New H0LiCOW result: Hubble constant is 73.3 +1.7-1.8 km/(s*Mpc)

[mfb]

TL;DR Summary

New measurement of the Hubble constant, increasing tension with measurements based on the CMB

The collaboration with the questionable acronym improved their measurement with a joint analysis of the whole dataset of six gravitationally lensed quasars. Measurements based on supernovae (measuring the Hubble constant "now") and measurements based on the cosmic microwave background (needing an extrapolation to get the current value) have been in disagreement for a while now. H0LiCOW has an independent method, its result fits well to the supernova results but not the CMB results.

H0LiCOW XIII. A 2.4% measurement of H0 from lensed quasars: 5.3σ tension between early and late-Universe probes

We present a measurement of the Hubble constant (H0) and other cosmological parameters from a joint analysis of six gravitationally lensed quasars with measured time delays. All lenses except the first are analyzed blindly with respect to the cosmological parameters. In a flat ΛCDM cosmology, we find $H_0=73.3^{+1.7}_{−1.8}$, a 2.4% precision measurement, in agreement with local measurements of H0 from type Ia supernovae calibrated by the distance ladder, but in 3.1σ tension with Planck observations of the cosmic microwave background (CMB). This method is completely independent of both the supernovae and CMB analyses. A combination of time-delay cosmography and the distance ladder results is in 5.3σ tension with Planck CMB determinations of H0 in flat ΛCDM. We compute Bayes factors to verify that all lenses give statistically consistent results, showing that we are not underestimating our uncertainties and are able to control our systematics. We explore extensions to flat ΛCDM using constraints from time-delay cosmography alone, as well as combinations with other cosmological probes, including CMB observations from Planck, baryon acoustic oscillations, and type Ia supernovae. Time-delay cosmography improves the precision of the other probes, demonstrating the strong complementarity. Using the distance constraints from time-delay cosmography to anchor the type Ia supernova distance scale, we reduce the sensitivity of our H0 inference to cosmological model assumptions. For six different cosmological models, our combined inference on H0 ranges from 73-78 km s⁻¹ Mpc⁻¹, which is consistent with the local distance ladder constraints.

 

[Arman777]

Very Interesting article. Thanks for sharing. I think it's clear that we need to change the flat lambda CDM model with something else to explain this discrepancy.
I am quoting "As this tension between early-Universe and late universe probes continues to grow..."
I have a question why the change in the $H_0$ is such a big deal? I remember that $H_0$ is not actually a constant but a parameter so its value should change in time and isn't it normal that it was low in the past and gets higher in the future?

 

[Bandersnatch]

I have a question why the change in the $H_0$ is such a big deal? I remember that $H_0$ is not actually a constant but a parameter so its value should change in time and isn't it normal that it was low in the past and gets higher in the future?

All the studies mentioned report the Hubble parameter now - that's what the 0 subscript indicates.
In case of CMB data, the measurements are taken at the early stages of the universe to determine model parameters back then, so that the model can be propagated forwards in time until the >now< moment is reached. The Hubble parameter decreases as the model propagates - it was much, much higher at the CMB emission (i.e. recombination) epoch.
It's the >now< $H_0$ that the model predicts that is in tension with the >now< $H_0$ measured from supernovae - and now also from lensing.
If everything were tip-top, the $H_0$ value predicted for today by the model would match the $H_0$ measured directly.

 

[Mordred]

And that really sources out the contention. At what range do we set $H_O$ Hubble constant today as opposed to the rate of expansion in the past ?

Using studies from CMB data is a huge timeline roughly 13 Gyrs. As you mentioned Hubble constant isn't constant, it is only constant for every location at a specified moment in time. So quite frankly studied such as this is a good plus with regards to measuring expansion rate via our locality.

I can double check using Jorrie's light one calculator which was developed with this site's involvement but if memory serves correct H at time of surface of last scattering was roughly 11000 times what is is today. Quite a change. Look in the $H/H_0$ column in the various pinned threads on this forum...

I am not sure how well the results would post on this site with current latex changes.. might be worth trying as it's a handy tool.

