Wednesday, 7 December 2011

More dark matter shenanigans

The internet is alive again with another cry that dark matter is dead (again) and the slashdot-eratti are getting themselves into the usual lather and "DM is BS" claims.

I've written my views on slashdot commenting previously, and will not reiterate them here, but will comment on the paper and the "meaning" of dark matter to astronomers.

OK. The paper can be read here and here's the press image that goes with it.
The picture is correct. If we just considered the gravitational attraction of stars, then their rotation speed should follow the red curve, but when we measure it (and we can measure it far outside the stellar disk by looking at the rotational velocities of HI gas) it actually follows the white curve.

So, the big question is why? The prevailing hypothesis is that there is more mass there than we can see, i.e. dark matter. But others suggest that dark matter is not there, and there is some other influence, usually by modifying the laws of physics (i.e. MOND) which accounts for the extra acceleration which is needed to give the larger speeds.

The current paper suggests that the extra acceleration comes from the attraction of mass in the local universe *outside* of the galaxy. Those that remember their classes on Newtonian physics will remember that if you sit inside a spherical shell of mass, you do not feel any gravitational pull from the mass. But this paper says "well matter is not smooth but is lumpy" and the author, Carati, tries to calculate the influence of this lumpiness.

This is all quite legit, but it turns out the calculations are very difficult, so instead of directly calculating the influence, he calculates some average effect. What happens is the "average" effect modifies the gravitational attraction in galaxies and so looks like this
It's the second term that modifies the gravitational attraction due to this average effect. Take a couple of rotation curves of galaxies, and viola
The left-hand panel has the data (the dots and error bars) and the effect of two components; the disks is due to the mass we see, and the halo is the influence of dark matter. The right-hand panel is the result of this new model, which only has the disk we see and then the correction due to the cosmological mass. Excellent. Let's have a press release!

But, hold your horses. Is everything as excellent as it seems? Well, no. Firstly, we knew we could get this model to work in some galaxies as it basically looks extremely similar to the mathematical form of MOND, and we know that works in some galaxies (and so, in some sense, we knew the answer beforehand). And remember that we have taken some sort of averaging effect, rather than calculating the actual effect, and, as it is stochastic should be stochastic and I can't imagine that it will give a smooth influence on the galaxy.

Are the religious, dark matter zealots up in arms, calling for Carati's head for daring to suggest that the god of dark matter may not exist? Well, no. While the majority of scientists will be nowhere near convinced by this one paper, Carati is free to do whatever research they want to, and I am sure they know that if they want to convince us that their model is viable, the maths needs to be worked out, and then they must show that their model explains everything that dark matter does; from gravitational lensing, big-bang nucleosynthesis, hot gas in clusters etc etc.

Again, scientists don't believe in dark matter. Currently all the evidence points to a material substance explaining what we see out there in the Universe, and so people are heavily weighted into using this particular model. But if we woke up tomorrow and some one has convincing proved that dark matter as a substance can be conclusively ruled out, there will be no wailing and gnashing of teeth, and most scientists will say "oh, that's interesting! what more does that tell us about the Universe". Science will move on. And I am sure the Slashdotters will tell us "we told you so!".


  1. As Bob Sanders pointed out, even if one thinks that dark matter is a better explanation than MOND, one still needs to get the MOND result in those many cases where it does work in some way which is not contrived or ad-hoc.

  2. I'm thinking of the rotation curves which were the initial motivation for MOND. While one can get an arbitrary rotation curve with arbitrarily complicated physics, mass distributions, angels etc, the point is that flat rotation curves must come out naturally of any theory, be it dark matter or anything else. MOND doesn't work everywhere and has no real theory, but is simple in that one free parameter fits a lot of stuff. Now that it has become clear that standard CDM isn't right in the details, any fix for that problem needs to appear natural and not ad-hoc.

    I'm not saying that MOND is correct, rather that it indicates something interesting which we don't completely understand, and that standard CDM doesn't (yet) get completely right.

  3. I agree that MOND fits flat rotation curves, but that, in part, was what it was designed to do. If rotation curves rose of fell then you can construct a simple MOND with a dependence on r.

    As you said, MOND is not really a theory (and, IMHO, is more epicyclic than DM)