Unresolved background of gravitational waves?

[bcrowell]

The aLIGO signal is a discrete event. In addition to such events that can be resolved into individual chirps, I would think that there would be a background of gravitational waves that would look like noise, but that could be distinguished from terrestrial noise (trucks passing by, etc.) because it's correlated between the Hanford and Livingston detectors. To some extent this would be analogous to the CMB, but for gravity rather than electromagnetism. I don't know if it would be primordial in origin, or whether it would be thermalized (I would guess not). But it seems like it would be of interest to measure its power spectrum, angular distribution, and polarization, just as we do with the CMB. Is this actually possible with aLIGO? It seems like it should be possible in principle, although conceivably when one attempted to measure it, it would be statistically consistent with zero. But I don't see how its strength could be estimated a priori, since we know so little about possible sources, including primordial ones. Since matter is transparent to gravitational waves, it would seem like this noise would probe an earlier era than the CMB does.

 

[Dr. Courtney]

That's a good point, and I think you may be right. Correlation between the two detectors will be key. It is unclear how much averaging would be needed to distinguish a background signal from the sources of noise. One may also need to consider common sources of noise between the detectors. At least in the frequency domain, I think both detectors have very similar noise, so a phase technique may be needed.

 

[PAllen]

I also think you are right. Hopefully, ongoing LIGO papers will discuss this.

 

[Vanadium 50]

The problem with this is that the signal travel time is unknown, so you are looking for correlations anywhere within a 10 ms window. You will get a LOT of false positives. This looks much more feasible if there is a network of at least 4 or 5 detectors.

 

[Dr. Courtney]

The problem with this is that the signal travel time is unknown, so you are looking for correlations anywhere within a 10 ms window. You will get a LOT of false positives. This looks much more feasible if there is a network of at least 4 or 5 detectors.

My understanding is that a given delay corresponds to a given region of space (or more precisely a given range of originating directions).

If we are truly talking about signals that are continuous or long-term background signals (rather than short discrete events), it seems like the approach of looking for the signal in data from different time windows would be effective for double checking false positives.

 

[Vanadium 50]

Dr. Courtney, if your argument is that one can generate a model background with almost infinite statistics by looking at out of time windows (for a 16 day run, there are about 100 billion 10 ms pairs) I agree. The problem with this technique, at least every time I have applied it (not in gravity waves) is that "the background" turns out not to be a completely uniform thing and if you use this technique to measure two background (say at different times), they also disagree significantly. So the background you are interested in is not measured nearly as well as the statistics would suggest.

 

[mfb]

I'm not sure if that can work reasonably well with just two detectors. Let's consider 100 Hz waves, roughly at the peak sensitivity. If we look at a direction of the sky where Hanford is 10 ms ahead of Livingston, we get a nice positive correlation if we shift the signal by 10 ms. If we look at a direction of the sky where Hanford is 5 ms ahead of Livingston, the same shift will lead to a negative correlation. Not as strong as the positive one because the source is not very coherent, but still a negative correlation. Well, we cannot choose where we look at, both signals will be present at the same time. You get some messy sum of all the correlations coming from all directions in the sky. I don't think you can properly get angular distributions, even in the complete absence of non-gravitational-wave noise. With 4-5 detectors: sure.

 

[Dr. Courtney]

Trying to ferret out additional information by additional data analysis of existing streams is infinitely cheaper than building 2-3 new detectors, so it is probably worth some concerted efforts.

 

[mfb]

Well, the additional detectors are built anyway, to locate isolated sources precise enough for follow-up observations of telescopes, to search for even weaker signals and to increase the overall uptime of the set of detectors.
VIRGO in Italy this year, KAGRA in Japan probably in 2018-2019, maybe one more in India.

 

[mrspeedybob]

Are there any estimates about how much energy there should be in the form of ambient gravity waves? I'm thinking if every mass that accelerates radiates a bit of gravity wave energy, the total could be large.

I'm thinking about random thermal movement of individual atoms and molecules. Probably no way to detect this radiation, but it could perhaps be observed as coiling that is not accounted for via EM radiation. I know gravity is many orders of magnitude weaker then EM, does that mean that thermal gravity radiation would be proportionally less, or could it perhaps be not quite that much less because EM has canceling charges, and gravity does not?

 

[mfb]

The cosmic energy inventory

The overall contribution is small, and most of it comes from massive objects moving at high speeds.

Thermal motion is completely irrelevant, because the motion of different particles averages out over the relevant timescale. As an example, a particle in a liquid or solid might change its direction every picosecond. That corresponds to 300 micrometers, where you have something like 10¹⁸ atoms to average over. Compared to electromagnetic thermal radiation (which escapes from the surface only), that gives additional 9 orders of magnitude reduction.

 

[Vanadium 50]

Are there any estimates about how much energy there should be in the form of ambient gravity waves? I'm thinking if every mass that accelerates radiates a bit of gravity wave energy, the total could be large.

It's not.

One reason is that gravity is very, very weak. One way to think of it is 40 orders of magnitude smaller than electromagnetism.

