************************************************
Virgo Consortium
Hubble Volume Simulations
Cluster Catalogs from Light Cone Datasets
************************************************
Contents:
- Light-cone surveys
- Cluster Catalog Files
- Cluster Finding Algorithm
+ iparent flag
- Additonal Cluster Data
-----------------------------------------------------------------------
For a decription of the cosmological models simulated, see the 'expert'
page at http://www.physics.lsa.umich.edu/hubble-volume/
-----------------------------------------------------------------------
==================
Light-cone Surveys
==================
Five separate light-cone 'sky' surveys are produced by each run. Their
properties are summarized in the following table -
Survey type angular R_max (Mpc/h) z_max
coverage tcdm/lcdm tcdm/lcdm
---------------------------------------------------------
octants (PO,NO) pi/2 ster 2000/3000 1.25/1.46
oct+wedge (POW *) 76x16 sqdeg 6000 - /6.8
spheres (MS,VS) 4pi ster 1000/1500 0.44/0.58
deep wedge (DW) 100 sqdeg 3400/5100 4.60/4.37
* available for lcdm only. The POW survey is a combination of the PO
octant (extending to z=1.46) plus a 16x76 sq deg wedge extending to
twice the simulation cube length (6000/h Mpc) using periodic
boundaries. The center of the wedge is co-aligned with the x-y
diagonal plane in the simulation cube.
The surveys are centered at the following locations in the original
simulation cube (mapped to 0-1 interval) -
PO/POW has origin (0,0,0) and is directed toward (1,1,1)
NO has origin (1,1,1) and is directed toward (0,0,0)
VS has origin (0,0,0) and uses octants from that location plus the
7 other vertices (periodic replications of the origin)
MS has origin (0.5,0.5,0.5)
DW has origin (0,0,0) and is directed toward (1,1,1)
=====================
Cluster Catalog Files
=====================
Cluster catalogs derived from a SPHERICAL OVERDENSITY method applied
to the light-cone particle data. Cluster lists are found in files of
the form
grpSO.d$delta_c/$model.$survey.msort$min
where
$delta_c = { 200, 500 } is the defining density threshold
$model = { tcdm, lcdm } is the cosmological model
$survey = { DW, VS, MS, NO, PO/POW } is the choice of sky survey
$min = { 12, 48 } is the minimum particle count of the cluster
Each file contains cluster information
# m zred sigma ip x y z vx vy vz
#
where
m = number of particles (particle mass is 2.22/2.25e12/h for tcdm/lcdm)
zred = redshift
sigma = measured 1D velocity dispersion
ip = parent flag (see below)
x,y,z = cluster location in SURVEY COORDINATES in 0-1 units
vx,vy,vz= physical peculiar velocity in km/s
In the cluster data sets, survey coordinate systems are defined so
that the survey origins are located at
PO/NO/DW : (0,0,0)
VS/MS : (0.5,0.5,0.5)
Cluster locations in the ORIGINAL cube volume can be recovered by
NO : x -> 1-x
VS : x -> mod( (x-0,5), 1 )
PO/DW/MS : no transformation required
=========================
Cluster Finding Algorithm
=========================
Clusters are identified using a spherical overdensity method with the
following steps:
1) measure distance to 8th nearest particle = r8
2) sort in r8 (increasing order) & identify first particle
3) center group on this particle
i) sort neightboring particles about this center
ii) cut particle list at overdensity threshold \delta_c (1)
iii) flag particles as belonging to this group (2)
4) find next particle in r8 list not flagged as a group member
5) repeat steps 3 & 4 until max r8 is reached (3)
Notes:
------
(1) \delta_c is RELATIVE TO CRITICAL DENSITY \rho_c(z) for both LCDM and tCDM
(2) particles can belong to more than one group
(3) defined by comparing density constrast using r8 to \delta_c threshold
### LIMITATIONS ###
No attempt is made to clean near boundaries of the datasets. Some
incompleteness near the planar and radial edges of the surveys is
present. If volume completeness is desired, stay interior to ~3/h Mpc of
any boundary.
+ IPARENT flag
------------
The issue of ovelapping clusters is handled as follows. A cluster
is defined to be a parent if it is the largest of any group of
clusters with which it overlaps. Isolated clusters are parents by
default. The `family' criterion employed is that the threshold spheres
of pair i & j overlap:
| cen_i - cen_j | <= Rvir_i + Rvir_j
Note that the center of any cluster cannot lie within the threshold
radius of any cluster other than itself.
======================
Additonal Cluster Data
======================
The catalog files are shortened versions of the original group finder
output. The group finder processes planes of cubes in the
Z-direction. Separate output files for each Z-slice are produced.
These files can be made available upon request.
For each slice iz, there are three output files:
z$iz.base : basic cluster data
z$iz.extra : internal cluster structure data
z$iz.iparent : list of binary values for iparent flag (see above)
The *.base files contain:
# M {X Y Z}_c {Vx Vy Vz V}_com sigma1D R R.5 S.5 {X Y Z}_mb {X Y Z}_com
#
where
M = cluster mass in particles (1 particle = 2.2e12/h Msol in both models)
{X Y Z}_c = cluster center (1)
{Vx Vy Vz V}_com = center-of-mass velocity components and speed (2)
sigma1D = one-dimensional velocity dispersion (2)
R = cluster radius (1)
R.5 = ratio of half mass radius to full radius R_{1/2}/R
S.5 = ratio of velocity dispersions within R_{1/2} and R
{X Y Z}_mb = position of most bound cluster member (1,3)
{X Y Z}_com = position of cluster center-of-masss (1)
Notes:
(1) in box units (L=1). Box is (2/3)000/h Mpc for (t/L)cdm models.
(2) in km/s
(3) determined using N^2 sum of potential among group members only.
The *.extra files contain moments of the internal density and velocity:
# {Rxx Rxy Rxz Ryy Ryz Rzz} {Vxx Vxy Vxz Vyy Vyz Vzz}
#
where
Rij = Sum_k (R-R_c)_i (R-R_c)_j / M
Vij = Sum_k (V-V_com)_i (V-V_com)_j / M
and the sum is over members k running from 1 to M.
The *.iparent files contain the integer iparent flag for each cluster.
original: 19-May-99, evrard
last update: 18-Jul-00, evrard