Write SO map based simulations paper

[zonca]

@simonsobs/mbs @jodunkley

Plan is to have a short summary by end of this week, so we can submit to publication committee, I'll post updates here.

Data release:

• The main product is 201906_highres_foregrounds_extragalactic_tophat
• The second is noise, unfortunately 201906_noise_no_lowell has both old hitmaps and outdated noise figures.
  I would like to have the new noise sims completed (see Plan new noise simulations #21), also I would like to produce them directly in the new pixelization (variable nside, see Increase resolution for next runs #18). At this point I would also rerun 201906_highres_foregrounds_extragalactic_tophat with the new pixelization, it is pretty quick to do.
• Should we release the simulations for Planck (201909_highres_foregrounds_extragalactic_planck_deltabandpass) as well? (I would release them)

[zonca]

I have a first draft, unfortunately I would like to submit this tomorrow, can you please take a quick look?

Read link: https://www.overleaf.com/read/vqzpytsvwmhh
Write/review link on #mbs on Slack

Also attach a PDF for convenience: Simons_Observatory_Map_Based_Simulations_paper_summary (1).pdf

[zonca]

I made a small update, added expected number of noise sims and section about analysis
Simons_Observatory_Map_Based_Simulations_paper_summary.pdf

Sent to TAC and @aiolasimone

[zonca]

Got positive feedback from @aiolasimone and @jodunkley, sent to SO publication panel now

[zonca]

got approval last month from publication panel. Deadline is end of June.

Need to provide timetable with expected completion dates of all the tasks to TAC.

[zonca]

Under request from @erminiacalabrese I started to layout the structure of the paper and assigned people sections of the paper to contribute.

For now I have @NicolettaK @marcelo-alvarez @smsimon @mhasself,
I'll email you the write access link, the read-only link is:

https://www.overleaf.com/read/vqzpytsvwmhh

No hurry, I will need them at the beginning of May.

[zonca]

See below an estimate of the timeline, I guess they are accurate to ~50%,
it depends also on how my other projects are going (I work on SO ~20%).
Based on this timeline:

• Delivery of foreground simulations, mid-May
• Noise simulations in CAR, beginning of June
• Paper draft, end of June

Estimated Timeline

Noise simulations

2 weeks if @msyriac has time to finalize the Pull Request, 4 weeks otherwise

• Generate noise simulations directly in CAR (finalize Fixes to CAR hitmap interpolation order and sky fraction calculation + homogenous scan strategy debug mode galsci/mapsims#39)

Foreground simulations

~1.5 months - includes validation, can be descoped and use tophat and done in 1 month, but I'd rather not

• Implement into PySM:
  • run foreground model just ~10 equally spaced points in bandpass
  • weight with full bandpass ~200 points
• Upgrade galactic and extragalactic models to 8192
• Rotate to CAR foreground simulations (finalize Simulations natively in CAR #13)

Paper writing

3 weeks

• First complete draft

[erminiacalabrese]

Excellent, thank you!

…

On 25 Mar 2020, at 01:34, Andrea Zonca ***@***.***> wrote: See below an estimate of the timeline, I guess they are accurate to ~50%, it depends also on how my other projects are going (I work on SO ~20%). Based on this timeline: Delivery of foreground simulations, mid-May Noise simulations in CAR, beginning of June Paper draft, end of June Estimated Timeline Noise simulations 2 weeks if @msyriac <https://github.com/msyriac> has time to finalize the Pull Request, 4 weeks otherwise Generate noise simulations directly in CAR (finalize galsci/mapsims#39 <galsci/mapsims#39>) Foreground simulations ~1.5 months - includes validation, can be descoped and use tophat and done in 1 month, but I'd rather not Implement into PySM: run foreground model just ~10 equally spaced points in bandpass weight with full bandpass ~200 points Upgrade galactic and extragalactic models to 8192 Rotate to CAR foreground simulations (finalize #13 <#13>) Paper writing 3 weeks First complete draft — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#25 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGBRWJC4FJK7AMWECTBAZ7TRJFNTDANCNFSM4JC7HADA>.

[zonca]

Due to COVID, time off I took, work on other projects, and issues found in the noise simulations, all the MBS work is heavily delayed.

Estimated timeline is now:

• Rerun of the noise simulations in HEALPix: by June 24th in time for the MBS tutorial
• Run of the noise simulations in CAR: by mid-July
• Foreground simulations: by end of August
• Paper draft: mid-September

[zonca]

Current status

Noise

It is composed of 1 realization on disk plus software (mapsims) to produce more: https://github.com/simonsobs/map_based_simulations/blob/master/202006_noise/README.md.

• In HEALPix this is ready
• the software should be tested by external people to check that it works fine and documentation is sufficient
• In CAR it is implemented, but I had a normalization issue trying to verify the results, I would need some help with a CAR expert to finalize this: Noise simulation in CAR #36

Foregrounds

I have an old release: https://github.com/simonsobs/map_based_simulations/blob/master/201906_highres_foregrounds_extragalactic_tophat/README.md

I would like to rerun this with variable NSIDE like the noise is.
The initial plan was to extend the current foreground models to NSIDE 8192 just extrapolating at the high end of the spectrum from the NSIDE 512 models included in PySM 2.

However, it is a widespread idea that those PySM 2 models have significant problems, in particular the dust (too much noise in maps and in spectral index, CIB contamination).

So I started working on new foreground components models for PySM 3, starting from dust (this is in collaboration with the Panexperiment galactic science group), see galsci/pysm#72.

So the work on simulations and the paper is on hold until I have foreground models for dust, synchrotron (started by Ben Thorne galsci/pysm#73), AME and free-free.

The problem is that developing these new models is taking a lot of time, so it will take months before I can wrap those up and run the simulation.

Paper

Once the simulations themselves are ready, I can start to work on the paper, which will explain all the inputs, the available data products and the available software, and the validation that has been performed on those.