Dipole pattern for Fine Structure Constant?

[mikenw]

Can someone describe for me the dipole pattern that was detected related to the fine structure constant? This thread talks about that pattern some. What I am trying to understand is how that dipole is oriented relative to the observer on Earth? Are we talking about a toroidal shape with a null(s) pointing in some opposite directions of the sky? Or is it something else? I am having trouble visualizing the actual shape of this pattern to the observer on Earth. Thank you! (first time post here too, hello to everyone! :)

 

[Chalnoth]

A dipole means that they measured the fine structure constant to be slightly lower in one specific direction, and slightly higher in the opposite direction.

However, this result is highly unlikely to hold up under scrutiny.

 

[mikenw]

Thank you Clanoth for clarifying that!
I am curious why you say "this result is highly unlikely to hold up under scrutiny?" Is there some other more likely explanation for this dipole pattern?

 

[windy miller]

Whether it will hold up or not is a matter of opinion. We shall see. If you want a talk that is relevant watch this:

 

[Chalnoth]

Thank you Clanoth for clarifying that!
I am curious why you say "this result is highly unlikely to hold up under scrutiny?" Is there some other more likely explanation for this dipole pattern?

That it's a spurious result based upon overfitting of the spectra of heavier elements in quasars.

 

[ruarimac]

I am curious why you say "this result is highly unlikely to hold up under scrutiny?" Is there some other more likely explanation for this dipole pattern?

While I don't believe that overfitting is the problem (as that can be tested in simulation), real systematics with the measurements have been called into question. It is in my opinion most likely spurious but we will have better data on it soon. The new ESPRESSO spectrograph will be mounted on the VLT in a couple of years, it's a much more carefully designed instrument than UVES or HIRES in terms of wavelength calibration.

http://arxiv.org/abs/1409.4467

 

[mikenw]

Hey folks,
Can you help me understand exactly what is meant by the term "dipole pattern" in this context? Is it simply the fact that one direction measured + and the opposite direction measured - ? As in this illustration:

This is throwing me for a loop because dipole field strength (at least in ham radio) goes up as you are at 90 degrees to the actual axis of the dipole, not 180. (180 would represent a + to + or - to - ). I normally think of 90 degrees as being the + to - range of field strength for a radio dipole. And yet the Alpha measurements were showing a 180 degree + to - strength change? Any help anyone can offer is much appreciated! (I am probably I am over-complicating it!).

 

[Chalnoth]

Hey folks,
Can you help me understand exactly what is meant by the term "dipole pattern" in this context? Is it simply the fact that one direction measured + and the opposite direction measured - ? As in this illustration:

This is throwing me for a loop because dipole field strength (at least in ham radio) goes up as you are at 90 degrees to the actual axis of the dipole, not 180. (180 would represent a + to + or - to - ). I normally think of 90 degrees as being the + to - range of field strength for a radio dipole. And yet the Alpha measurements were showing a 180 degree + to - strength change? Any help anyone can offer is much appreciated! (I am probably I am over-complicating it!).

I think so. As in, they claimed the fine structure constant was slightly higher in one direction compared to the opposite direction. This is averaged over observations in many different directions.

 

[mfb]

You get the 90° if you take the magnitude. In radio transmission, you are interested in intensity only, which depends on amplitude squared. Look at the "+-" direction and you see something (if there is a dipole dependence, which is highly questionable), rotate by 90° and you see nothing.

 

[mfb]

The dipole moment is defined as an integral over the whole solid angle. It is either zero or non-zero, but never a monopole. It has a direction, roughly telling you in which hemisphere something is larger than in the other. The monopole contribution is the average fine-structure constant. Or simply the fine-structure constant if it is constant.
Higher moments can exist in general, if the pattern is more complicated. If there is any variation, you would expect higher moments (quadrupole and so on), but those are harder to measure.