Wednesday, 29 February 2012

The structure of star clusters in the outer halo of M31

I am slowly recovering from the recent round of grant writing. The internal deadline for our Australian Research Council (ARC) Discovery Projects (DPs) was on Monday, and, as ever, it was late nights and early mornings to put together the 115 pages of science case, budgets, budget justifications, publication records, career histories etc etc etc. Now all I have to do is sit back until July (when the referees' reports come in) and then to November (when the results are announced). If I get my grant, I'll write and say how great the systems is. However, if I am the three in four who will not get funded, I'll have a rant.

Anyway, even though the days have been lost to grant writing, science has advanced, and we recently had a paper accepted for publication in MNRAS, using the Hubble Space Telescope to look at clusters around our nearest big galactic neighbour, Andromeda.

Many objects are globular clusters, just like the ones we orbiting our own Milky Way. With Hubble Space Telescope images, we can easily see the individual stars in the globular and work out how the stars are distributed. Here's an example -

In the top left, we have our Hubble picture of the sky, and the bottom left is the colour-magnitude diagram. The problem is that we have not only our globular's stars in there, but also contamination from galactic stars. The upper right are the stars identified from the colour-magnitude diagram, and the bottom right is the surface brightness distribution.

All well and good. Except, we have these in the halo of Andromeda as well
These are extended clusters (otherwise known as "faint fuzzies"). These things are weird. As you can see from the bottom right, the profile of these objects are a lot flatter, and more extended than standard globulars. Now, here's a cool plot from an earlier paper;
So, this shows the size of objects in the verses their brightness. This plot used to be quite simple with objects on the left, the globulars
and on the right are dwarf galaxies
Basically, the ones on the left are not dominated by dark matter, and have simple star formation histories, whereas those on the right have lots of dark matter and complicated histories.

But what has happened over the recent years is that the gap in between has been filled in. In the Milky Way there are ultra-faint dwarfs, thought by some (not by all) to be dominated by dark matter. Whereas, from the left, the faint fuzzies are seen not to be dark matter dominated.

It's actually all a little confuddling. Perhaps we should expect the gap to be filled, but not necessarily by two different kinds of things. And the very bizarre thing, there does seem to be any faint fuzzies orbiting the Milky Way. Why? I don't know, but we are finding the universe to be more complex than thought. Well done Nial :)

The structure of star clusters in the outer halo of M31

N. R. Tanvir, A. D. Mackey, A. M. N. Ferguson, A. P. Huxor, J. I. Read, G. F. Lewis, M. J. Irwin, S. Chapman, R. Ibata, M. I. Wilkinson, A. W. McConnachie, N. F. Martin, M. B. Davies, T. J. Bridges
We present a structural analysis of halo star clusters in M31 based on deep Hubble Space Telescope (HST) Advanced Camera for Surveys (ACS) imaging. The clusters in our sample span a range in galactocentric projected distance from 13 to 100 kpc and thus reside in rather remote environments. Ten of the clusters are classical globulars, while four are from the Huxor et al. (2005, 2008) population of extended, old clusters. For most clusters, contamination by M31 halo stars is slight, and so the profiles can be mapped reliably to large radial distances from their centres. We find that the extended clusters are well fit by analytic King (1962) profiles with ~20 parsec core radii and ~100 parsec photometric tidal radii, or by Sersic profiles of index ~1 (i.e. approximately exponential). Most of the classical globulars also have large photometric tidal radii in the range 50-100 parsec, however the King profile is a less good fit in some cases, particularly at small radii. We find 60% of the classical globular clusters exhibit cuspy cores which are reasonably well described by Sersic profiles of index ~2-6. Our analysis also reinforces the finding that luminous classical globulars, with half-light radii <10 parsec, are present out to radii of at least 100 kpc in M31, which is in contrast to the situation in the Milky Way where such clusters (other than the unusual object NGC 2419) are absent beyond 40 kpc.

Thursday, 23 February 2012

Citations to Australian Astronomy: 5 and 10 Year Benchmarks

Not everything that can be counted counts, and not everything that counts can be counted.
Albert Einstein, (attributed)
I don't know how widely it is known outside academic circles, but researchers these days are surveyed and counted continuously, in an effort to show that tax-payers money is being spent on research excellence and research impact. A number of countries have held such exercises, and here we are into the second round of the Excellence in Research for Australia.

