1: \documentclass[12pt,preprint]{aastex}
2: %\documentstyle[12pt,aasms4]{article}
3: %\documentstyle[11pt,aaspp4]{article}
4: %\documentstyle[aas2pp4]{article}
5: %\documentclass[12pt,aasms4]{aastex}
6: %\received{}
7: %\accepted{}
8: %\journalid{000}{}
9: %\articleid{}{}
10:
11: \begin{document}
12:
13: \title{Estimating Power Spectrum of Sunyaev-Zeldovich Effect
14: from the Cross-Correlation between WMAP and 2MASS}
15:
16: \author{Liang Cao, Jiren Liu and Li-Zhi Fang}
17:
18: \affil{Department of Physics, University of Arizona,
19: Tucson, AZ 85721}
20:
21: \begin{abstract}
22: We estimate the power spectrum of
23: SZ(Sunyaev-Zel'dovich)-effect-induced temperature fluctuations on
24: sub-degree scales by using the cross correlation between the
25: three-year WMAP maps and 2MASS galaxy distribution. We produced
26: the SZ effect maps by hydrodynamic simulation samples of the
27: $\Lambda$CDM model, and show that the SZ effect temperature
28: fluctuations are highly non-Gaussian. The PDF of the temperature
29: fluctuations has a long tail. More than $70\%$ power of the SZ
30: effect temperature fluctuations attributes to top $\sim 1\%$
31: wavelet modes (long tail events). On the other hand, the CMB
32: temperature fluctuations basically are Gaussian. Although the mean
33: power of CMB temperature fluctuations on sub-degree scales is much
34: higher than that of SZ effect map, the SZ effect temperature
35: fluctuations associated with top 2MASS clusters is comparable to
36: the power of CMB temperature fluctuations on the same scales.
37: Thus, from noisy WMAP maps, one can have a proper estimation of
38: the SZ effect power at the positions of the top 2MASS clusters.
39: The power spectrum given by these top wavelet modes is useful to
40: constrain the parameter of density fluctuations amplitude
41: $\sigma_8$. We find that the power spectrum of these top wavelet
42: modes of SZ effect on sub-degree scales basically is consistent
43: with the simulation maps produced with $\sigma_8=0.84$. The
44: simulation samples of $\sigma_8=0.74$ show, however, significant
45: deviation from detected SZ power spectrum. It can be ruled out
46: with confidence level 99\% if all other cosmological parameters
47: are the same as that given by the three-year WMAP results.
48:
49: \end{abstract}
50:
51: \keywords{cosmology: theory - large-scale structure of the
52: universe}
53:
54: \section{Introduction}
55:
56: The thermal Sunyaev-Zel'dovich (SZ) effect is due to the inverse
57: Compton scattering of the cosmic microwave background (CMB)
58: photons by hot electrons. This effect is proportional to the
59: pressure of the electron gases, $p =k_B nT$, where $n$ and $T$
60: being the number density and temperature of electrons. Therefore,
61: SZ effect can be used as a measurement of the hot baryon gases of
62: the universe (e.g. Birkinshaw 1999; Carlstrom et al. 2002). Since
63: clusters are the hosts of high pressure gases, the angular power
64: spectrum of the SZ effect is very sensitive to the amplitude of
65: mass density fluctuations parameter $\sigma_8$ (Refregier et al.
66: 2000; Bond et al. 2002). The amplitude of the SZ effect angular
67: power spectrum is found to be proportional to
68: $\sigma_8^{7}(\Omega_bh)^2$ and is almost independent of other
69: cosmological parameters (e.g. Seljak et al. 2001; Komatsu \&
70: Seljak 2002). Therefore, the amplitude of the SZ effect power
71: spectrum can be used as an effective constraint on the parameter
72: $\sigma_8$. The SZ effect power spectrum on scales of a few
73: arcminutes has been used for this issue (Mason et al.2003;
74: Readhead et al. 2004; Dawson et al. 2006).
75:
76: Recently released WMAP three-year data (WMAP III) refines most
77: results of cosmological parameters given by the WMAP first year
78: data. They found $\sigma_8=0.74\pm 0.05$ (Spergel et al. 2006),
79: which is significantly lower than $\sigma_8=0.84\pm 0.04$ of the
80: first year data. The new number of $\sigma_8$ is a challenge to the
81: cosmological parameter determined with samples of galaxies and
82: galaxy clusters, most of which yield $\sigma_8 \simeq 0.9-1$ if the
83: matter content of the universe $\Omega_m \leq 0.3$ (Reiprich \&
84: B{\"o}hringer 2002; Hoekstra et al. 2002; Refregier, et. al. 2002;
85: Van Waerbeke et. al. 2002; Bacon et al. 2003; Bahcall \& Bode 2003;
86: Seljak, et al. 2005; Viel \& Haehnelt 2006; McCarthy et al. 2006;
87: Cai et al. 2006). This problem motivates us to study the constraint
88: on $\sigma_8$ given by the power spectrum of the SZ effect.
89:
90: We try to extract the information of the SZ effect power spectrum on
91: sub-degree scales from the WMAP III data themselves. At the first
92: glance, it seems to be hopeless to extract such information from
93: noisy WMAP III data, because the power of SZ effect induced
94: temperature fluctuations on sub-degree scales is much less than that
95: of CMB temperature fluctuations (e.g. Cooray et al. 2004). A direct
96: analysis on the WMAP data also shows that the SZ effect contribution
97: to the CMB fluctuations on scale of the first acoustic peak should
98: be less than 2\% (Huffenberger et al. 2004; Spergel et al. 2006).
99:
100: This problem can be solved if considering that the SZ effect map is
101: highly non-Gaussian, while CMB is Gaussian. Although the mean power
102: of the thermal SZ effect on sub-degree scales is no more than $2\%$
103: of the power of CMB, the SZ temperature fluctuations at local
104: modes associated with clusters are comparable to the CMB temperature
105: fluctuations. Thus, the SZ temperature fluctuations power on sub-degree
106: scales would be detectable from noisy WMAP maps if one can identify such
107: local modes. To implement this detection, we developed a DWT(discrete
108: wavelet transform) algorithm for the power spectrum estimation.
109: The DWT analysis is powerful to pick up weak but non-Gaussian signals
110: from Gaussian background (Donoho 1995).
111:
112: We will use the cross-correlation between the WMAP III data and
113: 2MASS galaxies. Since the redshift of 2MASS XSC galaxies is up to
114: $z\simeq 0.1$, at which clusters are on angular scales of
115: sub-degree. Moreover, the thermal SZ effect on angular scales of
116: clusters has been significantly identified with the 2nd and 4th
117: order cross-correlations between the first year WMAP data and 2MASS
118: XSC galaxies(Afshordi et al. 2004, 2005; Myers et al. 2004; Cao et
119: al. 2006). The WMAP-2MASS cross-correlation would be effective to
120: extract the SZ power on sub-degree scales from the WMAP III maps.
121:
122: The paper is organized as follow. In \S 2, we analyze the
123: non-Gaussian features of the SZ effect maps generated with
124: cosmological hydrodynamic simulation. \S 3 describes the
125: identification of the SZ effect of 2MASS clusters from the
126: WMAP-2MASS cross correlation with the wavelet modes. \S 4 first
127: describes the algorithm of estimating the SZ power spectrum from the
128: WMAP III maps, and then presents the result of the SZ effect
129: constraint on $\sigma_8$. The discussion and conclusion are given in
130: \S 5.
131:
132: \section{The power spectrum of SZ effect fluctuations}
133:
134: \subsection{Maps of SZ effect}
135:
136: To develop the method of estimating power spectrum, we first study
137: the statistical properties of the SZ effect temperature
138: fluctuations. We produce the SZ effect maps with the WIGEON (Weno
139: for Intergalactic medium and Galaxy Evolution and formatiON) code,
140: which is a cosmological hydrodynamic $N$-body code based on the
141: Weighted Essentially Non-Oscillatory (WENO) algorithm (Harten et
142: al. 1986; Liu et al. 1994; Jiang \& Shu 1996; Shu 1998; Fedkiw,
143: Sapiro \& Shu 2003; Shu 2003). The details of this code can be
144: found at Feng et al. (2004). Some WIGEON samples have already been
145: used in cosmological studies (e.g. He et al. 2004, 2006; Zhang et
146: al. 2006; Liu et al. 2006).
