1: \def\boldsymbol{\bf}
2:
3:
4:
5: %\documentclass[12pt, preprint]{aastex}
6: %\documentclass[10pt, preprint]{emulateapj}
7: \documentclass{emulateapj}
8: %\usepackage{epsfig}
9: %\usepackage{changebar}
10: %\usepackage{natbib}
11: %use emulateapj for testing - remove for apj submission
12: %\usepackage{emulateapj5}
13: %\usepackage{graphicx}
14: %\usepackage{rotate}
15:
16: % JR's adds to make figures appear in better places
17: %
18: \renewcommand{\topfraction}{1}
19: \renewcommand{\bottomfraction}{1}
20: \renewcommand{\dbltopfraction}{1}
21: \renewcommand{\textfraction}{0}
22:
23: \topmargin 0.3in
24:
25: \newcommand{\vdag}{(v)^\dagger}
26: \newcommand{\myemail}{}
27: \bibpunct{(}{)}{;}{a}{}{,}
28:
29: %% You can insert a short comment on the title page using the command below.
30:
31: \slugcomment{To appear in ApJ}
32:
33:
34: \shorttitle{ACBAR Power Spectrum}
35: \shortauthors{Reichardt et al.}
36:
37: %% This is the end of the preamble. Indicate the beginning of the
38: %% paper itself with \begin{document}.
39:
40:
41: \usepackage{amsmath}
42: \usepackage{amssymb}
43:
44: \begin{document}
45:
46:
47: \title{High resolution CMB power spectrum from the complete ACBAR data set.}
48:
49:
50:
51: %% Use \author, \affil, and the \and command to format
52: %% author and affiliation information.
53: %% Note that \email has replaced the old \authoremail command
54: %% from AASTeX v4.0. You can use \email to mark an email address
55: %% anywhere in the paper, not just in the front matter.
56: %% As in the title, you can use \\ to force line breaks.
57:
58:
59: \author{ C.L. Reichardt\altaffilmark{1},
60: P.A.R. Ade\altaffilmark{2},
61: J.J. Bock\altaffilmark{1,3},
62: J.R. Bond\altaffilmark{4},
63: J.A. Brevik\altaffilmark{1},
64: C.R. Contaldi\altaffilmark{5},
65: M.D. Daub \altaffilmark{6},
66: J.T. Dempsey \altaffilmark{7},
67: J.H. Goldstein\altaffilmark{8,9},
68: W.L. Holzapfel \altaffilmark{6},
69: C.L. Kuo\altaffilmark{1,10},
70: A.E. Lange\altaffilmark{1,3},
71: M. Lueker\altaffilmark{6},
72: M. Newcomb\altaffilmark{11},
73: J.B. Peterson\altaffilmark{12},
74: J. Ruhl\altaffilmark{8},
75: M.C. Runyan\altaffilmark{1},
76: Z. Staniszewski\altaffilmark{8}}
77:
78: \altaffiltext{1}{Observational Cosmology, California Institute of Technology, MS 59-33, Pasadena, CA 91125}
79: \altaffiltext{2}{Department of Physics and Astronomy, Cardiff University, CF24 3YB Wales, UK}
80: \altaffiltext{3}{Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA 91109}
81: \altaffiltext{4}{Canadian Institute of Theoretical Astrophysics, University of Toronto, Toronto, Ontario, M5S3H8,
82: Canada}
83: \altaffiltext{5}{Blackett Laboratory, Imperial College, Prince Consort Road, London, SW7 2AZ, U.K.}
84: \altaffiltext{6}{Department of Physics, University of California, Berkeley, CA 94720}
85: \altaffiltext{7}{Joint Astronomy Centre, Hilo HI 96720}
86: \altaffiltext{8}{Department of Physics, Case Western Reserve University, Cleveland, OH 44106}
87: \altaffiltext{9}{Aret\'{e} Associates, Arlington, VA 22202}
88: \altaffiltext{10}{Department of Physics and KIPAC, Stanford University, Stanford, CA 94305}
89: \altaffiltext{11}{Yerkes Observatory, 373 W. Geneva Street, Williams Bay, WI 53191}
90: \altaffiltext{12}{Department of Physics, Carnegie Mellon University, Pittsburgh, PA 15213}
91:
92: \begin{abstract}
93:
94: In this paper, we present results from the complete set of cosmic
95: microwave background (CMB) radiation temperature anisotropy
96: observations made with the Arcminute Cosmology Bolometer Array
97: Receiver (ACBAR) operating at $150\,$GHz. We include new data from
98: the final 2005 observing season, expanding the number of
99: detector-hours by 210\% and the sky coverage by 490\% over that used
100: for the previous ACBAR release. As a result, the band-power
101: uncertainties have been reduced by more than a factor of two on
102: angular scales encompassing the third to fifth acoustic peaks as
103: well as the damping tail of the CMB power spectrum. The calibration
104: uncertainty has been reduced from 6\% to 2.1\% in temperature
105: through a direct comparison of the CMB anisotropy measured by ACBAR
106: with that of the dipole-calibrated WMAP5 experiment. The measured
107: power spectrum is consistent with a spatially flat, $\Lambda$CDM
108: cosmological model. We include the effects of weak lensing in the power spectrum model computations and find that this significantly improves the
109: fits of the models to the combined ACBAR+WMAP5 power spectrum.
110: The preferred strength of the lensing is consistent with theoretical expectations. On fine angular scales, there is weak evidence ($1.1\sigma$) for excess power above the level expected from primary anisotropies. We expect any excess power to be dominated by the combination of emission from
111: dusty protogalaxies and the Sunyaev-Zel'dovich effect (SZE).
112: However, the excess observed by ACBAR is significantly smaller than the excess power at $\ell >2000$ reported by the CBI experiment operating at $30\,$GHz.
113: Therefore, while it is unlikely that the CBI excess has a primordial origin;
114: the combined ACBAR and CBI results are consistent with the source of
115: the CBI excess being either the SZE or radio source contamination.
116:
117: \end{abstract}
118:
119: \keywords{cosmic microwave background --- cosmology: observations}
120:
121: \section{Introduction}\label{sec:intro}
122:
123: Observations of the cosmic microwave background (CMB) are among the most
124: powerful and important tests of cosmological theory. Measurements of the
125: angular power spectrum of CMB temperature anisotropies on angular scales $>
126: 10^\prime$
127: - corresponding to multipoles $\ell \lesssim 1000$ - \citep{spergel06} in conjunction with other cosmological probes \citep{burles01,cole05,tegmark06,riess07} have produced compelling evidence for the $\Lambda$CDM cosmological model. At higher multipoles, measurements probe the Silk damping tail of the power spectrum and provide an independent check of the cosmological model.
128:
129: At smaller angular scales, the primary CMB anisotropies originating
130: at redshift z = 1100 are exponentially damped by photon diffusion. This
131: effect, known as Silk damping, makes secondary anisotropies - those induced
132: along the line of sight at lower redshift - increasingly important at higher
133: $\ell$. At 150$\,$GHz, for example, the Sunyaev-Zel'dovich effect (SZE) is expected
134: to be brighter than the primary CMB anisotropy at $\ell \gtrsim 2500$. The amplitude
135: of the SZE depends sensitively on the amplitude of the matter perturbations,
136: scaling as $\sigma_8^7$. Measurements of the CMB power spectrum
137: with sufficient sensitivity on arcminute scales
138: not only extend tests of the $\Lambda$CDM
139: model's ability to accurately predict the features in the power spectrum of
140: primary CMB anisotropy, but also probe the epoch of cluster formation and
141: provide an independent measure of $\sigma_8$.
142:
143: In this paper, we present the complete results of observations of CMB
144: temperature anisotropies at 150 GHz with 5$^\prime$ resolution from the
145: Arcminute Cosmology Bolometer Array Receiver (ACBAR)
146: experiment at the South Pole station. Previous measurements of the CMB power
147: spectrum by ACBAR have been presented in \cite{kuo04} (hereafter K04)
148: and \cite{kuo07} (hereafter K07). In addition, the angular power spectrum
149: on these angular scales has been measured at 30 GHz by CBI \citep{readhead04}, VSA \citep{dickinson04}, and BIMA \citep{dawson06}, and at 100
150: and 150 GHz by QuAD \citep{quad07}.
151:
152:
153:
154:
155: To date, measurements at angular scales $<10^\prime$ have been
156: consistent with predictions of the primary anisotropy based on measurements
157: at larger angular scales, with one exception. Both CBI~\citep{mason03,bond05} and BIMA~\citep{dawson06} observe excess power for $\ell > 2000$ at 30 GHz compared to
158: the predictions of the $\Lambda$CDM model.
159: This excess can be explained by the SZE if $\sigma_8 \approx 1$,
160: but this value is in tension with the best-fit WMAP5 value of $\sigma_8 \approx 0.8$.
161: In K07, we found that while the frequency dependence of the excess is consistent with the SZE,
162: the ACBAR and CBI data could not be used to rule out radio source contamination or systematic errors
163: as the source of the CBI excess.
164: Careful measurements over a broad range of frequencies and angular scales are needed to provide a definitive answer.
165:
166:
167: Current estimates of the primordial power spectrum are consistent with the predictions of slow-roll inflation for a nearly scale-invariant spectrum which may also include a small running of the spectral index.
168: Sparked by the modest evidence for negative running in the WMAP first-year data, a number of authors have investigated how existing data sets limit the allowed inflationary scenarios \citep{peiris03,mukherjee03,bridle03,leach03}.
169: Small-scale data extend the range over which the primordial power spectrum is
170: measured and can potentially yield information about the mechanism of inflation.
171:
172: This is the third and final ACBAR power spectrum release. The first
173: release in K04 analyzed two fields from the 2001 and 2002 seasons with
174: a conservative field differencing algorithm. The second ACBAR power
175: spectrum, presented by K07, added two more fields from the 2002 season
176: and implemented an improved, undifferenced Lead-Main-Trail (no-LMT)
177: analysis of the dataset. The results presented here improve on the
178: previous work in two ways. First, we include seven additional fields
179: observed in the 2005 Austral winter. These fields double the total
180: number of detector hours and substantially improve the precision of
181: the band-power estimates. In particular, the new fields were selected
182: to dramatically expand ACBAR's sky coverage in order to reduce the
183: cosmic variance contribution to the uncertainty and to improve the multipole resolution on angular
184: scales below $\ell \lesssim 1800$. This angular range covers the
185: third to fifth acoustic peaks, making it especially interesting for
186: constraining cosmological models. Second, we implement a new
187: temperature calibration based on a comparison of CMB fluctuations as
188: measured by ACBAR and the WMAP satellite~\citep{hinshaw08}. This
189: improved calibration tightens constraints on cosmological models found
190: from the combination of high-$\ell$ ACBAR band-powers with low-$\ell$
191: results from other experiments.
192:
193:
194:
195: This paper is organized as follows. In \S~\ref{sec:instrument} we
196: review the ACBAR instrument and the CMB observation program.
197: The analysis algorithm is explained
198: in \S~\ref{sec:analysis}. Section \S~\ref{sec:calib} is an
199: overview of the calibration;
200: the details of cross-calibration between
201: WMAP5 and ACBAR are discussed in Appendix A. Information on ACBAR's beams can be found in \S~\ref{sec:beam}.
202: Systematic tests and foreground contamination are discussed in
203: \S~\ref{sec:sys}.
204: We present the band-power
205: results in \S~\ref{sec:results}, including a discussion of the scientific
206: interpretation.
207: The ACBAR band-powers are combined with the results of other experiments to place
208: constraints on the parameters of cosmological models in
209: \S~\ref{sec:parameters}.
210: In \S~\ref{sec:conclusion}, we summarize the main results of this paper.
211:
212:
213: \section{The Instrument And Observations}\label{sec:instrument}
214: The ACBAR receiver was designed to take advantage of the excellent observing conditions at the South Pole to make extremely deep maps of CMB anisotropies~\citep{runyan03a}. It observes from the Viper telescope, a 2.1m off-axis Gregorian with a beam size of $5^{\prime}$ at $150\,$GHz.
215: The beams are swept across the sky at near-constant elevation by the motion of an actuated flat tertiary mirror.
216:
217: The receiver contains 16 optically active bolometers cooled to 240 mK by a three-stage He$^3$-He$^3$-He$^4$ sorption refrigerator. The results reported here are derived from the 150 GHz detectors: there were 4-150 GHz bolometers in 2001, 8 in 2002 and 2004, and 16 in 2005. The detectors were background limited at 150 GHz with a sensitivity of approximately 340 $\mu K\sqrt{s}$.
218:
219: In total, ACBAR observed 10 independent CMB fields, detailed in Table~\ref{tab:fields}. The power spectrum derived from portions of four fields, CMB2/4, CMB5, CMB6, and CMB7 was reported in K07. Since then, we have completed the analysis of six new fields observed in 2005 as well as additional observations of the original four fields. Details of the instrument configuration and performance in the 2001 and 2002 seasons are given in \citet{runyan03a}, while additional details of the CMB observations, data reduction procedures, and beam maps can be found in K04 and K07.
