It's been a long few years for everyone, but we're excited to get back to searching the universe for gravitational waves! The LIGO/Virgo/KAGRA detectors, which were down for commissioning and upgrades since April 2020, have just started the next observing run. The LIGO detectors started the 4th observing run (O4) on May 24 2023, which will run for a year and a half. This also means new glitches are starting to roll in! See here more information. And in Gravity Spy 2.0, you can explore some of the other data the detectors are recording.

It's been a long few years for everyone, but we're excited to get back to searching the universe for gravitational waves! The LIGO/Virgo/KAGRA detectors, which were down for commissioning and upgrades since April 2020, have just started the next observing run. The LIGO detectors started the 4th observing run (O4) on May 24 2023, which will run for a year and a half. This also means new glitches are starting to roll in! See here more information. And in Gravity Spy 2.0, you can explore some of the other data the detectors are recording.


Frequently Asked Questions

Gravitational Waves FAQ

What are gravitational waves?

Gravitational waves were predicted by Albert Einstein’s theory of general relativity in 1916. He showed that accelerating masses distort spacetime as they move, sending out ripples in the fabric of spacetime itself. An (imperfect) analogy is imaging your finger moving through water. If your finger is still, no waves will be created on the water’s surface. However, if you start to move your finger through the bowl of water, ripples will propagate outwards. Gravitational waves move at the speed of light, and pass through matter. They have the effect of stretching and compressing spacetime as they move, causing distances to stretch and shrink.

What is LIGO?

‘LIGO’ stands for the Laser Interferometer Gravitational-Wave Observatory, the largest experiment to detect gravitational waves that is currently in operation. The LIGO Scientific Collaboration (LSC) consists of scientists from nearly one hundred institutions, all working together to develop the field of gravitational wave astronomy. LIGO operates two main observatories in the United States (LIGO Livingston in Louisiana and LIGO Hanford in Washington), and also collaborates with the Virgo observatory in Italy and the KAGRA observatory in Japan.

How do you detect gravitational waves?

In essence, the instruments LIGO uses to detect gravitational waves are called Michelson interferometers. They take advantage of the way gravitational waves distort spacetime– namely, they take advantage of the fact that gravitational waves distort spacetime in perpendicular directions. In reality, the detectors are much, much more complicated than a standard Michelson interferometer, but to get a taste of how this instrument can detect gravitational waves this simple picture will suffice.

A Michelson interferometer consists of two arms of equal length oriented perpendicular to each other, joined at the elbow (picture an L-shape). A laser beam is sent to the elbow and then split by a beam splitter. The beam splitter splits the laser beam at the elbow, causing separate beams to travel down each arm. At the end of each arm, there lies a reflective mirror, which causes the beam to be reflected back toward the elbow.

Because light travels at a constant speed (the speed of light) and the arms are equal in length, we know from basic physics that the light should return to the elbow of the interferometer at the same time.

Here is where the perpendicular orientation of the arms comes into play: when the laser beams both return to the elbow, the perpendicular orientation of the arms guarantees that the light will be perfectly out of phase an destructively interfere at a detector, such that the beams effectively cancel each other out. We therefore expect, under normal conditions, there to be no leftover laser light at all at the elbow. (We measure the amount of light leftover at the elbow with an instrument called a photodetector.)

However, we understand the physical effect of a gravitational wave to be the stretching and compressing of spacetime. If a gravitational wave were to pass through the interferometer throughout this process, we would, in fact, expect there to be some leftover laser light at the photodetector. This is because the gravitational wave would stretch one arm while compressing the other. Then, it would do the opposite (stretch the other arm while compressing the first). Essentially, the gravitational wave changes the lengths of the arms in opposite ways at the same time, thus causing our resultant beam of light at the photodetector to be nonzero in amplitude.

What do gravitational waves come from?

There are many cosmic sources of gravitational waves, but the main source LIGO is focused on detecting is the violent collision between highly massive and compact objects. The gravitational-wave signals detected so far have all come from collisions between black holes, but it is possible that collisions between two neutron stars (or between black holes and neutron stars) will eventually be detected. In fact, the names of the multiple workflows of the Gravity Spy project are all named after potential LIGO sources.