Such exercises have real impact, as some at The University of Sydney just found out; those not producing enough "research outputs" are in the firing-line for redundancies.

With the growth of online databases of papers and citations, it's now easy to get an assessment of someones research output, and with things like Google Scholar it's all nicely displayed; here's mine. People's careers get wrapped into a few metric, and the principle one is the h-index.

The idea is simple. Each paper you write is cited by others, and when you rank your papers in order of the number of citations it gets, your h-index is the number of the paper with the same number of citations. As an example, my h-index is 50, and so I have 50 papers with more than 50 citations.

This is taken to be a measure research impact. The more active you are, the more papers you produce. But you need people to read and cite them (this is the impact part). The bigger the h-index, the "better" the researcher (for some vague definition of better).

Of course, many people hate this kind of ranking system as many factors can affect your h-index, such as the field you work in, being ahead of your time etc. But it is common for people to look at h-indicies of similar people at similar level of their careers, at least as a starting point of assessing research impact.

Such ranking exercises have become a subfield, and a recent paper appeared which looked at the impact of Australian astronomers (in terms of their h-index and citations). Here's the abstract:

Citations to Australian Astronomy: 5 and 10 Year Benchmarks

Expanding upon Pimbblet's informative 2011 analysis of career h-indices for members of the Astronomical Society of Australia, we provide additional citation metrics which are geared to a) quantifying the current performance of b) all professional astronomers in Australia. We have trawled the staff web-pages of Australian Universities, Observatories and Research Organisations hosting professional astronomers, and identified 383 PhD-qualified, research-active, astronomers in the nation - 131 of these are not members of the Astronomical Society of Australia. Using the SAO/NASA Astrophysics Data System, we provide the three following common metrics based on publications in the first decade of the 21st century (2001-2010): h-index, author-normalised citation count and lead-author citation count. We additionally present a somewhat more inclusive analysis, applicable for many early-career researchers, that is based on publications from 2006-2010. Histograms and percentiles, plus top-performer lists, are presented for each category. Finally, building on Hirsch's empirical equation, we find that the (10-year) h-index and (10-year) total citation count T can be approximated by the relation h = (0.5+sqrt{T})/sqrt{5} for h > 5.
One of the things they do is produce ranked lists of the "top" astronomers in Australia. Here it is in terms of h-index.

I'm not going to comment too much about this, but it's nice to be on the list, and I'm in excellent company also.

Right, back to grant writing!

Monday, 20 February 2012

First Galaxies and Faint Dwarfs

Back in Oz after a week in the US at the University of California at Santa Barbara (UCSB). I was at a conference called First Galaxies and Faint Dwarfs at the Kavli Institute of Theoretical Physics.

The week started with us being told that this was going to be the best conference we would ever attend, and I must agree that it was. Instead of the usual barrage of 15min talks, we had 40mins of review talks, with discussion sessions, and it was great.

What was the point of the meeting? Well, that was summarized by the opening slide by Leon Koopmans;

Essentially, the goal was to bring together two communities, those that study the high redshift universe and the first forming galaxies, and those that look at the tiny dwarf galaxies in our local universe. Why? Because understanding galaxy formation and the links between the nearby and faraway will reveal the inner workings of the universe, especially the nature of dark matter.

I'm not going to go through the conference in detail (as I have to get to work), so here's few slides. All the talks are public, and there are podcasts so you can hear the discussions, and I recommend that you take a looksie.

Notice what there is not. There is not wailing and gnashing of teeth with screams of "oh what is dark matter, we don't know what it is, oh this is a disaster, ohhhh, woe is me". No. That is not where we are in the subject. The point is that we know dark matter (or, with a much lower prior, modified gravity) is out there, and we want to know what the consequences are.

So, what are the big questions. Well, they may seem quite technical, but  it is where the front lines of cosmological science currently are.

