Abstract
Next generation surveys will be capable of determining cosmological parameters beyond percent level. To match this precision, theoretical descriptions should look beyond the linear perturbations to approximate the observables in large scale structure. A quantity of interest is the Number density of galaxies detected by our instruments. This has been focus of interest recently, and several efforts have been made to explain relativistic effects theoretically, thereby testing the full theory. However, the results at nonlinear level from previous works are in disagreement. We present a new and independent approach to computing the relativistic galaxy number counts to second order in cosmological perturbation theory. We derive analytical expressions for the full second order relativistic observed redshift, for the angular diameter distance and for the volume spanned by a survey. Finally, we compare our results with previous works which compute the general distance-redshift relation, finding that our result is in agreement at linear order.
Export citation and abstract BibTeX RIS
Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
1. Introduction
Recent years have witnessed the beginning of the era of precision cosmology with future surveys such as BOSS [1], eBOSS [2], Euclid [3], MeerKAT, SKA, LSST [4–6],and WFIRST [7] improving and tightening constraints on observable cosmological parameters. Additionally, great theoretical advancements have been made in tackling nonlinear regimes to test cosmological models and general relativity.
For theoretical cosmological models, probing the relation between redshift and angular diameter or luminosity distance of a source is of significant value. This relation determines the parameters of the cosmological model, but when perturbations due to structure are included, new effects are revealed. One of these effects is lensing, which is observed along the line of sight. Another effect is the distortion in redshift space due to velocities and motion of the sources, giving rise to 'Doppler lensing'. The integrated Sachs–Wolfe (ISW) effect which arises from integrating along the full line of sight between the source and the observer.
Most of the known effects on the distance-redshift relation are calculated at linear order in cosmological perturbation theory (CPT) in references [8–12]. However, at second order other general relativistic effects must be considered. When structure is evolving, nonlinear modes come into play, and many of these go beyond Newtonian theory.
One of the main observables directly affected by the angular and luminosity distance estimation is the galaxy number density (dubbed often as number counts). Important examples of these effects have been calculated in references [9, 13–21]. The dominating terms of the full second order calculations have been reviewed in reference [22]. More recently in reference [23] the authors present second order relativistic corrections to the observable redshift. And even a 'pedagogical' approach to the lengthy calculations is provided in reference [24] to try to ease the tension between the different groups.
In this work, we present a new path to compute the second-order galaxy number in a Friedmann–Lemaître–Robertson–Walker (FLRW) universe. This follows the volume determination as defined in reference [25] instead of computing the luminosity distance as in reference [17]. We identify key effects, some of which will be observable with the next generation of cosmological surveys. To check the robustness of our results we confirm the consistency for the first order expressions with previous works.
This paper is organized as follows: in section 2 we provide all the definitions needed for the linear and nonlinear calculations in the context of CPT. In section 3 we compute the linear and nonlinear parts of the null geodesic equation, the observed redshift, and show the geometrical effects present at this level. In section 4 we compute the angular diameter distance and the physical volume that the galaxy survey spans, only in this section we make a conformal transformation that maps null geodesics from the perturbed FLRW metric to a perturbed Minkowski spacetime , how quantities transform under this map is discussed in further detail within this section. In section 5 we compute our main result, the galaxy number overdensity. In section 6 we make a check for the calculation performed in this paper with other results in the literature at linear order and find an exact agreement with all of them pertaining the right interpretation of variables. Finally, in section 7 we give a discussion of our result, some conclusions and future work.
Notation. We use indices μ, ν, ⋯ = 0, 1, 2, 3 in a general spacetime. In perturbed FLRW, the indices i, j, ⋯ = 1, 2, 3 denote spatial components. The derivative with respect to the conformal time is given by a dash
We use the notation
The derivative with respect to the affine parameter is
where ni represents the direction of observation. This last equation implies, for a scalar function X,
where ∇i X = ∂i X = X,i is the spatial part of the covariant derivative.
