---
title: Adapted Renormalized Volume
url: https://www.emergentmind.com/topics/adapted-renormalized-volume
type: topic
---

# Adapted Renormalized Volume

Adapted renormalized volume is not a single universally fixed invariant but a family of closely related constructions in which the renormalized volume is normalized relative to additional geometric data. In the asymptotically hyperbolic Einstein setting, the adaptation is a canonical choice of boundary representative satisfying a holographic slice condition \(v_n(h)=\int_N v_n(h)\,dvol_h\), which turns renormalized volume into a well-defined functional on conformal structures and identifies its differential with an exact cotangent datum [1211.6705]. In convex co-compact hyperbolic \(3\)-manifolds with compressible boundary, the term denotes a modified functional
\[
\widetilde{V_R}(X)=V_R(X)+\max_{m\in \Gamma^{comp}(\partial M)} \pi^3 \sum_{\gamma\in m}\frac{1}{\ell_\gamma(X)},
\]
designed to cancel the classical divergence under compressible pinching [2507.19291]. In minimal-submanifold and singular Yamabe problems, the adaptation is to boundary-adapted defining functions, tubular coordinates, or singularity-adapted cutoffs, so that the finite part and anomaly reflect the geometry of the embedded singular set [2111.04294].

| Context | Adaptation | Principal outcome |
|---|---|---|
| AHE manifolds, even boundary dimension | \(v_n(h)=\int_N v_n(h)\,dvol_h\) | Canonical representative and exact \(1\)-form on a slice |
| Convex co-compact \(3\)-manifolds with compressible boundary | Add \(\pi^3/\ell_\gamma\) penalty over compressible multicurves | Lower bounded functional with bounded differential and WP gradient |
| WP gradient-flow setting | Surgery near geodesic-boundary strata | “Adapted” flow, not a new \(V_R\) |
| Minimal submanifolds in PE manifolds | Boundary-adapted FG/Fermi coordinates and special bdf | Canonical finite part for even \(m\) |
| Higher-codimension singular Yamabe spaces | Tubular cutoff and fiberwise parity | Odd-\(n\) conformal invariant \(V_{n,k}\); even-\(n\) energy \(\mathcal A_{n,k}\) |

## 1. Foundational renormalized-volume framework

For an asymptotically hyperbolic Einstein manifold \((M,g)\) with conformal infinity \([h_0]\), a geodesic boundary defining function \(x\) associated to \(h_0\) puts the metric in Fefferman–Graham form \(g=x^{-2}(dx^2+h_x)\), and the renormalized volume is defined by Hadamard or Riesz regularization,
\[
\operatorname{Vol}_R(M,g;h_0):=\operatorname{FP}_{z=0}\int_M x^z\,dvol_g
\]
or equivalently \(\operatorname{FP}_{\varepsilon\to 0}\int_{x>\varepsilon} dvol_g\). When the boundary dimension \(n\) is even, \(\operatorname{Vol}_R(M,g;h_0)\) depends on the representative \(h_0\), while the logarithmic coefficient \(L=\int_N v_n(h_0)\,dvol_{h_0}\) is conformally invariant; when \(n\) is odd, \(\operatorname{Vol}_R(M,g)\) is independent of \(h_0\) [1211.6705].

A related invariant in even bulk dimension is the renormalized volume \(\mathrm{RenV}\) of a volume-renormalizable asymptotically hyperbolic metric. In that setting, partial evenness of the Graham–Lee expansion together with the trace condition \(\mathrm{tr}^{h_0}h_{n-1}=0\) guarantees that the finite part is independent of adapted special boundary defining functions. This independence is preserved by normalized Ricci flow, which also preserves the underlying evenness structure [1607.08558].

In dimension three, an additional normalization uses a fixed hyperbolic background. For asymptotically hyperbolic \(3\)-manifolds modeled on hyperbolic space, one defines an adapted renormalized volume by subtracting the hyperbolic volume on the same exhaustion,
\[
V(M,g)=\lim_{i\to\infty}\big(\operatorname{vol}(Q_i\cap M,g)-\operatorname{vol}(Q_i\cap M,g_{\mathbb H^3})\big),
\]
which coincides with the finite-part renormalized volume associated to the round conformal infinity [1305.6628].

