1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %% ws-p9-75x6-50.tex : 17-04-98
3: %% This Latex2e file rewritten from various sources for use in the
4: %% preparation of the (larger [9.75''x6.50'']) single-column proceedings
5: %% Volume, latest version by R. Sankaran with acknowledgements to Susan
6: %% Hezlet and Lukas Nellen. Comments/Suggestions to:rsanka@wspc.com.sg
7: %%
8: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
9: %
10: \documentclass{ws-p9-75x6-50}
11:
12: \newcommand{\hmpc}{$h^{-1}$Mpc}
13:
14: \begin{document}
15:
16: \title{Quantitative Analysis of the Void Size Distribution in the LCRS
17: and in CDM Models}
18:
19: \author{V. M\"uller, S. Arbabi-Bidgoli}
20:
21: \address{Astrophysikalisches Institut Potsdam, D-14482 Potsdam, Germany\\
22: E-mail: vmueller@aip.de, sarbabi@aip.de}
23:
24: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
25: % You may repeat \author \address as often as necessary %
26: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27:
28: \maketitle
29:
30: \abstracts{
31: We have analyzed the size distribution of voids in 2D slices of the Las Campanas
32: Redshift Survey (LCRS). The characteristic sizes show a scaling $D = D_0 +
33: \nu\lambda$ with the mean galaxy separation $\lambda$. Comparison with mock
34: samples of the LCRS in 2D and 3D, using various simulations, cosmologies and
35: galaxy identification schemes, gives a similar scaling, but with steeper slope
36: and a lack of large voids. Best results are obtained for dark matter halos in a
37: $\Lambda$CDM model.}
38:
39: \section{Voids in the LCRS}
40:
41: The appearence of large voids in the galaxy distribution of extended galaxy
42: surveys as the Las Campanas Redshift Survey (LCRS, Shectman et
43: al.~\cite{LCRS96}) was one of the most spectacular findings characterizing the
44: large-scale matter distribution of the universe. We expect large voids due to
45: the matter outflow from positive primordial potential
46: perturbations~\cite{MDGM98,LS98}. A second mechanism for the lack of galaxies
47: in voids is the lower probability of galaxy formation in underdense regions
48: known as biasing. Recently, Friedmann and Piran~\cite{FP00} presented a simple
49: model for void formation based on the probability distribution of spherical
50: underdense regions in the universe, and the less abundance of density peaks
51: therein. There it was aimed to explain that in recent redshift surveys voids
52: have radii between 13 and 30 \hmpc, occupy 50\% of the volume, and that the
53: probability of larger voids falls off exponentially. Since the theoretical
54: conclusions are based on a simple analytical model, and since galaxy surveys
55: before LCRS were not extended enough to get reliable void statistics, we have
56: undertaken a new quantitative study of voids in the LCRS and in numerical
57: simulations~\cite{MAET00}. Here this analysis is evaluated and compared with
58: theoretical expectations.
59:
60: In our analysis~\cite{MAET00} we have used automatic void finders in 2D. Voids
61: are identified as connected empty cells on a grid, where first square base voids
62: are identified, and then boundary layers are added if they are empty and larger
63: than 2/3 of the previous extension. The void finder starts with the search for
64: the largest voids. The algorithm was first described and tested in 2D using
65: slices through the CfA galaxy redshift survey by Kauffmann \&
66: Fairall~\cite{KF91}.
67:
68: From the LCRS we have selected 14 volume limited subsets in different radial
69: ranges of the six narrow slices, projected them into the central planes and
70: corrected for variations in the sampling fraction. In Fig.~\ref{f1} the void
71: distribution of such sets is shown which illustrates that we got a large number
72: of voids in each data set, and that the voids cover smoothly the surface of the
73: slice, with no obvious inhomogeneities or biasing by boundary effects or radial
74: gradients. To emphasize the importance of large voids, we constructed the size
75: distribution as cumulative coverage of the survey area with voids. Results are
76: shown in Fig.~\ref{f2}. In the left pannel we overplot the void distributions
77: of all 14 samples in normalizing them with the median void sizes from each
78: sample. The thick lines show the results from the best sampled part of the
79: survey. Altogether we get a nice convergence to a mean behavior shown by a
80: smooth parametric fit. Also the derivative is shown in comparison with the
81: distribution of voids in random point distributions which shows the main result
82: of the gravitational clustering of galaxies in overdense regions: The relative
83: abundence both of large voids and of small voids in the data is larger than in
84: the random point distributions (the dashed lines).