220:
221: %\clearpage
222: %\onecolumn
223: \input{tab1}
224:
225: %\twocolumn
226: %\clearpage
227:
228: \begin{figure*}[ht!]
229: \resizebox{\hsize}{!}{
230: \includegraphics[angle=90]{f1.eps}}
231:
232:
233: \caption{The ACBAR fields overlaid on the IRAS dust map. The position of each field is plotted and labeled with the field name.
234: The color coding indicates the year in which the observations occurred: red $\equiv$ 2001, orange $\equiv$ 2002, and yellow $\equiv$ 2005. The bulk of the 2005 season was targeted at large, comparatively shallow fields, increasing the total sky coverage by a factor of six.
235: The fields are plotted on top of the 100 $\mu$m IRAS dust map~\citep{schlegel98}. Each field
236: lies within the ``Southern hole'', a region of low dust emission visible from the South Pole. The CMB8 field (lower right corner) was targeted at the deep region of the B03 experiment as an alternative calibration path to the WMAP cross-calibration used for the results presented here.
237: }
238: \label{fig:acbarfields}
239: \end{figure*}
240: %\clearpage
241:
242: \section{Un-differenced Power Spectrum Analysis}\label{sec:analysis}
243:
244: Following the conventions of the previous data releases,
245: the band-powers ${\bf q}$ are reported in units of $\mu K^2$,
246: and are used to parameterize the power spectrum according to
247: \begin{equation}
248: \ell(\ell+1)C_{\ell}/2\pi\equiv{\cal D}_{\ell}=\sum_Bq_B\chi_{B\ell}\;,\label{dl}
249: \end{equation}
250: where $\chi_{B\ell}$ are tophat functions; $\chi_{B\ell}=1$ for $\ell \in B$, and
251: $\chi_{B\ell}=0$ for $\ell \not\in B$.
252: The ACBAR observations were carried out in a {\em lead-main-trail}
253: (LMT) pattern. Originally, the three fields were differenced according to the formula $M-(L+T)/2$ in order to remove time-dependent
254: chopper synchronous offsets. In K07, this conservative strategy was shown to be unnecessary and an un-differenced analysis algorithm was presented.
255: We continued to observe in a lead-trail or LMT pattern in 2005 in order to produce maps wider than the maximum range ($\sim 3^\circ$) of the chopping tertiary mirror. The un-differenced analysis presented in K07, and used for this paper's analysis, is outlined below with any differences in the application to the 2005 data set highlighted.
256:
257:
258: Let $d_\alpha$ be a measurement of the CMB temperature at pixel $\alpha$. The data vector can be represented as the sum of the signal, noise and chopper synchronous offsets: $d_\alpha = s_\alpha + n_\alpha +o_\alpha$. For example, although the chopping mirror moves the beams at nearly-constant elevation, the slight residual atmospheric gradient produces a chopper synchronous signal $o_\alpha$ which is an approximately quadratic function of chopper angle.
259: To remove these offsets, the data from each chopper sweep
260: are filtered with the ``corrupted mode projection'' matrix
261: ${\bf \Pi}$ to produce the cleaned
262: time stream ${\tilde{\bf d}}\equiv {\bf \Pi d}$.
263:
264:
265: The ${\bf \Pi}$ matrix projects out a third to tenth order polynomial which suppresses large angular scale chopper offsets. The order of the polynomial removed depends on the amplitude of atmosphere-induced cross-channel correlations.
266: As described in K07, small angular scale offsets can be be removed by subtracting an ``average'' chopper function.
267: For 2002 data, we removed a chopper synchronous offset from each data strip where the amplitude of the offset at each
268: sample in the strip is free to vary quadratically with elevation in the map.
269: The large fields observed in 2005 have up to four times the dec range of the fields observed in 2001 and 2002 ($\sim10^\circ$ vs. $\sim2.5^\circ$).
270: For 2005 data, we allow the offset to vary from a third to fifth order polynomial depending on the extent of the map in declination.
271: A zeroth order polynomial in elevation removes the average chopper function and the higher order terms effectively act as a high-pass filter on changes in the offset as a function of time or elevation.
272: This anisotropic filtering removes offset-corrupted modes while preserving most of the uncorrupted modes for the power spectrum analysis.
273: The loss of information at high-$\ell$ is small; the removed modes account for only a few percent of the total
274: degrees of freedom of the data.
275:
276:
277: The corrupted mode projection matrix $\Pi$
278: can be represented as the product of two matrices, $\Pi\equiv\Pi_2 \Pi_1$.
279: The operator $\Pi_1$ is the original $\Pi$ matrix referenced in K04 which
280: adaptively removes polynomial modes in RA.
281: The additional operator $\Pi_2$ removes modes in dec independently
282: for each of the lead, main, and trail fields and can be further decomposed into the product $\Pi_2 = \Pi_2^{Poly} \Pi_2^{LPF}$. The operator $\Pi_2^{Poly}$ performs the aforementioned polynomial projection in dec to remove small-scale chopper offsets. The second operator $\Pi_2^{LPF}$ imposes a low-pass filter (LPF) $\ell<3200$ on each dec strip. The dec strips are perpendicular to the scan direction; the timestreams have always had a LPF applied in the scan direction. The pixelation used when estimating the power spectrum is too large to resolve all of the noise power (at $\ell$ up to 10,800), causing out-of-band noise to be aliased into the signal band ($\ell<3000$) if a LPF is not applied. Eliminating this high-frequency noise reduces the contribution of instrumental noise to the reported band-powers.
283:
284:
285: Using the pointing model, the cleaned timestreams are coadded to create a map:
286: $$
287: {\bf \Delta}={\bf L}{\bf d}.
288: $$
289: The noise covariance matrix of the map can be represented as
290: $$
291: {\bf C}_N={\bf L}\langle {\bf n n}^t \rangle {\bf L}^t.
292: $$
293: where ${\bf n}$ is the timestream noise. The noise matrix is diagonalized as part of applying a high signal-to-noise transformation to the data. Eliminating modes with insignificant information content reduces the computational requirements of later steps in the analysis.
294:
295: In order to apply the iterative quadratic band-power estimator, we need to know the partial derivative $\frac{\partial {\bf C}_T}{\partial q_B}$ of the theory covariance matrix ${\bf C}_T$ with respect to each band-power $q_B$. The theory matrix can be calculated by considering the effects of the filtering on the raw sky signal. The signal timestream $s_\alpha$ is the convolution of the true temperature map ${\mathfrak T}({\bf r})$ with the instrumental beam function $B_\alpha({\bf r})$
296: $$s_\alpha = \int d^2r {\mathfrak T}({\bf r}) B_\alpha({\bf r}).$$
297:
298: The signal component of the coadded map will be ${\bf \Delta^{sig}}={\bf L}{\bf s}$ or
299:
300: $$
301: \Delta_i^{sig}=\int d^2r F_i({\bf r}) {\mathfrak T}({\bf r}),\label{tsig}
302: $$
303: where we have defined the pixel-beam filter function $F_i$
304: $$
305: F_i({\bf r})=\sum_\alpha L_{i\alpha}B_\alpha({\bf r}).
306: $$
307:
308: The theory covariance matrix can be calculated in the flat sky case to be
309: $$
310: C_{T\{ij\}} \equiv
311: \langle \Delta_i\Delta_j\rangle^{sig}=
312: \int\!\!\int d^2r d^2r' F_i({\bf r})F_j({\bf r}')
313: \langle {\mathfrak T}({\bf r}){\mathfrak T}({\bf r}')\rangle
314: $$
315: $$
316: =\int\!\!\int d^2r d^2r' F_i({\bf r})F_j({\bf r}') \int \frac{d^2\ell}{(2\pi)^2}
317: C_{\ell} \cdot e^{i{\bf l}\cdot ({\bf r}-{\bf r}')}
318: $$
319: \begin{equation}
320: =\int \frac{d^2\ell}{(2\pi)^2}C_{\ell} \cdot {\tilde F}_i^*({\bf l}){\tilde F}_j({\bf l}),
321: \label{ct}
322: \end{equation}
323: where ${\tilde F}_i({\bf l})$ is the Fourier transform of the pixel-beam filter function. The partial derivative of the theory matrix can be calculated in a straightforward manner from equations \ref{dl} and \ref{ct}.
324:
325: This algorithm does not require the instrument beams to remain constant.
326: The actual ACBAR beam sizes vary slightly with chopper angle \citep{runyan03a}. The measured beam variations can be fit to a semi-analytic function as described in K04 to create a more accurate representation of the true beam shape across the map. We use the corrected beam sizes when removing point sources. In K04 and K07, we found that the differences in the power spectra from using the map-averaged beam or exact beam for each pixel were negligible. For the band-powers reported in Table \ref{tab:bands}, an averaged beam is used for the entire map.
327:
328: As in K07, we calculate the full two dimensional
329: noise correlation matrix directly from the time stream data without using
330: Fourier transforms.
331: All the numerical calculations are performed on the
332: National Energy Research Scientific Computing Center (NERSC) IBM SP RS/6000.
333: The evaluation of the Fourier transform of $F_i({\bf r})$ is the most computationally expensive step of this analysis. We use an iterative quadratic estimator to find the maximum likelihood band-powers \citep{bond98}.
334: The resulting band-powers are presented in Table \ref{tab:bands} and Figure \ref{fig:acbar}.
335:
336:
337:
338:
339: \section{Calibration}\label{sec:calib}
340:
341:
342: We derive the absolute calibration of ACBAR by directly comparing the 2005 ACBAR maps to the WMAP5 V and W-band temperature maps \citep{hinshaw08}. We pass the WMAP5 maps through a simulated version of the ACBAR pipeline to ensure equivalent filtering and cross-spectra are calculated for each field. The ratios of the cross-spectra are used to measure the relative calibration after being corrected for the respective instrumental beam functions. We had initially applied this calibration scheme to the WMAP3 maps. Transitioning to the WMAP5 dataset lowered the calibration by 1.4\% in CMB temperature units and slightly reduced the overall uncertainty. The ACBAR band-powers are unchanged except for this calibration factor. Results for ACBAR's six largest fields (approximately $600\,{\rm deg}^2$ in area) are combined to achieve a calibration accuracy of 1.97\% for the 2005 data.
343:
344:
345: The 2005 calibration is transfered to 2001 and 2002 through a comparison of power spectra for overlapping regions observed by ACBAR in each year. The CMB5 field is used to extend the calibration of the 2005 season to the 2002 data. The CMB5 calibration is carried to other fields observed in 2002 by daily observations of the flux of RCW38. The calibration of the CMB4 field (observed in 2002) then is transfered to the 70\% overlapping CMB2 field (observed in 2001). According to the new calibration, results from the RCW38-based calibration used for the 2002 data in K07 need to be multipled by $0.959 \pm 0.032$.
346: Including the year-to-year calibration uncertainty, the final calibration has an uncertainty of 2.05\% in CMB temperature units (4.1\% in power). Additional details of this procedure are discussed in Appendix \ref{app:calib}.
347:
348:
349:
350:
351: \section{Beam Determination}\label{sec:beam}
352:
353: The beams are well-described by a symmetric Gaussian, with their main-lobe FWHM determined to $2.6\%$ by continuous measurements of the images of bright quasars located in the CMB fields. The beam sidelobes were measured to the level of 30 dB with observations of Venus made in 2002. Venus is extremely bright at millimeter wavelengths and with a diameter of $\lesssim 1^{\prime}$, is much smaller than ACBAR's beam size. However, there are extended periods during which ACBAR was unable to observe Venus. ACBAR observed RCW38, a bright HII region in the galactic plane, every day. We compare deep, coadded observations of RCW38 to constrain the temporal variability of the beam sidelobes when Venus was unavailable. The complex structure surrounding RCW38 makes it difficult to directly recover the beam shape $B({\bf r})$. Instead, we monitor ratios of the beam-smoothed RCW38 maps $ \int d^2r S^{RCW38}({\bf r}) B({\bf r})$. Any observed differences in the maps would indicate a change to the instrumental beam function as the morphology of RCW38's emission $S^{RCW38}$ is expected to be constant. We set an upper limit on the possible temporal variations in the map and use this to constrain temporal variations in the beam function. The estimated band-power uncertainty from the beam function is comparable to the overall calibration uncertainty and is plotted in Figure \ref{fig:blerr}.
354:
355: %\clearpage
356: \begin{figure*}[htp]
357: \begin{center}
358: \plotone{f2.eps}
359:
360: \caption{ACBAR Beam Uncertainty and Beam Function. {\it Solid line \& left axis}: The 1$\sigma$ envelope for uncertainty in the ACBAR beam function $B_\ell$. The increasing uncertainty above $\ell=1000$ reflects the 2.6\% uncertainty in the fitted Gaussian FWHMs. The behavior at $\ell<1000$ is a combination of the uncertainty in the measured sidelobes and the calibration method `pinning' the transfer function for $\ell \in [256,512]$. {\it Dashed line \& right axis}: The measured ACBAR beam function.