## 2. Canonical boundary representatives in AHE geometry

The most systematic use of the adjective “adapted” in higher-dimensional AHE geometry arises from conformal variation. Under a boundary conformal change \(h_0^s=e^{2s\omega}h_0\),
\[
\frac{d}{ds}\Big|_{s=0}\operatorname{Vol}_R(M,g;e^{2s\omega}h_0)=\int_N \omega\, v_n(h_0)\,dvol_{h_0}.
\]
With fixed boundary volume, the Euler–Lagrange equation is therefore \(v_n(h)=\mathrm{const}\). In boundary dimension \(n=2\), this is the constant-curvature equation; in \(n=4\), it is exactly the \(\sigma_2\)-Yamabe equation \(\sigma_2(\mathrm{Sch}_h)=\mathrm{const}\) [1211.6705].

The adapted representative is the metric in a conformal class satisfying
\[
v_n(h)=\int_N v_n(h)\,dvol_h.
\]
Near Einstein conformal classes with negative Ricci curvature and no conformal Killing fields, such representatives form a smooth slice \(S_0\). On that slice, the formally undetermined Neumann term \(h_n\) must be corrected by a local tensor \(F_n(h_0)\), yielding
\[
G_n:=-\frac14(h_n+F_n),\qquad \delta_{h_0}(G_n)=0,\qquad \mathrm{Tr}_{h_0}(G_n)=\frac12 v_n(h_0).
\]
The traceless part \(G_n^\circ\) defines a cotangent vector to \(\mathcal T(N)\), and the adapted renormalized volume has differential
\[
d\,\operatorname{Vol}_R(M,g;h_0)\cdot \dot h_0=\int_N G_n^\circ(h_0):\dot h_0\,dvol_{h_0}.
\]
This makes the map \(h_0\mapsto G_n^\circ(h_0)\) an exact \(1\)-form on \(S_0\), with \(\operatorname{Vol}_R\) as generating function, and identifies the set of adapted Cauchy data with a Lagrangian submanifold of \(T^*\mathcal T(N)\). In the two-ended case this yields a higher-dimensional analogue of McMullen’s quasifuchsian reciprocity [1211.6705].

A complementary adapted viewpoint uses totally geodesic compactifications. If \(n\ge 3\) is odd and \(g\) is any totally geodesic compactification of an AHE metric \(g_+\), then
\[
V_{g_+}=C_{n+1}\int_X v^{(n+1)}(g)\,dv_g.
\]
Within fixed conformal class and unit volume, the second variation of the resulting renormalized volume functionals shows that Einstein metrics of nonzero scalar curvature are local extrema; for the even-\(n\) renormalized volume \(V_{g_+}\), the sign depends on the scalar curvature and on \(n \bmod 4\) [1211.6422].

## 3. Hyperbolic \(3\)-manifolds and Teichmüller-theoretic adaptations

For convex cocompact hyperbolic \(3\)-manifolds, a standard normalization defines \(V_R(M)=W(\rho_M)\), where \(W(N)=\operatorname{vol}(N)-\frac12\int_{\partial N}H\,dA\) and \(\rho_M\) is the hyperbolic metric in the conformal class at infinity. The Takhtajan–Zograf/Takhtajan–Teo variation formula gives
\[
dV_R(\mu)=\operatorname{Re}\int_Y \phi_Y\,\mu,
\]
so the Weil–Petersson gradient is represented by \(\overline{\phi_Y}/\rho_Y\), with global bounds coming from the Kraus–Nehari estimate \(\|\phi_Y\|_\infty\le 3/2\) [2003.00337].

In one important usage, however, “adapted” does not redefine the functional at all. The Weil–Petersson gradient-flow work for relatively acylindrical deformation spaces introduces an adapted or surgered flow of \(-V_R\): away from geodesic-boundary strata it follows the WP gradient flow \(\dot Y_t=-\nabla_{WP}V_R(Y_t)\), and near a noded geodesic-boundary point it performs a Bonahon–Otal unbending surgery that strictly decreases \(V_R\). The resulting path is piecewise smooth, exists for all time, decreases \(V_R\) strictly, and converges to the unique geodesic-boundary structure \(M_{\rm geod}\). In that paper, “adapted” refers to the flow, not to a new renormalized volume [2003.00337].

