Is it possible that the LIGO results are anomalies?

Question

Since it was reported recently that the supermassive black holes at the centres of galaxies may in fact be as many as 20,000 smaller black holes I wondered if, if these black holes collide with one another, why they wouldn't be producing gravitational waves all the time. But then I also wondered that, if whilst accepting material, supermassive black holes must drastically increase the density of that material before it collides with the SMBH why this wouldn't also cause gravitational waves. Since GW's can (presumably) be caused by anything, since colliding neutron stars also emit these waves, why wouldn't any other source emit them also?

Also, why could they not be caused by seismic activity? Even if there's a minute probability of such events occurring surely it's possible that, given all the potential sources of a matching signal, that it would happen eventually? What are people's thoughts on this matter?

Answer 1

I want to address your main question, rather than the various subsidiary ones. You asked if it's possible that LIGO's results are anomalies.

No. There's no way that LIGO's results are anomalies.

The most vivid proof is the detection by LIGO of GW170817/GRB170817A, a binary neutron star merger that emitted a gamma ray burst. (Paper here.)

Binary neutron star mergers are incredibly violent explosions. When they occur, the merging systems emit both gravitational and electromagnetic radiation and eject a cloud of matter. The electromagnetic radiation comes in the form of a gamma ray burst. There is no brighter source of electromagnetic radiation in the cosmos.

As gravitational and electromagnetic radiation emerge from the explosion, they interact with the ejected matter. Electromagnetic radiation travels slower in media than in vacuum, in proportion to the amount of matter interacted with and in proportion to the strength of the interaction. This is also true of gravitational radiation, but the gravitational interaction is so much weaker that it can be neglected. Consequently, the gravitational waves escaped the explosion before the electromagnetic ones, and arrived here on Earth first.

On 17 Aug 2017, LIGO detected a large gravitational wave, and then, 2 seconds later, the Fermi gamma-ray telescope detected a gamma ray burst. Two seconds may not seem like much, but it's an eternity for these kinds of signals. The gravitational wave had already passed by the Earth when the gamma rays arrived.

With help from its sister instrument Virgo (which had a known blind spot where the explosion occurred), LIGO was able to triangulate the gravitational wave source well enough to confirm that they were receiving radiation from the same source.

That, to my mind, is a smoking gun. LIGO saw something in the sky, and when other detectors were pointed there, they saw it too.

Answer 2

From a philosophical point of view, yes, of course the LIGO results could be an anomaly. So could every scientific measurement ever made. Science admits the possibility that its conclusions are incorrect.

If you take a look at the LIGO papers, you’ll find that they present an error analysis to estimate the possibility that the signals were mere artifacts. For example, in the paper announcing the first measured black hole merger, the estimate is that a false alarm on the level of the signal they measured would occur less than once every 203000 years. So from that paper alone, the odds are pretty slim that the results are anomalous. Add in the results from the following two years, including the neutron star merger mentioned by @user1504 (false alarm rate <1 per 80000 years), and the chance of anomaly becomes even smaller. But it’s possible, and LIGO rightly admits as much through their use of language: “we infer the component masses...”, the $\gamma$-ray burst “corroborates the hypothesis”, subsequent observation “supports the interpretation of this event”, etc.

Judge for yourself!

Answer 3

Firstly you have misunderstood the recent article about the black holes at the centre of the galaxy. The finding is that there may be tens of thousands of stellar mass black holes orbiting the central supermassive black hole. The supermassive black hole typically has a mass of several million stellar masses, so the ten thousand or so smaller black holes orbiting it make very little difference.

You suggest that the ten thousand black holes in the centre of the galaxy should collide frequently, but direct black hole collisions are so improbably they essentially never happen. That's because black holes are tiny compared to the distances between them. The collisions happen when two black holes are orbiting each other and that orbit gradually decays due to the emission of gravitational waves. This orbital decay is very, very slow. Take for example the Hulse-Taylor binary. We can measure the orbital decay to high precision, and it's going to take about $300$ million years before the two stars eventually collide and merge. So ten thousand such systems would give a merger roughly every $30$ thousand years - don't hold your breath!

But back to the main point of your question. The merger of two black holes is now a well understood problem and the gravitational radiation emitted during the process can be calculated to a high accuracy. The shape of the GW pulse emitted depends on the masses of the two black holes, along with various other properties like their spins, so the masses can be calculated from the gravitational waves received by LIGO. The merger of a stellar mass black hole with the supermassive black hole would give a very different signal to the merger of two stellar mass black holes so it would be easily distinguished.

Finally, background noise is a very important consideration in LIGO and the team have gone to great lengths to eliminate it. That's why the observatories are a long way from each other. Seismic events are easily distinguished because they propagate at the speed of sound (in rock) while gravitational waves propagate at the speed of light. By comparing the times the signals are received at the three detectors sesimic events can be eliminated.