361: }\label{fig:blerr}
362: \end{center}
363: \end{figure*}
364: %\clearpage
365:
366: \section{Systematic Uncertainties and Foregrounds}\label{sec:sys}
367:
368:
369: \subsection{Jackknife Tests}\label{subsec:jackknife}
370: We performed a series of tests to search for and constrain
371: potential systematic errors in the power spectrum results. As described in K04, the data can be divided into two halves based on whether the chopping mirror is moving to the left or right. The ``left minus right'' jackknife is a sensitive test for errors in the transfer function correction, microphonic vibrations excited by the chopper motion, or the effects of wind direction. Maps with bright sources such as RCW38 can provide particularly sensitive tests of the transfer function (see \cite{runyan03a} for a description of ACBAR's transfer functions).
372: Similarly, the data can be split based on the time that the observation occurred. A non-zero signal could be produced in the ``first half minus second half'' jackknife by variation in the calibration, pointing, beam and sidelobe,
373: or any other time dependent variations in the instrument. In addition, the band-powers of each jackknife constrain
374: the mis-estimation of noise during that period.
375:
376: %\clearpage
377: \begin{figure*}[htp]
378: \centering
379: \includegraphics[width=3.5in]{f3.eps}
380:
381: \caption{Systematic tests performed on the ACBAR data. {\em Top}: Power spectrum ({\em red triangle}) for differenced maps from the first half of the season and second half of the season for each field, compared to the results of Monte Carlo simulations ({\em error bars}).
382: {\em Middle}: Power spectrum ({\em blue star}) derived from difference maps of the left- and right-going chopper sweeps for all ten fields.
383: {\em Bottom}: The undifferenced band-powers from Table \ref{tab:bands} ({\em black diamonds}) compared to both jackknife power spectra: the left-right jackknife ({\em blue star}) and first half-second half jackknife ({\em red triangle}).
384: }\label{fig:sys}
385: \end{figure*}
386: %\clearpage
387:
388: We applied the left-right jackknife to the 2005 CMB data and
389: found the band-powers were inconsistent with zero at 2.5$\sigma$ at high-$\ell$ ($\ell > 2100$). We reran a set of left-right jacknives dropping individual channels, and found that two channels stood out. With both channels excluded, the discrepancy in the left-right jackknife band-powers disappeared. We were unable to find evidence for unusual microphonic lines or transfer functions in the two problematic channels, but hypothesize that these two channels have subtle microphonic response in the signal bandwidth that are detectable only in a deep integration. Both channels are excluded from the 2005 data for all results reported in this paper.
390:
391:
392: We apply the first-second half jackknife test to the joint CMB power spectrum with the exclusion of the bad channels from the 2005 data.
393: The power spectrum of the chronologically differenced maps is compared to the band-powers of a set of Monte Carlo realizations of simulated difference maps in order to account for a number of effects such as the small filtering differences due to different scan patterns and the temporal uncertainty in the beam sidelobes (see \S\ref{sec:beam}). We find that the jackknife band-powers are consistent with the predictions of the Monte Carlo above $\ell = 400$. There is a 4$\sigma$ residual of $\sim$15 $\mu$K$^2$ in the first bin. Because the combined statistical and cosmic variance uncertainty in this bin is a factor of six larger, we assume that the band-power estimate will not be significantly biased.
394:
395:
396: We also perform the left-right jackknife on the joint CMB power spectrum found from the complete data set. The results are consistent with zero for $\ell > 900$. Statistically, the probability to exceed the measured $\chi^2$ for $\ell > 900$ is 15\%. The results are inconsistent with zero at a very low ($\sim$4 $\mu K^2$) level (Fig. \ref{fig:sys}) on larger angular scales. This residual could be due to a small noise mis-estimate at low-$\ell$, possibly caused by neglected atmospheric correlations. The jackknife failure of $\sim$4 $\mu K^2$ is much smaller than the band-power uncertainties (90 - 300 $\mu K^2$) in these $\ell$-bins which are dominated by cosmic variance. The first-second half jackknife is insensitive to discrepancies of this magnitude due to the greater uncertainties introduced by small pointing and filtering
397: differences.
398: We conclude that the complete ACBAR data set shows no significant residuals in the jackknife tests and we
399: expect no significant systematic contamination of the resulting power spectrum.
400:
401:
402:
403:
404: \subsection{Foregrounds}\label{subsec:foregrounds}
405:
406: The potential contribution of foreground emission must be considered in the interpretation of CMB temperature anisotropies. There are three foregrounds with significant emission at $150\,$GHz on the relevant angular scales: galactic dust, extragalactic radio sources, and dusty proto-galaxies. As an effectively single-frequency instrument, ACBAR depends on data from other experiments to construct and constrain foreground models. We use the methodology described in K04 to remove templates for radio sources and dust emission from the CMB maps without making assumptions about their flux.
407: The contribution from dusty protogalaxies is less certain; however, we do not expect the combined residual foreground
408: emission to significantly impact the power spectrum for $\ell < 2400$.
409:
410:
411: We remove modes from the CMB maps corresponding to radio sources in the $4.85\,$GHz Parkes-MIT-NRAO (PMN) survey \citep{wright94}.
412: Of the 1601 PMN sources with a flux greater than $40\,$ mJy that lie in the ACBAR fields, we detect 37 sources including the guiding quasars at greater than 3$\sigma$ with the application of an optimal matched filter. There are less than 2.2 false detections expected with this detection threshold. The measurement errors are estimated through sampling the distribution of pixels in a set of 100 Monte Carlo realizations of the CMB+noise for each field. Table \ref{tab:pmnsources} lists the parameters of the detected PMN sources.
413: Except for the few bright sources detected at $150\,$GHz, removing the PMN point sources does not
414: significantly affect the band-powers.
415:
416:
417: It is possible that faint radio sources, undetected at $150\,$GHz, could contribute to the observed band-powers. We parameterize the contribution as
418: \begin{equation}\label{eq:src}
419: {\cal D}^{\rm src}_\ell = q_{\rm src}\left(\frac{\ell}{2600}\right)^2 \, \mu K^2\,
420: \end{equation}
421: which is appropriate for unclustered point sources.
422: We compare the $150\,$GHz ACBAR point source number counts to the model in \citet{white04} based on WMAP Q-band data, $$\frac{dN}{dS_\nu} = \frac{80~deg^{-2}}{1~mJy} \left(\frac{S_\nu}{1~mJy}\right)^{-2.3},$$ to constrain the residual power contribution at $150\,$GHz.
423: We can use this model to estimate the residual band-power contribution due to sources too faint to be included
424: in the PMN catalog.
425:
426: Following the convention in that work, the spectral dependence of the fluxes is parameterized as $S_\nu \propto \nu^\beta$.
427: The number of PMN sources detected at $150\,$GHz in a logarithmic flux bin, $n^{obs}_B$, are compared to the predicted number counts from the same population of sources with a given $\beta$, $n^{(\beta)}_B + n^{noise}_B$. Here, $n^{(\beta)}_B$ is the modeled number counts and $n^{noise}_B$ is the expected number of false detections due to ACBAR's measurement uncertainty. The number counts are assumed to follow a Poisson distribution. Sources with estimated measurement errors greater than 140 mJy in the ACBAR maps are cut to reduce the $n^{noise}_B$ term. The modeled number counts $n^{(\beta)}_B$ are scaled by $(N_{tot}-N_{cut})/N_{tot}$ to compensate. All other sources with measured amplitudes greater than 350 mJy at $150\,$GHz are included in the calculation without consideration of the signal-to-noise.
428: If the sources found by ACBAR at $150\,$GHz are the same population found by WMAP, this implies a uniform spectral index of $\beta = 0.14\pm 0.15$. However, this small sample selected for high flux at $150\,$GHz is heavily biased toward sources with flat or rising spectra. We increase ACBAR's sensitivity to dimmer sources by binning all sources within a given PMN flux range, and look at the ratio of the average flux at $150\,$GHz to the average flux at 4.85 GHz within each bin. We find the ratio ($S_{150}/S_{4.85}$) increases with PMN flux from 0.07 for sources below 400 mJy to 0.41 for sources above 1600~mJy at 4.85~GHz.
429: This implies that the sources in the PMN catalog have a flux-dependent spectral index where dimmer objects typically have a more steeply falling spectrum.
430: The band-power contribution of the low-flux sources depends sensitively on the extrapolation of the 40~mJy flux cutoff in the $4.85\,$GHz PMN catalog to $150\,$GHz. Based on the observed flux ratios for PMN sources with $S_{4.85} < 400$ mJy, we conservatively assume $S_{150}/S_{4.85} = 0.1$ for a flux cutoff at $4\,$mJy at $150\,$GHz. This flux ratio corresponds to a spectral index of $\beta = -0.67$, well below $\beta = 0.14\pm 0.15$ estimated from the ACBAR detected sources and the WMAP Q-band source model.
431: Estimating the residual radio source band-power contribution at $150\,$GHz with a flux cutoff of $4\,$mJy gives $q_{\rm src}^{radio} = 2.2$. At this level, the residual contribution from radio sources will be negligible in the ACBAR data.
432:
433:
434:
435: The ACBAR fields are positioned in the ``Southern Hole,'' a region of exceptionally low Galactic dust emission (Figure~\ref{fig:acbarfields}). \citet{finkbeiner99} (FDS99) constructed a multi-component dust model that predicts thermal emission at CMB frequencies from the combined observations of IRAS, COBE/DIRBE, and COBE/FIRAS. Taking into account the ACBAR filtering, the FDS99 model\footnote{We use the default model 8 of FDS99.} predicts a RMS dust signal at the few $\mu K$ level primarily on large angular scales. The ACBAR maps can be decomposed as the sum of the CMB and dust signals $T_{CMB}+\xi T_{FDS}$. The dust amplitude parameter $\xi$ is predicted to equal unity by the FSD99 model. The ACBAR maps are cross-correlated with the dust templates $T_{FDS}$ to calculate the amplitude in each field. The errors are estimated by applying the same procedure to 100 CMB+noise map realizations for each field. The uncertainty in $\xi$ is dominated by CMB fluctuations. The best-fit amplitude from combining all the fields is $\xi = 0.1 \pm 0.5$. The estimated amplitudes of the individual fields are shown in Figure \ref{fig:dustamp}. The reduced $\chi^2$ of the measured amplitudes $\xi$s of the eight fields analyzed is 0.75 for the no-dust assumption of $\langle \xi \rangle = 0$ and increases to $\chi^2 = 1.12$ for the FDS99 model amplitude of $\langle \xi \rangle = 1$. Therefore, the ACBAR data slightly favor a lower amplitude than predicted by the FDS99 model. The dust signal is not detectable in any of the ACBAR fields, and removing the dust template has a negligible impact on the measured power spectrum.
436:
437: Dusty IR galaxies are the third and least constrained foreground in the ACBAR fields. This population of high-redshift, star-forming galaxies has been studied by several experiments at higher frequencies \citep{coppin06,laurent05,maloney05,greve04}. However, as discussed in K07, extrapolating the expected signal to $150\,$GHz remains highly uncertain, and there remain significant uncertainties in the number counts $\frac{dN}{dS}$ and spatial clustering of the sources. The frequency dependence can be empirically determined by comparing the measured number counts in overlapping fields observed at different frequencies. This comparison has been done with MAMBO (1.2 mm) and SCUBA (850 $\mu$m), leading to a spectral dependence of $S_\nu\propto \nu^{2.65}$ \citep{greve04}. A second method of estimating the index used nearby galaxy data to obtain $S_\nu \sim \nu^{2.6\pm 0.3}$ \citep{knox04}.
438: The uncertainty in the spectral dependence significantly affects the extrapolation of the flux of dusty galaxies to $150\,$GHz. We use estimates of the source number counts from the SHADES survey~\citep{coppin06} and Bolocam Lockman Hole Survery~\citep{maloney05}. We apply the formulas in \citet{scott99} to estimate the expected power spectrum for the source number counts, ignoring the clustering terms. In this limit, ${\cal D}_{\ell}$ will have the form in eq. \ref{eq:src}. Scaling the results to $150\,$GHz with the MAMBO/SCUBA prescription of $S_\nu\propto \nu^{2.65}$ leads to an estimated contribution of $q_{\rm src}^{dusty} = 17-29$. This range reflects the differences between the measured number counts, but does not include the uncertainty in the spectral dependence of the fluxes.
439: Combining the median index with earlier SCUBA data fit by two power laws in $S$ from
440: \citep{borys03}, we find an excess of 22 $\mu {\rm K}^2$ at $\ell = 2600$, within that range.
