[EDITED to accommodate info from the comments]

Among the local atmospheric electromagnetic potential sources of a signal capable of mimicking the waveform of a GW not sufficiently considered by LIGO and are sferics and radio bursts in relation with lightning and with terrestrial gamma-ray flashes associated to thunderstorms. As discussed in the comments in the presence of some charge in the interferometer it is plausible(see calculations in the comments below) for a radio wave in the 35-70 Hz range of frequency to produce a few oscillations and with a magnetic field just under what woud be detected by the magnetometers. Once this oscillatory force produce displacement on the mirrors it is also a possibility considered in the GW optomechanics literature, by the optical spring property that the detuned signal recycled cavity that forms between the signal recycling mirror and the other mirrors, that the laser radiation pressure anti-damping effect leads to a few more cycles with freq going up towards the optical cavity resonance before it is quickly damped. Precisely this signal recycling-optical spring property is exploited by the AdvancedLigo interferometer to get close to or even beat the standard quantum limit(SQL) of detectibility and enhance the sensitivity in the presence of an actual GW in these low frequencies(under 300 Hz is where amplitude sensitivity is maximized) associated to binaries spiraling and merging, but it is also known to be capable of producing parametric instabilities.(see for instance chapters 3, 11 and 12 of "Advanced interferometers and the search for gravitational waves" edited by Bassan or consult the arxiv for papers by Meers, Chen and Buonnano)

So given all this it is hard not to wonder why LIGO and other scientific groups independent of LIGO haven't apparently considered this kind of potential electromagnetic radiation sources of confusion, if only to critically scrutinize the interpretation of a single event put forth by a single team like it should always be done in science.

ADDED june 18th 2017: After more than a year and over 2000 citations finally one team makes the actual effort of going critically through the published data from LIGO and publish their conclusions: https://arxiv.org/abs/1706.04191 . Basically they find that when doing the analysis template-free it can be seen that the noise during the GW "signal"(with the analysis done either with the signal substracted or not given the weakness of the putative GW amplitude) is also lagged 6.9 ms between detectors, when it should be stochastic. And that "noise" could perfectly accomodate a signal of lightning related events of terrestrial origin(that could go perfectly unnoticed by the magnetometer as shown in the comments below), but certainly as a whole not the shape of a GW signal.


closed as off-topic by user10851, Kyle Kanos, CuriousOne, Gert, yuggib Mar 18 '16 at 14:20

This question appears to be off-topic. The users who voted to close gave this specific reason:

  • "We deal with mainstream physics here. Questions about the general correctness of unpublished personal theories are off topic, although specific questions evaluating new theories in the context of established science are usually allowed. For more information, see Is non mainstream physics appropriate for this site?." – Kyle Kanos, CuriousOne, Gert, yuggib
If this question can be reworded to fit the rules in the help center, please edit the question.

  • 2
    $\begingroup$ Fermi doesn't actually monitor TGF per se, these events simply trigger the GBM on-board. That said, I'd need to see a paper showing that a TGF (a) could produce a similar waveform and (b) appeared that runs along LA & WA (or vice versa) where the two LIGOs are located. $\endgroup$ – Kyle Kanos Mar 15 '16 at 11:44
  • $\begingroup$ How does a flash of light impinge on a mirror inside a concrete tunnel or differentially change the path length? Also how does your idea produce nearly coincident signals in detectors separated by 1000s of km? $\endgroup$ – Rob Jeffries Mar 15 '16 at 15:10
  • $\begingroup$ As commented before this is not in the visible range of EM radiation, it is gamma-rays with much much more energy per photon. As for the nearly coincident signals, these are EM wave, travel at c. Depending on the angle of incidence they will have different tranfer of momentum to the reflecting surface, producing slightly different amplitudes of oscillation. $\endgroup$ – bonif Mar 15 '16 at 15:35
  • $\begingroup$ @kyle a) I would also like to see such paper, if only to discard it based on solid physics. The actual weight of my question is to wonder why this hasn't been considered by LIGO, while they have papers on something like Schumann resonances for instance. It would be great if someone related to the Ligo collaboration clarified this. $\endgroup$ – bonif Mar 15 '16 at 15:52
  • 3
    $\begingroup$ I will write up an answer if you can give the source that shows that, Under the right circumstances of angle and intensity of the flash beam impulse a vibration of the mirrors on both LIGO sites could be produced of the form seen in GW150914 otherwise, this is just more crackpottery not with anyone's time. $\endgroup$ – Kyle Kanos Mar 15 '16 at 23:43

