3
$\begingroup$

I'm currently trying to understand Dennis Deturck's and Jerry Kazdan's proof of the existence of isothermal coordinates for $2$-dimensional Riemannian manifolds in Some regularity theorems in Riemannian geometry, p. 251, but I'm kinda stuck. On a Riemannian manifold $(M, g)$, coordinate functions $(x_1, \dots, x_n)$ are called isothermal if the metric is locally conformal to the Euclidean metric in these coordinates, that is, if $g$ is locally of the shape $d s^2 = \lambda(x_1, \dots, x_n) (d x_1^2 + \dots + d x_n^2)$ for some $\lambda(x_1, \dots, x_n) > 0$.

Firstly, Deturck and Kazdan prove that on any Riemannian manifold, harmonic coordinate functions (where the Lacplace operator of the function is $= 0$) exist everywhere locally. Now to prove the existence of isothermal coordinates in the $2$-dimensional case, this allows to choose a harmonic function $u$ with non-vanishing derivative as the first component. So far I can follow.

Now in the next step, they choose the second coordinate function $v$ using the Hodge star operator $\star$. Since $d u$ is a $1$-form and the Hodge star sends $1$-forms to $1$-forms, $\star d u $ is also a $1$-form. Now because $\star du$ is closed, the Poincaré lemma ensures the existence of a coordinate function $v$ satisfying $d v = \star du$. Finally, they show that $v$ is also harmonic through a simple calculation.

Now my question is: how can I see that these coordinate functions $(u, v)$ constructed this way are actually isothermal? Deturck and Kazdan don't seem to elaborate on that unless I have missed it somewhere, so maybe it's very easy and just I'm missing something trivial, but I've been trying to wrap my head around why the off-diagonal entries of the metric become $= 0$ when representing it with respect to these coordinate functions. I'm not sure how exactly one can deduce that from the fact that $u$ and $v$ are harmonic and that $d v = \star d u$.

| cite | improve this question | | | | |
$\endgroup$
3
$\begingroup$

The relation $dv = \star du$ implies that $du, dv$ are orthogonal, since (letting $\omega$ denote the Riemannian area form) the definition of the Hodge star gives $$\langle du, dv \rangle \omega= \langle du, \star du\rangle \omega = du \wedge du = 0.$$

In other words, the matrix $g^{ij}$ representing the inverse metric in the coordinates $(u,v)$ is diagonal. Since the inverse of a diagonal matrix is diagonal, this means that $g_{ij}$ is diagonal too.

| cite | improve this answer | | | | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.