441: This level is only a factor of two smaller than the instrumental noise of ACBAR and might
442: influence the interpretation of high-$\ell$ band-powers.
443: We tentatively assume that contamination from dusty proto-galaxies does not significantly effect the
444: resulting power spectrum.
445: The implications of relaxing this assumption for cosmological parameter estimation are
446: explored in \S~\ref{subsec:ressourcemarg}.
447:
448: %\clearpage
449: \input{tab2}
450:
451: \begin{figure*}[ht!]
452: \epsscale{1.}
453: \plotone{f4.eps}
454: \caption{Dust emission is not detected in the ACBAR fields. Parameterizing the dust signal as $T_{CMB}+\xi T_{FDS}$, a suite of Monte Carlo realizations of maps of CMB and noise is used to estimate the uncertainty in $\xi$. We find the upper limits in each field to be consistent with the FDS99 model ($\xi = 1$), but the data somewhat favor a lower dust amplitude. The reduced $\chi^2$ of the measured amplitudes $\xi$s is 0.75 under the assumption that $\langle \xi \rangle = 0$ (the dashed line). The reduced $\chi^2$ for the FDS99 model with $\langle \xi \rangle = 1$ is 1.12 (the dotted line).
455: }
456: \label{fig:dustamp}
457: \end{figure*}
458: %\clearpage
459:
460: \section{Results And Discussions}\label{sec:results}
461:
462: \subsection{Power Spectrum}\label{sec:ps}
463:
464: The power spectrum presented in Figure~\ref{fig:acbar} is produced by
465: the application of the analysis algorithm outlined in
466: \S~\ref{sec:analysis} to ACBAR data from the 2001, 2002 and 2005
467: austral winters. The resulting power spectrum is compared to the WMAP5
468: and B03 spectra in Figure~\ref{fig:acbar_ext}. The zero-curvature,
469: $\Lambda$CDM ``ACBAR+WMAP5" best fit model is shown in each figure for
470: reference. The decorrelated band-powers are tabulated in
471: Table~\ref{tab:bands}. Our choice of the decorrelation
472: transformations follows \citet{tegmark97b}. The band-powers can be
473: compared to a theoretical model using the window functions
474: \citep{knox99}. As in K04, we sample the likelihood function ${\cal
475: L}(\Delta) = \frac{1}{\sqrt{C}}e^{-(\Delta^t C^{-1} \Delta)/2}$ near
476: the maximum and fit the results with offset lognormal functions
477: \citep{bond2000}. The fit parameters ${\bf \sigma}, {\bf x}$ are
478: listed in Table~\ref{tab:bands} as well. The band-powers, likelihood
479: fit parameters, and window functions are available for download from
480: the ACBAR
481: website\footnote{http://cosmology.berkeley.edu/group/swlh/acbar/index.html}.
482:
483:
484: The ACBAR data extend the measurement of the temperature anisotropies
485: well into the damping tail with S/N $>$ 5 for $\ell \lesssim 2300$.
486: The fourth and fifth acoustic peaks are detected for the first time in
487: the ACBAR band-powers, providing additional support for the coherent
488: origin of anisotropy \citep{albrecht96}. The position of the third
489: acoustic peak is consistent with previous detections of the feature by
490: CBI \citep{readhead04}, B03 \citep{jones06}, ACBAR (K07), and QUaD
491: \citep{quad07}. The ACBAR band-powers are in excellent agreement with
492: the cosmological models constrained by observations on larger angular scales. The
493: probability to exceed the reduced $\chi^2$ between the ACBAR
494: band-powers and the WMAP3 only best-fit $\Lambda$CDM model is
495: 17\%. This probability increases to 62\% with the WMAP5 only best-fit model.
496: This serves as both a powerful confirmation of our basic cosmological model and an
497: indication of the quality of the ACBAR data set.
498:
499: %\clearpage
500: \input{tab3}
501: %\clearpage
502:
503: \begin{figure*}[ht!]
504: \resizebox{\hsize}{!}{
505: \plotone{f5.eps}}
506: \caption{The decorrelated ACBAR band-powers for the full data set. The
507: $1\sigma$ error bars are derived from the offset-lognormal fits to
508: the likelihood function. The band-powers are in excellent agreement
509: with a $\Lambda$CDM model. The damping of the anisotropies is
510: clearly seen with a S/N $>$ 4 out to $\ell=2500$. The third
511: acoustic peak (at $\ell \sim 800$), fourth acoustic peak (at $\ell
512: \sim 1100$), and fifth acoustic peak (at $\ell \sim 1400$) are
513: visible. The plotted lines are the best fits to the ACBAR and WMAP5
514: band-powers for a spatially flat, $\Lambda$CDM universe with no SZE
515: contribution. A lensed ({\it red}) and unlensed ({\it blue}) model
516: spectrum is shown for a fixed parameter set; the lensed spectrum is a significantly better fit
517: to the ACBAR data. }
518: \label{fig:acbar}
519: \end{figure*}
520: %\clearpage
521:
522: \begin{figure*}[ht!]
523: \resizebox{\hsize}{!}{
524: \plotone{f6.eps}}
525: \caption{\protect\small
526: The ACBAR band-powers plotted with those from WMAP5 \citep{hinshaw08}
527: and the 2003 flight of BOOMERANG \citep{jones06}. The three
528: experiments show excellent agreement in the region where they overlap.}
529: \label{fig:acbar_ext}
530: \end{figure*}
531: %\clearpage
532:
533: \subsection{Anisotropies at $\ell>2000$}
534: \label{sec:excess}
535:
536: Several theoretical calculations \citep{cooray00,komatsu02} and
537: hydrodynamical simulations \citep{bond02,white02} suggest that the
538: thermal Sunyaev-Zel'dovich effect power spectrum will exceed that of
539: the primary CMB temperature anisotropies for $\ell \gtrsim 2500$ at
540: 150 GHz. The amplitude of the SZE power spectrum is closely related to
541: the amplitude of matter perturbations which is commonly parameterized
542: as $\sigma_8$; the SZE power spectrum is expected to scale as
543: $\sigma_8^7$ \citep{zhang02}. To a lesser extent, the level of the SZE
544: will also depend on details of cluster gas physics and thermal
545: history. The non-relativistic thermal SZE ($\Delta T_{SZ}$) has a
546: unique frequency dependence
547: \begin{equation}
548: \frac{\Delta T_{SZ}}{ T_{CMB}} =
549: y \left(x\frac{e^x + 1}{e^x - 1} - 4\right),\,\label{szs}
550: \end{equation}
551: where $x = \frac{h\nu}{ k T_{\rm CMB}}$ = $\nu/56.8\,$GHz. The
552: variable y is the Compton parameter and is proportional to the
553: integrated electron pressure along the line of sight. The CBI
554: extended mosaic observations \citep{readhead04} detected more power
555: above $\ell = 2000$ than is expected from primary CMB anisotropies.
556: This excess could be the first detection of the SZE power spectrum
557: \citep{mason03,readhead04,bond02}. However, there are alternative
558: explanations for the observed power ranging from an unresolved
559: population of low-flux radio sources to non-standard inflationary
560: models \citep{cooray02,voids,Bfields} that produce higher than
561: expected CMB anisotropy power at small angular scales. The frequency
562: dependence of the excess power can be exploited to help discriminate
563: between the SZE and other potential explanations.
564:
565: The ACBAR band-powers reported in this paper are slightly larger at
566: $\ell > 2000$ than expected for the ``ACBAR+WMAP5" best fit model. We
567: subtract the predicted band-powers at $\ell>1950$ from the measured
568: band-powers in Table~\ref{tab:bands} and find an excess of
569: $22 \pm 20 ~\mu K^2$ in a flat band-power from $1950 < \ell < 3000$. This estimate ignores the band-power contribution from dusty proto-galaxies which is expected to be comparable (see \S\ref{subsec:foregrounds}).
570: The ACBAR band-powers at $150\,$GHz can be
571: used to place constraints on frequency spectrum of the larger CBI
572: excess measured at $30\,$GHz.
573: We parameterize the excess power for $\ell > 1950$ at the two frequencies as
574: $P_{30} = \alpha P_{150}$ and sample the likelihood surface for
575: $\alpha \in [0,10]$ and $P_{150} \in [0 ,300]$ $\mu K^2$. The beam
576: uncertainty and the calibration error for both experiments is taken
577: into account by Monte Carlo techniques. The likelihood function is
578: averaged over 1000 realizations under the assumption that each of the
579: errors has a normal distribution. The resulting likelihood function for $\alpha$
580: (after $P_{150}$ is marginalized) is shown in
581: Figure~\ref{fig:excess}. From the ACBAR and CBI frequency bands, we
582: expect $\alpha=4.3$ for power originating from the SZE.
583: If the excess is due to primary
584: CMB anisotropies, we expect $\alpha = 1$. We conclude that it is more than 5
585: times as likely that the excess seen by CBI and ACBAR is caused by
586: the thermal SZE than a primordial source. We expect the
587: contribution of radio sources to CMB power to be at least a factor of
588: ten higher at 30 GHz than at 150 GHz ($\alpha \ge 10$). Because of
589: the relatively weak detection of excess power by ACBAR, flat spectrum
590: radio sources are determined to be $\sim$10\% more likely than the SZE to
591: be the source of the excess.
592: The lower level of excess power seen by ACBAR argues against the
593: the CBI excess having a primordial origin, but is consistent
594: with either the SZE or radio source foregrounds.
595: When we include the expected contribution of dusty protogalaxies to the
596: ACBAR excess band-power, the likelihood of the CBI excess being due to radio sources
597: increases with respect to thermal SZE.
598:
599:
600: %\clearpage
601: \begin{figure*}[ht!]
602: \resizebox{\hsize}{!}{
603: \plottwo{f7a.eps}{f7b.eps}}
604: \caption{ACBAR results on the high-$\ell$ anisotropies. {\em Left}:
605: The ACBAR band-powers above $\ell=1000$ plotted against the best-fit
606: ACBAR+WMAP5 model spectrum. The latest CBI results at 30 GHz and the
607: previous ACBAR results are also shown. The ACBAR band-powers for
608: $\ell > 1950$ are consistently below the reported high-$\ell$
609: CBI band-power. {\em Right}: The likelihood distribution for
610: the ratio of the ``excess" power, observed by CBI at $30\,$GHz and
611: ACBAR at $150\,$GHz. The solid line assumes that dusty proto-galaxies do not contribute to the ACBAR band-powers, while the dash-dot line includes the predicted contribution of approximately 22 $\mu {\rm K}^2$ at $\ell = 2600$ from these sources (see \S\ref{subsec:foregrounds}). With the inclusion of dusty galaxies, the likelihood peaks at a higher power ratio ($\simeq 10$).
612: The excess power for each experiment is the difference between the
613: measured and model band-powers for each experiment in a flat band
614: with $\ell>1950$.
615: The likelihood for a given power ratio is found from Monte Carlo
616: simulations of the band-powers and uncertainties.
617: The vertical dashed line represents the expected ratio
618: (4.3) for the excess being due to the SZE, while radio foregrounds
619: would correspond to a ratio of $> 10$. If the excess
620: power seen in CBI is caused by non-standard primordial processes,
621: the ratio will be unity (blackbody), indicated by the dotted line.
622: It is considerably more likely that the excess seen by
623: CBI is caused by either the thermal SZE or radio foreground
624: contamination than a primordial source. }
625: \label{fig:excess}
626: \end{figure*}
627: %\clearpage
628:
629:
630: \section{Cosmological Parameters}\label{sec:parameters}
631:
632: \subsection{Cosmological Parameters and their ``Prior'' Measures}\label{subsec:basicvars}
633:
634: In this section, we estimate cosmological parameters for a minimal
635: inflation-based, spatially-flat, tilted, gravitationally lensed,
636: $\Lambda$CDM model characterized by six parameters, and then
637: investigate models with additional parameters to test
638: extensions of the theory. For our base model, the six parameters are:
639: the physical density of baryonic and dark matter, $\Omega_bh^2$ and
640: $\Omega_ch^2$; a uniform spectral index $n_s$ and amplitude $\ln A_s$
641: of the primordial power spectrum, the optical depth to last
642: scattering, $\tau$; and $\theta$, the ratio of the sound horizon at
643: last scattering to the angular diameter distance. The primordial
644: comoving scalar curvature power spectrum is expressed as ${\cal
645: P}_s(k) = A_s (k/k_\star)^{(n_s-1)}$, where the normalization
646: (pivot-point) wavenumber is chosen to be $k_\star = 0.05\, {\rm
647: Mpc}^{-1}$. The parameter $\theta$ maps angles observed at our
648: location to comoving spatial scales at recombination; changing
649: $\theta$ shifts the entire acoustic peak/valley and damping pattern of
650: the CMB power spectra.