The radiation pressure from a ~50 Hz (freq of the aLIGO wiggle) radio wave with an amplitude of 1pT (a typical Schumann wave) or 10pT (which is E=cB=3 mV/m) for a less frequent Q-burst (associated with Sprites) is very very small. Also, it would push the 40kg mirror mass in one direction only. If it arrived as an impulse, the mirror would swing at its pendulum freq of < 1 Hz and not make a ~50 Hz chirp back and forth waveform.

However, you may be looking in the correct place for a non-gravitational effect. It turns out aLIGO's end mirrors may be charged!!

If there was a ~50 Hz radio wave with a chirp pattern, it might be able to explain what aLIGO saw. At a recent aLIGO talk, I asked the speaker about the charge on the end mirrors. He said a charge may be there and was an active topic of investigation within aLIGO. For now we will have to estimate it. The position of the 40 kg x .34 m diam mirror is controlled by pushing against another mass hanging from the same pendulum stage. This adjustment is needed to precisely send the laser beam back down the 4 km to the splitter mirror. This pusher plate is 5 mm away, has concentric electrodes on it, and is divided into quadrants. As I understand from an aLIGO paper, up to +-280 volts (and up to an additional 500 volt offset) may be placed on these electrodes. If the average voltage on these electrodes is not zero, then the other plate of the capacitor (the mirror) charges up. The capacitance between the two plates is 160 pf. If (wild guess) the average voltage were 100 v, then the mirror would have q=CV= 16 nC of charge on it.

For a driving freq of 50 Hz, the mirror (< 1 Hz pendulum freq) behaves as a free mass. So the amplitude of its motion for a 10pT Q-burst is:

$$ x=\frac{qE}{m (2\pi \nu)^2}=\frac{(16*10^{-9}coul)(10^{-11}Tesla)(3*10^8m/sec) }{(40kg)(2\pi 50 sec^{-1})^2}=1.2*10^{-17}meters $$ $$ strain=\frac{1.2*10^{-17}meters}{4000meters}=3*10^{-21} $$ The 50 Hz radio wave would also be attenuated by ~1/3 by passing thru a 1 cm aluminum vacuum chamber wall. The aLIGO signal at both interferometers had an amplitude of $.5*10^{-21}$ so a Q-burst size signal, happening in the ionosphere between the two interferometers, is large enough given our assumption of 16 nC on the mirror.

A 10 pT amplitude signal would not have been vetoed by the aLIGO magnetometers. A LIGO paper said the magnetometers had a noise of 4pT/sqrt(Hz). Integrating this over the bandwidth 35-350 Hz that aLIGO filtered its strain signal, the magnetometer threshold for detecting a glitch was probably greater than 71 pT.

However, in googling the literature, I have found no atmospheric effect waveform that looks like the aLIGO chirp (increases in freq and amplitude as time progresses). The Q-burst mentioned above is a spike with some decaying oscillations of about the correct freq and does not look like the aLIGO chirp. Though aLIGO's detection doesn't look like a Q-burst, the above calculation shows (if 16 nC on the mirror is correct) that an electromagnetic wave could make the observed strain and escape the magnetometer's glitch veto. Perhaps LIGO has discovered some previously unknown, small, and infrequent electromagnetic atmospheric phenomena?

Maybe someone from within LIGO is on Physics Stack and can comment/ add info.

What the aLIGO signal is will become clearer as they see more events. Very exciting, and I too hope it is a gravitational wave for the window this would open on the universe!

Addendum 1: Calculation of the strain caused by a "~100 Hz laser spring oscillator" receiving a .005 sec (=1/2 period) impulse of EM radiation perpendicular to the mirror face and completely absorbed by the mirror. The 100 pT EM wave is probably just below what the magnetometer will veto as a glitch.

