Thursday, February 4, 2016

Tabby's Star for the Perplexed

Last Update: 5 August 2016

Update: Montet/Simon Preprint
Earlier this year, Brad Schaefer stated that Ben Montet was working on the question of secular fading over the four years of Kepler primary mission data when Tabby's Star was visible, and that he was seeing fading. A preprint came out last night confirming this, and in fact the fading was quite dramatic at times. There are lots of questions, and I suppose there will be controversy, but it's quite important if it holds up. We'll have more soon.

Expecting data soon from the the Kickstarter funded observations by the LCOGT. Stand by...

Related Wow! Signal Podcast Audio Links:
    Tabby's Star for the Perplexed, Part 1
    Tabby's Star for Perplexed, Part 2
    The Slow and Fast Dimming of Tabby's Star
    DASCH Photometry with Josh Grindlay
    Audio Interview with Tabetha Boyajian
    Catching Tabby's Star in the Act - Interview with AAVSO's Stella Kafka

When it comes to Tabby's Star (also known as KIC 8462852), we are all perplexed. This post is for those who are disinclined to read technical papers by professional astronomers, but would like to know just what the heck is going on. What is all this stuff about alien megastructures, swarms of giant comets, infrared excess, and old photographic plates? We'll lay all that out here for you in non-technical terms (or we'll explain the terms as we go). Please, if there are any questions, ask in the comments below, and we'll try and figure out an answer, if there is one. The post is richly hyperlinked, so if you want more detail, you can easily find it. I hope I have given credit wherever it is due.

Let me start by stating up front, that no one knows exactly what is going on with this star. What we'll try to lay out here is why this otherwise ordinary star is strange. If you have questions, or find errors, or know of updates I should include, please leave a comment here.

The Basic Background Facts

First, a few basic facts we should all know. Stars vary quite a lot in size, brightness and temperature, but most are small and relatively cool and burn for tens of billions of years. A smaller population of stars are like our own sun, which is called a G class star, and some stars are even bigger and brighter than the sun. Tabby's star (an informal name - it's known in some star catalogs as TYC 3162-665-1 or KIC 8462852)  is one such, and is called an F class star - a bit bigger, quite a bit hotter, and considerably brighter than our sun. It is just under 1500 light years away (so we are seeing it as it was not quite 1500 years ago), and is in the constellation Cygnus.

KIC 8462852 at the center of the image (visualized with Aladin)

Stars are born in a collapsing cloud of gas and dust, they burn steadily for a while, and then they die. While they are burning steadily, they are called Main Sequence stars, and when they die they change size and color and can vary quite a lot in brightness . Astronomers can study the light from a star and determine the approximate age, just as your vet can tell about how old a cat is by looking at his teeth. Tabby's star is neither very young or very old - it is a main sequence star, and should burn steadily.

It turns out that Tabby's Star was one of the many stars that the planet-hunting Kepler Space telescope was staring at for about four years (starting in 2009) in an effort to find planets around other stars, or exoplanets. Kepler is able to find exoplanets because for some subset of those stars it is keeping vigil on, its planets will pass across the face of the star from our vantage point, and we will see a very slight dimming of the star when this happens. Although the resulting variation in brightness is subtle, it should repeat in a rhythm that is a fingerprint for such transits, and this fingerprint can be recognized by careful analysis. More exoplanets have been discovered with Kepler than any other telescope - 1039 confirmed exoplanets at this writing. Big exoplanets are easier to detect than small ones, and planets close to their star with shorter period orbits repeat their transits more often, and so are easier for Kepler to spot.

A problem with the Kepler spacecraft in 2013 meant that it can no longer stare at the same patch of sky that it did at first, so it is no longer keeping watch on Tabby's star. However, the data produced by Kepler continues to be analyzed by astronomers, and notably, by a group of citizen scientists called Planet Hunters. The Planet Hunters study the subtle variations in stellar light curves by eye, augmenting the sophisticated computer analysis used to detect many of the exoplanet candidates Kepler has found, and flag unusual events in the light curves for follow up by the science team.

What the Planet Hunters Found

The first two planets found by Planet Hunters were reported in 2011, and since then they have found quite a few more, but on 17 October 2015, a puzzling finding by the Planet Hunters was reported by Boyajian, et. al (we'll just call this paper B2015 from now on).

