WMAP and Planck comparision

[Richard D. Saam]

It is noted that Planck data (no error bars)
does not fall within the WMAP(9yr) error bars
although none of this was known a few years ago,
the similarity is remarkable.

WMAP(9yr) Planck

H0 (km/s/Mpc) 69.7 2.4 67.15

Omega_b 0.0462 0.0026 0.049

Omega_c 0.237 0.026 .268

Omega_lambda 0.716 0.028 .683

t0 (Gyrs) 13.75 0.012 13.8

RDS

[Mod. note: non-ASCII characters removed. Please post in ASCII -- mjh]

[Jos Bergervoet]

On 3/23/2013 9:08 AM, Richard D. Saam wrote:
> It is noted that Planck data (no error bars)
> does not fall within the WMAP(9yr) error bars
> although none of this was known a few years ago,
> the similarity is remarkable.
>
> WMAP(9yr) Planck
>
> H0 (km/s/Mpc) 69.7 2.4 67.15
> Omega_b 0.0462 0.0026 0.049
> Omega_c 0.237 0.026 .268
> Omega_lambda 0.716 0.028 .683
> t0 (Gyrs) 13.75 0.012 13.8

They all seem to agree at almost exactly 1
sigma, except t0 which has 4 sigma.. But
only if we assume Planck data have no error
(as you do?) and treat the measured values
as uncorrelated (which is not likely).

And if they *are* correlated, the correlation
matrix, H, can give us small or large error
bars depending on whether we use
1/sqrt(H_ii) or sqrt((H^-1)_ii)
as explained in:
http://ned.ipac.caltech.edu/level5/Sept01/Orear/Figures/figure2.jpg

I would say the normal choice is to report
the large errors for each individual quantity
(treating the others as unknown). Replacing
this with the correct error ellipsoid would
then give even less agreement. (If you keep
treating Planck as error free, that is!)

We could also omit all errors. Your remark:
"the similarity is remarkable" would still
hold. :-)

--
Jos

[Phillip Helbig---undress to reply]

In article <mt2.0-25969...@hydra.herts.ac.uk>, "Richard D. Saam"

<rds...@att.net> writes:

> It is noted that Planck data (no error bars)
> does not fall within the WMAP(9yr) error bars
>

> WMAP(9yr) Planck
>
> H0 (km/s/Mpc) 69.7 2.4 67.15
>
> Omega_b 0.0462 0.0026 0.049
>
> Omega_c 0.237 0.026 .268
>
> Omega_lambda 0.716 0.028 .683
>
> t0 (Gyrs) 13.75 0.012 13.8

If these are 1-sigma WMAP error bars, no big deal. If they were 3-sigma
or more, it might be interesting.

> although none of this was known a few years ago,
> the similarity is remarkable.

What wasn't known? What similarity?

[Richard D. Saam]

I put "What wasn't known" in the context of the life of
Georges-Henri Lemaitre was born 17 July 1894
He died on the 20 June 1966.
He thought of an expanding universe with remnant energy in the 1920's.

In 1964, Arno Penzias and Robert Wilson, Robert H. Dicke, Jim Peebles,
and David Wilkinson discovered the background emanating from the
universe origin. Legend has it that Lemaitre heard the news on this
death bed. He was probably happy with this residual energy existence
without going into many details but surely anticipated the further logic
of COBE, WMAP and Planck.

Data in arXiv:1212.5226v1 [astro-ph.CO] 20 Dec 2012
was generally reported with 95% confidence level.

The Planck numbers are different enough from WMAP to warrant a new
perspective on universe origin in the case of large scale fluctuations.
http://sci.esa.int/science-e/www/object/index.cfm?fobjectid=51555

"While the observations on small and intermediate angular scales agree
extremely well with the model predictions, the fluctuations detected on
large angular scales on the sky -- between 90 and six degrees -- are about
10 per cent weaker than the best fit of the standard model to Planck
data. At angular scales larger than six degrees, there is one data point
that falls well outside the range of allowed models. These anomalies in
the Cosmic Microwave Background pattern might challenge the very
foundations of cosmology, suggesting that some aspects of the standard
model of cosmology may need a rethink."

Could these large scale fluctuations be an indication
of a second non homogeneous phase?

RDS

[Mod. note: non-ASCII characters, you know the drill -- mjh]

[Phillip Helbig---undress to reply]

In article <mt2.0-11720...@hydra.herts.ac.uk>, "Richard D. Saam"

<rds...@att.net> writes:

> The Planck numbers are different enough from WMAP to warrant a new
> perspective on universe origin in the case of large scale fluctuations.
> http://sci.esa.int/science-e/www/object/index.cfm?fobjectid=51555
>
> "While the observations on small and intermediate angular scales agree
> extremely well with the model predictions, the fluctuations detected on
> large angular scales on the sky -- between 90 and six degrees -- are about
> 10 per cent weaker than the best fit of the standard model to Planck
> data. At angular scales larger than six degrees, there is one data point
> that falls well outside the range of allowed models. These anomalies in
> the Cosmic Microwave Background pattern might challenge the very
> foundations of cosmology, suggesting that some aspects of the standard
> model of cosmology may need a rethink."

Right, but this lack of power at large scales is not something new with
Planck, but was known from WMAP and even from COBE. Cosmic variance is
large at large angular scales (low l) so it is not clear what this
means; deviation at smaller angular scales would be more serious.

[Nicolaas Vroom]

Op zondag 24 maart 2013 09:45:15 UTC+1 schreef Richard D. Saam het volgende:

>
> The Planck numbers are different enough from WMAP to warrant a new
> perspective on universe origin in the case of large scale fluctuations.
> http://sci.esa.int/science-e/www/object/index.cfm?fobjectid=51555
>

I have also performed a comparison between WMAP and Planck data.
The results are at:
http://users.telenet.be/nicvroom/WMAP%20Planck.htm

The comparison shows that both the positive and negative peaks
of the Planck data are higher than the WMAP results.
That the number of hot and cold spots is also higher.
For Cobe data the same results are lower.

The main question to answer is how can you calculate the cosmological
parameters based only on these CMBR data ?

Nicolaas Vroom
http://users.pandora.be/nicvroom/

[Phillip Helbig---undress to reply]

In article <mt2.0-6732...@hydra.herts.ac.uk>, Nicolaas Vroom

<nicolaa...@pandora.be> writes:

> I have also performed a comparison between WMAP and Planck data.
> The results are at:
> http://users.telenet.be/nicvroom/WMAP%20Planck.htm
>
> The comparison shows that both the positive and negative peaks
> of the Planck data are higher than the WMAP results.
> That the number of hot and cold spots is also higher.
> For Cobe data the same results are lower.

But this is all due to the increase in resolution from COBE through WMAP
to Planck, right?

> The main question to answer is how can you calculate the cosmological
> parameters based only on these CMBR data ?

Use CAMB, for example. Can you get different and perhaps better results
by incorporating non-CMB data? Sure. Can one calculate all the
interesting cosmological parameters with just the CMB data? Yes.

[Nicolaas Vroom]

Op vrijdag 12 april 2013 09:35:55 UTC+2 schreef Phillip Helbig---undress to reply het volgende:

This whole exercise consists of 2 steps:
1. First you have to calculate the Power Spectrum using CMBR data
2. Secondly you have to calculate the parameters using the
power spectrum
(Of course a method to calculate without the power spectrum is prefered)

I tried to calculate the Power Spectrum only using the temperatures
of the equator. I failed. I do not know if my method is correct.
Ofcourse in reality you should use many different circles.
If I supply a list of 360 temperatures would you
(or some one else ?) be able to tell me what the answer should be ?

To calculate the parameters is a much more difficult question.
You could use the program CAMB
See: http://lambda.gsfc.nasa.gov/toolbox/tb_camb_form.cfm
but that is not what I want.
First I would like to write my own. I think that that is very
difficult which immediate supports my claim that to understand
CMBR data completely is very difficult.
To write a program to calculate the same using SN 1A data is relatif
simple.
To use the program CAMB is also not that simple, because it means you
need to write your program in Fortran.
Using Camb Web interface:
When you try H = 70 age of universe = 13.738
When you try H = 70.1 age = 13.730
Which of the two is better compared with observations (step 1)
That is a difficult question to answer.
Next I change the parameter He Fraction. Same questions.
All in all a very complex exercise.

Nicolaas Vroom
http://users.telenet.be/nicvroom/friedmann%27s%20equation.htm#Q13

[craig.m...@gmail.com]

On Thursday, April 11, 2013 3:30:48 PM UTC-4, Nicolaas Vroom wrote:
> Op zondag 24 maart 2013 09:45:15 UTC+1 schreef Richard D. Saam het volgende:
>

> The comparison shows that both the positive and negative peaks
> of the Planck data are higher than the WMAP results.
> That the number of hot and cold spots is also higher.
> For Cobe data the same results are lower.

You are operated on JPG and PNG (consumer) images. Even worse, the
images have a false color scale. These images are totally uncalibrated
for scientific work. The best way to analyze images is in FITS format,
which retains the full numerical precision of the data.

WMAP data are available from the LAMBDA archive, and Planck from the
IPAC Planck archive. Unfortunately they are also in a very
non-intuitive sky projection called HEALpix. I think the free (and
professional!) image viewer DS9 will handle this projection.

Craig

[Phillip Helbig---undress to reply]

In article <mt2.0-27081...@hydra.herts.ac.uk>, Nicolaas Vroom

<nicolaa...@pandora.be> writes:

> This whole exercise consists of 2 steps:
> 1. First you have to calculate the Power Spectrum using CMBR data

Right, extract the observed power spectrum from the data.

> 2. Secondly you have to calculate the parameters using the
> power spectrum

Right, compare with theoretical power spectra for various values of the
cosmological models, minimize the differences and hence fit for the
cosmological parameters.

> (Of course a method to calculate without the power spectrum is prefered)

Why?

> To calculate the parameters is a much more difficult question.
> You could use the program CAMB
> See: http://lambda.gsfc.nasa.gov/toolbox/tb_camb_form.cfm
> but that is not what I want.
> First I would like to write my own. I think that that is very
> difficult which immediate supports my claim that to understand
> CMBR data completely is very difficult.

Indeed. CAMB represents several person-years of work by people highly
trained in physics.

> To write a program to calculate the same using SN 1A data is relatif
> simple.

That much is true.

> To use the program CAMB is also not that simple, because it means you
> need to write your program in Fortran.

I consider writing my program in Fortran to be easier. :-)

[Jos Bergervoet]

On 4/12/2013 5:19 PM, Nicolaas Vroom wrote:
...

>> Use CAMB, for example. Can you get different and perhaps better results
>> by incorporating non-CMB data? Sure. Can one calculate all the
>> interesting cosmological parameters with just the CMB data? Yes.
>
> This whole exercise consists of 2 steps:
> 1. First you have to calculate the Power Spectrum using CMBR data
> 2. Secondly you have to calculate the parameters using the
> power spectrum
> (Of course a method to calculate without the power spectrum is prefered)
>
> I tried to calculate the Power Spectrum only using the temperatures
> of the equator. I failed. I do not know if my method is correct.

I think you used colors of some Jpeg image of the
sky you found somewhere, but not the measured data!

> Of course in reality you should use many different circles.

> If I supply a list of 360 temperatures would you
> (or some one else ?) be able to tell me what the answer should be ?

If you have any list of temperatures for different
points of the sky, it would be quite easy to find
a least squares fit to spherical harmonics, which
is essentially the spectrum you want. The points
don't have to be exactly equidistant and it's
allowed to have different error bars for different
points.

Share with us the list of points, then I'll show
you how to set up the code!

>..

> To use the program CAMB is also not that simple, because it means you
> need to write your program in Fortran.

Doesn't every student get a Fortran course anymore,
these days?! And Fortran is one the 3 languages
fully supported by the most widely used compiler
suite (gcc) which is completely free to use. Would
you prefer having to write a Mathematica program,
after donating $1000 to Mr. Wolfram for a licence?

>..

> All in all a very complex exercise.

All in all, you may be right. Isn't that what makes
it interesting?

--
Jos

[Nicolaas Vroom]

Op zaterdag 13 april 2013 07:32:03 UTC+2 schreef Jos Bergervoet het volgende:

> On 4/12/2013 5:19 PM, Nicolaas Vroom wrote:
>
> > Of course in reality you should use many different
> > circles.
> > If I supply a list of 360 temperatures would you
> > (or some one else ?) be able to tell me what the answer
> > answer should be ?
>
> If you have any list of temperatures for different
> points of the sky, it would be quite easy to find
> a least squares fit to spherical harmonics, which
> is essentially the spectrum you want. The points
> don't have to be exactly equidistant and it's
> allowed to have different error bars for different>
> points.
>
> Share with us the list of points, then I'll show
> you how to set up the code!

That is what I have done.
The next link shows examples of 5 circles.
2 Circles are based on WMAP data and 3 of Planck data
http://users.telenet.be/nicvroom/WMAP%20Planck.htm#PSC
Each circle contains 360*4 temperature values.

At the beginning of that page you see an improved version
of the comparison between WMAP and Planck data.
Specific if you compare Picture 3 (WMAP) with Picture 4
(Planck) you see that Planck data shows much more detail.
In fact the difference between the two is rather large.
The main reason is that WMAP data is approximate between
-200 and +200 micro Kelvin, while Planck data is
between -500 and +500 micro Kelvin.

At the end of the document I have also tried to calculate
the sinus functions of each circle.
The sinus function is: R(l)*sin(l * alpha * pi/180 + beta(l))
with l going from 1 to 60 (or 1 to 500)
The object is to calculate R(l) and beta(l) which shows
the smallest error with observation.
The method I use is some type of Monte Carlo random selection
curve fitting approach.
When I repeat this method for the same circle the results are
more or less the same.
When I compare this method for identical circles of WMAP and
Planck data the results are also more or less the same.
When I compare different Planck circles (10,45 and 90) the
results are very difficult. This is more or less as expected.

Nicolaas Vroom
http://users.pandora.be/nicvroom/