The Kaiser Effect: How Coherent Infall Squashes Redshift-Space Clustering

What the Kaiser effect is: velocities masquerading as distance

We never measure a galaxy's distance directly in a large survey — we measure its redshift and convert it to distance assuming pure Hubble flow, r = cz/H₀. But a galaxy's redshift is the sum of two pieces: the cosmological expansion plus a Doppler shift from its own peculiar velocity (its motion relative to the smooth expansion). That contamination means the map we build is not real space but redshift space, and its distortions carry physical information.

On large scales, gravity pulls galaxies coherently toward growing mass concentrations — the near side of a supercluster falls away from us, the far side falls toward us. In redshift space these infall velocities push both sides toward the cluster's center along the line of sight, compressing the structure radially. Kaiser (1987) showed this coherent flow doesn't just distort shapes; it amplifies the measured clustering amplitude for modes aligned with the line of sight, turning a nuisance into a precision probe of structure growth.

The mechanism and Kaiser's linear formula

In linear perturbation theory, continuity ties the velocity field to the density field: the divergence of the peculiar velocity is proportional to the growth rate f = d ln D / d ln a, where D(a) is the linear growth factor. A Fourier density mode with wavevector k gets amplified in redshift space by a factor that depends on the angle to the line of sight.

• Real-space density contrast δ maps to a redshift-space contrast δ_s = δ (1 + β μ²), where μ = cos θ is the cosine of the angle between k and the line of sight.
• β = f / b is the redshift-distortion parameter: f is the growth rate, b the linear galaxy bias.
• Squaring for the power spectrum gives the famous Kaiser formula: P_s(k, μ) = (1 + β μ²)² · P_r(k).

Modes along the line of sight (μ = 1) are boosted by (1 + β)²; transverse modes (μ = 0) are untouched. This angular dependence is the entire signal — it flattens the correlation function contours perpendicular to the line of sight, the geometric opposite of the Fingers of God.

Key quantities: multipoles, β, and a worked example

Because the distortion depends only on μ, the redshift-space power spectrum expands cleanly in Legendre polynomials. Only three even multipoles survive in linear theory:

• Monopole: P₀ = (1 + (2/3)β + (1/5)β²) · P(k)
• Quadrupole: P₂ = ((4/3)β + (4/7)β²) · P(k)
• Hexadecapole: P₄ = (8/35)β² · P(k)

The quadrupole-to-monopole ratio P₂/P₀ isolates β independently of the overall amplitude, which is why it's the workhorse RSD estimator. As a worked example, take a present-day Ω_m,0 ≈ 0.31, so that the density parameter at z ≈ 0.5 has grown to Ω_m(z) ≈ 0.60; then f ≈ Ω_m(z)^0.55 ≈ 0.76. For a bias b ≈ 1.8 galaxy sample, β = f/b ≈ 0.42. The monopole is then boosted by 1 + (2/3)(0.42) + (1/5)(0.42²) ≈ 1.32 — a 32% enhancement of the measured clustering amplitude relative to real space, purely from coherent infall.

How it's observed: from 2dF to DESI

The Kaiser effect shows up as an anisotropy in the two-point statistics: plot the correlation function ξ(σ, π) as a function of separation transverse (σ) and along (π) the line of sight, and the iso-correlation contours are squashed in the π direction on large scales. The first robust detections came from the 2dF Galaxy Redshift Survey (Peacock et al. 2001), which measured β ≈ 0.43 ± 0.07 at z ≈ 0.1.

Modern surveys turn this into a growth-rate measurement. Because the bias b is degenerate with the amplitude σ₈, RSD analyses report the combination fσ₈. Landmark values include BOSS DR12: fσ₈ ≈ 0.44 ± 0.04 at z ≈ 0.57, and eBOSS quasars: fσ₈ ≈ 0.43 ± 0.08 at z ≈ 1.52. DESI and Euclid are now pushing these to sub-percent precision across 0 < z < 2, mapping the entire growth history in a single anisotropy signal.

Kaiser versus its cousins: Fingers of God, AP, and nonlinear RSD

The Kaiser effect is one of two redshift-space distortions, and they act in opposite directions:

• Fingers of God (FoG): inside virialized clusters, random galaxy velocities (σ_v ~ 300–500 km/s) smear galaxies radially, elongating structures along the line of sight and suppressing small-scale power. This is nonlinear and must be modeled — often with a Lorentzian or Gaussian streaming term multiplying the Kaiser factor.
• Alcock–Paczyński effect: a purely geometric distortion from assuming the wrong cosmology when converting redshifts to distances — distinct from the dynamical Kaiser signal, though the two are measured jointly.

The Kaiser formula is strictly linear: it breaks down below ~10 Mpc/h where FoG and mode-coupling dominate. Real analyses use extensions — the Scoccimarro / TNS models, effective field theory of large-scale structure, and Gaussian streaming — that reduce to Kaiser on large scales but capture the nonlinear damping on small scales.

Why it matters: testing gravity on cosmic scales

The deep payoff is that the growth rate f depends on the theory of gravity, not just the expansion history. General Relativity predicts f ≈ Ω_m(z)^γ with γ ≈ 0.55; many modified-gravity theories predict a different γ. Because the Kaiser effect measures f directly through β, RSD provides a growth-based test of GR that is independent of geometric probes like supernovae or the CMB.

Current open questions center on the 'growth tension': some RSD and weak-lensing measurements hint at slightly weaker structure growth (lower S₈ = σ₈√(Ω_m/0.3)) than the Planck CMB predicts, at the ~2–3σ level. Whether this reflects new physics, unmodeled nonlinear RSD, or bias systematics is actively debated. Kaiser passed away in 2023, but his 1987 paper — one of the most cited in cosmology — remains the linear backbone of every growth-rate result from DESI, Euclid, and the Roman Space Telescope.

Kaiser effect versus Fingers of God: the two redshift-space distortions and their opposite geometries

Property	Kaiser effect	Fingers of God
Scale	Large / linear (~10–100 Mpc/h)	Small / nonlinear (~1 Mpc/h, inside halos)
Velocity source	Coherent infall toward overdensity	Random virial motions in bound clusters
Redshift-space shape	Flattened perpendicular to line of sight ('pancakes')	Elongated along line of sight ('fingers')
Effect on clustering	Amplifies power, factor (1+βμ²)²	Suppresses small-scale power, radial smearing
Governing quantity	β = f/b, growth rate f	Velocity dispersion σ_v (~300–500 km/s)
Cosmological use	Measures fσ8, tests gravity	Systematic to be modeled/removed