Dust templates based on GNILC

[zonca]

Notebook to create dust templates based on GNILC.

• Check https://app.reviewnb.com/healpy/pysm/pull/68/files/ to leave line-by-line comments on the notebook.
• Check https://nbviewer.jupyter.org/github/zonca/pysm/tree/gnilc_dust/docs/preprocess-templates/ to view a read-only version of the notebooks
• Go to https://github.com/zonca/pysm/tree/gnilc_dust/docs/preprocess-templates to download the notebooks

The notebooks download the required input data so they can be directly executed, the notebook to process the templates required 10GB of RAM to execute (works fine on Jupyter@NERSC), the "compare" notebook about 4GB, they both need namaster.

I saved intermediate and final results, they are available at:

https://portal.nersc.gov/project/cmb/pysm-dev/dust-gnilc/template/ or at NERSC /global/project/projectdirs/cmb/www/pysm-dev/dust-gnilc/template

@brandonshensley @NicolettaK if you have fixes/suggestions, you can either leave comments or email me a modified version of the notebooks.

Next I'll work on spectral index and dust temperature. Afterwards we can fix a telecon to discuss them.

[review-notebook-app[bot]]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[zonca]

see https://github.com/healpy/pysm/issues/69 for dust spectral index and temperature

[zonca]

ok, I have created a single notebook that processes both I and QU just changing the value of the comp variable. In temperature this algorithm doesn't work well. I think the problem is that the GNILC template are not smoothed using a standard spherical harmonics approach and they retain extra power even after the supposed cutoff of the beam. So I think this needs a more sophisticated algorithm, also, instead of using the 80 arcminutes version, I think it would be nice to retain as much as possible the original map, and then only fill small scales after the resolution limit of the map.

The GNILC map also contains a map of the resolution in the FWHM column:

I think these notebooks are a good starting point and someone more expert in dust modelling could start from them and implement the full processing of the templates.

[zonca]

see #72