search

simonsobs / map_based_simulations

Map based simulations for the Simons Observatory
4 stars 1 forks source link

Run 2 simulations: highres foreground in equatorial, variable spectral index #6

Closed zonca closed 5 years ago

zonca commented 5 years ago

Rerun of 201903_highres_foregrounds in Equatorial

zonca commented 5 years ago

@NicolettaK @msyriac please review the README.md, maps are temporarily in /global/cscratch1/sd/zonca/simonsobs/mapsims_runs/201904_highres_foregrounds_equatorial, 512 are available, 4096 in the queue.

Plots at https://gist.github.com/zonca/0e6d1a16f4afdb1dcc7859f3262fdd9f

zonca commented 5 years ago

In this pull request I am preparing 3 simulations, check the readmes: https://github.com/simonsobs/map_based_simulations/pull/6/files?file-filters%5B%5D=.md

see above for location and plots for fixed spectral index in equatorial

NicolettaK commented 5 years ago

@zonca I've compared the power spectra of the new equatorial maps with those of the old galactic ones. In the attached file you can see the spectra for the maps at Nside=512 LAT 27 GHz. There are some differences at small scales, I'm waiting for the maps at Nside=4096 to check the behavior on those.

spectra_lat.pdf

zonca commented 5 years ago

@NicolettaK the job timed out, I only got the lowest frequencies at 4096 in the same folder, you can take a look at those.

zonca commented 5 years ago

output maps for the 3 simulations at NSIDE 512 are available at NERSC in:

/global/cscratch1/sd/zonca/simonsobs/mapsims_runs/
zonca commented 5 years ago

freefree_pysm3 @NicolettaK I have not identified the cause of that issue in the power spectrum, but I am rerunning this with PySM 3 and that issue is gone.

zonca commented 5 years ago

@NicolettaK I have replaced the old nside 512 maps with the new ones, would you check again the folder:

/global/cscratch1/sd/zonca/simonsobs/mapsims_runs/201904_highres_foregrounds_equatorial/512

the 4096 maps should be ready tomorrow

NicolettaK commented 5 years ago

@zonca, I see a sudden drop of the signal amplitude at ell~1000 in the new lat maps at Nside=512. spectra_pysm3.pdf I need to check whether this is present also at Nside=4096.

zonca commented 5 years ago

ok, nside 4096 are ready:

/global/cscratch1/sd/zonca/simonsobs/mapsims_runs/201904_highres_foregrounds_equatorial/4096

I am expanding alms up to nside*2, please check those maps, but I'm rerunning again up to 3*nside-1 which is the default for anafast

NicolettaK commented 5 years ago

ok then this is the reason for the dropping at ell~1000 for Nside=512 maps.

zonca commented 5 years ago

ok, new maps with ellmax to 3*nside-1 both 512 and 4096 are ready at:

/global/cscratch1/sd/zonca/simonsobs/mapsims_runs/201904_highres_foregrounds_equatorial

I moved the others in the subfolder ellmax_2nside

zonca commented 5 years ago

@NicolettaK ok, variable spectral index maps are ready as well:

/global/project/projectdirs/sobs/v4_sims/mbs/201904_highres_foregrounds_variable_spectral_index

if you check them I'll wrap up a release of these two and switch to work on extragalactic

NicolettaK commented 5 years ago

@Zonca I've checked the maps with spatially constant spectral index and now there is a prefect agreement with the old Galactic maps. I'll try to have a look at the spatially varying in the weekend.

zonca commented 5 years ago

ok, this is completed, if you find any problem please open an issue in the repo

NicolettaK commented 5 years ago

@zonca, I've just started to check the maps with variable spectral indices. I have realized that having different pivot frequencies for synchrotron temperature and polarization (for T is at 408 MHz, while for P is at 23 GHz) is not ideal, as this causes larger distortion in temperature when you apply a spatially varying scaling. Probably it is better to set the pivot frequency at 23 GHz for both T and P. I'll modify the model and do a PR. That means that once this is done we need to run the sims again.

zonca commented 5 years ago

ok, no problem, can you please add a note to the README in the meantime https://github.com/simonsobs/map_based_simulations/tree/master/201904_highres_foregrounds_variable_spectral_index and warn people not to use those maps?

NicolettaK commented 5 years ago

@zonca could you copy the following files: dust_T_ns4096_353GHz.fits dust_T_ns512_353GHz.fits synch_T_ns4096_23GHz.fits synch_T_ns512_23GHz.fits from /global/cscratch1/sd/krach/SObs/FGawg/so_template to the so_pysm_models template directory?

zonca commented 5 years ago

copied

NicolettaK commented 5 years ago

PR https://github.com/simonsobs/so_pysm_models/pull/38# should solve the problem of pivot frequency for spatially varying spectral indices. @zonca could you run the full sims starting with those at Nside=512?

zonca commented 5 years ago

@NicolettaK ok, theyare at:

/global/cscratch1/sd/zonca/simonsobs/mapsims_runs/201904_highres_foregrounds_variable_spectral_index/512

zonca commented 5 years ago

@NicolettaK also 4096 available at the same path