Internal Discrepancies Between Merger Tree Data

Lia Formenti
  • 8 Jun '15


I've been working with the raw Sublink HDF5 files to get a feel for how the data is stored. I noticed that even if the SubhaloMass and SubhaloMassInHalfRad have non zero masses, the SubhaloMassInRad will still sometimes be zero, especially at high redshift. How does this work? If the subhalo has mass in its stellar half mass radius, then it must still have mass in the double half mass radius since this field covers the half mass radius. Some clarification in case I am interpreting the definition of these fields wrong would be appreciated. Thank you,

Lia Formenti

Dylan Nelson
  • 20 Jun '15

Hi Lia,

In Sublink these values are copied directly from the (Subfind) group catalogs, so probably this same issue is there as well.

Perhaps you can write which subhalo ID, snapshot, and run you found this issue with, and I can check. Thanks!

Lia Formenti
  • 21 Jun '15

I described my issue incorrectly above: I have found most subhalos (especially ones with small masses) will have a subhaloMass but no mass in their stellar half mass radius or double stellar half mass radius. Three of these subhalos are 496609 at snap 56, 695888 at snap 64, and 293000 at snap 50. Thank you!

Dylan Nelson
  • 22 Jun '15

You're right in what you say, about these subhalos, and this is indeed the correct information as proposed by the Subfind algorithm.

I looked in particular at 496609 where the issue is that it has no stars. So, the stellar half mass radius is zero, and then there cannot be any particles/cells within such a radius.

So wherever halfmassrad_stars is zero, there is going to be such a problem.

Depending on what you're trying to do / searching for, perhaps you should impose a restriction of a minimum stellar mass?

Lia Formenti
  • 23 Jun '15

Thanks so much, that makes perfect sense.

  • Page 1 of 1