The momentum pmax transferred to the mirror oscillator is

$$ pmax=\frac{1}{c\mu_0} (E\times B)*Area=\frac{(3*10^8m/sec)(10^{-10}Tesla)^2(\pi (.17m)^2)(.005 sec)}{(3*10^8m/sec)(4\pi 10^{-7})}=3.6*10^{-18} kg-m/sec $$

After 1/4 of a period the mirror will have p=0 at its max amplitude of xmax

$$ xmax=\frac{pmax}{m(2\pi \nu)}=\frac{3.6*10^{-18} kg-m/sec}{(40kg)(2\pi100sec^{-1})}=1.4*10^{-22}m $$ $$ strain=\frac{1.4*10^{-22}m}{4000m}=3.6*10^{-26} $$ which is much less than the $10^{-21}$ peak strain aLIGO saw.

Addendum 2: Now consider if Terrestrial Gamma Flashes(TGFs) of ~1 MeV gamma rays might give a LIGO mirror enough impulse. Assume the 40 kg mirror is absorbing relativistic particles like photons so $p=\frac{E}{c}$. Calculate how much energy must absorbed by the "100 Hz laser spring osc" to account for the strain amplitude seen.

$$ E=c*pmax=c*m(2\pi \nu)xmax=(3*10^8 m/sec)(2\pi*100Hz)(10^{-21}*4000m)=3*10^{-5}joules $$ Convert this to MeV and divide by the area of the mirror $$ Flux=(\frac{3*10^{-5}joules}{1.6*10^{-13}joule/Mev})(\frac{1}{\pi (17cm)^2})=2*10^5Mev/cm^2 $$ The Fermi papers say the TGF events are <1/4 msec (so ~delta function impulse to excite our 10 msec period osc) but do not report the total energy deposited in the GBM BGO. What they do say is that the largest TGF events they have seen have ~300 gammas in the ~300 cm2 area of their BGO detectors, the energy spectrum falls ~$E^{-2}$, and ~40 MeV is the largest energy gamma they have seen in ~3000 events in 4 years of data. So, we can calculate a big overestimate of the MeV/cm2 they have seen in their rarest event: $$ Flux_{GBMmax}=\frac{300*40Mev}{300cm^2}=40Mev/cm^2 $$ This falls 4-orders of magnitude short of what we calculated as needed for the aLIGO signal. Yes, the question of distance from the TGF lightning has been ignored, but Fermi with its 450 mile orbital altitude probably has been as close (or closer) to a lightning storm in 4 years as the two interferometers (2000 miles apart/2=1000 miles to the lightning) were in two weeks of LIGO data taking. We have also ignored the shielding of the atmosphere which would attenuate the energy reaching LIGO even more.

  • $\begingroup$ When analyzing the possibility of an EM origin of GW150914 we of course first thought of ELF radiation(as this was also considered by the LIGO paper mentioned above) in the form of Q-bursts either from Schumann's resonances or other lightning related sources(as I said above TGFs are often accompanied by intense ELF radiation. But as you note there is no way to produce a waveform like the correlated one coupling the oscillating field of such low freq/energy wave to the mirrors. $\endgroup$ – bonif Mar 16 '16 at 10:36
  • $\begingroup$ I should clarify something about the aLIGO interferometers system of laser and mirrors, the latter rather than simple pendulums form with the laser beams an optical spring(akin to an spring pendulum) and because of the nature of the GW signals that were most likely expected to be detected(binary mergers) this optical spring is designed to be most sensitive in amplitude at frequencies between 80-250 Hz. It is possible for an energetic enough impulse to generate a few oscillations in this range in an optical spring pendulum. And if you look at the graph of the correlated waveforms you can.... $\endgroup$ – bonif Mar 16 '16 at 10:40
  • $\begingroup$ ...see that the exact correlation between the two sites begins just before the peak amplitude and around 100 Hz, even if the previous cycles are included as part of the putative GW signal they are compatible with noise filtered in the approximate desired frequency. $\endgroup$ – bonif Mar 16 '16 at 10:47
  • $\begingroup$ See fig.1 of the detection paper, one can observe what I comment above in the first right(superposed waveform) and third rows(residual noise with numerical relativity substracted) and specially bottom row with the chirp starting at around 0.36s $\endgroup$ – bonif Mar 16 '16 at 11:02
  • $\begingroup$ @bonif I added an addendum above. Using your idea of the laser light making an optical spring for the mirror mass, I calculate the oscillation amplitude that an EM impulse (of an amplitude just below the magnetometer veto) would cause by radiation pressure for this 100 Hz oscillator. It appears to come out 4-5 order of magnitudes smaller than what aLIGO saw. A 100 times larger amplitude EM pulse might do it, but that should show up in the magnetometer. $\endgroup$ – Gary Godfrey Mar 16 '16 at 19:49

