A consistency test of the cosmological model at the epoch of recombination using

DESI BAO and Planck measurements

Levon Pogosian,1 Gong-Bo Zhao,2, 3, 4 and Karsten Jedamzik5
1
Department of Physics, Simon Fraser University, Burnaby, British Columbia, Canada V5A 1S6∗
2
National Astronomical Observatories, Chinese Academy of Sciences, Beijing, 100101, P.R.China
3
University of Chinese Academy of Sciences, Beijing, 100049, P.R.China
4
Institute for Frontiers in Astronomy and Astrophysics,
Beijing Normal University, Beijing, 102206, P.R.China†
5
Laboratoire de Univers et Particules de Montpellier, UMR5299-CNRS,
Universite de Montpellier, 34095 Montpellier, France‡
arXiv:2405.20306v1 [astro-ph.CO] 30 May 2024

The value of the Hubble constant determined from CMB and BAO measurements is directly
dependent on the sound horizon at the photon-baryon decoupling. There has been significant interest
in the possibility of new physics at the epoch around recombination that could reduce the sound
horizon and increase the inferred value of H0 , thus helping to relieve the Hubble tension. One way to
determine if new physics is required would be to measure H0 from BAO and CMB without assuming
any model for computing the sound horizon. In this study, we use the recently released DESI Year
1 BAO data combined with the CMB acoustic scale θ⋆ and the Planck ΛCDM prior on Ωm h2 to
determine H0 while treating the sound horizon at baryon decoupling rd as a free parameter. We
find H0 = 69.88 ± 0.93 km/s/Mpc, which is ∼ 2σ larger than H0 = 67.44 ± 0.47 km/s/Mpc in the
Planck-best-fit ΛCDM where rd is derived using the standard recombination model. For comparison,
we perform the same analysis using the pre-DESI BAO data with θ⋆ and the same prior on Ωm h2 ,
finding H0 = 67.37 ± 0.96 km/s/Mpc. This difference derives from the notably larger value of the
product rd h measured by DESI. Future BAO data from DESI will help determine if the cosmological
model at the epoch of recombination model requires a modification.

I. INTRODUCTION a shorter comoving distance to the epoch of decoupling

The Λ Cold Dark Matter (ΛCDM) model has been
very successful in describing the detailed spectrum of
the Cosmic Microwave Background (CMB) anisotropies
and measures of cosmic structure formation with just a
few cosmological parameters. Notwithstanding this suc-
cess, the value of the Hubble constant derived from the
ΛCDM best fit to the Planck CMB data, H0 = 67.4 ± 0.5
km/s/Mpc [1], is lower than that obtained from “di-
rect” measurements of the present-day expansion rate,
most notably the H0 = 73.0 ± 1.0 km/s/Mpc value ob-
tained by the Supernovae H0 for the Equation of State
(SH0ES) collaboration using Cepheid-calibrated super-
novae (SN) [2]. SN-based analyses using alternative cali-
bration methods [3] also yield higher values of H0 , albeit
at a lower level of tension with CMB.
The value of H0 derived from CMB depends on the
comoving sound horizon at photon decoupling, r⋆ , which
sets the physical scale for the acoustic features imprinted
in the anisotropies. This is closely related to the sound
horizon at the baryon decoupling, rd , that sets the char-
acteristic scale of Baryon Acoustic Oscillations (BAO) in
the distribution of large scale structure. Both CMB and
BAO measure the angular size of the acoustic scale at
the respective redshifts, and a smaller rd would imply
∗ levon@sfu.ca
† gbzhao@nao.cas.cn
‡ karsten.jedamzik@umontpellier.fr
and, thus, a larger H0 .
Results from various independent ways of determining
H0 [4] support an interesting trend that measurements
which do not rely on a model of recombination yield val-
ues of H0 in the 69 − 74 km/s/Mpc range [5–9], while
estimates that use the standard treatment of recombi-
nation give H0 around 67 − 68 km/s/Mpc [10–12]. The
trend suggests that there could be a missing ingredient in
our current understanding of the universe prior and/or
during the epoch of recombination – an ingredient that
could affect the value of the sound horizon at the epoch
of photon-baryon decoupling. This motivates identifying
combinations of cosmological datasets that can test the
validity of the standard model for computing the sound
horizon at decoupling.
Assuming the post-recombination expansion history is
described by the flat ΛCDM cosmology, measuring the
angular sizes of the acoustic imprints in the distribu-
tion of galaxies at different redshifts allows one to deter-
mine the product rd H0 and the current fraction of non-
relativistic matter, Ωm , without making any assumptions
about the recombination epoch physics. To break the de-
generacy between rd and H0 , further assumptions have
to be made. A number of studies (e.g. [13]), including the
recent analysis of the Dark Energy Spectroscopic Instru-
ment (DESI) Year 1 data [14], combined the BAO data
with the Big Bang Nucleosynthesis (BBN) prior on the
baryonic density parameter, Ωb h2 , which was then used
in the standard recombination model to compute rd and
determine H0 . Although interesting as a consistency test
of some aspects of the ΛCDM model, this approach does

