Smoothing of stellar particles?

[kateharborne]

When working in a regime of low particle numbers along a spaxel, it would be useful to be able to reconstruct the distribution along the line of sight. This feature would require an assumption from the user that the stellar properties should be smoothed across some kernel relating to the local number density of particles.

Possible feature to be added to SimSpin directly:

• [ ] If PartType4 contains a SmoothingLength group (i.e. if the user has precomputed a smoothing length for each stellar particle), SimSpin can run the n_spawn generation of extra stellar particles to sample the smoothed distribution.

Required tests:

• [ ] Check that smoothing allows the user to approach the intended properties in the observation through comparison of high resolution models with low resolution smoothed models.

[kateharborne]

Generated new branch `dev-StellarSmoothing' for development of smoothing stellar particles to increase the spaxel resolution for models that are kinematically resolved, but observed using SimSpin with low particle numbers per pixel.

Planned changes:

• [ ] Add option to make_simspin_file to generate smoothed stellar distribution. Using FNN package (get.knn) to determine the smoothing length for each stellar particle. Necessary to add another dependency.
• [ ] Need to decide/test the appropriate number of nearest neighbours. For gas particles in the EAGLE simulation suite, k = 54?