This void catalog uses a novel algorithm to look for voids in 2D slices of the Sloan Digital Sky Survey (SDSS) galaxy distribution. Individual slices are then combined to create the final, 3D void catalog. This catalog enabled the first 5σ detections of void lensing (Clampitt and Jain 2015) and void autocorrelations (Clampitt, Jain, and Sanchez 2016). The galaxy sample used to define these voids is the DR7 Luminous Red Galaxy sample of Kazin et al. (2010).

If using this void catalog in your own work, please cite this paper which describes the void finding algorithm:
Clampitt, J., & Jain, B. 2015, MNRAS, 454, 3357

(random catalogs below)

Last Update: May 12, 2017
The only difference between this update and the old catalog is the removal of two small, disconnected patches with a few hundred voids. This ensures the geometry of the void catalog matches the random catalogs.

This table describes the columns available within the catalog FITS file.

Data Column	Definition	Units
ra	Void center (angular position)	degrees
dec	Void center (angular position)	degrees
z	Void center (redshift position)	–
r_los	Void center (redshift position)	comoving Mpc/h
R_v	radius projected on the sky	comoving Mpc/h
theta_v	radius projected on the sky	arcminutes
los_size	void size in the line-of-sight direction	comoving Mpc/h
dens_rand	random point density	(degrees)−2
f_vol	Volume overlap fraction with a larger neighbor	–

1. Require (dens_rand > 100) to remove voids near the survey edges.
2. In published papers I’ve always required the ratio of transverse to line-of-sight sizes to be within a factor of three:
   (1/3 ≤ R_v / los_size) AND (R_v / los_size ≤ 3)
   This cut removes long, pencil-shaped objects and flat, pancaked-shaped objects. These objects are more likely due to Poisson noise in the galaxy sampling rather than actual cosmic voids.

Clampitt and Jain (2015) describes the motivations for these particular cuts. In principle you could explore other cuts on dens_rand and void axis ratio.

The random catalogs for different cuts on void size and volume overlap are available below. The void size cuts are based on an effective radius that combines transverse and line-of-sight void sizes:

r_eff = (R_v²× los_size)^1/3

Make the same cuts on the void catalog (for both r_eff and f_vol) and use the corresponding random catalog. Also be sure to do cuts 1 and 2 above (under “Important cuts”) since these have also been applied to the random catalogs.

Download Random Catalog (15 Mpc/h < r_eff < 20 Mpc/h) AND (f_vol < 0.50)
Download Random Catalog (20 Mpc/h < r_eff < 25 Mpc/h) AND (f_vol < 0.50)
Download Random Catalog (25 Mpc/h < r_eff < 30 Mpc/h) AND (f_vol < 0.50)

Download Random Catalog (15 Mpc/h < r_eff < 20 Mpc/h) AND (f_vol < 0.90)
Download Random Catalog (20 Mpc/h < r_eff < 25 Mpc/h) AND (f_vol < 0.90)
Download Random Catalog (25 Mpc/h < r_eff < 30 Mpc/h) AND (f_vol < 0.90)

1. The redshift distribution of smaller voids does not match the randoms perfectly. If your measurement is very sensitive to the redshift, then the randoms for voids with reff < 25 may not be good enough.
2. There are at least 10 times as many randoms as voids for each sample. But for some samples there are as many as 20 times as many randoms as voids. So to use all voids with r_eff = 15-30 Mpc/h do not just append the random catalogs together. You must take the same proportion of random points from each subsample.