Arto Annila - Universe's expansion may be understood without dark energy

[TNick]

http://www.physorg.com/news/2011-10-supernovae-universe-expansion-understood-dark.html

Your input, please!

 

[Chronos]

This has been proposed before for a variety of reasons, but, dark energy is still considered the best fit to observational evidence to date.

 

[BillSaltLake]

Looks like a terrible fit. At z=1, the least-action predicted bightness is ~2/5 of a mag low, which is 1.4x too bright. If I remember right, the observed brightness deviation (cf a matter-dominated Universe) is only ~1.5x at z=1.

 

[phyzguy]

If I understand correctly, he is also saying that gravitational lensing in underpredicted by 5X (p 2947). But we've measured the gravitational deflection of the sun, and there's no way we're off by 5X. Or am I missing something?

 

[Chalnoth]

Looks like a terrible fit. At z=1, the least-action predicted bightness is ~2/5 of a mag low, which is 1.4x too bright. If I remember right, the observed brightness deviation (cf a matter-dominated Universe) is only ~1.5x at z=1.

Indeed, it is an absolutely terrible fit. Note how the line is below almost all of the data points.

 

[twofish-quant]

As far as I can tell, he is proposing a "tired light" model which doesn't work for lots of reasons (look up wikipedia and google).

Looking on his webpage, Arto Annila main expertise is in complex systems. He has no background in astrophysics and obviously has not read the wikipedia pages or done a google search on "tired light."

If you cut out the section in which he talks about supernova, then the paper is actually rather interesting.

 

[twofish-quant]

Just FYI for why tired light models have problems.

1) If you have light change speed, it's like going through a lens. If you go through a lens then there are a ton of effects that we don't see. In particular, if you are moving through a lens then things that are far away get blurry.

2) More to the point if there is something that causes supernova light to go funny on the way to the earth, that whatever causes the light to go funny will also affect anything beyond the supernova. Which will result in all sorts of lensing effects.

 

[Chronos]

This is merely populist bull, IMO. IOW, I agree with twofish.

 

[bill alsept]

Just FYI for why tired light models have problems.

1) If you have light change speed, it's like going through a lens. If you go through a lens then there are a ton of effects that we don't see. In particular, if you are moving through a lens then things that are far away get blurry.

2) More to the point if there is something that causes supernova light to go funny on the way to the earth, that whatever causes the light to go funny will also affect anything beyond the supernova. Which will result in all sorts of lensing effects.

How do (1) and (2) prove that light can't change speed?

How do we know "anything beyond the supernova" is not affected?

 

[Chalnoth]

How do (1) and (2) prove that light can't change speed?

Well, we know that the speed of light doesn't change to a high degree of accuracy. This doesn't necessarily mean it can't, but it certainly doesn't (at least not by any significant amount).

How do we know "anything beyond the supernova" is not affected?

Because when you have a change in the speed of light it impacts everything we see whose light was en-route to us before the change in speed. So, if there were changes that started to be detectable at, say, 1 billion light years, they would be even more apparent at 2 billion light years, and so on. But the speed of light is highly uniform all the way out to the CMB.

 

[RUTA]

Looks like a terrible fit. At z=1, the least-action predicted bightness is ~2/5 of a mag low, which is 1.4x too bright. If I remember right, the observed brightness deviation (cf a matter-dominated Universe) is only ~1.5x at z=1.

I posted this on another thread today, but it's relevant here, so ...

I did a fit using his version of mu (Eq 4 in Mon. Not. R. Astron. Soc. 416, 2944–2948 (2011)) over linearized Union2 Compilation data (http://supernova.lbl.gov/Union/) and compared it to LCDM and the flat, dust-filled model (which is just LCDM minus Lambda). First, let me show you how his mu vs z looks compared to LCDM (Ho = 70, OmegaM = 0.3) when I use T = 14Gy in his mu (T is age of the universe, his only parameter, and in his Fig 3 caption he says he used T = 13.7Gy):

http://users.etown.edu/s/stuckeym/Plot 14Gy.pdf [Broken]

Red is LCDM, green is Annila. Now look at this same comparison using T = 15Gy:

http://users.etown.edu/s/stuckeym/Plot 15Gy.pdf [Broken]

Much improved, but he still could've given us some fit info. The 14Gy plot shows why his curve in his Fig 3

http://users.etown.edu/s/stuckeym/Annila Figure 3.jpg

is a little low at high z, since LCDM doesn't suffer that fate. Anyway, here is what I have for (mu/5 - 8) vs log(z) with Union2 data:

Best fit line SSE = 1.95, R = .9955
Best fit flat, dust-filled SSE = 2.68 using Ho = 60.9
Best fit LCDM SSE = 1.79 using Ho = 69.2, OmegaM = 0.29, OmegaL = 0.71
Best fit Annila SSE = 1.95 using T = 14.9Gy
Annila SSE = 2.69 using T = 13.7Gy

So, I think he would've been better served to show his fit using T = 14.9Gy rather than T = 13.7Gy. Anyone else run numbers? It would be nice to double check.