Clustering properties of the CatWISE2020 quasar catalogue and their impact on the cosmic dipole anomaly

TL;DR

The paper tackles the cosmic dipole anomaly by reanalyzing the CatWISE2020 quasar catalogue with a multipole-aware Bayesian framework that explicitly models low-$\ell$ power and higher-multipole contamination. It develops a parameterization that links dipole, quadrupole, and octupole templates to angular power spectra, incorporates a non-Poissonian shot-noise model, and infers these components using No U-Turn Hamiltonian Monte Carlo with Bayesian model comparison. The results show a statistically robust ecliptic latitude trend and a preference for super-Poissonian statistics, but no decisive octupole detection; the dipole remains anomalously large, while the clustering dipole predicted by $\Lambda$CDM is subdominant. On small scales the power spectrum agrees with $\Lambda$CDM once a minor effective noise term is included, yet the large-scale dipole is not explained by clustering, implying a potential breakdown of isotropy at the largest scales. Together, these findings reinforce the cosmic dipole anomaly while constraining the role of low-level multipole power, with significant implications for the Cosmological Principle and future surveys.

Abstract

The cosmic dipole anomaly -- the observation of a significant mismatch between the dipole observed in the matter distribution and that expected given the kinematic interpretation of the cosmic microwave background dipole -- poses a serious challenge to the Cosmological Principle upon which the standard model of cosmology rests. Measurements of the dipole ($\ell=1$) in a given sample crucially depend on having control over other large-scale power ($\ell > 1$) so as to avoid biases, in particular those potentially caused by correlations among multipoles during fitting, and those by local source clustering. Currently, the most powerful catalogue that exhibits the cosmic dipole anomaly is the sample of 1.6~million mid-infrared quasars derived from CatWISE2020. We therefore analyse clustering properties of this catalogue by performing an inference analysis of large-scale multipoles in real space, and by computing its angular power spectrum on small scales to test for convergence with $Λ$CDM. After accounting for the known trend of the quasar number counts with ecliptic latitude, we find that any other large-scale power is consistent with noise, find no evidence for the presence of an octupole ($\ell=3$) in the data, and quantify the clustering dipole's proportion to be marginal. Our results therefore reaffirm the anomalously high dipole in the distribution of quasars.

Figures (12)

• Figure 1: Empirical one-point distribution function (EDF) of the CatWISE2020 quasar number densities (black) after correction for the ecliptic latitude trend (\ref{['eq:selection_ecl']}) and removal of the best-fit dipole. Top panel: Comparison of the EDF against a Poisson (dotted black), Generalised Poisson (blue), and negative-binomial (or Gamma Poisson) distribution (dashed red). The grey-shaded error band indicates Poisson errors per bin. Bottom panel: Difference of the respective distributions from the EDF normalised to the error per bin, indicating that both the super-Poissonian distributions provide excellent fits to the data, as opposed to the strongly deviant Poisson distribution.
• Figure 2: Posteriors of all parameters in model (ii) (${\sf DY}_{\sf ecl}$, see Table \ref{['tab:models']}) inferred using the Poisson and the Generalised Poisson likelihood, Eqs. (\ref{['eq:lnl_poisson']}) and (\ref{['eq:lnl_genpoisson']}). The dashed lines indicate the expected amplitude and direction for the kinematic dipole $\mathcal{D}_{\rm kin}$.
• Figure 3: Posteriors of select parameters in models (v) and (ii) (${\sf DY}_{\sf ecl}{\sf O}$ and ${\sf DY}_{\sf ecl}$, see Table \ref{['tab:models']}) inferred using the Generalised Poisson likelihood (Eq. (\ref{['eq:lnl_genpoisson']})). The posteriors are marginalised over the octupole angles and the ecliptic latitude trend's amplitude. The dipole amplitude remains anomalous (the dashed lines indicate the expected amplitude and direction for the kinematic dipole $\mathcal{D}_{\rm kin}$.) even after inclusion of a general octupole in the inference.
• Figure 4: The power spectrum of CatWISE2020 quasars from Secrest:2022uvx after correcting for the ecliptic latitude trend. Small markers denote the power computed using namaster (with $1\sigma$ errors derived from jackknife sampling after shot noise subtraction), while the large markers are from a template fit for multipoles $\ell \leq 4$ (model (vii), Table \ref{['tab:models']} using a Generalised Poisson likelihood, Eq. (\ref{['eq:lnl_genpoisson']})). The small gray markers are shown for comparison with other works, but should not be trusted, and are not included in the fit. The solid blue line is a $\Lambda$CDM matter power spectrum fitted to data in the range $10<\ell<1000$ and the blue shaded band indicates cosmic variance. Dashed and dotted red lines denote, respectively, indicate the level of shot noise and additional "noise" due to errors in quasar source finding (e.g. due to substructure, as discussed in Section \ref{['sec:lowmultipoles_onepoint']}).
• Figure 5: Stability of the dipole amplitude $\mathcal{D}$ across a range of cuts on flux density (our fiducial sample, Secrest:2022uvx, is listed as ${\rm W1}<16.5$). Posteriors of the dipole parameters are shown, using model (ii) (${\sf DY}_{\sf ecl}$) and the Generalised Poisson likelihood, Eq. (\ref{['eq:lnl_genpoisson']}). Dashed lines indicate the expected amplitude and direction for the kinematic dipole $\mathcal{D}_{\rm kin}$, which is roughly constant across cuts. The inset shows the evolution of the amplitude of the ecliptic latitude trend, $\mathcal{Y}_{\rm ecl}$, and the overdispersion parameter, $b$, of the generalised Poisson likelihood, demonstrating that both the ecliptic latitude trend as well as the overdispersion vanish towards brighter magnitudes and hence higher signal-to-noise.