1: \documentclass[useAMS,usenatbib]{mn2e}
2: \usepackage{graphics}
3: \usepackage{graphicx}
4: \usepackage{multirow}
5: \usepackage{amssymb}
6: \usepackage{color}
7: \newcommand{\be}{\begin{equation}}
8: \newcommand{\ee}{\end{equation}}
9: \def \vs {{\it vs.}\ }
10: \def \ie {{\it i.e.}\ }
11: \def \etal {{\it et al.}\ }
12: \def \LCDM {$\Lambda$CDM}
13: \def \msun {M$_{\odot}$}
14: \newcommand{\hmsun}{{\,\rm h^{-1}M}_\odot}
15: \newcommand{\hmpc}{{\,\rm h^{-1}Mpc}}
16: \def\bg{{\bf g}}
17: \def\br{{\bf r}}
18: \def\bx{{\bf x}}
19: \def\bff{{\bf f}}
20: \def\bv{{\bf v}}
21:
22: \newcommand\apjl{ApJ}%
23: % Astrophysical Journal, Letters
24: \newcommand\aap{A\&A}%
25: % Astronomy and Astrophysics
26: \newcommand\apj{ApJ}%
27: % Astrophysical Journal
28: \newcommand\mnras{MNRAS}%
29: % Monthly Notices of the RAS
30: \newcommand\aj{AJ}%
31: % Astronomical Journal
32:
33:
34:
35: \begin{document}
36:
37:
38: \title{Constrained simulations of the local universe: I. Mass and motion in the Local Volume}
39: \author[Martinez-Vaquero et al.]
40: {\parbox[t]\textwidth{Luis A. Martinez-Vaquero$^1$, Gustavo Yepes$^1$ and Yehuda Hoffman$^2$}
41: \vspace*{6pt} \\
42: $^1$Grupo de Astrof\'{\i}sica,
43: Universidad Aut\'onoma de Madrid,
44: Madrid E-280049, Spain
45: \\
46: $^2$Racah Institute of Physics,
47: Hebrew University,
48: Jerusalem 91904, Israel
49: \\
50: }
51: \date{\today}
52:
53: \maketitle
54:
55:
56:
57: \begin{abstract}
58: It has been recently claimed that there is no correlation between the
59: distribution of galaxies and their peculiar velocities within the Local Volume (LV), namely a sphere of $R=7\hmpc$ around the Local Group (LG). It has been then stated that this implies that either locally dark matter is not distributed in the same way as luminous matter, or peculiar velocities are not due to fluctuations in mass. To test that statement a set of constrained N-body cosmological simulations, designed to reproduce the main observed large scale structure, have been analyzed.
60: The simulations were performed within the flat-$\Lambda$, open and flat matter only CDM cosmogonies. Two unconstrained simulations of the flat-$\Lambda$ and open CDM models were performed for comparison. LG-like objects have been selected so as to mimic the
61: real LG environment. The local gravitational field due to all halos found within each LV is compared with the exact gravitational field induced by all matter in the simulation.
62: We conclude that there is no correlation between the exact and the local gravitational field obtained by pairwise newtonian forces between halos. Moreover, the local
63: gravitational field is uncorrelated with the peculiar velocities of halos. The exact gravitational field has a linear correlation with peculiar velocities but the proportionality constant relating the velocity with gravitational field falls below the prediction of the linear theory. Upon considering all matter inside the LVs, the exact and local gravitational accelerations show a much better correlation, but with a considerable scatter independent on the cosmological models. The main conclusion is that the lack of correlation between the local gravitation and the peculiar velocity fields around LG-like objects is
64: naturally expected in the CDM cosmologies.
65: \end{abstract}
66:
67: \begin{keywords}
68: methods: numerical --
69: galaxies: Local Group --
70: cosmology: dark matter
71: \end{keywords}
72:
73:
74:
75: \begin{figure}
76: \resizebox{\linewidth}{!}{\includegraphics{F1.ps}}
77: \caption {Projected dark matter distribution for the \LCDM\ constrained
78: simulation. The image shows a projected slice of 10 h$^{-1}$ Mpc thick
79: across the box centre in supergalactic coordinates X and
80: Y. The Local Supercluster is the filamente crossing the box
81: horizonally. The position of the Local Group and the Virgo and Fornax clusters
82: are shown. }
83: \label{boxL}
84: \end{figure}
85:
86: \section{Introduction}
87:
88: A key ingredient of the standard model of cosmology is that the large
89: scale structure (LSS) has emerged out of an otherwise a homogenous and
90: isotropic universe {\it via} gravitational instability
91: (e.g. \cite{pee80}).
92: One of the main consequences of gravitational instability is that the
93: growth of structure induces a non-vanishing velocity field. The
94: standard model of cosmology relates the large scale mass density and
95: peculiar velocity fields. It further assumes that the observed galaxy
96: distribution is closely related with the underlying matter density
97: field, even if this relation is biased in some yet unknown way. In an
98: interesting recent paper \cite{whi05} (W05) has
99: challenged
100: these basic ideas of the standard model by testing them against a study
101: of the galaxy distribution and their peculiar velocities in the Local
102: Volume (LV), defined as the sphere of radius $R=7 \hmpc$ (where $h$ is the
103: Hubble's constant in units of $100\,
104: \mbox{\rm km s}^{-1} \mbox{\rm Mpc}^{-1}$) centred on
105: the Local Group (LG). W05's main conclusion is that {\em 'Either dark matter
106: is not distributed in the same way as luminous matter in this region,
107: or peculiar velocities are not due to fluctuations in mass.'}
108: This is a
109: very important result and if it
110: was
111: true it would
112: put the standard model of cosmology on very shaky foundations.
113: The main goal of our paper is to
114: examine W05 methodology and claims
115: by analyzing the kinematics and dynamics of simulated LV
116: sytems in different cosmological models. As we will show in this
117: paper, our numerical results
118: do not support
119: W05's claims.
120:
121:
122: W05 carried out an analysis of the distribution and peculiar velocities
123: of 149 galaxies in the LV. Using the high-quality data of these
124: galaxies W05 has mapped the mass distribution within the LV, assuming
125: it is traced by the galaxies, and calculated the gravitational field
126: within the LV. The gravitational field has been calculated by summing
127: over the pairwise Newtonian interaction for each galaxy and by
128: weighting the galaxies by their K and B luminosity. W05's working
129: assumption is that the peculiar velocity of galaxy should be aligned
130: with the gravitational field it experiences and that the amplitude of
131: the peculiar velocity and gravity fields should be linearly connected.
132: The notion that 'the peculiar velocity field is linear with the
133: gravitational field' is strictly valid in the linear regime of the
134: gravitational instability in an expending universe. If such a linear
135: relation had been confirmed by the high-quality data that the LV
136: provides this would have validated the idea that the LSS is indeed induced
137: by gravitational instability. No clear correlation was found
138: between the velocity and local gravitational fields. This has
139: led W05 to conclude that, at least in the LV,
140: structure has not formed by means of the gravitational instability.
141:
142: The above conclusion should be tested carefully. The main goal of the
143: present paper is to apply W05 analysis to a set of N-body constrained
144: simulations (CSs) of the local universe. The CSs are designed to
145: reproduce the gross features of the nearby universe, namely the main
146: players of the nearby LSS, such as the Local Supercluster (LSC) and the
147: Virgo, are imprinted onto the simulations \citep{kra02,mat02,kly04,dol05,hof06}.
148: This is achieved by setting the initial conditions of
149: the simulations as constrained realizations of Gaussian fields, where
150: actual observational data is used as constraints. The CSs provide an
151: almost optimal laboratory for testing W05 algorithm, as they closely
152: mimic the dynamics of the LV. CSs have been run in the framework of the
153: benchmark flat-$\Lambda$ CDM model as well as in the open CDM (OCDM) and the
154: so-called flat matter only standard CDM (SCDM) models. In addition, W05
155: procedure has been tested against standard, non-constrained,
156: simulations of the same models.
157:
158:
159: This is a first in a series of papers that focuses on studying the
160: nearby universe by means of N-body and hydrodynamical CSs. In
161: particular issues concerning the coldness of the local flow, the mass
162: distribution in the LV, the mass accretion history of the LG and the
163: future of the nearby structure in a dark energy dominated universe are
164: to be addressed. The highlight of this project will be the full galaxy
165: formation high resolution simulation of the LG.
166: Our choice of the different models is dictated by our plan to study the nature of the velocity
167: field in the LV in the CDM cosmogony and in particular its dependence on the dark energy
168: and the dark matter. In this regard the SCDM model is considered to be an extreme case and is
169: taken for reference. Forthcoming papers in this series will focus on the problem of the coldness of the local Hubble flow.
170:
171: The structure of the paper is as follows:
172: a brief description of the simulations is given in
173: \S~\ref{sec:sim}.
174: The selection of LG-like objects is described in \S~\ref{sec:obj} and the
175: analysis itself in \S~\ref{sec:analysis}. The results of the analyses
176: are given in \S~\ref{sec:result}.
177: A final discussion of the results is given in \S~\ref{sec:disc}.
178:
179: \begin{table}
180: \begin{center}
181: \begin{tabular}{ccccc}
182: \hline
183: Model & $\Omega_m$ & $\Omega_\Lambda$ & h & $\sigma_8$ \\
184: \hline
185: \LCDM & 0.3 & 0.7 & 0.7 & 0.9 \\
186: OCDM & 0.3 & 0 & 0.7 & 0.9 \\
187: SCDM & 1.0 & 0 & 0.5 & 0.7 \\
188: \hline
189: \end{tabular}
190: \caption{Cosmological parameters used for the differents CDM
191: models}
192: \label{tabmodels}
193: \end{center}
194: \end{table}
195:
196:
197:
198: \section{Simulations}
199: \label{sec:sim}
200:
201:
202: W05 has analyzed one and only one particular patch of the universe,
203: namely the LV that extends around the LG. The LG is not a unique or an
204: unusual object in the universe, yet it has its own characteristics that
205: affect the outcome of any dynamical test that would
206: be applied to it. Our
207: main goal is to check the validity of the W05 approach, in
208: the context of the environment of the LG. This is to be achieved by
209: applying the W05 analysis to LG-like objects found in appropriate
210: N-body simulations. The key to a successful study is to properly select
211: LG-like objects from the simulation and analyze the LV around these
212: objects (see \S~\ref{sec:obj}). Yet the selection of the LG-like objects
213: is based on the properties of the LG itself and not on its environment
214: within the LV. This has encouraged us to use CSs as a 'laboratory' for
215: testing the W05 procedure in conditions very similar to the ones
216: prevail in the LV and its immediate surrounding. In addition we have
217: used unconstrained simulations for reference. The comparison of the
218: results obtained in the constrained and unconstrained simulations will
219: shed light on the question of whether the findings of W05 are naturally
220: expected in CDM dominatted cosmogonies.
221:
222:
223:
224: The data used to constrain the initial conditions of the simulations is
225: made of two kinds. The first data set is made of radial velocities of
226: galaxies drawn from the MARK III \citep{mark}, SBF \citep{sbf} and
227: the \cite{kar05} catalogs. Peculiar velocities are less affected by
228: non-linear effects and are used as constraints as if they were linear
229: quantities \citep{zar99}. This follows the CSs
230: performed by \cite{kra02} and \cite{kly03}. The other constraints are
231: obtained from the catalog of nearby X-ray selected clusters of galaxies
232: \citep{rei02}. Given the virial parameters of a cluster and assuming
233: the spherical top-hat model one can derive the linear overdensity of
234: the cluster. The estimated linear overdensity is imposed on the mass
235: scale of the cluster as a constraint. Different CSs with different
236: random realizations have been calculated and they all exhibit a clear
237: and unambiguous LSC-like structure that dominates the entire
238: simulation, much in the same way as in the actual universe in which the
239: LSC dominates the nearby LSS. The simulations do vary with respect to
240: the particular details of the LG-like object that is formed roughly in
241: its actual position. All simulations used here are based on the same
242: random number realization of the initial conditions.
243:
244:
245: Five simulations have been performed so far. Three constrained ones have
246: been performed within the framework of the \LCDM, OCDM and SCDM models.