147:
148: The samples are simulated in a 100 $h^{-1}$ Mpc cubic box with
149: 512$^3$ meshes for gas, and equal number of particles for dark
150: matter. We use the standard $\Lambda$CDM model, which is specified
151: by the matter density parameter $\Omega_{\rm m}=0.27$, baryon
152: density parameter $\Omega_{\rm b}=0.044$, cosmological constant
153: $\Omega_{\Lambda}=0.73$, Hubble constant $h=0.71$. In order to
154: test the effect of the amplitude of density fluctuations, we
155: produce two samples with $\sigma_8=0.74$ and 0.84, respectively.
156: The ratio of specific heats is $\gamma=5/3$. The transfer function
157: is calculated using CMBFAST (Seljak \& Zaldarriaga 1996). In order
158: to correct the underestimate of rare massive galaxy clusters due
159: to the finite box size, the initial conditions are generated by
160: sampling a convolved power spectrum following the method of Pen
161: (1997) and Sirko (2006), which accurately model the real-space
162: statistics, such as $\sigma_8$.
163:
164: The atomic processes including ionization, cooling and heating are
165: modeled as the method in Theuns et al. (1998). We take a
166: primordial composition of H and He ($X$=0.76, $Y$=0.24) and use an
167: ionizing background model of Haardt \& Madau (1996). Star
168: formation and their feedback due to supernova explosions and AGN
169: activities were not taken into account.
170:
171: Since the redshift of 2MASS galaxies is up to 0.1, we produce the
172: SZ maps by integrating the pressure of gas from $z=0$ to 0.1. The
173: simulation data is stored for every redshifts corresponding to
174: comoving distance $n\times$100$h^{-1}$Mpc, $n$ being integral.
175: Three simulation boxes are stacked together and integrated along
176: different angular direction to compose a SZ map. Two 2-D maps of
177: the Compton parameter $y$ for $\sigma_8=0.74$ and 0.84 are shown
178: in Figure 1. Both maps have a size of $19.1^{\circ}\times
179: 19.1^{\circ}$, with resolution of 512$\times$512. We produce 40
180: independent maps. The total area, $40\times (19.1^{\circ})^2$, is
181: comparable with the area of 2MASS data we used (\S 3.1).
182:
183: Figure 2 gives the angular power spectrum for two samples with
184: $\sigma_8=0.74$ and 0.84, respectively. We see that the amplitude
185: of the power spectrum is significantly dependent on $\sigma_8$.
186: Comparing Figure 2 with other simulations (e.g. Refregier et al.
187: 2000; Seljak et al. 2001; Komatsu \& Seljak 2002), our result
188: shows about the same power as others on scales equal to and larger
189: than $0^{\circ}.5$, while it is much lower than others on scales
190: $\ll 0^{\circ}.5$. This is expected, as the power on scales less
191: than $0^{\circ}.5$ mainly comes from clusters with redshift $>
192: 0.1$. If considering all the contribution from $z=6$, our
193: simulation samples show the same behavior as others, i.e.,
194: declining around $0^{\circ}.1$.
195:
196: \subsection{The DWT power spectrum of SZ effect map}
197:
198: As shown by Figure 1, the value of $y$ is very small in most
199: space, while high value of $y$ appears only in small and localized
200: areas. One can then take the advantage of DWT analysis to describe
201: the field $y({\bf x})$ (${\bf x}=(x_1,x_2)$) with DWT
202: decomposition. The DWT variables of a map $y(x_1,x_2)$ in the area
203: of $0\leq x_1\leq L_1$ and $0\leq x_2\leq L_2$ are given by the
204: decompositions of scaling functions $\phi_{\bf j,l}({\bf x})$ and
205: wavelet functions $\psi_{\bf j,l}({\bf x})$. We use Daubechies 4
206: wavelet (Daubechies (1992), Fang \& Thews (1998). The functions
207: $\phi_{\bf j,l}({\bf x})$ and $\psi_{\bf j,l}({\bf x})$ can also
208: be found in Yang et al. (2001).
209:
210: Briefly, The DWT scaling function, $\phi_{\bf j,l}({\bf x})$, is a
211: window function for cells (modes) in the area $L_1l_1/2^{j_1} < x_1<
212: L_1(l_1+1)/2^{j_1}$ and $L_2l_2/2^{j_2} < x_2< L_2(l_2+1)/2^{j_2}$.
213: The scale index ${\bf j}=(j_1,j_2)$, $j_1$ and $j_2$ can be any
214: positive integer. The position index ${\bf l}=(l_1,l_2)$, $l_1$ and
215: $l_2$ can be 0, 1... $2^{j_1}-1$, and 0, 1... $2^{j_2}-1$,
216: respectively. On the other hand, the wavelet function is admissible,
217: satisfying $\int \psi_{\bf j,l}({\bf x})d{\bf x}=0$.
218:
219: For the field $y({\bf x})$, the DWT variables are
220: %eq1
221: \begin{equation}
222: y_{\bf j,l}=\frac{1 }
223: {\int\phi_{\bf j,l}({\bf x})d{\bf x}}
224: \int y({\bf x})\phi_{\bf j,l}({\bf x})d{\bf x},
225: \end{equation}
226: %eq2
227: \begin{equation}
228: \tilde{\epsilon}^y_{\bf j,l}= \int y({\bf x}) \psi_{\bf j,l}({\bf
229: x})d{\bf x},
230: \end{equation}
231: Since $\phi_{\bf j,l}({\bf x})$ is a window function of cell
232: $\bf(j,l)$, $y_{\bf j,l}$ is the mean of $y(\bf x)$ in cell
233: $\bf(j,l)$. Because $\psi_{\bf j,l}({\bf x})$ is admissible,
234: $\tilde{\epsilon}^y_{\bf j,l}$ measures the fluctuation of field
235: $y({\bf x})$ with respect to the mean, $y_{\bf j,l}$, in the
236: mode $({\bf j,l})$. For a given ${\bf j}$, the mean of
237: $\langle \tilde{\epsilon}^y_{\bf j,l} \rangle$ over all modes is
238: zero (Fang \& Thews 1998). Since the basis $\psi_{\bf j,l}$ are
239: orthogonal and complete, the DWT variables of the field $y({\bf x})$
240: do not lost information and cause false correlations.
241:
242: With the DWT variables, the power spectrum of the 2-D field
243: $y({\bf x})$ can be measured as (Pando \& Fang 1998; Fang \& Feng
244: 2000):
245: %eq3
246: \begin{equation}
247: P_{\bf j}=\langle |\tilde{\epsilon}_{\bf j,l}|^2\rangle\equiv
248: \frac{1}{2^{j_1+j_2}}
249: \sum_{l_1=0;l_2=0}^{2^{j_1}-1; 2^{j_2}-1}
250: |\tilde{\epsilon}^y_{j_1,j_2;l_1,l_2}|^2,
251: \end{equation}
252: $|\tilde{\epsilon}^y_{j_1,j_2;l_1,l_2}|^2$ is the power at mode
253: $({\bf j,l})=(j_1,j_2;l_1,l_2)$. Therefore, the DWT power spectrum
254: $P_{\bf j}$ actually is the mean of local powers
255: $|\tilde{\epsilon}_{\bf j,l}|^2$ over all $2^{j_1+j_2}$ modes on
256: scale ${\bf j}$. It has been shown that the DWT power spectrum can
257: be expressed by the Fourier power spectrum as (Fang \& Feng 2000;
258: He et al. 2005a):
259: %eq4
260: \begin{equation}
261: P_{\bf j} = \frac{1}{2^{j_1+j_2}}
262: \sum_{n_1 = - \infty}^{\infty}
263: \sum_{n_2 = - \infty}^{\infty}
264: |\hat{\psi}(n_1/2^{j_1})\hat{\psi}(n_2/2^{j_2})|^2 P(n_1,n_2),
265: \end{equation}
266: where $\hat{\psi}(n)$ is the Fourier transform of the basic
267: wavelet $\psi(x)$. Because we have the normalized relation
268: $\sum_{n_1,n_2}|\hat{\psi}(n_1/2^{j_1})\hat{\psi}(n_2/2^{j_2})
269: |^2/2^{j_1+j_2}=1$, the DWT power spectrum $P_{\bf j}$ is equal to
270: the Fourier power spectrum banded on scales around
271: $(L_1L_2/2^{j_1+j_2})^{1/2}$. Figure 3 shows both the angular
272: power spectrum $C_l$ and DWT power spectrum $P_{\bf j}$ for the
273: simulation samples. The angular scale of $j$ is $19.1/2^j$ degree,
274: and therefore, $l\simeq 9.4\times 2^j$. Figure 3 shows that DWT
275: power spectrum is the same as the angular power spectrum $C_l$. We
276: will use the DWT analysis on angular scales of sub-degree, on
277: which the effect of spherical surface is not significant.
278:
279: \subsection{Non-Gaussianity of SZ effect map}
280:
281: The DWT power spectrum can effectively detect the non-Gaussian
282: feature of a field (Jamkhedkar et al. 2003). Taking this advantage,
283: one can reveal the non-Gaussianity of the SZ effect field $y({\bf
284: x})$ with the DWT power spectrum of the SZ field $y({\bf x})$
285: defined as
286: %eq5
287: \begin{equation}
288: P^{\rm{top}}_{\bf j}=\frac{1}{2^{j_1+j_2}} \sum_{\rm{top}}
289: |\tilde{\epsilon}_{j_1,j_2;l_1,l_2}|^2,
290: \end{equation}
291: where summation of ``top'' goes over only modes having high values
292: $|\tilde{\epsilon}_{j_1,j_2;l_1,l_2}|^2$ among the $2^{j_1+j_2}$
293: modes for a given ${\bf j}$. In Figure 4, we plot the $P_{\bf j}$
294: and $P^{\rm{top}}_{\bf j}$ of the simulation samples, in which top
295: power spectrum, $P^{\rm{top}}_{\bf j}$, is given by top 1\%,
296: 0.3\%, 0.1\% among the $2^{j_1+j_2}$ modes. We see that all the
297: power spectra $P^{\rm{top}}_{\bf j}$ are close to $P_{\bf j}$. For
298: samples with $\sigma_8=0.84$, there is more than $70\%$, $50\%$,
299: $35\%$ power attributes from top 1\%, 0.3\%, 0.1\% modes,
300: respectively. That is, the power $
301: |\tilde{\epsilon}^{\rm{top}}_{\bf j,l}|^2$ of each mode among the
302: top 1\% modes should be as high as about 70 times of the mean
303: $\langle{|\tilde{\epsilon}_{\bf j,l}|^2}\rangle=P_{\bf j}$. These
304: results show that the field $y({\bf x})$ is highly non-Gaussian.
305: From Figure 4, one can see that for sample with $\sigma_8=0.74$,
306: all the power spectra $P^{\rm{top}}_{\bf j}$ of top 1\%, 0.3\%,
307: 0.1\% are less close to $P_{\bf j}$. This is because the
308: clustering of $\sigma_8=0.84$ sample is stronger than that of
309: $\sigma_8=0.74$.
310:
311: Because $\langle \tilde{\epsilon}_{\bf j,l} \rangle =0$, the value
312: of $\langle{|\tilde{\epsilon}_{\bf j,l}|^2}\rangle^{1/2}$ actually
313: is the standard deviation($\sigma$) of the probability
314: distribution function (PDF) of $\tilde{\epsilon}_{\bf j,l}$. Thus,
315: the events of 70$\langle{|\tilde{\epsilon}_{\bf j,l}|^2}\rangle$
316: actually are $\sim$ 8$\sigma$ events. All the top 1\% wavelet
317: modes are 8$\sigma$ events. All the top 0.3\% and 0.1\% modes
318: correspond, respectively, to $\sim
319: 170$$\langle{|\tilde{\epsilon}_{\bf j,l}|^2}\rangle$, and $\sim
320: 350$$\langle{|\tilde{\epsilon}_{\bf j,l}|^2}\rangle$, or 13
321: $\sigma$ and 19 $\sigma$ events. For a Gaussian field, the 10
322: $\sigma$ events should be no more than $10^{-23}$. Therefore, the
323: PDF of $\tilde{\epsilon}_{\bf j,l}$ has a very long tail. In other
324: words, a significant part of the SZ effect DWT power spectrum is
325: given by the long tail events. These long tail events make the SZ
326: effect on sub-degree scales to be detectable. Although the mean of
327: SZ power $\tilde{\epsilon}_{\bf j,l}$ is less than about 2\% of
328: the power of CMB, for long tail events, the SZ power is comparable
329: with that of CMB.
330:
331: One can estimate the power spectrum of SZ effect temperature
332: fluctuations with the top modes (long tail event) of
333: $\tilde{\epsilon}_{\bf j,l}$. For observed maps, however, not all
334: the top modes of $\tilde{\epsilon}_{\bf j,l}$ are given by SZ
335: effect. To reduce this uncertainty, we add a condition that the
336: useful modes $({\bf j,l})$ for top $\tilde{\epsilon}_{\bf j,l}$
337: should also be the modes significantly showing SZ effect by the
338: WMAP-2MASS cross correlation. It is equal to assume that the power
339: of SZ effect temperature fluctuations mainly comes from clusters
340: (Komatsu \& Kitayama 1999; Komatsu \& Seljak 2002). We check this
341: assumption with the relation between variables $y_{\bf j,l}$ and
342: $\tilde{\epsilon}_{\bf j,l}$. We found that the top modes of
343: $|\tilde{\epsilon}_{\bf j,l}|$ are largely coincident with the top
344: modes of the Compton parameter $y_{\bf j,l}$. For sample of
345: $\sigma_8=0.84$, $84\%$ top modes of $y_{\bf j,l}$ on the angular
346: scale of $0^{\circ}.5$ are coincident with the top modes of
347: $|\tilde{\epsilon}_{\bf j,l}|$ on the same scale. This result
348: indicates that modes with high $y_{\bf j,l}$ generally are also the
349: modes of high local power of SZ effect temperature fluctuations.
350: High $y_{\bf j,l}$ generally is related to clusters.
351:
352: The fact that the coincidence between top $y_{\bf j,l}$ and top
353: $\tilde{\epsilon}_{\bf j,l}$ is not 100\% is due partially to the
354: so-called warm-hot intergalactic medium (WHIM) (Dave et al 2001; He
355: et al. 2004, 2005b). It shows that a significant fraction of hot
356: baryon gases ($T>10^5$) are not located in collapsed structures with
357: high density contrast $\rho/\bar{\rho}> 100$, but in the areas with
358: a median overdensity $10< \rho/\bar{\rho} < 30$. The WHIM is also
359: the source of SZ effect. Nevertheless, the major part of top modes
360: of $|\tilde{\epsilon}_{\bf j,l}|$ is associated with clusters.
361: Therefore, a proper method to estimate the SZ power spectrum would
362: be to identify the modes with top numbers of both $y_{\bf j,l}$ and
363: $|\tilde{\epsilon}_{\bf j,l}|$. The details of this method will be
364: given in the following two sections.
365:
366: \section{SZ effect identified from WMAP-2MASS correlation}
367:
368: \subsection{Data}
369:
370: The data of galaxies in the 2MASS extended source catalog (XSC,
371: Jarrett et al. 2000) used for cross-correlation analysis is the
372: same as Guo et al. (2004) and Cao et al. (2006). That is, we use
373: galaxies of ${\rm K\_m\_k20fe}$, which measures the magnitude
374: inside a elliptical isophote with surface brightness of 20 mag
375: ${\rm arcsec^{-2}}$ in $K_s$-band. There are approximately 1.6
376: million extended objects with $K_s<14.3$. To avoid the contaminant
377: of stars, we use a latitude cut of $|b| > 10^{\circ}$. We also
378: removed a small number of bright ($K_s<9$) sources by the
379: parameters of the XSC confusion flag (${\rm cc\_flag}$) and visual
380: verification score for source (${\rm vc}$). To eliminate duplicate
381: sources and have a uniform sample, we use ${\rm use\_src=1}$ and
382: ${\rm dup\_src=0}$ \footnote{The notations of the 2MASS parameters
383: used in this paragraph are from the list shown in the 2MASS Web
384: site http://www.ipac.caltech.edu/2mass/releases/allsky/doc.}. We
385: also use a cut of $10.0<K_s<14.0$ to ensure the sample is complete
386: greater than $90\%$. This sample contains 987,125 galaxies with
387: redshift up to $z \sim 0.1$.
388:
389: For the CMB temperature fluctuations, we use the full resolution
390: coadded 3 year sky map (Hinshaw et al. 2006) for $W$ band. The
391: maps have had the ILC(Internal Linear Combination) estimate of the
392: CMB signal to highlight the foreground emissions(Hinshaw et al.
393: 2006). We also use the three annual maps to construct three
394: difference maps between the 1st and 2nd, 2nd and 3rd, 3rd and 1st
395: year, which are useful to check statistical significance.
396:
397: As in Cao et al. (2006), we subject both the 2MASS and WMAP maps
398: to an equal-area projection. We then have 2-D maps of galaxy
399: number density, $n({\bf x})$, and CMB temperature fluctuations,
400: $\Delta T({\bf x})$ with coordinate ${\bf x}=(x_1,x_2)$. As has
401: been shown, the equal-area projection will not affect statistical
402: analysis on sub-degree scales, but helpful for using the DWT
403: method. We analyze the cross correlation between $n({\bf x})$ and
404: $\Delta T({\bf x})$ in two $123^{\circ}.88 \times 123^{\circ}.88$
405: areas, which are, respectively, in the northern and southern sky.
406: The 2MASS XSC galaxies are resolved to 10 arcsec. The beam size of
407: WMAP $W$ band map is
408: $0^{\circ}.22$\footnote{http://lambda.gsfc.nasa.gov/product},
409: while the scale of galaxy clusters at the median redshift of the
410: 2MASS samples is $0^{\circ}.5$.
411:
412: In order to apply the method of DWT power spectrum (\S 2.2), we also
413: use the DWT variables to describe the maps $n({\bf x})$ and
414: $\Delta T({\bf x})$. They are wavelet variables
415: $\tilde{\epsilon}^g_{\bf j,l}=
416: \int n({\bf x}) \psi_{\bf j,l}({\bf x})d{\bf x}$,
417: $\tilde{\epsilon}^T_{\bf j,l} = \int \Delta T({\bf x})\psi_{\bf
418: j,l}({\bf x})d{\bf x}$, and scaling function variables
419: $n_{\bf j,l}=\int n({\bf x})\phi_{\bf j,l}({\bf x})d{\bf x}/
420: \int\phi_{\bf j,l}({\bf x})d{\bf x}$,
421: $\Delta T_{\bf j,l}=
422: \int \Delta T({\bf x})\phi_{\bf j,l}({\bf x})d{\bf x}/
423: \int\phi_{\bf j,l}({\bf x})d{\bf x}$. The mode $({\bf j,l})$
424: occupy a area of $123^{\circ}.88/2^{j_1} \times
425: 123^{\circ}.88/2^{j_2}$ at position around
426: $[l_1(123^{\circ}.88)/2^{j_1}, l_2123^{\circ}.88)/2^{j_1}]$. The
427: angular distance between modes ${\bf l}$ and ${\bf l'}$ at scale
428: $j$ is given by $\theta =123.88^{\circ}|{\bf l-l'}|/2^j$.
429:
430: The variables $\tilde{\epsilon}^T_{\bf j,l}$ and
431: $\tilde{\epsilon}^g_{\bf j,l}$ describe, respectively, the
432: fluctuations of the CMB temperature and galaxy number density in
433: the mode $({\bf j, l})$, while the variables $\Delta T_{\bf j,l}$
434: and $n_{\bf j,l}$ describe, respectively, the mean temperature and
435: the mean number density of galaxies in the mode $({\bf j,l})$.
436: Therefore, $n_{\bf j,l}$ and $\Delta T_{\bf j,l}$ can be seen as
437: the maps of galaxies and CMB temperature functions smoothed on
438: scale ${\bf j}$.
439:
440: \subsection{DWT clusters}
441:
442: We picked up top clusters on scale ${\bf j}$ by the top members of
443: $n_{\bf j,l}$. These clusters are called DWT clusters. The details
444: of $n_{\bf j,l}$ identification of clusters have been studied with
445: simulation and real samples (Xu et al. 1999, 2000). It showed that
446: the clusters identified by DWT $n_{\bf j,l}$ are statistically the
447: same as the clusters identified by the friend-of-friend method if
448: the mean size of the friend-of-friend identified clusters is the
449: same as DWT clusters. The clusters identified by the
450: friend-of-friend method usually have very irregular shapes (Jing \&
451: Fang 1994) and are inconvenient to estimate the statistical
452: significance of the cross correlation. On the other hand, the DWT
453: variables of 2MASS galaxies $n_{\bf j,l}$ and WMAP map $\Delta
454: T_{j,l}$ are in the same cell, the statistical significance of the
455: $n_{\bf j,l}$-$\Delta T_{\bf j,l}$ cross correlation is unambiguous.
456: Scaling functions are orthogonal with each other, different DWT
457: clusters consist of different galaxies. This is also necessary for
458: statistical analysis which we will take below. With this method, we
459: identify top clusters on scales ${\bf j}=(8,8)$, corresponding to
460: scale $\simeq 1.8$ h$^{-1}$ Mpc of the 2MASS samples.
461:
462: \subsection{Thermal SZ effect of 2MASS clusters}
463:
464: The SZ effect of 2MASS galaxies can be measured with the
465: cross-correlation between the WMAP and 2MASS clusters defined as
466: Cao et al.(2006):
467: %eq6
468: \begin{equation}
469: \Delta T^{\rm wmap-2mass}(|{\bf l-l'}|)=
470: \langle C_{\bf j,l}\Delta T_{\bf j',l'}\rangle,
471: \end{equation}
472: where the variable $C_{\bf j,l}$ is given by:
473: %eq7
474: \begin{equation}
475: C_{\bf j,l}= \left \{ \begin{array}{ll}
476: 1 & {\rm if \ ({\bf j,l}) \ is \ a \ cell \ of
477: \ identified \ clusters}, \\
478: 0 & {\rm otherwise}.
479: \end{array} \right .
480: \end{equation}
481: The average $\langle ...\rangle$ covers all the DWT clusters on
482: scale ${\bf j}$. Therefore, $\Delta T^{\rm wmap-2mass}(|{\bf
483: l-l'}|)$ measures the mean of the CMB temperature fluctuations on
484: a distance $|{\bf l-l'}|$ from identified DWT clusters of 2MASS
485: data. When $|{\bf l-l'}|\gg 1$, $\Delta T^{\rm wmap-2mass}(|{\bf
486: l-l'}|)$ actually is the mean of $\Delta T$ on a large area, and
487: therefore, it should approach to zero. Figure 5 presents the cross
488: correlation $\Delta T^{\rm wmap-2mass}(|{\bf l-l'}|)$ between top
489: 100 ${\bf j}=(8,8)$ DWT clusters and WMAP map, in which the scales
490: ${\bf j'}$ of $\Delta T_{\bf j',l'}$ are taken to be ${\bf
491: j'}=(8,8)$, ${\bf j'}=(8,7)$, and ${\bf j'}=(7,7)$ from top to
492: bottom. In all case $\Delta T_{\bf j',l'}$ is from the full
493: resolution coadded 3 year WMAP map of $W$ band. The error bars are
494: given by the 1-$\sigma$ standard deviation of the 100 CMB
495: simulations.
496:
497: Figure 5 shows typical anti-correlation of SZ effect. At the
498: position of DWT clusters, i.e. $|\bf l-\bf l'|=0$, the mean of
499: temperature change $\Delta T_{\bf j',l'}$ is always negative,
500: which is expected from the thermal SZ effect on $W$ band. For
501: ${\bf j'}=(8,8)$, $\Delta T^{\rm wmap-2mass}(|{\bf l-
502: l'}|=0)\simeq -25 \pm 15$ $\mu$K, while for ${\bf j'}=(8,7)$ and
503: ${\bf j'}=(7,7)$, the temperature decreases are $-20 \pm 16$
504: $\mu$K and $-18 \pm 14$ $\mu$K respectively. These results are
505: consistent with that given by cross-correlation of the first year
506: WMAP map and 2MASS galaxies (Myers et al. 2004;
507: Hern\'andez-Monteagudo et al. 2004; Cao et al. 2006).
508:
509: To further check the SZ signals, we calculate the cross-correlation
510: of eq.(6) by using the maps given by the difference between the 1st
511: and 2nd, 2nd and 3rd, 3rd and 1st year. The results are shown in
512: Figure 6. All the anti-correlations shown in Figures 5 completely
513: disappear in Figure 6. It confirms that the temperature decreases at
514: the positions of the DWT clusters in the three annual maps are about
515: the same and not due to noise.
516:
517: To view the clusters richness dependence, Figure 7 plots $\Delta
518: T^{\rm wmap-2mass}(|{\bf l-l'}|=0)$ against the number of top
519: clusters, $n$. As expected, the higher the richness, the stronger
520: the SZ effect. $\Delta T^{\rm wmap-2mass}(|{\bf l-l'}|=0)$ rapidly
521: decreases with the number of top clusters, and approaches to the
522: noise level at $n>300$. We should, however, point out that the
523: positive correlation between the clusters richness and SZ effect
524: is statistical. That is, the order of clusters richness is not,
525: one-by-one, corresponding to the order of the amount of SZ effect
526: temperature decrease. Consequently, the relation of richness-SZ
527: effect in the range $n<100$ is not very stable. When $n \geq 100$,
528: the relation of richness-SZ effect, as shown in Figure 6 is very
529: stable. Therefore, in all the statistics below, the number of top
530: clusters should be $>100$.
531:
532: In the DWT analysis, wavelet modes are orthogonal from each other.
533: Therefore, the area fraction of top 300 ${\bf j}=(8,8)$ DWT
534: clusters is $300/(2\times256^2)=0.23\%$. As shown in \S 2, the
535: mean power of SZ effect of top 0.3\% modes would be 170 times
536: higher than mean power of total mode. On the other hand, on
537: sub-degree scales, the CMB power is higher than the SZ effect
538: power by a factor $\sim 10^2$. Therefore, we can expect that the
539: mean of SZ temperature decreases of the top 0.3\% modes should be
540: larger than the CMB temperature fluctuations by a factor $\sim
541: \sqrt{170/100}\simeq 1.3$. Figure 8 plots the richness-dependence
542: of the S/N ratio, i.e. the ratio between SZ effect temperature
543: decrease and the $1-\sigma$ error bar shown in Figure 7. Since the
544: error bar of Figure 7 is given by the simulation of CMB
545: temperature fluctuations, and therefore, it actually is to measure
546: the the CMB temperature fluctuations. We see from figure 8 that
547: the S/N ratio is about 1.3 - 1.6. This result is consistent with
548: the estimation mentioned above.
549:
550: \section{Estimation of DWT power spectrum of SZ effect}
551:
552: \subsection{Method}
553:
554: The observed CMB temperature fluctuations are given by:
555: %eq8
556: \begin{equation}
557: \Delta T({\bf x}) = \Delta T^{\rm cmb}({\bf x})+ \Delta T^{\rm
558: sz}({\bf x})+ \Delta T^{\rm second}({\bf x}),
559: \end{equation}
560: where $\Delta T^{\rm cmb}({\bf n})$ is the primeval temperature
561: fluctuations, $\Delta T^{\rm sz}({\bf x})$ is the SZ effect and
562: $\Delta T^{\rm second}({\bf n})$ is due to secondary effects other
563: than the SZ effect, such as the ISW effect and radio point
564: sources. Since the random fields $\Delta T^{\rm cmb}({\bf x})$,
565: $\Delta T^{\rm sz}({\bf x})$ and $\Delta T^{\rm second}({\bf x})$
566: basically are uncorrelated from each other, the DWT power
567: spectrum, $P_{\bf j}$, of the observed $\Delta T({\bf x})$
568: consists of three terms as:
569: %eq9
570: \begin{equation}
571: P_{\bf j} = P^{\rm cmb}_{\bf j}+ P^{\rm sz}_{\bf j}+P^{\rm
572: second}_{\bf j},
573: \end{equation}
574: where
575: %eq10
576: \begin{equation}
577: P^{\rm cmb, sz, second}_{\bf j}=\frac{1}{2^{j_1+j_2}}
578: \sum_{l_1=0;l_2=0}^{2^{j_1}-1; 2^{j_2}-1}
579: \left |\tilde{\epsilon}^{\rm cmb, sz,
580: second}_{j_1,j_2;l_1,l_2}\right|^2.
581: \end{equation}
582:
583: Since $P^{\rm cmb}_{\bf j}$ is about $10^2$ times larger than
584: $P^{\rm sz}_{\bf j}$ on angular scale of $\simeq 0^{\circ}.5$, one
585: cannot extract $P^{\rm sz}_{\bf j}$ from noisy map $\Delta T({\bf
586: x})$ if the field $\Delta T^{\rm sz}({\bf x})$ is Gaussian. As
587: shown in \S 2, the SZ effect DWT power spectrum is largely
588: given by the top modes. The local power of the top DWT modes can
589: be as large as $10^2$ times of their average. Therefore, the SZ
590: effect local power of the top modes is comparable to $P^{\rm
591: cmb}_{\bf j}$. On the other hand, the primeval temperature
592: fluctuations $\Delta T^{\rm cmb}({\bf x})$ are Gaussian, and the
593: PDF of local temperature fluctuations $\tilde{\epsilon}^T_{\bf
594: j,l}$ is Gaussian with zero mean and standard deviation $(P^{\rm
595: cmb}_{\bf j})^{1/2}$. The primeval temperature fluctuations map
596: does not have the modes with very high local power. Thus, the
597: local power of modes associated with top clusters would be
598: measurable from the noisy map of $\Delta T({\bf x})$.
599:
600: The measurability of power of top modes can be further tested as
601: follows. Subjecting eq.(8) to a wavelet transform, we have:
602: %eq11
603: \begin{equation}
604: \tilde{\epsilon}_{\bf j,l}= \tilde{\epsilon}^{\rm cmb}_{\bf j,l}+
605: \tilde{\epsilon}^{\rm sz}_{\bf j,l}+\tilde{\epsilon}^{\rm
606: second}_{\bf j,l},
607: \end{equation}
608: For most modes $({\bf j,l})$ on scales of clusters,
609: $|\tilde{\epsilon}^{\rm cmb}_{\bf j,l}|$ is much larger than
610: $|\tilde{\epsilon}^{\rm sz}_{\bf j,l}|$, and therefore, the latter
611: is negligible. But for modes corresponding to the top clusters,
612: $|\tilde{\epsilon}^{\rm sz}_{\bf j,l}|$ would be comparable to
613: $|\tilde{\epsilon}^{\rm cmb}_{\bf j,l}|$. On these modes, the term
614: $|\tilde{\epsilon}^{\rm second}_{\bf j,l}|$ is also negligible, as
615: statistically the positions of secondary effect are not coincident
616: to top clusters. Thus, on these modes, we have:
617: %eq12
618: \begin{equation}
619: |\tilde{\epsilon}_{\bf j,l}|^2=
620: |\tilde{\epsilon}^{\rm cmb}_{\bf j,l}|^2+ |\tilde{\epsilon}^{\rm
621: sz}_{\bf j,l}|^2,
622: \end{equation}
623: and therefore,
624: %eq13
625: \begin{equation}
626: \frac{1}{N_{\rm top}}\sum_{\rm top}|\tilde{\epsilon}^{\rm sz}_{\bf
627: j,l}|^2 = \frac{1}{N_{\rm top}}\sum_{\rm
628: top}|\tilde{\epsilon}_{\bf j,l}|^2 -\frac{1}{N_{\rm top}}\sum_{\rm
629: top}|\tilde{\epsilon}^{\rm cmb}_{\bf j,l}|^2,
630: \end{equation}
631: where the summation ``top'' is over modes with top
632: $\tilde{\epsilon}_{\bf j,l}$ and also the top DWT clusters.
