Files for "Observed Extra Mixing Trends in Red Giants are Reproduced by the Reduced Density Ratio in Thermohaline Zones"

This repository contains most of the files used to create the content in the paper "Observed Extra Mixing Trends in Red Giants are Reproduced by the Reduced Density Ratio in Thermohaline Zones" by A. Fraser, M. Joyce, E. Anders, J. Tayar and M. Cantiello. Pre-print at arXiv:2204.08487

Running the MESA grids

The code (inlists, etc.) used to run the various MESA grids in this paper can be found in the mesa_models/ folder. The workflow for creating the main grids of models (spanning 4 thermohaline mixing prescriptions in a [Fe/H] x Mass grid) is as follows:

1. cd mesa_models, then run python3 generate_run_directories.py. This creates a new folder ("work_thermohaline_...") for each choice of mixing prescription, mass, metallicity, etc. To do so, a copy of the template_work_thermohaline folder is created, and the MESA inlist is modified with the appropriate choices as specified in the Python file.
2. For each run directory WORK_DIR, cd WORK_DIR then follow the full workflow laid out in the evan_pleiades_job_submit file. This includes:
   1. Cleaning, making, and running the MESA model (./clean, ./mk, ./rn). Sometimes things go wrong or weird with the copying or syncing to github, so you may have to make these files executable (e.g., chmod +x mk). We ran all models using MESA version mesa-r21.12.1.
   2. Decreasing the amount of disk space required to store the MESA outputs (python3 mesa_to_hdf5.py, python3 hdf5_to_lite_hdf5.py). The first command creates an hdf5 file of all of the MESA outputs, which reduces disk space of outputs by a factor of ~5 (filename hdf5_LOGS.h5). The second file further reduces disk space by only preserving some of the profile fields for only a small portion of the star around the relevant region for the analysis in this paper (hdf5_lite.h5).
   3. Post-processing the lite hdf5 outputs (python3 find_r_vs_time.py), which produces the movie shown in e.g., figure 7 of the paper as well as a file r_vs_time.h5 which contains time traces of r, R0, tau, log g, etc.
   4. Removal of the original MESA LOGS/ directory to clean up disk space.
   5. Creation of movies as in e.g., fig 7 using a the png2mp4 script originally created by Keaton Burns.
3. After this process is completed and r_vs_time.h5 is created, we move the work directory into the post directory (cd .., cd WORK_DIR post/).
4. Once the full grid has been moved into post/, we run python3 generate_csv_data.py, which generates the various csv files which are stored in the Empirical-Magnetic-Thermohaline/figures/mesa_spread.csv folder.

To conduct the resolution test shown in appendix B, all of the same steps were followed, but using the generate_run_directories.py file in mesa_models/resolution_test/. In step 3, finished work directories are moved to post/resolution_test/, and the csv file is created using the generate_csv_data.py file there. The output csv files are moved to figures/resolution_test/.

The python workflow depends on the numpy, scipy, and hdf5 packages.

Creating figures

To create figures 1, 2, 4-6 in the manuscript, navigate to the figures/ directory then run ./generate_figures.sh. Figure 3 is created using elder magic (IDL) and files to create it were uploaded by J. Tayar in commit 3736ee6. Figure 7 is created during the MESA post-processing pipeline described above.