2. Basis for the definition of the galaxy number density
2.1. Metric perturbations
The perturbed FLRW spacetime is described in the longitudinal gauge by [26]
where η is the conformal time, a = a(η) is the scale factor and δij is the flat spatial metric, and we have neglected the vector and tensor modes, we also allow for first and second order anisotropic stresses. From now on we consider perturbations around an FLRW metric up to second-order.
2.2. Matter velocity field and peculiar velocities
The components of the four-velocity uμ = dxμ /dη up to and including second order using the perturbed metric are given by
where vi = ∂i v, with v the velocity potential.
2.3. Photon wavevector
In a redshift survey galaxy positions are identified by measuring photons produced at the source, denoted by s, and detected by an observer labelled o. In a general spacetime, we consider a lightray with tangent vector kμ and affine parameter λ, that parametrises the curve the lightray follows. The source, λs, and the observer, λo, are represented by given values for the affine parameter, as illustrated in figure 1. The components of the photon wavevector can be written as
where the overbar denotes background quantities, ni is the direction of observation 3 pointing from the observer to the source, and following the normalisation condition: ni ni = 1.
The tangent vector is null
and geodesic
where ∇ν is the covariant derivative defined by the metric given in equation (2.1). In general, the perturbed wavevector can be written as
where δ(n) gives the nth order perturbation, and we are following the usual notation for the temporal component, that is k0 ≡ ν [24].
The affine parameter of the geodesic equation is also related to the comoving distance (χ) by
and in terms of the redshift this is
2.4. Observed redshift
The photon energy measured by an observer with four-velocity uμ is
From equation (2.12) the observed redshift of a source (e.g. a galaxy) can be defined as
From this definition there will be a Doppler effect on the redshift due to the velocities uμ and the observed redshift is in fact a function of the velocity and the wavevector, i.e. z = z(kμ , uμ ).
2.5. Angular diameter distance
For a given bundle of lightrays leaving a source, the bundle will invariantly expand and create an area in between the lightrays that conform it, this area can be projected to a screen space, perpendicular to the trajectories of the photons and the four-velocity of the observer, as illustrated in figures 2 and 3.
Download figure:
Standard image High-resolution imageDownload figure:
Standard image High-resolution imageThe area of a bundle in screen space, , defines the angular diameter distance dA , and is directly related to the null expansion θ [27] defined in section 4,
where λ is the affine parameter defined in equation (2.6). Using equation (2.14) we can compute how the area of the bundle changes along the geodesic trajectory that the photons are following from the source towards the observer.
2.6. Physical volume
Number counts relate to the number of sources detected in a bundle of rays, for a small affine parameter displacement λ to λ + dλ at an event P. This corresponds to a physical distance
in the rest frame of a comoving galaxy at said point in space P, if kμ is a tangent vector to the past directed null geodesics (so that kμ uμ > 0).
The cross-sectional area of the bundle is
if the geodesics subtend a solid angle dΩ at the observer, this is shown in figure 4.
Download figure:
Standard image High-resolution imageFrom equations (2.15) and (2.16) the corresponding volume element at a point P in space is (see e.g. [25])
These covariant definitions lead to the expressions we compute in the following sections at first and second order in CPT.
3. Perturbed null geodesics and redshift
3.1. Geodesic equation
Let us now look at solutions to the geodesic equation. First, from equation (2.7) and the normalisation of ni we obtain the null condition of the photon wavevector
For the geodesic equation (2.8), we get the propagation equations for the temporal and spatial perturbations
where are the connection coefficients at any given order (n). Their expansion up to second order is provided in appendix
where equation (3.4) there is a term related to the ISW effect defined below. The expressions for the solution at second order are given in appendix
3.2. Observed redshift
We now expand the photon energy up to second order,
Using equations (2.1)–(2.3), (3.4), (3.5), (C1) and (C2), we find that , and
The perturbations to the photon energy given in equations (3.7) and (3.8) are written explicitly in terms of the metric potentials in appendix
The observed redshift is given by equation (2.13), then up to second order that is
Note that in general, the expansion of the perturbed quotient up to second order is
Using equations (3.11) and (3.12), and ignoring the background redshift, we obtain
From equation (3.13) the redshift of a source s is, at first order,
where the integral runs w.r.t. the affine parameter λ along the line of sight. In equation (3.14) we identify the following elements
- (a)The Doppler redshift, which depends on the difference between the peculiar velocities of the source and the observer
- (b)The Gravitational redshift, which describes the change of energy the photon experiences when it travels from a region with potential Φ1|s to a region with potential Φ1|o
- (c)The ISW, which describes the change in energy when a photon travels through a potential well between the source and the observer, this effect is only non-zero when the gravitational potential evolves during and along the photon's trajectory, so that the energy gained by going down the gravitational potential is not cancelled by the energy lost by climbing out the potential at the other end.
