Multiscale Probability Mapping (MSPM)
MSPM is a multi-scale mapping algorithm suited to coarse-graining the galaxy distribution**, aiding studies of the universe's structure from the not-so-large to the largest scales. The grains are gravitationally-bound groups and clusters, and together they can be used to identify filaments, voids and who knows what else. As far as we know it's the first algorithmic approach to identify filaments with a false discovery rate of less than a half.
The basic approach is to coarse-grain the galaxy distribution in order to make it easier to interpret. Some information is lost in this process, but the user is able to set the grain size, the idea being that if the user is okay with losing information below 1 Mpc/h scales (say), setting that grain size won't be a problem. Coarse-graining in this instance is distinct from smoothing because we do aim to produce discrete grains, which doesn't always happen when smoothing. Making grains at all locations indiscriminately can also be a problem, so part of the process is to look for statistically-significant overdensities.
So we get these grains from the galaxy distribution. Then what? The grains may be interesting in themselves if they are gravitationally-bound, which can be handy. One can go on to treat the grains as particles (possibly weighted by measured properties like velocity dispersion and radius) to identify structures with sizes above the grain scale. For example, underdensities might be voids and elongated configurations might be filaments. We did these and can show you here!
|Visualisations of Groups, Clusters, Filaments and Voids|
We have applied MSPM to the Seventh Data Release of the Sloan Digital Sky Survey (SDSS). The MSPM group, cluster and filament catalogues are presented in this paper.
Group and Cluster Catalogue (text)
Group and Cluster Catalogue (pdf)
Filament Catalogue (text)
Figures showing all 53 filaments
Void Catalogue (text)
Universe In A Box
The universe is rumoured to be homogeneous on this sort of a scale.
|A cube of side 160 Mpc/h at 0.06 < z < 1.12.
Left: Render showing groups and clusters. Right: Same view with the smoothed galaxy distribution (Gaussian, σ = 5 Mpc/h) clearly showing the filamentary network, and a large void. See our visualisation page for a description of what information these show. Click on the images for rotating three-dimensional animations!
**which is a landscape, yes? (from xkcd) -