I think I hear a clock
I can remember when a 100x100 particle matrix motion program ran for 39 hours....and now they can simulate entire galaxies...
A group of astrophysicists from the Harvard-Smithsonian Center for Astrophysics and the University of Wisconsin-Madison say they've resolved a long-standing question: how long do spiral arms in galaxies like our own last? The boffins aren't just thinking about our very own Milky Way: the paper, published in The Astrophysical …
When I was kid I depended on a kaleidoscope to see stuff like this
To see anything remotely interactive I'd head off to the planetarium, which has now been almost killed off by Google Sky and other similar applications, like Starwalk for tablets.
It's awesome to watch rendering sequences like this, it shows that technology brings science to the masses via mediums like YouTube.. and let's not overlook the programming that went into the production
And to think my kids believe I spend all my "Internet time" on YouTube time watching cheesy 80's musics videos...
More at the Galaxy Zoo
This actually reminded me that I have This Olden Book From 1996 lying around. So long ago... when EDO RAM was fresh...
While this is great, and I will admit right away I have not had time to read up on the full method, there are more things than stars in galaxies were these included in the simulation.
Also they put up information about transparency and changing it on the particles the further from the centre they moved. This seems to me a bit like cheating. To be honest, if they did not vary the transparency it looks to me like it would still be a disc rather than a spiral.
Is this due to the time frame this occurred over and then extrapolating the data? Please someone in the know put me out of my misery. Okay I actually have to work and don't have time to read up on it, but I will look later.
Icon for the hopefully technical answer.
The transparency was explained in the vid: "The variable transparency allows for the spiral structure to be easily observed." Or in other words, if they didn't make the inner particles (stars) more transparent than the outer ones the middle would be just one big white mess and no structure would be visible.
> there are more things than stars in galaxies were these included in the simulation
From the paper:
The galaxies in our study consist of dark matter halos and rotationally supported disks of stars. The parameters describing each component are independent and the models are constructed in a manner similar to the approach described in previous works (Hernquist 1993; Springel 2000; Springel et al. 2005).
2.2. Dark Halo
We model the dark matter mass distribution with a Hernquist (1990) profile:
rho_dm = M_dm/ (2*Pi) * a / (r * (r + a)³)
which has a cumulative mass distribution M (< r) = M_dm * r² / (r + a)² , where a is the radial scale length and M_dm is the total halo mass here set to 9.5 * 10¹¹ M_solar.
In the past, models of disk galaxies run in isolation and used to study the properties of spiral arms employed only a few million particles to sample both the stellar disk and the dark matter halo. In such experiments, randomly-placed particles produce fluctuations in the halo potential. Even if the disk is initially featureless, the Poisson noise owing to such discretization of the mass in N-body experiments is inevitably swing amplified, producing trailing multi-armed spiral patterns in the disk (Toomre 1977; Fujii et al. 2011; Sellwood 2012).
In order to suppress the development of artificial features in all the N-body experiments that follow, we set up a live disk of stars embedded in a rigid dark matter potential. We employ simulations with a sufficiently large number of particles in the disk, i.e. 100 million, so that the disks are essentially featureless when evolved without any perturbers acting on them. These simulations serve as “controls,” making it possible to identify the response of the disk to imposed perturbations. In this manner, we will be able to separate the sources responsible for exciting features in the disk from the stars which react to the perturbations, unlike previous experiments in which the stars themselves acted as perturbers, complicating the interpretation of the experiments, as emphasized by Toomre (1990).
Looking forward to gas + full stellar evolution simulations for added beauty.
When I did astronomy at the Kapteyn Institute (in the 1980s), a similar theory floated about, which suggested the spiral arms are essentially a compression wave running through the gas and dust, triggering star formation in spiral patterns, and the hot, short-lived stars lit up the surrounding gas with their UV radiation. Shortly after the wave passed, the bright stars burnt up, and the amount of light decreased, without the total mass density changing much. The theory behind this simulation runs along a similar pattern, although in the 1980s simulating anything this big was impossible, of course.
