Could LIGO discovery be due to e.g. earthquakes or have a terrestrial source? [duplicate]

This question already has an answer here:

I mean, could it have been earthquakes or anything else?

marked as duplicate by Kyle Kanos, user36790, Bill N, ACuriousMind♦, GertFeb 13 '16 at 20:26

• Coincidence of signals for one... – Kyle Kanos Feb 12 '16 at 13:06
• Hi Sidarth. Just a simple answer for you: the two stations picked up exactly the same signal at once. You see? They are very far apart. It could be that incredibly there was a "passing freight train" or something at the exact same instant, speed, etc etc in both areas - but it's extremely unlikely. At heart, the issue is that simple. – Fattie Feb 12 '16 at 13:42
• I deleted my earlier comment and I posted an answer below explaining why it could not have been terrestrial perturbations. – user106422 Feb 12 '16 at 23:54

I want to add one very compelling argument which clearly shows that Ligo could not have been because of earthquakes or terrestrial phenomena, or at least their probabilities would be outrageously low.

What was announced yesterday was that two different detectors picked up almost entirely identical signals with a spacing of a few milliseconds difference. The distance between the two detectors in Livingston, Lousiana and Hanford, Washington as the crow flies on the surface of the Earth is about 3042 km.

But gravitational waves would not care about the ATCF distance along the surface of the earth but the actual geodesic distance between the two points which is much smaller (about 3000km).

For an earthquake to travel about 3042km, the minimum time to do so assuming a uniform terrestrial medium between the two detector sites is about 3000km/ (8km/s) (8km/s is a respectable upper bound on the speed of a seismic wave, according to Wikipedia.) which gives us about 375 to 380, assuming that the earthquake is so powerful that it can maintain the same energy density at both the locations. This is already highly unlikely that the seismic wave does not lose energy but we now have definitive proof that the same seismic wave is not reponsible for the events measured on the gravitational detectors 3042km apart.

Finally, what if there was more than 1 such event? The USGS and other earthquake related databases indicate that the strongest possible earthquake which can most probably affect the Hanford detector site is from the faultlines from California which have approximately a 1/50 chance that the next earthquake will be a magnitude 7 earthquake in San Fransisco which is still a good ~1100km from Hanford. The closest possible predictive earthquake chance near Lousiana is a 1/300 chance of a magnitede 8 earthquake in Missouri which is again about 1000km away from Livingston in Lousiana. Now, assuming that one would need really strong earthquakes to contribute to the data (which as @anna mentions, is nearly impossible because the mirrors and the beam sources were suspended from highly sophisticated suspensions), the probability of two earthquakes happening at two different locations simultaneously such that the same energy is felt at two different locations which are over 3000km apart, thereby leading to the same waveform being measured on detectors is almost 0. I don't need to be a geologist to claim that the probably of this happening is a bit more outrageous than the probability of a celestial event which is capable of producing detectable gravitational waves.

Finally, not to mention - If I were doing an experiment which might involve millions/billions of dollars worth of effort, time and equipment over a period of 40 years which in principle could produce false positive data from earth quakes, I would also allow for data analysis and tools which removes any seismic effects from the final data in addition to the intricate suspension technologies. And if I can be wise enough to contemplate this, I'm pretty sure highly experienced scientists and analysts are already so.

• that would have had to been a really REALLY powerful earthquake which all the media is covering right now. also, seismic waves lose energy and there are technologies implemented to cancel these effects in LIGO. – user106422 Feb 13 '16 at 9:32
• You have speculated extensively about earthquake possibility, so I assumed this option is not ruled out "by definition" especially that the signal recorded was on the order of $10^-21$ – bright magus Feb 13 '16 at 9:38
• It isn't really speculation. I read off the data provided by geological surverys for possibility of strong earthquakes on the day that the signal was detected. The key idea here is not to focus on these kind of "what if" scenarios. I mentioned the sophistication of the suspension devices which effectively screen out seismic waves. EVEN IF there were seismic contributions, the probability that they/it would produce identical results in 2 different experiments exact to the order of $10^{-21}$ is practically 0. And I don't think nature gets that lucky to a significance of 5.1$\sigma$. – user106422 Feb 13 '16 at 11:52
• "The key idea here is not to focus on these kind of "what if" scenarios." Well, I always assumed this is exactly the key idea in science. And as to the nature getting lucky - the nature was "lucky" enough to produce homo sapiens capable of discovering its ways. – bright magus Feb 13 '16 at 12:02
• This is getting too philosophical for me. You're addressing this a priori, in which case yes it is the essence of science. I'm talking from a a posteriori perspective. I'm not looking to answer the question "How can we eliminate terrestrial effects from data?" but rather "How do we know that the data that we have following rigourous analysis could not have been caused by an earthquake?" As for the anthropic argument, I think we should continue it elsewhere because it deviates from the essence of this thread. – user106422 Feb 13 '16 at 12:08

