RoSE-Sim: Roman Semi-resolved Galaxy Simulator

RoSE-Sim is a Python package designed for creating image simulations of semi-resolved dwarf galaxies for the Nancy Grace Roman Space Telescope.

Citation

If you use Rosesim in your work, please cite it:

@software{jiaxuan_li_2026_18274779,
  author       = {Jiaxuan Li},
  title        = {AstroJacobLi/Rosesim: v1.0},
  month        = jan,
  year         = 2026,
  publisher    = {Zenodo},
  version      = {v1.0},
  doi          = {10.5281/zenodo.18274779},
  url          = {https://doi.org/10.5281/zenodo.18274779},
}

Installation

1. Install the Package

You can install Rosesim locally by cloning the repository and running:

git clone https://github.com/AstroJacobLi/Rosesim.git
cd Rosesim
pip install -e .

When installing romanisim, you have to set the following for the CRDS cache:

export CRDS_PATH="$gpfs/Data/temp/crds_cache" # or any other directory
export CRDS_SERVER_URL="https://roman-crds.stsci.edu"

To run the unit tests to make sure everything is installed correctly, run the following command under the root directory of the repository:

pytest

Caution

The customized romanisim and artpop are required for this package, which will be automatically installed when you run pip install -e .. If you don't wanna cause conflicts with your existing packages, please make a new Python environment and install them there.

2. Set Up Environment Variables

Rosesim requires a dedicated data directory to store large files (e.g., isochrones, sky models).

1. Create a directory for the data (e.g., /scratch/gpfs/user/Rosesim_Data).

2. Set the ROSESIM_DATA_PATH environment variable pointing to this directory. Add the following line to your shell configuration file (e.g., ~/.bashrc or ~/.zshrc):

   export ROSESIM_DATA_PATH="/path/to/your/data/directory"

3. Reload your shell configuration:

   source ~/.bashrc

3. Download Required Data

Once the package is installed and the environment variable is set, you can easily download the necessary data files using the built-in fetch function:

import rosesim
rosesim.fetch_data()

This command downloads the required isochrones and sky models to your ROSESIM_DATA_PATH.

Data Description

Stellar Population Synthesis Models

Rosesim uses PARSEC isochrones for stellar population synthesis.

• Isochrones: Pre-packaged for Roman filters. Note that Roman isochrones are provided in Vega magnitudes.
• Conversion: Zeropoint offsets to convert to AB magnitudes are available in rosesim.Roman_zp_AB_Vega_mist.
• AGB Stars: Special attention has been given to the resolution of thermal pulse cycles (ninTPC=20 in COLIBRI tracks) to strictly resolve AGB stars (see Lee et al. 2025).

Available Grid: (unfortunately, artpop doesn't support the PARSEC isochrones to be interpolated yet)

• Filters: F062, F087, F106, F129, F146, F158, F184, F213
• Log(Age/yr): [8.0, 8.1, 8.2, 8.3, 8.4, 8.5, 8.6, 8.7, 8.8, 8.9, 9.0, 9.1, 9.2, 9.3, 9.4, 9.5, 9.6, 9.7, 9.8, 9.9, 10.0, 10.1]
• Metallicity [Fe/H]: [-2.00, -1.75, -1.50, -1.25, -1.00, -0.75, -0.50, -0.25, 0.00, +0.25]

Milky Way Star Model

A Milky Way star catalog is generated using the TRILEGAL model.

• Default Field: RA = 10h, Dec = 2.27 deg
• FoV: 1 square degrees
• Depth: Limiting magnitude of 30 mag in the 4th filter.
• Location: Stored in ROSESIM_DATA_PATH/TRILEGAL/.

I have also downloaded the TRILEGAL models for NGC 253 (0.7925372h, -25.2888000 deg); Cen A (NGC5128, 13.4246944h, -43.0166667deg).

Background Sky Model

To simulate realistic observations, Rosesim includes a background sky model comprising:

1. Background Galaxies: From the JAGUAR mock catalog.
2. Foreground Stars: From the TRILEGAL model.

Caution

JAGUAR fluxes are currently based on JWST filters, and are converted to Roman filters using simple stellar population models. Also the background galaxies do not have any spatial clustering.

