Hubble Tension and Cosmic Acceleration: A measurement artifact?

[ritesh]

TL;DR

The mean separation between two reported values of a normal distribution should be 0.79 sigma. However, it was found to be 2.07 sigma in the case of extra-galactic distances. Does it mean that the error bars in cosmological distances are underestimated?

By analyzing 91,742 reported extra-galactic distances and their one sigma uncertainties for 14,560 galaxies, it was found that pairs of reported extra-galactic distances of the same galaxy differ from each other by 2.07 the reported uncertainties on average.

In my opinion, this indicates that the uncertainties in extra-galactic distances are widely under-estimated.

If correct, could this mean that our confidence in the local Hubble-Lemaitre constant (which is directly derived from extra-galactic distances and recession velocities) is over-estimated?

If so, it would mean that the Hubble Tension is an artifact of this underestimation of uncertainties. It could also reduce the certainty of Cosmic Acceleration since the Acceleration is just observation of marginal reduction in the value of the Hubble-Lemaitre constant with distance (negative time).

I would be grateful to get the opinion of other thinkers on this forum.

More information about the result:
Paper - https://link.springer.com/article/10...09-021-04006-5
Full Text - https://rdcu.be/cAre5

 

[PAllen]

As a basis for comparison and discussion, I found a very recent review of these issues:

https://arxiv.org/abs/2103.01183

 

[berkeman]

As a basis for comparison and discussion, I found a very recent review of these issues:

https://arxiv.org/abs/2103.01183

I can't tell if the paper you link to refutes the OP's position or if you are agreeing that it is an active area of research. Can you comment please? Thanks.

 

[PAllen]

I can't tell if the paper you link to refutes the OP's position or if you are agreeing that it is an active area of research. Can you comment please? Thanks.

The paper I linked argues that systematic errors are unlikely but not wholly ruled out (it argues you need multiple independent error sources of similar magnitude). I linked it for background, and that it reviews the whole topic of Hubble tension in detail, with latest data. The OP linked paper, if I understand, claims that there are actually correlated error bars in seemingly independent elements of the distance ladder. I definitely think it is an open research area, with disagreement among experts.

 

[ohwilleke]

I definitely think it is an open research area, with disagreement among experts.

I agree. I've seen credible papers in the literature arguing both ways. There are multiple possible solutions (which aren't necessarily mutually exclusive).

Some sort of systemic or theory error is one of them.

Some of this is involves straight out claims of a small number of inaccurately estimated measurements uncertainties. For example, someone claims that the resolution of a key telescope is X when that is really only under optimal conditions that aren't present in reality.

Another is that measurements are failing to adjust for factors that are often immaterial, but matter in this case.

For example: gravitational lensing and red shifting of the light used to assess distances and redshift by "near misses" of objects between us and the source that are hard to see (like filament dark matter, and black holes), or from effects like absorption and re-emission of light by dust between us and the source, or from small systemic variations within a class of objects being treated as standard candles when they really aren't.

Another possibility is the notion that the Hubble Constant is itself conceptually unsound because it implicitly assumes a cosmology model that is simpler than reality. Maybe, for example, the Hubble "Constant" is really a function that varies over time as many alternative theories suggest.

 

[PAllen]

@ritesh, have you tried calculating Hubble using only the smallest of available measured distances for each galaxy, versus using only the largest?

 

[ritesh]

@ritesh, have you tried calculating Hubble using only the smallest of available measured distances for each galaxy, versus using only the largest?

Thank you Dr. Allen for your interesting and very important question. I believe that one of the kindest things a person can do is to pay complete attention to another person. Thank you so much for that.

Yes, I have spent considerable time investigating the question asked by you. The results are intriguing to say the least.

Consider the Type Ia Supernova SN 13651 (selected as randomly as is humanly possible). It has two distance measurements reported in 2018 (https://iopscience.iop.org/article/10.1088/1538-3873/aab4e0):

1. 1706 ± 128 Mpc
2. 2109 ± 195 Mpc

In other words, the 2nd distance measurement lies at (2109-1706)/128 = 3.15 σ₁ or 3.15 times the reported uncertainty in measurement 1.

Whereas, the 1st distance measurement lies at (2109-1706)/195 = 2.07 σ₂ or 2.07 times the reported uncertainty in measurement 2.

For SN 13651, the mean separation between distance measurements as a multiple of their reported uncertainties, a quantity which I have called as mean Δ(σ) = (3.15+2.07)/2 = 2.61.

The conclusion of my work is simple: it only says that since the mean separation between pairs of distance moduli measurements to the same galaxy differ by 2.07 times the reported uncertainty on average, it seems that the uncertainties are widely underestimated. I found this mean separation being 2.07 times the reported uncertainty by analyzing 91,742 reported extra-galactic distances and their one sigma uncertainties for 14,560 galaxies.

Now, coming to your very interesting question Dr. Allen.

The reported error in the 1st measurement is 7.5% and that in the 2nd measurement is 9.2%. If the measurements are combined, the error bars need to be increased to at least 20%.

However, even if one measurement is completely discarded, still the error bar in the Hubble-Lemaitre constant determined from this measurement should at least be 7.5% (assuming 0% uncertainty in the speed of recession, which is unrealistic).

In Cosmology, the standard value of the Hubble-Lemaitre constant is taken to be 70 km/s/Mpc. So, even if we discard one measurement, and assume 0% uncertainty in the speed of recession, the Hubble-Lemaitre constant determined from the 1st measurement alone should be 70 ± 7.5% = 70 ± 5.25 km/s/Mpc. This range encompasses the reported value of the Hubble-Lemaitre constant from the Cosmic Microwave Background (CMB) measurements by the Planck observatory.

Where is the "Hubble Tension" or the "Crisis in Cosmology"?