In very broad language, we don't know (and hear me out before you judge me)! But then what is science? Science is the process of producing models that get us to understand the universe better and make predictions about it. We have a model of gravitational waves that was produced using general relativity. This model predicts a specific signal that we would detect if this model is accurate. The signal is shown in the paper of gravitational waves published yesterday. It compares our model with the signal detected:

(Image from: Observation of Gravitational Waves from a Binary Black Hole Merger by B. P. Abbott et al. (LIGO Scientific Collaboration and Virgo Collaboration) in Phys. Rev. Lett. 116, 061102, doi:10.1103/PhysRevLett.116.061102)

Then after we see this amount of matching, we do the statistical math and calculate, what is the probability of this happening by coincidence in two stations? The probability is measured by how many "sigmas" we're far from our model. Then we make a publication like the one linked above, and we say: We made an observation that is consistent with gravity waves. Then other experiments in the future repeat the measurement again, and again, and again, and every other experiment confirms what we had. If only LIGO would measure gravitational waves, and bigger binary black holes merge in the future and we see nothing, then we start doubting what happened and question whether what we measured is gravitational waves. More experiments reveal more evidence and solid proof. This is how science works.

The big deal about this is that LIGO and Virgo were the first ever to detect such solid evidence like the signals you saw in the pictures. So we're quite certain this is gravitational waves.

• How do we know that "bigger binary black holes merge in the future" if "we see nothing"? – bright magus Feb 12 '16 at 19:12
• @brightmagus are gravitational waves the only way to recognize black holes? – The Quantum Physicist Feb 12 '16 at 19:25
• How many merging black holes have been observed so far? – bright magus Feb 12 '16 at 19:27
• @brightmagus Why are you playing this game with me? Is my explanation wrong? If it's, then say it's; if it's not, then ask your question politely without applying your own mind models as if they're absolutely true. I don't have time for games! – The Quantum Physicist Feb 12 '16 at 19:32
• What are you talking about? I'm asking politely how do you expect to verify if the interferometer correctly detects gravitational waves created by the merging of black holes. If you don't know or do not wish to answer then that's OK with me. – bright magus Feb 12 '16 at 20:03

I will address this main question:

Could LIGO discovery be due to e.g. earthquakes or have a terrestrial source?

The short answer is , NO. The reason the observation happened in September and the rumors rose just a month ago is because the researchers themselves were double checking all the numbers.

Anybody who watches the presentation given for the press can see in a simplified manner that vibrations from trucks ( :) ), earthquakes etc are isolated by the suspension of the detectors as pendulums, to dampen any high frequency changes. The characteristic signal takes miliseconds, one can barely hear it in the demonstration. This discussion in Motl's blog will help.

Now of course a model is used to identify the signal with two black holes merging, as the other answer says. No competing physical model has been proposed so , if it walks like a duck and it quacks like a duck, why, it IS a duck.

• " ... if it walks like a duck and it quacks like a duck, why, it IS a duck." Well, perhaps ... "She is a witch! She looks like one. Burn her! – bright magus Feb 12 '16 at 22:12
• @brightmagus there is no burning in making the identification of a duck. did I talk about shooting ? you are distorting the classical analogy – anna v Feb 13 '16 at 4:03
• You'd need to have a consulting seismologist with deep mantle specialism on the LIGO team. It's at least possible a deep mantle earthquake could register identically at the two locations. Was there a seismologist on the team? – Lucy Meadow Mar 22 '16 at 10:48
• @LucyMeadow The suspension takes care of this . did you watch the video? – anna v Mar 22 '16 at 11:23

In addition to the answers already given, let me add one thought: These people aren't just doing some out of the box experiment and they aren't amateurs.

The experiments have been going on for decades and the detectors used in the experiment were build in 1999 and have been measuring in 2002. They are constantly refined using the latest technology which is also probed and used in other experiments of a similar type (LIGO explicitly mentions GEO 600 an their report). Hundreds of people are and were involved during that time and because of independent experiments, you will have people thinking about similar matters independently (meaning there is redundancy).

Therefore, you can be sure that the people know their experiments. When the experiments were started, you can be sure that a lot of signals were recorded (and still are), each with its own characteristics. Then the experimenters started to learn what these signals mean. Often, they will find natural causes: maybe there were earthquakes at the time or maybe the signal pops up at the exact same time a train is going by close-by. Then they find out more and more and make sure that this is really the cause and they try to eliminate these kind of signals by making the design even better. It is not unusual that such signals cannot be explained for several months and eventually, a natural cause is found.

To sum it up: The experiments are extremely carefully designed and the experimenters go to great length to understand every single detail of their experiments. Of course this doesn't eliminate the chance that the signal is just noise as others have explained, but it is highly unlikely.

• I thought the detection that was made had only been possible for 2 days? – Lucy Meadow Mar 22 '16 at 10:50
• Yes, but the advanced LIGO doesn't come out of nowhere. It's an enhanced system of what has been running for years. All parts of the new experiment have also been tested individually, etc. – Martin Mar 22 '16 at 12:05