‘A movie camera to watch the whole universe’; an artist’s conception of the LSST inside its dome. Courtesy Wikimedia
For most of human history, the distant ‘celestial sphere’ was regarded as perfect and unchanging. Stars remained in place, planets moved predictably, and the few rogue comets were viewed as atmospheric phenomena. This began to change with the Danish astronomer Tycho Brahe’s observation of the supernova of 1572 – apparently, a new star – and his studies of the Great Comet of 1577, which he proved was actually a distant object. Nonetheless, the impression of permanence is strong. There are very few astronomical objects that noticeably vary to the naked eye: only the brightest comets, novae and supernovae. For observers in the northern hemisphere, the last naked-eye supernova was in 1604.
Modern telescopic studies tell a quite different story. Today, we know of roughly a half-million variable stars in our galaxy, and identify thousands of transient objects each year. Although many stars vary in predictable ways, the Universe is also full of unpredictable violence. When two stars orbit close to each other, mass can flow from one to the other. If one of the stars is an old, collapsed white dwarf, the gas it pulls from its companion can accumulate until the dwarf undergoes a sudden thermonuclear explosion – a supernova like the one seen by Tycho. There is also another, more common type of supernova produced by the deaths of solitary stars more than about 10 times the mass of the Sun.
Supernovae show a broad range of behaviours that depend on the detailed properties of the system at the time of the final, fatal cataclysm. The atoms that emerge from supernova explosions have provided the raw material for all planets, including our own. Astronomers are understandably eager to learn more about them, but the two classes of supernovae combined happen only about once per century in our galaxy.
Obviously, for events occurring on time scales of a century, searching for them in our galaxy alone is not terribly profitable. Fortunately, our galaxy is only one of about a trillion galaxies in the visible Universe. If you monitor millions of galaxies all the time, it is possible to find many supernovae each and every day. This is one of the most exciting challenges of modern high-speed astronomy.
Other than supernovae, there are only a few variable sources luminous enough to be seen at the great distances to other galaxies, even using powerful telescopes. By far the most common is the random variability of quasars. Quasars consist of a supermassive black hole, millions to billions of times the mass of our Sun, which shine as material falls towards the black hole, heats up and radiates energy.
Today we think that essentially every galaxy contains a supermassive black hole at its centre, and something like 1 per cent of them are accreting mass fast enough to be seen as luminous quasars. The supermassive black hole at the centre of our own galaxy is essentially ‘off’. On rare occasions, though, such a black hole rapidly turns itself ‘on’. The most fascinating cause is a so-called ‘tidal disruption event’ in which a star like the Sun passes too close to the black hole and is ripped apart by the black hole’s tides. Some of the debris then falls into the black hole to power a transient flare. These tidal disruption events are far rarer than supernovae, occurring only about once every 10,000 years in any particular galaxy. In the distant Universe, the study of variability is essentially the study of black holes and supernovae.
This gives you some sense of the remarkable astronomical zoo of variable and transient objects. The challenge for the professional astronomer is to find and characterise all these different sources not only for how they work individually, but also to determine their overall demographics and statistics. To find large numbers of them, you need a big telescope that can detect the much more numerous distant, faint objects. In general, however, bigger telescopes see only smaller pieces of the sky. This frustrating rule can be bent only by spending large sums of money.
If your scientific goal is to find the largest possible number of transients, and to study their evolution across the cosmic history of the Universe, then you want to use a big telescope that covers as much of the sky as you can afford. This is fundamentally the goal of the Large Synoptic Survey Telescope (LSST). Located in Chile, LSST is (effectively) a 6.7-metre diameter telescope, scheduled to start full science operations in 2022.
LSST will be the closest astronomers have ever come to creating a movie camera to watch the whole universe. It will survey approximately half the sky using a camera that spans more than 40 times the area of the full Moon. But LSST can obtain a new image of each patch of that sky only once every three nights. LSST can detect transients 30 million times fainter than visible to the naked eye, making it a phenomenal project for finding huge numbers of faint transient sources across the visible Universe – LSST should find some 1,000 supernovae per day! But this capability comes at a cost: roughly $600 million just for construction, plus a significant operation cost as well.
At the other limit from LSST is a project I am working on: the All-Sky Automated Survey for Supernovae (ASAS-SN). By the end of this year, ASAS-SN will consist of 20 14-cm aperture telescopes spread across the globe, and costing roughly $3.5 million for both construction and operation through to 2022. With such small telescopes – big telephoto camera lenses, really – ASAS-SN can find only bright transients, roughly 25,000 times fainter than are visible to the human eye. Even so, it should still find about one supernova a day. And because ASAS-SN is comprised of small telescopes, it can image the sky far faster than LSST. The combined ‘image’ from all the ASAS-SN telescopes spans 1,600 times the area of the Moon. This allows them to survey the entire visible sky every night.
The two projects are highly complementary, essentially balancing a trade-off between ‘quantity’ and ‘quality’. LSST provides ‘quantity’: the large numbers of faint sources needed for statistical studies of distant sources, and for studying the evolution of transient sources across cosmic time. However, the typical LSST transient is faint and hard to study in detail for long periods of time, even with the world’s largest telescopes. ASAS-SN provides ‘quality’. The bright sources found by ASAS-SN are the ones that best survey the nearby Universe, and that can be studied in the greatest detail and for the longest periods of time using larger telescopes.
One of the most important tools for astronomers is the spectrum of an object: how much light is emitted as a function of its colour. A spectrum is the best way to classify the velocities, temperatures, elemental composition and type of an object (eg, which type of supernova? What were its unique properties?). Because you must chop up the light into narrow bins of colour, you need far more light to make a spectrum of an object than to get an image of it. LSST is already a large telescope, so it will be difficult or impossible to get a spectrum of the typical, faint LSST transient.
Even for the minority of LSST sources bright enough to obtain one spectrum, the source will quickly fade and become too faint to get another spectrum to study how it evolves with time. Therefore, a negligible fraction of LSST discoveries will be studied by this fundamentally important astronomical tool. The ASAS-SN transients are far fewer in number but are far brighter, so a very large fraction of ASAS-SN transients can be studied spectroscopically, and they can be studied for long periods of time even as they fade away.
Projects like LSST and ASAS-SN are continuing the revolution begun by Tycho, revealing the variable and sometimes violent events that light up the highly imperfect, ever-changing celestial sphere.