wasral / so_noise_models

Simons Observatory N(ell) noise models for public release

so_noise_models

This repository hosts publicly released N(ell) noise curve projection code for the Simons Observatory. The intention is that the full history of noise models will be available here, to supplement published projections and simulations.

This main home of the repository is: https://github.com/simonsobs/so_noise_models.

The noise model was originally described and applied in the publication:

Simons Observatory Collaboration. "The Simons Observatory: science goals and forecasts". JCAP 1902 (2019) 056. arxiv:1808.07455

The paper may be found at the following links:

• https://dx.doi.org/10.1088/1475-7516/2019/02/056
• https://arxiv.org/abs/1808.07445

Note that only code and results marked with version code v3.0.0 will correspond exactly to the results shown in the paper. For the LAT component-separated noise spectra (LAT_comp_sep_noise/v3.1.0/), this repository contains only the results for the v3.1.0 noise model. The v3.1.x model incorporates a fix to the low-ell noise.

This repository is organized as follows:

• so_models_v3/: Python 3 package providing the noise model code used in publications and publicly released simulations. This consists of several independent sub-modules, representing each version of the noise code. The usage of the models can vary substantially from version to version -- please consult code in demos/ for typical usage patterns.
• demos/: Code that demonstrates usage of the noise models, such as by producing noise curve plots.
• LAT_comp_sep_noise/ - Effective noise power spectra for SO LAT component-separated CMB T, E, B, and Compton-y maps. See dedicated README within.
• LAT_lensing_noise/ - Lensing noise curves from SO LAT component-separated CMB T, E, B maps. See dedicated README files within.

The so_models_v3 package is pure python and thus can be imported from the root level of this repository. But you might want to install it into your Python environment. Conda users can simply run:

pip3 install .

If you're not using Conda and want to install the package to your local user package folder, run:

python3 setup.py install --user