Weil–Petersson Volumes Overview

• Weil–Petersson volumes are symplectic measures that define moduli spaces of hyperbolic Riemann surfaces with geodesic boundaries, cusps, or cone points.
• They encode intersection numbers on compactified moduli spaces via even polynomial expressions and recursive formulas established by Mirzakhani.
• They connect hyperbolic geometry, asymptotic analysis, and matrix-model frameworks, offering computational tools for complex surface invariants.

Weil–Petersson volumes are symplectic volumes of moduli spaces of hyperbolic Riemann surfaces, possibly with geodesic boundary, cusps, or cone points. For the moduli space $\mathcal M_{g,n}(\mathbf L)$ of genus $g$ surfaces with $n$ labelled geodesic boundary components of prescribed lengths $\mathbf L=(L_1,\dots,L_n)$, a standard normalization is

$$V_{g,n}(\mathbf L)=\int_{\mathcal M_{g,n}(\mathbf L)} \frac{\omega_{WP}^{\,3g-3+n}}{(3g-3+n)!},$$

while for cusped moduli the cohomology class satisfies $[\omega_{WP}]=2\pi^2\kappa_1$, so the volumes are directly expressible in tautological intersection theory on $\overline{\mathcal M}_{g,n}$ (Do, 2011, Chen et al., 7 Mar 2026). This places Weil–Petersson volume theory at the intersection of hyperbolic geometry, Teichmüller theory, recursion formalisms, asymptotic analysis, and mathematical physics.

1. Geometric setting and analytic definition

The basic geometric object is a hyperbolic surface of type $(g,n)$, with moduli space $\mathcal M_{g,n}(\mathbf L)$ for prescribed geodesic boundary lengths, and $\mathcal M_{g,n}$ in the cusped case $g$0. In Fenchel–Nielsen coordinates associated to a pants decomposition, one has length variables $g$1 and twist variables $g$2, and the Weil–Petersson symplectic form is

$g$3

This form is invariant under the mapping class group and descends from Teichmüller space to moduli space (Do, 2011).

An analytic description identifies the cotangent space at a pointed curve $g$4 with

$g$5

and the Petersson pairing

$g$6

with $g$7 the complete hyperbolic metric defines the Weil–Petersson Hermitian metric. Its imaginary part is the Kähler form $g$8. For cone metrics with $g$9, the same formula is used with the conical hyperbolic metric $n$0 (Chen et al., 7 Mar 2026).

This dual symplectic–analytic description explains why Weil–Petersson volumes admit both geometric and algebro-geometric treatments. In hyperbolic terms they measure the size of moduli spaces; in algebro-geometric terms they encode intersection numbers on compactifications of moduli of curves.

2. Polynomiality and intersection-theoretic structure

Mirzakhani’s polynomiality theorem states that $n$1 is an even polynomial in the boundary lengths, of total degree $n$2. More precisely,

$n$3

where $n$4 are cotangent-line classes and $n$5 is the first Mumford–Morita–Miller class (Do, 2011).

The cohomological source of this polynomiality is the identity

$n$6

In the cusped case only the $n$7-term remains, yielding

$n$8

The survey literature treats this as the starting point of the modern theory (Chen et al., 7 Mar 2026).

Several low-complexity examples already display the structure: $n$9

$\mathbf L=(L_1,\dots,L_n)$0

These formulas show that only even powers of boundary lengths occur, and that coefficients are rational multiples of powers of $\mathbf L=(L_1,\dots,L_n)$1 (Do, 2011).

Mirzakhani’s proof passes through symplectic reduction. On an auxiliary moduli space with a marked point on each boundary, the torus action rotating boundary points has moment map

$\mathbf L=(L_1,\dots,L_n)$2

and the relevant circle-bundle Chern classes satisfy $\mathbf L=(L_1,\dots,L_n)$3. This turns variation in boundary lengths into linear variation of the reduced symplectic form, explaining both polynomiality and the appearance of $\mathbf L=(L_1,\dots,L_n)$4-classes (Do, 2011).

