Volume limited sample using BGS survey (clustering catalog & fastspecfit vac)

Question

Hello,

I am attempting to create a volume-limited sample by applying an absolute r-band magnitude cut to BGS galaxies.

First, I use the 'flux_r_dered' and redshift columns from the clustering catalog. 

Then, by matching 'TARGETID' between the catalog and the FastSpecFit VAC, I retrieve 'ABSMAG10_DECAM_R' and 'KCORR10_DECAM_R' from the VAC. 

However, I cannot reproduce the absolute magnitude using the K-correction values and flux_r_dered. (M_abs = m_r - DM - K_corr ?)

I found the FastSpecFit documentation here, but it only provide short explanations for each column:

https://fastspecfit.readthedocs.io/en/latest/fastspec.html

1. Is there a more comprehensive reference that explains the details of each column in the FastSpecFit VAC?

2. In the clustering catalog, which column is typically used when constructing a volume-limited sample?

Thank you in advance.

Answer 1

To fully answer your question, it would be helpful to know if they are almost all off by some constant factor, or if they are scattered, or something else.

A known issue that should only affect a small percentage of the data is that the LSS catalogs use the tile-based redshifts and fastspecfit uses the healpix-based ones. But, that would result in exact agreement for almost all of the data (because most are single tile spectra), some small disagreement for a small fraction (because we don't expect BGS redshifts to change much from increased S/N), and large disagreement for a very small fraction (~< 1% of the multi-tile spectra).

(From your description, I doubt that is your issue, but might face it at some point.)

Comments

thank you for your comment!

Answer 2

In order to minimize the dependence of the rest-frame photometry on the best-fitting spectral-energy distribution model, FastSpecFit computes the K-correction for each object from the observed-frame bandpass which minimizes the K-correction, not from the same bandpass as you've assumed. You can find more discussion of this issue (which still needs to be documented more properly---sorry) in this ticket--https://github.com/desihub/fastspecfit/issues/202

If you have any additional questions for your specific use case, please feel free to follow up here or to open a new ticket.

Comments

Thanks a lot! As you pointed out, I reproduce the absolute magnitude by using different reference bands (e.g. M_R = m_G - DM - K_R).