Hacker News Reader

Comment by pdonis | Hacker News Reader

One factor that this article does not take into account is the signal correlation between detectors. LIGO only counts signals as a detection of an event if the same signal appears in multiple detectors with the appropriate time delays (indicating light speed travel time in a particular direction through the Earth). The same signal in multiple detectors rules out detector errors, and the relative timing rules out terrestrial origin: no signal of terrestrial origin will travel at anywhere close to the speed of light through the Earth (earthquake waves travel orders of magnitude slower).

View on HN

> the relative timing rules out terrestrial origin: no signal of terrestrial origin will travel at anywhere close to the speed of light through the Earth

Well, there is still the chance that similar terrestrial events might happen at the same time near each detector. Especially while we only have 2 or 3 active LIGO detectors.

I don't know how closely "fingerprinted" each event is to rule out such coincidences. It might be purely theoretical.

Then again there is still the theoretical chance that an event will occur at an equidistant point from the LIGO stations and produce a signal of the same signature at all places at the same time.

Probably absurdly unlikely, even as a conscious prankster attack.

(Sorry for being so difficult...)

> I don't know how closely "fingerprinted" each event is to rule out such coincidences.

Such a coincidence would have to produce the same specific signal in each detector. That's highly unlikely.

> there is still the theoretical chance that an event will occur at an equidistant point from the LIGO stations and produce a signal of the same signature at all places at the same time.

You're misunderstanding the relative time criterion. It is not that the signals are seen at the same time in all the detectors. It is that the signals are seen at just the right different times at the detectors that would be expected from an outside signal that travels at the speed of light through the Earth, hitting the different detectors at different times as it passes through. An event that really is seen at the same time in all the detectors would obviously be of terrestrial origin and would not be reported as a detection of an outside signal. Even events that are at different times, but the wrong different times for a light speed signal coming from outside, would not be reported as a detection.

Moreover, it is through these time delays, and knowledge of the angular "antenna pattern" of the detectors, that it is possible to localize the sky position of a gravitational wave source.

> no signal of terrestrial origin will travel at anywhere close to the speed of light through the Earth (earthquake waves travel orders of magnitude slower)

To be fair, electromagnetic (i.e. radio) waves do travel at the speed of light through the Earth, though almost none can significantly penetrate the Earth itself, except extremely low frequencies. Electromagnetic interference should be quite easy to measure and look for correlations though; if scientists did their homework (likely), it has been ruled out.

> electromagnetic (i.e. radio) waves do travel at the speed of light through the Earth, though almost none can significantly penetrate the Earth itself, except extremely low frequencies

Yes, but they won't jiggle the test masses in LIGO, so they won't produce a signal.

LIGO does not record jiggling test masses. It records electrical signals that can be generated by jiggling test masses, and also by other means.

Every false alarm should be investigated to see what produces such frequent apparent events. They need to distinguish better real events from ersatz events, independently of simultaneous reports elsewhere or optical confirmations.

> LIGO does not record jiggling test masses. It records electrical signals that can be generated by jiggling test masses, and also by other means.

No, it records interference patterns at a detector which are produced by jiggling of the test masses. There are no other means of producing those interference patterns--certainly not by low frequency EM radiation happening to pass through the apparatus, which was what the post I was responding to was talking about.

The question is what makes the test masses jiggle. A gravitational wave is one possibility, but certainly not the only one. Seismic disturbances, a truck driving by the facility, someone banging on the wall of the vacuum chamber, etc.

The electrical signals you are talking about come from the interference detector. It is possible that there could be electrical noise elsewhere in the circuit, but that's something the experimenters already allow for and take great precautions to minimize, just as the test masses are isolated to minimize any source of jiggling of them other than gravitational waves.

> They need to distinguish better real events from ersatz events

What do you think they could do better?

They need to discover better ways to reject apparent events without relying on the other detectors. The frequency of apparent real events needs to approach that of actual events, for corroboration from other detectors to mean anything.

> They need to discover better ways to reject apparent events without relying on the other detectors.

They already have plenty of these. Comparison of signal arrival times at multiple detectors is not even close to being the only criterion they use. It's just an important one that is not mentioned in Hossenfelder's article.

Is the index of refraction of the Earth at low radio frequencies really 1? I'd expect it to be higher, and so there to be some slowdown for radio waves.

Depends on how low the frequency is :)

But indeed I'm not an expert by any means on magma properties, perhaps someone else can chime in. If magma is conductive then even less radiation goes through.

Keep in mind at extremely low frequencies the whole Earth is in mid-field or near-field of emissions, so it doesn't behave attenuated like light rays.

Magma is, for practical purposes, infinitely conductive. Only atmospheric EM events are plausible causes of spurious correlated events.

Even an infinitely conductive sphere does not screen the quasi-static near field very well. Although the delay is probably going to be weird, greater than direct visibility. Atmospheric EM events indeed are what comes to mind.

> The same signal in multiple detectors rules out detector errors

Only if the cross correlation frequency is higher than the autocorrelation frequency. In other words, how often do signals with signature X appear, in each detector? If it's frequent then the likelihood of a by chance cross correlation is higher.

Figure 4 in the original detection paper shows a histogram of this false alarm rate (SNR versus detection frequency) and the detection as a significant outlier:

https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.11...

Then why were 9 events for which alerts were sent later “downgraded because they were identified as likely of terrestrial origin”?

It is always possible for a random signal to appear by coincidence in several detectors simultaneously. LIGO also gathers data from a huge number of "auxiliary channels" that monitor the local environment (seismometers, microphones, radio receivers, etc). Possible detections are heavily scrutinized by ensuring that the detectors were operating normally during the time of detection, and no local disturbances are seen in the auxiliary channels.

Because looking at the signal correlation is not the only test that is applied. But it is a significant test that was not discussed in Hossenfelder's article.

Because analyzing them takes time, and they want to inform telescopes as soon as possible so optical/EM data can be captured while the event (possibly) occurs.

A false positive is a lot better than a false negative here.

Reply on news.ycombinator.com