3. Recursion, Laplace transforms, and spectral curves

A major computational tool is Mirzakhani’s recursion formula, derived by integrating a generalized McShane identity over moduli space and unfolding the resulting sums over mapping-class-group orbits. In its structural form, the recursion expresses $\mathbf L=(L_1,\dots,L_n)$5 in terms of integrals involving $\mathbf L=(L_1,\dots,L_n)$6, products $\mathbf L=(L_1,\dots,L_n)$7, and lower-boundary terms $\mathbf L=(L_1,\dots,L_n)$8. This recursion computes all Weil–Petersson volumes inductively and, because the coefficients of $\mathbf L=(L_1,\dots,L_n)$9 are intersection numbers, also gives an algorithm for tautological intersections on $$V_{g,n}(\mathbf L)=\int_{\mathcal M_{g,n}(\mathbf L)} \frac{\omega_{WP}^{\,3g-3+n}}{(3g-3+n)!},$$0 (Do, 2011).

The Laplace transform reformulation reveals a second structure. Defining

$$V_{g,n}(\mathbf L)=\int_{\mathcal M_{g,n}(\mathbf L)} \frac{\omega_{WP}^{\,3g-3+n}}{(3g-3+n)!},$$1

Eynard and Orantin showed that the transformed recursion is exactly a matrix-model topological recursion. The corresponding spectral curve is

$$V_{g,n}(\mathbf L)=\int_{\mathcal M_{g,n}(\mathbf L)} \frac{\omega_{WP}^{\,3g-3+n}}{(3g-3+n)!},$$2

and the paper identifies the generating function of Weil–Petersson volumes with a Kontsevich-type tau-function. In the same framework,

$$V_{g,n}(\mathbf L)=\int_{\mathcal M_{g,n}(\mathbf L)} \frac{\omega_{WP}^{\,3g-3+n}}{(3g-3+n)!},$$3

so the closed-surface volume can be extracted from the one-boundary volume polynomial (0705.3600).

Recent generalizations keep the recursive backbone while modifying the geometry. For hyperbolic surfaces with tight distinguished boundaries, the tight Weil–Petersson volumes $$V_{g,n}(\mathbf L)=\int_{\mathcal M_{g,n}(\mathbf L)} \frac{\omega_{WP}^{\,3g-3+n}}{(3g-3+n)!},$$4 remain polynomial in $$V_{g,n}(\mathbf L)=\int_{\mathcal M_{g,n}(\mathbf L)} \frac{\omega_{WP}^{\,3g-3+n}}{(3g-3+n)!},$$5, and their generating functions satisfy a deformed topological recursion on the spectral curve

$$V_{g,n}(\mathbf L)=\int_{\mathcal M_{g,n}(\mathbf L)} \frac{\omega_{WP}^{\,3g-3+n}}{(3g-3+n)!},$$6

When the defect parameter $$V_{g,n}(\mathbf L)=\int_{\mathcal M_{g,n}(\mathbf L)} \frac{\omega_{WP}^{\,3g-3+n}}{(3g-3+n)!},$$7 is set to $$V_{g,n}(\mathbf L)=\int_{\mathcal M_{g,n}(\mathbf L)} \frac{\omega_{WP}^{\,3g-3+n}}{(3g-3+n)!},$$8, this reduces to the classical Weil–Petersson spectral curve and the ordinary Mirzakhani recursion (Budd et al., 2023).

These recursive and spectral-curve formalisms explain why Weil–Petersson volumes are simultaneously geometric invariants, generating series for intersection numbers, and outputs of integrable or matrix-model structures.

4. Large-genus and large-$$V_{g,n}(\mathbf L)=\int_{\mathcal M_{g,n}(\mathbf L)} \frac{\omega_{WP}^{\,3g-3+n}}{(3g-3+n)!},$$9 asymptotics

For fixed $[\omega_{WP}]=2\pi^2\kappa_1$0, the large-genus regime $[\omega_{WP}]=2\pi^2\kappa_1$1 is governed by rigid ratio asymptotics. One has

$[\omega_{WP}]=2\pi^2\kappa_1$2

hence

$[\omega_{WP}]=2\pi^2\kappa_1$3

up to polynomial factors in $[\omega_{WP}]=2\pi^2\kappa_1$4 (Mirzakhani, 2010).

