The Crowd and the Cosmos: Adventures in the Zooniverse

by Lintott, Chris

  improve the efficiency of citizen science projects. As well as pay-

  ing different levels of attention to individuals based on their per-

  formance, we can start thinking about how to combine human

  and machine classification, or even how to manipulate what

  people see to make them more likely to stay on the site. As you

  might imagine, this gets tricky quickly, and to think about the

  possibilities we need to return to space, and to looking up at the

  night sky.

  6

  FROM SUPERNOVAE

  TO ZORILLAS

  One of the things I like about Oxford is that the sky is still pretty dark. I don’t live far from the centre, and a relatively

  benign street-lighting policy means that as I wheel my bike into

  the garden after a late night in the office (or the Lamb and Flag) I

  can look up and see the stars. Partly, I like being able to mark the

  passing of time in the changing display of constellations, but I’m

  looking too for a glimpse of unchanging infinity. It’s nice to be

  assured that whatever crises are happening here on Earth, the

  vastness of the Universe is there, beautiful and silent and unchan-

  ging. It’s clearly not just me that feels like this; from Immanuel

  Kant, filled with ‘admiration and awe’ by the starry heavens above

  him, through Walt Whitman’s protagonist who braves the ‘mys-

  tical moist night air’ to look up ‘in perfect silence at the stars’,

  plenty have gazed on the sky for a bracing dose of cosmic per-

  spective.

  The effect isn’t limited to looking at the sky directly, either. A

  good 8 per cent of respondents to the first survey we carried out

  of Galaxy Zoo volunteers said they were participating in the pro-

  ject because they enjoyed the opportunity to ‘think about the

  154 From Supernovae to Zorill aS

  vastness of the Universe’. The bad news is that modern astro-

  physics is hell bent on making it clear that the unchanging, ever-

  enduring sky is anything but that. It changes, and increasingly

  astronomers are paying attention to this changing sky, an atti-

  tude which is bringing exciting discoveries. Whereas in the twen-

  tieth century hacking at the frontier of observational astronomy

  meant stretching to new wavelengths, to radio astronomy, and to

  high energies, in the twenty-first it means paying attention to

  how things change with time. This would have surprised our pre-

  decessors. Of course, a few things have always been known to

  change—the phases of the Moon, the positions of what the ancient

  Greeks called ‘wandering stars’ which we now call planets—but at

  least regular and increasingly predictable patterns could be dis-

  cerned. The idea of sudden, violent change would have been

  deeply shocking; it seems to have been at least one reason that

  comets, appearing without warning and then vanishing again,

  were seen as omens.

  The shock must have been worse when apparently simple,

  straightforward, and familiar objects misbehaved. Comets belong

  to the Solar System, and whizz past with their message of doom,

  but supernovae are another thing altogether. They appear as new

  stars, shining brightly for a period of a few weeks or months

  before fading forever. In May 1006, for example, the otherwise

  obscure constellation of Lepus—the hare at Orion’s feet—

  suddenly boasted a bright new jewel. Bright enough to be easily

  visible in daylight, it was recorded by observers in Europe, China,

  and Japan.

  For a few short months it was the brightest thing in the sky

  other than the Sun and Moon, exceeding even the most brilliant

  apparition of Venus. Similar events were noted in 1054, in 1572,

  and in 1604. This last event was observed by Kepler, who reached

  for an explanation in terms familiar to anyone at the time,

  From Supernovae to Zorill aS 155

  pointing to the event as a possible analogue of the star that the

  Bible tells us shone over Bethlehem.

  That 1604 supernova is still the most recent to have been

  observed within the Milky Way; centuries of instrumental advance

  since the invention of the telescope have been denied the chance to

  study such an event close up. The closest we’ve come is a 1987 event

  in the Large Magellanic Cloud, the largest of the Milky Way’s sat-

  ellite galaxies, which was brilliant enough to be seen with the

  naked eye and was observed with every instrument available.

  Supernovae are, though, powerful enough to be seen in dis-

  tant galaxies. Most such events mark not the birth of a new star,

  but the death of an old one much more massive than the Sun.

  Most stars spend most of their life converting hydrogen to

  helium, before moving through stages where they burn heavier

  and heavier elements, all the way up to iron for the most massive

  stars. When such a star runs out of fuel at its core, or when the

  temperature is insufficient to start the next set of reactions, the

  nuclear fusion that has sustained it ceases. This is a problem as

  the light produced in normal nuclear reactions will stream out-

  wards, encountering atoms in the stellar atmosphere and produ-

  cing a pressure which supports the star’s outer layers.

  Once this radiation pressure which normally prevents gravity

  from taking its course vanishes, the star collapses in on itself.

