Discrepancy in the Hubble Paramater

[windy miller]

As I understand Planck has a value for Hubble as 68 kilometers per second per megaparsec and the supernova data is saying 71 . I also thought the time since the big bang (I think cosmologists are wrong to call it the age of the universe since we don't really know whether the universe did begin at the big bang or had some prior history) was determined by the value you get for Hubble. So why do people always talk about 13.8 billion years? Why isn't there two lengths of time stated if there are two radically different values for Hubble?

 

[Orodruin]

Please provide appropriate references.

 

[windy miller]

DO i really have to? Isn't this widely known in the community ?

 

[Orodruin]

If you want to discuss something, you are responsible for providing references as the basis for that discussion. This is particularly true when somebody explicitly asks you to provide references. Your reaction is completely uncalled for.

Edit: In addition, even if it is known in the community it is no excuse for not providing the references when asked. References lay out a common basis for the discussion and even if I know the answer to your questions it is helpful to know what you have been reading.

 

[windy miller]

If you want to discuss something, you are responsible for providing references as the basis for that discussion. This is particularly true when somebody explicitly asks you to provide references. Your reaction is completely uncalled for.

Edit: In addition, even if it is known in the community it is no excuse for not providing the references when asked. References lay out a common basis for the discussion and even if I know the answer to your questions it is helpful to know what you have been reading.

if I was at a seminar and I asked this question and the answerer rather than answering it asked me to supply my references for a fact that was well know in the community. i would take it the person was just being difficult for the sake of being difficult. but i appreciate the argument you made and apologise. here are the references:
https://arxiv.org/pdf/1708.07030.pdf
"The tension between the direct and indirect measurements is well known in the liter- ature [50]. However, there was no disagreement between indirect measurements until the advent of Planck data. The WMAP 7 year result [51] has given such values of H0 − Ωm which can accommodate the LSS results. On the other hand, MCMC analysis done with Planck data, as described in the previous section shows some tension with LSS result obtained sim- ilarly with the Planck prior set on {As,ns,τreion}. Planck analysis gives the value of H0 to be 67.91 ± 0.89 with Ωm = 0.305 ± 0.012 where as the best-fitted value for H0 turns out be 70.58 ± 0.83 and Ωm is 0.270 ± 0.005 for LSS data at 1-σ level. We see that there is clear discordance between these best-fit parameter "

 

[Orodruin]

if I was at a seminar and I asked this question and the answerer rather than answering it asked me to supply my references for a fact that was well know in the community.

This is a very different situation. When you are in a specialist seminar people will expect you to be familiar with most of the literature.

Planck analysis gives the value of H0 to be 67.91 ± 0.89 ... best-fitted value for H0 turns out be 70.58 ± 0.83

There is only a slight tension between those results, assuming that the errors are given at 1 sigma.

I also thought the time since the big bang (I think cosmologists are wrong to call it the age of the universe since we don't really know whether the universe did begin at the big bang or had some prior history) was determined by the value you get for Hubble.

To answer this question, the Hubble parameter is indeed relevant for the age. However, it also depends on the distribution of the energy content of the Universe and to get the age you need to explicitly consider the different dynamics of the expansion depending on what components is dominating. Of course, people should typically include the errors in the numbers when they talk about the age (or any other measurement) and state what assumptions and data sets go into the analysis. Of course, at a popular or overview level this might not always happen.

 

[Vanadium 50]

(70.58 ± 0.83) - (67.91 ± 0.89) = 2.67 ± 1.22, which is a 2.2 sigma discrepancy. It's an important number, and this level of disagreement is annoying, but not particularly rare. About one in 30 such measurement pairs will have this level of disagreement, more if you take an ensemble of measurements first and then decide what axis you will categorize them on second. So it's not very significant.

The situation has improved over time. Forty years ago, measurements were clustered around 100 and around 55.

 

[windy miller]

This is a very different situation. When you are in a specialist seminar people will expect you to be familiar with most of the literature.There is only a slight tension between those results, assuming that the errors are given at 1 sigma.To answer this question, the Hubble parameter is indeed relevant for the age. However, it also depends on the distribution of the energy content of the Universe and to get the age you need to explicitly consider the different dynamics of the expansion depending on what components is dominating. Of course, people should typically include the errors in the numbers when they talk about the age (or any other measurement) and state what assumptions and data sets go into the analysis. Of course, at a popular or overview level this might not always happen.