247: In addition, unconstrained simulations of the \LCDM and OCDM cosmologies
248: have been performed for the sake of comparison and benchmarking.
249: Table \ref{tabmodels}
250: presents the cosmological parameters
251: of the different simulations.
252: All simulations correspond to a periodic cubic box of 64 h$^{-1}$ Mpc on a
253: side. We made a random realization of the corresponding power
254: spectrum for each cosmological model with a large number of particles
255: ($2048^3$).
256: At the same time, a constrained realization of the density
257: field for the same power spectra was done in an eulerian mesh of
258: $256^3$ grid points. After fourier transforming both constrained and
259: unconstrained density fields, we substitute the fourier modes of
260: the constrained field into the unconstrained one. Finally, we used
261: Zeldovich approximation to compute the 3D displacement field for an
262: initial redshift of z=60. Once we got the displacements for a refined
263: mesh of $2048^3$ grid points, we used it to estimate the initial
264: conditions for a resampling of $256^3$ dark matter particles in
265: total. In this way, we are able to zoom into a particular area of the
266: simulations and resimulate them with much higher resolution,
267: up to the maximum resolution possible $(2048^3)$,
268: having the same structures as in the low resolution simulations
269: (see \citet{kly01} for a more detailed information about zoomed
270: simulation techniques). Thus,
271: all the numerical experiments that are reported here have the
272: same number of particles ($256^3$). This translates into a mass per
273: particle of $1.3 \times 10^7 h^{-1} M_\odot$ for \LCDM and OCDM and $4.3
274: \times 10^7 h^{-1} M_\odot $ for SCDM simulations.
275:
276: We have used the parallel TREEPM N-body code GADGET2 (\cite{gadget2})
277: to run these simulations. An uniform mesh of $512^3$ grid points was
278: used to compute the long-range gravitational force by means of the
279: Particle-Mesh algorithm. A constant comoving Plummer equivalent
280: gravitational smoothing scale of 20 $h^{-1}$ kpc was set at high
281: redshift and we changed it to 5 $h^{-1}$ kpc physical scale
282: since z=3 till z=0. The number of timesteps to complete the evolution
283: from z=60 till z=0 ranges from 5000 to 7000 depending on the
284: simulation. We employed a variety of parallel computer architectures
285: (SGI-ALTIX, IBM-SP4, Opteron-clusters) during
286: the course of this work. Using 16 processors simultaneously, we
287: completed one run in about 2 cpu days.
288:
289: In what follows, we will use the name \LCDM, OCDM and SCDM
290: for the simulations with Constrained Initial conditions in
291: the different cosmological models. The names {\LCDM}u and OCDMu
292: will refer to the two different unconstrained realizations
293: in the \LCDM\ and OCDM models respectively.
294: As an example of how the simulations look like, we show
295: in Figure \ref{boxL} a projection of the dark matter distribution in
296: the \LCDM\ simulation box at z=0.
297:
298:
299:
300:
301:
302: \section{Selection of LG Candidates}
303: \label{sec:obj}
304:
305:
306: \begin{table}
307: \begin{tabular}{|ll|c|}
308: \hline \hline
309: \multirow{4}{*}{Components} & Kind & MW + M31 \\
310: & Mass & $125 \leq V_c \leq 270$ km/s \\
311: & Separation & $s \leq 1$ Mpc/h \\
312: & Relative velocity & $V_r < 0$ \\
313: \hline
314: \multirow{2}{*}{$\nexists$ neighbours} & Distance to LG & $d_{neigh} < 3$ Mpc/h \\
315: & Mass & $V_c \geq V_{c, comp}$ \\
316: \hline
317: \multirow{2}{*}{Virgo halos} & Distance to LG & $5 \leq d_{Virgo} \leq 12$ Mpc/h \\
318: & Mass & $500 \leq V_c \leq 1500$ km/s\\
319: \hline \hline
320: \end{tabular}
321: \caption{Constrains used to find LG candidates
322: following the Macci\`{o} \etal (2005) criterion.
323: The circular velocity has been used for the mass contrains.}
324: \label{LGc}
325: \end{table}
326:
327:
328: Dark matter halos were found in simulations using two object finding
329: methods: The Bound Density Maxima (BDM) algorithm (\cite{bdm}) is
330: based on finding local center of mass in spheres of variable radius
331: starting from randomly selected particles in the simulation.
332: The Amiga Halo Finder (\cite{amiga}), on the contrary, finds local
333: density maxima from an adaptive mesh hierachy.
334: In both cases, an iterative procedure to find local centre of mass from
335: density maxima is used. Particles that are not gravitatonally bound to
336: the halo potential are also removed. We took halos composed by more than
337: 100 dark matter particles, which translates into a minimum mass per halo of
338: $M_{min}= 1.3 \times 10^9 h^{-1} M_\odot $ for the \LCDM\ and OCDM simulations
339: and $M_{min}= 4.3 \times 10^9 h^{-1} M_\odot $ for the SCDM simulation.
340: We indentified the same objects with both methods. For the work
341: reported here we have used the halo catalogues obtained by the public
342: available AMIGA halo finder code
343: (http://www.aip.de/People/AKnebe/AMIGA/).
344:
345: To identify LG candidates from the halo distribution,
346: we selected those objects that fulfill the strict requirements
347: as given in \cite{gov97} and
348: \cite{mac05}. They are summarized in Table \ref{LGc}. In brief,
349: we searched for two halos similar to Milky Way and M31 galaxies,
350: without neighbours with masses as high as any of the LG members
351: and with a Virgo-like halo at an appropiate distance.
352: A few tens of LG-like objects have been
353: identified in each simulation ($23$ objects in \LCDM, 34 in $\Lambda$CDMu, $41$
354: in OCDM, $58$ in OCDMu and 37 in SCDM). One of the LG-like object
355: for the \LCDM simulation which closely resembles the actual LG in
356: terms of its mass and position, is presented in Figure \ref{spec_c}.
