Skip to content

Visualization

Make a density-weighted temperature map of a TNG50 galaxy cutout at redshift zero. THOR reconstructs a Voronoi mesh from the gas cells and traces parallel rays through it.

1. Configure

Use the galaxy cutout from Prepare. Download projection.yaml into your working directory.

projection.yaml 
# Project a TNG50 cutout onto a temperature map. [AI-Codex]
dataset_type: pointcloud_voronoi
driver_type: raytracer
device: cpu

pointcloud_voronoi:
  hilbert_bits: 8
  loader: gadget
  gadget:
    path: ./tng50-halo625-gas-stars-v2.hdf5
    fields: [PositionX, PositionY, PositionZ, Temperature, Density]
    zoom_box: [[0.4945, 0.5055], [0.4945, 0.5055], [0.4919, 0.5029]]

raytracer:
  dist_max: 1.0
  max_step: 0.01
  outputpath: temperature.zr
  overwrite: true
  operators:
    projections:
      fields: [Temperature]
      mode: manual
      view: orthogonal
      widths: [0.264, 0.176]
      npixels: [768, 512]
      position: [0.5, 0.5, 0.01]
      direction: [0.0, 0.0, 1.0]
      up: [0.0, 1.0, 0.0]
      use_weighting: true
      perform_averaging: true

zoom_box encloses the cutout in a cube. Camera positions and widths use fractions of that domain. The view looks along the z-axis, centered on the galaxy, and covers about 150 × 100 kpc at 768 × 512 pixels.

use_weighting and perform_averaging together give the density-weighted mean temperature along each ray: \(\int \rho T\,dl / \int \rho\,dl\).

2. Run

thor projection.yaml

A successful run ends with Thor finished. and writes temperature.zr. Rerunning replaces that output (overwrite: true).

3. Inspect the result

Download plot_projection.py and thorplot.py into the same directory, then run:

uv run --python 3.11 --with zarr --with numpy --with matplotlib plot_projection.py

Density-weighted temperature map of the TNG50 cutout Density-weighted temperature map of the TNG50 cutout

The colors show density-weighted temperature in kelvin on a logarithmic scale. The scalebar marks 20 physical kiloparsecs. The array is stored at projections/temperature inside the Zarr output.

4. Things to try

Increase the resolutionRender a finer map of the same galaxy.

In projection.yaml, change npixels: [768, 512] to npixels: [1536, 1024]. Run the simulation and plotting command again. This traces four times as many rays over the same field of view, giving a finer image without changing the gas mesh.

Map H I column densitySee where neutral hydrogen is concentrated.

Neutral-hydrogen column-density map Neutral-hydrogen column-density map

Copy projection.yaml to hi-column.yaml and set raytracer.outputpath to hi-column.zr. Add HIdensity to pointcloud_voronoi.gadget.fields, then change these settings under raytracer.operators.projections:

fields: [HIdensity]
use_weighting: false
perform_averaging: false

This gives \(N_{\mathrm{HI}} = \int n_{\mathrm{HI}}\,dl\) in \(\mathrm{cm^{-2}}\), using the cutout's neutral-hydrogen abundances.

Run THOR with hi-column.yaml, then download plot_hi_column.py and run:

uv run --python 3.11 --with zarr --with numpy --with matplotlib plot_hi_column.py

The output image hi-column.png shows the neutral gas column for comparison with the temperature map.

Continue in the Cookbook

Turn a single projection into a movie by moving the camera around or through the galaxy.

See the projection settings for camera and field options. For particle-based 360 images, see native equirectangular SPH.