Abstract: The combined CfA2/SSRS redshift catalog has recently been improved in a number of ways and made public. We compute the redshift-space power spectrum of this catalog using the regions at -2.5 < delta < 50 north (20h < alpha < 3h) and south (9h < alpha < 17h, b < -13) of the Galactic plane, where it is 98% complete down to Zwicky magnitude 15.5 and contains 13,681 galaxies. Our analysis uses Heavens-Taylor mode expansion, Karhunen-Loeve data compression and the Fisher matrix technique to compute quadratic band power estimates. This allows an exact calculation of window functions, including the integral constraint, as well as the production of a power spectrum with uncorrelated error bars. Our results with this larger data set agree well those of previous studies. We analyze 101 Mpc/h and 130 Mpc/h volume-limited subsets as well as the full magnitude-limited sample, and our results are well fit by, e.g., Lambda-CDM models with sigma8 = 1.2, 1.4 and 1.4, respectively. We estimate the effect of extinction using the Schlegel, Finkbeiner and Davis dust map.
Introduction
To understand the evolution of the universe, cosmologists must accurately measure the power spectrum $P(k)$, which describes matter clumping across various scales. However, extracting $P(k)$ from galaxy surveys is complicated by survey geometries, selection functions, and the "integral constraint"—the uncertainty in the true average density of the universe. This paper applies advanced statistical techniques, originally developed for Cosmic Microwave Background (CMB) analysis, to the Updated Zwicky Catalog (UZC) to produce a high-precision power spectrum with uncorrelated error bars.
The Dataset and Survey Geometry
The primary dataset is the Updated Zwicky Catalog (UZC), an enhanced version of the CfA2 and SSRS2 surveys. It contains 13,681 galaxies and achieves 98% completeness. The analysis focuses on specific northern and southern regions of the Galactic plane to minimize the impact of our own galaxy's dust and stars.
To ensure robustness, the authors also analyzed volume-limited subsets (101 $h^{-1}$Mpc and 130 $h^{-1}$Mpc) and utilized the Schlegel, Finkbeiner & Davis (1998) dust map to account for Galactic extinction.
Methodology
The study employs a sophisticated four-step pipeline to transform raw galaxy coordinates into a clean power spectrum:
- Heavens-Taylor (HT) Pixelization: The galaxy distribution is expanded into "spherical waves" (spherical harmonics and Bessel functions), naturally accounting for the survey's spherical geometry.
- Karhunen-Loève (KL) Compression: This technique identifies the most informative data modes, reducing computational complexity while preserving the cosmological signal.
- Quadratic Band-Power Estimation: The Fisher matrix technique is used to estimate power in 52 discrete $k$-bands.
- Fisher Decorrelation: A matrix transformation ensures that error bars for each $k$-bin are independent, facilitating easier model fitting.
Additionally, an Integral Constraint Correction is implemented to immunize results against uncertainties in the total number of galaxies expected in the survey volume.
Key Results and Cosmological Implications
The measured power spectrum is remarkably consistent with a flat $\Lambda$CDM "concordance" model ($\Omega_\Lambda \approx 0.7, \Omega_m \approx 0.3$). To match these models, the UZC galaxies require a bias factor ($b$) between 1.2 and 1.4, indicating that galaxies are slightly more clumped than the underlying dark matter.
Key insights from the analysis include:
- Compression Efficiency: KL compression demonstrates that most cosmological information is packed into a small fraction of data modes, promising for future large-scale surveys.
- Prior Robustness: The pipeline is robust against incorrect initial guesses (priors) for the power spectrum.
- Large-Scale Sensitivity: The largest scales (low $k$) are highly sensitive to radial selection function modeling, requiring careful interpretation.
Conclusion and Limitations
While the results provide tight constraints, the analysis was limited to large scales ($k < 0.3 h \text{ Mpc}^{-1}$) to avoid non-linear distortions like the "Fingers of God" effect. Future work may incorporate more complex stochastic biasing models or higher-order statistics, such as the bispectrum, to extract further information from non-linear galaxy clumping on smaller scales.
Summary updated