  • Where are the dark matter subhalos? Basically, our picture of the large scale structure of the Universe, one driven by dark matter and dark energy, works really well in predicting where galaxies should be. But on small scales, it predicts that a galaxy like the Milky Way should be surrounded by a host (thousands of..) dwarf galaxies. But we only see a few 10s of these little things out there. So where are they? If you believe Carlos Frenk, the problem is solved - the first burst of star formation in the universe blew the gas out of the halos, and so they could not form stars. The basic result of this would be that we surrounded by dark halos, loads of them. But how can you see if their there? Well, we can use gravitational lensing, and we're working on it. Just as an example, here is a simulated gravitational lens, followed by an actual lens, the Horseshoe. By studying the details, we'll be able to measure the the amount of subclumps of mass out there. Cool eh?
  • Are dark matter halos cored or cusped?  This is quite technical, but our cosmological models make some firm predictions on the distribution of mass in galaxies, predicting that the mass density should increase rapidly towards the centre. This peaky mass distribution is called cusped and so if we can measure it, we can directly test whether our models are correct. If the mass does not rise so quickly (a cored distribution) then this would be a bit of a problem. Except, things are not so simple, as the existence of baryons (to you and me, basically means gas and atoms and stars and stuff) can apparently change cusps into cores. But gas physics is hard, especially including them in the complex computer simulations of structure formation in the Universe. If halos are cored, rather than cusped, then it might means that dark matter is not cold, which brings me to the next question.
  •  Is dark matter hot, warm or cold? Is dark matter cold (i.e. composed of things moving slowly) or hot (things moving relativistically), or somewhere in between (i.e. warm)? This is related to the above two questions, as warm dark matter doesn't form the little halos we expect around the Milky Way, and is likely to influence the core/cusp question. Currently, we don't know.
  • Are dwarf galaxies we see today like the dwarf galaxies that formed the Milky Way? Again, this is a tricky one. The Milky Way was built of little galaxies that fell in. But are the ones falling in now like the ones that fell in long ago? The answer is "we don't really know", but there are some clues, namely coming from the mass-metallicity relationship, which means that smaller galaxies have less metals (which, to an astronomer, is all elements heavier than helium. But stellar chemistry is quite complex, and the observations challenging, but there is some really good ground being made in this area.
There are more, of course, but these are the key questions. At the meeting we had the usually argy-bargy about theorists and observers not believing or trusting each other, but it's all part of the fun.

On a closing note, I've now visited three Kavli Institutes around the world, the one in Santa Barbara, one in Bejing and another in Cambridge. These are wonderful research environments, with great programs bringing people together. But similar to the Perimeter Institute in Canada, and the Keck Telescope, these are funded by private endowments, rather than being solely funded by government.  Any Australian philanthropists out there, it would be nice to have a similar institute here :)

Friday, 10 February 2012

Mapping Growth and Gravity with Robust Redshift Space Distortions

A quick post this evening, as I have been at a workshop for the SAMI instrument, and am off to Santa Barbara for the First Galaxies and Faint Dwarfs: First Galaxies and Faint Dwarfs conference next week, but a couple of things to post. The first is a paper by my ex-phd student, Juliana Kwan, who is now a postdoc in the US at the Argonne National Laboratory.

The paper is quite complex, and focuses on redshift space distortions. This can be difficult to understand, but here goes. We've mentioned a couple of times that matter in the Universe is arranged on a cosmic web, with clusters, clumps, filaments and voids. In fact, it looks something like this:
Our Milky Way galaxy is just a little dot in there. But the detail of the way the mass is distributed is a probe of our Universe, as its present structure carries the imprint of the forces that created it, including the make up of the Universe, the cosmic evolution, and even the nature of gravity itself.

What do we see when we look out into the Universe? Well, we can measure the position to a galaxy on the sky very accurately, but distance is not. But we can easily measure the redshift, or the amount features in the spectrum are moving to longer wavelength, and use our cosmology to turn this into a distance using the famous Hubble law.

However, there is a problem. The redshift we see is a mixture of two parts, one due to the cosmic expansion (the Hubble law bit) and one due to the `peculiar velocity, or how much the galaxy is whizzing about. By comparing to the Microwave Background, we know that our Milky Way is moving with a speed of about 500 km/s.

