Flawed second LIGO gravitational wave discovery

In a previous post I listed the many problems associated with the LIGO team’s claim of having made the first direct observation of gravitational waves in history.  Now they have released a second claim that makes all but one of the problems with the initial claim seem irrelevant.  That problem is the methodology used to try to recover a signal from what is essentially unidentified and random background noise.  This “mystery noise” has continued to plague the LIGO detectors, particularly at low frequencies.  But a recent Science magazine online article reports that David Reitze, the executive director of LIGO at the California Institute of Technology “says he’s cautiously optimistic that physicists can eliminate the noise and reach design sensitivity by 2019.”  The director is quoted as saying “I won’t say with 100% confidence that we will get there, but I won’t say that we won’t either.”  Such a Bayesian answer is not surprising but still does leave one guessing.

As shown in the LIGO team’s own published paper, the signal for GW151226 is buried in the background noise and almost totally lost in the Livingston detector (see image at top).  The reason given is simply “due to the different sensitivities of the detectors at the time”.  Regardless of the reason, the fact remains that the signal was so weak it could not be picked up in a generic transient search like the first signal.  It was only found after months of pattern searching (“matched filtering”) using hundreds of thousands of different “waveform templates” as the patterns.  The background noise was searched for patterns that supposedly appear in both LIGO detector locations within 15 milliseconds of each other.  It is repeatedly emphasized in the published paper that GW151226 was not a coincident random match.  But consider that if every 15 milliseconds of incoming noise was searched using 250,000 different waveform templates, with no overlaps, that would equal 1 billion (1,000,000,000) separate pattern searches per minute of recorded noise.  Using such a robust search methodology certainly guarantees an eventual result; it should only take a matter of months, as happened in this case, not a thousand years as quoted in the published paper.

But even using such a large bank of waveform templates does not guarantee a perfect match.  In the end the “matched” waveform had to be tweaked to best fit it to the background noise pattern.  In fact the entire event had to be tweaked to make it at least marginally plausible.  The LIGO team does not mention this in their published paper but they have no choice but to admit it when they are supposedly providing their data to the public so they can openly duplicate the LIGO team’s results.  It is clearly stated in bold letters at the beginning of each of the LIGO event tutorials that “the results obtained here will not match precisely with numbers in our papers, due to various subtleties in the analysis that are discussed further down.”  Further down the tutorial page for GW151226 we are reminded again:

“As noted above, the results won’t be identical to what is in the LIGO-Virgo papers, since we’re skipping many subtleties, such as combining many consistent templates.”

It is clear subtleties are being skipped because their only other mention is close to the end of the tutorial page:

“The full analysis produces a Bayesian posterior result using many nearby templates. It does a more careful job estimating the ASD, and includes effects of uncertain calibration.  As a result, our parameters (SNR, masses, spins, D_eff) are somewhat different from what you will see in our papers.”

So basically the LIGO team is admitting to publishing a manipulated final result and no real exact means for anyone outside of the LIGO project to duplicate it.  It also appears the team tried to bolster and perhaps even justify their reporting of this highly questionable second event by including a third one they had previously dismissed.  In their published paper this third event is labeled LVT151012 and is described as “the third most significant binary black hole candidate event in the observing period.”  However when the event was originally reported on the LIGO LSC website it was described as follows:

“LVT151012, was found in the data with a fairly low probability of being a false alarm (about 2%) which is enough to get our attention, but not enough to claim a detection. (LIGO detections are named “GW” followed by the date in YYMMDD format. LIGO/Virgo candidates start with a “LVT” for “LIGO-Virgo Trigger” followed by the date.  LVT151012 is an interesting candidate, but is not significant enough to “graduate” and become a “GW”).”

This insignificance is not described in the LIGO team’s published paper, apparently in an attempt to add legitimacy to the third event and consequently to all of their gravitational wave discovery claims.  The pressure to legitimize their claims and for that matter the entire LIGO project itself is only going to increase later this year when Advanced LIGO starts its second observation run and is eventually joined by the Virgo detector in Italy.  It will be truly interesting to see how the LIGO collaboration will attempt to pattern match background noise across four detectors instead of just two.  And it will be truly remarkable to see them able to do so with such efficiency as to eventually be able to detect one event per day as estimated by LIGO team member Stephen Fairhurst of Cardiff University in the same aforementioned Science article.

Each of these issues should be legitimate concerns for the LIGO collaboration and for the scientific community as a whole.  But what should really concern the public in general is the question of the continuation of LIGO if it fails to live up to expectations and all of the hype.  Let’s not forget that LIGO is still primarily funded with public money.  If it has already resorted to questionable pattern searches of random instrument noise and entered the realm of pseudoscience should the public have to continue to fund it?  I personally don’t have a problem with public funds being used in the pursuit of science and would even like to see an increase in spending.  But after 40 years and billions spent on LIGO and only vagaries in return I would like to see the time and money finally spent elsewhere.

Shannon Sims
Curator and Host