357:
358:
359:
360:
361: \begin{figure}
362: \resizebox{\linewidth}{!}{\includegraphics{F2.ps}}
363: \caption{Projected dark matter distribution around the best LG candidate
364: in the \LCDM\ simulation in supergalactic coordinates. The outer
365: circle delimits the Local Volume and the inner circle represents
366: the LG position }
367: \label{spec_c}
368: \end{figure}
369:
370: \begin{figure}
371: \resizebox{\linewidth}{!}{\includegraphics{F3a.ps}\includegraphics{F3b.ps}}
372: \caption{Plots of $v$ \vs $g_G$ and $v$ \vs $g_L$ for the candidate of Figure \ref{spec_c}.
373: }
374: \label{spec}
375: \end{figure}
376:
377:
378:
379:
380:
381: \section{Analysis}
382: \label{sec:analysis}
383:
384: In order to study the dynamics of the LG-like objects found in the
385: simulations, we have first computed the local Hubble flow
386: in spheres of $7\hmpc$ around each candidate,
387: and have also estimated the local overdensity within these spheres
388: from the total mass inside them. All DM halos
389: within the LV around each LG-like objects have been identified and the
390: gravitational field acting on each halo has been calculated in two
391: different ways. First, the local gravitational field field is
392: calculated like in W05 by the summation over the pairwise newtonian
393: interaction. Namely, the field acting on the i-th halo is given by \be
394: \tilde{\bg}{^{tot}_{l,i}} = -G \sum_{j\neq i} M_j \frac{ {\bf r}_j- {\bf r}_i
395: }{[( {\bf r}_j-{\bf r}_i )^2+A^2]^{3/2}} {\bf \cdot \hat{r}}_i,
396: \label{eqloc}
397: \ee where, following W05, a softening parameter of $A=1.2\hmpc$ is
398: introduced. Then we calculate also the 'true' gravitational field,
399: namely the field calculated by the N-body code of the full mass
400: distribution in the computational box. This is defined as the global
401: gravitational field, (see the appendix for further information).
402:
403:
404: As in W05, the local gravitational field is decomposed into two
405: terms, the contribution of a smooth background of matter and a
406: fluctuating part given by the point mass distributions. Since we are
407: interested in deviations from the average, we have to substract the
408: linear term contributed by the background, $\bg_{bg}$, from the local
409: gravitation field
410: \be
411: \tilde{\bg}_l = \tilde{\bg}{^{tot}_l} - \tilde{\bg}_{bg}.
412: \ee The
413: calculation of the background term is done in two ways. First,
414: following W05, the background solution is fitted by a linear term in
415: ${\bf r}$. Alternatively, the background solution is calculated by the
416: exact solution for the unperturbed universe:
417: \be
418: \tilde{\bg}_{bg} = -G \frac{4 \pi}{3} \Omega _m \rho_c {\bf r}
419: \ee
420: The two methods give virtually identical
421: results and the fitting method has been used here, so as to be
422: consistent with analysis of W05.
423: We have also substracted the anisotropic background estimated
424: from the tidal field (see \S{} \ref{sec:a-v}).
425:
426:
427: As we have mentioned earlier, the purpose of the present study
428: is to compare gravitational accelerations and velocities.
429: To facilitate such a comparison the gravitational field is
430: scaled by the linear theory prediction,
431: \be
432: \bg_x=\frac{2f(\Omega_m,\Omega_\Lambda)}{3H_0\Omega_m}\ \tilde{\bg}_x
433: \ee
434: where $x$ stands
435: here for the local or global field. The velocity-gravity scaling
436: factor is given by
437: \be f(\Omega_m,\Omega_\Lambda) \approx
438: \Omega_m^{0.6}+\frac{1}{70}\Omega_\Lambda\left(1+\frac{1}{2}\Omega_m \right)
439: \ee
440: (e.g. \cite{pee80}, \cite{lah91}). Throughout the paper the gravitational
441: fields will be represented by their scaled versions.
442:
443:
444: Our analysis is exemplified by Figure \ref{spec_c} which shows
445: one of the LG-like objects in the \LCDM\ constrained simulation.
446: The figure shows the matter distribution in a box of
447: $20 \hmpc$ centered on the simulated LG
448: and projected on the Supergalactic plane. In Figure \ref{spec}, two
449: scatter plots show the scatter of the local and global gravitational
450: accelerations,
451: \vs the peculiar velocity of all DM halos in this
452: particular LV. In what follows, all quantities shown correspond to
453: projections along the line of sight with the observer located in the
454: center of mass of the LG candidates.
455:
456: \begin{figure*}
457: \resizebox{\textwidth}{!}{ \includegraphics{F4a.ps}\includegraphics{F4b.ps}\includegraphics{F4c.ps}}
458: \caption {
459: Local \vs global scaled acceleration for one of the LG
460: candidates in the constrained \LCDM, OCDM and SCDM simulations respectively.
461: }
462: \label{loc_g}
463: \end{figure*}
464:
465:
466: \begin{figure*}
467: \resizebox{\textwidth}{!}{\includegraphics{F5.ps}}
468: \caption {Slopes (first row) and Pearson's correlation coefficient
469: (second row) of the linear fit of $g_L$ \vs $g_G$ as a function of
470: overdensity calculated within the simulated Local Volumes for all
471: candidates found in each simulation. The median and the average
472: slope and correlation coefficient with their one sigma errors
473: for all LGs in each simulation are also shown.
474: }
475: \label{loc_g_cor}
476: \end{figure*}
477:
478:
479:
480:
481: \section{Results}
482: \label{sec:result}
483:
484: \subsection{Relation between local and global accelerations}
485: \label{sec:loc-glob}
486:
487: Figure \ref{loc_g} shows scatter plots
488: of local \vs global gravitational fields for some of the Local
489: Volume candidates extracted from simulations. In order to study the
490: relation between both fields, a linear fit to the $g_L$ \vs $g_G$ distribution has been made
491: for all the LG candidates found in the different simulations.
492:
493:
494:
495: Figure \ref{loc_g_cor} shows the least square fit slopes and
496: correlation coefficients of the local to the global gravitational fields
497: as a function of the overdensidity $\delta \rho / \rho$ measured within each
498: of the simulated LV's. We find that the local and global field are
499: uncorrelated, showing an extremely small correlation coefficient and a
500: mean and median slope of roughly $0.1$ rather than unity.
501:
502:
503:
504: We have tested the effect of taking only haloes with high mass
505: (more than $10^{11}\hmsun$).