651: Additional parameters are derived from this basic
652: set. These include: the energy density of a cosmological constant in
653: units of the critical density, $\Omega_\Lambda$; the age of the
654: universe; the energy density of non-relativistic matter, $\Omega_m$;
655: the {\it rms} (linear) matter fluctuation level in $8h^{-1}$Mpc
656: spheres, $\sigma_8$; the redshift to reionization, $z_{re}$; and the
657: value of the present day Hubble constant, $H_0$, in units of km
658: s$^{-1}$Mpc$^{-1}$.
659:
660: Single-field models of inflation predict the existence of a
661: gravitational wave background characterized by a primordial power law
662: ${\cal P}_t \sim k^{n_t}$. We
663: characterize the strength by the tensor-to-scalar ratio $r={\cal
664: P}_t/{\cal P}_s$ evaluated at a pivot point $0.002\, {\rm
665: Mpc}^{-1}$. We relate the tilt to $r$ using the approximate
666: consistency relation $n_t \approx -r/8/(1-r/16)$. (We find little
667: difference in the parameters if we just fix $n_t$ to be zero, as has
668: often been assumed when $r$ is included, but using this relation is
669: superior since it is motivated by inflation physics.)
670:
671: A small running of the spectral index is also expected in slow-roll
672: inflation and we test for this by extending the basic $\Lambda$CDM
673: power law model to include a scale dependence of the scalar spectral tilt,
674: $d n_s/d\ln (k)$.
675:
676: We have also added non-zero curvature $\Omega_k$ to our basic six
677: parameters. The results are consistent with the flat case, but with the
678: standard geometrical degeneracy relating $\Omega_k$ and
679: $\Omega_\Lambda$ expressed through $\theta$ leading to a
680: near-degenerate tail to $\Omega_k< 0$.
681:
682: The ACBAR spectrum includes band-powers at $\ell> 2000$ where the
683: signal due to secondary CMB anisotropies associated with
684: post-recombination nonlinear effects should become significant. In
685: particular, the thermal Sunyaev-Zel'dovich effect and the contribution
686: of unresolved radio sources and dusty galaxies will ultimately
687: dominate over the primary anisotropy damping tail; the only question
688: is at what multipole crossover occurs. Without including such
689: secondary effects, the parameters we derive from the primary
690: anisotropy power spectrum could be biased. To account for this, we
691: have added to the primary anisotropy power spectrum (1) the SZE
692: template power spectrum ${\hat {\cal D}}_\ell^{\rm SZ}$ used in K07
693: and \citep{bond05,goldstein03} which was derived from cosmological
694: hydrodynamics simulations and (2) an unclustered point source
695: template, as in eq.~\ref{eq:src}. Each template is scaled by an
696: overall amplitude parameter, $q_{\rm SZ}$ and $q_{\rm src}$, which we
697: assume have uniform prior measures with a range much larger than
698: required by the ACBAR data. The white-noise form for ${\cal D}^{\rm
699: src}_\ell$ given by eq.~\ref{eq:src}, is appropriate for the
700: statistically-averaged power of a distribution of unclustered sources.
701: The clustering of radio sources is not a large effect, but we do
702: expect sub-mm sources associated with dusty galaxies at lower flux
703: levels to be clustered. As mentioned in \S~\ref{subsec:foregrounds},
704: in spite of great strides in sub-mm observations in recent years,
705: significant uncertainties remain in source fluxes and clustering at
706: $150\,$ GHz. Theoretical models suggest both will be important for a
707: complete treatment, but the approximation adopted here should be
708: sufficient for the ACBAR data set. In the parameter tables below, we
709: show results including these secondary templates. We find the basic
710: parameter central values and uncertainties change little whether we
711: marginalize over either of the two template amplitudes or set them
712: both to zero.
713:
714: The parameter constraints are obtained using a Monte Carlo Markov
715: Chain (MCMC) sampling of the multi-dimensional likelihood as a
716: function of model parameters. The pipeline is based on the publicly
717: available {\textsc CosmoMC}\footnote{http://cosmologist.info/cosmomc}
718: package \citep{Lewis:2002ah}. CMB angular power spectra and matter
719: power spectra are computed using the {\textsc CAMB} code
720: \citep{lewis00}. As described in Section~\ref{sec:results}, we
721: approximate the full non-Gaussian band-power likelihoods with an
722: offset lognormal distribution \citep{bond2000}. Our standard {\textsc
723: CosmoMC} results include the effects of weak gravitational lensing
724: on the CMB \citep{seljak96,lewis00}. Lensing effects in the
725: temperature spectrum are expected to become significant at scales
726: $\ell > 1000$, hence it is important to include this effect when
727: interpreting the ACBAR results. The major effect of lensing is a
728: scale-dependent smoothing of the angular power spectrum which
729: diminishes the peaks and valleys of the spectrum. Inclusion of
730: lensing in the model improves the fit to the data for all experiment
731: combinations.
732:
733: The typical computation consists of eight separate chains, each having
734: different initial, random parameter choices. The chains are run until
735: the largest eigenvalue of the Gelman-Rubin test is smaller than 0.01
736: after accounting for burn-in. Uniform priors with very broad
737: distributions are assumed for the basic parameters. The standard run
738: also includes a weak prior on the Hubble constant ($45 < H_0 < 90$
739: km\, s$^{-1}$\, Mpc$^{-1}$) and on the age of the universe ($>10$
740: Gyrs), but these have negligible effects. We also investigate the
741: influence of adding Large Scale Structure (LSS) data from the 2 degree
742: Field Galaxy Redshift Survey (2dFGRS) \citep{cole05} and the Sloan
743: Digital Sky Survey (SDSS) \citep{tegmark06}. When including the LSS
744: data, we use only the band-powers for length scales larger than $k
745: \sim 0.1 ~h~$Mpc$^{-1}$ to avoid non-linear clustering and
746: scale-dependent galaxy biasing effects. We marginalize over a
747: parameter $b^2_g$ which describes the (linear) biasing of the
748: galaxy-galaxy power spectrum for $L_\star$ galaxies relative to the
749: underlying mass density power spectrum. We adopt a Gaussian prior on
750: $b^2_g$ centered around $b_g=1.0$ with a very large width. We have
751: also tried restricting the width to $\delta b_g = 0.3$, but the
752: cosmic parameters are insensitive to this width.
753:
754:
755: \subsection {Base Parameter Results}\label{sec:basic}
756:
757: The results for the basic spatially flat tilted $\Lambda$CDM
758: parameters are presented in Table~\ref{tab:basic}. The confidence
759: limits are obtained by marginalizing the multi-dimensional likelihoods
760: down to one dimension. The median value is obtained by finding the
761: 50\% integral of the resulting likelihood function while the lower and
762: upper error limits are obtained by finding the 16\% and 84\%
763: integrals, respectively. The CMBall data combination includes the
764: ACBAR results presented here and other CMB data sets with published
765: band-powers and window functions: the WMAP 5 year angular power
766: spectra \cite{nolta08}, and for comparison the WMAP 3
767: year spectra \cite{hinshaw06}; the CBI extended mosaic results \citep{readhead04} and
768: polarization results \citep{Readhead04b,Sievers05}, combined in the
769: manner described in \citet{Sievers05};\footnote{We exclude the band-powers below $\ell=600$ from the CBI extended mosaic results to
770: reduce the correlation with the TT band-powers of the CBI
771: polarization dataset which influence the sample-dominated end of the
772: spectrum.} the DASI two year results \citep{halverson02}; the DASI
773: EE and TE band-powers \citep{Leitch04}; the VSA final results
774: \citep{dickinson04}; the MAXIMA 1998 flight results \citep{hanany00};
775: and the TT, TE, and EE results from the BOOMERANG 2003 flight
776: \citep{jones06, piacentini06, montroy06}. Only $\ell > 350$
777: band-powers are included for BOOMERANG because of overlap with WMAP
778: (although inclusion of the lower $\ell$ results leaves the parameter
779: results essentially unchanged). While ACBAR and BOOMERANG are both
780: calibrated through WMAP, this is a small contribution to the total
781: uncertainty in the ACBAR calibration and we treat the calibration
782: uncertainties as independent in our parameter analysis. Although the
783: DASI, CBI, and BOOMERANG 2003 EE and TE results for high-$\ell$
784: polarization are included, they have little impact on the values of
785: the parameters we obtain.
786:
787: The latest WMAP likelihood code found at http://lambda.gsfc.nasa.gov/
788: has been used in our analyses. When this ACBAR paper was submitted,
789: all parameter analyses were done using WMAP3 \citep{hinshaw06};
790: the new WMAP5 results came out a few months later.
791: The WMAP5 data allowed for an improved cross calibration of the ACBAR
792: and WMAP band-powers and resolved the ($<$ 1-$\sigma$) parameter tensions
793: that existed between the WMAP3 and ACBAR results.
794: We note that the WMAP5 team's parameter analysis \citep{dunkley08}
795: made use of the ACBAR band-powers with the $1.4\%$ higher temperature calibration
796: from the WMAP3 and ACBAR cross-calibration.
797: The marginalized one-dimensional likelihood
798: distributions for the basic parameter set we obtain are shown in
799: Figure~\ref{fig:basic}. Note the contrast between the parameter
800: determinations for WMAP3 and WMAP5. The primary improvement of WMAP5
801: over WMAP3 was a better understanding of the beam, which resulted in an
802: improved measurement of the third acoustic peak.
803: Other improvements in WMAP5 were updated point-source
804: correction, stimulated by \citep{huffenberger06}, and foreground
805: marginalization on large angular scales.
806:
807: %\clearpage
808: \begin{figure*}[th!]
809: \resizebox{\hsize}{!}{
810: \plotone{f8.eps}}
811: \caption{Basic parameter marginalized 1-dimensional likelihood
812: distributions for the following data combinations; WMAP3-only
813: (green, dashed), WMAP5-only (black, solid), ACBAR + WMAP5 (red,
814: long-dashed), CMBall (blue, dot-dashed). All runs include
815: lensing.}
816: \label{fig:basic}
817: \end{figure*}
818: %\clearpage
819: \begin{figure*}[th!]
820: \resizebox{\hsize}{!}{
821: \plotone{f9.eps}}
822: \caption{68\% and 95\% 2D marginalized contours in $\theta$ and
823: $\Omega_m$ for a number of data combinations. The results for $\theta$
824: from WMAP are largely driven by the determination of the third peak.
825: The value preferred by the WMAP5 data is slightly higher than that from WMAP3,
826: and more consistent with that found when ACBAR is added to either WMAP data
827: set.}
828: \label{fig:theta}
829: \end{figure*}
830: %\clearpage
831: The addition of LSS data has little impact on the mean values and
832: errors in the cosmic parameters. The largest shift is $<$ 1-$\sigma$ in
833: $\Omega_ch^2$, from $0.111^{+0.005}_{-0.005}$ to
834: $0.107^{+0.004}_{-0.004}$. Our LSS results only include information on
835: the shape of the density power spectrum, not its overall amplitude
836: since we marginalize over the galaxy bias factor. If results from weak
837: lensing are included in the LSS data, then there is a slight increase
838: in $\sigma_8$, but it depends somewhat on which
839: lensing results are included. The effect is to slightly increase the
840: CMB+LSS result and improve the consistency with the CMBall results.
841: The third peak is well determined by both ACBAR and WMAP5, and this defines
842: the dark matter density;
843: the inclusion of the LSS data does not significantly improve the constraints.
844: We have also found that the results do not change significantly from
845: CMBall+LSS when SN1a data are included (in this case from the
846: \citet{riess04} gold set), so we have not included a separate
847: column. Including SN1a data would be crucial if we were attempting to constrain
848: the equation of state of dark energy.
849:
850: All parameter results listed in Tables~\ref{tab:basic} and
851: \ref{tab:nrun} include the effects of weak lensing of the CMB on the
852: resulting power spectrum. For every case including CMB lensing, we
853: have performed an identical calculation neglecting the effects of
854: lensing. Including lensing improves the fit of the model to the
855: observed band-powers compared for all data combinations.
856: This can be quantified by the log-ratio of the lensed to
857: no-lensed Bayesian evidence, $\Delta \ln {\cal E}=\ln[P({\rm
858: lens}|data,theory)/P({\rm no-lens}|data,theory)]$. The evidence
859: $P({\rm lens}|data, theory)$ is an integral of the product of the {\it
860: a priori} probability (the parameters' measure) and the likelihood
861: of data given those parameters; it appears in the denominator in the
862: Bayesian chain to ensure the {\it a posteriori} probability has unity
863: normalization. The resulting number is a conditional probability given
864: the data and the assumptions about the parameters.
