Wide Angle Redshift Distortions Revisited

Short version: the algebra checks out and the trigonometric expansion in Eq. 11 is the real contribution. The framing around the B coefficients claims more than the data can deliver, but that is a presentation problem, not a math problem, and the paper itself points to Eq. 11 as the practical route. Worth a referee's time; not a desk reject and not a landmark either. It is a clean theory note in a narrow subfield, and it says so in its own conclusion (theory only, no data, practical issues deferred). The classifier's 'no limitations found' is wrong on the text: Section 2 explicitly drops the selection-function gradient and the rocket effect, and Section 5 is candid about sky coverage and noise. Credit where due — the author is not hiding anything. What is actually new: the zero-total-angular-momentum tripolar expansion, the demonstration that only five B's survive (Eq. 6), and the two-angle trigonometric form (Eq. 11). The plane-parallel cross-check against Kaiser and Hamilton in Eq. 10 is the right kind of validation, and the equivalence to Szalay et al. (1998) in Eq. 13 is a real consistency check, not a hand-wave. Where it is soft: the stress-test concern lands. The B's are deterministic functions of separation given P(k) and f, not stochastic quantities estimated from a map, and the orthogonality in Eq. 7 needs the full 3-sphere while the physical correlation function lives on coplanar triples. Property (vi) extends the domain by flipping x₂, but that stays inside the coplanar set. So 'analogous to the C_l's' and 'a natural basis for maximum likelihood estimation' are overstated; the honest object for estimation is Eq. 11, which the paper does present. Two smaller things: the inversion of Eq. 11 is withheld (the author says it is a fourth-order polynomial in f and too messy to print — fair, but a reader wanting to use this will have to redo it), and the typo correction to Szalay et al. (4/15 to 8/15) is asserted rather than shown. Also, the text cites 'Figure 3' when only one figure exists — cosmetic. Who should read it: anyone doing wide-angle RSD or building a KL/maximum-likelihood pipeline for a wide survey, and anyone who wants the Szalay et al. result rederived from symmetry instead of brute force. Recommendation: send it out. Ask the referee to press on the C_l analogy and to request the Eq. 11 inversion coefficients, or at least a Mathematica notebook.

The paper's central claim is the tripolar expansion and the five non-zero B's (Eq. 6). I checked this algebra and it holds: the F_0, F_2 structure of Eq. 2 yields exactly B000, B220, B022=B202, B222, B224, and the plane-parallel limit of Eq. 11 does reduce to Eq. 10 (I verified the constant, mu^2, and mu^4 terms). The concern is not with the derivation but with the interpretation. The paper says the B's are 'analogous to the C_l's of the angular power spectrum' and 'form a natural basis for maximum likelihood estimation' (Section 2, after Eq. 6). But the C_l's are stochastic quantities estimated from a full-sky map, whereas the B's here are deterministic functions of x given P(k) and f. More importantly, the orthogonality in Eq. 7 requires integrating over the full 3-sphere dOmega_1 dOmega_2 dOmega_3, while the physical correlation function is only defined on coplanar triples (property iv). The paper's property (vi) extends the domain by flipping x_2, but this only extends within the coplanar set; it does not make the physical data cover the full 3-sphere. So the B's are not directly estimable from the data in the way C_l's are. The paper's practical route is the 2D trigonometric expansion (Eq. 11), which is orthogonal on the 2D domain and is the right object for inversion. The claim that the B's themselves are a 'natural basis for maximum likelihood estimation' is therefore somewhat overstated, though the paper does present Eq. 11 as the practical inversion route, so this is a presentational issue rather than a correctness error.

The paper's algebra is correct. I verified the five non-zero B's in Eq. 6 and the plane-parallel limit of Eq. 11, which does reduce to Eq. 10. The reader's weakest assumption (linear theory, neglect of selection function gradient and rocket effect) is an acknowledged limitation stated explicitly in Section 2, so it is not a new finding and is somewhat misdirected for a theory paper that explicitly frames itself as linear theory. The most load-bearing concern is the interpretation of the B's as analogous to C_l's and as a natural basis for maximum likelihood estimation, which is overstated because the B's are deterministic functions of x and the orthogonality in Eq. 7 is over the full 3-sphere, while the physical data only covers coplanar configurations. However, the paper does present Eq. 11 as the practical inversion route, so this is a presentational issue rather than a correctness error. The reader's CONDITIONAL verdict is reasonable, but the specific red flags the reader cites (withheld inversion, asserted typo correction, missing figure) are secondary: the inversion is a fourth-order polynomial that the paper honestly says is too complicated to list, the typo correction is a cross-check claim that does not affect the paper's own derivation, and the figure does appear (the text says 'Figure 3' but the figure is labeled 'FIG. 1', a minor numbering typo). I would leave the verdict unchanged.

If correct, the result gives a compact, orthogonal, finite basis for wide-angle redshift distortions, with the plane-parallel limit recovered and cross-checks against three independent published results (Kaiser 1987, Hamilton 1993, Szalay et al. 1998). The trigonometric expansion of Eq. (11) is a genuinely compact and invertible expression, and the reduction to five non-zero B's is a real simplification over the formally infinite expansions used elsewhere. The paper is a theory paper with no data application, which is appropriate for the journal; the main strengths are the clean angular-momentum derivation and the explicit cross-checks against published limits.

