Most of what I do at Relative Gravity Labs doesn’t involve a laser interferometer. It involves a notebook, a timestamp, and a witness who is usually more rattled than reliable. But every field report I file eventually has to cross a desk that does involve a laser interferometer — Project Horizon’s — and in five years of doing this job, I’ve learned that the single biggest reason a promising sighting report goes nowhere isn’t a bad witness. It’s a bad timestamp.
This piece is not a physics paper. Elena and Liam write those. This is the primer I wish someone had handed me on my first week: what a gravitational wave actually is, how we detect something a billion times smaller than a proton, and why getting the clock right matters more than almost anything else a field investigator does.
What a Gravitational Wave Actually Is
In 1916, Einstein’s general theory of relativity predicted that massive accelerating objects — two black holes spiraling into each other, for instance — should send ripples through spacetime itself, the same way a dropped stone sends ripples across a pond. Unlike a stone’s ripples, these don’t move through space. They move space, stretching and squeezing it by an almost absurdly small amount as they pass, at the speed of light.
For decades this was a theoretical curiosity. Indirect evidence arrived in the 1970s, when astronomers Russell Hulse and Joseph Taylor tracked a binary pulsar whose orbit was decaying exactly as predicted if it were losing energy to gravitational radiation — work that earned them the 1993 Nobel Prize. Direct detection didn’t come until September 14, 2015, when the twin LIGO detectors in the United States picked up the signature of two black holes, roughly 29 and 36 times the mass of the Sun, merging 1.3 billion light-years away. The event is catalogued as GW150914. It remains the cleanest example of what we’re all looking for.
How We Detect Something a Billion Times Smaller Than a Proton
LIGO, and its partner observatories Virgo (Italy) and KAGRA (Japan), work on the same principle our own Project Horizon sensors scale down: laser interferometry. Split a laser beam, send each half down a long arm at right angles to the other, bounce both off mirrors, and recombine them. If nothing has disturbed the arms, the beams cancel out in a predictable pattern. If a gravitational wave passes through, it stretches one arm by a fraction of a proton’s width while squeezing the other — and the recombined beams shift out of that pattern in a way a sufficiently sensitive detector can register.
“Sufficiently sensitive” undersells it. LIGO’s full-scale arms are 4 kilometers long and can register a strain — a fractional change in length — of about 1 part in 10²¹. That’s the equivalent of measuring the distance to the nearest star to within the width of a human hair. Our own Horizon array, built from modified commercial interferometry kits, is nowhere near that sensitive. What it can do is flag ambient strain anomalies at each of our three sites and let us cross-reference the timing against public data.
The waveform itself has a signature shape investigators should recognize even without the math: a “chirp” — rising in both frequency and amplitude as two massive objects spiral closer together, peaking at the moment of merger, then dying off in a brief ringdown. Seismic noise, trucks, and construction don’t chirp. That shape is most of what separates a real candidate from ambient junk.
Why This Matters to Field Investigators
Here’s the part that’s actually my job. Public gravitational-wave data — from LIGO, Virgo, and KAGRA’s joint observing runs — is freely available through the Gravitational Wave Open Science Center (GWOSC). Every confirmed event is timestamped to the second. So is every strain anomaly Project Horizon logs at our own sites. And so, if we’ve done our job, is every field sighting report that comes across my desk.
When those three timelines line up, we have something worth escalating to Tesseract for modeling. When they don’t — and honestly, most of the time they don’t — we’ve still ruled something out cleanly, which is its own kind of progress.
The problem is that “when they line up” only works if the sighting timestamp is trustworthy in the first place. A witness who says “around 11, maybe 11:30” gives us a 30-minute window to search against, during which dozens of unrelated strain fluctuations and precisely zero confirmed astrophysical events might have occurred. A witness who checked their phone the moment it happened gives us a window measured in seconds. The difference between those two reports, in terms of what we can actually do with them, is enormous.
A Standard Field Timestamp Protocol
Every field investigator at RGL now works from the same checklist, and I’d recommend it to anyone doing similar work independently:
- Log in UTC, not local time. Convert immediately, on-site, before memory does the converting for you later.
- Get the second, not just the minute. If the witness has a timestamped photo, video, notification, or call log, that beats any verbal estimate.
- Note anything nearby that could register as ground noise — passing vehicles, construction, weather — that a detector site might also pick up and that could produce a coincidental match.
- Check GWOSC’s public event list before filing an anomaly candidate. If a confirmed astrophysical merger already accounts for the timing, that’s a more parsimonious explanation than anything else on offer, and it should be logged as such.
- File the report whether or not anything lines up. A clean non-match is data. We’ve thrown away too many of those in the early years by only keeping the interesting ones.
None of this proves anything about what any given witness saw. What it does is make sure that if Horizon, Tesseract, or some future project ever finds a real correlation, it’ll be sitting in a dataset clean enough to trust — instead of buried under thirty years of “around 11, maybe 11:30.”