Update: The 2015 Boyajian, et. al. preprint referred to repeatedly below as B2015 is now published in the MNRAS. Here is the correct cite: (SIMBAD code: 2016MNRAS.457.3988B)
T.S. Boyajian, et. al., Planet Hunters X. KIC 8462852 -where's the flux? - Mon. Not. R. Astron. Soc., 457, 3988-4004 (2016) - 14.06.16 11.07.16 April(III) 2016 2016-04

Here is what the Planet Hunters saw:

KIC 8462852 Kepler Light Curve reported by Boyajian,
What you are seeing in this chart is a plot of the Kepler measurement of the brightness of Tabby's star (with the normal brightness equal to 1) over the roughly four years the telescope was staring at the star. The downward spikes at around Day 800 (henceforth called the D800 dip) and just after Day 1500 (D1500 dip) are real data, not mistakes, and they are much larger than the normal dips in magnitude observed when a planet transits a large star - typically close to 1%. Two of these dips (numbered 5 and 8) are around 20%, which is only achievable by an object more star sized than planet sized. Dips 9 and 10 are also considerably larger than any planetary transit. There are smaller dips as well, and the B2015 paper identifies the 10 biggest events in Table 1 of that paper. Here's a zoomed-in view of the same graph, in which the ten events are shown:

Normalized Kepler data for KIC 8462852 with dipping events numbered.
Based on its temperature (which can be derived from its color) and type, Tabby's Star is probably more than 2 million kilometers in diameter - more than 50% larger than our sun. By comparison, Jupiter is 138,350 km in diameter - so even a large planet would barely cause a dip in luminosity - a few percent at most. Whatever passed in front of Tabby's Star was big - the size of a small star itself, and yet no other close star is in evidence.

Another important thing to note is the time scale, which you can see for yourself by examining Figure 1 in B2015. The events took place on a time scale of days, and some of them about 10 days. This is a big part of the puzzle.

I would also note something else I don't know how to interpret, but the team found some evidence that a 48.4 day period is involved. Something orbiting at that period would be about 44 million kilometers from the star, which is quite close. Something that close to such a bright star would be hot, with an equilibrium temperature of 1070 degrees Kelvin, if I did my arithmetic right, and would glow fairly brightly in the infrared at a wavelength of about 3 microns. As we'll see, no evidence of an object emitting infrared light around that wavelength has been found.

The Planet Hunters Science Team's Follow Up

There is a lot of information packed into the 17 pages of B2015, and I am going to try and unpack it for you, although I am leaving out many details and interesting wrinkles.

The science team followed up the Planet Hunter's discovery by asking all the obvious questions. Could the data be in error? Are we looking at two or more stars instead of one?  Is there something unusual about this star? Does the electromagnetic energy coming from the star give us a clue as to what could cause the dips? It turns out there wasn't a whole lot of astronomical literature on this particular star. There are after all, billions of stars that telescopes can see, and most have not been the subject of close scrutiny.

The first thing they did was to check with the Kepler team. Had they seen anything like this before? Could it be something wrong with the telescope, or the sophisticated instrument on the telescope's focal plane that measured the brightness of the stars? the Kepler scientists took yet another close look, and shook their heads - the data was real, and wasn't showing up in other star's  light curves, as you would expect if it were a problem with the instrument. The Kepler scientists could find nothing wrong with the data - they believe the brightness curve - sharp dips and all - is real.

The next thing the team did was take a closer look at the data. They tried more sophisticated analysis techniques to try to understand if anything in this complex pattern repeated. If you have multiple things repeating at different intervals, they can "beat" against each other and cause patterns that at first glance have too fast or too slow a rhythm.

They found that there is a strong component at a 0.88 day period, which the B2015 team argue strongly is caused by the rotation rate of the star itself. Other data in the observational follow-up is consistent with this. Some other interesting bumps were noticed in the frequencies, including a 10-20 day pulsation, but no one knows exactly how to interpret them. They may just be natural variations in the star's brightness.

Following up with ground based telescopes

Boyajian's team followed up the Kepler observations with ground based telescope observations. An examination of observations with the United Kingdom Infrared Telescope (UKIRT) revealed a tiny speck that may or not be a small companion star. Images from the giant Keck telescope reveal that there is another, fainter star apparently close to Tabby's Star, but it is not clear if it is a distant companion or just another star passing close to the same line of sight. Observations were able to rule a close companion or a bright companion, but it is possible that a red dwarf star circles Tabby's star at a distance.