506: The results are very similar to the previous ones, because the
507: lightest haloes do not contribute much to the gravitational field.
508: The overdensity was calculated by computing all matter inside the local
509: volumes. When using only mass in halos to estimate the overdensity,
510: results do not change at all.
511:
512:
513:
514: In the preceding analysis the local gravitational field is calculated
515: by summing over the pairs of DM halos, assuming they carry all the mass
516: in the LV
517: and that their mass is spread out over a relatively large area
518: described by the $A$ smoothing parameter.
519: The poor correlation between
520: the local and global gravitational fields has led us to relax
521: the assumption that the mass is traced by the DM halos and calculate
522: the local field contributed by all dark matter particles in the LV.
523: This is presented in Figure \ref{all_partic},
524: which shows a very tight correlation between $\bg_L$ and $\bg_G$ for
525: one particular LG-like object in the \LCDM\ CS.
526: A better correlation with a slope very close to unity are
527: obtained when we include all the particles that
528: constitute the inter-halo medium, rather than by the intra-halo
529: particles only, as can be seen in Figure \ref{fig7}.
530: Moreover, the only gravitational smoothing done in this case
531: correspond to that included in the simulation, which is of the order of
532: kiloparsec scale, contrary to the strong smoothing of 1.2 Mpc used in
533: W05 analysis and in our previous estimates.
534: The horizontal branch around the origin observed in Figure \ref{all_partic} is
535: mainly due to particles not bound to halos.
536: These particles are more affected by the external field than particles bounds to halos, and therefore for these the local acceleration is much smaller that the global one.
537:
538:
539: Obviously, the correlation between the local and global acceleration improves as the LV increases.
540: This has been checked for the LG-like object of Figure \ref{all_partic}. Assuming a LV of
541: a 30 $\hmpc$ radius the correlation improves to $S=0.47$ and $r^2=0.58$. However this is still
542: considerably worse than the $S=0.89$ and $r^2=0.91$ obtained by considering all the particles
543: within the 7 $\hmpc$ LV (Figure \ref{all_partic}).
544:
545:
546: The lack of correlation between the local and global gravitational
547: accelerations, strongly leads to expect the absence of correlation
548: between the local gravitational field and the peculiar
549: velocities. This is clearly confirmed by the analysis presented in
550: \S{} \ref{sec:a-v}.
551:
552:
553:
554:
555: \begin{figure}
556:
557: \resizebox{\linewidth}{!}{\includegraphics{F6.ps}}
558: \caption { Local \vs\ global gravitational accelerations computed
559: taking into account all the particles within the LV
560: for the same candidate in the \LCDM\ simulation as in Figure
561: \ref{loc_g}. The small points represent the individual dark matter
562: particles. Thick solid points correspond to halos in which
563: accelerations have been computed by the average of all particles
564: inside them. The straight line shows the equivalence between both accelerations.}
565: \label{all_partic}
566: \end{figure}
567:
568:
569:
570:
571: \begin{figure}
572: \resizebox{\linewidth}{!}{\includegraphics{F7.ps}}
573: \caption{Slopes (first row) and Pearson's correlation
574: coefficients (second row) for the $g_L$ \vs $g_G$ least square
575: fits for all the LG
576: candidates found in the \LCDM\ and OCDM simulations as a function of
577: overdensity. All particles within simulated LVs have been taken
578: in the computation of the local accelerations.
579: A gravitational smoothing parameter of 5 $h^{-1}$ kpc is
580: assumed to compute the newtonian pairwise forces between particles.
581: }
582: \label{fig7}
583: \end{figure}
584:
585:
586:
587:
588:
589:
590:
591: \subsection{Accelerations and velocities}
592: \label{sec:a-v}
593:
594:
595:
596:
597:
598: Figures \ref{v_g} and \ref{v_loc} present the peculiar velocities as a
599: function of the scaled global and local gravitational accelerations
600: around some selected LG-like objects in the different simulations. The
601: slopes and the correlation coefficients of the linear fits for the
602: relation between the peculiar velocities and the gravitational
603: accelerations acting on halos in the LV around all LG-like objects are
604: given in Figures \ref{v_g_cor} and \ref{v_loc_cor}. The least square fit
605: analysis confirms the visual impression of Figures \ref{v_g} and
606: \ref{v_loc}. The peculiar velocity of halos in the LV of LG-like
607: objects is clearly correlated with the global acceleration but with a
608: non-negligible scatter, with a mean and median correlation
609: coefficient around $0.5$ for the constrained simulations
610: and somewhat lower for the
611: unconstrained ones (Figure \ref{v_g_cor}).
612: The mean and the median slope of the
613: linear relation between the velocity and scaled global acceleration is
614: in the range of $0.4$ to $0.6$. The distribution of the fitted slopes
615: shows some dependence on the mean overdensity ($\delta \rho /\rho$) in the LV. The
616: width of the distribution decreases with the overdensity. At low $\delta \rho
617: /\rho$'s the slope ranges from roughly $0.2$ to almost unity but at $\delta \rho
618: /\rho \gtrsim 1$ the slope shows a narrow scatter around its mean value of $\approx
619: 0.5$. Upon scaling of the gravitational field the linear theory
620: predicts the slop to be unity. It follows that the amplitude of the
621: peculiar velocities is smaller than what is expected by the linear
622: theory.
623:
624:
625:
626:
627: The peculiar velocities do not show any correlation with the local
628: accelerations (Figures \ref{v_loc} and \ref{v_loc_cor}). Their linear
629: fit yields extremely low correlation coefficients and the fitted slope
630: has no meaning. One should recall here that this lack of correlation
631: has already been anticipated from the lack of correlation we found
632: between the local and global accelerations (\S\ \ref{sec:loc-glob}).
633:
634: \begin{figure*}
635: \resizebox{\textwidth}{!}{ \includegraphics{F8a.ps}\includegraphics{F8b.ps}\includegraphics{F8c.ps}}
636: \caption {
637: Peculiar velocity \vs global scaled acceleration of halos inside the
638: LV for the same candidates as in Figure \ref{loc_g}.
