So, the media is abuzz about a BLC1, a candidate signal around Proxima. I’ve been all over Twitter about this, so I’m collecting my thoughts here.

But first, a disclaimer: as a member of the Breakthrough Listen Advisory Board and a Breakthrough Listen member’s current PhD adviser, I have a little bit more information than the public than this, but I am not a BL team member and have not seen the data. My comments here are purely general and, while they can provide context for what’s going on, they do not actually add anything to what’s known about the actual candidate signal beyond what is already in the press.

The Breakthrough Listen team uses radio telescopes to look for signs of radio technology in the form of (among other things) narrowband radio signals of the sort that can only be caused by technology. This is the sort of thing they’re looking for:

This is not the data from Proxima, it is an example.

This plot, from Howard Isaacson’s paper on the topic, shows the actual signal of extraterrestrial technology beaming a radio signal to the Earth. In this case, it’s not aliens: it’s Voyager II.

The vertical axis is time, going up.  Each bin is 10 seconds or so.  The horizontal axis is frequency, and each bin is a few Hz.  Note a few things about this signal:

1. At any given moment, almost all of the power is concentrated into a single frequency bin. This is how we know the signal must be artificial. Radio signals from space come from electrons or atoms or molecules, which always have some temperature.  They also tend to come from large clouds of gas, which have lots of internal motions. Both thermal and bulk  motions generate Doppler shifts that blur out the frequencies they radiate at. Even the narrowest masers, like water or cyclotron maters, must have widths 4 orders of magnitude broader than the signal above.
2. The signal is not perfectly narrow band.  There are two faint “sidelobes” visible on either side (there are bigger ones, too, outside the plot). This is due to signal modulation, illustrating that the signal contains information—it is not a pure “dialtone” or “doorbell”.
3. The signal’s frequency is shifting towards lower frequencies as time increases (upwards). This is how we know the signal is not from Earth: the telescope is on the Earth, which is rotating. As this happens the telescope is first moving towards the source (as the source rises), then moving away (as the source sets).  This creates an ever increasing redshift, making the signal “drift” to lower and lower frequencies during the observation. A source on the surface of the Earth would not be moving with respect to the telescope, and so would show no Doppler shift.
   Note that this shift is the change in the Doppler shift—we can’t calculate the total Doppler shift without knowing the transmission frequency.

The problem is that the spectrum is filled with these sorts of signals. Every now and then one is from an interplanetary probe, but most are from Earth-orbiting satellites and terrestrial sources.  Breakthrough Listen employs sophisticated software that sorts through the millions of signals they can detect and find the ones from space.  The way the team rules out signals from anything other than their celestial object is by nodding the telescope. If the signal is from something on Earth, then they’ll see it no matter which way the telescope is pointing. If it’s from space, it will only appear when they are pointing at the target. For instance, here’s a pernicious false positive from Emilio Enriquez’s paper on the topic:

This is not the data from Proxima, it is an example.

This signal is apparently modulated in a nasty way: it was strongly detected only when they were pointed at the star HIP 65352 (the first, third, and fifth rows) but not when they pointed away (in the second and fourth rows). It also has a slight drift: the signal has shifted by a few Hz by the end of the series of observations.

But, you’ll notice, the signal is also present in the off pointings. It’s very weak then, below their threshold I suspect, which is why the algorithm flagged this as interesting. But if it were really coming from HIP 65352 there’s no way it could be present in those off pointings. This is probably something terrestrial with a poorly stabilized oscillator putting power into the sidelobes of the telescope or something. The exact nature of this signal is not important—all that matters for SETI is that it is not from HIP 65352.

And the radio spectrum is filled with these sorts of false positives! Sifting through them all is hard, and takes a lot of time, and has trained the team to get very good at identifying radio frequency interference. Indeed, no signal has ever survived the four tests I described above after human inspection: narrow band, drifting signal, only present in the on pointings, never present in the off pointings.

Until now!

The Breakthrough Listen project uses the Parkes radio telescope in Australia as one of its tools to search for technosignatures. In this case, they were “piggybacking” on observations of Proxima, the nearest star to Earth, which were looking for radio emissions from stellar flares. These were long stares for many hours per day, for many days. The signals they were looking for are broadband, with complex frequency and temporal structure—basically, if you tuned into it with a radio receiver like the ones we use for FM or AM transmission, it would be present at every frequency, and sound like very complicated static.

But the equipment on the telescope can also be used for SETI, and so the BL team was using the telescope “commensally” to do a SETI experiment simultaneously to the flare study.

And in these data, a signal has apparently survived all of their tests!

Now, this does not mean it’s aliens, as the team has pointed out. It means they have, for the first time, a signal that can’t be easily ruled out as RFI. It’s probably RFI of some pernicious nature, but we don’t know what. Pete Worden of the Breakthrough Listen team says it is “99.9% likely” to be RFI.

