A Box of Universe
Watch the cosmos evolve in a cube one billion light-years wide
Evolution of a Universe
The first results from the Bolshoi simulation appear in a series of papers by Klypin, Primack and their colleagues. (The first two papers have been published in Astrophysical Journal and are also available as preprints at the Bolshoi web site, http://hipacc.ucsc.edu/Bolshoi.) The data generated by the simulations, including the snapshots, catalogs of halos, and “merger trees” tracing the provenance of halos, are also being made publicly available through a repository called the MultiDark database.
What does the Bolshoi universe look like, and how does it compare with the world we see when we look out the window? Early in the course of the simulation, some five million halos evolved, and the number later grew to 12 million as additional particles clumped together under their own gravitational attraction. But the halos too experience these forces, and so they tend to merge, leaving fewer but larger halos. At the end of the run, there were about 10 million halos left (many of which are classified as “subhalos,” existing inside other halos but retaining their identity). Halo sizes range from 1010 solar masses (comparable to a small galaxy) up to 1015 solar masses (equivalent to a large cluster of galaxies).
Statistics on the halos and the larger structures they form suggest opportunities for comparing the simulation with reality. This work is just beginning. An example concerns a survey of satellite galaxies. The Milky Way has two prominent satellites, the Large and Small Magellanic Clouds of the Southern Hemisphere. Observational data from the Sloan Digital Sky Survey suggest that this situation is somewhat rare; only about 10 percent of galaxies in the mass range of the Milky Way have two such companions. A study of the Bolshoi results by Michael T. Busha and others yields a similar probability. The model and real-world data also agree about the probability of zero satellites or one satellite.
But not every cross-check between model and reality comes out so neatly. Sebastian Trujillo-Gomez and his colleagues have compared a sample of real galaxies with simulated dark-matter halos classified according to rotational velocity, a property that is closely correlated with total mass and luminosity. Over most of the range, the two samples are in accord. But Bolshoi predicts a few too many galaxies at the highest rotational velocities. More worrisome is a larger discrepancy at the other end of the scale: The simulation predicts an overabundance of dwarf galaxies by a factor of almost 10.
For now this conflict remains unresolved. There are several candidate explanations. The cause might turn out to be a problem in the simulation. On the other hand, the observational survey could have some unrecognized selection bias, and the Bolshoi result could be correct. Or the cause might lie at a deeper theoretical level. For example, the difference might be explained if the dark matter pervading the universe is not cold but lukewarm. If this last possibility begins to look promising, then we could be on the cusp of yet another major shift in cosmological thinking.