Nearly Full-Sky Low-Multipole Cosmic Microwave Background Temperature Anisotropy: III. CMB Temperature Anomalies

Technical Summary: Nearly Full-Sky Low-Multipole Cosmic Microwave Background Temperature Anisotropy: III. CMB Temperature Anomalies

Problem Statement
Despite the success of the $\Lambda$CDM model in describing most cosmological data, several "anomalies" persist in the Cosmic Microwave Background (CMB) temperature fluctuations at large angular scales. These features, observed in data from COBE, WMAP, and Planck, include a lack of large-scale angular correlation, parity asymmetry, quadrupole-octopole alignment, low northern variance, and hemispherical local-variance asymmetry. The statistical significance of these features is typically in the range of $2\sigma$ to $3\sigma$. A primary challenge in assessing these anomalies is distinguishing between genuine physical deviations from statistical isotropy and artifacts introduced by foreground contamination or the aggressive sky masking required to remove galactic foregrounds. Previous analyses often relied on the Planck "common mask," which excludes approximately 26% of the sky, potentially biasing the significance of large-scale features.

Methodology
This study revisits five specific CMB anomaly tests using new, low-resolution ($1^\circ$), foreground-cleaned CMB temperature maps derived from WMAP and Planck data, as detailed in a companion paper (Nofi et al. 2025a). These maps utilize a morphological cleaning technique that fits external foreground templates (synchrotron, free-free, CO, and dust) to the target maps, allowing for a minimal sky mask of only 1%.

The authors compare results from these new "foreground-cleaned" maps (at 70, 94, 100, and 143 GHz) against the standard Planck 2018 component-separated maps (Commander, NILC, SEVEM, SMICA). The analysis employs three distinct sky cuts:

1. Full Sky: No galactic mask applied.
2. 1% Mask: The minimal mask derived from the new cleaning procedure.
3. Common Mask: The standard Planck 2018 mask (downgraded to match the resolution of the new maps), covering ~26% of the sky.

The study evaluates the significance of the anomalies using $10^5$ Gaussian statistically isotropic $\Lambda$CDM simulations generated from the Planck 2018 best-fit model. To assess the impact of the cleaning procedure itself, a subset of simulations ($10^4$) underwent the same morphological cleaning as the real data to check for chance correlations between CMB and foreground morphologies.

Key Contributions

• New Data Products: The application of morphologically cleaned, low-resolution CMB maps that enable analysis with a significantly larger sky fraction (99% vs. 74%) than previously possible.
• Systematic Assessment: A direct comparison of anomaly significance across different foreground cleaning methods (morphological vs. component separation) and sky masking strategies (1% vs. 26%).
• Look-Elsewhere Effect Analysis: An exploration of how the significance of specific statistics (e.g., $S_\mu$ and $R_{\ell_{max}}$) varies with their definition parameters (integration limits and maximum multipole), highlighting the need for look-elsewhere corrections.

Results
The study quantifies the significance of five anomalies using Probability to Exceed (PTE) values and corresponding Gaussian $\sigma$ levels:

1. Low Correlation ($S_{1/2}$):

   • Common Mask: Yields a moderate significance of $\sim 3\sigma$ ($p \approx 0.19\% - 0.24\%$), consistent with previous literature.
   • 1% Mask/Full Sky: The significance drops notably to $\sim 2\sigma$ ($p \approx 5.8\% - 7.3\%$).
   • Conclusion: The choice of mask has a substantial impact; the low correlation anomaly is less significant when a larger fraction of the sky is included.

2. Parity Asymmetry ($R_{27}$):

   • Results are consistent across all maps and sky cuts, showing mild significance of $\sim 2\sigma$ ($p \approx 2.8\% - 5.9\%$).
   • The significance is sensitive to the choice of $\ell_{max}$; a look-elsewhere correction would likely reduce this further.

3. Quadrupole-Octopole Alignment ($S_{QO}$):

   • Shows moderate significance of $\sim 3.2\sigma - 3.5\sigma$ ($p \approx 0.04\% - 0.13\%$) for the full sky and 1% mask.
   • The 1% mask results agree well with full-sky expectations, indicating that foreground contamination does not significantly drive this alignment. The common mask yields lower significance ($\sim 1.5\sigma - 1.8\sigma$), likely due to the removal of the aligned power in the galactic plane.

4. Low Northern Variance ($\sigma^2_{16}$):

   • Shows moderate significance of $\sim 3.1\sigma - 3.2\sigma$ with the common mask.
   • With the 1% mask, the significance decreases slightly to $\sim 2.9\sigma$. The authors note that the selection of the northern hemisphere is an a posteriori choice, suggesting a look-elsewhere correction would further reduce significance.

5. Local-Variance Asymmetry ($A_{LV}$):

   • Common Mask: Yields moderate significance of $\sim 3.2\sigma$.
   • 1% Mask/Full Sky: Significance drops to $\sim 2.2\sigma - 2.3\sigma$ for the foreground-cleaned maps.
   • Conclusion: Similar to $S_{1/2}$, the anomaly is less significant when using the larger sky fraction enabled by improved cleaning.

Significance and Claims
The paper concludes that improved foreground cleaning, which allows for the inclusion of a larger fraction of the sky (99% vs. 74%), reduces the statistical significance of two of the five studied anomalies: the low real-space correlation function ($S_{1/2}$) and the local-variance asymmetry ($A_{LV}$). For these two features, the significance drops from $\sim 3\sigma$ (with the common mask) to $\sim 2\sigma$ (with the 1% mask).

For the remaining three anomalies (Parity Asymmetry, Quadrupole-Octopole Alignment, and Northern Variance), the results remain qualitatively consistent with previous studies, confirming their moderate significance ($\sim 2\sigma - 3\sigma$) even with the larger sky fraction.

The authors assert that these findings suggest the CMB anomalies are not primarily caused by galactic foreground contamination, as the anomalies persist (though with varying significance) even with cleaner maps. However, the choice of the galactic mask significantly influences the measured significance of specific statistics.

Ultimately, the paper argues that the individual significance of these anomalies is insufficient to strongly favor alternative physical models over the standard $\Lambda$CDM model with statistically isotropic Gaussian fluctuations. For an alternative model to be convincingly favored, it would need to simultaneously explain multiple anomalies or provide a better description of other independent measurements. The authors also emphasize that the a posteriori nature of these tests implies that a rigorous look-elsewhere correction would likely further diminish the significance of individual anomalies.