Suspicious Signals: Newsletter Challenge (July 2018)

False positive transit events can occur when the star is a member of a close binary system and the transit is another star.

"...experience with transit photometry has shown that the method tends to produce "false positives" -- instances when a binary star is mistaken for a planet orbiting a star. Normally, the difference a planet transiting a star and one star transiting another is one of degree: the planet, being smaller, will create a much shallower dip in a star's luminosity than a transiting star would."

http://www.planetary.org/explore/space-topics/exoplanets/transit-photometry.html

An occultation of a distant star by an object that is part of our solar system (Kuyper Belt or Oort Cloud object) could be mistaken as being due to an exoplanet around a distant star.

Also possible, though less likely, is that one or more of the 'comparison' stars, used to calibrate a given star's steady state luminance, underwent a flare event causing a momentary brightening. This would have had the effect of making the observed star appear to go momentarily dim, as though an exoplanet passed in front of the star.

This would have had the effect of making the observed star appear to go momentarily dim, as though an exoplanet passed in front of the star.

Yeah, it would be rather suspect if all stars under observation had a planetary occultation at the same time!

???? What? No...

Objects in the Kuiper Belt / Oort cloud are tiny.

You aren't that clueless. You should know better.

I have assumed (dangerous, I know) that they must observe at least two occultations having similar characteristics before declaring the occultation to be caused by a planet. Two occultations by any given planet can only occur over a period of one full orbit of that planet. That time can be very short, but if our solar system is anything close to typical, the average orbital period is several Earth years.

Kuiper belt and Oort cloud objects are so far from the sun that their orbital period is measured in hundreds of Earth years. I'm not sure when astronomers first started measuring occultations, but it was well under a hundred years ago, so IF my assumption is correct, then the false positive can't be due to a Kuiper belt or Oort cloud object. No one has ever observed two occultations by any Kuiper belt or Oort cloud object. And of course, no single person ever will, except by using historical records of some kind.

I would think you would need at least 3 occultations, evenly spaced, to be sure the occultations are due to a planet and not unconnected events.

The occultation method is used in conjunction with the spectrographic method. Motion of the star due to the planet can be detected by a Doppler shift of emission lines. Occultation depth gives a ratio of the planet and star size. (The star size is known from spectrographic data.) The amount of Doppler shift and period give the planet's mass, so the density (type of planet) can be determined. Absorption lines in the spectra of the star can even identify the composition of the planet's atmosphere.

Doppler method:

Diagram showing how a smaller object (such as an extrasolar planet) orbiting a larger object (such as a star) could produce changes in position and velocity of the latter as they orbit their common center of mass (red cross).

https://en.wikipedia.org/wiki/Doppler_spectroscopy

Both the occultation method and spectrographic method are most sensitive to planets close to the star. (The sun subtends 1/2 degree from the earth, so the chance of detecting the earth by the occultation method would be less than 0.44%.)

Possibly the answer to this puzzle is that there was no confirmatory spectrographic data, the density and planet size were incompatible.

From Planetary Society:

"Transit Photometry

A Method for Finding Earths

This method detects distant planets by measuring the minute dimming of a star as an orbiting planet passes between it and the Earth. The passage of a planet between a star and the Earth is called a "transit." If such a dimming is detected at regular intervals and lasts a fixed length of time, then it is very probable that a planet is orbiting the star and passing in front of it once every orbital period.

The dimming of a star during transit directly reflects the size ratio between the star and the planet: A small planet transiting a large star will create only a slight dimming, while a large planet transiting a small star will have a more noticeable effect. The size of the host star can be known with considerable accuracy from its spectrum, and photometry therefore gives astronomers a good estimate of the orbiting planet's size, but not its mass. This makes photometry an excellent complement to the spectroscopic method, which provides an estimate of a planet's mass, but not its size. Using both methods, combining mass and size, scientists can calculate the planet's density, an important step towards assessing its composition."

http://www.planetary.org/explore/space-topics/exoplanets/transit-photometry.html

In order for us to be able to observe an occultation in the first place, the plane of the ecliptic of the solar system being observed must be pretty close to passing through the observer's location, which is equivalent to saying that the axis of the solar system being observed must be pretty close to perpendicular to the line-of-sight. Any idea what fraction of solar systems satisfy that requirement?

Most galaxies appear to be pretty flat, so I suspect that the ecliptics of many of the solar systems in a given galaxy have very similar ecliptics, having been mostly formed originally from a single spinning mass of material. If this is correct, then if we can detect planets around one star of a given galaxy, we should be able to detect many from that galaxy. But what fraction of galaxies have their axis of spin perpendicular to our line-of-sight?

If d is the size in radians of the star as seen from the planet (1/2 degree in the case of earth), then you can imagine a band around a unit sphere (r=1) of width "d radians". "d" equals the diameter of the star divided by the orbital radius of the planet. The direction of observation can be any direction of equal probability, so the probability of seeing an occultation is the ratio of the area of this band to the area of the sphere,

(r=1)

Area ratio = (Area of band)/(Area of sphere) = (2*pi*d)/(4*pi) = d/2

(d is small compared with the radius of the sphere).

The plane of the ecliptic is 60 degrees from the galactic plane, so I would guess that other planetary systems would be oriented in random directions. Since the solar system condensed from a tiny amount of gas compared to the entire galaxy, it's initial total angular momentum was apparently not that close to the angular momentum of the galaxy as a whole.