 

[mfb]

At what range do we set $H_O$ Hubble constant today as opposed to the rate of expansion in the past ?

We don't. You cannot measure it locally, so technically every measurement needs an extrapolation. That can be tiny for nearby galaxies, however.

A problem with the extrapolation from the early universe to today looks like the most plausible source of the discrepancy, but it doesn't have to be the cause. There were also some questions about the cosmological distance ladder before we got measurements independent of it .

 

[Mordred]

Agreed, that was how I had understood the scenario. This is one of the pluses to trying different methodologies on different sampling ranges. So it's not surprising to see differences between datasets from methodologies such as this to Planck or WMAP.

All the added data simply enables us to better fine tune our models. Yet it's curious that Planck tends to be low compared to other datasets I have come across. Though I cannot recall which ones they were specifically. They have been posted on this forum for the past few years.
One truth ( from my point of view) is that these discrepancies between datasets and models always lead to greater accuracy through further studies.

 

[Arman777]

All the studies mentioned report the Hubble parameter now - that's what the 0 subscript indicates.
In case of CMB data, the measurements are taken at the early stages of the universe to determine model parameters back then, so that the model can be propagated forwards in time until the >now< moment is reached. The Hubble parameter decreases as the model propagates - it was much, much higher at the CMB emission (i.e. recombination) epoch.
It's the >now< $H_0$ that the model predicts that is in tension with the >now< $H_0$ measured from supernovae - and now also from lensing.
If everything were tip-top, the $H_0$ value predicted for today by the model would match the $H_0$ measured directly.

Thanks for the explanation.

I also find a good picture from another recent article about the discrepancy on the $H_0$ value.

https://arxiv.org/abs/1903.07603

 

[windy miller]

I would have thought that an independent check on the Hubble parameter would support the idea that the tension is due to new physics, maybe varying dark energy rather than a cosmological constant, if it agree with the supernova result. Whereas if it agreed with the CMB that would be supportive that the super nova measurements have an error. Would I be right in such thoughts?

 

[PAllen]

I would have thought that an independent check on the Hubble parameter would support the idea that the tension is due to new physics, maybe varying dark energy rather than a cosmological constant, if it agree with the supernova result. Whereas if it agreed with the CMB that would be supportive that the super nova measurements have an error. Would I be right in such thoughts?

Generally, yes, and that is what is argued in the post just prior to yours.

 

[Mordred]

Well I can see papers coming out arguing this paper in support of a varying cosmological constant. It is already known that the Hubble parameter varies. It is only constant at a particular time. Which is a consideration when one considers we see into the past where is would have differing values the further you look.

 

[phyzguy]

FYI, I built the attached plot to show how far we are extrapolating from the CMB and how large the discrepancy is. It would only take a small error in the parameters extracted from the CMB to cause the observed discrepancy. I'm not saying that there may not be new physics, but I still think the most likely explanation is systematic errors in one set of measurements or the other. Time will tell.

 

[Mordred]

I too suspect systematic errors. As well as that graph posted here is a handy formula that can calculate the evolution of the Hubble parameter as a function of Z
$H_Z=H_O [\lambda_{mO}(1+Z)^3+\lambda_\Lambda+\lambda_{r0} (1+Z)^4]^{\frac{1}{2}}$

As you can see it accounts for the evolution of matter and radiation density while keeping the cosmological constant the same. I seriously doubt this paper will cast doubt on this formulas accuracy

 

[Mordred]

Why the bugger isn't latex working I tried $$ as well as [tex] and [itex]

 

[Ibix]

Why the bugger isn't latex working I tried $$ as well as [tex] and [itex]

You need to close the brace after r0 in $\lambda_{r0}$. Then you get this:

$H_Z=H_O [\lambda_{mO}(1+Z)^3+\lambda_\Lambda+\lambda_{r0} (1+Z)^4]^{\frac {1}{2}}$

 

[Mordred]

Doh thank you missed that error corrected thanks again

 

[phyzguy]

This paper came out today. A local measurement more in line with Planck. I added it to the graph below. The jury is still out on this issue.