I thought gravitational waves were found, aren't we done with this?

Far from! The detections that LIGO have made so far confirm the existence of gravitational waves, but the significance of gravitational-wave observation goes far beyond simply proving Einstein right. With the LIGO, Virgo and KAGRA detectors, it will be possible to detect a range of astronomical events never before directly recorded. For example, in addition to being the first gravitational-wave detection, LIGO’s first discovery was also the first direct detection of merging black holes made in any sort of observatory.

Can you actually hear gravitational waves?

We are certainly not able to hear gravitational-wave signals the way we hear normal sounds– they are waves in the medium of spacetime itself, while what we traditionally experience as sound consists of compressions of the air around us.

However, the frequencies of the gravitational waves that LIGO can detect have the same frequencies as sounds we can hear. Because of this, you may have heard a recording of the ‘sound of a gravitational wave’. These recordings are translations of LIGO data into sounds we can hear, and allow another way for us to conceptualize the way gravitational waves move.

How is the discovery of gravitational waves significant for the future of astronomy?

The discovery of gravitational waves opens the door to an entirely new field of astronomy. Previously, (almost) all of our observational information about astrophysical objects has relied on electromagnetic information, i.e., light. However, with gravitational-wave astronomy, we will be able to combine electromagnetic information with the information we receive from gravitational waves, leading to a new form of multi-messenger astronomy. In addition, we will be able to probe compact objects that emit no electromagnetic radiation, but do emit gravitational waves, such as black holes.

How many scientists are working on LIGO?

As of May 2021, the LIGO Scientific Collaboration consists of over 1400 members from 127 member institutions (universities and other research institutes). These members represent 19 different countries across the world.

More scientists are working on the companion observatories of Virgo and KAGRA. Much of the data analysis is shared, so scientists from each work together as part of the LIGO Scientific, Virgo and KAGRA Collaboration.

Where can I get more information about gravitational waves?

There are plenty of resources on the web for learning more about gravitational waves. Here are a few:

Recap of the first gravitational wave detection


LIGO Science Summary

Glitches FAQ

Why do we need to classify glitches, since they are not gravitational waves?

Glitches can obscure or mimic a true gravitational-wave signal. Glitches also make our searches for gravitational waves less efficient, and generally make the LIGO instruments less sensitive to astrophysical events. It is important that we know as much about glitches as possible, in order to separate them from true signals, remove them from the data, or in the best case eliminate them from the detector entirely. Additionally, while some types of glitches have already been matched to a physical cause, others are still mysteries. Finding out what physical effects cause these unidentified glitches may allow us to eliminate data when they occur and/or remove them from the detectors.

What is 'normalized energy'?

Normalized energy can be thought of as the amplitude of the signal, or how ‘strong’ a signal is in the detectors. In the images you will be classifying, the color of the image represents the normalized energy at a specific time and frequency. A more yellow color indicates higher normalized energy, while a blue indicates lower normalized energy. Though the normalized energy color bar only goes up to 25, glitches can have normalized energy much higher than this value (you can think of it like an over-saturated image that is exposed to too much light so everything looks white in color). This is typically done in glitch imaging so one can compare glitches of a wide range of energies.

What are glitches and what causes them?

To detect gravitational waves from events like binary black hole inspirals, LIGO needs to be sensitive to changes in distance on the order of 10^-19 meters (i.e., about 10,000 times smaller than the width of a proton). The extreme sensitivity required to make such detections is achieved through exquisite isolation of all sensitive components of Advanced LIGO from non-gravitational-wave disturbances. Nonetheless, Advanced LIGO is still susceptible to a variety of instrumental and environmental sources of noise that contaminate the data. These sources of noise are called glitches. There are many different sources for glitches, some we know, and some yet to be discovered. You can check out the field guide for more details about known glitch types.

How many classes of glitches are there?

