October science update: Part II
In this newsletter, we will share the latest updates on the science side of the operation. We have spent the last month tweaking and playing with the specifics of our observing campaign. Also, with winter coming, we have been setting up some preparations for when the star is traditionally out of range.
Updates to the Observations
Astronomy is limited as a science in that we typically cannot do much in a laboratory setting. We can't build a baby star and scale it up, we can't repeat the Big Bang and watch it play out, and we definitely cannot go out to WTF and see with our own eyes what is going on. Instead, we get really clever with how we use the light coming from these objects. You could then guess that astronomers want ALL THE STARLIGHT. This is kind of true, but it is possible to have too much starlight. When this happens in an image the star or object from which we counted too much light is called saturated. It looks at its worst like this:
This is an image from ESO showing bleed lines (vertical stripes) on some bright stars. As you can probably guess, we have experienced some saturation issues in our data, but nothing as blatant as that image. Ours looks more like this:
Do you see it? Not really, no. A better, and more analytical way to quantify this saturation is to look at a slice through the star and see what the profile of the counts looks like. We would expect a bell curve, a normal distribution. If the star was saturated, we would expect a plateau or a mesa in the counts. Here is one of the slices demonstrating saturation:
So, how did this happen? Excellent question, internal dialogue! This is because we planned our observations with a very particular setup. We wanted as much light as possible without saturating. This required de-focusing (deliberately making our stars bigger and fuzzier on the chip) to avoid the saturation. Due to a hiccup in the telescope following the focus commands it was sent (which apparently affected other users as well), our images were actually focused rather than unfocused. These *focused* images were saturated, making light curves that look like this:
This plot was put together quickly in AstroImageJ, which is freely available if you want to analyze some data on your own. Simple GUI and beginner-friendly :p. We identified the saturated images in the above graph based on the brightness levels of the pixels illuminated by the stars. Note the value at the top of the plot window for the RMS (root mean square) is 0.057, or about a 6% random fluctuation.
This looks not good: there are many images which are affected and the data are very scattered. All is not lost though. This only affected a small fraction of the images taken and was primarily in the longer wavelengths. These images are also not worthless. Some of these can be rescued by using different comparison stars or by cleverly applying some tricks with apertures and recalibration. We have also appropriately adjusted our observation parameters so that there is no longer the threat of saturation.
Let's end this on a lovely positive note though. Here is what all of our unsaturated data in r-band looks like on the same scale as before. Note that the data are much more closely gathered:
Note that the RMS scatter is now 0.0037, or ~0.4%; a significant improvement. This is also on the same vertical scaling as the saturated data in the graph above, so that the comparison is fair.
Looking at the last ten or so observations, the uncertainties are noticeably larger. The particularly large one is likely a poor exposure due to weather or tracking, and might be removed later if there is a reason to do so. These exposures have a larger uncertainty because we have less light in each image. Our uncertainty in our measurement decreases approximately with the square root of the number of counts. As the exposures are anywhere from half to a third of the time, the uncertainties should be about 1.5 (sqrt(2)) times as large.
Updates on future observations and LCOGT
Unfortunately, as winter is rolling around, we will soon lose the ability to observe our target from Earth. All is not lost though. This gives us a chance to buckle down and polish up our analysis and current data. By the time the star is visible again early next year, the funds raised from our Kickstarter program will "kick" in to observe the star.
We are fortunate enough to be able to continue the monitoring during this time from space with the Swift Telescope. This telescope is designed to study Gamma-ray bursts, but when it is not observing these objects, it can do other science. Swift's mission control is at Penn State - which happens to be Jason Wright's stomping grounds - and Jason has been successful in proposing for Swift time. Yeah! Here is a preview of what that data set we have for the star already looks like. The different colors plotted indicate different filters used to make the observations (a description of the Swift filter set is here).
We also want to make a quick note that the Las Cumbres Observatory Global Telescope Network (LCOGT) has changed its name to simply Las Cumbres Observatory (LCO). They have also updated their web address to lco.global. They have a brief update about this change on their website that is worth a read. To us, and to misquote Mister Shakespeare, a telescope by any other name operates just as sweet.
Thanks for tuning in!