WMAP Data Reduction

[TravisM]

Hi gang. In response to Nerid's request that people post questions about WMAP in a non ATM section, I came up with a line of though that may or may not be important. I've numbered them here for convienence.
1. I wondered how the data reduction is being done. I sort of understand the dipole and little less the quadrapole, then I get lost. I know about the large scale anisotropies (and was actually supprised when I first saw the raw map of COBE with the perfect division of hemispheres.)
2. I also want to know what order they've subtracted the motions of the earth about the sun, the sun about the galaxy and the galaxy toward the virgo cluster.
3. I want to know if there is a difference in the resulting map if the order of subtraction is changed.

Thanks.

[Ken G]

I don't know the technical details, but the motion of the Sun and Earth are not "subtracted", the data is "transformed" out of the moving frame. The Doppler shift is well understood, so is easy to transform away. Note there would be no need to transform the Sun's motion separately from the Earth's, so there is no "order" to worry about, you just want to remove a single resultant velocity from the data, and I presume this is done so as to leave the least imprint of motion on the data, there's no independent way to know what the relative speed "should be".

[TravisM]

Alright. So, they take the data and do a transform against vector, which itself is the sum of various motions.
Do I have it right?

[ngc3314]

Alright. So, they take the data and do a transform against vector, which itself is the sum of various motions.
Do I have it right?

As far as I (think I) know the process - they start with differential measurements (differences) between two fields of view about 60 degrees apart, which scan the sky as the spacecraft spins and precesses. This works well because not only are differential approaches good for small-scale variations, but instruments on COBE gave excellent measurements of the mean CMB temperature. These data then have to be processed to yield the all-sky pattern of relative brightness (hence temperature, since multiple frequencies are observed), folding in passes across each location in various directions. (There are various schools of thought on the best ways to do this, especially the filtering of long-term drifts in the detectors - "prewhitening"). This sky map still needs to have foreground signals removed - from galactic dust and synchrotron emission. These also use the multiple frequencies, since these have quite different and (AFAWCT) well-behaved spectral signatures. Them you see a dominant dipole, which will be the resultant of the soalr motion through the Milky Way, the Milky Way Center-of-mass motion with respect to the center of the Local Supercluster, andthe motion of the LS with respect to the mean of all the mass in the shell which emitted the CMB. (I'm not sure whether the precision is great enough for it to make sense to remove the Earth's orbital motion, which will enter in a complex way because the various spin passes over a piece of sky will come in groups, with the groups perhaps months apart). So what we end up knowing is the vector combination of all these motions. So any truly cosmological dipole anisotropy is swallowed up in the dipole due to relative motions.

[Tim Thompson]

There is no short answer to the question of how WMAP data are reduced. So I suggest the following. First, look at the WMAP Observatory Overview webpage. Follow the links from there, and you can find discussions on how the hardware works, why it was designed that way, scan strategies that define how the data are taken, and how the calibration & map making are done. You really need to know something about the instrument & what the data really are, before you can appreciate why the data reduction is done the way it is.

Next, go the the First Year Technical Papers webpage. There you can download copies (in PDF or PostScript) of the technical papers that will tell you exactly how the data reduction is done. If you aren't used to reading papers like this, it can be a slow process. Going over the webpages that describe the instruments first, will make it easier.

[Nereid]

In light of how complex this mission is, it may be a good idea to take a step by step look at how WMAP

*collects photons/microwaves and feeds them into the receivers (the 'optics', the telescope and feed design)
*amplifies, 'differences', detects, and produces raw (digital) data (the 'receivers')
*compresses and transmits the raw digital data to the ground stations.

We could then look at design aspects such as:
-> frequency coverage
-> angular resolution
-> noise sensitivity
-> scan strategy
-> orbit selection.

Then we can move on to map making and map calibration.

Finally (!), we can go to the second linked webpage in Tim's post. I would recommend that we go through them in this order:
#2 (Data Processing Methods and Systematic Error Limits)
#5 (Galactic Signal Contamination from Sidelobe Pickup)
#6 (Foreground Emission)

From there it depends upon posters' interests.

This is all pretty daunting stuff, and I don't know of an easy way to summarise it, without simplifying so much that really important points get short shrift.

Perhaps it might be better to approach WMAP from COBE? The latter had a much easier to appreciate (in a relatively short time) approach.

[Fortunate]

Talk about homework!