The B2015 team were mainly interested in detailed analysis of the star's light, using the science of spectroscopy, which is crucial to our entire understanding of the universe. Spectroscopy can not only tell us what sort of elements stars are made of, but since we know what the light spectrum looks like from laboratory work, it can also tell us how fast that material is moving either towards or away from us, since the wavelengths shift up (away) or down (towards) due to the Doppler effect. Over many generations of studying spectra, astronomers have learned to infer a great deal about a star from its spectrum and measurement of the Doppler effect on it.

One thing they learned from the spectroscopy was that the star is in fact an unremarkable star, of a fairly bright, yellow-white class called F, which constitutes about 3% of main sequence stars. The star appears to be rotating fairly fast, with a period of 0.88 days (compare to our Sun's 25 days). They were able to estimate the mass and size of the star, and determine that it is likely to be neither a very young F star, nor very old. If it is in the middle of its life cycle, its brightness should be steady.

Radial Velocity Measurements

Radial velocity is the movement toward or away from an observer, and can be determined from looking at the spectrum of light coming from a star or galaxy. It is the same technique Edwin Hubble used in the early 20th century to show that the universe is expanding. The radial velocity can do couple of things for us:
  • Although no one thinks it's at all likely anyway, it can show that the star is not moving at ridiculously high velocities towards or way from us, which could have weird effects.
  • If Tabby's star has a large, dark companion that orbits closely, you would expect to see wobbles in the radial velocity.
The radial velocity measurements performed on Tabby's Star were not the most accurate possible (you need a bigger telescope and more time), but they were good enough to show that there is no large, dark companion close to Tabby's star, and that its movement through space is nothing extraordinary. We can't however, rule out a big dark companion object further away from the star, but it could not cause both sets of dips, since it would take it too long to orbit the star.

Infrared Excess 

If some solid (or liquid) object is absorbing 20% of the energy of the light from Tabby's star in our line of sight, it should heat up, and then basic physics says that energy will be emitted in infrared light, which is just light with a longer wavelength than the human eye can see. So, if we plot the energy we're seeing vs. wavelength of light, then we should see the normal light from Tabby's star, and then a bump out in the infrared, or even in what are called millimeter waves if the object is far from the star and thus not getting as hot. The "bump" is what is called an "infrared excess", and the B2015 team looked for one in the space telescope data available to them, and with a ground based telescope. They did not see any excess.
The heat emission spectrum from an object at 300 degrees Kelvin, from Wolfram Alpha

The B2015 team and others looked in the survey data from the WISE infrared space telescope, and later followed up with data from the Spitzer Space Telescope and with both an infrared and a millimeter wave telescope on Mauna Kea, and found no clear evidence of anything absorbing the light from Tabby's star and re-emitting it as waste heat. There were some little hints that something might be there in the Spitzer data at 4.5 microns wavelength, but we shouldn't get our hopes up.Where we would expect to see the infrared spectrum peak depends on your hypothesis about how close to the star the absorber is.

B2015 examines some possible explanations 

So, what could explain the dips, and everything we know about them? The lack of an infrared excess ruled out large amounts of solid matter passing in front of a star, as you might have resulting from a major planetary catastrophe, such as two big planets colliding. Also, if this were the case, you would expect the first dip to be much messier, and it is nice and sharp, although asymmetric. The second set of dips is more complex, though.

The B2015 team examined all the more obvious explanations: 
  • the star itself exhibits variability
  • big clumps of dust orbiting the star
  • catastrophic collisions
  • planets in the process of formation, possibly with very large ring systems
  • a swarm of large comets
All of these are ruled out by the data, except for the last one. No one knows where such a large swarm of very big comets would come from, although it is at least conceivably consistent with what is known. Comets are can obscure light in a way that would not produce as large an infrared excess - they just boil their material off into space for the most part. B2015 did not consider this explanation a hill to die on, as the MSM seemed to imply in their usually oversimplified reporting, but just the only one they had found so far that they couldn't completely rule out.

Enter the Alien Megastructures - Jason Wright's Comments

Jason Wright is Penn State astronomer who has been studying ways to seek evidence of technological ET civilizations using astronomical instruments, like infrared telescopes. He heads up the GHAT project, which has done some some interesting work in this area. He also maintains a blog called AstroWright, and at about the same time as B2015 came out, he posted a blog entry on the same topic.  He raises there the possibility that it could be a very large artificial structure, something that he and the GHAT team had been studying.