633: $N_{\rm top}$ is the number of top modes. Because CMB field is
634: statistically isotropic (uniform), Gaussian and uncorrelated with
635: 2MASS data, the average of the primeval temperature fluctuations
636: power $|\tilde{\epsilon}^{\rm cmb}_{\bf j,l}|^2$ done by
637: $(1/N_{\rm top})\sum_{\rm top}$ should be about the same as the
638: average over all modes. We have then:
639: %eq14
640: \begin{equation}
641: \sum_{\rm top}|\tilde{\epsilon}^{\rm cmb}_{\bf j,l}|^2\simeq
642: N_{\rm top}\langle |\tilde{\epsilon}_{\bf j,l}|^2\rangle,
643: \end{equation}
644: where $\langle |\tilde{\epsilon}_{\bf j,l}|^2\rangle$ is the mean
645: of $|\tilde{\epsilon}_{\bf j,l}|^2$ over all modes on scale ${\bf
646: j}$. Thus, considering $|\tilde{\epsilon}^{\rm sz}_{\bf
647: j,l}|^2>0$, eq.(13) yields:
648: %eq15
649: \begin{equation}
650: \frac{1}{N_{\rm top}}\sum_{\rm top}|\tilde{\epsilon}_{\bf j,l}|^2
651: > \langle |\tilde{\epsilon}_{\bf j,l}|^2\rangle,
652: \end{equation}
653: Eq.(15) requires that the mean local power over modes at top
654: clusters should be larger than the mean on all modes.
655:
656: We test Eq.(15) by the cross correlation defined similar to
657: eq.(6):
658: %eq16
659: \begin{equation}
660: \Upsilon_{\bf j'}(|{\bf l-l'}|)= \frac{
661: \langle C_{\bf j,l}|\tilde{\epsilon}_{\bf j',l'}|^2 \rangle}
662: {\langle |\tilde{\epsilon}_{\bf j',l'}|^2\rangle},
663: \end{equation}
664: where $C_{\bf j,l}$ is given by eq.(7). The numerator on the
665: r.h.s. of eq.(16) is to measure the mean local power of CMB
666: temperature fluctuations over the modes of DWT clusters on scale
667: ${\bf j'}$. The denominator of eq.(16) is for normalization. The
668: inequality in eq.(15) requires $\Upsilon_{\bf j'}(|{\bf l-l'}|)>1$
669: at $|\bf l-\bf l'|=0$, i.e. the local power of the modes, which
670: shows significant SZ temperature decrease, should also show power
671: larger than the mean $\langle |\tilde{\epsilon}_{\bf
672: j',l'}|^2\rangle$ over the whole sky. The result of $\Upsilon_{\bf
673: j'}(|{\bf l-l'}|)$ is plotted in Figure 9, in which the scale
674: ${\bf j'}$ of local power $|\tilde{\epsilon}_{\bf j',l'}|^2$ is
675: taken to be ${\bf j'}=(8,8)$, ${\bf j'}=(8,7)$ and ${\bf
676: j'}=(7,7)$. For each scale, we count 0.11\% modes with top
677: $|\tilde{\epsilon}_{\bf j',l'}|^2$, and all the counted modes are
678: located in the positions of the 300 top 2MASS clusters,
679: corresponding to 150, 75, and 38 top modes of the scales ${\bf
680: j'}=(8,8)$, ${\bf j'}=(8,7)$ and ${\bf j'}=(7,7)$. Figure 9 shows
681: that all the mean powers of CMB temperature fluctuations on the
682: three scales are significantly larger than the mean
683: $\langle|\tilde{\epsilon}_{\bf j',l'}|^2\rangle$. This power
684: excess is due to the SZ effect of the 2MASS clusters, which are
685: larger than the noise of WMAP III data.
686:
687: \subsection{Results}
688:
689: With these preparation, we can estimate the SZ effect power
690: spectrum given by modes associated with the top DWT clusters of
691: 2MASS galaxies. It is,
692: %eq17
693: \begin{equation}
694: P^{\rm top}_{\bf j}=\frac{1}{2^{j_1+j_2}} \sum_{\rm
695: top}|\tilde{\epsilon}^{\rm sz}_{\bf j,l}|^2,
696: \end{equation}
697: where the ``top'' is taken the same modes as shown in Figure 9.
698: Using eqs.(13) and (14), eq.(17) gives,
699: %eq18
700: \begin{equation}
701: P^{\rm top}_{\bf j} \simeq \frac{1}{2^{j_1+j_2}}\left [ \sum_{\rm
702: top}
703: |\tilde{\epsilon}_{\bf j,l}|^2-N_{\rm top}
704: \langle |\tilde{\epsilon}_{\bf j,l}|^2\rangle \right ].
705: \end{equation}
706: This is the basic formula of extracting the SZ temperature
707: fluctuations power spectrum of 2MASS clusters from the WMAP III
708: map. The result is presented in Figure 10, in which we show the
709: $P^{\rm top}_{\bf j}$ given by 0.11\% top modes(black points). The
710: scales of ${\bf j}=(8,8)$, (8,7) and (7,7) is corresponding to
711: angular $0^{\circ}.48$, $0^{\circ}.68$ and $0^{\circ}.96$,
712: respectively. The error bars are given by jack-knife. As a
713: comparison, Figure 10 also shows the simulation results of both
714: $P_{\bf j}$ (dashed lines) and $P^{\rm top}_{\bf j}$ for top
715: 0.11\% modes(solid lines). The error bars are given by the 40
716: independent simulation maps using jack-knife.
717:
718: The shape of all the power spectra shown in Figure 10 are similar.
719: Their amplitudes are, however, very different. The extracted
720: $P^{\rm top}_{\bf j}$ of SZ effect from the 0.11\% top modes
721: associated with top 2MASS DWT clusters is much higher than $P^{\rm
722: top}_{\bf j}$ of the 0.11\% top modes of the simulation map with
723: $\sigma_8=0.74$. The data points are about the same, and even
724: higher than $P_{\bf j}$ of sample $\sigma_8=0.74$. Since $P_{\bf
725: j}$ is given by all modes of the sample, while $P^{\rm top}_{\bf
726: j}$ is only from the top 0.11\% mode. From eqs.(3) and (17) we
727: must have $P_{\bf j}\geq P^{\rm top}_{\bf j}$. Therefore, the result of
728: sample $\sigma_8=0.74$ is unacceptable with confidence level 99\%.
729: On the other hand, for sample of $\sigma_8=0.84$, the extracted SZ
730: power spectrum (data points) is safely lower than the simulation
731: $P_{\bf j}$, and it is also basically consistent with 0.11\% top
732: modes $P^{\rm top}_{\bf j}$ of the simulation map.
733:
734: Therefore, we can conclude that the power spectrum of the SZ
735: temperature fluctuations on sub-degree scales extracted from the
736: WMAP III shows a significant deviation from sample with the
737: parameter of density fluctuations amplitude $\sigma_8=0.74$, but
738: is consistent with the simulation sample produced with
739: $\sigma_8=0.84$, if all other cosmological parameters are the same
740: as that given by the three-year WMAP.
741:
742: \section{Discussions and conclusions}
743:
744: We studied the SZ-effect-induced temperature decrease and
745: fluctuations of the three-year WMAP maps caused by the 2MASS
746: galaxies. The DWT variables are very useful to study this problem,
747: as the two sets of DWT variables, given by the decompositions of
748: the scaling function and wavelet, are just suitable to describe,
749: respectively, the SZ temperature decrease and temperature
750: fluctuations.
751:
752: We show that the maps of both SZ temperature decrease and
753: temperature fluctuations are highly non-Gaussian. The PDF of the
754: DWT local power has a long tail. More than $70\%$ power of the SZ
755: effect on clusters scales is given by $1\%$ top modes associated
756: with top clusters. On the other hand, the field of CMB temperature
757: fluctuations is Gaussian. One can measure the long tail events of
758: the SZ effect from the noisy WMAP III maps. These long tail modes
759: yield an useful estimation of SZ effect power spectrum. It gives,
760: at least, a robust lower limit of the power spectrum of the SZ
761: temperature fluctuations.