not allow for the possibility of model extensions that pre-
dict smaller values of rd . While such extensions may be
unable to fully resolve the Hubble tension [15], they can
significantly reduce its statistical significance by elevat-
ing the Hubble constant derived from the combination
of CMB and BAO to H0 ∼ 70 km/s/Mpc. It is, there-
fore, important to find ways to determine rd and H0 from
BAO without assuming a model for computing rd .
In [16], we proposed treating rd and H0 as indepen-
dent parameters and determine their values from BAO
combined with a Gaussian prior on the physical matter
density parameter Ωm h2 . Such an approach is justified
for two reasons: a) in many extensions of the standard
model that succeed in reducing the sound horizon, while
maintaining a good fit to the CMB spectra, the best-fit
value of Ωm h2 is very close to that of ΛCDM1 ; b) one can
check if the values of H0 and rd derived from BAO with
the Planck prior on Ωm h2 agree with the Planck-best-fit
ΛCDM values obtained assuming standard recombina-
tion, thus providing a consistency test of the model used
to compute rd .
Other determinations of H0 without using a recom-
bination model explored the dependence of the mat-
ter power spectrum on the scale set by the horizon
at radiation-matter equality. These included combining
CMB lensing with uncalibrated supernovae (which pro-
vide a measurement of Ωm ) [24], combining the matter
power spectrum with a prior on Ωm from supernovae [25],
and combining BAO with the CMB and galaxy weak lens-
ing data [16]. Another possibility is to combine BAO with
the cosmic chronometers data [16, 26]. The uncertainty
in H0 from these measurements is still sufficiently large
to make them consistent with both Planck and SH0ES
values.
In this paper, we use the DESI Year 1 BAO data [14]
along with the Planck measurement of the CMB acous-
tic scale θ⋆ and the Planck prior on Ωm h2 to determine
the values for rd and H0 . We find H0 = 69.88 ± 0.93
km/s/Mpc, which is ∼ 2σ larger than the Planck-best-
fit ΛCDM value of 67.44 ± 0.47. For comparison, we also
perform the same analysis using the pre-DESI BAO data,
finding H0 = 67.37 ± 0.96 km/s/Mpc. This difference de-
rives from the notably larger value of the product rd h
measured by DESI, 102.33 ± 1.27 Mpc, compared to the
pre-DESI BAO value of rd h = 99.98 ± 1.21 Mpc.
In what follows, we describe our methods in Section II
before presenting the results in Section III.
II. METHODS
There are three types of BAO observables [27]: the an-
gular size of the acoustic feature measured using correla-
1 One notable exception is the class of Early Dark Energy (EDE)
models [17], that require a larger Ωm h2 , which tends to worsen
the agreement with the galaxy weak lensing data [18–23].
2
tions in the direction perpendicular to the line of sight,
β⊥ (z) = DM (z)/rd , (1)
the same feature measured in the direction parallel to
the line of sight, and the angle-averaged or “isotropic”
measurement. We explain our method using β⊥ as our
example, but the same arguments also apply to the other
two BAO observables. Assuming a flat ΛCDM cosmology
and ignoring the radiation density2 , which is negligible at
redshifts of interest, β⊥ measured at a given redshift z
can be written as
Z z
2998 Mpc dz ′
β⊥ (z) = p , (2)
′ 3
0 rd h Ωm (1 + z ) + 1 − Ωm
where h = H0 /(100km/s/Mpc). From Eq. (2) it is clear
that having BAO measurements at multiple redshifts al-
lows one to measure two numbers: rd h and Ωm . The de-
generacy between rd and h can be broken with a prior on
any combination of Ωm and h, such as a prior on Ωm h2 .
To derive our constraints on rd and H0 , we use
CosmoMC [28] modified to work with rd as an independent
parameter. The cosmological parameters we vary are rd ,
H0 and Ωm h2 .
We use the recently published DESI Year 1 (DESI
Y1) BAO data [29], including the BAO distance mea-
surements derived from the DESI Bright Galaxy Sample
(BGS, 0.1 < z < 0.4) [30], Luminous Red Galaxy Sample
(LRG, 0.4 < z < 1.1) [31], Emission Line Galaxy Sam-
ple (ELG, 1.1 < z < 1.6) [32], Quasar Sample (QSO,
0.8 < z < 2.1) [33] and the Lyman-α Forest Sample
(Lyα, 1.77 < z < 4.16) [34]. Obtaining measurements
of different types and over a broad range of redshifts is
important, as the intersection of the degenerate bands
in the Ωm -rd H0 plane corresponding to individual BAO
data points yields a precise determination of both quan-
tities.
The angular size of the CMB acoustic scale,
θ⋆ = r⋆ /DM (z⋆ ) , (3)
where z⋆ is the redshift of the photon decoupling, can be
⋆
“converted” into a perpendicular BAO observable β⊥ as
⋆ r⋆
β⊥ = . (4)
θ⋆ rd
⋆
We derive β⊥ using 100θ⋆ = 1.04104 ± 0.00025, r⋆ =
144.73 ± 0.23 Mpc and rd = 147.45 ± 0.23 Mpc mea-
sured by fitting ΛCDM to the combination of Planck
Hillipop, Lollipop, low-ell TT and CMB lensing. This
gives rd /r⋆ = 1.0188 ± 0.0024. As the uncertainty in
rd /r⋆ is an order of magnitude larger than that in θ⋆ ,
⋆
it dominates the uncertainty in β⊥ , giving
⋆
β⊥ = 94.286 ± 0.217 , (5)
2 Radiation density is included in our numerical analysis.
 3

