obs days, the midMJD was subtracted from 53965.

17759137	17753692	17764426
17759737	17759032	17756492

Non-Variable Candidates over the observed ~2.4 days.

17753613	17764772	17759295

• Step 8-  Determining the period for MatchID=17753692 (Also know as V043, Bernard et al. 2010 with a period of 1.3 days)