The other reason is that it is quadrupole in nature. That means you need a changing quadrupole to get it (skinny-fat-skinny-fat) instead of a dipole (up-down-up-down). That both reduces the ability of an arbitrary configuration to generate GR and the energy carried by GR.

 

[atyy]

To some extent this would be analogous to the CMB, but for gravity rather than electromagnetism. I don't know if it would be primordial in origin, or whether it would be thermalized (I would guess not). But it seems like it would be of interest to measure its power spectrum, angular distribution, and polarization, just as we do with the CMB.

Planck and eLISA (which I think is like LIGO) hope to find evidence for primordial gravitational waves.

Planck
http://www.esa.int/Our_Activities/Space_Science/Planck/Science_objectives

eLISA
https://en.wikipedia.org/wiki/Evolved_Laser_Interferometer_Space_Antenna
https://www.elisascience.org/articles/science-context-2028/dark-energy-and-cosmic-microwave-background

According to the Wikipedia article on eLISA: "Other gravitational wave antennas, such as LIGO, VIRGO, and GEO 600, are already in operation on Earth, but their sensitivity at low frequencies is limited by the largest practical arm lengths, by seismic noise, and by interference from nearby moving masses."

 

[bcrowell]

One of atyy's links referenced this paper:

Cosmological Backgrounds of Gravitational Waves and eLISA/NGO: Phase Transitions, Cosmic Strings and Other Sources
Pierre Binétruy, Alejandro Bohé, Chiara Caprini, Jean-François Dufaux
http://arxiv.org/abs/1201.0983

 

[FactChecker]

The aLIGO signal is a discrete event. In addition to such events that can be resolved into individual chirps, I would think that there would be a background of gravitational waves that would look like noise, but that could be distinguished from terrestrial noise (trucks passing by, etc.) because it's correlated between the Hanford and Livingston detectors.

I think they are detecting "terrestrial noise" by detecting inertial acceleration, whereas the gravitational waves will have none.

From http://journals.aps.org/prl/abstract/10.1103/PhysRevLett.116.061102 ,
"To monitor environmental disturbances and their influence on the detectors, each observatory site is equipped with an array of sensors: seismometers, accelerometers, microphones, magnetometers, radio receivers, weather sensors, ac-power line monitors, and a cosmic-ray detector [65]."

 

[mfb]

They even actively cancel some terrestrial noise, but the cancellation is not perfect. The remaining uncertainty on this noise is still a relevant limit for low frequencies.

 

[Jeff Rosenbury]

What about directionality? For EM waves we can build a very directional horn antenna or whatnot. Trying to tell the noise from the backgound radiation by phase delay between sites when the noise is (notionally) coming from all directions seems difficult to me.

But perhaps I'm missing some math trick? Or is there some way to shield against gravitational waves?

 

[mfb]

There is no way to shield gravitational waves. The relative amplitudes and phase delays between different detectors is all we have. For individual signals like binary black hole mergers this works nicely, for continuous sources like rotating neutron stars it is more complicated (3+ detectors will help).

 

[bcrowell]

What about directionality? For EM waves we can build a very directional horn antenna or whatnot. Trying to tell the noise from the backgound radiation by phase delay between sites when the noise is (notionally) coming from all directions seems difficult to me.

I think this is what mfb was discussing in #7 and #9.

 

[ProfChuck]

There are several important questions that need to be asked.
Considering that general relativity predicts that G waves propagate at light speed a confirmation of this would be yet another bolster for the model. Can LIGO perform this test or is it circular reasoning? In other words how much of LIGO's operation is based on the assumption that Vg=c?
It is assumed that matter is transparent to G waves. However, the operation of LIGO is based on the apparent relative motion of test masses. Is this motion "real" or is the observable consequence of Space-Time distortion by the passing wave? Or is that even a legitimate question?
Do G waves constructively interfere? If so it may be possible, at least in principal, to construct a device that can generate G waves under controlled conditions that are powerful enough to be detected by LIGO. This would open the door to laboratory experimentation.

 

[mfb]

A gravitational wave has four degrees of freedom relevant to speed measurements: polarization, amplitude and two for the direction. And speed as fifth parameter, if we want to measure this. With two detectors, LIGO can measure two amplitudes and one time delay, which gives some sensitivity to the speed and a strict upper limit, but the speed cannot be measured accurately. Together with VIRGO, we will get three amplitude measurements and two time delays. That allows to measure all five parameters.

Is this motion "real" or is the observable consequence of Space-Time distortion by the passing wave?

What does real mean?

Do G waves constructively interfere?

Depends on their phase.

If so it may be possible, at least in principal, to construct a device that can generate G waves under controlled conditions that are powerful enough to be detected by LIGO.

If you convert the whole Earth to a source of gravitational waves, maybe. But without significantly improved detectors (beyond all planned detectors), a detection of self-made gravitational waves is unrealistic.

 

[Billmyk]

just curious what this gravitational wave discovery does for the graviton?

 

[bcrowell]

just curious what this gravitational wave discovery does for the graviton?

That doesn't relate to the topic of this thread. You might want to look at this thread: https://www.physicsforums.com/threads/gravity-wave-discovery-and-gravitons.856907/