2
$\begingroup$

Let $\bar{X}$ be a compact Riemann surface of genus $g$, and let $D:=\{p_{1},p_{2},\cdots,p_{n}\}$ be $n$ distinct points on $\bar{X}$. Define $X:=\bar{X}-D$ to be the punctured Riemann surface given by the complement of the divisor $D$. Assume that $2g-2+n>0$, which ensures that the surface $X$ supports a complete hyperbolic metric $h_{0}$.

Let $\sigma$ and $\rho$ be two smooth hyperbolic metrics on the $C^{\infty}$ surface $\bar{X}$. It is known that there exists a unique harmonic diffeomorphism from $(\bar{X},\sigma)$ to $(\bar{X}, \rho)$ in the isotropy class of the identity map $\rm{id}\colon\bar{X}\rightarrow \bar{X}$. This follows from work of Eells and Sampson, Hartman, Schoen and Yau. Moreover, Lohkamp extended this unique existence result to a punctured surface. That is, given a complete hyperbolic metric $h$ on the $C^{\infty}$ surface $X$, there is a unique diffeomorphism $f$, of $X$ homotopic to the identity map, such that $f$ is a harmonic map from $(X,h_{0})$ to $(X,h)$. An alternative proof is provided by Biswas, Ares-Gastesi and Govindarajan using parabolic Higgs bundle ([1], p.g. 1557). Below is the proof.

Proof. Let $(V, \nabla)$ be the flat rank two vector bundle given by the Fuchsian representation for the Riemann surface $(X,h)$. Let $H$ be the harmonic metric on $V$ given by the main theorem of [2] (p.g. 755 ) for the flat bundle $(V, \nabla)$ on the Riemann surface $(X,h_{0})$. In other words, $H$ gives a section, denoted by $s$, of the associated fiber bundle with fiber ${\rm SL}(2,\mathbb{R})/{\rm SO}(2)=\mathbb{H}$, where $\mathbb{H}$ is the upper half plane. This section $s$ gives the harmonic map $f$ mentioned above.

Question: How can we obtain the harmonic map $f$ from the section $s$ ?

  1. I. Biswas, P. Ares-Gastesi and S. Govindarajan, Parabolic Higgs bundles and Teichmuller spaces for punctured surfaces. Trans. Amer. Math. Soc. 163 (1972), 261--275.

  2. C. Simpson, Harmonic bundles on noncompact curves. J. Amer. Math. Soc. 3 (1990), 713--770.

$\endgroup$
5
  • $\begingroup$ You can get a harmonic map, but you would not know if it is a diffeomorphism (without quoting a suitable uniqueness theorem). Is this what you wanted to know? $\endgroup$ Commented Nov 30 at 16:02
  • 3
    $\begingroup$ In any case, the strongest existence results for harmonic diffeomorphisms are due to Vlad Markovi\'c who proved that every quasiconformal homeomorphism between Riemann surfaces (possibly of infinite type) is homotopic to a harmonic quasiconformal diffeomorphism; the uniqueness in this theorem was known earlier (I think, by Li and Tam). $\endgroup$ Commented Nov 30 at 16:18
  • $\begingroup$ Thank you for your comment. The section s is a section of a fiber bundle over X whose fibers are isomorphic to the upper half-plane. I would like to know how to obtain a map from X to X using the section s. $\endgroup$
    – Yu Feng
    Commented Dec 1 at 11:58
  • 2
    $\begingroup$ I guess you real question is: Given a flat fiber bundle $F\to E\to B$ and a section $s$ of this bundle, how does one obtain an equivariant map from the universal cover of $B$ to the fiber $F$. In your special case, $F$ happens to be the universal cover of $B$ and the holonomy representation $h$ of your flat bundle is discrete and faithful. Can you now put these two things together to get a map $B\to F/h(\pi_1(B))$? By the way, it is a terrible idea to treat Teichmuller space as the space of metrics on a fixed surface. Sadly, some people do this. $\endgroup$ Commented Dec 1 at 12:06
  • $\begingroup$ Got it, thank you. $\endgroup$
    – Yu Feng
    Commented Dec 1 at 13:15

0

You must log in to answer this question.