85:
86: \begin{figure}[t]
87: \epsfxsize=35pc
88: \vspace{-1cm}
89: \hspace{-2cm}
90: \epsfbox{f1.eps}
91: \vspace{-2cm}
92: \caption{Large voids in a well sampled LCRS slice. Voids are empty regions
93: marked by square base voids and rectangular extensions. Galaxies are shown by
94: dots connected with the minimal spanning tree to emphasize the overdense
95: regions}
96: \label{f1}
97: \end{figure}
98:
99: \section{Comparison with simulations}
100:
101: We have compared the LCRS voids with cold dark matter simulations in spatially
102: flat $\Lambda$-term cosmology ($\Lambda$CDM, $\Omega_0=0.3$) and in open models
103: (OCDM, $\Omega_0=0.5$). PM simulations were performed in 500 $h^{-1}$Mpc boxes
104: using $300^3$ particles and a threshold bias prescription to identify `galaxies'
105: with single particles. In addition we analyzed P3M simulations in 280
106: $h^{-1}$Mpc boxes and with $256^3$ particles, a friend-of-friend halo finder, and a
107: procedure to split off unvirialized halos that result from particle clumps
108: merged by numerical effects. The mock samples in the different simulations fit
109: well the 2-point correlation function of the LCRS as already shown in Tucker et
110: al.~\cite{Tu97}. In the right pannel of Fig.~\ref{f3}, we compare the real
111: space correlation function of the mock samples with the reconstructed 3D
112: correlation function of APM galaxies~\cite{B96}. The mock samples provide good
113: fits to the observations especially on large scales.
114:
115: The cumulative void size distribution of ten OCDM mock samples is shown in the
116: right pannel of Fig.~\ref{f2} together with the variance range. Obviously this
117: model has difficulty in matching the void distribution for a large range from 15
118: to 35 \hmpc. The $\Lambda$CDM fits better, but it also seems to underestimate
119: the size of the largest voids.
120:
121: \begin{figure}[t]
122: \epsfxsize=30pc % will enlarge or reduce the figures based on the xsize
123: \epsfbox{f2.eps} % postscript image file name
124: \caption{Cumulative area covered by voids as function of the size $D$ compared
125: to voids in a random point distribution ({\it dashed lines}). ({\bf a}) LCRS
126: void sizes measured in terms of the median size together with a fit ({\it solid
127: line\,}) and its derivative. ({\bf b}) LCRS histograms compared to mock samples
128: of OCDM ({\it dash-dotted line with errors}) and $\Lambda$CDM ({\it thick line})}
129: \label{f2}
130: \end{figure}
131:
132: \section{Scaling Relation}
133:
134: The 14 analyzed volume-limited data sets show variations in the galaxy densities
135: which are due to different absolute magnitude cuts and different sampling
136: fractions. We already showed~\cite{MAET00} that differences in the magnitude
137: ranges have negligible influence on the size distribution of voids. So we
138: ascribed the differences in the parameters of the void distributions to
139: variations in the galaxy sampling densities. We measured it by the mean galaxy
140: separation $\lambda$. In analyzing both data and mock samples, we found a
141: characteristic scaling relation for the mean void sizes that cover 25\%, 50\%
142: and 75\% of the survey area. This scaling is a simple linear relation $D = D_0
143: + \nu\lambda$, with size $D_0=(6,12,17)\,h^{-1}{\mbox{Mpc}}$ and slope
144: $\nu=(0.9,1.1,1.5)$ for 25\%, 50\% and 75\% of void coverage, cp.
145: Table~\ref{voids}.
146:
147: In Fig.~\ref{f3} we show that the scaling relations in simulated OCDM mock
148: samples are steeper, $\nu_{model} > \nu_{data}$. It is even steeper for the
149: $\Lambda$CDM model as seen in Table~\ref{voids}. According to our analysis, the
150: median size of voids is larger than reported from earlier surveys~\cite{FP00}.
151: The largest voids have a size of $D_{max}\approx 50\,h^{-1}{\mbox{Mpc}}$ not met
152: in simulations. Tests with mock samples indicate that the 2D-void size
153: distributions in LCRS slices have a fixed ratio to the 3D-voids. Further, we
154: have shown that large voids increase in size by 10 \% in redshift space, but
155: small voids remain unchanged, cp. also~\cite{SRM01}. Our mock samples were
156: always analyzed in redshift space.
157:
158: \begin{table}[t]
159: \caption{Parameters of the similarity relation of the void size distribution.