639: }
640: \label{v_g}
641: \end{figure*}
642:
643:
644:
645:
646:
647: The calculation of the local gravitational acceleration neglects the
648: contribution of the tidal field, which is induced by the inhomogenous
649: matter distribution outside of the LV. Obviously, for objects like the
650: LG, with the Virgo cluster located just outside of the LV, the tidal
651: field cannot be neglected. The local gravity-velocity
652: correlation should be improved by adding the tidal field into the
653: fitting procedure. Indeed, we follow W05 and extend the fitting
654: procedure to
655: \begin{equation}
656: \sigma^2=\frac{1}{N}\sum (g_{L}-v+{\bf v}_0 \cdot {\bf \hat{r}}+{\bf \hat{r}\cdot H \cdot r})^2,
657: \label{sig2}
658: \end{equation}
659: where ${\bf v}_0$ is an unknown vector and ${\bf H}$ is a symmetrical
660: tensor with six unknown quatities. The nine free parameters are found
661: by minimizing the scatter. Note that in the limit of the linear theory
662: and for a small LV ${\bf v}_0$ should be set to zero and the symmetric
663: tensor $H$ should be traceless. As the above assumptions do not hold
664: for the LG we allow for a finite ${\bf v}_0$ and for $H$ to have a
665: trace.
666:
667: Figure \ref{v_loc_cor_m} shows that indeed adding the
668: ${\bf v}_0$ and ${\bf H}$ terms improves somewhat the correlation
669: between the local gravity and peculiar velocities. Yet, they are weakly
670: correlated with $r^2 \approx (0.1 \ - \ 0.2)$ for most LG-like objects. The
671: minor improvement is not surprising. The Virgo cluster is located at a
672: distance of about $10 \hmpc$ and the LV is defined by a sphere of
673: radius of $7\hmpc$. Hence, modeling the tidal field by a spatial linear
674: expansion, as is implicitly assumed in Eq. \ref{sig2} constitutes a
675: poor fit of the tidal field.
676:
677:
678:
679: \begin{figure*}
680: \resizebox{\textwidth}{!}{\includegraphics{F9a.ps}\includegraphics{F9b.ps}\includegraphics{F9c.ps}}
681: \caption {
682: Peculiar velocity \vs local scaled acceleration for halos within the LV
683: for the same candidates as in Figure \ref{loc_g}.
684: }
685: \label{v_loc}
686: \end{figure*}
687:
688:
689:
690: \begin{figure*}
691: \resizebox{\textwidth}{!}{\includegraphics{F10.ps}}
692: \caption {Same as Figure \ref{loc_g_cor} but for the peculiar velocity
693: \vs global scaled aceleration fits.
694: }
695: \label{v_g_cor}
696: \end{figure*}
697:
698:
699: \begin{figure*}
700: \resizebox{\textwidth}{!}{\includegraphics{F11.ps}}
701: \caption {
702: The same as Figure \ref{loc_g_cor} but for the peculiar velocity \vs local scaled acceleration fit.
703: }
704: \label{v_loc_cor}
705: \end{figure*}
706:
707:
708:
709:
710: \section{Discussion}
711: \label{sec:disc}
712:
713: We have clearly demonstrated in this paper
714: that the W05's analysis is not
715: expected
716: to yield a simple
717: and clear correlation between the gravitational field
718: calculated from the mass distribution within the local volume and the
719: peculiar velocities of halos within that volume. This result
720: invalidates W05's basic assumption that the peculiar velocity field
721: traces the gravitational field in a simple fashion, and therefore the
722: statements concerning the role of the gravitational instability in the
723: LV are
724: not valid.
725:
726:
727:
728: Setting aside issues concerning the practical limitations posed by
729: observations of the LV and the uncertainty in estimating the dynamical
730: mass of luminous galaxies there are three main theoretical reasons why
731: W05's analysis fails. The most obvious one is that the gravitational
732: field is assumed to be traced by the galaxies, treated as point-like
733: particles. Figs. \ref{loc_g} and \ref{loc_g_cor} show the poor
734: correlation between the actual gravitational field and the local field
735: induced by the galaxies. Now, this poor correlation is mostly due to
736: the sampling of the field by the galaxies, as manifested by Fig.
737: \ref{all_partic} which shows a clear correlation between the global
738: field and the local field that is induced by all dark matter particles
739: in the LV.
740:
741: The other reasons for the breakdown of the simple gravity - velocity
742: relation of the linear theory are both related to the tidal field,
743: hence the sheer of the velocity field. The simpler reason is that in
744: solving the Poisson equation one should not neglect the homogenous
745: solution, namely the tidal field. Now, in principle this can be easily
746: corrected by adding a (spatial) linear term to the gravitational field
747: that scales with the traceless shear tensor (Eq. \ref{sig2} and Figure
748: \ref{v_loc_cor_m}). Thus by adding six free
749: parameters to the fitting procedure one might be able to account for
750: the tidal field. However, the size of the LV is such that the spatial
751: linear expansion of the tidal field would fail and lead to an incorrect
752: estimation of the gravitational field. For the LV centered on the LG,
753: the tidal effect of the LSC cannot be represented by a linear term.
754:
755: The other reason for the inadequacy of the linear theory is more
756: subtle. It has been shown that in the quasi-linear regime the growth of
757: the density contrast depends on the magnitude of the shear tensor
758: \citep{hof86,hof89,zar93,van94,ber94}.
759: The shear dependence introduces a non-local term in the equations that
760: govern the growth of structure in the quasi-linear regime. Indeed,
761: Fig. \ref{v_g} shows a tight linear relation between the peculiar
762: velocities and the scaled global gravitation field. Yet, the constant
763: of proportionality is less than unity for all the LG-like objects in
764: all the simulation (except of one single object) as is predicted by the
765: linear theory prediction (in agreement with \citet{hof89}). This
766: behavior implies that under the optimal conditions of a full knowledge
767: of the gravitational field a linear relation between the gravity and
768: velocity field is expected within the LV around LG-like objects.
769: However, using the slope of the relation as a way of measuring $\Omega_m$
770: would underestimate its true value.
771:
772: W05 attempted to find a simple linear relation between the peculiar
773: velocities and local gravitational field of galaxies in the LV. A
774: careful analysis of the dynamics within the LV around LG-like objects
775: identified in constrained simulations of the local universe and in
776: unconstrained simulations in flat-\LCDM, OCDM and flat-matter only CDM
777: cosmologies shows that a lack of correlation is to be expected. Hence
778: we cannot support the claim that {\em 'either dark matter is not distributed
779: in the same way as luminous matter in this region, or peculiar
780: velocities are not due to fluctuations in mass.'}
781:
782:
783: \begin{figure*}
784: \resizebox{\textwidth}{!}{\includegraphics{F12.ps}}
785: \caption {
786: Same as Figure \ref{v_loc_cor}, but including the linear tidal field
787: into the fitting procedure (see Eq. \ref {sig2})}
788: \label{v_loc_cor_m}
789: \end{figure*}
790:
791:
792:
793:
794: \section{Acknowledgements}
795: We appreciate very much the comments and discussions with Yago
796: Ascasibar, Anatoly Klypin, Andrey Kravtsov, Stefan Gottl\"ober and
797: Yaniv Dover. We would like to thank Astrophysikalisches Institut Potsdam for
798: allowing us to use the Sansoucci supercomputer opteron clusters and to
799: host us several times during the course of this work.
800: We also thank CIEMAT (Spain)
801: to allow us to use their SGI-ALTIX supercomputer and to NIC J\"ulich
802: (Germany) for the access to the IBM-Regatta p690+ JUMP supercomputer.
803: GY would like to thank also MCyT for financial support under
804: project numbers AYA2003-07468 and BFM2003-01266.
805: LAMV acknowledges financial support from MCyT (Spain) under project
806: BFM2003-01266 and from Comunidad de Madrid through a PhD fellowship.
807: YH acknowledges the support of ISF-143/02, the Sheinborn
808: Foundation and the DFG for a Mercator Gastprofessur at
809: Potsdam University.
810:
811:
812:
813:
814:
815:
816: \section{Appendix: The GADGET estimation of gravitational accelerations}
817:
818: Throughout the paper the global and local gravitational accelerations
819: have been compared.
820: The local accelerations within the LV were
821: calculated by summing the Newtonian pairwise field.
822: This can be repeated for the global field by summing over all the
823: particles of the simulation. This direct sum is computationally very
824: costly and does not take into account the contribution of the infinite
825: periodic boxes. We used instead the particle accelerations calculated
826: by the TREE-PM algorithm by the GADGET code. The relation between
827: the GADGET acceleration with $\bg_G$ is as follows:
828:
829: The physical $\textbf{r}$ and comoving $\textbf{x}$ coordinates are related by:
830: $\textbf{r} = a\textbf{x}$. The global gravitational field equals
831: the physical acceleration of an object
832: $\ddot{\textbf{r} }= \bg$.
833:
834: Now, the GADGET code provides an acceleration-like term defined as:
835: \begin{equation}
836: \bff_p=\frac{1}{a} \frac{d}{dt}\left( a\cdot \bv_p\right)=\dot{\bv}_p + H\cdot \bv_p,
837: \label{fpdef}
838: \end{equation}
839: where $\bv_p$ is the peculiar velocity and $a$ is the expansion scale factor.
840:
841: It follows that
842: \begin{equation}
843: \ddot \br = \bff_p + \br \frac{\ddot a}{a}.
844: \label{r_f_a}
845: \end{equation}
846: Recalling that $\frac{\ddot a}{a}=-\frac{4\pi G\rho_c\Omega_0}{3}$ one gets at the end
847: \begin{equation}
848: \bg_G = \bff_p - \frac{1}{2}H^2\Omega_0 \cdot \br.
849: \end{equation}
850: This is the value we used for the global acceleration of each dark
851: matter particle. The total acceleration of halos was computed by
852: averaging this quantity over the entire number of particles belonging
853: to each halo.
854:
855:
856:
857:
858: \begin{thebibliography}{}
859:
860: \bibitem[\protect\citeauthoryear{{Bertschinger} \& {Jain}}{{Bertschinger} \&
861: {Jain}}{1994}]{ber94}
862: {Bertschinger} E., {Jain} B., 1994, \apj, 431, 486
863:
864: \bibitem[\protect\citeauthoryear{{Dolag}, {Hansen}, {Roncarelli} \&
865: {Moscardini}}{{Dolag} et~al.}{2005}]{dol05}
866: {Dolag} K., {Hansen} F.~K., {Roncarelli} M., {Moscardini} L., 2005,
867: \mnras, 363, 29
868:
869: \bibitem[\protect\citeauthoryear{{Gill}, {Knebe} \& {Gibson}}{{Gill}
870: et~al.}{2004}]{amiga}
871: {Gill} S.~P.~D., {Knebe} A., {Gibson} B.~K., 2004, \mnras, 351, 399
872:
873: \bibitem[\protect\citeauthoryear{{Governato}, {Moore}, {Cen}, {Stadel}, {Lake}
874: \& {Quinn}}{{Governato} et~al.}{1997}]{gov97}
875: {Governato} F., {Moore} B., {Cen} R., {Stadel} J., {Lake} G., {Quinn}
876: T., 1997, New Astronomy, 2, 91
877:
878: \bibitem[\protect\citeauthoryear{{Hoffman}}{{Hoffman}}{1986}]{hof86}
879: {Hoffman} Y., 1986, \apj, 308, 493
880:
881: \bibitem[\protect\citeauthoryear{{Hoffman}}{{Hoffman}}{1989}]{hof89}
882: {Hoffman} Y., 1989, \apj, 340, 69
883:
884: \bibitem[\protect\citeauthoryear{{Hoffman}, {Romano-D{\'{\i}}az},
885: {Faltenbacher}, {Jones}, {Heller} \& {Shlosman}}{{Hoffman}
886: et~al.}{2006}]{hof06}
887: {Hoffman} Y., {Romano-D{\'{\i}}az} E., {Faltenbacher} A., {Jones} D.,
888: {Heller} C., {Shlosman} I., 2006, in {Mamon} G.~A., {Combes} F.,
889: {Deffayet} C., {Fort} B., eds, EAS Publications Series {Constrained
890: Simulations of Dark Matter Halos}.