One the point about the seminar . Its well taken but are you suggesting if you were giving a popular level talk to the public and someone asked you this question rather than answer it would you ask the the questioner for their source? If someone asked you what is the universe expanding into, would you ask them , what's your source that the universe is expanding? I appreciate its very important to provide sources if someone came on here with a controversial claim. but not for one that is well agreed upon . It seem like pedantry to me. Anyway maybe we should move past this distraction. I apologise for being grumpy about it and appreciate your help , although i can't help it nagging at me .

on to the issue at hand

So are you saying that even if the discrepancy is genuine then the time since the big bang comes out the same as the energy content changes in the different approaches? So in that sense it doesn't matter whether you have 68 or 71?? You still get 13.8 billion years? Also As I understand the sigma level fo the discrepancy is getting higher as time goes on:

"Meanwhile, Planck, a European Space Agency (ESA) mission that has imaged the CMB at higher resolution and greater temperature sensitivity, has settled on 67.8. In statistical terms, the two values are separated by a gulf of 3.4 sigma— not quite the 5 sigma that in particle physics signals a significant result, but getting there. "That, I think, is hard to explain as a statistical fluke," says Chuck Bennett, an astrophysicist at Johns Hopkins who led the WMAP team."
http://www.sciencemag.org/news/2017...eed-expansion-universe-could-lead-new-physics

 

[windy miller]

(70.58 ± 0.83) - (67.91 ± 0.89) = 2.67 ± 1.22, which is a 2.2 sigma discrepancy. It's an important number, and this level of disagreement is annoying, but not particularly rare. About one in 30 such measurement pairs will have this level of disagreement, more if you take an ensemble of measurements first and then decide what axis you will categorize them on second. So it's not very significant.

The situation has improved over time. Forty years ago, measurements were clustered around 100 and around 55.

Although the Hubble has converged into the late 60,s early 70's range. As I understand it the sigma discernacy between Planck and Super Nova data has grown and not gone down , this is not what we would expect if it was a fluke. See my source above now claiming 3.4 sigma different.

 

[Orodruin]

if you were giving a popular level talk to the public and someone asked you this question rather than answer it would you ask the the questioner for their source?

I would argue that this is also a different situation. It is easier to have a face-to-face discussion and break down the answer if it is not understood.

So are you saying that even if the discrepancy is genuine then the time since the big bang comes out the same as the energy content changes in the different approaches? So in that sense it doesn't matter whether you have 68 or 71?? You still get 13.8 billion years?

I do not have the numbers and errors readily at hand. You would have to explicitly look into what different datasets the 13.8 billion years refer to.

As I understand it the sigma discernacy between Planck and Super Nova data has grown and not gone down , this is not what we would expect if it was a fluke.

Well, you would also not expect a fluke in the first place (by definition of fluke). However, sometimes statistics do funny things for you. People generally seem to over-interpret "but it has not gone down" as "it is really there" beyond the level of evidence that is actually available. Still, 3 sigma results are not that uncommon. (Note that a 3 sigma result does not mean that there is a 99.7% probability of actually having found something significant.)

 

[windy miller]

I would argue that this is also a different situation. It is easier to have a face-to-face discussion and break down the answer if it is not understood.I do not have the numbers and errors readily at hand. You would have to explicitly look into what different datasets the 13.8 billion years refer to.Well, you would also not expect a fluke in the first place (by definition of fluke). However, sometimes statistics do funny things for you. People generally seem to over-interpret "but it has not gone down" as "it is really there" beyond the level of evidence that is actually available. Still, 3 sigma results are not that uncommon. (Note that a 3 sigma result does not mean that there is a 99.7% probability of actually having found something significant.)

Id rather not get bogged down in the levels of sigma. i accept that this discrepancy could be a fluke. My question is more about what the different Hubble parameters imply for the time since the big bang. It seems you are unable to answer that, but I appreciate the discussion so far, thanks.

 

[Orodruin]

It seems you are unable to answer that, but I appreciate the discussion so far, thanks.

Well, the only answer that can be given is that it is complicated. Again, you have to look at the datasets that are used and the results obtained (with error bars) in order to know if they are compatible or not. Many cosmological observables are correlated and it is difficult to answer the question without going into the gritty details. Your question as it is can only be answered by looking into the literature (or doing the gritty details yourself). Something that I currently do not have time to do.

 

[russ_watters]

One the point about the seminar . Its well taken but are you suggesting if you were giving a popular level talk to the public and someone asked you this question rather than answer it would you ask the the questioner for their source?

I might, yes. But it is certainly more reasonable here, where conversations are more drawn-out and writing things down enables easier transfer of information. You should also consider that you are arguing against the rules in a highly successful, 18 year old forum. You're not going to change our rules.