 

[George Jones]

I would have thought that an independent check on the Hubble parameter would support the idea that the tension is due to new physics, maybe varying dark energy rather than a cosmological constant, if it agree with the supernova result. Whereas if it agreed with the CMB that would be supportive that the super nova measurements have an error. Would I be right in such thoughts?

Standard sirens and LIGO might soon provide an independent way of measuring the Hubble parameter:

https://physicsworld.com/a/merging-...nstant-crisis-sooner-than-previously-thought/https://www.nature.com/articles/s41550-019-0820-1https://arxiv.org/abs/1802.03404

 

[phyzguy]

I wonder how independent these local measurements really are. Don't they all depend on knowing the distance to the objects being used for the H0 determination? Are the "distance ladders" being used truly independent? Does anyone know?

 

[Mordred]

I know Planck doesn't just rely on its own measurements but also includes other datasets in its calibration procedures. About the others I couldn't really say at this time.
I will have to read these other papers after work but it's becoming obvious the jury is still out as you mentioned

 

[mfb]

I wonder how independent these local measurements really are. Don't they all depend on knowing the distance to the objects being used for the H0 determination? Are the "distance ladders" being used truly independent? Does anyone know?

H0LiCOW is independent of the distance ladder, this new red giant measurement says it is independent of "the Cepheid distance scale" - whatever that might mean for other measurements.
The red giant measurement is too central and with an uncertainty too large to help right now, but it is a promising approach if they can reduce the uncertainty in the future. Their uncertainty is largely from systematics - most of it from stellar models as far as I understand table 2.

 

[Mordred]

Having getting a chance at reading the new findings your assessment is pretty accurate. In a way these different methodologies and studied is particularly good in that one can assess the pros and cons of each study which greatly helps us remove or minimize systematic errors.

I found some of the theories mentioned in the new paper rather interesting with regards to the 6 km/MPC error margin. I hadnt heard of those particular studies until that paper

 

[cmb]

This unit, km/s/Mpc, always seems a bit odd to me when it reduces down to a s^-1 term.

This value can be re-written as 1/(13.3855 Ga), which seems more "physical" to me.

 

[George Jones]

This unit, km/s/Mpc, always seems a bit odd to me when it reduces down to a s^-1 term.

This value can be re-written as 1/(13.3855 Ga), which seems more "physical" to me.

Which is more "physical" depends on the context. Sometimes, it is intuitive to express the Hubble parameter as an inverse time; other times, units of (km/s)/Mpc are more intuitive.

If $D$ is the (now) proper distance between us and a distant galaxy, then

$$D \left( t \right) = R \left( t \right) \chi,$$
where $R \left( t \right)$ is the scale factor for the universe, $\chi$ is the constant comoving coordinate difference between us and the galaxy, and $t$ is cosmological time.

If the comoving coordinate is constant in time (zero peculiar velocity), the rate of change of proper distance is given by

$$\begin{align} \frac{dD}{dt} &= \frac{dR}{dt} \chi \nonumber \\
&= \frac{\frac{dR}{dt}}{R} R \chi \nonumber \\
&= H D, \nonumber
\end{align}$$
where the Hubble parameter (a function of time) is give by

$$H = \frac{\frac{dR}{dt}}{R}. $$
Astronomers typically express distances like $D$ in Mpc, and "recessional speeds" like $dD/dt$ in familiar units of km/s. In order to make these units work out in the above equation, $H$ is expressed in units of (km/s)/Mpc.

 

[mfb]

There was a conference about it in the last days: KITP Conference: Tensions between the Early and the Late Universe

As far as I understand there are six largely independent measurement types:
* Supernovae
* H0LiCOW (quasar lensing)
* TRGB (red giants)
* Neutron star mergers + gravitational waves
* Baryonic acoustic oscillations + Big Bang nucleosynthesis
* Cosmic microwave background

The first two measure today's universe, agree with each other and favor a larger Hubble constant today, the last two measure the early universe, agree with each other and favor a smaller Hubble constant today. The difference between the two groups is quite significant.