We know the signal was present for around three hours, present in 5, 30-minute “on” pointings and not at all in the interspersed “off” pointings. We also know it has a positive drift rate, it appears at 982.002 MHz, and that it appears to be unmodulated.

Other than that, we don’t know much!  But there are some things we can conclude based on this little bit of information.

I cannot speak for the team but I know they’re committed to transparency and scientific rigor. They also think hard about how to convey results to the media, and are careful about things like press releases and peer review of results.

Unfortunately, this news leaked out before the team had finished their analysis, so we’re left to read tea leaves and parse vague newspaper statements instead of reading their paper on the topic (which does not exist because they’re not done with their analysis!)

Someone in the “astronomical community” (we don’t know if they are even a member of the team) leaked the story to the Guardian. Their hand having been forced, the team then gave interviews to Scientific American and NatGeo with some more details, emphasizing that the signal is probably RFI.

Now, I’m pretty grumpy about this. SETI has extensive post-detection protocols that were not followed by the leaker, exactly to avoid this sort of situation. Especially since the team was definitely going to announce this, there’s no need for the leak.

But really what I’m grumpy about is that the team did not get to announce this on their own terms in a way that made clear what was going on. Instead we have lots of speculation and questions that not even the team can answer (because they haven’t finished their analysis yet!)

As Pete Worden tweeted:

Read what we said! No one is claiming it’s A techno signature. We are in the process of following the agreed protocols. At this point we have some interesting signals we believe are interference but as of yet have not been able to track down the source.

— S. Pete Worden (@worden) December 19, 2020

And in the SciAm article:

“The most likely thing is that it’s some human cause,” says Pete Worden, executive director of the Breakthrough Initiatives. “And when I say most likely, it’s like 99.9 [percent].”

It’s unclear how to interpret this.

The fact that it drifts at all is consistent with a non-terrestrial origin. The fact that it drifts more than you’d expect from the motion of Parkes by itself means that the source is either “chirping” its signal to go up in frequency, or that it is not correcting for its own acceleration, and it is accelerating towards the Earth (not directly towards the Earth, like it’s coming for us or something, but just that we’re in the same hemisphere of its sky as the direction it’s accelerating).

Some SETI practitioners expect that a signal would be non-drifting in the frame of the Solar System barycenter, meaning that after we correct for our motion, the signal would have just one frequency. This defies that expectation.

It also can’t be from the rotation of a planet that hosts the transmitter—those shifts would also be negative.  But it could be from the orbital motion of a planet, or from a free-floating transmitter, or from a transmitter on a moon.

The most likely explanation is probably that it is a source on the surface of the earth whose frequency is, for whatever reason, very slowly changing.

Until we know more about the drift, though, there’s not much we could say.

Not completely. If it’s ground-based interference, it’s definitely not coming from that direction. If it’s really from space, it could actually be coming from any place in a 16 arcminute circle around Proxima—about half the width of the full moon.

I’m no expert, but apparently 982.002 MHz is in a relatively unused part of the radio spectrum, where there is not a lot of radio frequency interference. It is in or near what radio astronomers call L band (I guess it’s technically UHF because it’s below 1 GHz), which has long been favored as a place to do SETI because it is in the broad minimum between noise and opacity from the electrons throughout the Galaxy and Earth’s ionosphere on one side, and that from water and other molecules in Earth’s atmosphere, and the cosmic microwave background on the other.

From seti.net. The x-axis is in GHZ, so 982 MHz is just to the left of the 1.

It’s also not far from the “water hole” favored for a long time as the place to look in radio SETI.

Some have pointed out that the signal is suspiciously close to an integer value of MHz, which would argue for a terrestrial origin (since aliens presumably would not use Hz as a standard. The deviation from an exact number of MHz is also consistent with imperfect oscillators in typical radio equipment).

Maybe? Until we know more we really can’t say. The team itself does not appear to have favored this idea (it seems to me to have come from the authors of the newspaper articles) and indeed they have privately communicated to me that they have not analyzed this possibility because they’re focused on the RFI origin right now.

We also don’t know the orbital inclination or rotational properties of Proxima b, so we don’t know what acceleration signal it would provide. Without a good model for that planet and without knowing what the team has seen, we can only speculate.

That said, if the signal repeats and turns out to be from Proxima, and if the signal is not being inherently modulated, then we could use the drifts to infer the accelerations of the transmitted, and possibly determine whether it’s on the surface of a planet, and determine the rotational and orbital period of that planet.

The original Guardian article had a misguided take on this one:

“The chances against this being an artificial signal from Proxima Centauri seem staggering,” said Lewis Dartnell, an astrobiologist and professor of science communication at the University of Westminster. “We’ve been looking for alien life for so long now and the idea that it could turn out to be on our front doorstep, in the very next star system, is piling improbabilities upon improbabilities.