Note

A quick note on the PSF. romanisim now takes psftype as an argument, which can be either galsim, epsf, or stpsf. If your star is bright, we recommend using psftype='galsim' as its PSF box is large enough to cover the PSF wings and spikes far enough. Turning on fastpointsources=True will speed up the simulation, without compromising the PSF quality. For faint stars, we recommend using psftype='epsf' and fastpointsources=True, as it is much faster.

Usage

Command Line Interface

Rosesim provides command-line scripts for common simulation tasks. The scripts are located in rosesim/scripts/.

We strongly encourage to use psftype='galsim' for simulating bright stars, as its PSF box is larger. For dwarf galaxies, it might be better to use psftype='epsf' and fastpointsources=True.

1. Simulate a Sky Model

To generate an empty sky (for noise-only or background-free simulations):

rosesim_sky \
  --obs_ra=150 --obs_dec=0 --size=5001 --prefix='empty_sky' --exptime=642 --filters="['F106', 'F129', 'F158']" --seed=42 --include_bkg=False --include_star=False

rosesim_sky \
  --obs_ra=150 --obs_dec=0 --size=4096 --prefix='empty_sky' --exptime=600 --nexp=1 --filters="['F106', 'F129', 'F158']" --seed=42 --include_bkg=False --include_star=False

rosesim_sky \
  --obs_ra=150 --obs_dec=0 --size=5001 --prefix='empty_sky' --exptime=10272 --filters="['F106', 'F129', 'F158']" --seed=42 --include_bkg=False --include_star=False

rosesim_sky \
  --obs_ra=150 --obs_dec=0 --size=5001 --prefix='empty_sky' --exptime=5136 --filters="['F106', 'F129', 'F158']" --seed=42 --include_bkg=False --include_star=False

Generate a full sky model including background galaxies and Milky Way stars, e.g., for the sky around NGC 253:

rosesim_sky \
  --obs_ra=11.8880580 --obs_dec=-25.2888000 --size=5001 --prefix='sky_jaguar_trilegal_ngc253' \
  --exptime=642 --nexp=6 --filters="['F106', 'F129', 'F158']" \
  --seed=42 --include_bkg=True --include_star=True \
  --fastpointsources=True --psftype='galsim' --trilegal_file="trilegal_Roman_30mag_2deg2_NGC253.dat"

This corresponds to the default HLWAS depth (2 pointings, 3 dithers per pointing, and each exposure takes 107s). If you wanna simulate a deeper depth (e.g., 8 times longer exposure time), you can use the following command:

rosesim_sky \
  --obs_ra=11.8880580 --obs_dec=-25.2888000 --size=4096 --prefix='sky_jaguar_trilegal_ngc253' \
  --exptime=600 --nexp=1 --filters="['F106', 'F129', 'F158']" \
  --seed=42 --include_bkg=True --include_star=True \
  --fastpointsources=True --psftype='galsim' --trilegal_file="trilegal_Roman_30mag_2deg2_NGC253.dat"

rosesim_sky \
  --obs_ra=11.8880580 --obs_dec=-25.2888000 --size=5001 --prefix='sky_jaguar_trilegal_ngc253' \
  --exptime=5136 --nexp=48 --filters="['F106', 'F129', 'F158']" \
  --seed=42 --include_bkg=True --include_star=True \
  --fastpointsources=True --psftype='galsim' --trilegal_file="trilegal_Roman_30mag_2deg2_NGC253.dat"

This is 16 times longer exposure time:

rosesim_sky \
  --obs_ra=11.8880580 --obs_dec=-25.2888000 --size=5001 --prefix='sky_jaguar_trilegal_ngc253' \
  --exptime=10272 --nexp=96 --filters="['F129', 'F158']" \
  --seed=42 --include_bkg=True --include_star=True \
  --fastpointsources=True --psftype='galsim' --trilegal_file="trilegal_Roman_30mag_2deg2_NGC253.dat"

If you wanna exclude large galaxies that have R_e > 0.15 arcsec by using --exclude_size_thresh=0.15,

rosesim_sky \
  --obs_ra=11.8880580 --obs_dec=-25.2888000 --size=5001 --prefix='sky_jaguar_trilegal_ngc253_sizecut' \
  --exptime=642 --nexp=6 --filters="['F106', 'F129', 'F158']" \
  --seed=42 --include_bkg=True --include_star=True \
  --fastpointsources=True --psftype='galsim' --trilegal_file="trilegal_Roman_30mag_2deg2_NGC253.dat" \
  --exclude_size_thresh=0.15

