Simulation shows Milky Way's quirks

It may be our home but just how special is the Milk Way?

That’s the question a team including Oxford University scientists have been looking to answer using simulations of our galaxy and our neighbours, the Magellanic Clouds.

Their findings, reported in a paper in The Astrophysical Journal could help in the hunt for dark matter. I asked one of the paper’s authors, Phil Marshall of Oxford University’s Department of Physics, about Universal assumptions, starless galaxies, and telltale gamma rays…

OxSciBlog: What made cosmologists assume that the Milky Way is an 'ordinary' galaxy?
Phil Marshall: Basically, we had to start somewhere! The cosmological principle states that we do not live in a special place in the Universe, one that has a special viewpoint. Asserting this principle allows us to make many wide-ranging inferences about the Universe, even though we can only observe it from one location (Earth). But it's important to test our assumptions, so we asked whether the galaxy we live in was, in fact, special - at least in one respect.

OSB: How can the Magellanic Clouds reveal if the Milky Way is special?
PM: A galaxy's neighbours - its ‘satellite galaxies’ - are one of its observable features. We wondered if having these two very nearby neighbours, the Magellanic Clouds, made the Milky Way special.

So we looked in the Sloan Digital Sky Survey [SDSS] sky survey at thousands of galaxies that have the same brightness as the Milky Way, and asked how many of them have two nearby neighbours like our Magellanic Clouds. It turns out that only about 4% of them do - so the Milky Way is a little unusual, but not very unusual. It's a one-in-twenty-five galaxy, rather than one in a million.

OSB: How did you use simulations to see how the Milky Way relates to its neighbours?
PM: We did the same thing in a simulated sky survey, counting neighbouring objects around Milky Way-like objects. If the simulated Milky Way galaxies don't have as many satellites as the SDSS galaxies, then the simulation needs more work.

We used a simulation called ‘Bolshoi’ that followed the formation of about 100,000 galaxies, and picked out the ones that were about as bright as the Milky Way. This is tricky to do actually, because the simulated galaxies don't have any simulated stars in them! They are just dark matter ‘halos’ - blobs of dark matter that would contain gas and stars in real life. The simulation doesn't include stars and gas, because it's too difficult to simulate them. Dark matter structures are easier to model - for them, it's only gravity you have to understand, and not the complicated physics and chemistry of how stars are made.

What we do is match the simulated dark matter halos to the real SDSS galaxies, one by one, most massive halo to most luminous galaxy and so on. You end up with a model Universe full of dark matter halos with bright galaxies ‘painted on’ - and it turns out this painted Universe looks very similar to the real one indeed. Then we can select all the model galaxies that are as bright as the Milky Way, and count their neighbours.

OSB: What can you infer about how 'odd' our home galaxy is?
PM: We found that, just like in the real Universe, Magellanic Clouds occur in about 5% of Milky Way galaxies. So the simulation matches the SDSS sky survey very well, right down to the smallest galaxies it contains, the Magellanic Cloud-like satellites.

Actually we can say quite a lot about our home galaxy without doing all the matching I just described: If we just look in the simulation for halos that have 2 subhalos that are the same mass as the Magellanic Clouds, and that are at the same distance from their host galaxy as our Magellanic Clouds are from us, and that are moving at the same speeds as our Magellanic Clouds are, we can collect a group of model halos that really resemble our own halo very closely.

We call these halos ‘analogs’, and they show us some possibilities for what our own dark matter halo is like. For example, they weigh about a trillion solar masses each, so we can say that this is probably what our halo weighs. Likewise, looking at the formation histories of each our analogs, we can infer that our Magellanic Clouds probably arrived quite recently (within the last billion years), and they probably arrived together.

OSB: How might such simulations help in the hunt for dark matter?
PM: Understanding the distribution of dark matter in our own galaxy is very important, especially when searching for the very faint glow expected if dark matter turns into something else.

The idea is that dark matter particles in our galaxy could, very occasionally, collide with each other, and ‘annihilate’, in a very faint flash of gamma rays. These flashes may be so faint that knowing where the dark matter is likely to be, ahead of time, from its gravity, would really help in interpreting the gamma rays that telescopes, like Fermi, detect.

Dr Phil Marshall is based at Oxford University’s Department of Physics