Brian Shaw and Denise Stephens, Department of Physics and Astronomy
Over the past year I have been looking at pictures filled with dots and running them through computers. On the surface that doesn’t seem quite exciting but once one understands what is happening between the pixels, the humdrum atmosphere disappears. Each of those dots represent an actual star in our galaxy. And by analyzing the pixels, I can determine properties of these stars. The Orson Pratt Observatory atop the Eyring Science Center on BYU campus has been collection data of potential transits provided by the KELT-North Project. I have been taking that data and reducing it to the point that it can be understood. In this report I will illustrate how that process works and also summarize some of the more interesting results that I found.
The first thing that I did was reduce noise, or extra data, from the pictures I analyzed. Noise can come from many things, like bright lights at night or electrical current from the telescope. I ran the pictures through the program IRAF, which turns hazy pictures to crisp shots of star fields. The next step I did was run the clean pictures through the program AstroImageJ. This converted pixel counts on the stars into flux, a measurement of brightness. Flux in this context is defined as the amount of light that passes through space to reach us during a certain amount of time. After this, I converted flux into relative flux. “Normal” stars are ones that have constant flux over the night. By comparing the relative flux of my target star to normal ones, I could see if the target had a constant flux as well. This is done by averaging out chosen comparison stars and subtracting them from the target. If a star did have a transiting planet, that relative flux would change over the night. If it did, I would move to the fourth step. I took the relative flux measurements and graphed them together. If 100 pictures were taken of the star over the night, the graph shows 100 flux counts over that time. How the flux changes tells us what is happening to the star. However, not all relative flux changes mean a transiting planet. It could be a binary system and to rule that possibility out I compared filters. Possible transits were recorded on the telescope multiple times, and in multiple light filters. If the flux dips lower in one filter than the other, it means the star is actually two eclipsing binary stars. Finally, when a transit was confirmed, I analyzed the graph to find how long the transit was and how much the flux changed.
Figures 1 and 2 are both graphs of KC05 and its change in flux in two different filters. Since the flux of both graphs dip the same amount, I knew it was not an eclipsing binary. No difference in filters means the eclipsing object is cold; it doesn’t give off any light of its own. That makes it a planet instead of another star. KC04 (Figure 3) was interesting because it shows the transit coming out of the dip, or the wing of the transit. Moreover, the transit had already started by the time the telescope started collecting data. Variations in predicted transit time implies that there could be other planets besides the first that are gravitationally pulling on it, affecting its orbit. All three of the flux changes in my table are in the decimal places. This appears to be a small difference but it is nonetheless detected by our telescopes. Given the proportional sizes of stars to planets, this small decimal amount is justifiable. It also makes sense why excessive noise could drown the transit out. Figure 4 shows us an example that even with noise reduction, some data and images are still unusable. There was a forest fire on the night KC20 was watched and the smoke in the sky ruined the data.
Apart from confirming transiting exo-planets, several other good things have come from my research. I learned two computer programs and several methods useful for conduction astronomy research. I also have been approved to go to the upcoming AAS Meeting in Seattle in January 2015. There, I will be assisting Dr. Stephens in her presentation of educational research for undergraduate astronomy majors.
1 To preserve integrity of the KELT-North Project, star names have been altered and star coordinates have been omitted. 2 The change of flux is defined as how far the graph dips down from the “normal” constant flux