This is too hand waving, it needs geographical and timing information of thunderstorms.

To get a similar strength in the two LIGOs it would have to be in the middle, and a unique crush. Thunderstorms have time sequences, very seldom there is only one isolated bolt in time, the way the LIGO signal is.

To get serious attention by the LIGO team one could work on a preprint with the estimates of frequencies in the region between the LIGO detectors. I am sure thunder storms are recorded in some data base and a coincidence in timing could be searched.

  • $\begingroup$ TGFs are associated to thunderstorms but are NOT the actual visible lightning bolt. That's why they are called "dark lightning". As TGFs are not completely understood yet (and even less known by many physicists according to my experience) I recommend reading something about how they are thought to form. $\endgroup$ – bonif Mar 15 '16 at 15:18
  • $\begingroup$ And yes my team is working on a preprint with the modelling and calculations. $\endgroup$ – bonif Mar 15 '16 at 15:20
  • $\begingroup$ About the thunderstorm recordings, yes, there is such a worlwide monitoring watch. However it is obviously not exhaustive and one feature of TGFs is that they are not associated to great thunderstorms wich are the ones mostly monitored. They are observed associated to modest thunderstorms frequently. $\endgroup$ – bonif Mar 15 '16 at 15:57

If there are thousands of TGR events per day worldwide then the fact they proposed signals in the detector are rare pretty much rules them out: you'd see (on average) a handful of candidate events per day as singles in each detector and a non-trivial number per month as candidate coincidences. From the single:coincidence ratio you could infer then range over which TGRs were mimicing the expected signal (which must be on order of the distance between the detectors if this hypothesis is to be true).

  • $\begingroup$ Over a thousand is the estimate worldwide, yes. The time the Ligo run lasted was about 15 days(a month on and off). They got at least one signal, but there's talk (rumors) of from at least one more to seven or six weaker signals more. The fact is as long as there is not a serious modelling of the amount of thunderstorms in the area capable of producing a flash with the adequate orientation for the two sites it is very hard to say what would be the statistics(and the big iff is whether this is actually a possibility, I've only run order of magnitude estimates). $\endgroup$ – bonif Mar 15 '16 at 15:26
  • $\begingroup$ Besides, the kind of TGFs most likely to produce a signal would be from the lower atmosphere and there's no way to really know about these since basically only those from the upper atmosphere make it through to the gamma-ray telescope detectors. $\endgroup$ – bonif Mar 15 '16 at 15:28
  • $\begingroup$ You can't have it both ways on penetration. Either the events generate radiation that penetrates many hundreds of kilometers of atmosphere (for one such even to trigger both detectors) or ones in the lower atmosphere are blocked and not seen by the satellites. $\endgroup$ – dmckee Mar 15 '16 at 15:31
  • $\begingroup$ Those are two different possibilities. What I'm saying is that those that are produced higher up are the ones that Fermi telescope more easily detect. And those suspected to occurr in the lower atmosphere are hardly detected by Fermi but would be more likely capable of making it to a surface sensitive enough detector. $\endgroup$ – bonif Mar 15 '16 at 15:40
  • $\begingroup$ By the way TGFs are usually accompanied by afterglows, in the form of X rays and also ELF EM waves. Fermi detected X rays 0.4 seconds after the signal that didn't trigger the monitor but was subsequently dug up from the data, it has been published and widely talked about but pointing to a possible GRB. I have knowledge on the other hand of the detection of intense ELF radiation coincident with the GW signal. $\endgroup$ – bonif Mar 15 '16 at 15:48

Not the answer you're looking for? Browse other questions tagged or ask your own question.