which we use as our “BAO” data point at z⋆ = 1090.

Note that the comoving distance DM (z) asymptotes to
a constant value at high redshifts and, for all practical
purposes, is not sensitive to the precise value of z⋆ . The
ratio rd /r⋆ is also largely the same across different ex-
tensions of ΛCDM model3 . Thus, we can use β⊥ ⋆
as a
“model-independent” BAO measurement at z⋆ .
We supplement our BAO dataset with a Gaussian prior
on Ωm h2 based on the ΛCDM value measured by Planck,
Ωm h2 = 0.142 ± 0.001.
In addition, we also perform our analysis using the
BAO data that was available prior to DESI. Our pre-
DESI BAO dataset, denoted as SDSS+ , includes the
Date Release (DR) 16 of the extended Baryon Oscillation
Spectroscopic Survey (eBOSS) [12] BAO measurement
from the full-shape auto- and cross-power spectrum of
the eBOSS LRGs and ELGs [36, 37], the BAO measure-
ment from the QSO sample [38], and from the Lyman-α
forest sample [39], combined with the low-z BAO mea-
surements by 6dF [40] and the SDSS DR7 main Galaxy FIG. 1. Marginalized 68% and 95% confidence level CL con-
sample (MGS) [41]. tours for H0 and rd from the SDSS+ and DESI Y1 BAO data,
and the two BAO datasets combined with the CMB acoustic
scale θ⋆ and a prior on Ωm h2 . Shown also are the contours for
III. RESULTS the best fit Planck LCDM model.