TRGB gets a value that is in between the groups and compatible with both.
Gravitational waves are not yet sensitive enough to contribute but could get there in the near future.

There seem to be more ideas around but they don't have a result yet or the uncertainty is so large that it is not yet relevant.

 

[Arman777]

There was a conference about it in the last days: KITP Conference: Tensions between the Early and the Late Universe

As far as I understand there are six largely independent measurement types:
* Supernovae
* H0LiCOW (quasar lensing)
* TRGB (red giants)
* Neutron star mergers + gravitational waves
* Baryonic acoustic oscillations + Big Bang nucleosynthesis
* Cosmic microwave background

The first two measure today's universe, agree with each other and favor a larger Hubble constant today, the last two measure the early universe, agree with each other and favor a smaller Hubble constant today. The difference between the two groups is quite significant.

TRGB gets a value that is in between the groups and compatible with both.
Gravitational waves are not yet sensitive enough to contribute but could get there in the near future.

There seem to be more ideas around but they don't have a result yet or the uncertainty is so large that it is not yet relevant.

Thanks a lot for sharing this. I am going to watch all of them ( Right now I am watching the Tommaso Treu's talk)

The problem is that as you said there's a 5.3 sigma difference between late and early measurements. But I think the biggest issue is that The LCDM explains most of the observed things in great accuracy ( in general ). When we try to leave the LCDM model to explain the tension in the Hubble parameter ( for instance changing w for dark energy ) we are altering the other observed quantities (especially in the CMB) which they were consistent with the LCDM model.

I think this means one thing. We need a new cosmological model

 

[phyzguy]

When I look at the H0LiCOW measurements, I continue to wonder whether the systematics are truly understood. Look at the graph of the actual measurements, below. Four of the six measurements are clearly consistent with Planck at 67. Only the two high points pull the average over to where it disagrees with Planck. If they hadn't measured those two, the conclusion would be very different. Can you really look at this and say, "It is clearly inconsistent with the CMB, and we need new physics." I can't.

 

[Arman777]

In the conference videos somone in the audicance asked the same question. He replied that , " Its expected for 6 objects. I don't think there's anything special with that one". Also there are large uncertanities in each data ( from $±2.9$ to $±8.0$.

Tommaso also mentioned about the systematic errors as you mentioned. In the future they will test up to 40 systems which will most likely give better results.

Even if the data were close to the Planck values, it still does not solve the Hubble parameter problem. (At least in my opinion).

Many independent experiments on both side (early and late) show discrepency. It seems its more than some systematic error. If the difference would have been 1-2 that's kind of okay but 4.4-5.3 ... i think that's a huge gap.

 

[George Keeling]

I wanted to find out more about the the Hubble constant. I asked Wikipedia and I now understand why the units are (km/s)/Mpc. There's a table at the end which gives a history of Hubble constant measurements (including his own of 500 in 1929!). This latest of $H_0={73.3}_{-1.8}^{+1.7}$ is not there. Perhaps somebody would like to add it. I am insufficiently qualified to fill in the other columns.

 

[mfb]

Added it. The list is getting long, it will probably need some new solution soon with the rapid rate of new measurements.

 

[Buzz Bloom]

The chart in post #26 seems to identify Bayesian a priori ranges for parameters H₀ and Ω_m. I would appreciate someone explaining why Ω_k is ignored. A diagram showing a probability range of values for Ω_k was included in the Planck Legacy 2018 release

https://arxiv.org/pdf/1911.02087v1.pdf​

on page 4. On page 5 the corresponding model values are given:

Ω_m= 0.35,​

Ω_Λ= 0.65, and​

H₀= 65 km/s/Mpc .​

Does anyone know whether or not the model computations leading to approximately 74 km/s/Mpc included any non-zero Ω_k value, and if not, whether including a negative Ω_k like the one in Planck Legacy 2018 might yield a H₀ value closer to the 65 km/s/Mpc value.

If the Ω_k data I am asking about is actually in a paper which includes approximately the H₀ = 74 value, I would hope that if someone posts an appropriate citation, I will be able to understand reading the cited reference.