A genuinely new adapted renormalized volume appears for hyperbolic \(3\)-manifolds with compressible boundary. Classical \(V_R\) can diverge to \(-\infty\) when a compressible curve is pinched. The adapted functional
\[
\widetilde{V_R}(X)=V_R(X)+\max_{m\in\Gamma^{comp}(\partial M)} \pi^3 \sum_{\gamma\in m}\frac1{\ell_\gamma(X)}
\]
absorbs that divergence. The maximum may be taken over maximal compressible multicurves; it is attained, so \(\widetilde{V_R}\) is continuous and smooth away from loci where multiple multicurves realize the maximum. The functional is bounded from below, its differential is uniformly bounded in \(L^1\) and \(L^\infty\) wherever it exists, its Weil–Petersson gradient is uniformly bounded, and it stays at uniformly bounded distance from the convex core volume. It extends continuously to the strata in the boundary of the Weil–Petersson completion corresponding to compressible multicurves, using pointed convex cocompact limits and cusp counterterms near marked boundary points [2507.19291].

These \(3\)-dimensional constructions sit alongside a broader classical theory in which renormalized volume is continuous under geometrically finite limits and attains its minimum at the geodesic class in the acylindrical setting [1605.07986].

## 4. Minimal submanifolds and higher-codimension singular Yamabe spaces

For a minimal submanifold \(Y^m\subset M^{n+1}\) in a Poincaré–Einstein manifold, the renormalized volume is the constant term in the expansion of \(\mathrm{Vol}(Y_\varepsilon)\) or, equivalently,
\[
\mathcal V(Y):=\operatorname{FP}_{z=0}\int_Y x^z\,dA_Y.
\]
The analysis is carried out in boundary-adapted FG/Fermi coordinates, where \(Y\) is represented as a normal graph over the boundary cylinder \(\gamma\times \mathbb R^+\). The expansion
\[
u^i(s,x)=u_2^i(s)x^2+\cdots
\]
has leading coefficient
\[
u_2^i(s)=\frac{1}{2(m-1)}H_\gamma^i(s).
\]
In this context, “adapted renormalized volume” means the use of the ambient special boundary defining function \(x\) or the intrinsic special defining function \(x_Y\) on \(Y\). For even \(m\),
\[
\operatorname{FP}_{z=0}\int_Y x_Y^z\,dA_Y=\operatorname{FP}_{z=0}\int_Y x^z\,dA_Y,
\]
so the renormalized volume is canonical; for odd \(m\), the two differ by a local boundary anomaly. The first variation is
\[
D\mathcal V|_Y(\dot S)=\int_\gamma \big[dx(\dot S)\sqrt{\det \overline h}\,\big]^{(m)}\,dA_\gamma,
\]
and in codimension \(1\) with even \(n\),
\[
\frac{d}{dt}V(Y_t)\Big|_{t=0}=-(n+1)\int_\gamma \dot\phi_0(s)\,u_{n+1}(s)\,dA_\gamma,
\]
so the renormalized volume recovers the Dirichlet-to-Neumann datum \(u_{n+1}\) [2111.04294].

A different higher-codimension theory starts from a singular Yamabe metric \(g^+=u^{-2}g\) on \(M^{n+k}\setminus \Sigma^n\), normalized by
\[
R_{g^+}=(k-n-2)(n+k-1).
\]
Using tubular Fermi coordinates and the distance \(t\) to \(\Sigma\), one obtains
\[
\mathrm{Vol}_{g^+}(M\setminus N_\varepsilon(\Sigma))
= c_0\varepsilon^{-n}+c_1\varepsilon^{1-n}+\cdots+c_{n-1}\varepsilon^{-1}
+\mathcal A_{n,k}\log(1/\varepsilon)+V_{n,k}+o(1).
\]
Fiberwise parity on the unit normal sphere bundle forces \(c_j=0\) for all odd \(j\). Hence, if \(n\) is odd, \(\mathcal A_{n,k}=0\) and \(V_{n,k}\) is an absolute conformal invariant; if \(n\) is even, \(\mathcal A_{n,k}\) is conformally invariant and takes the form
\[
\mathcal A_{n,k}=\int_\Sigma I_{n,k}\,dV_h.
\]
For \(n=2\) and \(k\neq 4\), the paper gives an explicit formula for \(I_{2,k}\) involving \(|H|^2\), \(\operatorname{tr}(P|_{T\Sigma})\), \(|\mathring A|^2\), and \(R_h\). In the knot case \((n,k)=(1,2)\), \(V_{1,2}\) is a global conformal invariant of knot embeddings in \(S^3\), and for the equatorial unknot one has \(V_{1,2}=-4\pi^2\) [2408.01882].