Wright did not claim that the dips were caused by a transiting megastructure, but just that Tabby's Star was in interesting candidate to study more closely, and a good new SETI target. Note that the lack of infrared excess is a problem for the most common type of concept we have for this, which is a Dyson swarm.

The Dyson swarm (also known as a "Dyson sphere", although this is a misnomer), was first proposed by physicist Freeman Dyson in 1960. It comprises a ring or rings of very large artificial satellites in a circular orbit about a star, absorbing a significant fraction of the star's energy for use by an ever-expanding civilization. While the swarm is under construction (or long after it is no longer in use), it may have large gaps along its orbit path, and its orbit plane may not be oriented so that we would see it transit frequently, if ever. The Dyson swarm would necessarily radiate waste heat in large amounts, at infrared wavelengths around 10 to 20 microns (your eye can't see light with a wavelength of 1 micron or longer), and so far no evidence of this has been seen, as we noted above.

Of course, Wright's cautious comments were blown out of proportion by the media, and B2015 doesn't even mention the possibility of artificial structures. For a couple of weeks, the headlines proclaimed alien megastructures, until more infrared results came in, in which the authors very cautiously concurred with B2015 that the comet swarm hypothesis was the only one anyone had thought of yet that was consistent with their data. The headlines now read something like, "It's Not Aliens, It's Comets!". Sigh. No responsible person ever said it was either.

SETI Searches in Radio and Optical 

Shortly after the  publication of Wright's blog post and B2015, the SETI Institute (for whom I have great respect, I want to emphasize) swung into action, using the Allen Telescope Array to look for radio signals from Tabby's Star. Of course, this was a long shot, that 1500 years ago someone orbiting that star should decide to beam a signal our way, in the frequency bands they were analyzing, over the short window of time over which they looked, but still, they had to give it a shot. After all, "they" would have been looking at the Earth at 1000 B.C., and would have seen no evidence of a civilization that mastered microwaves. Nothing was seen, but the array was not at full sensitivity, and they didn't look for a long time. Even with the future, far more sensitive SKA radio telescope, we will have to be lucky to eavesdrop on an ET civilization, so the SETI Institute really would only detect a beacon deliberately transmitted in our direction.

There were also optical searches looking for laser flashes form Tabby's Star. Again, a long shot, and again, they came up empty.

The negative SETI searches on Tabby's star to date are a perfect example of a case in which absence of evidence is not evidence of absence. I hope there will be more searches in the future with more powerful telescopes, but we would have to be very lucky to see anything.

Bradley Schaefer's Claims that Tabby's Star is Slowly Dimming

Astronomer Bradley Schaefer decided to look in the historical archive to find out what it told us about the brightness of Tabby's Star. He is an expert in deriving object brightnesses from old photographic plates, and went to the Harvard library, which holds a large collection of sky survey photographic plates dating back into the late 19th Century.

A research group called Digital Access to a Sky Century @Harvard (DASCH) has digitized more than 130,000 of these plates and determined billions of object brightnesses and derived millions of light curves, including for Tabby's Star. Anyone can go to the DASCH website and plot light curves for themselves, although it takes some insight into the process to exclude plates that could skew the data. Recent improvement to the DASCH processing pipeline have improved brightness measurement errors to about 0.1 magnitudes (a logarithmic measure of observed brightness that increases as a star dims.). Sometimes the error will be less than this, sometimes more. A little high school arithmetic will tell you that this error equals about a 9% change in brightness, so we might be lucky and catch Tabby's star acting oddly one or more of the plates. B2015 had examined the DASCH data briefly, but had not seen anything unusual. Schaefer wanted to take a closer look.

Schaefer's preprint came out in January 2016. This is a paper submitted to the Astrophysical Journal, but not yet accepted for publication. His primary finding was to my knowledge completely unprecedented, and I immediately e-mailed him and requested an interview, which he was kind enough to grant. I had hoped for five or ten minutes, but ended up with over 40 minutes of good material, and it became Episode 3 of Season 3 of the Wow! Signal.

What Schaefer found was that this main sequence star had dimmed by roughly 20% over the century for which Harvard had plates. He found the dimming trend in both in the DASCH data and his own manual measurements directly from the plates (although he has not documented the latter in depth). In both cases, the measurements are made by comparison to known stars on the same plate, so the fact that different telescopes and cameras were used over that century was accounted for. Plates with known problems were excluded from data.