865: The parameters and their measures are
866: exactly the same, so the ratio is a robust indicator of preference. For
867: WMAP5 alone it is $\Delta \ln {\cal E}=2.04$; it increases to 2.89
868: with ACBAR included; and is 2.63 for CMBall. Naively relating this to
869: a Gaussian translates to a significance of $\sim 2.3 \sigma$ for CMBall. From
870: Fig.~\ref{fig:acbar}, it is clear that the difference in the
871: power spectra of lensing and no-lensing is small; for this plot, the
872: best-fit parameters from the lensed analysis were fixed and used to
873: compute a lensed and unlensed spectrum.
874: The difference, $\Delta
875: C_\ell^{\rm lens} \equiv C_\ell^{\rm lens} - C_\ell^{\rm no-lens}$, is
876: explicitly shown in the inset of Figure~\ref{fig:lensfit}. We find
877: only small shifts in the median value of the cosmic parameters when
878: lensing is included; e.g., for the ACBAR+WMAP5 data combination, we
879: find $\sigma_8=0.79^{+0.03}_{-0.03} \rightarrow 0.80^{+0.03}_{-0.03}$
880: and $\Omega_ch^2 = 0.109^{+0.006}_{-0.006}\rightarrow
881: 0.111^{+0.006}_{-0.006}$ when going from non-lensed to lensed models
882: respectively.
883:
884: %\clearpage
885: \input{tab4}
886: %\clearpage
887:
888: We now test whether the strength of
889: the lensing modification is consistent with expectations for lensing,
890: by multiplying the lensing template $\Delta
891: C_\ell^{\rm lens}$, which varies with cosmic parameters, by a
892: variable strength $q_{\rm lens}$:
893: \begin{equation} C_\ell^{\rm lens} = C_\ell^{\rm no-lens} + q_{\rm
894: lens} \Delta C_\ell^{\rm lens} \, .
895: \end{equation}
896: The normalization is such that $q_{\rm lens}=1$ gives the normal
897: lensed CMB spectrum, while $q_{\rm lens}=0$ gives the no-lensing case.
898: An accurate determination of this subtle effect requires
899: highly accurate window functions for the bands. We use a flat prior
900: probability for $q_{\rm lens}$, allowing it to vary from 0 to 10.
901: With WMAP5 alone, we obtain $q_{\rm lens}=1.34^{+0.27 (+1.51)}_{-0.26
902: (-0.85)}$; WMAP5+ACBAR gives $q_{\rm lens}=1.23^{+0.21
903: (+0.83)}_{-0.23 (-0.76)}$; CMBall gives $q_{\rm lens}=1.21^{+0.24
904: (+0.82)}_{-0.24 (-0.76)}$. We have also listed the 2-$\sigma$ errors
905: (in brackets), which are far from twice the 1-sigma values, reflecting
906: the highly non-Gaussian nature of the marginalized likelihoods evident
907: in the figure. Although we emphasize that $q_{\rm lens}$ is not in any
908: sense an independent parameter, it does illustrate that lensing
909: of the expected strength is preferred.\footnote{\cite{slosar08} also undertook a lensing analysis of
910: ACBAR temperature power spectrum, but, instead of $q_{\rm lens}$ as
911: defined here, they used a multiplier $A_{\rm L}$ of the lensing
912: potential power spectrum, defined to be unity for normal lensing;
913: they found $A_{\rm L}=3.0^{+0.9}_{-0.9}$ for WMAP5+ACBAR. Repeating our
914: analysis with this parameterization, we find lower values, $A_{\rm
915: L}=1.60^{+0.55 (+1.79) }_{-0.26 (-0.99)}$. With the highest 3 ACBAR bins cut out,
916: where secondary effects might have an impact on the determination of
917: the lensing strength parameter, the results are essentially unchanged. }
918: The significance of the detection is somewhat less than the WMAP3/NVSS/SDSS
919: cross-correlation results of \citet{smith07} and \citet{hirata08}. In
920: Figure~\ref{fig:lensfit}, we show the marginalized distribution for
921: the $q_{\rm lens}$ parameter using various combinations of CMB data.
922:
923: We have also run a limited set of non-flat model chains. The models in
924: this case do not include the effect of weak lensing and we keep the
925: same weak prior on $H_0$. When $\Omega_k=0$ is not enforced, the weak
926: prior on $H_0$ has a significant effect on the result as it restricts
927: the extent of the geometrical degeneracy which is present in this
928: case. For WMAP5 only we obtain $\Omega_k=-0.018^{+0.027}_{-0.026}$ which
929: becomes $\Omega_k=-0.013^{+0.019}_{-0.029}$ when ACBAR is included.
930: %\clearpage
931: \begin{figure*}[th!]
932: \resizebox{\hsize}{!}{
933: \plotone{f10.eps}}
934: \caption{The 1-dimensional marginalized likelihood distribution for the lensing
935: amplitude $q_{\rm lens}$ for the combinations of CMB data
936: shown. Although only $q_{\rm lens}=1$ is physically meaningful,
937: these distributions indicate that the data prefer lensing of about
938: the right magnitude. The ratio of the value
939: at $q_{\rm lens}=1$ to that at $q_{\rm lens}=0$ is the
940: Bayesian evidence that lensing at the right amplitude is preferred
941: over no-lensing. The inset figure shows $\Delta {\cal D}_\ell^{\rm
942: lens} \equiv {\cal D}_\ell^{\rm lens} - {\cal D}_\ell^{\rm
943: no-lens}$, with both spectra computed using the best-fit parameters
944: form the lensed analysis of the
945: WMAP5+ACBAR data. Notice how small this correction is relative to
946: ${\cal D}_\ell^{\rm lens}$, shown as the dashed blue curve.
947: The cosmological parameters used to construct $\Delta {\cal D}_\ell^{\rm
948: lens}$ have been allowed to vary as the data requires in
949: constructing the marginalized likelihoods for $q_{\rm lens}$. Each
950: parameter variation leads to a slightly different difference
951: template than that shown in the inset. }
952: \label{fig:lensfit}
953: \end{figure*}
954: %\clearpage
955: \subsection{Residual source marginalization}\label{subsec:ressourcemarg}
956:
957: As discussed in \S~\ref{subsec:foregrounds}, the ACBAR band-powers at
958: $\ell>1950$ marginally exceed the predictions of the best-fit models
959: for the primary CMB.
960: We have repeated the basic parameter runs including an unclustered source
961: contribution described by equation~\ref{eq:src} and marginalize over a wide uniform
962: prior in $q_{\rm src}$ from 0 to $4600~\mu K^2$, i.e. more than 100
963: times the power required to fit the high-$\ell$ points.
964: The results for runs including this marginalization are shown in
965: Tables~\ref{tab:basic} and~\ref{tab:nrun}.
966: The template amplitude $q_{\rm src}$ is not well constrained by the fit,
967: and the effect of the marginalization on the basic parameters is
968: negligible. The addition of a source template improves the fit to the
969: high-$\ell$ points and therefore increases, albeit modestly, the
970: best-fit likelihood: for the ACBAR+WMAP5 combination the change in
971: likelihood is $\Delta \ln L=0.24$.
972: The value and uncertainty for $q_{src}$ given in Table~\ref{tab:basic} spans the range of
973: predictions for the contribution from dusty protogalaxies.
974:
975: Contributions from point sources or the SZE are nearly
976: degenerate in the high-$\ell$ ACBAR band-powers. However, the CMBall
977: combination is potentially sensitive to an SZE contribution because of
978: the SZE frequency dependence. The CBI data has a more significant excess
979: at high-$\ell$, however, through substantial modeling and vetoing, the CBI band-powers
980: are expected to be free of radio source contamination.
981: We assume that the residual contribution of the unresolved source background to the CBI
982: band-powers is sufficiently low that we do not require a source
983: template for that data, only the SZE template.
984:
985: \subsection{Sunyaev-Zel'dovich template extension}\label{sec:sz}
986:
987: The amplitude of the SZE signal depends strongly on the overall matter
988: fluctuation amplitude, $\sigma_8$. We have modified our parameter
989: fitting pipeline to allow for extra frequency dependent contributions
990: to the CMB power spectrum and have implemented it in a simple analysis
991: using a fixed template ${\hat{\cal D}}_\ell^{\rm SZ}$ for the shape of
992: the thermal SZE power spectrum. The template was obtained from two
993: large, hydrodynamical simulations of a scale-invariant ($n_s=1$)
994: $\Lambda$CDM model with $\Omega_bh = 0.029$ and with $\sigma_8=0.9$
995: and $\sigma_8=1.0$ . (See \cite{bond05} for a detailed description of
996: the simulations.) Recently the WMAP team have used a different SZE
997: template based on analytic estimations of the power spectrum
998: \citep{spergel06}. It is characterized by a slower rise in $\ell$ than
999: the simulation-based template, which cut nearby clusters out of the
1000: power spectrum. There has been no fine-tuning of either spectra to
1001: agree with all of the X-ray and other cluster data. This may have an
1002: effect on shape, especially at high-$\ell$.
1003:
1004: The SZE contribution, ${\cal D}_\ell^{\rm SZ} = (q_{\rm SZ})f_\nu {
1005: \hat{\cal D}}_\ell^{\rm SZ}$, added to the base six-parameter model
1006: spectrum has a frequency-dependent SZE pre-factor $f_\nu$. Including
1007: this SZE template with all model parameters free to vary is
1008: complementary to the analysis of \S~\ref{sec:excess} which directly
1009: compared the residual CBI and ACBAR band-powers at $\ell > 1950$ for
1010: the best fit WMAP5+ACBAR model power spectrum. In that more
1011: restrictive analysis, the primary power spectrum is fixed and $f_\nu$
1012: is allowed to vary as well as a broad-band excess power. We found the
1013: ratio of excess power seen by CBI and ACBAR to be compatible with
1014: the ratio of the frequency prefactors $f_\nu$ at 30 GHz and 150 GHz
1015: for the Sunyaev-Zel'dovich effect, although foreground contamination of
1016: both experiments could also contribute to the observed excess.
1017: In this section, we assume that all the excess power seen by the CBI experiment
1018: is due to the SZE.
1019: In the hydrodynamical simulations used to derive the SZE template, the
1020: amplitude was shown to scale as $q_{\rm SZ} = (\sigma_8/0.9)^{7}
1021: (\Omega_bh/0.029)^2$. We consider two cases: (1) the scaling
1022: parameter $q_{\rm SZ}$ is slaved to the results from the hydrodynamical
1023: simulations, which means that it is primarily determined by
1024: the primary CMB data; (2) $q_{\rm SZ}$ is allowed to float freely and
1025: an independent $\sigma_8^{\rm SZ}$ is derived, to be compared with the
1026: $\sigma_8$ that is derived from the basic six parameters. We use a
1027: uniform prior in $q_{\rm SZ}$ in this case with limits $0\le q_{\rm
1028: SZ}\le 4.0$.
1029:
1030: Regardless of the data combination, we find that including an SZE
1031: component in the model has little effect on the values of most basic
1032: cosmological parameters (see Table~\ref{tab:basic}), whether $q_{\rm
1033: SZ}$ is related to cosmic parameters through $q_{\rm SZ} =
1034: (\sigma_8/0.9)^{7} (\Omega_bh/0.029)^2$ or is allowed to float freely.
1035: Note that the SZE results break the $A_s e^{-2\tau}$ near-degeneracy
1036: (as does weak lensing, though not as strongly).
1037:
1038: With the combination of the ACBAR and WMAP5 data, for which there is only a weak
1039: indication of excess power, we
1040: find a freely-floating SZE amplitude results in $q_{\rm
1041: SZ}=0.94^{+0.35}_{-0.93}$, with no effective lower bound. We can use
1042: the above relation of $q_{\rm SZ}(\sigma_{8},\Omega_b h)$ to estimate
1043: a corresponding $\sigma_{8}^{({\rm SZ})}= 0.97^{+0.09 (+0.13)}_{-0.13
1044: (-0.31)}$, where in brackets we have indicated the $2\sigma$ errors.
1045: This result is higher than, but compatible within the
1046: uncertainties, to values obtained from the primary CMB fits alone: the
1047: ACBAR + WMAP5 fits in Table~\ref{tab:basic} give $\sigma_8 =
1048: 0.80^{+0.03}_{-0.03}$. The confidence limits of the derived
1049: $\sigma_{8}^{({\rm SZ})}$ depend strongly on the choice of measure,
1050: which is here taken to be uniform in the amplitude $q_{\rm SZ}$ (This
1051: is evident in the translation of the relative flatness of the
1052: likelihood at low $q_{\rm SZ}$ to a ``1-sigma detection'' in
1053: $\sigma_8^{({\rm SZ})}$). When the SZE contribution is slaved to the
1054: $\sigma_8$ and $\Omega_bh$ values from the primary spectrum, there is no effect
1055: on $\sigma_8$.
1056: The high-$\ell$ excess power is not significant
1057: enough to change the well constrained value of $\sigma_8$ to the weakly preferred
1058: higher value.
1059: The effect of slaved and unslaved fits can be seen in
1060: Fig.~\ref{fig:best_fit}.