Our results are presented in Table I and Figures 1 and

2. In Fig. 1 we compare the constraints on H0 and rd
from the SDSS+ and DESI Y1 BAO data by themselves
and in combination with the CMB acoustic scale θ⋆ and
a prior on Ωm h2 . As explained in Sec. II, BAO data by
itself only constrains the product rd h, corresponding a
band in the rd -H0 plane. One can see that the DESI Y1
band is shifted upwards relative to the SDSS+ band, cor-
responding to a larger value of rd h. As shown in Table I,
while SDSS+ BAO gives rd h = 99.98 ± 1.21 Mpc, in ex-
cellent agreement with the Planck-best-fit ΛCDM value
of 99.44 ± 0.82 Mpc, DESI Y1 yields rd h = 102.33 ± 1.27
Mpc.
Adding the Gaussian prior on Ωm h2 and the CMB
acoustic scale θ⋆ allows one to constrain both H0 and
rd , as also shown in Fig. 1. For SDSS+ , this combination
yields H0 = 67.37±0.96 km/s/Mpc, in perfect agreement
with the Planck-best-fit value of 67.44 ± 0.47 km/s/Mpc,
while the same combination for DESI Y1 gives a value
that is ∼ 2σ higher, H0 = 69.88 ± 0.93 km/s/Mpc.
Further observations can be made from from Table I
and Fig. 2. Without θ⋆ , combining the BAO data with the
Ωm h2 prior gives similar results for SDSS+ and DESI Y1,
69.4 ± 1.8 and 70.1 ± 1.7 km/s/Mpc, respectively. Adding
the θ⋆ data point to DESI Y1 halves the uncertainty in
H0 while leaving the mean value effectively the same. Do-
ing so for SDSS+ shifts the mean value to the lower end of
the 1σ range. This suggests that the DESI Y1 BAO is in a
better consistency with the CMB acoustic scale when one
treats rd as a free parameter, as opposed to computing it
using the standard recombination model. The value of H0
inferred in this way, 69.88 ± 0.93 km/s/Mpc, is at a ∼ 2σ
tension with the Planck-best-fit ΛCDM value, 67.44 ± 47
km/s/Mpc, based on the standard recombination model.
On the other hand, SDSS+ is more consistent with the
Planck-best-fit ΛCDM model with standard recombina-
tion, but in slight tension with the CMB acoustic peaks
if no recombination model assumed.
The differences between the SDSS+ and DESI Y1 BAO
data and the possible reasons for them have been dis-
cussed extensively in [14, 29] and examining them fur-
ther is beyond the scope of this paper. There are notable
differences in the BAO extracted from the Luminous Red
Galaxy samples in the redshift bins centred at z = 0.5
and z = 0.7. Resolving these differences would likely have
to wait until BAO derived from DESI data covering larger
volumes is available.
IV. CONCLUSIONS

3
The rd /r⋆ ratio is found to be effectively the same across many
extensions of ΛCDM that relieve the Hubble tension by reducing
r⋆ , including additional relativistic degrees of freedom, massive
neutrinos, primordial magnetic fields, early dark energy and oth-
ers. We thank Helena Garcı́a Escudero for confirming this for the
models studied in [35].
The Hubble tension and attempts to resolve it by mod-
ifying the physics of (or at) recombination motivate find-
ing ways to determine H0 and the sound horizon at the
epoch of baryon decoupling rd without relying on a re-
combination model or on the SN calibration. We have
analyzed the recent BAO data from DESI Year 1 without
 4

Ωm h2 rd h [Mpc] Ωm rd [Mpc] H0 [km/s/Mpc]

Planck LCDM 0.142 ± 0.001 99.44 ± 0.82 0.3126 ± 0.0064 147.44 ± 0.23 67.44 ± 0.47
SDSS+ BAO - 99.98 ± 1.21 0.295 ± 0.016 - -
SDSS+ BAO + Θ⋆ - 99.28 ± 1.08 0.312 ± 0.009 - -
SDSS+ BAO + Ωm h2 Planck prior 100.01 ± 1.21 0.294 ± 0.015 144.1 ± 2.5 69.4 ± 1.8
SDSS+ BAO + Θ⋆ + Ωm h2 Planck prior 99.35 ± 1.07 0.312 ± 0.009 147.5 ± 0.8 67.37 ± 0.96
DESI Y1 BAO - 102.33 ± 1.27 0.290 ± 0.014 - -
DESI Y1 BAO + Θ⋆ - 102.34 ± 1.03 0.290 ± 0.008 - -
DESI Y1 BAO + Ωm h2 Planck prior 102.46 ± 1.25 0.288 ± 0.014 146.2 ± 2.1 70.1 ± 1.7
DESI Y1 BAO + Θ⋆ + Ωm h2 Planck prior 102.3 ± 1.0 0.290 ± 0.008 146.5 ± 0.8 69.88 ± 0.93

TABLE I. The mean parameter values and 68% CL uncertainties derived from the considered combinations of datasets.