A main sequence star like Tabby's Star should not dim on these time scales. There is no precedent for it, and well verified models of stellar evolution provide no explanation for it. Schaefer argues that since the star shows anomalous 20% dips on short (days) time scales and over a century, that the same or at least closely related mechanism must account for it. However, this essentially rules out the comet swarm hypothesis because of the the very large number of very large comets required to make this happen.

However, is this slow dimming real? how likely is it, given the actual data?


The Hippke Paper says there's probably no dimming

Shortly after the Schaefer preprint came out, a rebuttal of sorts was posted on Arxiv by Michael Hippke and Daniel Angerhausen. They argue that Tabby's Star is probably not dimming, becasue they found several other F class main sequence stars in the DASCH database that were also dimming - more than 18 of the 28 stars they were looking at. Shortly thereafter, Schaefer posted a rebuttal to this on the blog Centauri Dreams. This essentially amounts to a claim that the DASCH photometry is not well calibrated.

Schaefer argued that Hippke was an inexperienced user of the DASCH photometry, and had made mistakes in the selection of stars to be included in his analysis, and that Hippke's stars were not dimming at all.

I decided to call up the authority on this, Dr. Josh Grindlay, who is the PI of DASCH. In an interview on the Wow! Signal, he told me that he concurred that Hippke's analysis was in error, and that the stars Hippke found were dimming were not dimming. The DASCH photometry has been validated against the so-called Landolt Standard Stars, and they show flat light curves for these standards, demonstrating that the DASCH photometry is in fact well calibrated, to the advertised error bars of 0.1 magnitudes.

However, Grindlay also told me that he was skeptical of Schaefer's results showing that Tabby's star is dimming, since he find that eliminating all but the data points known to be trouble-free results in a flat light curve for Tabby's Star. In his view, Schaefer has more work to do make a convincing case.

Here is the plot I get from the DASCH website if I turn off all the flagged points, and use the photometric calibrations developed for the Kepler project (the Kepler Input Catalog, hence the acronym KIC). Only 178 points survive all the filters. Here the curve appears to my eye to have a slight downward slope, but the paucity of early points makes it difficult to judge.

The unflagged points for Tabby's star using the Kepler calibration (not B magnitude)

From this, I think we have to conclude that the claim of an unexplained long term dimming may be premature, although it remains at least plausible, and we are going to have to wait for further analysis. B2015 presented recent observations and found a blue magnitude ("B" in astronomical parlance) of 12.26, which is substantially dimmer (recall that higher magnitudes are dimmer), than any of the unflagged points early from the 20th Century in the DASCH database. Hippke has revised the initial paper, and significantly adjusted the finding since the initial storm of criticism.

My own, nonprofessional guess is that Schaefer is right, but he has to persuade his own professional colleagues, not me. If he is right, the mystery of the missing infrared excess deepens.


So, Where are we Now?

At present, everyone is stumped, and we need more data. The James Webb Space Telescope is still about 3 years away from on-orbit commissioning, and its "first light" image is unlikely to be Tabby's Star. It could be some time before it studies the system in depth. However, skilled amateur astronomers are out in clear nights, keeping a watch on Tabby's star for anomalous dimming events. Since these events last on the order of days, there will be time to swing a big telescope onto the star and attempt to get detailed spectra in the visible and infrared, while smaller telescopes measure the fluctuations in brightness. This could give us a much better idea of what it is that is blocking Tabby's Star. Whether this will solve or deepen the mystery, no one knows.

Some Frequently Asked Questions

Could it be a black hole causing the dimming?

Dr. Boyajian told me that she frequently gets e-mails proposing that the dimming of KIC 8462852 is due to a black hole.

No. There are three reasons, all of which are show stoppers to that idea:
  1. Black holes aren't nearly wide enough. A typical stellar mass black hole is only about 20 km in diameter, which is at least 4 orders of magnitude too small.
  2. Black holes have a lot of mass. The smallest possible black hole is more massive than Tabby's Star, and could be considerably more massive than that. It and Tabby's Star wold be orbiting around the mutual center of mass, and so astronomers should see that in their radial velocity measurements, unless it were orbiting a large distance from the star - in which case, we wouldn't get two events 700 days apart. They simply don't see the radial velocity varying in this way.
  3. A black hole passing in front of the star from our our point of view would make the star brighter, not dimmer, due to a phenomenon known as gravitational lensing, a phenomenon predicted by general relativity, and which astronomers observe all the time at different scales.