But regardless of the practical issues, clearly it was the right thing to do here, since the numbers you initially stated were wrong!

 

[windy miller]

You are ri

I might, yes. But it is certainly more reasonable here, where conversations are more drawn-out and writing things down enables easier transfer of information. You should also consider that you are arguing against the rules in a highly successful, 18 year old forum. You're not going to change our rules.

But regardless of the practical issues, clearly it was the right thing to do here, since the numbers you initially stated were wrong!

You are right I initially posted the wrong values I have now corrected this. But I don't see what the relevance is to the question. the question was do the different value of the Hubble parameter mean we should have two values for the time since the big bang? We've had more than a dozen posts in this thread now and no one has actually given a clear answer answer to the question in hand. Perhaps then you can see why someone gets frustrated by side shows like these. And i agree physics forums is excellent. But that doesn't mean there aren't cases of unnecessarily distracting the conversation or that things can't be improved. You didn't provide a reference that physics forums was 18 years old. since the numbers you initially stated were wrong! the forum was founded in 2001. here's a reference for you :
https://www.physicsforums.com/help/aboutu
Now the reference i have only gives us the year of founding: 2001 . it might have been dec 31 2001 for all i know in which case it is more like 16.2 years old. Which is further out on a percentage basis that I was on my values for the Hubble. Perhaps you can see how this sort of pedantry can get frustrating and waste people times?

Dont get me wrong I am all in favour of providing references. But there are cases when these are required and cases when they are not. to ask for the reference and pick apart your mistake about the age of physics forums is pedantry and I am doing it as a satirical way of making a point. . the truth in the point you make isn independent of wether the forums 18 years old, or whatever. So to focus on it is just silly. Same with my questions. To focus on the values of Hubble or even the sigma is to miss the point of the question. . of course we should correct false information but to do so without answering the questions in hand is just annoying. I think if someone asks a simple question on a forum about something that has been widely reported and not disputed then it would be simpler and easier to just answer the question. something I'm still waiting for.

 

[Orodruin]

We've had more than a dozen posts in this thread now and no one has actually given a clear answer answer to the question in hand.

How is this

the Hubble parameter is indeed relevant for the age. However, it also depends on the distribution of the energy content of the Universe and to get the age you need to explicitly consider the different dynamics of the expansion depending on what components is dominating.

not a clear answer?

To reiterate: The Hubble parameter is important, but it is not the only measurement that is relevant for determining the age of the Universe. You really cannot get more information than that without doing the gritty details.

 

[windy miller]

How is this

not a clear answer?

To reiterate: The Hubble parameter is important, but it is not the only measurement that is relevant for determining the age of the Universe. You really cannot get more information than that without doing the gritty details.

And noe one has done this?

 

[Vanadium 50]

First, it's difficult to have a quantitative discussion if the numbers are shifting.

Second, discussing a statistical discrepancy without discussing the systematics is misleading and meaningless. When comparing two results, the relevant quantity is the total uncertainty.

Third, it's been pointed out that the connection between H0 and t0 is complicated. But suppose with more data, the community adopts the position that 71 is the right value, and that as a consequence 13.80 Gy turns into a number more like 13.75 Gy. So what?

 

[Orodruin]

And noe one has done this?

Obviously, that is how you get the age of the Universe in the first place. As I already said, what you get depends on the data set you take into account. If you want to know what has been done and how it was done, I suggest following the appropriate references in the paper you linked. That is what references are for.

 

[windy miller]

First, it's difficult to have a quantitative discussion if the numbers are shifting.

Second, discussing a statistical discrepancy without discussing the systematics is misleading and meaningless. When comparing two results, the relevant quantity is the total uncertainty.

Third, it's been pointed out that the connection between H0 and t0 is complicated. But suppose with more data, the community adopts the position that 71 is the right value, and that as a consequence 13.80 Gy turns into a number more like 13.75 Gy. So what?

but does it go to 13.75 ? it might go to 12 billion years for all I know. No one has provided any information as to what the number would be if it turns out that the Planck number is the wrong one. If it turns out we think Planck is wrong then we could have the age crisis all over again and we might have a more serious problem in cosmology as a whole if that turns out to be true. So it could be important, of course it might not be important. Yes it could be a fluke or it could be it doesn't change things very much even if its not a fluke. But maybe its not a fluke and maybe it does change things and reveal something interesting. Thats why its interesting to know the answer to the question, how much time since the big bang has elapsed if the super nova data turns out to be correct?

 

[windy miller]

Obviously, that is how you get the age of the Universe in the first place. As I already said, what you get depends on the data set you take into account. If you want to know what has been done and how it was done, I suggest following the appropriate references in the paper you linked. That is what references are for.