At second order the redshift (C8) can also be decomposed as above
where the nonlinear contribution from squared first order quantities denoted by 'NL' is given by
Here, just as in the linear case, we have
- (a)Nonlinear Doppler redshift,
- (b)Nonlinear gravitational redshift,
- (c)Nonlinear ISW,
Here the second order contribution to the effects described above are evident, plus the product of first order contributions:
- (a)Gravitational redshift squared
- (b)Gravitational redshift × ISW
- (c)Gravitational redshift × Doppler
- (d)Doppler squared
- (e)Doppler × ISW
- (f)Integrated ISW
4. Distance determinations and the observed volume
4.1. Angular diameter distance
To measure the angular diameter distance (dA ) we must define a projector into the screen space perpendicular to the light ray as shown in figure 3. The screen space is orthogonal to the light ray and to the observer four-velocity. In fact the tensor
where gμν is the metric, uμ is the four-velocity and nμ is the four-direction of observation 4 , projects four-vectors onto screen space, as can be seen in figure 2. The projector tensor satisfies the relations
The null expansion, θ, and null shear, Σμν , are optical properties given in terms of the tangent vector kμ by [26, 27, 29]
Here θ describes the rate of expansion of the projected area of a bundle of light rays and Σμν describes its rate of shear illustrated in figure 3. Note that the wavevector can be obtained from a scalar potential (S), i.e. kμ = ∇μ S, and thus there is no null vorticity, that is ω ≡ ∇[μ kν] = 0 [30].
The 'null evolution' is given by the Sachs propagation equations (see e.g. [29] for full derivation)
where Cμρνσ is the Weyl tensor.
Equations (4.5) and (4.6) allow us to compute the angular diameter distance as a parametric function depending only on the affine parameter, λ, in contrast to previous works where the dependency is on the redshift [13–15] or the conformal time [17]. The advantages of maintaining this dependency are discussed in section 7.
From equations (2.14) and (4.5) we obtain a second order differential equation for the angular diameter distance,
We require appropriate initial conditions to solve (4.7). These can be found from the series expansion of the squared distance given by Kristian and Sachs in [31]:
from where we obtain the boundary conditions at the observer
In this section we will define a conformal metric , useful to compute the angular diameter distance. In our notation, a hat () denotes quantities on the physical spacetime, while quantities on the conformal spacetime have no hat. The background of the metric gμν is Minkowski spacetime, which simplifies both the equations and the calculations. Conformal maps preserve both angles and shapes of infinitesimally small figures, but not their overall size [32]. The conformal transformation maps the null geodesic equation of the perturbed FLRW metric to a null geodesic on the perturbed Minkowski metric gμν [33] and the angular diameter distance transforms as The affine parameter transforms as so that the photon ray vector transforms as [34]. For the four-velocity we have Finally, the energy transforms as In Minkowski spacetime we normalize .
Hereafter, and until the end of this section, we will be working in a perturbed Minkowski spacetime, in order to finally conformally transform our result back to an FLRW spacetime.
In the Minkowski background, equation (4.7) simplifies to
since and the shear vanish in the background. The solution is then
The initial conditions given in equation (4.9) yield C1 = 0 and C2 = −1, so that
Mapping this into the FLRW background we obtain, for the angular diameter distance,
which can be expressed in terms of the comoving distance (2.10) as
Here we have used the definition of the scale factor and the fact that the comoving distance depends on the redshift as given in equation (2.11).