  This dramatic event can liberate enormous energies; a typical

  supernova shines with a hundred billion, billion, billion, billon

  watts. I don’t have a good comparison for this, but it’s enough to

  match the power produced by all of Earth’s power stations for at

  least a couple of billion years, or by more than a billion billion

  suns; at its peak a single supernova will easily outshine the col-

  lective might of an entire galaxy’s stars.

  It’s this brightness that means that supernovae can be seen

  from distant galaxies, and that makes them extremely useful.

  156 From Supernovae to Zorill aS

  One particular kind of supernova, which astronomers call ‘type

  1a’ supernovae, can be used to measure the expansion of the

  Universe itself. These rare events have a slightly more compli-

  cated story than most, being the product of two stars rather than

  one. The most likely scenario for the formation of a type 1a

  supernova starts with a binary system consisting of a pair of

  Sun-like stars.

  Normally, stars the size of the Sun don’t go supernova. When

  hydrogen is exhausted at their cores, they can begin burning

  helium. When this happens, the equilibrium between gravity

  and the pressure pushing outwards is disturbed, and the star will

  expand to become a red giant. The core’s helium too will eventu-

  ally be exhausted, and the star will shed its outer layers, produ-

  cing at the end of its life nothing more than a transient, if

  beautiful, planetary nebula from the gas lost in these outer layers

  and the dense, cooling remnant of the core known as a white

  dwarf. Left to its own devices, such a relic would cool slowly;

  nuclear reactions will have ceased.
<
br />   In a binary system, a more interesting future is possible. The

  more massive of the two stars will undergo evolution as normal,

  ending up as a white dwarf. If the second star enters a red giant

  phase and is sufficiently close to the white dwarf, material may

  be pulled by gravity from the still-shining red star onto the sur-

  face of its dead companion.

  At first, not much happens, but as the mass of material accu-

  mulating on the surface of the white dwarf increases there will

  come a point where it will reignite, and the star will, briefly, shine brightly once more. This isn’t a slow and stable process, but

  rather a chain reaction in which all the accreted material rapidly

  becomes involved. From almost nothing, the star will shine more

  brilliantly than it ever has before, appearing just for a short time

  as a spectacular supernova.

  From Supernovae to Zorill aS 157

  What’s more, the timing of this runaway nuclear reaction

  will mostly depend on how much fuel has built up; there will

  be a common mass required to ignite new reactions whenever

  and wherever in the Universe this happens. That means that

  whenever a type 1a supernova explodes, it will do so with the

  same brightness, shining out into the Universe with a standard

  luminosity.

  In the few sentences above, I’ve ridden roughshod over about

  twenty years’ worth of work of many of my colleagues. While

  type 1a supernovae are always roughly the same luminosity,

  exactly how bright any particular supernova will be depends on

  many different factors, including the composition of the material

  involved. Luckily, by looking at the details of how the supernova

  brightens and fades, we can adjust (we think!) for most of these

  effects.

  Astronomers like such objects. If you know how bright some-

  thing really is (how luminous it is), and can measure how bright

  it appears, then you have a measure of distance. If I told you that

  I was holding a lamp containing a sixty-watt light bulb, you’d be

  able to hazard a sensible guess as to whether it was a metre or a

  mile away from you. Such ‘standard candles’ are the main tools

  by which we measure distances; the Cepheids discussed in

  Chapter 2 are a famous example. Type 1a supernovae are merely

  the latest in a long line of such objects, but they are especially

  valuable because of their great luminosity. Once astronomers

  had captured enough of them it turned out they had a surprise

  for us.

  Two big research efforts in the 1990s set out to systematically

  discover and then follow up these supernovae for cosmological

  purposes. One was led by Saul Perlmutter in Berkeley, and the

  other by Adam Riess at Johns Hopkins and Brian Schmidt at the

  Australian National University, though like much of today’s

  158 From Supernovae to Zorill aS

  astrophysics there were large teams at work in both cases. The

  idea was to measure enough distant type 1a supernovae to get a

  sense of how the expansion of the Universe was changing.

  Hubble’s constant—the speed of expansion of the Universe—

  isn’t really a constant. The speed of expansion has been changing

  since the beginning, affected, for starters, by gravity. The pull

  of the matter in the Universe acts as a drag on the universal

  expansion, and so both teams of supernova hunters expected

  to be measuring a slowing down of this expansion.* Instead,

  they found a remarkable, deceptively simple result. To their

  great puzzlement and surprise, both groups found that many

  of the type 1a supernovae were fainter than expected. Stranger

  still, the effect increased with distance; the further away a

  supernova was, the fainter it was compared to the expected

  brightness.

