Field:
Physics incl. High-energy and Astrophysics
Session Topic:
Virtual Galaxies
Chair:
Junichiro Makino/National Astronomical Observatory of Japan
Large-scale computer simulation has become more and more important in the study of
astronomical objects such as galaxies for several reasons. One is that recent
observations of cosmic microwave background and other cosmological "standard
candles" such as Type Ia supernovae determined basic cosmological parameters such
as the amount of dark matter and dark energy with high accuracy. This means we
now roughly know what is the initial condition of our universe.
The second reason is that advances in the computer technology and simulation
techniques made it possible to follow evolution of complex systems such as a galaxy
consisting of numerous sub structures or clusters of galaxies consisting of numerous
galaxies.
In our current understanding, galaxies and larger structures such as groups and
clusters of galaxies are all formed through gravitational instability of the expanding
universe, and these structures are formed "bottom up", since the initial density
fluctuation has the peak amplitude at very small mass. Thus, to understand the
formation and evolution of a galaxy, we need to understand how it was formed
through merging of smaller structures. Recent simulations with more than billions of
particles allows us to follow this merging history from sub-galactic scale up to the
scale close to that of the visible universe, if we limit the simulation to dark matter
particles which interact only through gravity.
On the other hand, to really simulate the formation of a galaxy, we need to model how
baryonic matter behaves within a galaxy. This means we need to solve hydrodynamic
equation for interstellar and intergalactic gas, coupled with radiation transfer, and we
need to model how stars are formed through gravitational instability and radiative
cooling, and how elements other than hydrogen and helium are formed through
nuclear reaction within stars and in particular by supernovae.
To simulate all these processes with sufficient resolution and accuracy is beyond the
current ability of current computer code on current computers, but we might reach
the goal in not so far future.
Masao Mori will present the state-of-the-art simulation of the early phase of galaxy
formation, in which all the above processes listed above are incorporated with best
model we can use now. This simulation was done on the Earth Simulator, which was
the fastest general-purpose computer in the world until 2004.
J. Devriendt will present semi-analytic model, in which gravitational structure
formation was solved by N-body simulation and for individual objects we apply more
approximate single-zone models for star formation and chemical evolution. Sine our
ability is still far from calculating the evolution of many galaxies, such approximate
techniques are essential to make useful comparison between the theory and
observation.