891: pp 15--18
892:
893: \bibitem[\protect\citeauthoryear{{Karachentsev}}{{Karachentsev}}{2005}]{kar05}
894: {Karachentsev} I.~D., 2005, \aj, 129, 178
895:
896: \bibitem[\protect\citeauthoryear{{Klypin}, {Gottl{\"o}ber}, {Kravtsov} \&
897: {Khokhlov}}{{Klypin} et~al.}{1999}]{bdm}
898: {Klypin} A., {Gottl{\"o}ber} S., {Kravtsov} A.~V., {Khokhlov} A.~M.,
899: 1999, \apj, 516, 530
900:
901: \bibitem[\protect\citeauthoryear{{Klypin},{Kravtsov}, {Bullock} \&
902: {Primack}}{{ Klypin} et~al.}{2001}]{kly01}
903: {Klypin}, A., {Kravtsov}, A.~V., {Bullock}, J.~S., {Primack}, J.~R., 2001,
904: \apj, 554, 903
905:
906:
907: \bibitem[\protect\citeauthoryear{{Klypin}, {Hoffman}, {Kravtsov} \&
908: {Gottl{\"o}ber}}{{Klypin} et~al.}{2003}]{kly03}
909: {Klypin} A., {Hoffman} Y., {Kravtsov} A.~V., {Gottl{\"o}ber} S., 2003,
910: \apj, 596, 19
911:
912: \bibitem[\protect\citeauthoryear{{Klypin}, {Rhee}, {Valenzuela}, {Holtzman} \&
913: {Moorthy}}{{Klypin} et~al.}{2004}]{kly04}
914: {Klypin} A., {Rhee} G., {Valenzuela} O., {Holtzman} J., {Moorthy} B.,
915: 2004, in {Prada} F., {Martinez Delgado} D., {Mahoney} T.~J., eds, ASP
916: Conf. Ser. 327: Satellites and Tidal Streams {The Rotation Curves of Dwarf
917: Galaxies: a Problem for Cold Dark Matter?}.
918: pp~3--+
919:
920: \bibitem[\protect\citeauthoryear{{Kravtsov}, {Klypin} \& {Hoffman}}{{Kravtsov}
921: et~al.}{2002}]{kra02}
922: {Kravtsov} A.~V., {Klypin} A., {Hoffman} Y., 2002, \apj, 571, 563
923:
924: \bibitem[\protect\citeauthoryear{{Lahav}, {Lilje}, {Primack} \& {Rees}}{{Lahav}
925: et~al.}{1991}]{lah91}
926: {Lahav} O., {Lilje} P.~B., {Primack} J.~R., {Rees} M.~J., 1991, \mnras,
927: 251, 128
928:
929: \bibitem[\protect\citeauthoryear{{Macci{\`o}}, {Governato} \&
930: {Horellou}}{{Macci{\`o}} et~al.}{2005}]{mac05}
931: {Macci{\`o}} A.~V., {Governato} F., {Horellou} C., 2005, \mnras, 359, 941
932:
933: \bibitem[\protect\citeauthoryear{{Mathis}, {Lemson}, {Springel}, {Kauffmann},
934: {White}, {Eldar} \& {Dekel}}{{Mathis} et~al.}{2002}]{mat02}
935: {Mathis} H., {Lemson} G., {Springel} V., {Kauffmann} G., {White} S.~D.~M.,
936: {Eldar} A., {Dekel} A., 2002, \mnras, 333, 739
937:
938: \bibitem[\protect\citeauthoryear{{Peebles}}{{Peebles}}{1980}]{pee80}
939: {Peebles} P.~J.~E., 1980, {The large-scale structure of the universe}.
940: ~Princeton, N.J., Princeton University Press, 1980.~435 p.
941:
942: \bibitem[\protect\citeauthoryear{{Reiprich} \& {B{\"o}hringer}}{{Reiprich} \&
943: {B{\"o}hringer}}{2002}]{rei02}
944: {Reiprich} T.~H., {B{\"o}hringer} H., 2002, \apj, 567, 716
945:
946: \bibitem[\protect\citeauthoryear{{Springel}}{{Springel}}{2005}]{gadget2}
947: {Springel} V., 2005, \mnras, 364, 1105
948:
949: \bibitem[\protect\citeauthoryear{{Tonry}, {Dressler}, {Blakeslee}, {Ajhar},
950: {Fletcher}, {Luppino}, {Metzger} \& {Moore}}{{Tonry} et~al.}{2001}]{sbf}
951: {Tonry} J.~L., {Dressler} A., {Blakeslee} J.~P., {Ajhar} E.~A., {Fletcher}
952: A.~B., {Luppino} G.~A., {Metzger} M.~R., {Moore} C.~B., 2001, \apj,
953: 546, 681
954:
955: \bibitem[\protect\citeauthoryear{{van de Weygaert} \& {Babul}}{{van de
956: Weygaert} \& {Babul}}{1994}]{van94}
957: {van de Weygaert} R., {Babul} A., 1994, \apjl, 425, L59
958:
959: \bibitem[\protect\citeauthoryear{{Whiting}}{{Whiting}}{2005}]{whi05}
960: {Whiting} A.~B., 2005, \apj, 622, 217
961:
962: \bibitem[\protect\citeauthoryear{{Willick}, {Courteau}, {Faber}, {Burstein},
963: {Dekel} \& {Strauss}}{{Willick} et~al.}{1997}]{mark}
964: {Willick} J.~A., {Courteau} S., {Faber} S.~M., {Burstein} D., {Dekel} A.,
965: {Strauss} M.~A., 1997, \apj, 109, 333
966:
967: \bibitem[\protect\citeauthoryear{{Zaroubi} \& {Hoffman}}{{Zaroubi} \&
968: {Hoffman}}{1993}]{zar93}
969: {Zaroubi} S., {Hoffman} Y., 1993, \apj, 414, 20
970:
971: \bibitem[\protect\citeauthoryear{{Zaroubi}, {Hoffman} \& {Dekel}}{{Zaroubi}
972: et~al.}{1999}]{zar99}
973: {Zaroubi} S., {Hoffman} Y., {Dekel} A., 1999, \apj, 520, 413
974:
975: \end{thebibliography}
976:
977:
978: \end{document}
979:
980:
981: