reliable minimum radius given gravitational softening
Melika Sarrami
24 Jun
Dear Dylan,
I am wondering at which softening radii the DM density profile and its inner slope are considered reliable in the TNG100 simulation. I frequently encounter references to a threshold of approximately twice the softening length. Could you tell me the reliable radii, or how I should determine them?
Thanks
Dylan Nelson
25 Jun
It depends what you mean by "reliable".
I assume you mean that the resulting value i.e. density is converged (to some tolerance) if one were to change the softening length. The only quantitative answer would be to perform a resolution study using the different resolution boxes. You can find an example of such a study in Pillepich+ (2019) regarding stellar sizes.
Mathematically, gravity becomes unsoftened at 2.8 times the gravitational softening length. You can find the softening length of dark matter in the documentation, for a given TNG simulation. However, even below this value the magnitude of softening is very small, and results could certainly be "reliable". You can find a more detailed discussion of this in Springel+ (2024) the Gadget4 paper, for example.
As a rule of thumb, I would suggest to plot radial profiles to r=0 and draw a line indicating the softening length. Caution should be used in interpreting or using values at or near this radius.
Dear Dylan,
I am wondering at which softening radii the DM density profile and its inner slope are considered reliable in the TNG100 simulation. I frequently encounter references to a threshold of approximately twice the softening length. Could you tell me the reliable radii, or how I should determine them?
Thanks
It depends what you mean by "reliable".
I assume you mean that the resulting value i.e. density is converged (to some tolerance) if one were to change the softening length. The only quantitative answer would be to perform a resolution study using the different resolution boxes. You can find an example of such a study in Pillepich+ (2019) regarding stellar sizes.
Mathematically, gravity becomes unsoftened at 2.8 times the gravitational softening length. You can find the softening length of dark matter in the documentation, for a given TNG simulation. However, even below this value the magnitude of softening is very small, and results could certainly be "reliable". You can find a more detailed discussion of this in Springel+ (2024) the Gadget4 paper, for example.
As a rule of thumb, I would suggest to plot radial profiles to r=0 and draw a line indicating the softening length. Caution should be used in interpreting or using values at or near this radius.