160: \label{voids}}
161: \begin{center}
162: \footnotesize
163: \begin{tabular}{|c|cc|cc|rc|}
164: \hline
165: {\rule[-1pt]{0mm}{3ex} data} &
166: \multicolumn{2}{|c|}{median} & \multicolumn{2}{|c|}{lower quartile} &
167: \multicolumn{2}{|c|}{upper quartile} \\
168: \hline
169: {\rule[-1pt]{0mm}{3ex}}
170: & $D_0$ & $\nu$ & $D_0$ & $\nu$ & $D_0$ & $\nu$ \\
171: & \hmpc & & \hmpc & & \hmpc & \\
172: \hline
173: {\rule[-1pt]{0mm}{3ex}}
174: LCRS & $11.8 \pm 2.9$ & $1.1 \pm 0.2$ & $5.7 \pm 1.6$ & $0.9 \pm 0.1$
175: & $16.8 \pm 2.9$ & $1.5 \pm 0.2$ \\
176:
177: Poisson & $0.9 \pm 1.0$ & $1.8 \pm 0.1$ & $0.9 \pm 0.3$ & $1.2 \pm 0.1$
178: & $2.1 \pm 0.5$ & $2.1 \pm 0.1$ \\
179: $\Lambda$CDM & $7.6 \pm 0.9$ & $1.5 \pm 0.1$ & $2.9 \pm 1.2$ & $1.1 \pm 0.1$
180: & $14.2 \pm 1.2$ & $1.8 \pm 0.1$ \\
181: OCDM & $10.7 \pm 1.8$ & $1.2 \pm 0.2$ & $4.8 \pm 0.6$ & $1.0 \pm 0.1$
182: & $15.7 \pm 2.7$ & $1.8 \pm 0.2$ \\
183: \hline
184: \end{tabular}
185: \end{center}
186: \end{table}
187:
188: \begin{figure}[t]
189: \epsfxsize=30pc % will enlarge or reduce the figures based on the xsize
190: \epsfbox{f3.eps} % postscript image file name
191: \caption{({\bf a}) Median ({\it squares}) and quartiles ({\it triangles}) of
192: the void sizes versus the galaxy separation $\lambda$ and the scaling relations
193: as thick lines. The circles ({\it filled for the median}) for the OCDM
194: mock samples show smaller void sizes for low $\lambda$.
195: ({\bf b}) Comparison of the real space two-point correlation function in the APM
196: survey compared with the $\Lambda$CDM (solid line) and OCDM (dashed-dotted line)
197: mock samples.}
198: \label{f3}
199: \end{figure}
200:
201: \section{Discussion}
202:
203: The void statistics provide important characteristics of the galaxy distribution
204: in low to medium dense regions where the correlation function is less sensitive.
205: The self-similarity of the size distribution of samples with different mean
206: galaxy separations culled by random dilutions and the scaling relation of the
207: percentile sizes represent a genuine feature of hierarchical clustering. The
208: discrepancy of the simulated void distribution to observations requires a
209: stronger biasing of simulated galaxies and a higher concentration of them in
210: superclusters, as has also been seen by analyzing overdensity regions with the
211: minimal spanning tree technique illustrated in Fig.~\ref{f1}, cp. Doroshkevich
212: et al.~\cite{DMRT99}.
213:
214:
215: \begin{thebibliography}{99}
216:
217: \bibitem{LCRS96} S.~A. Shectman, S.~D. Landy, A. Oemler, D.~L. Tucker, H. Lin,
218: R.~P. Kirshner, P.~L. Schechter , {\sl ApJ} {\bf 470}, 172 (1996).
219:
220: \bibitem{MDGM98} S. Madsen, A.G. Doroshkevich, S. Gottl\"ober, V. M\"uller,
221: {\sl A\&A} {\bf 329}, 1 (1998).
222:
223: \bibitem{LS98} J. Lee, S. Shandarin, {\sl ApJ} {\bf 505}, L75 (1998).
224:
225: \bibitem{FP00} Y. Friedmann, T. Piran, {\sl ApJ} in press, astro-ph/0009320
226: (2000).
227:
228: \bibitem{MAET00} V. M\"uller, S. Arbabi-Bidgoli, J. Einasto, D. Tucker,
229: {\sl MNRAS} {\bf 325}, 280 (2000).
230:
231: \bibitem{SRM01}
232: J.D. Schmidt, B.S. Ryden, A.L. Melott, {\sl ApJ} {\bf 546}, 609 (2001).
233:
234: \bibitem{KF91} G. Kauffmann, A.P. Fairall 1991, {\sl MNRAS} {\bf 248}, 313 (1991).
235:
236: \bibitem{Tu97} D.T. Tucker, A. Oemler, R.P. Kirshner, H. Lin, S.A. Shectman,
237: S.D. Landy, P.L. Schechter, V. M\"uller, S. Gottl\"ober, J. Einasto,
238: {\sl MNRAS} {\bf 285}, L5 (1997).
239:
240: \bibitem{B96} C.M. Baugh, {\sl MNRAS} {\bf 280}, 267 (1996).
241:
242: \bibitem{DMRT99} A.G. Doroshkevich, V. M\"uller, J. Retzlaff, V. Turchaninov,
243: {\sl MNRAS} {\bf 306}, 575 (1999).
244:
245: \end{thebibliography}
246:
247: \end{document}
248: