,
16 tweets,
8 min read
Read on Twitter
Lady Gaga as instrumental noise glitches in LIGO data: A thread.
An instrument designed to detect extremely tiny Gravitational Wave ripples, @LIGO is susceptible to a great deal of instrumental and environmental noise. Researchers in the detector characterization group at LIGO work night and day to understand and help mitigate this.
The LIGO detector data can be best visualized in blue time-frequency spectrogram images. Any noise (or perhaps a Gravitational Wave signal!) higher than expected levels (blue) appears as a yellow burst in these plots. 
Loud, transient bursts of noise in the data are termed as ‘glitches’. They show a variety of morphologies in the spectrograms, providing clues on what causes them. Similar glitches may come from similar sources.
The citizen science and machine-learning based tool ‘Gravity Spy’ @GravitySpyZoo helps classify these using creative labels. You too can participate here, it’s fun! zooniverse.org/projects/zooni…
We start with some extremely loud glitches which can threaten to knock the interferometer out of observing mode. They can happen due to a variety of reasons.
A famous loud glitch was caused due to saturation in the digital-to-analog converter of the feedback signal controlling the position of the mirrors. It occurred right on top of the first binary neutron star signal in the Livingston detector!
‘Whistle’ glitches are caused by radio frequency beats, and are so named because of the way they sweep up and down in frequency similar to the sound of a whistle. (The scorpion in the middle cannot be resolved.)
Sometimes, whistles look like nasty horns in the spectrograms...
‘Blip’ glitches are short spikes that occur quite frequently but no one knows what causes them. They are particularly wicked since they can mimic actual gravitational wave signals coming from a high mass ratio compact binary merger.
Sometimes blip-like spikes repeat in a melee as if a cat has been scratching the spectrogram. Hence, ‘scratchy’ glitches.
‘Tomte’ glitches are conical bursts that I helped investigate during a detector characterization noise sprint workshop this January. Our efforts are summarized in the Gravity Spy blog: blog.gravityspy.org
Some glitches seen during the second observing run resembled the DNA double helix.
At the end of the day, a complete understanding of LIGO noise can help us distinguish it from the beautiful Gravitational Wave signals from merging compact binaries, called ‘chirps’, like this one here from GW150914, the first black hole merger detected.
This thread was inspired by @JedMSP's wonderful thread on Prince as common population genetics visualizations.