In general, at first order equation (4.7) takes the form
where we use take the background equivalence between the affine parameter and the conformal time dλ = dη, so that
This relation is only fulfiled in the background. Once perturbations are introduced the relation between the affine parameter and time becomes non-trivial.
In Minkowski spacetime, equation (4.15) simplifies to
where we used the background solution for dA
(4.12), the first order perturbation of the Ricci tensor δ(1)
Rμν
given in appendix
The solution to (4.17) is, upon several integrations by parts,
From equation (4.9), in general
In the absence of anisotropic stress, Φ1 = Ψ1 (see, e.g. reference [26]), we recover in equation (4.18) the fully relativistic lensing convergence, usually denoted as κ [9, 17, 30], at first order, which includes Sachs–Wolfe, ISW and Doppler terms in addition to the standard lensing integral
At second order, equation (4.7) takes the form
where δ(1)Σμν is the linear perturbation to the shear. Using equation (4.4) we obtain
Without loss of generality, we set the perturbation of the shear at the observer δ(1)Σμν |o = 0, and integrating along the line of sight from the observer to the source (λo to λs), we obtain
The contraction δ(1)Σij δ(1)Σij is given by
In equation (C9) we find the second order part of the angular diameter distance, using the background solution for , and the full expression δ(2)
Rμν
, the second order perturbation of the Ricci tensor given in appendix
Thus, the total area distance as a function of the affine parameter in a perturbed FLRW spacetime is given by
where the solutions for , δ(1) dA (λs) and δ(2) dA (λs) are given in equations (4.12), (4.18) and (C9), respectively. From here onwards, we abandon the conformal Minkowski spacetime and return to FLRW.
4.2. Area distance as a function of observed redshift
In order to compare with previous work done in the literature, we can convert the angular diameter distance in terms of the affine parameter to a function of the observed redshift. To do so, we need to perturbatively invert z(λ) into λ(z) and substitute this into equation (4.25). This means we need dA on surfaces of constant observed redshift z rather than on surfaces of constant affine parameter λ, which is not observable.
We expand the affine parameter in perturbation theory as
where ς is the affine parameter in redshift space corresponding to the redshift , as if there were no perturbations [14]. We define ς using as an anchor the background relation
with this relation we can fix δ(1) λ and δ(2) λ, since it should always hold, and if there are any perturbations, they should cancel since equation (4.27) is only valid in the background. To begin with, we see that at any redshift , the derivatives of a with respect to ς are
We now expand the scale factor a about λ up to and including second order perturbations as defined in equation (4.26), we have
Using equations (3.11) and (4.30) we find that,
From the background relation given in equations (4.27) and (4.31) we then find that the perturbations to the affine parameter must follow the following relations:
Finally, using these relations to substitute for , we find that the area distance (4.25) becomes
Up until here we have corrected the scale factor from the affine parameter λ to ς. Now we need to convert the first order contributions because they bring additional second order contributions. We introduce that, for a general first order quantity δ(1) X, converting to ς gives
where δ(1) X(ς) is to be understood as substituting ς in the expression for δ(1) X(λ), i.e. δ(1) X(λ → ς). The factor ∂λ δ(1) X|ς is multiplied by a first order quantity, so the derivative is evaluated in the background. Thus, we can write dς δ(1) X. With this, equation (4.34) finally becomes
We can now write the angular diameter distance as a function of the redshift, although it is written in terms of integrals over the comoving distance χ = ςo − ς, which depends on the redshift itself by equation (2.11).
Using equation (4.36) written in terms of observable quantities such as the observable redshift and comoving distance, the angular diameter distance will possibly, in principle, be measured with great accuracy by the upcoming surveys, and should complement to the known luminosity distance measurements quite well.
