ConceptComplete

Curvature of Riemannian Manifolds

Curvature measures how a Riemannian manifold deviates from flat Euclidean space. It comes in several flavors -- sectional, Ricci, and scalar -- each capturing different aspects of the geometry.


The Riemann Curvature Tensor

Definition6.6Riemann curvature tensor

The Riemann curvature tensor of a Riemannian manifold (M,g)(M, g) with Levi-Civita connection \nabla is the (1,3)(1,3)-tensor

R(X,Y)Z=XYZYXZ[X,Y]Z,R(X, Y)Z = \nabla_X \nabla_Y Z - \nabla_Y \nabla_X Z - \nabla_{[X,Y]} Z,

for vector fields X,Y,ZX, Y, Z. In components: Rlijk=iΓjkljΓikl+ΓimlΓjkmΓjmlΓikmR^l{}_{ijk} = \partial_i \Gamma^l_{jk} - \partial_j \Gamma^l_{ik} + \Gamma^l_{im}\Gamma^m_{jk} - \Gamma^l_{jm}\Gamma^m_{ik}. The curvature measures the failure of parallel transport to be path-independent.

RemarkSymmetries of the curvature tensor

The Riemann tensor satisfies:

  1. Rijkl=RjiklR_{ijkl} = -R_{jikl} (skew-symmetry in first pair).
  2. Rijkl=RijlkR_{ijkl} = -R_{ijlk} (skew-symmetry in second pair).
  3. Rijkl=RklijR_{ijkl} = R_{klij} (pair symmetry).
  4. Rijkl+Riklj+Riljk=0R_{ijkl} + R_{iklj} + R_{iljk} = 0 (first Bianchi identity).

These reduce the independent components of RR from n4n^4 to n2(n21)12\frac{n^2(n^2-1)}{12} in dimension nn.


Types of Curvature

Definition6.7Sectional curvature

For a 2-plane σ=span(u,v)TpM\sigma = \operatorname{span}(u, v) \subset T_pM, the sectional curvature is

K(σ)=K(u,v)=g(R(u,v)v,u)g(u,u)g(v,v)g(u,v)2.K(\sigma) = K(u, v) = \frac{g(R(u,v)v, u)}{g(u,u)g(v,v) - g(u,v)^2}.

This is the Gaussian curvature of the surface obtained by exponentiating σ\sigma. The sectional curvatures determine the full curvature tensor.

Definition6.8Ricci and scalar curvature

The Ricci curvature is the trace of the curvature tensor: Ric(u,v)=i=1ng(R(ei,u)v,ei)\mathrm{Ric}(u, v) = \sum_{i=1}^n g(R(e_i, u)v, e_i) for an orthonormal basis {ei}\{e_i\}. In components: Rij=RkikjR_{ij} = R^k{}_{ikj}.

The scalar curvature is the trace of the Ricci tensor: S=gijRij=iRic(ei,ei)S = g^{ij} R_{ij} = \sum_{i} \mathrm{Ric}(e_i, e_i).

ExampleConstant curvature spaces

The three model geometries:

  1. Rn\mathbb{R}^n (Euclidean): K=0K = 0 everywhere. Flat space.
  2. Sn(r)S^n(r) (sphere of radius rr): K=1/r2K = 1/r^2 everywhere. Positive curvature.
  3. Hn\mathbb{H}^n (hyperbolic): K=1K = -1 everywhere. Negative curvature.

By Schur's lemma, if the sectional curvature KK is pointwise constant on a connected manifold of dimension 3\geq 3, then KK is globally constant.


Curvature and Topology

Theorem6.4Gauss-Bonnet theorem (dimension 2)

For a compact oriented Riemannian 2-manifold (M,g)(M, g),

MKdvolg=2πχ(M),\int_M K \, \mathrm{dvol}_g = 2\pi \chi(M),

where KK is the Gaussian curvature and χ(M)\chi(M) is the Euler characteristic. This remarkable formula connects local geometry (curvature) to global topology (Euler characteristic).

RemarkHigher-dimensional Gauss-Bonnet

The Chern-Gauss-Bonnet theorem generalizes this to even-dimensional manifolds: χ(M)=MPf(Ω)\chi(M) = \int_M \mathrm{Pf}(\Omega), where Pf(Ω)\mathrm{Pf}(\Omega) is the Pfaffian of the curvature form. This is a prototype of index theorems connecting analysis and topology.