## 5. Regulators, anomalies, and boundary transgressions

A broad anomaly-based version of renormalized volume treats metrics singular along hypersurfaces. If \((M,c)\) carries an embedded hypersurface \(\Sigma\) with defining density \(\sigma\), and the singular metric is \(g^o=g/\sigma^2\), the regulated volume has an expansion
\[
\mathrm{Vol}_\varepsilon=\sum_{k=1}^{d-1} v_k\,\varepsilon^{-k}+\mathrm{Vol}_{ren}+A\log\varepsilon+\varepsilon R(\varepsilon).
\]
The anomaly
\[
A=\frac{(-1)^d}{(d-1)!}\int \delta^{(d-1)}
\]
is regulator independent, whereas the divergences encode the regulator dependence. For singular Yamabe solutions, the anomaly becomes the integral of an extrinsic \(Q\)-curvature, yielding higher-dimensional generalizations of Willmore-type functionals. The first variation of the anomaly recovers the obstruction density for the singular Yamabe problem [1603.07367].

When the singular region itself has boundary, the anomaly splits into hypersurface and edge contributions. For a region \(D\) intersecting a separating hypersurface \(\widetilde\Sigma\), the anomaly takes the form
\[
\mathcal A=\frac{1}{(d-1)!(d-2)!}\left[\int_\Sigma Q^\sigma_\tau+\int_{\partial\Sigma}T^\sigma_\tau\right].
\]
The renormalized volume is then the primitive of the corresponding anomaly operator. Under the singular Yamabe condition, this produces generalized Willmore energies with explicit boundary transgression terms; in the surface case \(d=3\), the anomaly contains \(\pi\chi_\Sigma\), the trace-free second fundamental form squared, and angle-dependent boundary curvature terms [1611.08345].

An orbifold version replaces boundary cusps and cone points by Schottky \(3\)-orbifolds with geodesic lines of conical singularity reaching the conformal boundary. After subtracting the cut-off area term, the universal Euler-characteristic logarithm, the cone-line logarithms, and the puncture logarithms, one obtains a renormalized hyperbolic volume \(V_{ren}\) satisfying the exact holographic identity
\[
V_{ren}=-\frac14\,\check{\mathscr S}_{\boldsymbol m}.
\]
This identifies \(V_{ren}\) with an orbifold Liouville-type functional, proves a Polyakov anomaly formula for \(V_{ren}\), and shows that \(V_{ren}\) is a Kähler potential for a specific combination of the Weil–Petersson and Takhtajan–Zograf metrics on generalized Schottky space [2412.19137].

## 6. Evolution, extremality, and conceptual distinctions

Under normalized Ricci flow,
\[
\partial_t g=-2(\mathrm{Rc}(g)+(n-1)g),
\]
volume-renormalizable and asymptotically Poincaré–Einstein structures provide yet another adapted setting. For even bulk dimension, the renormalized volume satisfies
\[
\frac{d}{dt}\,\mathrm{RenV}(M^n,g(t))
=
-\,{}^R\!\!\int_{M^n}\big(S(g(t))+n(n-1)\big)\,dV_{g(t)},
\]
and for APE metrics the same derivative formula appears without additional boundary terms because the special defining function and parity kill the would-be contribution from varying the cutoff. Consequently, \(\mathrm{RenV}\) is nonincreasing whenever \(S(g(t))+n(n-1)\ge 0\), and it is constant on Einstein solutions [1607.08558] [1307.4788].

Across these literatures, a recurrent misconception is that “adapted renormalized volume” denotes one canonical formula. The record is more differentiated. In the AHE boundary-value problem, adaptation means choosing the slice \(v_n=\mathrm{average}\); in the compressible \(3\)-manifold problem, it means adding a precise \(\pi^3/\ell_\gamma\) correction; in the WP-gradient-flow setting, it means adapting the flow rather than the functional; and in singular Yamabe or minimal-submanifold theory, it means adapting the defining function, cutoff, or tubular geometry to the singular set [1211.6705] [2507.19291] [2003.00337].

This suggests a common structural principle rather than a universal definition. The finite part produced by renormalization is rarely canonical before auxiliary data are fixed. The role of adaptation is to select those auxiliary data — boundary representatives, special defining functions, surgery rules, or explicit correction terms — for which the renormalized volume acquires sharper properties: conformal naturality, exact differential, Lagrangian interpretation, boundedness under degeneration, Polyakov anomaly formulas, or local extremality.

Source: https://www.emergentmind.com/topics/adapted-renormalized-volume