There are a couple dozen types of glitches that we currently know of. Over time, some glitches have come and gone in the detectors as their causes have been identified (such as a Air Compressor glitch, which was in the data from the first observing run but should not be around anymore as data from future observing runs rolls in), and some glitches have remained a mystery and stayed persistent in the detectors (such as Blip glitches). As of now, the Gravity Spy project has over 20 different options for glitch morphologies. The ‘none of the above’ category was included to account for the glitches that are not classified under one of our already predetermined classes. If enough glitches of a certain morphology are found in the none of the above category, our experts can decide to create a new glitch classification option. As Gravity Spy evolves, and more and more images are classified, we hope to discover more of these classes. Volunteers have already identified several new classes!

How did glitch classes get their names?

Though it may seem like it, glitches are not named randomly. Many of the glitches get their names from their shape and structure. For example, the class ‘tomte’ was named because those glitches look like hat of a tomte gnome. Other glitches get their names based on the physical properties of the signal. For example, the ‘low frequency burst’ class are usually quick glitches with lower frequencies.

How often do glitches occur?

Glitches occur at different rates based on what is happening within the detector and in the environment around the detector. At their highest rate, they can occur at about 3 times per second in the detectors.

How do you fix a glitch?

If we can figure out the cause of a glitch, LIGO scientists can potentially find a way to eliminate it from the detectors. Also, 'vetoes' can be created to remove bad data that is particularly afflicted by glitch activity. As one example of how a glitch was fixed, at LIGO Hanford, the Air Compressor glitches were found to be related to air compressor motors switching on and off at the end stations. These were solved at LIGO Hanford by replacing the vibration isolators (rubber feet) on air compressors.

Gravity Spy Project FAQ

Why are you getting regular people to do this? Wouldn't you want someone more knowledgeable? Or some research underling to do this?

The volume of glitches that need to be classified is much too large to be categorized by the relatively few researchers directly involved with Gravity Spy (and even the large number of scientists in LIGO, for that matter, as over a million relevant glitches were identified during LIGO’s first observing run alone). Crowd-sourcing this process allows us to have as much human input as possible, which helps improve our computer-based methods of glitch classification.

In order to facilitate this, the method of data processing used to create the images you will see when you are classifying glitches has been designed to make the relevant features of the glitches as apparent as possible, so that anyone can help out– not just trained researchers.

How do my classifications help computer algorithms to better classify?

The computer algorithms used, broadly known as machine learning, rely on a large, labeled data set created by humans. You can think of it as teaching the computer to find glitches by itself. With such large, labeled sets, computers have the benefit of being able to classify large amounts of data systematically. However, algorithms have a tough time identifying new categories of glitches that appear in the detectors. The amount of examples we provide it only increases its potential to classify glitches accurately.

How are my classifications helping the LIGO team?

Each new classification moves us one step further to improving our understanding of gravitational-wave detectors and how to analyze their data. Your classifications help LIGO members reduce detector noise, allowing searches through the data to unveil even quieter gravitational-wave signals from the Universe. Understanding the structure and tendencies of glitches is also important for data-quality vetoes which remove bad data from the searches, or even for determining the causation of a specific glitch class.

The Gravity Spy project has proved extremely helpful to LIGO since its launch in 2016. The ability of volunteers to spot new types of glitches is now well established, and our machine-learning algorithm has proved so useful that it is now considered one of the core tools used by those characterizing the behavior of the detectors.

Why are my friends seeing a different Gravity Spy workflow than I am?

Gravity Spy uses multiple workflows to help train citizen scientists in glitch classification. When a user classifies a handful of glitches correctly, they will be able to move to more difficult workflows, which have more glitch classes, more interface options, and present glitches that computers were less certain about.

What do the names of the different workflows represent?

All the workflow names represent astrophysical signals that LIGO may be able to detect. As you progress to more advanced workflows, the gravitational-wave signal that the workflow is named after can be detected by LIGO at larger and larger distances in the Universe.