1061:
1062: When the high-$\ell$ band-powers of CBI and BIMA are included in the
1063: analysis, there is a significant detection of excess power. Both the
1064: CBI and BIMA band-powers are from $30\,$GHz interferometric
1065: observations and have higher $f_\nu$ values than ACBAR. For the slaved
1066: case the $\sigma_{8}$ value and errors are unchanged. For the floating
1067: case, we find $q_{\rm SZ}= 0.69^{+0.11}_{-0.11}$ which maps to
1068: $\sigma_{8}^{({\rm SZ})}=0.93^{+0.04}_{-0.05}$.
1069:
1070: These central values and uncertainties are computed by transforming
1071: integrals of the likelihood $L(q_{\rm SZ}, \Omega_bh , ... )$ over the
1072: prior measures to confidence limits for $\sigma_{8}^{(SZ)}$. Exactly
1073: what measure to place on $q_{\rm SZ}$ and therefore on
1074: $\sigma_{8}^{(SZ)}$ is debatable. A measure uniform in
1075: $\alpha_{SZ}=q_{\rm SZ}^{1/2}$, as we used in K07 and
1076: \citep{goldstein03} translates into a measure $\propto q_{\rm
1077: SZ}^{-1/2}dq_{\rm SZ}$ which favors lower values of
1078: $\sigma_{8}^{(SZ)}$ than the measure uniform in $q_{SZ}$ that we have
1079: adopted.
1080:
1081:
1082: As the cosmological parameters vary, the SZE template may depend on
1083: $\sigma_{8}$ and $\Omega_b h$, and certainly depends on the spectral
1084: index $n_s$ and astrophysical issues such as the history of energy
1085: injection into the cluster system. In a more complete treatment than
1086: that presented here, the shape should be modified along with the base
1087: cosmological parameters in the MCMC runs.
1088:
1089: We caution that the derived $\sigma_{8}^{({\rm SZ})}$ depends on the
1090: SZE template shape, its extension into the higher $\ell$ regime probed
1091: by BIMA, and the prior measure placed upon $q_{\rm SZ}$.\footnote{We
1092: also note that the non-Gaussian nature of the SZE signal was included
1093: in the BIMA results, but not in the CBI results. The non-Gaussian
1094: effect increases the sample variance and tends to open up the
1095: allowed range towards lower $\sigma_8$ values
1096: \citep{goldstein03,readhead04}.} The modest decrease in log likelihood for
1097: the fit to the model when the SZE is taken into account, is $\Delta \ln
1098: L = 0.26$ for the ACBAR+WMAP5 combination and increases to $\Delta \ln
1099: L = 1.91$ for CMBall+BIMA.
1100:
1101: Regardless of the assumed prior, the interpretation of excess power
1102: as being due to the SZE results in a non-zero $\sigma_{8}^{(SZ)}$
1103: for the CMBall+BIMA combination.
1104: Uncertainties for $\sigma_{8}^{({\rm SZ})}$ are about a factor of two
1105: larger than for the $\sigma_8$ determined from the primary CMB data
1106: and there is a tension at about the 2-sigma level between the two
1107: median values. A visual summary of the results is shown in
1108: Fig.~\ref{fig:SZ} where we plot both $\sigma_{8}$ and
1109: $\sigma_{8}^{({\rm SZ})}$ against the spectral index for a number of
1110: data combinations. The addition of LSS data does not significantly
1111: change these results. When the ACBAR dusty point source or SZE template
1112: marginalization is included, $\sigma_8$ decreases slightly for the
1113: CMBall dataset.
1114:
1115: Recent weak lensing results are in basic agreement with the primary
1116: $\sigma_8$ values when $\Omega_m = 0.26 \pm 0.03$ from
1117: Table~\ref{tab:basic} is used. With 100 deg$^2$ of lensing data from
1118: the combination of the CFHT weak lensing legacy, RCS, Virmos-Descart
1119: and GaBaDos surveys, \cite{benjamin07} get $\sigma_8
1120: (\Omega_m/0.26)^{0.59} = 0.80 \pm 0.05$. From the CFHT weak lensing
1121: legacy survey alone, \cite{fu07} get $\sigma_8 (\Omega_m/0.26)^{0.64}
1122: = 0.753 \pm 0.043$, and get $\sigma_8 (\Omega_m/0.26)^{0.53} = 0.82
1123: \pm 0.084$ if only large-scale linear-regime results are used. These
1124: weak lensing numbers are lower than past published results because of
1125: improved treatments of the redshift distribution of the lensed
1126: sources.
1127:
1128: As shown in Fig.~\ref{fig:best_fit}, the marginal excess power in ACBAR
1129: is consistent with the combination of a point source contribution at the upper limit
1130: of the $150\,$GHz extrapolations and an SZE template at the level predicted from the
1131: primary anisotropy $\sigma_8$ value. In this scenario, the cosmological
1132: parameters are virtually unchanged, but no explanation is provided for the
1133: CBI excess. The SZE
1134: analyses with a free-floating amplitude have not included additional
1135: foreground sources for CBI, BIMA, or ACBAR. The effect of radio
1136: sources extrapolated to 30~GHz was included in the original CBI and BIMA
1137: results, and is unlikely to be an important contaminant for
1138: ACBAR. Dusty proto-galaxies should not effect CBI and BIMA, but will
1139: contribute to the high-$\ell$ ACBAR band-powers.
1140: Given the uncertainty in the source contamination
1141: for ACBAR, the weak detection of excess power
1142: does not significantly support the SZE interpretation of the CBI+BIMA excess.
1143:
1144: \subsection{Running Spectral Index}\label{sec:nrun}
1145:
1146: There has been interest in the running of the spectral index
1147: $dn_s/d\ln k$ since the first release of WMAP data, which showed
1148: evidence for a significant negative running when combined with LSS and
1149: Lyman alpha forest observations \citep{spergel03}. With the precision
1150: of the new ACBAR data, we might expect improved constraints on running
1151: of the spectral index. To the basic six parameters in the minimal
1152: model, we add running of the spectral index $dn_s/d\ln k (k_\star)$
1153: around the pivot point $k_\star=0.05$ Mpc$^{-1}$. We adopt the
1154: conventionally-used uniform prior in $dn_s/d\ln k (k_\star)$, although
1155: in usual slow-roll-inflation models, the spectral index fluctuation
1156: $\delta n_s \propto \ln (k/k_\star ) dn_s/d\ln k (k_\star)$ is
1157: typically restricted by $\vert 1-n_s\vert$. Table~\ref{tab:nrun}
1158: summarizes the parameter values when running is allowed and
1159: demonstrates that its inclusion has only a small effect on the other
1160: parameters. For WMAP5 only, we find $dn_s/d\ln k
1161: (k_\star)=-0.031^{+0.029}_{-0.028}$. The main tendency for the
1162: negative value comes from the low-$\ell$ WMAP5 data. When the ACBAR
1163: data is added, the median value is similar, $dn_s/d\ln k
1164: (k_\star)=-0.037^{+0.023}_{-0.022}$, with a reduced error. The results
1165: are nearly identical for CMBall, but more compatible with no-running
1166: when LSS is added, $-0.016^{+0.019}_{-0.018}$. The precise
1167: measurement of the high-$\ell$ CMB power spectrum provided by ACBAR is
1168: a potentially powerful constraint on running, but may also include
1169: significant contributions from secondary anisotropies and foregrounds.
1170: Therefore, in the last two columns we have included the effect of
1171: marginalizing over the SZE or point source templates. In both cases,
1172: the mean value of the running becomes more negative and more
1173: significant at the ~2-sigma level. A visual representation of the
1174: impact of adding the ACBAR data is given in Fig.~\ref{fig:nrun} which
1175: shows the correlation between $n_s$ and $dn_s/d\ln k$. The scalar
1176: spectral index, $n_s(k_\star ) =0.918\pm 0.032$ for the WMAP5 + ACBAR
1177: combination, depends on the choice of pivot point $k_\star$; a smaller
1178: value would yield a higher result while a higher one would give an
1179: even lower result.
1180:
1181: %\clearpage
1182: \input{tab5}
1183: %\clearpage
1184:
1185: \subsection{Tensor modes}\label{sec:tensor}
1186:
1187: We have run a limited number of cases including tensor modes,
1188: characterizing their strength relative to scalar perturbations by $r$
1189: and fixing the tensor tilt by $n_t \approx -r/8(1-r/16)$. The most
1190: stringent upper limit for $r$ is given by the CMBall combination which
1191: yields $r < 0.40$ (95\% confidence). This assumes a uniform prior
1192: measure for $r$, as is conventional in parameter estimation, although
1193: without much justification except that it is conservative. Adding LSS
1194: tightens this limit. Our result can be compared with those obtained by
1195: \citet{dunkley08}, $r<0.43$ (95\% confidence) for WMAP5 alone.
1196: %
1197:
1198: %\clearpage
1199: \begin{figure*}[th!]
1200: \resizebox{\hsize}{!}{
1201: \plotone{f11.eps}}
1202: \caption{68\% and 95\% 2D marginalized contours in $n_s$ and $n_{run}
1203: =dn_s/d\ln k$. The left panel shows the results for the WMAP5 and
1204: ACBAR+WMAP5 combinations. The right panel shows the effect of
1205: adding the rest of the CMB data and marginalizing over the SZE
1206: template. The basic parameters are essentially unchanged
1207: if we marginalize over one or the other of the two template
1208: amplitudes.}
1209: \label{fig:nrun}
1210: \end{figure*}
1211: %\clearpage
1212:
1213:
1214: \begin{figure*}[ht!]
1215: \centering
1216: \includegraphics[width=5.5in]{f12.eps}
1217: \caption{ Best-fit models for the ACBAR + WMAP5 combination. Only the
1218: $\ell>1000$ ACBAR band-powers are shown. The arrows are indicative of
1219: the possible (coherent) shift in the 1-$\sigma$ confidence limits
1220: due to beam error. The top panel shows fits with just primary CMB
1221: (black, solid) and with an extra SZE contribution. The case where the
1222: SZE template is scaled independently as $q_{\rm SZ}$ gives the best
1223: fit to the high-$\ell$ excess (red, short-dashed) with $q_{\rm
1224: SZ}=0.82$. The case where the SZE amplitude is slaved to the
1225: cosmological parameters $\sigma_8^7(\Omega_bh)^2$ (blue,
1226: long-dashed) does not yield enough power to fit the excess given the
1227: best-fit value of $\sigma_8=0.81$. The bottom panel shows the case
1228: when a point source contribution scaled by $q_{\rm src}$ is included
1229: for ACBAR together with the slaved SZE contribution (blue,
1230: dotted). In this case the best-fit model has $\sigma_8=0.81$
1231: determining the sub-dominant SZE contribution and $q_{\rm src}=31\mu
1232: K^2$ determining the point source contribution.}
1233: \label{fig:best_fit}
1234: \end{figure*}
1235: %\clearpage
1236:
1237: \begin{figure*}[ht!]
1238: \resizebox{\hsize}{!}{
1239: \plotone{f13.eps}}
1240: \caption{The figure contrasts the one and two sigma contour intervals
1241: for $\sigma_8$ determined from the primary anisotropy component of
1242: the CMB (left) with the value inferred from the SZE template
1243: transformation of $q_{\rm SZ}$ into $\sigma_{8}^{({\rm SZ})}$
1244: (right), assuming a uniform prior measure in $q_{\rm SZ}$. Allowing
1245: for a point source contribution would decrease the tension between $\sigma_8$
1246: and $\sigma_8^{SZ}$ for the ACBAR+WMAP5 case. These
1247: panels also demonstrate the strength of the deviation of $n_s$
1248: from unity for the flat $\Lambda$CDM model. }
1249: \label{fig:SZ}
1250: \end{figure*}
1251: %\clearpage
1252: \section{Conclusions}\label{sec:conclusion}
1253:
1254: We have used the complete ACBAR $150\,$GHz data set to measure the
1255: CMB temperature anisotropy angular power spectrum. Over
1256: three seasons of observation, ACBAR dedicated 85K detector-hours to
1257: CMB observations at 150 GHz and covered 1.7\% of the sky. The data are
1258: calibrated by comparing CMB temperature maps for the largest ACBAR
1259: fields with those produced by WMAP5. This calibration is found to
1260: be consistent with the previous planet-based and RCW38-based
1261: calibrations, but with the temperature uncertainty reduced to
1262: $2.05\%$. In the original preprint of this paper, the ACBAR band-powers
1263: were calibrated through comparison with WMAP3.
1264: The new WMAP5 results became available during the review process and we
1265: have updated the ACBAR band-powers and cosmological parameters to take
1266: advantage of this new information. The ACBAR band-powers are otherwise
1267: unchanged. The new WMAP5 parameters and calibration subtly changed the
1268: best fit WMAP5 and ACBAR model resulting in a decrease of the significance
1269: of the high-$\ell$ excess in the ACBAR data.
