Raphaël Enthoven,Jacques Perry-salkow
(The validation of) A great discovery requires a genuinely independent analysis of data
To date, the LIGO collaboration has detected three gravitational wave (GW) events appearing in both its Hanford and Livingston detectors. In this article we reexamine the LIGO data with regard to correlations between the two detectors. With special focus on GW150914, we report correlations in the detector noise which, at the time of the event, happen to be maximized for the same time lag as that found for the event itself. Specifically, we analyze correlations in the calibration lines in the vicinity of 35 Hz as well as the residual noise in the data after subtraction of the best-fit theoretical templates. The residual noise for the other two events, GW151226 and GW170104, exhibits similar behavior. A clear distinction between signal and noise therefore remains to be established in order to determine the contribution of gravitational waves to the detected signals.
(Submitted on 13 Jun 2017 (v1), last revised 9 Aug 2017 (this version, v2))
A debate about how to sift the astrophysical wheat from the terrestrial chaff
Recent claims in a preprint by Creswell et al. of puzzling correlations in LIGO data have broadened interest in understanding the publicly available LIGO data around the times of the detected gravitational-wave events. We see that the features presented in Creswell et al. arose from misunderstandings of public data products. The LIGO Scientific Collaboration and Virgo Collaboration (LVC) have full confidence in our published results, and we are preparing a paper in which we will provide more details about LIGO detector noise properties and the data analysis techniques used by the LVC to detect gravitational-wave signals and infer their waveforms.
News from LIGO Scientific Collaborationundated (between 7 July and 1 August 2017)
In our view, if we are to conclude reliably that this signal is due to a genuine astrophysical event, apart from chance-correlations, there should be no correlation between the "residual" time records from LIGO's two detectors in Hanford and Livingston. The residual records are defined as the difference between the cleaned records and the best GW template found by LIGO. Residual records should thus be dominated by noise, and they should show no correlations between Hanford and Livingston. Our investigation revealed that these residuals are, in fact, strongly correlated. Moreover, the time delay for these correlations coincides with the 6.9 ms time delay found for the putative GW signal itself...
During a two-week period at the beginning of August, we had a number of "unofficial" seminars and informal discussions with colleagues participating in the LIGO collaboration... Given the media hype surrounding our recent publication, these meetings began with some measure of scepticism on both sides. The atmosphere improved dramatically as our meetings progressed.
The focus of these meetings was on the detailed presentation and lively critical discussion of the data analysis methods adopted by the two groups. While there was unofficial agreement on a number of important topics - such as the desirability of better public access to LIGO data and codes - we emphasize that no consensus view emerged on fundamental issues related to data analysis and interpretation.
In view of unsubstantiated claims of errors in our calculations, we appreciated the opportunity to go through our respective codes together - line by line when necessary - until agreement was reached. This check did not lead to revisions in the results of calculations reported in versions 1 and 2 of arXiv:1706.04191 or in the version of our paper published in JCAP. It did result in changes to the codes used by our visitors.
There are a number of in-principle issues on which we disagree with LIGO's approach. Given the importance of LIGO's claims, we believe that it is essential to establish the correlation between Hanford and Livingston signals and to determine the shape of these signals without employing templates. Before such comparisons can be made, the quality of data cleaning (which necessarily includes the removal of non-Gaussian and non-stationary instrumental "foreground" effects) must be demonstrated by showing that the residuals consist only of uncorrelated Gaussian noise. We believe that suitable cleaning is a mandatory prerequisite for any meaningful comparisons with specific astrophysical models of GW events. This is why we are concerned, for example, about the pronounced "phase lock" in the LIGO data.
James Creswell, Sebastian von Hausegger, Andrew D. Jackson, Hao Liu, Pavel NaselskyAugust 21, 2017
Disentangling the man-made detectors from the Earth-shaped one
As the LIGO detectors are extremely sensitive instruments they are prone to many sources of noise that need to be identified and removed from the data. An impressive amount of efforts were undertaken by the LIGO collaboration to ensure that GW150914 signal was really the first detection of gravitational waves with all transient noise backgrounds being under a good control [4, 5, 6].
It was claimed, however, in a recent publication  that the residual noise of the GW150914 event in LIGO’s two widely separated detectors exhibit correlations that are maximized for the same 7 ms time lag as that found for the gravitational-wave signal itself. Thus questions on the integrity and reliability of the gravitational waves detection were raised and informally discussed [8, 9]. It seems at present time it is not quite clear whether there is something unexplained in LIGO noise that may be of genuine interest. It was argued that even assuming that the claims of  about correlated noise are true, it would not affect the 5-sigma confidence associated with GW0150914 . Nevertheless, in this case it will be interesting to find out the origin of this correlated noise.
Correlated magnetic fields from Schumann resonances constitute a well known potential source of correlated noise in gravitational waves detectors [11, 12, 13]... Schumann resonances are global electromagnetic resonances in the Earthionosphere cavity [14, 15]. The electromagnetic waves in the extremely low frequencies (ELF) range (3Hz to 3 kHz) are mostly confined in this spherical cavity and their propagation is characterized by very low attenuation which in the 5 Hz to 60 Hz frequency range is of the order of 0.5-1 db/Mm. Schumann resonances are eigenfrequencies of the Earth-ionosphere cavity. They are constantly excited by lightning discharges around the globe. While individual lightning signals below 100 Hz are very weak, thanks to the very low attenuation, related ELF electromagnetic waves can be propagated a number of times around the globe, constructively interfere for wavelengths comparable with the Earth’s circumference and create standing waves in the cavity.
Note that there exists some day-night variation of the resonance frequencies, and some catastrophic events, like a nuclear explosion, simultaneously lower all the resonance frequencies by about 0.5 Hz due to lowering of the effective ionosphere height . Interestingly, frequency decrease of comparable magnitude of the first Schumann resonance, caused by the extremely intense cosmic gamma-ray flare, was reported in . Usually eight distinct Schumann resonances are reliably detected in the frequency range from 7 Hz to 52 Hz. However five more were detected thanks to particularly intense lightning discharges, thus extending the frequency range up to 90 Hz .
... For short duration gravitationalwave transients, like the three gravitational-waves signals observed by LIGO, Schumann resonances are not considered as significant noise sources because the magnetic field amplitudes induced by even strong remote lightning strikes usually are of the order of a picotesla, too small to produce strong signals in the LIGO gravitational-wave channel .
Interestingly enough, the Schumann resonances make the Earth a natural gravitational-wave detector, albeit not very sensitive . As the Earth is positively charged with respect to ionosphere, a static electric field, the so-called fair weather field is present in the earth-ionosphere cavity. In the presence of this background electric field, the infalling gravitational wave of suitable frequency resonantly excites the Schumann eigenmodes, most effectively the second Schumann resonance . Unfortunately, it is not practical to turn Earth into a gravitational-wave detector. Because of the weakness of the fair weather field (about 100 V/m) and low value of the quality factor (from 2 to 6) of the Earth-ionosphere resonant cavity, the sensitivity of such detector will be many orders of magnitude smaller than the sensitivity of the modern gravitational-wave detectors.
However, a recent study of short duration magnetic field transients that were coincident in low-noise magnetometers in Poland and Colorado revealed that there was about 2.3 coincident events per day where the amplitudes of the pulses exceeded 200 pT, strong enough to induce a gravitational-wave like signal in the LIGO gravitational-wave channel of the same amplitude as in the GW150914 event ...
The main source of the Schumann ELF waves are negative cloud-toground lightning discharges with the typical charge moment change of about 6 Ckm. On Earth, storm cells, mostly in the tropics, generate about 50 such discharges per second.
The so-called Q-bursts are more strong positive cloud-to-ground atmospheric discharges with charge moment changes of order of 1000 Ckm. ELF pulses excited by Q-bursts propagate around the world. At very far distances only the low frequency components of the ELF pulse will be clearly visible, because the higher frequency components experience more attenuation than the lower frequency components...
In  Earth’s lightning hotspots are revealed in detail using 16 years of space-based Lightning Imaging Sensor observations. Information about locations of these lightning hotspots allows us to calculate time lags between arrivals of the ELF transients from these locations to the LIGO-Livingston (latitude 30.563◦ , longitude −90.774◦ ) and LIGO-Hanford (latitude 46.455◦ , longitude −119.408◦ ) gravitational-wave detectors...
We have taken Earth’s lightning hotspots from  with lightning flash rate densities more than about 100 fl km−2 yr−1 and calculated the expected time lags between ELF transients arrivals from these locations to the LIGO detectors... Note that the observed group velocity for short ELF field transients depends on the upper frequency limit of the receiver . For the magnetometers used in  this frequency limit was 300 Hz corresponding to the quoted group velocity of about 0.88c. For the LIGO detectors the coupling of magnetic field to differential arm motion decreases by an order of magnitude for 30 Hz compared to 10 Hz . Thus for the LIGO detectors, as the ELF transients receivers, the more appropriate upper frequency limit is about 30 Hz, not 300 Hz. According to (2), low frequencies propagate with smaller velocities 0.75c-0.8c. Therefore the inferred time lags in the Table1 might be underestimated by about 15%...
If the strong lightnings and Q-bursts indeed contribute to the LIGO detectors correlated noise then the distribution of lightning hotspots around the globe can lead to some regularities in this correlated noise. Namely, extremely low frequency transients due to lightnings in Africa will be characterized by 5-7 ms time lags between the LIGO-Hanford and LIGO-Livingston detectors. Asian lightnings lead to time lags which have about the same magnitude but the opposite sign. Lightnings in North and South Americas should lead positive time lags of about 11-13 ms, greater than the light propagation time between the LIGO-Hanford and LIGO-Livingston detectors.
(Submitted on 27 Jul 2017)