Could it be something in our own solar system blocking the light from Tabby's Star?

The short answer is no, because these events are on the order of days in duration. Asteroids and planets in our own solar system DO block the light from stars, but these events are much briefer, and they wouldn't look like the D1500 events. Comets close to our sun can partially obscure the light from a star for a while longer, but we would know about them, and Kepler does not look close to the sun.


Does Tabby's Star Have Planets?

It might very well, since statistically, it seems that nearly all stars have planets. All we know about planets for sure in this case is that Tabby's Star almost certainly doesn't have very large planets close in to the star.  The radial velocity data does seem to rule that out, but there could be other planets. However, there don't appear to be any other transiting planets  - the sort of planets Kepler is designed to detect. Kepler was only watching the star for four years, so transiting planets with longer than a four year orbit period may be out there.


What about Gravity Darkening?

In gravity darkening, a fast spinning star is brighter at the poles than the equator. It's not at all clear that without a close, massive companion, gravity darkening would work as an explanation here, even though the star is spinning fairly fast - but not fast enough for gravity darkening to be significant. B2015 did not consider this a candidate explanation, and I think mainly because it is a non-starter.


Since the SETI searches didn't find anything, are aliens ruled out?

No, not at all. The lack of observed IR excess is far harder to explain if we think the dips are caused by aliens than the (so far) negative SETI results. The searches to date haven't been very long or very sensitive, and would have only seen signals if they were sent toward us deliberately - almost 1500 years ago. If someone with advanced technology was there at that time, then they would not know that there was a technological civilization on Earth capable of receiving signals, even if they wanted to send them. It would be a long shot that they'd send them, and even longer shot that we'd be listening when the signals finally arrived.


A compendium of exotic speculations

So, while we're waiting, let's have some fun. Almost all our speculations are likely to be wrong, but out of these toy models, perhaps some better ideas will emerge. 


The highly efficient mirror

This is actually the best "alien megastructure" idea I have seen, and as far as I know was devised by Charles Engelke. Some people have proposed that an efficient mirror is positioned close to Tabby's star, and floats on the light pressure - the momentum of the enormous number of photons bouncing off the mirror.  The waste heat signal from such a mirror may not be easy to see, as most of the energy is not absorbed, but redirected out into interstellar space, where another large mirror is using it to propel a spacecraft (or some other application we haven't thought of). The waste heat may be hard to spot, not only because there is less of it, but because the temperature of the mirror is very close to that of the star, and so mimics it.

My question is: would it be possible to see the receiver? It could, after all, absorb or reflect as much 20% of the light from an F class star, so it might be quite bright, depending on how it is oriented with respect to us.

The swarm under construction

If Schaefer is right, then a Dyson Swarm could be under construction, but is still too small to see in IR when it is not transiting. Is that possible, if it is capable of obstructing 20% of the star? As far as I know, no one has considered all the possibilities, but significant infrared excess at a wavelength around 10-20 microns should be observed if 20% of the star's energy is finding its way into waste heat.

Caltech's WISE survey included wavelength bands of 12 and 22 microns, so WISE should have seen it. Although the WISE 22 micron wavelength band wasn't as sensitive, my own very crude calculation is that WISE would have no difficulty seeing the IR from such a structure. A quick look at the WISE catalog shows a very weak signal in this band (W4) for Tabby's star, which may be no better than an upper limit. So, either the structure is very cold, far from the star, or it is incredibly well aligned with our line of sight (so that is is blocking well less than 20% of the star's light overall), or some combination of the two.

The fragmentary swarm

This depends on Schaefer being wrong - that the star has not been dimming on average in the last century, but dimming only takes place during brief events. Here may have once been a Dyson Swarm, but it is no longer in use and has largely fallen apart, been disrupted by collisions, deliberately destroyed,or alternatively, has not been completed yet. They would still absorb light from the star when passing in front of it. but would not produce much IR excess at other times, especially since the surface orientation may be away from us. So, when WISE and Spitzer performed their observations, the signal may have been too weak to pick up.

Weird physics is being used to modulate the brightness

A very advanced technological civilization may be able to control stars use exotic techniques, like a massive pulse of neutrinos.