I have been investigating methods of doing photometry when there are no good comparison stars within a few degrees. In these cases a star will appear to fade as it sets and there is no good way to correct for it (as yet).
It is therefore hard to tell if the star has short term variability. I was playing with data from Alpha and Beta Peg and trying to correct the fade off with a simple 1/cos(zenith angle) correction. I am aware that this correction is inexact as it assumes the Earth is flat. A good link on this was pointed out by Noah previously as
http://cfa-www.harvard.edu/icq/ICQExtinct.html .
Tonight I found an indication (OK, found a NEW indication) that atmospheric extinction is not the only factor that causes a star to dim as it nears the horizon. Even though the extinction correction I applied was inexact, the correction factor does not appear large enough to explain the obseved fading. Another factor is operating, and I assume that this factor is vignetting -- that the lens itself passes less light for off-axis objects. Now vignetting through a fish-eye lens is a strange effect, and might be well understood and corrected for by a simple correction factor. But the point of this post is just to point out that such vignetting appears to exist. It might turn out that the only way to correct for object dimming due to a different altitude is empirically.
- RJN
Vignetting
Here is an idea that appears to get rid of both vignetting and atmospheric absorption effects in the photometry. For a comparison star, use the SAME star on the previous night. Take the ratio of photometric values on the two nights at the same SIDEREAL time, not the same SOLAR time. At the same sidereal time, all the stars in the frames will appear exactly in the same place -- only the Sun and Moon will appear to have moved. The practical way of doing this with NSL photometry files is to match altitude for the star for each night.
This is uniquely possible with CONCAM data since CONCAMs all start a new integration every 3m56s. Therefore a frame taken one night has the same sidereal time as the "next" frame on the previous night.
This appeared to work for Alpha Peg. The odd thing was that the ratio appeared to be not exactly one, but rather 1.05 with a standard deviation of less than 0.5 (taking out the bad data points). Most of the ratios were above 1 -- too many for chance. So either the star changed its brightness by about 5 percent between the two nights, or one night was 5 percent cloudier, or something else completely.
But one thing was clear -- there was no systematic fading as the star neared the horizon. If the fading was NOT atmospheric/vignetting but rather intrinsic to the star, the variability would surely NOT be identical on two consecutive nights. So the constant plot indicates that Alpha Peg was constant to within 0.1 magnitudes for the four hours of the plot(s).
- RJN
This is uniquely possible with CONCAM data since CONCAMs all start a new integration every 3m56s. Therefore a frame taken one night has the same sidereal time as the "next" frame on the previous night.
This appeared to work for Alpha Peg. The odd thing was that the ratio appeared to be not exactly one, but rather 1.05 with a standard deviation of less than 0.5 (taking out the bad data points). Most of the ratios were above 1 -- too many for chance. So either the star changed its brightness by about 5 percent between the two nights, or one night was 5 percent cloudier, or something else completely.
But one thing was clear -- there was no systematic fading as the star neared the horizon. If the fading was NOT atmospheric/vignetting but rather intrinsic to the star, the variability would surely NOT be identical on two consecutive nights. So the constant plot indicates that Alpha Peg was constant to within 0.1 magnitudes for the four hours of the plot(s).
- RJN