Combining equations (3.14) and (4.20) with (4.36), we have that at linear order the diameter distance as a function of redshift is given by,
The full expression for in terms of the metric potentials is given in appendix
4.3. Physical volume
The area distance the light-ray bundle creates, changes along the line of sight as seen in figure 2, and we are interested in computing the volume that these hypersurfaces enclose, since therein lie the overdensities we are accounting for.
The volume element (2.17) can be rewritten in terms of the quantities we have computed in the previous sections; the angular diameter distance in equations (4.20) and (C9), and the energy in equations (3.7) and (3.8). It is given up to second order by,
The volume element is given in terms of the affine parameter λ, but we need to express our result in terms of the observed redshift z, and so we need to take the volume in bins of dz instead of dλ. To do so we use the fact that we can write the affine parameter as a function of redshift, i.e. λ(z), and using equations (4.26), (4.32) and (4.33), we obtain the relation
modifying equation (4.38) into
We now give an expression for the volume element order by order. In the background we have
From equation (4.40) and using equations (C5) and (4.37), we have that the first order perturbation to the physical volume is
and using equations (C6) and (C9) we find that the second order perturbation to the physical volume is
The equivalent expression in terms of the metric potentials is given in appendix
5. Galaxy number density
In this section we present our main result, the galaxy number overdensity at second order. As a first element, we take V(ni , z) as the physical survey volume density per redshift bin per solid angle given by (2.17), where ni is the direction of observation and z = z(λs). The volume is a perturbed quantity since the solid angle of observation as well as the redshift bin are distorted between the source and the observer
In equations (4.43) and (C10) we provide the first and second order perturbations to the volume, respectively. Note that we use where other authors in the literature use (see, e.g. reference [9]).
In a galaxy redshift survey, we measure the number of galaxies in direction ni at redshift z. Let us call this N(ni , z)dΩn dz, where dΩn is the solid angle the survey spans. Then one must average over the angles to obtain their redshift distribution, ⟨N⟩(z)dz, where the angle brackets correspond to this angular average [35]
where the integral is over the solid angle the survey spans.
We can then build the matter density perturbation, density contrast, in redshift space, i.e. the perturbation variable [9]
and expand it up to second order as
Our aim in the following is to compute the observed matter density perturbation since the density of sources is proportional to the number of the sources within a given volume, i.e.
and expanding equation (5.5) we show that at any order
The observed quantity is the perturbation in the number density of galaxies, Δ, and it is defined as
and we thus have
In order to compute the above, let us first relate δz (ni , z) to the matter density quantity δ(xi , η) and the perturbations on the redshift computed in section 2. The redshift density up to second order in redshift space is
Structure in the universe is formed from dark matter and baryons which at large scales are modelled by a single pressureless component, which evolves with redshift as
Thus we have that
so using equation (3.14), the redshift density perturbation at first order is given by
Combining (4.43) and (5.13) we find that the galaxy number density fluctuation in redshift space as defined in equation (5.8) is, at first order,
From equation (5.11), we have that the second derivative of the background density is
so using equations (3.14), (C7) and (5.15) in equation (5.10) we find that the redshift density perturbation at second order is given by
where the full expression in terms of the metric potentials is given in appendix
Which is the main result of this work. In the following section we make a comparison of our result with others in the literature [13–15].
6. Comparison with previous work
In this section we compare our linear result given in equation (5.14) with those in the literature given in references [13, 14, 17] which also compute second order corrections and in particular with Dio et al [15]. We do this to verify that our result is correct, since the number counts are well established to linear order with references. [9, 10]. In a companion paper [36] we perform a comparison of the leading terms of the second order expansion of the galaxy number counts.