Neutron Star Mountain: Neutron stars are dense, remnant cores of massive stars that are about the size of Manhattan but as massive as the Sun! They are so dense that a spoonful of neutron star material would weigh as much as an entire mountain here on Earth! Neutron stars can often spin rapidly and have very strong magnetic fields. The effect of these magnetic fields on the neutron star material can effectively 'lift' up a mountain on the surface, and as the neutron star rotates, this asymmetry creates a continuous source of gravitational waves.

Galactic Supernova: A supernova (to be more specific, a core-collapse supernova, which is the explosive death of a massive star) is an expected source of gravitational waves if the explosion happens in an asymmetric fashion. Interestingly, if the supernova explosion was completely spherically symmetric, it would not emit gravitational waves. However, simulations indicate that these explosions should have some asymmetry to them.

Neutron Star Merger: If two neutron stars (see the description above on neutron stars) are in a tight binary system, the emission of gravitational waves will cause the binary to inspiral and eventually collide. The merging of compact binary systems, such as two neutron stars, a black hole and a neutron star, or two black holes, are the main gravitational-wave target for the LIGO detectors.

Black Hole Merger: Black holes are much more compact and can be much more massive than neutron stars. The energy emitted in gravitational waves from a compact binary merger scales with how massive the objects are and how fast they are moving (which is directly related to how compact the objects are and how close they can get before touching), so the merger of black holes emits higher energy gravitational waves than the merger of neutron stars, and can thus be detected at much larger distances.

Universe Cosmic Background: The emission of gravitational waves from many astrophysical sources can create a stochastic background of gravitational waves - essentially the superposition of many events such as compact binary mergers. So far, LIGO has been able to put upper limits on such a background, but has yet to make a detection of it.

How many scientists are working on this project?

Currently, there are about a dozen scientists in LIGO, social science, and machine learning that are part of the project. Check out our ‘The team’ section, under the ‘About’ tab to learn more about the Gravity Spy team!

What is the benefit of logging in?

Logging in allows you to use the Zooniverse Talk forum, make collections of glitches, and favorite glitches that you want to look at later. Also, without logging in the Gravity Spy project cannot upgrade you to more advanced workflows, meaning you will be stuck classifying blips and whistles instead of moving on to more exciting images.

How can I change the workflow that I am in?

Click on the ‘Gravity Spy’ tab to change your workflow. If you have not unlocked all of the workflows, keep classifying!

Gravity Spy Interface FAQ

Does clicking 'Done & Talk' classify a glitch differently from clicking 'Done'?

No - the classification will be recorded the same regardless of which you click on. The ‘Done & Talk’ feature allows you to immediately access our ‘Talk’ section, where you can comment additional information about the glitch, and discuss it with other Gravity Spy users and the Gravity Spy team.

What do I do if I see more than one glitch in a single image?

If you see more than one glitch in a single image, focus your classification on the one at the center of the image (at t=0). However, certain categories, like 'Repeating Blips', 'Scratchy', and 'Paired Doves' are associated with other glitches happening nearby in time. If you zoom out and see that a glitch resembles these cases, you should take into account the glitches offset from the center of the image as well.

What do you mean by classifying something at 'time=0'?

We define time=0 to be the time when the glitch occurred. In the glitch images you will be looking at, time=0 is at the center on the horizontal axis.

How do I favorite a glitch?

At the bottom right of the image, there is a heart icon. Click this button will favorite that particular glitch, allowing you to see it again later in Talk.

How do I make a collection, or add a glitch to a preexisting collection?

At the bottom right of the image, you will see a 'list' icon (stacked horizontal lines). Clicking this icon will access your collections and allow you to add the current glitch into one of your preexisting collections, or start a new collection.

Where can I find my collections?

Under the ‘Collect’ tab there is an option to access ‘My Collections’.

Is the frequency dependence of a glitch important?

Yes. The frequency of the glitches is a key feature of which class they belong too. For example, 'Power Lines' are glitches that always occur at 60 Hz and harmonics of this frequency (120 Hz, 180 Hz, etc). Seeing a glitch with the same structure as a Power Line, but at a different frequency means the glitch is actually part of a different class. Be sure to take the frequency into account when making your classifications!