From the paper:
A long-standing controversy over the nature of spiral arms is whether they correspond to density enhancements in the background stellar distribution (density waves), or are made up of stars that always remain in the arm and are just more concentrated than the stars outside the arm (material arms). In the early studies, the arms were assumed to be density waves, because if they were material they would quickly wind up as the galaxy rotates. Thus, both the swing amplification and the static density wave theories argue that the arms are overdense regions of the disk moving around at a different speed relative to the stars themselves. Stars thus continuously move in and out of the spiral arms. However, recent investigations using numerical simulations of stellar disks have challenged this claim and and argue that the arms might be material structures (Grand et al. 2012). To investigate this in the context of our simulations, we identify a patch of stars along the arm in the stellar disk after the arms are fully developed (after 100 Myrs) as displayed in top panels of Fig. 6, where the patch is colored in black. Then, we follow the positions of the stars originally in the patch forward in time and display the outcome after two galactic years (bottom panels). We note that the stars initially in the patch spread out, confirming that the spiral patterns in our simulations are density waves and not material structures. This is shown in polar coordinates in the bottom panel of Fig. 6 where it is clear that the patch is being sheared out by differential rotation and the pitch angle of the patch differs significantly from that of the spiral features.
Looking forward to Matt Bryant complaining how Suns are shit.
It would be a correct simulation if and only if at the very least 30% to 50% of the whole galaxy was known down to a single atom. If and only if all unknown forces other than the known ones here on Earth were known.
Creating any simulation based on assumptions that rely on practically no data, because what is officially known about just our galaxy is a 0.000000...[0s]....0001% or less ... it's a childish fantasy. It's just wrong data.
It's just like creating any simulation in any videogame and that's it. Nothing more than that.
It's not the truth. It's not even a bad approximation of reality.
Where did you come up with those numbers? Off the back of a Cornflakes packet?
Do you need to simulate every atom in a snooker ball to come up with a good mathematical model of how it bounces off a cushion? No. Do you need to simulate every atom in a galaxy to have a decent idea of how discrete clumps of 10^30kg matter (stars) behave in aggregation? No.
What you do need is a lot of simulated stars, a few clouds and a lot of computing power. They only put in the laws of physics, and visualisation techniques, and – presto – out pops spiral structures.
Kudos to them.
Back in the day when I studied astronomy at uni, it was widely believed that spiral arms were density waves, and could only really be understood as being examples of emergent behaviour. An n-body gravitational problem is a tough nut to crack when n is 10^11. So it's nice to see computing power making such simulations possible.
> It's just like creating any simulation in any videogame and that's it. Nothing more than that.
Implying the potemkin village of a videogame does any "simulation".
You are right insofar as numerical cutoffs (gotta stop subdividing time and space at some point) and the floating-point rounding and logarithmic scaling will influence the computation. These influences must be quantified and reasons must be given why such effects do not unduly affect the overall results. But in the end it's just an experiment in statistics, like weather pattern computation. You don't care about any exact results at all, just an representative result of an ensemble.
Joerg is correct to point out that "creating a simulation based upon assumptions" is a childish fantasy.
The underlying problem is the assumption that the laws of physics are immutable; laid down on stone tablets.
There is at least one much better explanation of the forces involved with the formation of the arms of a galaxy that has been out there for more than a decade; but it has not been reviewed because the new thinking challenges the assumptions laid down on those stone tablets.
In a sense. If the universe ITSELF rotates, then this will be apparent in the distribution of galactic rotation axis (no conclusive effect has been found AFAIK). I think it also means CLOSED TIMELIKE CURVES are possible, and one might have an UNIVERSE THAT CREATES ITSELF BY CURVING AROUND.
More on this in papers that are way over my paygrade.
Do I have to spell everything out for you first-graders?
Following Gödel, we can interpret the dust particles as galaxies, so that the Gödel solution becomes a cosmological model of a rotating universe. Besides rotating, this model exhibits no Hubble expansion, so it is not a realistic model of the universe in which we live, but can be taken as illustrating an alternative universe which would in principle be allowed by general relativity (if one admits the legitimacy of a nonzero cosmological constant). A less well known solution of Gödel's exhibits both rotation and Hubble expansion, and has the other qualities of his first model, so Gödel's model is really killed by the inconvenient observations that the universe is not rotating. The quality of these observations improved continually up until his death, and he would always ask "is the universe rotating yet?" and be told "no, it isn't."