Re: How to keep FITS header when stacking?

Posted by karenacollins on
URL: http://astroimagej.170.s1.nabble.com/How-to-keep-FITS-header-when-stacking-tp725p733.html

Hi Luca,

I've seen formulae for estimating scintillation noise for single aperture photometry, but I'm not sure how to reliably
do it for differential photometry. My thinking is that if comparison stars are near the target star, the scintillation
noise will be less of an effect, and if the comparison stars are far from the target stars, the effect will be greater.
Do you have a reference on how to calculate the effect of scintillation noise in differential photometry? Beyond that,
ground based light curves almost always have correlated (red) noise of some sort too, which can't really be calculated.
When doing modeling using light curves and RV data, MCMC algorithms typically scale errors such that the reduced
Chi-square of the data fit to the model is 1.0, so as long as the relative values of the calculated errors within a
dataset are reasonable correct, the scaled errors will be reasonably good representations of the actual errors
(including scintillation and red noise).

If you are taking short exposures of focused images, inter-pixel sensitivity variations may be the dominate factor in
the scatter, unless you have perfect guiding. If you can defocus slightly and extend the exposure time a bit to 30-60
seconds, the light curve scatter will likely be relatively close to the formal calculated error. If you are observing a
crowded field, then obviously this technique should be avoided.

All of that being said, I have considered adding the single aperture scintillation noise calculations to AIJ and then
just propagating those through the differential photometry error calculations, but I just haven't had time, and it would
likely over-estimate the errors for high-quality differential photometry. Maybe someday I can get this done when I have
spare time after my "day job".

Karen