Galaxy surveys that span a wide angle on the sky see peculiar velocities distort the two-point correlation function in a way that depends on two lines of sight and the separation between the pair, not just on the angle to a single line of sight. Szapudi asks whether that three-direction angular dependence can be organized by symmetry instead of by an infinite series, and argues that the natural basis is the zero-total-angular-momentum tripolar spherical harmonics $S_{l_1l_2l_3}(x_1,x_2,x)$. Rotational invariance forces total angular momentum $L=0$, and the quadratic structure of linear redshift distortions then leaves only five non-zero coefficients $B_{l_1l_2l_3}$, each a function of pair separation alone and built from spherical Bessel moments $xi_ell(x)$ of the linear power spectrum together with $f=Omega^{0.6}/b$. If this is right, a wide-angle distortion analysis becomes a small orthogonal fitting problem with a handful of invertible coefficients, and the standard distant-observer result of Kaiser and Hamilton is recovered as the limit where the two lines of sight coincide.

What carries the argument. The carrier is the family $S_{l_1l_2l_3}(hat{x}_1,hat{x}_2,hat{x})$, defined by contracting three normalized spherical harmonics with a Wigner 3j symbol of zero lower row. These functions are a subset of the tripolar spherical harmonics and form a complete orthogonal basis for spherically symmetric functions of three unit vectors; the paper gives the normalization explicitly, with the integral of $S_{l_1l_2l_3}S_{j_1j_2j_3}$ over all three spheres equal to $delta_{l_1j_1}delta_{l_2j_2}delta_{l_3j_3}(4pi)^{3/2}/sqrt{(2l_1+1)(2l_2+1)(2l_3+1)}$. Two structural facts do the work: the plane-parallel reduction $S_{l_1l_2l_3}(hat{x}_1,hat{x}_1,hat{x})=binom{l_1 l_2 l_3}{0 0 0}P_{l_3}(hat{x}_1cdothat{x})$, which is what makes contact with Kaiser and Hamilton, and the vanishing of every $S$ with a single zero index except the $S_{0ll}$, $S_{l0l}$, $S_{ll0}$ family, which is what prunes the expansion to five terms. The Gaunt integral, the Rayleigh expansion of the plane wave, and the addition theorem for Legendre polynomials are the three standard identities that carry the algebra from Eq. 2 to Eq. 6.

Core claim. The paper's central claim is that the redshift-space two-point correlation function, written in linear theory as a sum of a monopole term plus two cross terms plus a quadrupole-squared term, expands in zero-angular-momentum tripolar harmonics with only five non-zero coefficients: $B_{000}=(1+frac{1}{3}f)^2xi_0$, $B_{220}=frac{4}{9sqrt{5}}f^2xi_0$, $B_{022}=B_{202}=-(frac{2}{3}f+frac{2}{9}f^2)sqrt{5}xi_2$, $B_{222}=frac{4sqrt{10}}{9sqrt{7}}f^2xi_2$, and $B_{224}=frac{4sqrt{2}}{sqrt{35}}f^2xi_4$. Because the three unit vectors are coplanar, the angular dependence can be parametrized by two angles, and the author uses that freedom to write two equivalent forms: a trigonometric polynomial in two angles (Eq. 11) with four independent amplitudes $a_{00},a_{02}=a_{20},a_{22},b_{22}$, which he calls the most compact expression of wide-angle distortions; and a double Legendre expansion (Eq. 13) in the variables of Szalay et al., which he shows reproduces their result once a typographical factor $4/15$ is corrected to $8/15$. Setting the two lines of sight parallel returns the Kaiser-Hamilton expression $(1+frac{2}{3}f+frac{1}{5}f^2)xi_0-(frac{4}{3}f+frac{4}{7}f^2)xi_2P_2+frac{8}{35}f^2xi_4P_4$.

The premise it rests on. The whole construction starts from linear theory: the Jacobian is truncated at first order in the velocity divergence, and the gradient of the selection function and the rocket effect are explicitly dropped, so the five coefficients describe the correlation function only where velocities are linear and the survey selection is locally uniform.

Wide-angle redshift distortion measurements can be turned into a maximum-likelihood fit for five separation-dependent coefficients rather than an infinite angular series, with the $C_ell$-like $B_{l_1l_2l_3}$ playing the role of the summary statistics.

Because the basis functions are orthogonal over the full sphere, the coefficients of the trigonometric expansion (Eq. 11) can be extracted by integration or by fit and inverted for $f$, $xi_0$, $xi_2$, and $xi_4$; the paper notes the inversion is a quartic in $f$ and better done numerically.

The double Legendre form (Eq. 13) gives a direct, quantitative handle on how far a survey can open up before the plane-parallel approximation fails; the paper's Figure 1 puts that threshold near 15 to 20 degrees half-opening angle for common distortion estimators.

The same symmetry pruning should extend to higher-order statistics such as the three-point function and cumulant correlators, which the author defers to later work.

The five-coefficient structure suggests a practical diagnostic: any measured set of $B$'s that requires more than five non-zero terms at a given separation is signalling physics beyond linear theory, selection-function gradients, or survey geometry rather than a failure of the basis.

Because the coefficients separate into $f$-polynomials times $xi_ell(x)$, the ratio $B_{220}/B_{000}$ at small separations is nearly independent of the power spectrum shape, which is the kind of quantity that could be calibrated against mocks before being trusted on data.

The author's remark that the trigonometric form survives selection-function gradients while the Legendre form does not implies the two expansions are not interchangeable in practice; a survey analysis should pick the trigonometric route if the radial selection varies appreciably across the pair.

The paper is purely theoretical and presents no measurement; the natural next test is to run the five-coefficient fit on existing wide-angle catalogues and check that the recovered $f$ agrees with the plane-parallel estimate once the opening angle is restricted.

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review