# metrics compatible with conformal structures

I have three related questions:

(1) How does one describe the possible Riemannian metrics that are compatible with a conformal structure on a two dimensional surface?

(2) Can all conformable structures be realized through embeddings of the surface in Euclidean 3-space?

(3) How does one understand solutions to the equation

-exp(f) = Laplacian(f), i.e., $$- \mathrm{exp} (f) = \Delta f$$

where $f$ is a real-valued function of two variables in an open domain?

-
You should probably split this question up. In particular, the first question is too elementary and should be asked either locally or on math.stackexchange.com. The second question is a good one. I don't understand what you want in the third. – Deane Yang Apr 1 '11 at 17:03
I tried to make the posting more readable without changing the questions substantively. Roll-back or correct as you see fit. – Joseph O'Rourke Apr 1 '11 at 20:23
The answer to question (2) is an old thread: mathoverflow.net/questions/53999/… – Ryan Budney Apr 1 '11 at 22:14

## 1 Answer

I don't have much to say about (1) or (2), but (3) is, of course, a classical equation and the usual interpretation is this:

I'm assuming that you are talking about a domain in the $z$-plane, also known as the $xy$-plane, and that, by $\Delta f$ you mean the classical $f_{xx} + f_{yy}$, not the negative of this or this times some metric scalar (as some sources interpret $\Delta$). In this case, the general solution of $\Delta f = -\exp(f)$ can locally be written in the form

$$f = \log\left({8\ h'(z)\overline{h'(z)}}\over{(1+h(z)\overline{h(z)})^2}\right)$$

where $h$ is a (locally defined) holomorphic function on the domain. In the case that the domain is simply connected, one can take $h$ to be globally defined, but you may have to allow it to be meromorphic.

Geometrically, what is going on is that the conformal metric $ds^2 = \exp(f/2)(dx^2+dy^2)$ has constant curvature $K = 1$ and hence must be constructed by pulling back the standard metric on the $2$-sphere by a conformal map (which one can take to be orientation preserving (in the weak sense), so that it is actually a holomorphic map to the $2$-sphere, thought of as $\mathbb{C}\mathbb{P}^1$ and hence $ds^2$ is the pullback via $h$ of the $K=1$ conformal metric

$${{4}\over{(1+w\overline{w})^2}}\ dw\circ d\overline{w}\ .$$

-