As we measure redshifts, not distances, these peculiar velocities distort the distances we calculate via the Hubble law. So, this happens
The blue on the right is the actual positions of galaxies in the cosmic web (in a simulation of the Universe). The green on the left show the effects of peculiar velocity, and things are stretched and squished from the space position.

In fact, clusters of galaxies, where velocities are typically several thousands of km/s, get stretched out into what are known as Fingers of God - although what they have to do with the Higgs boson, I don't know (and no, that's not a serious statement). Here's a real set of observations;
Now, the details of these Redshift Space Distortions allow us learn even more information about the Universe, but it is very hard to untangle. What Juliana's paper does is to look at the possible ways that can be used to extract science, and shows what needs to be done if you want to get "robust" measures. I'll write more on on what robust means later, but for now, I'll finish by saying "Well done Juliana!"

Mapping Growth and Gravity with Robust Redshift Space Distortions

 Juliana Kwan, Geraint F. Lewis, Eric V. Linder
Redshift space distortions caused by galaxy peculiar velocities provide a window onto the growth rate of large scale structure and a method for testing general relativity. We investigate through a comparison of N-body simulations to various extensions of perturbation theory beyond the linear regime, the robustness of cosmological parameter extraction, including the gravitational growth index, \gamma. We find that the Kaiser formula and some perturbation theory approaches bias the growth rate by 1-sigma or more relative to the fiducial at scales as large as k > 0.07 h/Mpc. This bias propagates to estimates of the gravitational growth index as well as \Omega_m and the equation of state parameter and presents a significant challenge to modelling redshift space distortions. We also determine an accurate fitting function for a combination of line of sight damping and higher order angular dependence that allows robust modelling of the redshift space power spectrum to substantially higher k.

Sunday, 5 February 2012

Brydon and the Royal Australian Observatory

Now, I love a good comedy (and a few bad ones) as much as the next person, but it's grant writing time, and so I am going to be grumpy here. One thing that gets me hot under the collar is terrible portrayals of science, especially astronomy and astrophysics, in the media. Which brings me to exhibit number 1. The TV comedy, Supernova, starring welsh comedian, Rob Brydon.

The story is simple. Brydon plays Dr Paul Hamilton, who moves to work at the fictional Royal Australian Observatory, and the laughs come thick and fast in the usual "fish out of water" situations. Well, they don't really, the series was a flop, and lacks a lot of laughs,  but, amazingly, a second season was made.

I don't know how I missed this when it was originally shown in 2005/2006, but its portrayal of astronomy and astronomers is, well, rubbish.

I would write pages and pages on this, but let's start with some basic things.

The Observatory: You can tell by the image above, the Royal Astronomical Observatory is sitting there, out in the Outback, in the middle of nowhere (well, close to Broken Hill, where Mad Max 2 was filmed). I'm sure this was to get an Australian "feel" to the show, and make the fish-out-of-water more fish-out-of-watery. Now, you might want to build a radio telescope out there (and that's what they have done at Narrabri, although not so deserty). There is a real optical observatory in Australia, at Siding Spring, not that far from Broken Hill. What does that look like?
It's up a mountain, where you expect to find an optical observatory. Clearly, it's a bit too green for a fish-out-of-water welsh astronomer to be in. But it is what an Australian optical observatory looks like. And there are aboriginals living there too (but, of course, people want to see aboriginals living like wild warriors out in the desert rather than think about the real challenges facing these people).

The Telescope: This is the telescope at the Royal Australian Observatory (in the TV show).
The telescope is puny and many amateurs would have better. This is the mighty Anglo-Australian Telescope
Notice the scale on this image. I remember the first time I went into the dome, I thought I was in a James Bond set. It's huge, it's impressive, it's a real telescope! The show makes another cardinal sin with regards to astronomy. Astronomers don't sit around in the dome with lights blazing, computers humming etc when they are observing. The telescope is in total darkness, away from sources of light which would cause the air to shake and blur the images. The astronomers sit in control rooms, away from the telescope, often with music thumping to keep them awake. Given the nature of the site of the Royal Australian Observatory, a crappy telescope with lights on would seem to fit.