FIG. 2. Left panel: constraints on H0 , rd , rd h and Ωm derived from the DESI Y1 BAO data, from DESI Y1 BAO combined
with a prior on Ωm h2 , from DESI Y1 BAO combined with the CMB acoustic scale θ⋆ and a prior on Ωm h2 , and the best fit
Planck LCDM. Right panel: same analysis as in the right panel using the pre-DESI BAO data.

using a model to compute rd . We confirmed that DESI
Y1 BAO data prefers a larger value of the product rd h
that is in mild tension with the Planck-best-fit ΛCDM
value and the pre-DESI BAO data. Supplementing the
BAO data with a prior on Ωm h2 allows us to measure
H0 and rd while treating the latter as a free parameter,
without assuming a recombination model.
Combining DESI Year 1 BAO with the CMB acous-
tic scale θ⋆ as another ”BAO” measurement, and the
ΛCDM Planck prior on Ωm h2 , we find H0 = 69.88 ± 0.93
km/s/Mpc. This is ∼ 2σ away from both the SH0ES
and Planck H0 values, while in good agreement with the
values of H0 obtained from supernovae calibrated using
alternative methods [3]. SDSS+ BAO is more consistent
with the Planck LCDM model with standard recombina-
tion, but in slight tension with the CMB acoustic scale if
no recombination model is assumed. DESI Y1 is less con-
sistent with Planck LCDM with standard recombination,
but in perfect agreement with the CMB acoustic scale if
no recombination model is assumed.
The value of H0 determined from the 5 year DESI
BAO data with a prior on Ωm h2 and not assuming a
recombination model would have an uncertainty that
is comparable to the uncertainty in the Planck-best-fit
ΛCDM [16]. This will help determine if standard recom-
bination within ΛCDM is the correct model, or if an
amended model with a reduced sound horizon will be
required, with the benefit of relieving the Hubble ten-
sion.
ACKNOWLEDGMENTS
We thank John Peacock for a useful discussion and
Helena Garcı́a Escudero for checking the values of rd /r⋆
in alternative models. We gratefully acknowledge using
GetDist [42]. This research was enabled in part by sup-
port provided by the BC DRI Group and the Digital
Research Alliance of Canada (alliancecan.ca). L.P. is
supported in part by the National Sciences and Engi-
neering Research Council (NSERC) of Canada. G.B.Z. is
supported by the National Key R & D Program of China