762:
763: We found that the WMAP temperature shows significant decrease at
764: the positions from 100 up to $\sim$ 300 top clusters of 2MASS
765: galaxies. We also found that there is significant power excess at
766: the positions of these top clusters. This power excess is a
767: measurement of the SZ effect power spectrum. We find that the
768: power spectrum given by the 0.11\% top modes associated with top
769: clusters of the 2MASS galaxies, shows even higher amplitude than
770: that of simulation sample with $\sigma_8=0.74$. On the other hand,
771: the power spectrum of these top modes is bascially consistent with
772: $\sigma_8=0.84$ simulation SZ effect map. Therefore, the SZ effect
773: power spectrum on sub-degree scales seems to favor parameter
774: $\sigma_8>0.74$. This result supports the direct measurement of
775: the SZ effect power excess on arcminute scales from data of CBI
776: (Mason et al.2003; Readhead et al. 2004) and BIMA (Dawson et al.
777: 2006).
778:
779: \acknowledgments
780:
781: This work is in partial supported by the NSF AST-0507340. We
782: acknowledge the use of the HEALPix software and analysis package
783: for producing simulation maps of the CMB. Liang Cao acknowledges
784: the fellowship provided by the International Center for
785: Relativistic Astrophysics.
786:
787: \begin{references}
788:
789: \reference{} Afshordi, N., Loh, Y.-S.,\& Strauss, M. A. 2004,
790: Phys. Rev. D, 69, 083524
791:
792: \reference{} Afshordi N., Lin Y.-T., \& Sanderson A. J. R. 2005,
793: \apj, 629, 1
794:
795: \reference{} Bacon, D. J., Massey, R. J., Refregier, A. R., \&
796: Ellis, R. S., 2003, \mnras, 344, 673
797:
798: \reference{} Bahcall, N. A. \& Bode, P. 2003, \apjl, 588, L1
799:
800: \reference{} Birkinshaw, M. 1999, Phys.Rept., 310, 97
801:
802: \reference{} Bond, J. R. et al. 2002, \apj, 626, 12
803:
804: \reference{} Cai, Y. C., Pan, J., Zhao, Y. H., Feng L. L. \& Fang,
805: L.Z. 2006, astro-ph/0609411
806:
807: \reference{} Cao, L., Chu, Y. Q., Fang, L. Z. 2006, \mnras,
808: 369,645
809:
810: \reference{} Cooray, A., B. Daniel, Sigurdson, K. 2004,
811: astro-ph/0410006
812:
813: \reference{} Carlstrom, J. E., Holder, G. P., Reese, E. D. 2002,
814: Ann.Rev.Astron.Astrophys. 40, 643
815:
816: \reference{} Dawson, K.S., Holzapfel, W.L., Carlstrom, J.E.,
817: Joy, M. \& LaRoque, S.J. 2006, astr-ph/0602413
818:
819: \reference{} Daubechies I. 1992, Ten Lectures on
820: Wavelets,(Philadelphia: SIAM)
821:
822: \reference{} Dave, R. et al. 2001, \apj, 552,47300 \apj, 508, 472
823:
824: \reference{} Donoho, D.L., 1995, IEEE Transf. Inf. Theory, 41, 613
825:
826: \reference{} Fang, L.Z. \& Feng, L.L. 2000, \apj, 539, 5
827:
828: \reference{} Fang, L.Z. \& Thews, R. 1998, Wavelets in Physics,
829: World Scientific,(Singapore)
830:
831: \reference{} Feng, L.L., Shu, C. W. \& Zhang, M. P. 2004, \apj,
832: 612,1
833:
834: \reference{} Fedkiw, R.P., Sapiro, G., \& Shu, C.W. 2003, J.
835: Comput. Phys., 185, 309
836:
837: \reference{} Guo,Y.C., Chu,Y.Q. , \& Fang, L.Z. 2004, \apj, 610,
838: 51
839:
840: \reference{} Harten, A., Osher, S., Engquist, B., \& Chakravarthy,
841: S. 1986, Appl. Numerical Math., 2, 347
842:
843: \reference{} He, P., Feng, L.L., \& Fang, L.Z. 2004, \apj, 612, 14
844:
845: \reference{} He, P., Feng, L.L., \& Fang, L.Z. 2005a, \apj, 628,
846: 14
847:
848: \reference{} He, P., Feng, L.L., \& Fang, L.Z. 2005b, \apj, 623,
849: 601
850:
851: \reference{} He, P., Liu J. R., Feng, L.L., Shu C. W., Fang, L.Z.
852: 2006,\prl, 96, 051302
853:
854: \reference{} Hern\'andez-Monteagudo, C., Genova-Santos R. \&
855: Atrio-Barandela, F. 2004, \apj, 613, L89
856:
857: \reference{} Hinshaw, G. et al. 2006, astro-ph/0603451
858:
859: \reference{} Hoekstra, H., Yee H. K. C., \& Gladders, M. D. 2002,
860: \apj, 577, 595
861:
862: \reference{} Huffenberger,K.M., Seljak,U.,\& Makarov, A. 2004,
863: Phys. Rev. D,70, 063002
864:
865: \reference{} Jamkhedkar, P., Feng, L.L., Zheng, W., Kirkman, D.,
866: Tytler, D. \& Fang, L.Z. 2003, \mnras, 343, 1110
867:
868:
869: \reference{} Jarrett, T. H.,Chester, T.,Cutri, R., Schneider, S.,
870: Skrutskie, M.,\& Huchra, J. P. 2000, \aj, 119, 2498
871:
872: \reference{} Jiang, G. \& Shu, C.W. 1996, J. Comput. Phys., 126,
873: 202
874:
875: \reference{} Jing, Y.P.,\& Fang, L.Z. 1994 \apj, 432,438
876:
877: \reference{} Komatsu, E. \& Seljak, U. 2002, \mnras, 336, 1256
878:
879: \reference{} Komatsu, E. \& Kitayama, T. 1999, \apjl, 526,1
880:
881: \reference{} Liu, X.D., Osher, S., \& Chan, T. 1994, J. Comput.
882: Phys., 115, 200
883:
884: \reference{} Liu, J.R., Jamkhedkar,J., Zheng,W., Feng, L.L., \&
885: Fang, L.Z. 2006,\apj, 645,861
886:
887: \reference{} Mason, B. S. et al. 2003, \apj, 591,540
888:
889: \reference{} McCarthy, I.G., Bower, R.G., Balogh, M.L. 2006,
890: astro-ph/0609314
891:
892: \reference{} Myers, A. D.; Shanks, T.; Outram, P. J.; Frith, W.
893: J.; Wolfendale, A. W. 2004 \mnras ,347,67
894:
895: \reference{} Pando, J.,\& Fang, L.Z. 1998, A\&A, 340, 335
896:
897: \reference{} Pen, U.L. 1997, \apj, 490, L127
898:
899: \reference{} Readhead, A. C. S. et al. 2004, \apj, 609, 498
900:
901: \reference{} Refregier, A., Komatsu, E., Sperge,l D.N., Pen U. L.
902: 2000, Phys.Rev. D, 61, 123001
903:
904: \reference{} Refregier, A., Rhodes, J. \& Groth, E. J. 2002,
905: \apjl, 572, L131
906:
907: \reference{} Reiprich, T.H., B{\"o}hringer, H. 2002, \apj, 567,
908: 716
909:
910: \reference{} Seljak, U., \& Zaldarriaga, M. 1996, \apj, 469, 437
911:
912: \reference{} Seljak, U., Burwell J., Pen, U. L. 2001, Phys.Rev. D,
913: 63, 063001
914:
915: \reference{} Seljak, U. et al. 2005, \prd, 71, 103515
916:
917: \reference{} Shu, C.W. 1998, in Advanced Numerical Approximation
918: of Nonlinear Hyperbolic Equations, ed. A. Quarteroni (Berlin:
919: Springer), 325
920:
921: \reference{} Shu, C.W. 2003, Int. J. Comput. Fluid Dyn., 17, 10
922:
923: \reference{} Sirko,E. 2005,\apj,634,728
924:
925: \reference{} Spergel, D. N. et al. 2006, astro-ph/0603449
926:
927: \reference{} Theuns, T. et al. 1998, \mnras, 301, 478
928:
929: \reference{} Van Waerbeke, L., Mellier, Y., Pell{\'o}, R., Pen,
930: U.L., McCracken, H.J. \& Jain, B. 2002, \aap, 393, 369
931:
932: \reference{} Viel, M. \& Haehnelt, M.G. 2006 ,\mnras, 365, 231
933:
934: \reference{} Xu, W. Fang, L.Z., Deng, Z.G. 1999, \apj, 524, 1
935:
936: \reference{} Xu, W., Fang, L.Z., Wu, X.P. 2000, \apj, 508, 472
937:
938: \reference{} Yang, X.H., Feng, L.L., Chu, Y.Q. \& Fang, L.Z. 2001,
939: \apj, 553, 1
940:
941: \reference{} Zhang, T. J., Liu, J. R., Feng, L.L., Fang, L.Z.