  This sounds like it should be some systematic effect, rather

  than anything real. For a while, plenty of people believed that the

  results could be explained by light being affected by its passage

  towards us, perhaps absorbed by dust, but observations of super-

  novae at different wavelengths ruled that out. Remember, too,

  that when we look deeper into the Universe, we’re looking back

  in time, covering billions of years of history. Maybe type 1a super-

  novae were different back then. Certainly, the stars which create

  them will have been different from those around us today, hav-

  ing formed from pristine material produced in the first few min-

  utes after the Big Bang which is almost entirely hydrogen and

  * The technical details don’t matter too much here, but the secret is to measure the distance to the galaxy hosting the supernova in two ways. Measuring the apparent brightness of the supernova is one way, but we can also measure the redshift in the light from the galaxy caused by the expansion of the Universe while it was en route. Combining the two for supernovae at different distances gives a measure of how the expansion rate is changing.

  From Supernovae to Zorill aS 159

  helium; the heavier elements are all produced in stars and mixed

  back into their surroundings upon the death of the star. Early in

  the Universe’s history there just hadn’t been much time for stars

  to live out their lives and then die, and therefore pollutants such

  as carbon and oxygen are relatively rare.

  We (more or less!) understand stars, though, and how this

  change in composition would affect them. No matter how people

  tried, it didn’t seem like astronomers could get away with blam-

  ing the unexpected results on stellar properties. Lots of careful

  work instead led cosmologists to a more radical conclusion. The

  supernovae, it seems, are fainter than expected because they

  really are further away than expected. The expansion of the

  Universe has sped up during the time that light has been travel-

  ling from these distant events to Earth. Instead of the universal

  expansion slowing down under the influence of gravity, it is

  speeding up.

  To say this is confounding is to understate the case. We need

  some sort of ‘anti-gravity’ capable of acting on the largest of

  scales and speeding up the expansion. Worse, the measurements

  from large surveys of type 1a supernovae, alongside many other

  strands of evidence suggest that this anti-gravity is utterly dom-

  inant, accounting for about 70 per cent of the energy (more tech-

  nically, the energy density) in the Universe today. This accelerating force has come to be known as ‘dark energy’, presumably because

  we need a name as confusing as the thing itself (I don’t know

  what it means for energy to be ‘dark’), and understanding it is, in

  my opinion, the largest outstanding problem in physics.

  I admit that the quest to reveal the nature of dark energy is a

  long way from everyday concerns, but the entire fate of the

  Universe depends on it. Its influence increases as the Universe

  gets larger, which means that because of dark energy the accel-

  eration we observe today will continue. Rather than collapsing

  160 From Supernovae to Zorill aS

  back in on itself towards a Big Crunch,* or slowing its expansion

  as it approaches so
me maximum size, the Universe will continue

  expanding forever.

  While this seems cheering at first, what we face on a cosmic

  scale is a long, sober, and increasingly dull retirement. More stars

  are now dying than are being born each year, and as the expan-

  sion continues we will eventually reach the moment when the

  last star is born. All that remains after that is the long, slow dying of the light as first the most massive and then the smaller stars die one by one, their lights extinguished. Our Universe’s far future

  contains nothing more than an ever-expanding sea of radiation.

  Worse, from an astronomer’s (admittedly long-term) point of

  view, is the fact that the acceleration due to dark energy pushes

  objects over the cosmic horizon; the proportion of the Universe

  which we can see is decreasing due to the increased rate of expan-

  sion.† So what is going on?

  We don’t know. Until the discovery of dark energy, physicists

  had been content with four fundamental forces (gravity, electro-

  magnetism, and the strong and weak nuclear forces), but none

  can explain accelerated expansion. A natural candidate for a

  repulsive force emerges from some of the ways that quantum

  physics describes the behaviour of empty space, but this ‘vacuum

  energy’ has to either cancel out or else have a minimum strength

  around a thousand billion, billion, billion, billion, billion, billion, billion, billion, billion, billion, billion, billion, billion times

  stronger than we observe. Unless something’s wrong with the

  basics of quantum theory, we can rule this out as a source of dark

  energy by observing that the Universe has yet to tear itself apart.

  * I’ve always liked Douglas Adams’ suggestion that such an event should be called a Gnab Gib, as it’s a Big Bang but backwards.

  † This is one reason why it’s important to invest in astronomical research right now.

  From Supernovae to Zorill aS 161

  Faced with celebrating this, the largest error ever achieved by

  science, my theoretical colleagues have not been idle. There is no

  broadly accepted theory (it would be cruel but not that inaccur-

  ate to say there are few detailed theories that have more than one

  or two adherents), but there are plenty of ideas. The flourishing

  of creative ideas in response to a mystery is heartening, and

  browsing the research journals you can take your pick from ideas