(2023YFA1607800, 2023YFA1607803), NSFC grants (No.
11925303 and 11890691), a CAS Project for Young Sci-
entists in Basic Research (No. YSBR-092), a science re-
[1] N. Aghanim et al. (Planck), Astron. Astrophys. 641,
A6 (2020), [Erratum: Astron.Astrophys. 652, C4 (2021)],
arXiv:1807.06209 [astro-ph.CO].
[2] A. G. Riess et al., Astrophys. J. Lett. 934, L7 (2022),
arXiv:2112.04510 [astro-ph.CO].
[3] W. L. Freedman and B. F. Madore, JCAP 11, 050 (2023),
arXiv:2309.05618 [astro-ph.CO].
[4] E. Abdalla et al., JHEAp 34, 49 (2022), arXiv:2203.06142
[astro-ph.CO].
[5] D. Pesce et al., Astrophys. J. 891, L1 (2020),
arXiv:2001.09213 [astro-ph.CO].
[6] K. C. Wong et al., (2019), arXiv:1907.04869 [astro-
ph.CO].
[7] A. J. Shajib et al. (DES), Mon. Not. Roy. Astron. Soc.
494, 6072 (2020), arXiv:1910.06306 [astro-ph.CO].
[8] D. Harvey, (2020), 10.1093/mnras/staa2522,
arXiv:2011.09488 [astro-ph.CO].
[9] M. Millon et al., Astron. Astrophys. 639, A101 (2020),
arXiv:1912.08027 [astro-ph.CO].
[10] M. M. Ivanov, M. Simonović, and M. Zaldarriaga, JCAP
05, 042 (2020), arXiv:1909.05277 [astro-ph.CO].
[11] S. Aiola et al. (ACT), JCAP 12, 047 (2020),
arXiv:2007.07288 [astro-ph.CO].
[12] S. Alam et al. (eBOSS), (2020), arXiv:2007.08991 [astro-
ph.CO].
[13] G. Addison, D. Watts, C. Bennett, M. Halpern, G. Hin-
shaw, and J. Weiland, Astrophys. J. 853, 119 (2018),
arXiv:1707.06547 [astro-ph.CO].
[14] A. G. Adame et al. (DESI), (2024), arXiv:2404.03002
[astro-ph.CO].
[15] K. Jedamzik, L. Pogosian, and G.-B. Zhao, Commun. in
Phys. 4, 123 (2021), arXiv:2010.04158 [astro-ph.CO].
[16] L. Pogosian, G.-B. Zhao, and K. Jedamzik, Astrophys. J.
Lett. 904, L17 (2020), arXiv:2009.08455 [astro-ph.CO].
[17] V. Poulin, T. L. Smith, T. Karwal, and
M. Kamionkowski, Phys. Rev. Lett. 122, 221301
(2019), arXiv:1811.04083 [astro-ph.CO].
[18] J. C. Hill, E. McDonough, M. W. Toomey, and
S. Alexander, Phys. Rev. D 102, 043507 (2020),
arXiv:2003.07355 [astro-ph.CO].
[19] M. M. Ivanov, E. McDonough, J. C. Hill, M. Simonović,
M. W. Toomey, S. Alexander, and M. Zaldarriaga,
(2020), arXiv:2006.11235 [astro-ph.CO].
[20] G. D’Amico, L. Senatore, P. Zhang, and H. Zheng,
(2020), arXiv:2006.12420 [astro-ph.CO].
5
search grant from the China Manned Space Project with
No. CMS-CSST-2021-B01, and by the New Cornerstone
Science Foundation through the XPLORER prize.
[21] R. Murgia, G. F. Abellán, and V. Poulin, (2020),
arXiv:2009.10733 [astro-ph.CO].
[22] T. L. Smith, V. Poulin, J. L. Bernal, K. K.
Boddy, M. Kamionkowski, and R. Murgia, (2020),
arXiv:2009.10740 [astro-ph.CO].
[23] F. Niedermann and M. S. Sloth, Phys. Rev. D 103,
103537 (2021), arXiv:2009.00006 [astro-ph.CO].
[24] E. J. Baxter and B. D. Sherwin, (2020),
arXiv:2007.04007 [astro-ph.CO].
[25] O. H. E. Philcox, B. D. Sherwin, G. S. Farren, and
E. J. Baxter, Phys. Rev. D 103, 023538 (2021),
arXiv:2008.08084 [astro-ph.CO].
[26] W. Lin, X. Chen, and K. J. Mack, Astrophys. J. 920,
159 (2021), arXiv:2102.05701 [astro-ph.CO].
[27] D. J. Eisenstein et al. (SDSS), Astrophys. J. 633, 560
(2005), arXiv:astro-ph/0501171.
[28] A. Lewis and S. Bridle, Phys. Rev. D 66, 103511 (2002),
arXiv:astro-ph/0205436 [astro-ph].
[29] A. G. Adame et al. (DESI), (2024), arXiv:2404.03000
[astro-ph.CO].
[30] Schlegel et al., in preparation (2024).
[31] R. Zhou et al. (DESI), Astron. J. 165, 58 (2023),
arXiv:2208.08515 [astro-ph.CO].
[32] A. Raichoor et al., Astron. J. 165, 126 (2023),
arXiv:2208.08513 [astro-ph.CO].
[33] E. Chaussidon et al., Astrophys. J. 944, 107 (2023),
arXiv:2208.08511 [astro-ph.CO].
[34] Bault et al., in preparation (2024).
[35] H. G. Escudero, J.-L. Kuo, R. E. Keeley, and
K. N. Abazajian, Phys. Rev. D 106, 103517 (2022),
arXiv:2208.14435 [astro-ph.CO].
[36] G.-B. Zhao et al., (2020), arXiv:2007.09011 [astro-
ph.CO].
[37] Y. Wang et al., (2020), 10.1093/mnras/staa2593,
arXiv:2007.09010 [astro-ph.CO].
[38] J. Hou et al., (2020), arXiv:2007.08998 [astro-ph.CO].
[39] H. du Mas des Bourboux et al., (2020), arXiv:2007.08995
[astro-ph.CO].
[40] F. Beutler, C. Blake, M. Colless, D. Jones, L. Staveley-
Smith, L. Campbell, Q. Parker, W. Saunders, and
F. Watson, Mon. Not. Roy. Astron. Soc. 416, 3017
(2011), arXiv:1106.3366 [astro-ph.CO].
[41] A. J. Ross, L. Samushia, C. Howlett, W. J. Percival,
A. Burden, and M. Manera, Mon. Not. Roy. Astron.
Soc. 449, 835 (2015), arXiv:1409.3242 [astro-ph.CO].
[42] A. Lewis, (2019), arXiv:1910.13970 [astro-ph.IM].