We investigate several techniques to identify voids in the galaxy distribution of matter in the universe. We utilize galaxy number counts as a function of apparent magnitude and Wolf plots to search a two- or three-dimensional data set in a pencil-beam fashion to locate voids within the field of view. The technique is able to distinguish between voids that represent simply a decrease in density as well as those that show a build up of galaxies on the front or back side of the void. This method turns out to be primarily useable only at relatively short range (out to about 200 Mpc). Beyond this distance, the characteristics indicating a void become increasingly difficult to separate from the statistical background noise. We apply the technique to a very simplified model as well as to the Millennium Run dark matter simulation. We then compare results with those obtained on the Sloan Digital Sky Survey. We also created the Watershed Void Examiner (WaVE) which treats densities in a fashion similar to elevation on a topographical map, and then we allow the "terrain" to flood. The flooded low-lying regions are identified as voids, which are allowed to grow and merge as the level of flooding becomes higher (the overdensity threshold increases). Void statistics can be calculated for each void. We also determine that within the Millennium Run semi-analytic galaxy catalog, the walls that separate the voids are permeable at a scale of 4 Mpc. For each resolution that we tested, there existed a characteristic density at which the walls could be penetrated, allowing a single void to grow to dominate the volume. With WaVE, we are able to get comparable results to those previously published, but often with fewer choices of parameters that could bias the results. We are also able to determine the the density at which the number of voids peaks for different resolutions as well as the expected number of void galaxies. The number of void galaxies is amazingly consistent at an overdensity of −0.600 at all resolutions, indicating that this could be a good choice for comparing models.



College and Department

Physical and Mathematical Sciences; Physics and Astronomy



Date Submitted


Document Type





large-scale structure of universe, dark matter, methods: n-body simulations, surveys