How do I get rid of a filter that I applied?

Under the classification options there is a small button that says ‘Clear filters’. Also, you can also remove each filter option individually by clicking 'Clear' after click on the filter of choice.

How do I freeze the image at a single duration rather than having the images constantly cycle?

Click the pause button in the bottom left corner. From there, you can toggle the duration of the image by clicking the circles at the bottom of the image.

Can I see more example images of a glitch category?

Yes, when you click on one of the categories you can see several example images by clicking on the circles below the image. Additionally, you can look at the Field Guide, which is a pull-out tab to the right of your screen.

What is the field guide, and how can I find it?

The Field Guide is a handbook for learning more about the glitch categories. Information about the shape, structure and frequency of the glitches, as well as information about where the glitch may originate from are all included in the Field Guide. The tab to access the field guide can be found on the right hand side of the page.

LIGO detectors FAQ

Why is there more than 1 gravitational-wave detector?

There are multiple reasons for having more than one detector. First, having two detectors is a first-order check to whether something is astrophysical rather than a noise artifact. As gravitational waves travel at the speed of light and pass straight through the Earth, we would expect a gravitational-wave signal to hit both of the LIGO detectors at nearly the same time (separated in time by less than the light travel time between the detectors, or about 10 milliseconds). Noise, however, is not expected to be correlated in time since the detectors are separated by thousands of miles. Also, having detectors in different locations around the globe allow us to pinpoint the sky location of an event more accurately. Interferometers are omnidirectional, meaning if there were only one detector we would have no idea where in the sky a signal originated. Adding more detectors to the network betters the localization of an event, and improves how confident scientists are with a detection.

Why are the detectors so big?

The strain produced by a gravitational wave is the distance change caused by a gravitational wave over some length. If you have a longer arm, the change in distance caused by a gravitational-wave event of a given strain increases. Therefore - a longer arm means a larger change in distance. LIGO is detecting changes in distances on the order of 10^-21 meters with detector arms that are 4 km long (though the laser bounces around in the arm about a hundred times before recombining, so the effective length that the light travels is about 100 times more than this).

Why are they planning on making gravitational-wave detectors in space?

LIGO is at its most sensitive in only a relatively small frequency region of gravitational waves. Just like electromagnetic telescopes that look at light, different instruments are required to observe radio waves as opposed to visible light as opposed to x-rays. LIGO looks at 'high-frequency' gravitational waves, and cannot access frequencies below 10 Hz because of the rumbling and grumbling of Earth. Space-based gravitational-wave detectors circumvent this issue by, well, leaving Earth! For example, LISA, a future space-based gravitational-wave detector, will be sensitive to much lower frequencies, meaning LISA will be able to detect different kinds of gravitational-wave sources than LIGO does, such as the merging of supermassive black holes at the centers of galaxies.

What kind of lasers do they use?

The laser that enters LIGO’s interferometers has 200 watts of power. This is 40,000 times than your typical off-the-shelf laser pointer. The laser in the interferometers has a wavelength of 1064 nanometers. This means that the laser is infrared and outside the visible spectrum of light, and if you were sitting inside the LIGO arms (please don’t), you would not be able to see the laser.

How does interference work?

Interference is a property of waves. Waves can either interfere destructively or constructively. Destructive interference occurs when the peak of one wave, meets the trough of another, causing the net amplitude of the combined wave to equal zero. Constructive interference occurs when the peaks of both waves meet, causing the net amplitude to equal the sum of the amplitudes.

If a gravitational wave stretches the distance between the LIGO mirrors, doesn't it also stretch the wavelength of the laser light?

This is an interesting question that LIGO scientists often get. A gravitational wave does stretch and squeeze the wavelength of the light in the arms. But the interference pattern doesn't come about because of the difference between the length of the arm and the wavelength of the light. Instead it's caused by the different arrival time of the light wave's "crests and troughs" from one arm with the arrival time of the light that traveled in the other arm (see the above question on interference). So the laser light is acting not so much as a ruler, but as a stopwatch.