"The galactic plane is tilted at an angle of 63 degrees to the celestial equator and at an angle of 60 degrees to the ecliptic (the path of the Sun on the sky). All three coordinate systems have poles which point in different directions on the sky."

astronomy.swin.edu.au/cosmos/G/Galactic+Plane

Thanks. That confirms your prior value of 0.44% probability for the Earth/Sun system.

I did not realize/remember that there was such a large angle between the solar system axis and the galactic axis. That definitely shoots down one of my assumptions, and significantly increases my estimate of the fraction of solar systems that might be eligible for occultation as seen from Earth, especially for galaxies that we don't see edge-on.

Where did you get the notion that a star is a half a degree? The Sun is about a half a degree, as is the Moon.

But stars are essentially mathematical points. The 'Airy Disk' of a star might be a half of an arc second, but that's due to diffraction in an optical system. Add in 'seeing' due to atmospheric turbulence, and a star might be a fuzzy ball a couple arc seconds across.

Thus it would be possible for a distant Oort cloud object to transit a star and produce the dip in the light curve like an exoplanet.

Sorry for the confusion.

In #11, I was describing the limited percentage of planets that could be detected by the transit method, and I used the earth as an example. The sun as seen from earth subtends about 1/2 degree, so you could only detect the earth from a distant star if the star were +/- 1/4 degree angle of the ecliptic plane.

Consider a sphere with the sun at the center. Around the equator (ecliptic plane) is a band 1/2 degree wide. If you divide the area of this band by the total area of the sphere you get the probability that the earth will transit the sun as seen from a distant star in any random direction.

That is only true for an actual occultation or transit. I believe the Kepler telescope predominantly detected planets by near occultation and near transits. I don't believe Tess is using a different analysis approach than Kepler did. Particularly since Tess is looking at the whole sky and not a tiny section as Kepler did.

Think about our view of Venus. Venus is brightest in our sky when only one quarter of her illuminated surface is facing us. We are close enough to Venus and Sol that this is an appreciable angular distance in the sky between these two bodies when this alignment happens. Between any star and its exo-planet there will be virtually no angular difference for any but the narrowest of viewing angle telescope. Tess is certainly not a narrow angle telescope. Thus a brightness and dimming from a planet will happen regardless of a true occultation or transit.

Interesting, "hot Jupiters" have about as much reflected light as is blocked by an earth size transit...

"In addition to transits, planets orbiting around their stars undergo reflected-light variations—like the Moon, they go through phases from full to new and back again. Because Kepler cannot resolve the planet from the star, it sees only the combined light, and the brightness of the host star seems to change over each orbit in a periodic manner. Although the effect is small—the photometric precision required to see a close-in giant planet is about the same as to detect an Earth-sized planet in transit across a solar-type star—Jupiter-sized planets with an orbital period of a few days or less are detectable by sensitive space telescopes such as Kepler. In the long run, this method may help find more planets than the transit method, because the reflected light variation with orbital phase is largely independent of the planet's orbital inclination, and does not require the planet to pass in front of the disk of the star. In addition, the phase function of a giant planet is also a function of its thermal properties and atmosphere, if any. Therefore, the phase curve may constrain other planetary properties, such as the particle size distribution of the atmospheric particles.^[87] "

https://en.wikipedia.org/wiki/Kepler_(spacecraft)#cite_note-89

https://arxiv.org/abs/astro-ph/0305473

This is precisely why I say the crux of this challenge question lies in the hidden implication enough data existed to announce a new exo-planet discovery happened and after an even shorter amount of time since starting to gather data retracting the initial discovery.

There are billions (...and billions...) of Oort and Kuiper Belt objects. Their orbital speed is extremely slow, but the spacecraft is moving relatively fast by comparison. Thus it's possible that first one, and then a second distant (yet solar system) object would appear to pass between a very distant star and the spacecraft.

Not unlike, while driving on a freeway, seeing a distant tree pass in front of an even more remote silo.

Or more likely, the star under observation could turn out to be a variable star.

A whole variety of false positive scenarios are possible. I suspect the most common but rarely reported false positive are the human errors we all catch before publishing. (Too much tequila on Tuesday can make Wednesday prone to errors.) Most of these errors will show themselves in this case by all stars having identical dips in luminance.

The other plausible false positive that comes to my mind is a random ionizing radiation discharge in the CCD pixel that produces the anomalous dip in brightness.

What will be interesting hear is what clues led to a false positive conclusion.

Then again, the occulting death star may have just jumped to another star system.

Well, after the hubble telescope showed that the universe is more crowed than thought, I imagine that there are also a lot more optical/sensory illusions going on out there. With all the lensing, black holes, dark matter/energy, and other "interference" out there, I think it's amazing that anyone can keep it all straight, at all.

There was a fly on the lens!

Bazzer

When I was a kid, there used to be a theory that the Moon was made of 'Blue Cheese'.

After all that research, and money spent on it, we now know it isn't.

But never mind, a good story at the time, however wide of the mark, kept the programme running.

All we need is a good theory........

Oops, the original data set was called into question when it was determined that the telescope hasn't been active long enough (launched April 18, 2018) to get a second occultation on a transit.

It was later determined that one of the National Park Service's toy dogs sent to the ISS was taken out for "walkies" after the trip and did some business that drifted across the telescopes line of sight, generating the obviously odorous occultation in the data.

Answer is posted.

So I guess with the exception of my facetious Death Star explanation, we are all correct.