exclusion radius for DLA with respect to background quasar

[londumas]

Following the discussion started in ticket #25, we might need to decide on an exclusion radius around the background quasar, where no DLA can lie.
The best way to assess if the peak at Delta_v ~ 0 is too sharp is to use the updated randoms once they are fixed (ticket #35).
Maybe the most straightforward exclusion radius around the background quasar would be of the size of the pixel of the large scale field, or maybe 2 or 3 times. What is the size of this @jfarr03? i.e. what is the resolution of the large scale field? Something like 2Mpc/h?

The two following plots are the distribution of the velocity difference between the backgroud quasar and the DLA from the same line-of-sight.

[londumas]

Update from v6.0/v6.0.0:

[andreufont]

Hi @londumas - As we discussed in the issue #25, it's not clear to me that there is a problem in the mocks, in the sense that it is unclear what the true distribution should look like.

The "data" plot is affected by observational artifacts: I would imagine that the difficulties in modelling the quasar continuum near the emission lines, and the presence of BAL-like features, makes it very difficult to find DLAs near the quasar.

I think any analysis should not use DLAs that are near the quasar, but that is not something that should go into the mock-making.

Do you have a measurement of the 3D cross-correlation of DLAs and quasars? It would be interesting to compare it to the prediction from linear theory, to see how large the excess is.

I would then only worry if we saw that this excess was very large, if it was an important fraction of the total number of DLAs is in the same cell than a quasar.

[londumas]

@andreufont, I really agree on your remark on the shape of the distribution. We should not aim at reproducing the shape. However, I really don't understand how we can have DLA in the same cell as a quasar. If I understood correctly the cell size is of order ~2Mpc/h, then can we set a DLA at less than 2Mpc/h of the host quasar? I think that doing so, we basically add mock QSOs to the catalog of DLAs.
Anyway, this is in no way a show stopper and we don't have to do anything for that.

[andreufont]

Well, quasars leave in massive haloes, and DLAs are common in massive haloes as well. It is not crazy to think that we'd have a fair amount of DLAs at small separations from quasars, either in infalling filaments or in satellite galaxies.

But yes, we can also add a hard cut at the minimum velocity separation between DLAs and quasars, and set it to something like 300 km/s (roughly 3 Mpc/h).

[londumas]

Here is the status in /global/projecta/projectdirs/desi/mocks/lya_forest/develop/london/v7.3/v7.3.0/.
I would advocate to not fix it, but leave this ticket open, so we just remember it exists.