Detecting Neutrino Mass by Combining Matter Clustering, Halos, and Voids

Abstract

We quantify the information content of the nonlinear matter power spectrum, the halo mass function, and the void size function, using the Quijote N-body simulations. We find that these three statistics exhibit very different degeneracies among the cosmological parameters, and thus the combination of all three probes enables the breaking of degeneracies, in turn yielding remarkably tight constraints. We perform a Fisher analysis using the full covariance matrix, including all auto- and cross correlations, finding that this increases the information content for neutrino mass compared to a correlation-free analysis. The multiplicative improvement of the constraints on the cosmological parameters obtained by combining all three probes compared to using the power spectrum alone are: 137, 5, 8, 20, 10, and 43, for Ω m , Ω b , h, n s , σ 8, and M ν , respectively. The marginalized error on the sum of the neutrino masses is σ(M ν ) = 0.018 eV for a cosmological volume of 1 h − 1 Gpc 3 , using k max = 0.5 h Mpc − 1 , and without cosmic microwave background (CMB) priors. We note that this error is an underestimate insomuch as we do not consider super-sample covariance, baryonic effects, and realistic survey noises and systematics. On the other hand, it is an overestimate insomuch as our cuts and binning are suboptimal due to restrictions imposed by the simulation resolution. Given upcoming galaxy surveys will observe volumes spanning ∼ 100 h − 1 Gpc 3 , this presents a promising new avenue to measure neutrino mass without being restricted by the need for accurate knowledge of the optical depth, which is required for CMB-based measurements. Furthermore, the improved constraints on other cosmological parameters, notably Ω m , may also be competitive with CMB-based measurements.

Cosmology with One Galaxy?

“…However, we do not know what estimator would allow us to extract the maximum amount of information for non-Gaussian density fields, such as the matter and galaxy distribution on nonlinear scales. Quantifying the information content from different estimators is currently a very active area of research (Coulton et al 2019;Li et al 2019;Marques et al 2019;Ajani et al 2020;Allys et al 2020;Banerjee et al 2020;Dai et al 2020;de la Bella et al 2021;Friedrich et al 2020;Giri & Smith 2020;Hahn et al 2020;Villaescusa-Navarro et al 2020;Uhlemann et al 2020;Zhang et al 2020;Banerjee & Abel 2021aBayer et al 2021aBayer et al , 2021bCheng and Ménard 2021;Gualdi et al 2021aGualdi et al , 2021bHahn & Villaescusa-Navarro 2021;Harnois-Déraps et al 2021;Kuruvilla 2021;Kuruvilla & Aghanim 2021;Massara et al 2021;Naidoo et al 2021;Porth et al 2021;Samushia et al 2021;Valogiannis & Dvorkin 2021;Harnois-Déraps et al 2022).…”

Multifield Cosmology with Artificial Intelligence

“…Numerous studies have shown that extracting information using the power spectrum in this regime is suboptimal (e.g. [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19].…”

Cosmology with cosmic web environments I. Real-space power spectra

“…Recent analyses with alternative approaches like the marked power spectrum and the wavelet scattering transform (hereafter WST) were employed on the very same sets of simulations and of cosmological parameters in Massara et al (2021) and Valogiannis & Dvorkin (2021) respectively. Alternatively, Bayer et al (2021) propose an analysis using the information from the halo mass function (HMF) and the void size function (VSF) in real space. All three analyses, together with ours, consider a maximum scale of k max = 0.5 h/Mpc and can hence be naturally compared.…”

“…The information contained in the power spectra bins in cosmic web environments provides the tightest constraints on Ω m , Ω b , h, and n s , and competitive values for constraints on M ν and σ 8 which are however both best constrained by the coefficients derived from the WST. Compared to the partial information of extreme environments from voids and nodes performed in Bayer et al (2021), our environment-dependent statistics are able to tighten the constraints with improvement factors of {0.5, 4.1, 2.9, 3.1, 1.4, 1.7} on the parameters {Ω m , Ω b , h, n s , σ 8 , M ν }, hence performing better than the VSF+HMF statistic in all cases, except for Ω m . These comparisons should however be nuanced in multiple aspects.…”

“…The measure of voids' clustering, sizes, shapes, counts, bias and corresponding evolution with redshift are hence key quantities probing the underlying cosmological model (van de Weygaert & Platen 2011;Hamaus et al 2014Hamaus et al , 2015Massara et al 2015;Schuster et al 2019;Kreisch et al 2019). The constraints brought by these two extreme environments that are voids and nodes are for instance combined in Bayer et al (2021) and Kreisch et al (2021) showing that the information provided by the halo mass function and the void size function leads to considerable improvement over the matter power spectrum constraints in real space.…”