20180424_dc_maps [Probe Mission Study Wiki]

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
**Data Challenge Maps I**

Apr 24 2018, Clem Pryke

For CMB-S4 project we have made simulations using a number of different foreground models
plus lensed-LCDM, noise and tensors.
These are described at
[[https://cmb-s4.org/CMB-S4workshops/index.php/Data_Challenges | Data challenge summary page]]
and in
[[https://cmb-s4.org/CMB-S4workshops/index.php/Simulation_and_Forecasting_Logbook | a series of logbook postings]].

I have exploited this work for PICO to make equivalent sims.

Everything below is available on NERSC under /project/projectdirs/pico/

I first made [[https://arxiv.org/abs/1608.02841 | PySM]] input maps for the PICO
band centers as listed in the v3.2 spreadsheet at [[imageroptions]].
I did this for delta function bandwidths to keep things simple.
Everything is nside=512.

Under sky_yy we have the sky models where yy designates the sky model number:

  * 91=PySM a1d1f1s1
  * 92=PySM a2d4f1s3
  * 93=PySM a2d7f1s3
  * 96=Brandon's MHD model taken from /global/homes/b/bhensley/mhd_maps/maps_v1 on 180424
  * cmb = links to the cl's and alm's from which the LCDM component are generated (shared with Plank ffp10 sims)

Under expt_xx we just have single file 90/params.dat which specifies the instrument parameters for this
round as taken from the v3.2 spreadsheet.

Under data_xx.yy we have the sets of simulated experimental maps.
90.00 contains the lensed-LCDM (llcdm), noise (noise) and tensor (tenso) components for each band.
Noise levels are also as per the v3.2 spreadsheet.
The signal components have beam smoothing applied with beam widths as
per the v3.2 spreadsheet.
There are also combined lLCDM+noise+foreground+tensor maps (comb).
These come as four flavors. Straight "comb" has full lensing signal. The "comb_AL" variants
have the lensing signal artificially suppressed to the given levels of lensing power.
So "comb_AL0p15" is the amount of lensing PICO is supposed to have post de-lensing.
"comb_AL0p1" and "comb_AL0p03" are also provided and might be useful.

For the 90.00 case the "foreground" is just Gaussian realizations of dust and synchrotron
with uniform amplitude over the whole sky set to equal the observed amplitude in the BICEP/Keck
patch.
This is not a serious model and is intended only for test purposes.
In particular it enables:
a) To run full sky (unmasked) harmonic analysis - no need to deal with E/B mixing problems.
b) The dust and synchrotron SED's are simple and uniform over the full sky so they can
be fit as such. beta_dust=1.6, T_dust=19.6K, beta_sync=-3.1.

90.91, 90.92, 90.93 and 90.96 contains the combined maps for the proper sky models.

There are multiple realizations of each xx.yy.
For 90.00 there is a fresh foreground realization for each one.
For 90.9y only a single foreground realization is available so this is re-used and
only the LCDM, noise etc vary.
In both cases every even numbered realization has r=0.003 tensors injected.

Here are power spectrum check plots at low/mid/high frequencies (where the foreground is model 91).
To get a sensible looking plot for the foreground one has to cut out
the galactic plane.
The combined maps is not shown as this just looks like the foreground.
The task is take the set of multi band maps, clean out the foreground,
and take the power spectrum with an estimator which preserves E/B purity
on the cut sky.
Good luck!

{{:20180424:fig2a.gif}}
{{:20180424:fig2b.gif}}
{{:20180424:fig2c.gif}}

Here are plots of the four sky models with Planck real sky for comparison -
probably open them as tabs in your browser and flips between.
We see that MHD model looks something like the galaxy but lacks the synchrotron spurs.

{{:20180424:fig1_0.gif}}
{{:20180424:fig1_1.gif}}
{{:20180424:fig1_2.gif}}
{{:20180424:fig1_3.gif}}
{{:20180424:fig1_6.gif}}
User Tools

Page Tools