Our result, as given in equation (5.7) is
6.1. Dio et al
Rewriting the result from reference [15], in Poisson gauge, allowing for anisotropic stress. At first order, reference [15] have
where is the Hubble parameter, the 's' denotes source and the 'o' denotes observer, and
which rewriting in our notation is
Computing the difference between equations (6.2) and (5.14), we have
where the first line cancels out from the definition of the comoving density perturbation (), and the last integral cancels out from the definition of the angular operator Δ2, both given in reference [15], and we find,
6.2. Bertacca, et al
In references [13, 14], their result is written in terms of 'cosmic rulers' and it is given by
which in Poisson gauge, translates into
where we omitted the terms with the evolution bias be. We must rewrite our own result taking Ψ1 = Φ1 in equation (5.14) to make the comparison, so we have that
Computing the difference between equations (6.8) and (5.14), we have
where both are derivatives of first order terms with respect to background quantities, and in the background dχ = dς. Note that the direction in the sky, ni , is from references [13, 14], so
6.3. Yoo & Zaldarriaga
The galaxy overdensity in [17], is given by
which in Poisson gauge, allowing for anisotropic stress takes the form
where we did not use the evolution bias or the running and slope of the luminosity.
The difference between equations (6.13) and (5.14) is then,
where the first line is zero from the definition of , and without loss of generality we take δτo = 0, and the integrals in the second line cancel from the definition of κ in reference [17], so
7. Conclusions & future work
In this paper we have provided a new and independent approach to calculate the galaxy number overdensity. We present the galaxy number counts in a general form depending on the affine parameter which allows for simple plotting along the line of sight if the potentials are known, the potentials can be calculated either using the field equations or N-body simulations. Future surveys will provide us with information on large and small scales and our results will help to analyse the data and compare theoretical number counts with observed quantities.
We present our main result in equation (5.17), the galaxy number counts up to and including second order in CPT. We use scalar perturbations in longitudinal gauge allowing for non-zero anisotropic stress. We assume a flat FLRW background universe filled with a pressureless fluid.
As mentioned earlier, we are not the first group to perform this calculation. We compared our result for the galaxy number overdensities with others published in the literature, at first order. Since other groups use different notations and approaches, e.g. conformal time instead of affine parameter, we adapted the results of the other groups to our notation in order to make a clear comparison possible. We find that we are in agreement at linear order with previous works. Nevertheless, the approaches taken by other groups lead to differences in the results at second order. Given the size of the expressions involved and the complexity of rewriting the results of the other groups, we leave for a follow up paper the comparison of second order results. In reference [36], we tackle this issue by performing the full comparison in an Einstein–de Sitter universe.
Acknowledgments
The authors are grateful to Pedro Carrilho, Chris Clarkson, Obinna Umeh, Roy Maartens and Julian Larena for useful discussions and comments. JF acknowledges support of studentship funded by Queen Mary University of London as well as CONACYT Grant No. 603085. KAM is supported in part by the STFC under Grants ST/M001202/ and ST/P000592/1. JCH acknowledges support from research Grant SEP-CONACYT CB-2016-282569. The tensor algebra package xAct [37] and its subpackage xPand [38] were employed to derive the results presented.
Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).
Appendix A.: Connection coefficients
The connection coefficients in an FLRW spacetime, in longitudinal gauge, up to second order are
including only scalar perturbations. To translate the FLRW coefficients into Minkowski spacetime we just set .
Appendix B.: Perturbed Ricci tensor Rμν
The perturbed Ricci tensor components in an FLRW spacetime, in longitudinal gauge, up to second order are
including only scalar perturbations. To translate the FLRW components into Minkowski spacetime we just set .
Appendix C.: Second order in terms of the metric potentials
C.1. Geodesic equation
Solving equation (2.8) at second order gives
Using equation (3.1), and the integrated version of equations (3.4) and (3.5), we rewrite equations (C1) and (C2) purely in terms of the metric potentials,
where we integrate along the line of sight from λo to λs.
C.2. Energy
The perturbed energy in terms of the metric potentials is given by
C.3. Observed redshift
At second order the redshift is
and using equations (3.4), (3.5), (C1) and (C5), in terms of the metric potentials is
C.4. Angular diameter distance
Using equations (3.4), (3.5), (C3), (C4), (C6), (4.18), (B1)–(B3) and (4.24) we find that the second order perturbation to the angular diameter distance becomes
C.5. Physical volume
The second order perturbation to the physical volume is
C.6. Redshift density
Footnotes
- 3
- 4
Note that in the background, nμ = a−1[0, ni ].