Next Contents Previous

D. About simulations

Ever since the first simulations by Aarseth, Gott and Turner we have gazed upon and admired simulations looking "as good as the real thing". We were impressed by the gravitational growth of clustering and we were impressed by the fact that the two-point correlation function exhibited a power law of approximately the right slope.

Subsequent developments explored the dependency of the results on initial conditions and extended significantly the range of length scales over which we could apply our value-judgements. There has also been a clearer discrimination between dark matter (the stuff of simulations) and the luminous matter (the stuff we observe). To this has been added exceptional computer graphics to render the simulations as "observed samples". They look as good as the real thing.

Several caveats apply. First, simulations provide three space and three velocity coordinates for each mass point at each time. Data provide two (angular) space coordinates and a redshift, which is made up of two terms, one proportional to the third spatial coordinate (distance) and one representing motion of the point (galaxy or cluster) relative to uniform cosmic expansion. These can be separated only within some model of what real (rather than N-body) clusters ought to be doing in the way of a Virial theorem or some other way of parcelling out potential and kinetic energy among the mass points.

Second, between the simulations of what the (mostly dark) matter is doing and data on what luminous galaxies are doing lies all of what one might call gaseous astrophysics (or even gastrophysics). The intermediate territory includes inflow of baryons into the potential wells, star formation and wind energy input, supernovae (which add both kinetic energy and heavy elements, which change how gas cools and condenses), galactic winds, on-going infall into the wells, systematic gas flow within galaxies, shocking of baryons plus heating and/or triggered star formation when halos interact, collide, and merge, energy input from black hole accretion, and so forth. Most of these currently defy real calculation and are represented by parameters and proportionalities. Thus the statement that some particular set of cosmological parameters, initial conditions, and prescriptions for star formation evolve forward in time to "fit the data" is not equivalent to being able to say that this is the way nature did it.

Next Contents Previous