Useful color dependence detected?
Posted: Sun Sep 19, 2004 8:55 pm
I have been playing with MK data for Alpha Cas (a red K-star) and Beta Cas (a green F star) for 040918 and 040917. In another thread I discussed how I was able to show that both stars were stable at about the 0.3 percent level between the two nights. This gave confidence that the photometric measurements did not have any large systematic errors.
So I decided to check, among other things, (C16-B)/(C5-B) for each star on 040918. (C16 is the average of the 16 brightest pixels for Alpha Cas, while B means an independent background level computed over the whole frame simultaneously.) Curiously, this term drifted appreciably for both stars over the night, with the larger drift for Alpha Cas! Strange!
I then compared the nightly averages of (C16-B)/(C5-B) over the night for both stars. They were significantly different. Alpha Cas has (C16-B)/(C5-B) = 0.5056 +/- 0.0017 (standard error), while Beta Cas has (C16-B)/(C5-B) = 0.4604 +/- 0.0021. So, formally, the difference is significant at about the 20 sigma level. And the stars occupy a range of altitude from 29 to 50 degrees.
One possible reason for the difference is the difference in the color of the stars. Chromatic aberation through the lens might cause stars of different colors to have (slightly) different point spread functions (PSFs) in CONCAM3 data. Another possibility is the different amount of star trailing that occurs with position on the CCD, but I need to think more about that one.
If chromatic aberation is the reason, it might become a useful tool for crudely guessing the colors of transients and even the amount of aeresols in the Earth's atmosphere for its effects on the colors of stars near the horizon.
- RJN
So I decided to check, among other things, (C16-B)/(C5-B) for each star on 040918. (C16 is the average of the 16 brightest pixels for Alpha Cas, while B means an independent background level computed over the whole frame simultaneously.) Curiously, this term drifted appreciably for both stars over the night, with the larger drift for Alpha Cas! Strange!
I then compared the nightly averages of (C16-B)/(C5-B) over the night for both stars. They were significantly different. Alpha Cas has (C16-B)/(C5-B) = 0.5056 +/- 0.0017 (standard error), while Beta Cas has (C16-B)/(C5-B) = 0.4604 +/- 0.0021. So, formally, the difference is significant at about the 20 sigma level. And the stars occupy a range of altitude from 29 to 50 degrees.
One possible reason for the difference is the difference in the color of the stars. Chromatic aberation through the lens might cause stars of different colors to have (slightly) different point spread functions (PSFs) in CONCAM3 data. Another possibility is the different amount of star trailing that occurs with position on the CCD, but I need to think more about that one.
If chromatic aberation is the reason, it might become a useful tool for crudely guessing the colors of transients and even the amount of aeresols in the Earth's atmosphere for its effects on the colors of stars near the horizon.
- RJN