1270:
1271: The ACBAR band-powers reported in Table~\ref{tab:bands} are the most
1272: sensitive measurements to date of CMB temperature anisotropies for
1273: multipoles between $\ell \sim 900$ and 3000. In this data, the fourth and fifth
1274: acoustic peaks are significantly detected for the first time.
1275: These precise measurements of the CMB temperature
1276: anisotropies at high-$\ell$ are consistent with a spatially flat, dark
1277: energy-dominated $\Lambda$CDM cosmology. Including the
1278: effects of CMB weak lensing in the computation of model power spectra improves
1279: the fits to the combined ACBAR+WMAP5 data.
1280: The excellent fit of the $\Lambda$CDM cosmological model to
1281: the combined ACBAR+WMAP5 data at $\ell \lesssim 2000$ is a strong
1282: confirmation of the standard cosmological paradigm and gives us
1283: confidence in the resulting parameter values. The ACBAR data favor
1284: higher median values of $\sigma_8$, $\Omega_m$, and $\theta$ than
1285: those preferred by WMAP3; however, with the improved WMAP5 results for
1286: $\ell> 650$, this tension has been resolved. The
1287: parameter values remain stable with the inclusion of additional CMB
1288: and LSS data sets, with or without the marginalization over an SZE or point
1289: source template for ACBAR. For example, $\sigma_8\sim
1290: 0.80$ holds for all parameter variations in Table~\ref{tab:basic},
1291: and even in Table~\ref{tab:nrun} with the inclusion of a running
1292: spectral index.
1293:
1294: We have performed strict jackknife tests with the data, and find that the
1295: results are free of significant systematic errors. We have projected
1296: out templates derived from the FDS99 dust model and the PMN radio source
1297: catalog from the ACBAR maps before estimating the band-powers and find
1298: the residual contributions from these foregrounds to be negligible at
1299: the current sensitivity. The contribution of dusty
1300: proto-galaxies is expected to be insignificant for all but the few
1301: highest-$\ell$ band-powers, but remains poorly constrained due to our incomplete
1302: knowledge of the spectral dependence of these sources.
1303:
1304: Secondary anisotropies are expected to become important at small
1305: angular scales. The ACBAR band-powers are slightly higher
1306: (1.1$\sigma$) than expected for the primary CMB anisotropy at
1307: multipoles above $\ell \sim 2000$.
1308: We expect that some of this power results from contamination by dusty proto-galaxies.
1309: However, the combined signal is considerably smaller than the significant detection of
1310: excess power reported by the CBI experiment at $30\,$GHz.
1311: A joint analysis of the CBI and ACBAR band-powers in the
1312: multipole range of $2000\lesssim \ell \lesssim 3000$
1313: argues strongly against the CBI excess having the spectrum of primary CMB
1314: anisotropy.
1315: These results are consistent with the excess power seen by CBI
1316: being due to either the Sunyaev-Zel'dovich effect or radio source
1317: contamination.
1318: Higher sensitivity observations over a broad range of frequencies are
1319: necessary in order to fully characterize CMB secondary anisotropies and eliminate
1320: potential foreground contamination.
1321:
1322:
1323: \vspace{1cm}
1324:
1325: The ACBAR program has been primarily supported by NSF office of polar
1326: programs grants OPP-8920223 and OPP-0091840. This research used
1327: resources of the National Energy Research Scientific Computing Center,
1328: which is supported by the Office of Science of the U.S. Department of
1329: Energy under Contract No. DE-AC03-76SF00098. Some computations were
1330: performed on the Canada Foundation for Innovation funded CITA
1331: Sunnyvale cluster. Chao-Lin Kuo acknowledges support from a NASA
1332: postdoctoral fellowship and Marcus Runyan acknowledges support from a
1333: Fermi fellowship. Christian Reichardt acknowledges support from a
1334: National Science Foundation Graduate Research Fellowship. Some of the
1335: results in this paper have been derived using the HEALPix
1336: \citep{gorski05} package. We thank members of the BOOMERANG team, in
1337: particular Brendan Crill, Bill Jones, and Tom Montroy for providing
1338: access to the B03 data, the pipeline used to generate simulation maps,
1339: and assistance with its operation. We thank Antony Lewis for
1340: discussions about ways to parameterize tests for weak lensing in the
1341: data.
1342:
1343:
1344:
1345:
1346:
1347: \appendix
1348:
1349: \section{CALIBRATION }\label{app:calib}
1350:
1351: The calibration used in K07 was linked to the Boomerang03 (B03)
1352: calibration with observations of RCW38. In this section, we describe
1353: a new calibration using an $a_{lm}$-based comparison of CMB structure
1354: observed by WMAP5 and ACBAR in 2005. This method is inspired by the
1355: calibration scheme used to calibrate B03 to WMAP. The 2005 calibration
1356: is carried to other years by an ACBAR-ACBAR power spectrum comparison
1357: on fields observed in both years. The WMAP-ACBAR cross-calibration
1358: method is described below, with a detailed accounting of uncertainty
1359: in Table~\ref{tab:calerr}.
1360:
1361:
1362: \paragraph{WMAP-ACBAR Calibration}
1363:
1364: Calibrating with the CMB temperature anisotropies has two main
1365: advantages. The first is that the calibration of the WMAP temperature
1366: maps (at 0.2\% in temperature) is an order of magnitude more precise
1367: than the flux calibration of the calibration sources ACBAR used in
1368: previous releases. The second advantage is that the anisotropies have
1369: the same spectrum as what is being calibrated, rendering the large
1370: frequency gap between WMAP and ACBAR irrelevant.
1371:
1372:
1373: %\clearpage
1374: \begin{figure*}
1375: \epsscale{.9}
1376: \plotone{f14.eps}
1377: \caption{ A comparison of observations of the CMB8 field made with B03
1378: and ACBAR. This field lies in the deep region of the B03 map. The
1379: top two maps are from B03. The bottom two maps are from ACBAR. In
1380: the {\em top left} panel, the B03 map of the CMB8 field. The
1381: dynamic range of this map is greater than that of the other three
1382: figures. The increased noise at one edge marks the edge of the B03
1383: deep coverage. The ACBAR filtering is applied to the B03 map to
1384: create the map in the {\em top right} panel. Directly below it in
1385: the {\em bottom right} panel is the ACBAR map of same region. Note
1386: the clear correspondence between the CMB anisotropies observed by
1387: B03 and ACBAR. Three bright point sources have been masked. An ACBAR
1388: left-right sweep difference map is shown in the {\em bottom left}
1389: panel. The power spectrum of this map (and the other 9 fields) is
1390: plotted in Fig. \ref{fig:sys}. }
1391: \label{fig:cmb5}
1392: \end{figure*}
1393: %\clearpage
1394: The two experiments have different scan patterns, noise, beam widths, and spatial
1395: filters that will effect the measured flux. In this analysis, we assume that the WMAP5 maps are effectively unfiltered except for the instrumental beam function. The two maps can then be represented as:
1396: \[ S^{WMAP}_i = \int{T(x) B_{WMAP}(x_i - x)dx} +N^{WMAP}_i\]
1397: \[ S^{ACBAR}_i = F_{ij} \int{T(x) B_{ACBAR}(x_j - x)dx} +N^{ACBAR}_i\]
1398: where T is the underlying CMB signal, N is the instrumental noise, B is the beam function, and $F_{ij}$ is the ACBAR filter matrix as defined in Section \ref{sec:analysis}. We reduce the filtering differences by resampling the WMAP map using the ACBAR pointing information and applying the ACBAR spatial filtering to generate an `ACBAR-filtered' WMAP map.
1399: \[ S^{WMAP-equivalent}_i = F_{ij} (\int{T(x) B_{WMAP}(x_j - x)dx} +N^{WMAP}_i)\]
1400: The results of applying this algorithm to the B03 map is shown in Figure \ref{fig:cmb5}. We choose to do the absolute calibration via cross-power spectra rather than a direct pixel-to-pixel comparison of the maps. Using cross-spectra significantly reduces the impact of the noise model on the result. The significant beam differences between the experiments are more naturally dealt with in multipole space than in pixel space. We construct the ratio from the filtered maps:
1401: \[ R = \Re\left( \left \langle \frac{a_{\ell m}^{WMAP-X *} * a_{\ell
1402: m}^{ACBAR-Z}}{a_{\ell m}^{ACBAR-Y *} * a_{\ell m}^{ACBAR-Z}
1403: (B_{\ell}^{WMAP-X}/B_{\ell}^{ACBAR})} \right \rangle \right) \]
1404: where X can denote either the V- or W-band map for WMAP and Y/Z marks
1405: either of two noise-independent ACBAR combinations. There is a narrow
1406: $\ell$-range from 256-512 useful for calibration. The range is limited
1407: at high-$\ell$ by the rapidly falling WMAP beam function and at
1408: low-$\ell$ by the ACBAR polynomial filtering which acts as a high-pass
1409: filter. We choose to use the WMAP V \& W bands to take advantage of
1410: their smaller beam size.
1411:
1412: %\clearpage
1413: \input{tab6}
1414: %\clearpage
1415:
1416: Monte Carlo simulations are used to determine the transfer function of
1417: this estimator. We generate CMB sky simulations convolved with the
1418: respective instrumental beam functions using the
1419: Healpix\footnote{http://healpix.jpl.nasa.gov} library. We resample
1420: each realization and apply the ACBAR filtering matrix described above
1421: to generate equivalent maps for each field. We expected and found a
1422: small intrinsic bias as the beam convolution and filtering operations
1423: do not commute: $B^{ACBAR} * F_{ij} B^{WMAP} \ne B^{WMAP} * F_{ij}
1424: B^{ACBAR}$. We correct the real data by the $\ell$-dependent transfer
1425: function measured in these simulations. The technique is easily
1426: adapted to estimate the error caused by pointing uncertainties and to
1427: confirm that the estimator is unbiased with the inclusion of
1428: noise. The derived error in the transfer function is listed in
1429: Table~\ref{tab:calerr}.
1430:
1431: Foreground sources have the potential to systematically bias a
1432: calibration bridging 60 to 150 GHz. Radio sources, synchrotron
1433: emission, dust, and free-free emission all have a distinctly different
1434: spectral dependence than the CMB, which could lead to a calibration
1435: bias. This risk is ameliorated by the positioning of the ACBAR fields
1436: in regions of exceptionally low foregrounds. Bright radio sources
1437: detected in either experiment are masked and excluded from the
1438: calibration. The calibration proved insensitive to the exact threshold
1439: for source masking. We use the MEM foreground models in
1440: \citet{hinshaw06} to estimate the RMS fluctuations of each foreground
1441: relative to the CMB fluctuations and find that the free-free and
1442: synchrotron fluctuations are less than 0.1\% of the CMB fluctuations
1443: in all frequency bands while dust emission can reach 1.5\% of CMB
1444: fluctuations in the 150 GHz maps. We test the effects of the most
1445: significant foreground, dust, by adding the FDS99 dust model
1446: \citep{finkbeiner99} to a set of CMB realizations. The resultant maps
1447: are passed through a simulated pipeline as outlined in the previous
1448: paragraph. We find that the addition of dust does not introduce a
1449: detectable bias with an uncertainty of 0.2\%.
1450:
1451: We perform a weighted average of the measured calibration ratio across all $\ell$-bin, field and band combinations after correcting for the estimated signal-only transfer
1452: functions. We estimate the calibration error to be 1.97\% for the 2005 data.
1453: Table~\ref{tab:calerr} tabulates the contributing factors and error budget.
1454: We then propagate this $a_{\ell m}$-based calibration to the CMB observations
1455: done in 2001 and 2002.
1456:
1457:
1458:
1459: \paragraph{ACBAR 2001-2002 and 2002-2005 Cross Calibrations}
1460:
1461: We propagate the 2005 calibration into 2001 and 2002 by comparing the
1462: 2001 observations of the CMB2 field to the overlapping 2002 CMB4 field, and the 2002 observations of the CMB5 field to the 2005 observations of the CMB5 field. A power spectrum is
1463: calculated for each overlapping region and the ratio of the band-powers is used to derive a cross calibration. The procedures used are outlined in more detail in K07. We use the same relative calibration for 2001 as K07: $T_{2001}/T_{2002} = 1.238 \pm 0.067$. We find cross-calibration factor for 2002 to be $T_{2005}/T_{2002} = 1.035 \pm 0.025$. We apply these corrections to the data
1464: and determine the overall calibration uncertainty to be $2.05\%$ (in temperature units) based primarily on the uncertainties associated with WMAP/ACBAR-2005 cross calibration.
1465:
1466:
1467:
1468: \bibliographystyle{apj}
1469: \bibliography{merged}
1470:
1471:
1472: \end{document}
1473:
1474:
1475:
1476:
1477:
1478:
1479:
1480:
1481:
1482:
1483: