Skip to main content

Photometry of rapidly varying stars

We recently released a significant update to the backend systems for the AAVSO website. While most of the bugs introduced by this update have been fixed, there may still be problems we haven't fixed. If you run into a problem, please email webmaster@aavso.org
petike1's picture
petike1
Offline
Joined: 2011-06-24

I try to do photometry of RU Peg, and I'm not sure I'm doing it right. I have 2x3 images in every session (3 R and 3 I). I have the following questions:

1) Since the magnitude is clearly varying during the session I think it's best to report them separately instead of stacking - the SNR is not too big (~100-200) but I think important information would be lost during stacking. Is that correct?

2) According to the Uncertainty class if I have more than one image, I should report the stddev. of check star of similar magnitude. If I report 3 values for each filter, is it correct if all magnitudes per filter has the same uncertainity, the stddev. of the check star in that filter?

3) As I read it's good practice to transform the magnitudes to a standard magnitude. Since the magnitude is varying during the session the values in (r-i) would be the inst. magnitudes for /different/ real magnitudes. What's the best practice to solve this problem? How to transform the magnitudes of rapidly varying stars?

4) In VPHOT, during Single Image Photometry at the "photometry report" page there is a "Transform" checkbox. If I understand transformations right, I need at least two filters for that. How does that Transform checkbox work with single images?

5) In VPHOT some of the iTelescope.net scopes have their transformation coefficients filled in. Is that information reliable (e.g. regularly refreshed?)

thanks, Peter Bagyinszki (BPEA)

Need professional input on some things here
KTC's picture
KTC
Offline
Joined: 2010-12-08

I try to do photometry of RU Peg....

What are your science goals for this observing project?  Determine the supercycle interval?  Monitor for outburst/superoutburst and then report that fact in a timely manner?  Other?

Your science goals will drive your observing project.  (And depending on the professional...some may not desire filtered data.  CBA almost always works with unfiltered data because most of the phenomena from cataclysmic variables that CBA studies are 'white')

1) Since the magnitude is clearly varying during the session I think it's best to report them separately instead of stacking - the SNR is not too big (~100-200) but I think important information would be lost during stacking. Is that correct?

Again, what are the science goals?  CBA may often organize observing campaigns on CV's...that require many hours of time-series data from widely scattered longitudes...over the course of weeks, months, years, and maybe even decades.

What can you accomplish with three images in R and I?  (And why not V, or why not unfiltered?)

3) As I read it's good practice to transform the magnitudes to a standard magnitude. Since the magnitude is varying during the session the values in (r-i) would be the inst. magnitudes for /different/ real magnitudes. What's the best practice to solve this problem? How to transform the magnitudes of rapidly varying stars?

Some professional scientists do not attempt to transform such measurements.  Instead, they (if there is overlap with multiple observers) add/subtract offsets to 'align' everyone's data.  This is often done when searching for periodic behavior in large data sets of cataclysmic variables. 

See my next comments....

4) In VPHOT, during Single Image Photometry at the "photometry report" page there is a "Transform" checkbox. If I understand transformations right, I need at least two filters for that. How does that Transform checkbox work with single images?

Single-filter 'pseudo transformed magnitudes'...when you only have data/images from one filter.  You need the transform coefficient for that filter (R, for example) and associated color index (V-R, or R-I).  You need to know the R mags and color index values for your comp stars (that is getting easier in the post-APASS world)...and you need to either know, or assume a color index for your target star.  (If it's a well-studied type of star...then this should not be a problem.)  VPHOT can perform this single-filter correction...if you provide comp star magnitudes, color index values, and assume a color index value for the target.

Many cataclysmic variables have a color index B-V that is close to zero.  But if you assume a color index for the target, place that information in the comments of your reports.

5) In VPHOT some of the iTelescope.net scopes have their transformation coefficients filled in. Is that information reliable (e.g. regularly refreshed?)

I recommend you contact iTelescope personnel for that information.

I hope this helps.

Photometry of Rapidly Varying Stars
WBY's picture
WBY
Offline
Joined: 2010-07-24

"2) According to the Uncertainty class if I have more than one image, I should report the stddev. of check star of similar magnitude. If I report 3 values for each filter, is it correct if all magnitudes per filter has the same uncertainity, the stddev. of the check star in that filter?"

If you re-examine the uncertainty "preference tree" in the course outline of the class you will notice the criteria at the to is "Do you have more than one image THAT YOU WILL REPORT AS A SINGLE DATA POINT."

If I understand correctly, you plan to report the magnitudes from your individual obsrvations separately. Therefore, you would NOT be using the uncertainty of the comparison star determined from the multiple images. You would use the other branch of the chart. The method you use in the right branch of the tree depends on whether you performed ensemble photometry or single comp star photometry. If used an ensemble of several stars you use the standard deviations of the comp stars from their sequence values in a particular image. This is essentially giving you your zero point uncertainty  for the mmagnitude measurement in that image, and zero point uncertainty dominates except in very low SNR situations.  If you used a single comp star, you can use 1/SNR of the target in the image since SNR is greater than 100 or, if your program supplies it for you, use the CCD error equation result since it takes into account more error sources. These techniques underestimates the uncertainty in your measurements because they don't take into account zero point error which is included when you bin the measurments from multiple images or use the standard deviations of the comp stars from the sequence value.

In my opinion if you are using 1/SNR  you should report the error as SQRT(1/SNRtarget^2+1/SNRcomp^2) since the noise in both the target magnitude measurement and the comp measurement contribute to the uncertainty in the differential magnitude. For example, If the comp and target have equal SNRs (or CCD error equation uncertainties), the uncertainty that results from subtracting the two instrumental magnitudes is SQRT(2) times the uncertainty of the individual instrumental magnitudes. If the comp star has lower SNR than the target, its contribution dominates the uncertainty in the differential magnitude. If the Target star has lower SNR, its contribution dominates.

I don't know over what time period you made you observations in a session. Normally if I want to transform my results I take images in blocks that result in as close to a common midpoint when I bin the measurements. In your case I would take IRRI blocks. That way when you average the  results of 1 or more blocks, the average airmass, is the same for all filters. That minimizes extinction differences.  

The thing you have to be careful about with single filter "pseudo" transformations is that the color indexes of variable stars can change over very short periods of time depending on where you are in the light curve. For example, look at the light curve of RU peg over the last 10 days. The observed B-V has changed as it enters outburst from about B-V ~ 0.8 to B-V ~0.3. So you need a transformed color index for the filters you are using for the point in the light curve applicable to your observations. It may take more time to do the iresearch to find out if such data exists than it does to take and process images in multiple filters.

I don't know the exposure times you need, but at magnitude 13 with a 10" telescope you should be able to get SNR over 100 in V,R,or I in  2 minutes or less with a 10" aperture telescope for stars with B-V color indexes that are >/= zero. So in about 10 minutes or less you can take a IRRI bock of images (assuming you have a filter wheel). Even during the fast part of the rise, that should be fast enough for RU Peg. Even in the steepest part of the rise, the visual magnitude doesn't appear to change more than approximately 0.05 magnitudes per hour based on previous observations. therefore obsevations over 10 minutes taken so that they average to the same point in time should give you good data that allows you to transform.

AAVSO 49 Bay State Rd. Cambridge, MA 02138 aavso@aavso.org 617-354-0484