Introduction

Structure formation and clustering evolution

Through existing deep surveys with HST and various ground-based telescopes, as well as the results of cosmological numerical simulations, we now have a rough qualitative picture of how galaxies and larger structures formed throughout the history of the Universe.

In hierarchical structure formation models such as the CDM theory, small density fluctuation grow and collapse to form virialized objects. In the early Universe, galaxy-scale density fluctuations with the highest peaks collapse earlier and form stars rapidly and efficiently due to the high density of the Universe. These early galaxy formation processes are related to larger-scale density fluctuations, since the highest galaxy-scale peaks tend to lie on the highest peaks over larger scale if the phase distribution of the initial density fluctuations is random (as assumed normally). The strong clustering of Lyman Break Galaxies as well as the existence of old massive galaxies seen in present-day cluster cores strongly support this picture. On the other hand, galaxy formation proceeds more slowly in lower-density regions, and galaxy infall from the low-density to high-density regions causes the apparent evolution of the galaxy population in high-density region (i.e., the Butcher-Oemler effect). Various other observational results, such as the evolution of galaxy clustering strength with redshift and the number density evolution of field galaxies, also support this idea.

Existing deep survey data is poorly suited to addressing this fundamental, global issue due to the limited and non-contiguous volume coverage. The SXDS is a unique opportunity to map the general history of structure formation.

At z > 1, the contrast of a galaxy cluster above the field galaxies is substantially reduced, making optical (or near-infrared) color-selection impractical. In addition, at high redshift cluster members may still be undergoing substantial star formation which will affect their optical-infrared colors -- consequently, optical cluster searches (which assume that cluster members form a coeval, evolved population) may be substantially incomplete. Finally, the large scatter in the X-ray luminosity-optical cluster richness plane means that optical selection may not be a suitable way to detect large concentrations of mass.

X-ray imaging provides a simpler way to detect clusters of galaxies, since a typical core diameter of 500 kpc corresponds to ~60'' at z=2, much larger than the XMM-Newton PSF of ~12''. This means that even the most distant clusters will be flagged as extended objects, provided enough source photons are collected. Since the sensitivity of the XMM-Newton observations is high enough to detect clusters with Lx=5×10⁴³ erg/s at z~2, we have the opportunity to examine in detail the same volume of the Universe at optical/infrared and X-ray wavelengths to see the evolution of galaxy clustering. We have the following specific questions to be answered by the survey:

1. Mass mapping of the distant Universe by X-rays/galaxies/weak lensing.
   We can compare the distribution of galaxy concentrations detected by multicolor analysis (using the `matched filter' method), X-ray clusters, and clusters detected by weak-lensing S/N mapping. Comparing results from the weak-lensing method (for which the excellent image quality of Subaru is important) with others is very important since the lensing method provides the real `mass' mapping and we can directly follow the M/L evolution of the dense regions of the Universe.
2. Revealing the environmental dependency of the galaxy population at high redshift.
   As a result of our extensive photometric data and high photo-z accuracy, we can split our galaxy catalogue according to further criteria, such as environment and/or color. We can therefore investigate the evolution of the galaxy population in different environmental regions.
3. Evolution of the galaxy clustering strength.
   Modern cosmological models include star formation prescriptions and calculate the observed colors of galaxies. Our large survey area and accurate photometry (at 26th magnitude, the limits of most surveys, our photometric uncertainty will be only a few per cent) will enable us to produce large galaxy samples over a wide luminosity range to compare with the predictions of models.