For the sky around Cen A:

rosesim_sky \
  --obs_ra=201.3704160 --obs_dec=-43.0166667 --size=4096 --prefix='sky_jaguar_trilegal_cena' \
  --exptime=600 --nexp=1 --filters="['F106', 'F129', 'F158']" \
  --seed=42 --include_bkg=True --include_star=True \
  --fastpointsources=True --psftype='galsim' --trilegal_file="trilegal_Roman_30mag_2deg2_CenA.dat"

rosesim_sky \
  --obs_ra=201.3704160 --obs_dec=-43.0166667 --size=5001 --prefix='sky_jaguar_trilegal_cena' \
  --exptime=642 --nexp=6 --filters="['F106', 'F129', 'F158']" \
  --seed=42 --include_bkg=True --include_star=True \
  --fastpointsources=True --psftype='galsim' --trilegal_file="trilegal_Roman_30mag_2deg2_CenA.dat"

If you wanna simulate a deeper depth (e.g., 8 times longer exposure time), you can use the following command:

rosesim_sky \
  --obs_ra=201.3704160 --obs_dec=-43.0166667 --size=5001 --prefix='sky_jaguar_trilegal_cena' \
  --exptime=5136 --nexp=48 --filters="['F106', 'F129', 'F158']" \
  --seed=42 --include_bkg=True --include_star=True \
  --fastpointsources=True --psftype='galsim' --trilegal_file="trilegal_Roman_30mag_2deg2_CenA.dat"

16 times longer exposure time:

rosesim_sky \
  --obs_ra=201.3704160 --obs_dec=-43.0166667 --size=5001 --prefix='sky_jaguar_trilegal_cena' \
  --exptime=10272 --nexp=96 --filters="['F106', 'F129', 'F158']" \
  --seed=42 --include_bkg=True --include_star=True \
  --fastpointsources=True --psftype='galsim' --trilegal_file="trilegal_Roman_30mag_2deg2_CenA.dat"

Check this notebook if you wanna make your own sky model.

2. Simulate a Single Dwarf Galaxy

Inject a specific dwarf galaxy into an empty sky model:

rosesim_gal --obs_ra=150 --obs_dec=0 --distance=10 --log_age=10.0 --log_m_star=5 --exptime=642 --sky_model=$ROSESIM_DATA_PATH/empty_sky/

rosesim_gal --obs_ra=150 --obs_dec=0 --distance=10 --log_age=10.0 --log_m_star=5.5 --exptime=10272 --sky_model=$ROSESIM_DATA_PATH/empty_sky/

You don't need to specify the number of exposures because that is already encoded in the sky model. Make sure that your input RA, Dec matches the sky model. See rosesim_gal --help for more options.

A full list of options for simulating the dwarf galaxy is as follows:

simulate_galaxy(
    obs_ra=150,
    obs_dec=0,
    log_m_star=6,
    distance=30,
    log_age=9.0,
    feh=-1.5,
    abs_mag_lim=-1,
    filters=["F106", "F129", "F158"],
    exptime=642,
    n=0.8,
    theta=100,
    ellip=0.3,
    sky_model=DATA_PATH + "sky_jaguar_trilegal/",
)

Check this notebook if you wanna tune your dwarf galaxy's properties, such as size, age, metallicity, etc. Currently the galaxy size is fixed to follow the average mass-size relation in Carlsten+21.

Python API

You can also use the Python API to inspect or manipulate data.

Example: inspect the sky model

import rosesim

# Read the simulated ASDF file
sky_dm = rosesim.read_L3_asdf('./F158_642s.asdf')

# Convert to FITS for inspection (e.g., with DS9)
rosesim.asdf_to_fits(sky_dm, 'F158_642s.fits', subtract_bkg=True)
# This also subtracts the median background level from the image

Requirements

The package relies on the following libraries:

• numpy, matplotlib, astropy, astroquery, asdf, astrocut, roman_datamodels
• romanisim (Modified version required: GitHub)
• artpop (Modified version required: GitHub)

Future Plans

• Add stellar population info to ASDF meta.
• Simulate L2 images with different MA tables.
• Support more diverse sky background options.
• Support composite stellar populations and user-defined structural parameters.
• Expand documentation and examples.
• Enable photometry on the simulated images.

License

MIT