Mirzakhani and Zograf proved a complete asymptotic expansion of the form

$[\omega_{WP}]=2\pi^2\kappa_1$5

with coefficients initially known to lie in $[\omega_{WP}]=2\pi^2\kappa_1$6. A 2025 result proves the stronger statement conjectured by Mirzakhani and Zograf: the asymptotic coefficients are actually polynomials in $[\omega_{WP}]=2\pi^2\kappa_1$7 (Huang, 11 Jan 2025).

A distinct asymptotic regime fixes genus $[\omega_{WP}]=2\pi^2\kappa_1$8 and lets the number of cusps grow. For fixed $[\omega_{WP}]=2\pi^2\kappa_1$9 and fixed number of boundary components $\overline{\mathcal M}_{g,n}$0,

$\overline{\mathcal M}_{g,n}$1

as $\overline{\mathcal M}_{g,n}$2, where $\overline{\mathcal M}_{g,n}$3 is the modified Bessel function and $\overline{\mathcal M}_{g,n}$4 is the first positive zero of $\overline{\mathcal M}_{g,n}$5. This is the large-$\overline{\mathcal M}_{g,n}$6 analogue of the large-genus “sinh approximation” and is built on the asymptotic theory of Manin and Zograf (Hide et al., 2023).

These asymptotic regimes are structurally different. Large genus produces factorial growth with corrections organized in inverse powers of $\overline{\mathcal M}_{g,n}$7; large $\overline{\mathcal M}_{g,n}$8 at fixed genus produces an explicit boundary-length dependence through a product of Bessel functions. Together they show that Weil–Petersson volumes admit sharp asymptotic descriptions in more than one direction of moduli-space complexity.

5. Cone points, chamber decompositions, and wall-crossing

Weil–Petersson volume theory extends beyond geodesic boundaries to hyperbolic cone surfaces. In the notation of the survey literature, $\overline{\mathcal M}_{g,n}$9 corresponds to a cone point of angle $(g,n)$0, and the existence of a metric requires the Gauss–Bonnet constraint

$(g,n)$1

For cone angles $(g,n)$2, the parameter

$(g,n)$3

places the problem inside Hassett’s space of weighted pointed stable curves (Chen et al., 7 Mar 2026, Anagnostou et al., 2023).

The crucial structural fact is that cone-surface volumes are not globally polynomial in the cone angles. Instead they are piecewise polynomial, with pieces indexed by Hassett stability chambers. For each chamber $(g,n)$4,

$(g,n)$5

where

$(g,n)$6

In the maximal chamber $(g,n)$7, this recovers Mirzakhani’s polynomial. Crossing a wall $(g,n)$8 changes the chamber polynomial by an explicit wall-crossing term

$(g,n)$9

with

$\mathcal M_{g,n}(\mathbf L)$0

The resulting volume is continuous across walls, differentiable across walls, and vanishes at $\mathcal M_{g,n}(\mathbf L)$1 in light coordinates (Anagnostou et al., 2023).

This chamberwise description corrects a common oversimplification. The original Mirzakhani polynomial does not globally extend the actual volume function on all cone-angle chambers. The survey literature records an explicit counterexample,

$\mathcal M_{g,n}(\mathbf L)$2

showing that naive global analytic continuation fails (Chen et al., 7 Mar 2026).

A related wall-crossing phenomenon appears in the moduli of weighted points on $\mathcal M_{g,n}(\mathbf L)$3. For

$\mathcal M_{g,n}(\mathbf L)$4

the Weil–Petersson volume coincides with the CM degree, and the main continuity theorem states that the Fano CM volumes converge to the geometric volume computed by McMullen when $\mathcal M_{g,n}(\mathbf L)$5. In the four-point case,

$\mathcal M_{g,n}(\mathbf L)$6

which is recovered as the limit of the Fano expression (Tambasco, 2021).