“If there is intelligent life there, it would almost certainly have spread much more widely across the galaxy. The chances of the only two civilisations in the entire galaxy happening to be neighbours, among 400bn stars, absolutely stretches the bounds of rationality.”

This is wrong, because it’s based on a lot of unexamined priors and assumptions.

First, it assumes that signals of this sort must be very rare coming from only a handful of stars in the Galaxy. While that is certainly very plausible, the idea that nearly every star might have some sort of technology around it is older than SETI itself! Indeed, it is at the heart of the Fermi Paradox, which asks: since interstellar spaceflight is possible with ordinary rockets, and since the Galaxy can be populated by such rockets in less time than it’s been around, why aren’t aliens here in the Solar System right now?

One answer is “they don’t exist.”  Another is “they don’t spread around very much”. Another is “they are most places, but avoid the Solar System for some reason, perhaps because life is present here.” Another is “they have been here in the Solar System but aren’t here now.”  Another is “there is alien technology in the Solar System but we haven’t noticed it”.

Dartnell’s “improbabilities upon improbabilities” presumes that the second answer above is correct, but there is plenty of heritage in the SETI literature that explores the other answers, as well.

But even if it’s true that interstellar travel of creatures is rare and Parnell is right that it’s therefore unlikely that Proxima is inhabited, there is still a good argument to be made that Proxima is the most likely star to send us signals—perhaps even the only such star!

If there exists a Galactic community, either a diaspora or a lot of stars with technological life, or even just a single planet with life that has sent its technology everywhere, then it might set up a communication network. This is, after all, what SETI hopes to find.

But when you want to communicate with many places over very large distances, point-to-point communication is a poor way to go about it. When you call your friend on your mobile phone, your phones aren’t sending radio signals to each other. That would require way too much power and complexity. Instead, your phone sends its signal to the nearest cell tower. This makes the power requirements of your phone (and the tower) much more reasonable. This tower then sends the signal, via many means, on a complex route through many central nodes until it arrives at your friend’s nearest cell tower, and they get the signal that way.

By this logic, Proxima is the most likely place for the “last mile” portion of any message to the Solar System. Indeed, it may be the only star transmitting to us!

And note that this scheme does not assume that the message is meant for us—the Solar System may just be one stop in a network.

But if they were trying to get our attention, then they need to do something we would find obvious to look for, which means they’d have to guess which stars we’d guess to search for their signal. There are a lot of stars to choose from—which is the most obvious place for us to look?  It’s hard to argue for a better target than Proxima.

Now, this could all be wrong, but the point is we don’t know what sort of luminosity function or spatial distribution transmitters might have, and it’s easy to construct plausible scenarios where Proxima or some other very nearby star is the first one we’d detect.

I don’t really know. I’m not an expert in RFI, and even if I were, I haven’t seen the data.

Jonathan McDowell and I have had some fun on Twitter exploring an interesting possibility:

Anything in Molniya orbit is going to spend time at -63 deg moving slowly at apogee, but not sure if any are L-band. Russian GLONASS nav sats are L-band and in 65 deg orbits but 12hr circular, so wouldn't stay in the beam as long.

— Jonathan McDowell (@planet4589) December 19, 2020

There’s a special kind of orbit that takes satellites way out to +/- 63 degrees declination and sort of hang there at apogee for a while in a long elliptical orbit.  Such satellites would also have a positive draft rate, since they’re accelerating towards the Earth. Jonathan, who keeps careful track of everything artificial in space (literally) has been trying to see if any actual satellite might do this in the direction of Proxima, but he didn’t find any in his database.

Mainly, we wait for the team to finish their work and present their results.

Things that I imagine the team are and will be doing include:

• Pointing Parkes at Proxima a lot to see if the signal repeats! Unfortunately, there are not a lot of facilities in the Southern Hemisphere that can do this work. MeerKAT may be up to the task soon, but is hard to get time on. Depending on the strength of the signal it may be possible to point smaller telescopes at Proxima to search for it as well.
• Scouring all of their data for other examples of this signal. If it’s RFI, there’s a good chance they’ve seen it before when not pointing at Proxima
• Searching carefully for other signals from Proxima. If there is one signal, there may be many more.
• Considering lots of sources of RFI—what devices transmit at 982 MHz? Could any satellite or train of satellites stay in the Parkes beam for 3 hours? Could it be a hoax?

If it never repeats and if the team can’t find a good RFI explanation then I’m afraid it will be another Wow! Signal; an intriguing “Maybe?” that we’ll just have to wonder about forever. We can’t study it if it’s so ephemeral that we never get a good look at it again!

But mostly, we talk about how cool SETI is and we wait!