Riemannian Metrics on Smooth Manifolds

Subjects: Differential Geometry
Links: The Tensor Bundles, Correlations, Musical Isomorphisms, Embedded Smooth Submanifolds, Local and Global Sections of Vector Bundles

Def: Let M be a smooth manifold with or without boundary. A Riemannian metric on M is a smooth symmetric covariant 2-tensor field on M that is positive definite at each point. A Riemannian manifold is a pair (M,g), where M is a smooth manifold and g is a Riemannian metric on M. A Riemannian manifold with boundary is defined similarly.

If g is a Riemannian metric on M, then for each pM, the 2-tensor gp is an inner product on TpM. Because of this we use the notation v,wg to denote the real number gp(v,w) for v,wTpM.

In any smooth local coordinates (xi), a Riemannian metric can be written as $$g = g_{ij} dx^i \otimes dx^j,$$but we can make use of the symmetric product to get $$g = g_{ij} dx^i \otimes dx^j = g_{ij} dx^i dx^j. $$
Def: The simplex example of a Riemannian metric is the Euclidean metric g¯ on Rn, given in standard coordinates by $$\bar g = \delta_{ij} dx^i dx^j $$

Def: If (M1,g1) and (M2,g2) are Riemannian manifolds, then product M1×M2 has a natural Riemannian metric g=g1g2, called the product metric, defined by $$g(X_1 + X_2, Y_1 + Y_2) := g_1(X_1, Y_1) + g_2(X_2, Y_2),$$where Xi,YiTpiMi, under the natural identification T(p1,p2)(M1×M2)=Tp1M1×Tp2M2.

Obs: Any local coordinates (x1,,xn) for M1 and (xn+1,,xn+m) for M2 give coordinates (x1,xn+m) for M1×M2. In terms of these coordinates the product metric has the local expression g=gijdxidxj, where gij is a block diagonal matrix $$(g_{ij}) = \begin{pmatrix}
(g_1){ij} & 0 \
0 & (g_2)

\end{pmatrix}$$
Existence of Riemannian Metrics: Every smooth manifold with or without boundary admits a Riemannian metric.

We have a few of the geometric constructions that can be defined on a Riemannian manifold (M,g) with or without boundary.

Def: Let (M,g) be a an n-dimensional Riemannian manifold with or without boundary. We say that a local frame (E1,,En) for M on an open subset UM is an orthonormal frame if the vectors (E1|p,,En|p) from an orthonormal basis for TpM at each point pU, or equivalently if Ei,Ej=δij.

Prop: Suppose (M,g) is a Riemannian manifold with or without boundary, and (Xj) is a smooth local frame for M over an open subset UM. Then there is a smooth orthonormal frame (Ei) over U such that span(E1|p,,Ej|p)=span(X1|p,,Xj|p) for each j=1,,n and each pU.

Existence of Local Orthonormal Frames: Let (M,g) be a Riemannian manifold with or without boundary. For each pM, there is a smooth orthonormal frame on a neighbourhood of p.

Pullback Metrics

Def: Suppose M, N are smooth manifolds with or without boundary, g is a Riemannian metric on N, and F:MN is smooth. The pullback Fg is a smooth 2-tensor field on M. If it is positive definite, it is a Riemannian metric on M, calle the pullback metric determined by F.

Prop: Suppose F:MN is a smooth map and g is a Riemannian metric on N. Then Fg is a Riemannian metric on M iff F is a smooth immersion.

Def: If (M,g) and (M~,g~) are both Riemannian manifolds, a smooth map F:MM~ is called a Riemannian isometry if it is a diffeomorphism that satisfies Fg~=g. More generally, F is called a local isometry if every point pM has a neighbourhood U such that F|U is an isometry of U onto an open subset of M~.

If there exists a Riemannian isometry between (M,g) and (M~,g~), we say that they are isometric as Riemannian manifolds. If each point of M has a neighbourhood that is isometruc to an open subset of (M~,g~), then we say that (M,g) is locally isometric to (M~,g~). The study of properties of Riemannian manifolds that are invariant under (local or global) isometries is called Riemannian geometry.

A Riemannian n-manifold (M,g) is said to be a flat Riemannian manifold, and g is a flat metric, if (M,g) is locally isometric (Rn,g¯).

Def: An isometry φ:(M,g)(M,g) is called an isometry of M. the set of isometries of M is a group, called the isometry group of M; it is denoted I(M).

Th: Let (M,g) be a Riemmanian manifold. The following statements are equivalent:

Lemma: Suppose U,VRn are connected open sets, φ,ψ:UV are Riemannian isometries, and for some pU they satisfy φ(p)=ψ(p) and dφp=dψp, then φ=ψ.

Prop: The set of maps from Rn to itself given by the action of E(n), the Euclidean Group, on Rn is the full group of Riemannian isometries of (Rn,g¯).

Def: Let (M,g) be a Riemannian manifold. A smooth vector field V on M is called a Killing vector field for g, named after the late nineteenth/early twentieth-century German mathematician Wilhelm Killing, if the flow of V acts by isometries of g.

Obs: We see that a smooth vector field is a Killing vector field iff Lvg=0.

Prop: The set of all Killing vectors on M constitutes a Lie subalgebra of X(M).

Prop: Let (M,g) be a Riemannian manifold, and V is a smooth vector field on M. V is a Killing vector field iff it satisfies the following equation in each smooth local coordinate chart$$V^k \frac{\partial g_{ij}}{\partial x^k} + g_{jk} \frac{\partial V^k}{\partial x^i} + g_{ik} \frac{\partial V^k}{\partial x^j} = 0$$
Prop: Let KX(Rn) denote the Lie algebra of Killing vector fields with respect to the Euclidean metric. Then we see that a vector field is a Killing vector field iff $$\frac{\partial V^i}{\partial x^ j}+ \frac{\partial V^j}{\partial x^i} = 0.$$This means that the following vector fields form a basis for K: $$\frac{\partial}{\partial x^i}, \quad 1 \le i \le n; \qquad x^i\frac{\partial}{\partial x^j}- x^j\frac{\partial}{\partial x^i}, \quad 1\le i < j \le n. $$These vector fields represent translation, and rotations. If we consider K0K denote the subspace consisting of fields that vanish at the origin. Then the map $$V \mapsto \left(\frac{\partial V^i}{\partial x^j}(0)\right) $$is an injective linear map from K0 to so(n).

Submanifolds

Pullback metrics are particularly important for submanifolds.

Def: If (M,g) is a Riemannian manifold with or without boundary, every submanifold SM automatically inherits a pullback metric ιg, where ι:SM is inclusion. We know that ιg is just the restriction of g to pairs of vectors tangent to S. With this metric, S is called a Riemannian submanifold (with or without boundary) of M.

Example: The metric g˙=ιg¯ induced on Sn by the usual inclusion ι:SnRn+1 is called the round metric, or the standard metric, on the sphere.

Computations on a submanifold are usually most conveniently carried out in terms a local parametrization: this is an embedding of an open subset URn into N, whose image is an open subset of M.

Example: If X:URn is a parametrization of a submanifold MRn with the induced metric, the induced metric in standard coordinates (u1,,un) on U is just $$g = \sum_{i = 1}^n (dX^i)^2 = \sum_{i = 1}^n \left(\frac{\partial X^i}{\partial u^j} du^j\right)^2.$$
Induced Metrics on Surfaces of Revolution: Let C be an embedded 1-dimensional submanifold of the half plane {(r,z)r>0}, and let SC be the surface of revolution generated by C is given by $$S_C := \left{(x, y, z) \in\Bbb R^3 ;\left\rvert; \left(\sqrt{x^2+ y^2}, z\right)\right.\in C\right}. $$To compute the induced metric on SC, we can choose a smooth local parametrization γ(t):=(a(t),b(t)) for XC, and note that the map X(t,θ)=(a(t)cosθ,a(t)sinθ,b(t)) yields a smooth local parametrization of SC, provided (t,θ) is restricted to a sufficiently small open subset of the plane. Then $$\begin{align*}
X^* g &= d((a'(t) \cos\theta ,dt - a(t) \sin\theta ,d\theta)^2 + d((a'(t) \sin\theta dt + a(t) \cos\theta d\theta)^2 + (b'(t) dt)^2 \
&= (a'(t)^2 + b'(t)^2) dt^2+ a(t)^2d\theta^2.
\end

In particular, if $\gamma$ is a *unit-speed curve*, meaning that $|\gamma'(t)| = 1$, then it reduces to $X^*g = dt^2 + a(t)^2 d\theta^2$. **Flatness Criterion for Surfaces of Revolution:** Let $C\subseteq H$ be a connected embedded $1$-dimensional submanifold of the half-plane $H:= \{(r, z) \mid r > 0\}$, and let $S_C$ be the surface of revolution generated by $C$. The induced metric on $S_C$ is flat iff $C$ is part of the a straight line. **Cor:** The round or standard metric on $\Bbb S^2$ not flat. # The Normal Bundle Suppose $(M, g)$ is an $n$-dimensional Riemannian manifold with or without boundary, and $S\subseteq M$ is a $k$-dimensional submanifold with or without boundary. Just as we did for submanifolds of $\Bbb R^n$, for any $p\in S$ we say that a vector $v\in T_p M$ is *normal to $S$* if $v$ is orthogonal to every vector in $T_p S$ with respect to the inner product $\langle \cdot, \cdot \rangle_g$. The *normal space to $S$ at $p$* is the subspace $N_p S \subseteq T_p M$ consisting of all vectors that are normal to $S$ at $p$, and the *normal bundle of $S$* is the subset $NS \subseteq TM$ consisting of the union of all the normal spaces at points of $S$. The projection $\pi_{NS}:NS \to S$ is defined as the restriction to $NS$ of $\pi: TM \to M$. **The Normal Bundle to a Riemannian Submanifold:** Let $(M, g)$ be a Riemannian $n$-manifold with or without boundary. For any immersed $k$-dimensional submanifold $S\subseteq M$ with or without boundary, the normal bundle $NS$ is a smooth rank-$(n-k)$ subbundle of $TM|_S$. For each $p\in S$, there is a smooth frame for $NS$ on a neighbourhood of $p$ that is orthonormal with respect to $g$. **Def:** If $S\subseteq M$ is a Riemannian submanifold, we define the *normal bundle* to $S$ as $$NS := \coprod_{p\in S} N_p S.

Existence of Adapated Orthonormal Frames: Let SM be an embedded Riemannian submanifold of a Riemannian manifold (M,g). For each pS, there is a smooth adapted orthonormal frame on a neighbourhood p in M.

Prop: Suppose (M,g) is an Riemannian manifold and SM an immersed k-dimensional manifold. The ambient tangent bundle TM|S is isomorphic to the Whitney sum TSNS, where NSS is the normal bundle.

Def: Suppose π:NM is a smooth covering map. A covering transformation, or deck transformation is a smooth map φ:NN such that πφ=π. If g is Riemannian metric on M, then h:=πg is a Riemannian metric on N that is invariant under all covering transformations. In this case h is called a covering metric, and π is called Riemannian covering.

Prop: If (M,g) is a Riemannian n-manifold with or without boundary, let UMTM be the subset {(x,v)TM|v|g=1}, called the unit tangent bundle of M. Then UM is a smooth fibre bundle over M with model fibre Sn1.