Looking at the Light Curve III: Thar Be an Eclipse!

This is the third post in a series about the epsilon Aurigae light curve. For background, consult part 1 and part 2 and/or our tutorial on how to read a light curve.

To the right is a sample of data from the Quick Look file over the past 3 days (as of this posting). As you can see, the dataset as a whole is fainter than the data sets in the previous installments of this series. Now let's look at the light curve:

The data in this light curve are visual observations only. The red line is a 2-day average with errorbars. (We had an interesting discussion of error bars in the comments section of the last installment, click here and scroll to the bottom to check it out.) It does seem as if the star has definitely dimmed somewhat over the timescale in this plot (50 days). But it is a slow dimming. So let's look at a longer term light curve to see if this is periodic or something new:

According to this light curve, the dimming has been consistent since epsilon Aurigae emerged from the glare of the Sun (that is the source of the gap at the start of the light curve). Note that I increased the bin size of the average mean curve. Since we have much more data (because this covers many more days) we can make the bins larger, thus improving our statistics. There is no hard and fast rule on how to determine your bin size. Basically you just go with trial and error, taking into consideration how much data you have (and how it is spread out) and what kind of features you are looking for.

So now let's look to see what epsilon Aurigae was like before it went into the solar gap:

This is a 6-month light curve. And look at that! Epsilon Aurigae was getting brighter before the Sun blocked it out. This could be a result of the cycle-to-cycle variations we talked about in the last post. Epsilon Aurigae is known to have a background variation. It averages around 65 days, but it varies a lot and is not strictly periodic. In fact, the period is getting shorter and is  one of the most intriguing mysteries about this star. But this looks it has a period of more like 120 days or so (measured by eyeball).

If this long term (~4 month) increase and then decrease was caused by the background periodicity, then the minima (when the star reaches the faintest part of its light curve) must have occured during the solar gap (hence why we can't see it). A way to test it would be to take our data over the last few years, find the best periods in it and plot a phase diagram of it to see what the shape of the period is. Then we could overlay the shape on this light curve to see what the star's brightness was expected to be last summer. But that's not what this post is about, so I'll move on. Plotting phase diagrams is going to be a function of the VStar software and we'll have similar tutorials on it sometime this winter.

Back to our light curve:L I don't think it is periodic variation. We can see that the current dip is lower than the minima from the last recorded dip (just prior to JD 2454948). And if you look at the really long light curve, you'll see that the background variation rarely dips below 3.1. But it isn't substantially fainter and well within the error. So while the visual data looks like an eclipse has started, we can't be 100% certain based on this data alone.

However, remember the place all this in context. What is our core research question? It is: Is the eclipse starting? So let's look at all the additional info we have available. First, here is the visual data plus photometric V:

The downward trend in visual observations is backed up by the photometric V-band data. Also, remember that the eclipse is predicted to start now (actually, it may be somewhat overdue) based on centuries of past light curve data (namely, the average period between the previous eclipses) and also Robin has some spectroscopic evidence of the beginning. That begs the next question: Does our current light curve reflect what we expect the beginning to look like? That is, is the slope of the decline similar to past eclipses?

This is a light curve of the start of the 1982 eclipse. Using just my eyeballs to make an estimate of the light curve, it looks like the drop takes about 100-130 days and is roughly consistent (perhaps, and only just perhaps, there is a slight pause in the middle). That would make for a slow decline at a rate that I think is more or less what we're seeing now. So, yeah, my money right now is on this being the real start of the eclipse. (Of course, that's easy for me to say. As a graduate student - I have no money!) At this rate, I think we could make a pretty definitive statement in about 3-4 weeks, when it has declined about another tenth of a magnitude.

Of course, this is based on a big assumption: that the current eclipse will be similar to the last one. With this weird system, that may not be the case at all. There has been lots of variation recorded between epsilon Aurigae eclipses in the past. So don't let this extrapolation fool you. The only way to know for sure is to go out there and make an honest observation.

So there we have it. Most likely the party has begun. A more accurate subject line would be "Thar be the beginning of an eclipse...". We have almost two years of excitement ahead of us.

p/s: Remember to get out there and observe! The more quality data we have, the more we'll know about what is actually going on right now. The star is well positioned for observing around 10-12pm now, so no excuses! (and keep looking at the other 10 stars too, in the future we'll be looking at those light curves as well)