Overview
A Riemannian metric is a generalization of the inner product for a manifold. This is important since each tangent space might not be the βsameβ for different points on a manifold, so we need a notion of how inner products (and thus angles of linearizations) varies smoothly around the manifold.
Definition
Let be a smooth manifold. A Riemannian metric (usually denoted ) is a smooth assignment of inner products on tangent spaces for .
Formally, is a covariant 2 tensor field that is
-
Symmetric. for
-
Positive definite for such that .
Yet another way to think about a Riemannian metric is to think of it as a non-degenerate smooth section of the 2nd symmetric power .
With coordinates
Let be a chart on . Using this chart, we have the coordinate frame on such that a vector field .
A Riemannian metric is determined by
where .
Given Einstein notation, we have
Differences from symplectic form
On a vector space, we can always use the Gram Schmidt process to find an orthonormal basis . This acts as a global chart for the vector space, and the inner product in these coordinates is simply the identity matrix.
For a Riemannian manifold, , we can find an orthonormal frame for the tangent bundle, but donβt expect it to be a coordinate frame. In other words, there is no guarantee that we will always find coordinates that will all be orthonormal under .
This is because there can be local invariants in Riemannian manifolds. This is different from symplectic manifolds which by Darboux thoerem says that there are no local invariants and there will always be a local chart that makes the symplectic form into the βstandardβ symplectic form.
Examples
Consider the manifold . This is the normal βdot productβ or inner product on the vector space.
Therefore,
Using spherical coordinates
then
again!
Now using stereographic projection:
todo ADD EQUATIONS HEREβ¦
Using the above formulas, we have
Therefore, this gives the metric (after some algebraic manipulation)