942: 2006, \apj,642,625
943: \end{references}
944:
945: \clearpage
946: %Fig1
947: \begin{figure}
948: \centerline{\includegraphics[width=12.0cm,angle=0]{f1a.eps}}
949: \vspace{2mm}
950: \centerline{\includegraphics[width=12.0cm,angle=0]{f1b.eps}}
951: \caption{The SZ effect Compton parameter $y$ maps of simulation
952: samples with redshift $z\leq 0.1$ and $\sigma_8=0.74$(top) and
953: $\sigma_8=0.84$(bottom).}
954: \end{figure}
955:
956: %Fig2
957: \begin{figure}
958: \figurenum{2}\epsscale{1.0}\plotone{f2.eps} \figcaption{Angular
959: power spectra of SZ effect $y$ maps for samples given by
960: hydrodynamic simulations with parameter $\sigma_8=0.74$(solid) and
961: 0.84(dashed).}
962: \end{figure}
963:
964: %Fig3
965: \begin{figure}
966: \figurenum{3}\epsscale{1.0}\plotone{f3.eps} \figcaption{A comparison
967: between the angular power spectra of SZ effect parameter $y$
968: maps given by the Fourier modes (top panel) and DWT modes (bottom panel).
969: In each panel, the solid and dashed lines represent $\sigma_8=0.74$
970: and 0.84 respectively. $l$ at top panel is the index of spherical
971: harmonic function. Scale $j$ at the bottom panel corresponds
972: to angular scale of $19.1^{\circ}/2^j$.}
973: \end{figure}
974:
975: %Fig4
976: \begin{figure}
977: \figurenum{4}\epsscale{1.0}\plotone{f4.eps} \figcaption{DWT power
978: spectrum of SZ effect for simulation samples of $\sigma_8=0.74$(top)
979: and 0.84(bottom). In each panel, we show the DWT power spectra of
980: whole modes (solid), and 1\%(dot-dashed), 0.3\%(dashed) and
981: 0.1\%(dotted) modes with top power. Scale $j$ is corresponding to
982: angular scale of $19.1^{\circ}/2^j$.}
983: \end{figure}
984:
985: %Fig5
986: \begin{figure}
987: \figurenum{5}\epsscale{1.0}\plotone{f5.eps} \figcaption{Cross
988: correlation $\Delta T^{\rm wmap-2mass}(|\bf l-l'|)$[eq.(6)] between
989: the top 100 ${\bf j}=(8,8)$ DWT clusters of 2MASS galaxies and
990: $\Delta T_{\bf j',l}$ given by the full resolution coadded 3 year
991: sky map for $W$ band. The sacle of $\Delta T_{\bf j',l}$ is taken to
992: be ${\bf j'}=(8,8)$ (top), $(8,7)$ (middle) and $(7,7)$ (bottom). In
993: each panel, the error bars are given by the 1-$\sigma$ error of
994: cross correlation between the top 2MASS DWT clusters with 100 CMB
995: simulations. The angular scale of $|\bf l-l'|$ is $|{\bf
996: l-l'}|123.88^{\circ}/2^8$ degree.}
997: \end{figure}
998:
999: %Fig6
1000: \begin{figure}
1001: \figurenum{6}\epsscale{1.0}\plotone{f6.eps} \figcaption{Cross
1002: correlation $\Delta T^{\rm wmap-2mass}(|\bf l-l'|)$[eq.(6)] between
1003: the top 100 ${\bf j}=(8,8)$ DWT clusters of 2MASS galaxies and
1004: $\Delta T_{\bf j',l'}$ given by map of the differences between 1st
1005: and 2nd years (solid), 2nd and 3rd years (dotted), 3rd and 1st years
1006: (dashed) of WMAP data on $W$ band. The sacle of $\Delta T_{\bf
1007: j',l}$ is taken to be ${\bf j'}=(8,8)$ (top), $(8,7)$ (middle) and
1008: $(7,7)$ (bottom). The angular scale of $|\bf l-l'|$ is $|{\bf
1009: l-l'}|123.88^{\circ}/2^8$.}
1010: \end{figure}
1011:
1012: %Fig7
1013: \begin{figure}
1014: \figurenum{7}\epsscale{1.0}\plotone{f7.eps} \figcaption{The cluster
1015: richness dependence of the SZ effect temperature decrease $\Delta
1016: T^{\rm wmap-2mass}(0)$, in which $n$ is the number of top ${\bf
1017: j}=(8,8)$ clusters of 2MASS galaxies. The scale of $\Delta T_{\bf
1018: j',l}$ is taken to be ${\bf j'}=(8,8)$ (top), $(8,7)$ (middle) and
1019: $(7,7)$ (bottom). The error bars are given by the 1-$\sigma$ error
1020: of cross correlation between the top 2MASS DWT clusters with 100 CMB
1021: simulations.}
1022: \end{figure}
1023:
1024: %Fig8
1025: \begin{figure}
1026: \figurenum{8}\epsscale{1.0}\plotone{f8.eps} \figcaption{The
1027: signal-to-noise ratio of the SZ effect temperature decrease
1028: $\Delta T^{\rm wmap-2mass}(0)$ against clusters richness, in which
1029: $n$ is the number of top ${\bf j}=(8,8)$ clusters of 2MASS
1030: galaxies. The scale of $\Delta T_{\bf j',l}$ is taken to be ${\bf
1031: j'}=(8,8)$ (top), $(8,7)$ (middle) and $(7,7)$ (bottom).}
1032: \end{figure}
1033:
1034:
1035: %Fig9
1036: \begin{figure}
1037: \figurenum{9}\epsscale{1.0}\plotone{f9.eps}
1038: \figcaption{$\Upsilon_{\bf j'}(|\bf l-l'|)$ [eq.(16)]. The average
1039: is taken over 0.11\% modes with top $|\tilde{\epsilon}_{\bf
1040: j',l'}|^2$ in the positions of 300 top cluster with ${\bf j}=(8,8)$.
1041: The scales ${\bf j'}$ is taken to be ${\bf j'}=(8,8)$ (top), ${\bf
1042: j'}=(8,7)$ (middle) and ${\bf j'}=(7,7)$(bottom). The angular scale
1043: of $|\bf l-l'|$ is $|{\bf l-l'}|123.88^{\circ}/2^8$.}
1044: \end{figure}
1045:
1046:
1047: %Fig10
1048: \begin{figure}
1049: \figurenum{10}\epsscale{1.0}\plotone{f10.eps} \figcaption{The DWT
1050: power spectrum of SZ effect temperature fluctuations at W band.
1051: The results from eq.(18) with 0.11\% top modes associated with the
1052: top DWT clusters is shown by black point with error bar. The
1053: dashed and solid lines show the power spectrum of simulation
1054: sample with all modes and 0.11\% modes respectively. The top panel
1055: is for sample of $\sigma_8=0.74$, the bottom is for
1056: $\sigma_8=0.84$. Scale $j$ is corresponding to angular scale of
1057: $123.88^{\circ}/2^j$. The three black points of scales of ${\bf
1058: j}=(8,8)$, (8,7) and (7,7) have angular scale of $0^{\circ}.48$,
1059: $0^{\circ}.68$ and $0^{\circ}.96$, respectively.}
1060: \end{figure}
1061: \end{document}
1062: