Differences

This shows you the differences between two versions of the page.

--- 20180424_dc_maps [2018/04/24 16:46] – created pryke
+++ 20180424_dc_maps [2019/01/18 14:38] (current) – pryke
@@ Line 1: / Line 1: @@
+**Data Challenge Maps I**
 Apr 24 2018, Clem Pryke
@@ Line 9: / Line 11: @@
 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 and beam smoothing is applied to signal components with beam widths as
+Everything is nside=512.
-per the v3.2 spreadsheet.
-Noise levels are also as per the v3.2 spreadsheet.
-Everyting below is available on NERSC under /project/projectdirs/pico/.
 Under sky_yy we have the sky models where yy designates the sky model number:
@@ Line 23: / Line 25: @@
   * 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.
+.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.
+.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

Differences

Page Tools