6. Combinatorics, random surfaces, and metric models

Weil–Petersson volumes have direct probabilistic consequences for random hyperbolic surfaces sampled with respect to the Weil–Petersson measure. In large genus, asymptotic volume estimates imply that short simple closed geodesics occur with positive probability, while separating systoles are typically much longer: for $\mathcal M_{g,n}(\mathbf L)$7,

$\mathcal M_{g,n}(\mathbf L)$8

The same analysis yields a positive lower bound for the Cheeger constant of a random large-genus surface and logarithmic bounds for typical diameter (Mirzakhani, 2010).

In the many-cusps regime, the large-$\mathcal M_{g,n}(\mathbf L)$9 asymptotic formula for bordered volumes becomes the engine for new spectral and geodesic-counting results. For fixed genus and large $\mathcal M_{g,n}$0, a random surface in $\mathcal M_{g,n}$1 has linearly many small Laplacian eigenvalues with high probability, and for any $\mathcal M_{g,n}$2 with $\mathcal M_{g,n}$3,

$\mathcal M_{g,n}$4

with high probability, so most closed geodesics of those lengths are non-simple (Hide et al., 2023).

The subject also exhibits exact combinatorial correspondences. The number of border-strip decompositions of the $\mathcal M_{g,n}$5 rectangle equals the sequence $\mathcal M_{g,n}$6 governing the Weil–Petersson volume of $\mathcal M_{g,n}$7: $\mathcal M_{g,n}$8 Since

$\mathcal M_{g,n}$9

this identifies a specific volume sequence with a ribbon-tiling enumeration problem (Alexandersson et al., 2018).

A different combinatorial model comes from irreducible metric maps. The volumes $g$00 of essentially $g$01-irreducible metric maps are symmetric polynomials in $g$02, homogeneous of degree $g$03 in $g$04, and satisfy string and dilaton equations. For $g$05 and $g$06,

$g$07

For $g$08 the identity fails, but the generating functions remain closely parallel (Budd, 2020).

7. Matrix models, JT gravity, supergeometry, and deformations

The modern physics interface begins with Jackiw–Teitelboim gravity, where connected amplitudes with $g$09 asymptotic boundaries are genus sums whose geometric building blocks are the bordered Weil–Petersson volumes $g$10. In one formulation,

$g$11

In the regime $g$12 and $g$13, a conjectured asymptotic formula is

$g$14

and the universal random-matrix limit of the spectral form factor imposes explicit linear relations among the coefficients of the two-boundary polynomial

$g$15

providing independent constraints on Weil–Petersson volume data (Kimura, 2020, Weber et al., 2022).

Computationally, recent matrix-model methods replace full topological recursion by ordinary differential equations. For $g$16, the combination of the string equation for $g$17 and the Gel'fand–Dikii equation for the diagonal resolvent $g$18 yields an efficient genus-by-genus algorithm. In this framework each $g$19 is a total derivative in $g$20, so the volume reduces to boundary data at the Fermi surface after Laplace transform. The method reproduces, for example,

$g$21

and extends to $g$22, $g$23, and small and large $g$24 JT supergravity models (Ahmed et al., 24 Jul 2025).

The supergeometric extension replaces ordinary intersection theory by the Norbury class $g$25. The super Weil–Petersson volume is written as

$g$26

and recent work proves a complete large-genus asymptotic expansion generalizing Mirzakhani–Zograf, with asymptotic coefficients polynomial and recursively computable (Huang, 14 Jan 2025).

Several further deformations show that the theory is still expanding. A $g$27-analogue of Mirzakhani’s recursion produces symmetric polynomials with coefficients in a $g$28-zeta algebra and recovers the classical Weil–Petersson volumes in the rescaled limit $g$29 (Do et al., 14 Oct 2025). For super-Riemann surfaces with Neveu–Schwarz boundaries and Ramond punctures, a random-matrix construction computes closed-form formulae for $g$30 and identifies a spectral curve whose topological recursion re-derives the same data (Johnson, 8 Jun 2026).

Across these developments, a stable theme persists: Weil–Petersson volumes are simultaneously hyperbolic volumes, tautological intersection numbers, recursive amplitudes, asymptotic invariants, and matrix-model observables. That multiplicity of realizations is the defining feature of the subject.