Doing science: Now, this is the bit that really drove me crazy. The show is filled with sciencey buzzwords, and Rob Brydon's character, an expert on gas, spouts all kind of stuff. But their crappy telescope can so magic, it can see round corners and to see patches of gas at the birth of the Universe. The notion of writing a paper, the staple of research, is all very bizarre, as is giving presentations etc. I don't know what more I can say about this, it just is nothing like real science.

OK, OK  - I know it's a comedy show. I know I should not be so critical. But seeing that they were filming in Sydney, which has a high density of astronomers, perhaps they could have asked for some advice to look slightly like a real astronomer. Heck, they could have even filmed up at the AAT, and had a realistic telescope dome, one more impressive than the one they showed. It may not have helped the laughs, but it would have at least looked slightly more "scientific".

As an aside, I am a fan of Rob Brydon. In fact, he was raised in Baglan, next to Neath, the town I was born in. And as a link to science fiction, Baglan is part of Port Talbot, home to the Port Talbot Steel Works.

which was apparently the inspiration for Ridley Scott's vision of a future city in Blade Runner. I've see it at night, it's a startling sight!

The cosmic history of the spin of dark matter haloes within the large scale structure

What's the goal of astronomy and astrophysics? Pretty Hubble pictures? Whacky statements about multi-verses and black hole entropy? Nope, not really. The goal is to understand the Universe in terms of the laws of physics.

Like "real" science, there is a strong link between theory and observation and experimentation, each feeding off each other, and as I noted a little while ago, the development of new instrumentation on telescopes often opens a new window on the Universe. Instruments like SAMI will let us do something new, namely measure the spins and chemistry of a large number of objects.

But now the key question - If you were going to measure the spins of a whole load of galaxies, what would you expect to see? Would they be randomly orientated? Would patches of galaxies spin together in unison? Would nearby galaxies spin in opposite directions? And how would you even begin to answer this question?
There are paper-and-pen approaches to answering this question, basically known as tidal torquing theory, which basically says that as the mass that forms galaxies starts to collapse, nearby masses tug unevenly on each other, causing them to start to spin.

The problem is that we know that galaxies don't simply form from essentially isolated collapsing masses. It is a lot messier, with galaxies crashing together, and small systems being cannibalized by larger galaxies, and what we end up with at the end of the day is a cosmic web!
We see clusters (bright yellow), dark voids and filaments connecting it all. Just how are galaxy spins aligned in this mess?

Well, you can just measure it (and the just in there does some serious disservice to how tricky this really is to do). You can find the collapsed masses, and look what they are doing and where they live, and measure what their spins are.

Minor interlude: Here is a fantastic little intro from Andrew Pontzen on what we actually do when we mean galaxy here.

PhD student, Holly Trowland, Joss Bland-Hawthorn and myself have just submitted a paper to The Astrophysical Journal doing such spin measurements on cosmological simulations, finding that there are a whole range of correlations of the spins with environment, with mass and over cosmic history. In fact, it shows that if you simply take the paper-and-pen tidal torque theory, it just fails when you get into the mess of the cosmic web.

One issue is that our simulations have been looking at only dark matter, the dominant mass in the Universe. But galaxies are made of atoms and we need to know how these spin within the spinning dark matter halos. This is trickier than it sounds, but Holly is working on it, and we should have some results later in the year. But for now, well done Holly!

The cosmic history of the spin of dark matter haloes within the large scale structure

Holly E. Trowland, Geraint F. Lewis, Joss Bland-Hawthorn
We use N-body simulations to investigate the evolution of the orientation and magnitude of dark matter halo angular momentum within the large scale structure since z=3. We look at the evolution of the alignment of halo spins with filaments and with each other, as well as the spin parameter, which is a measure of the magnitude of angular momentum. It was found that the angular momentum vectors of dark matter haloes at high redshift have a weak tendency to be orthogonal to filaments and high mass haloes have a stronger orthogonal alignment than low mass haloes. Since z=1, the spins of low mass haloes have become weakly aligned parallel to filaments, whereas high mass haloes keep their orthogonal alignment. This recent parallel alignment of low mass haloes casts doubt on tidal torque theory as the sole mechanism for the build up of angular momentum. We find a significant alignment of neighboring dark matter haloes only at very small separations, r<0.3Mpc/h, which is driven by substructure. A correlation of the spin parameter with halo mass is confirmed at high redshift.