Improved Measurement of the Angular Power Spectrum of Temperature Anisotropy in the CMB from Two New Analyses of BOOMERANG Observations

TL;DR

This paper presents a refined measurement of the CMB angular power spectrum from BOOMERANG-98 using two independent analysis pipelines, MADCAP and FASTER, over 20 bands spanning $\ell \sim 50$ to $1000$. The identical data are processed with distinct methodologies, yielding highly consistent power spectra and enabling rigorous cross-checks of systematics. The analysis detects multiple acoustic features consistent with adiabatic inflationary CDM models and yields cosmological parameters that robustly favor a flat $\Lambda$CDM universe, with parameters such as $\Omega_b h^2$, $\Omega_c h^2$, $n_s$, and $\Omega_\Lambda$ constrained and largely immune to analysis differences. The work demonstrates the reliability of CMB-derived cosmological constraints and highlights the value of independent pipelines for validating high-precision cosmological inferences.

Abstract

We report the most complete analysis to date of observations of the Cosmic Microwave Background (CMB) obtained during the 1998 flight of BOOMERANG. We use two quite different methods to determine the angular power spectrum of the CMB in 20 bands centered at l = 50 to 1000, applying them to 50% more data than has previously been analyzed. The power spectra produced by the two methods are in good agreement with each other, and constitute the most sensitive measurements to date over the range 300 < l < 1000. The increased precision of the power spectrum yields more precise determinations of several cosmological parameters than previous analyses of BOOMERANG data. The results continue to support an inflationary paradigm for the origin of the universe, being well fit by a 13.5 Gyr old, flat universe composed of approximately 5% baryonic matter, 30% cold dark matter, and 65% dark energy, with a scale invariant initial density perturbations.

Figures (15)

• Figure 1: The sky covered by CMB observations; the color scale indicates the depth of coverage (diagonal component of the noise covariance matrix) in a $7'$ Healpix pixel, in the map produced by the MADCAP analysis described below. The region enclosed by the solid line is that used for the power spectrum estimation. The three circles show the locations of three bright known quasars; data within a $0.5^\circ$ radius of the quasars is not used in the power spectrum estimation.
• Figure 2: The maps of CMB temperature produced by MADCAP (top) and FASTER (bottom). For comparison, both maps are pixelized at $7'$; in practice we use a $7'$ ($3.5'$) pixelization in the MADCAP (FASTER) analysis. The strikingly different appearance of the maps, with the MADCAP map preserving more information on large scales, illustrates some of the significant differences in the two analysis methods, as described in the text.
• Figure 3: Angular power spectra derived from the FASTER pipeline. The solid black circles in each panel show the reference FASTER spectrum, which is derived from a $3.5'$ pixelized map using S+N weighting, spatially filtered as described in the text to remove constant-declination stripes. In the top panel the reference spectrum is compared with a spectrum derived using a single beam window function, as in MADCAP. The second panel shows the effect of using $7'$ pixelization. The third panel illustrates the effect of removing the constant declination stripes; the primary effect is to increase the error in the first bin. The bottom panel shows the result of using a uniformly weighted map and neglecting to remove the constant declination stripes. The top three panels show excellent agreement with the reference spectrum, while the bottom panel shows good agreement except at very high $\ell$.
• Figure 4: (1dps-2dps) difference maps, both at $7'$ pixelization to facilitate map comparisons by eye, for the region of sky where these scans overlap. The color scale is the same as for the previous figures. Note that the consistency test power spectra are calculated on these maps divided by two, (1dps-2dps)/2. Top panel: The MADCAP difference map. This map is not destriped, since in that pipeline the constant-declination stripes are ignored (by introducing a constraint matrix) in the derivation of the angular power spectrum. Bottom panel: The destriped FASTER difference map. Note that the MADCAP input timestream contains additional low frequency information that is removed by an additional highpass filter in the FASTER pipeline.
• Figure 5: MADCAP and FASTER angular power spectra and (1dps-2dps)/2 difference map power spectra. Top Panel: the FASTER (filled blue circles) and MADCAP (filled red squares) angular power spectra, and their respective (1dps-2dps)/2 difference map power spectra (open symbols). The effects of constant-declination stripes have been removed in each of these analyses. Bottom panel: the difference map angular power spectra plotted on a magnified scale. There is a systematic effect near $\ell \sim 200$ in the FASTER power spectrum, which is absent in the MADCAP treatment. The level of these residuals is much smaller than the statistical errors on the full power spectrum, shown in the top panel.