I am not a cosmologist, just someone interested in the subject and its hard for me to find the right information, that's what i thought forums were for.

 

[Orodruin]

I am not a cosmologist, just someone interested in the subject and its hard for me to find the right information, that's what i thought forums were for.

I believe we have given you about as much information as you can expect from a forum. We are not going to sit down and gather the data and solve the differential equations for you. In order to find people who have, you should follow the references, this should not be too difficult if you are indeed interested in the subject. If I remember correctly, there are even online calculators that let's you compute cosmological history based on the current energy content and other observables.

 

[Vanadium 50]

but does it go to 13.75 ? it might go to 12 billion years for all I know.

That's not possible.

1. The present uncertainty is a bit less than 100 My. That includes the H0 uncertainty.
2. Back when the Hubble constant was unknown to a factor of 2, there was only about a billion year uncertainty.

 

[Bandersnatch]

Using Jorrie's calculator (link) to naively plug in different values for H0, and keeping all the parameters of the model as from PLANCK 2015 data, the H0 value of 70.5 netts 13.26 Gy.
It should be noted that the discrepancy with LSS data is relatively low. Riess et al. 2016 gives 73.24 +/- 1.74 from combining a range of datasets.
Plugging that in brings the age down a full billion of years.

The question would be whether there is any value in such naive plugging in of different values of H0 obtained from local observations, given how all the other parameters are from PLANCK data.

(I think cosmologists are wrong to call it the age of the universe since we don't really know whether the universe did begin at the big bang or had some prior history)

It is generally understood that life doesn't pop up from nothing at conception, but is a continuation of other living cells with their own history. But nobody would argue that giving the age of a cat or a human is wrong, poorly defined or not useful.

(edit: typo)

 

[PeterDonis]

I initially posted the wrong values I have now corrected this. But I don't see what the relevance is to the question. the question was do the different value of the Hubble parameter mean we should have two values for the time since the big bang? We've had more than a dozen posts in this thread now and no one has actually given a clear answer answer to the question in hand

They have now. See below.

1. The present uncertainty is a bit less than 100 My. That includes the H0 uncertainty.

2. Back when the Hubble constant was unknown to a factor of 2, there was only about a billion year uncertainty.

@windy miller, this responds to the substantive question you were asking.

However, that does not mean all of your remarks quoted above were warranted. You started a thread with specific numerical claims. You were asked to provide references for those specific numerical claims. The proper reponse would have been to do so--which, as you admitted, would have resulted in your changing the claims since you initially posted the wrong values. Irrelevant complaints about being asked for references only prolonged that process.

Also, if your question was not dependent on any specific values for the Hubble parameter, you shouldn't have quoted specific values in your OP. You should have just asked the question that you really wanted the answer to. If you post specific numbers, people are going to assume you care about those specific numbers, and are going to want to see where you got them. The general idea that cosmologists are trying to pin down the value of the Hubble parameter might be "widely known", but that does not mean the specific numbers you quoted are. In short: don't quote facts unless you are prepared to back them up. If you find yourself wanting to say "I heard somewhere" or "I read somewhere" or "I understand", that should be a red flag to you that you should check first before posting.

Please bear these remarks in mind for future threads.

 

[kimbyd]

As I understand Planck has a value for Hubble as 68 kilometers per second per megaparsec and the supernova data is saying 71 . I also thought the time since the big bang (I think cosmologists are wrong to call it the age of the universe since we don't really know whether the universe did begin at the big bang or had some prior history) was determined by the value you get for Hubble. So why do people always talk about 13.8 billion years? Why isn't there two lengths of time stated if there are two radically different values for Hubble?

It's important to look at the errors, not just the values.

Also, as others pointed out, please provide references. The Planck numbers are easy enough to examine, but there are a large number of supernova data releases. Which specific ones are you referring to?

That said, supernova data isn't very good at constraining the Hubble constant. When doing the analysis, the Hubble constant is degenerate with the intrinsic brightness of the supernovae. So if they get their calibration of supernova brightness off by a few percent, then their estimate of the Hubble constant will be off by the same amount.

Anyway, Wikipedia has a list of a bunch of different observed values of the Hubble constant:
https://en.wikipedia.org/wiki/Hubble's_law

All of the most precise values cluster very close to the Planck mission's value.

 

[Vanadium 50]

Using Jorrie's calculator (link) to naively plug in different values for H0, and keeping all the parameters of the model as from PLANCK 2015 data

That's actually the wrong thing to do. The H0 data is more constraining to the composition of the universe (the "other parameters") than t0. I am not an expert in this, and haven't solved all the coupled differential equations myself, but understand that H0 is more constraining on the composition of the universe (the various Omega terms) than the age, once you include all the other data.

 

[windy miller]

That's not possible.

1. The present uncertainty is a bit less than 100 My. That includes the H0 uncertainty.
2. Back when the Hubble constant was unknown to a factor of 2, there was only about a billion year uncertainty.

Thanks you for this . Can I just clarify what you mean by the "present uncertainty" because the error bars quoted Planck and the Super Nova data are both less than 1 km/s/Mpc but the difference between the two methods is 2.67 . So which are you referring to when you get the 100 million year figure?

 

[windy miller]

Using Jorrie's calculator (link) to naively plug in different values for H0, and keeping all the parameters of the model as from PLANCK 2015 data, the H0 value of 70.5 netts 13.26 Gy.
It should be noted that the discrepancy with LSS data is relatively low. Riess et al. 2016 gives 73.24 +/- 1.74 from combining a range of datasets.
Plugging that in brings the age down a full billion of years.

The question would be whether there is any value in such naive plugging in of different values of H0 obtained from local observations, given how all the other parameters are from PLANCK data.It is generally understood that life doesn't pop up from nothing at conception, but is a continuation of other living cells with their own history. But nobody would argue that giving the age of a cat or a human is wrong, poorly defined or not useful.

(edit: typo)

Thanks you Bandersnatch this is very interesting.

 

[Bandersnatch]

That's actually the wrong thing to do.

I know. That's why I wrote it's a naive approach of questionable value. I was restating what I think the OP had in mind.

 

[Vanadium 50]

Can I just clarify what you mean by the "present uncertainty" because the error bars quoted Planck and the Super Nova data are both less than 1 km/s/Mpc but the difference between the two methods is 2.67

See message #7.

Furthermore, as we keep telling you (and you keep ignoring) a change of x% in H0 does not correspond to a change of x% in t0.

 

[Haelfix]

It is expected that future LIGO/VIRGO measurements will over time (say a decade), be able to arbitrate between these competing stories.

 

[alantheastronomer]

I understand your confusion Windy, two different values for the Hubble constant should result in two different ages for the age of the universe. The supernovae study found something very interesting however; it discovered that the expansion of the universe is accelerating as time goes on. This means that the universe was expanding slower at earlier times. If we use the present day value of the Hubble constant to calculate the age of the universe - the accelerated value - then we'll underestimate the age of the universe because the universe was expanding slower earlier on.
I'm sorry, I don't have a calculated value for the age of the universe for you. For that I think you'll have to look for publications by Adam Reiss.
The values you mention, 68 and 71, while not in agreement are much better than the way it used to be. Before, the discrepancy used to be between 50 and 100!
This is because, while both used the cosmological distance ladder to reach their conclusions, they disagreed on one crucial factor; the distance to our nearest neighbor - the Andromeda galaxy. This galaxy is located near the plane of the Milky Way in the sky and so it might be dimmed by dust in the plane of the Milky Way. This dimming, by the way, is called "extinction". deVaucolours, who championed a value of 100 for the Hubble constant, didn't use any extinction, and so estimated a distance that was further away because it was dimmer. Sandage on the other hand tried to compensate for the dimming due to dust and got a closer distance to Andromeda. A closer distance means a slower expansion, and his value for the Hubble constant was 50.

 

[Orodruin]

If we use the present day value of the Hubble constant to calculate the age of the universe - the accelerated value - then we'll underestimate the age of the universe because the universe was expanding slower earlier on.

This is, at best, misleading. In a flat universe containing only a cosmological constant the Hubble constant is just that, a constant. In our actual universe, the Hubble rate is actually decreasing. As has already been stated in this thread, to find the age of the universe you need to solve the Friedmann equations backwards for the actual composition of the Universe.

The values you mention, 68 and 71, while not in agreement are much better than the way it used to be. Before, the discrepancy used to be between 50 and 100!

Again, as has already been stated, that the values are closer does not mean that the discrepancy is worse. You have to look at the error bars relative to the discrepancy.

the distance to our nearest neighbor - the Andromeda galaxy

The Andromeda galaxy is within our local cluster, which is a gravitationally bound structure. It is not affected by the expansion.

 

[alantheastronomer]

I wasn't saying that the Andromeda galaxy is affected by the expanding universe; it is a crucial part of the cosmological distance ladder that caused the large discrepancy in the two values found for the Hubble constant that is a part of the history of science.

 

[alantheastronomer]

I never said that the discrepancy was worse the closer the values were.