Trace coordinates of Teichmüller spaces of Riemann surfaces

Using an algebraic formalism based on matrices in SL(2,R), we explicitly give the Teichmüller spaces of Riemann surfaces of signature (0,4) (X pieces), (1,2) ("Fish" pieces) and (2,0) in trace coordinates. The approach, based upon gluing together two building blocks (Q and Y pieces), is then extended to tree-like pants decomposition for higher signatures (g,n) and limit cases such as surfaces with cusps or cone-like singularities. Given the Teichmüller spaces, we establish a set of generators of their modular groups for signatures (0,4), (1,2) and (2,0) in trace coordinates using transformations acting separately on the building blocks and an algorithm on dividing geodesics. The fact that these generators act particularly nice in trace coordinates gives further motivation to this choice (rather then the one of Fenchel-Nielsen coordinates). This allows us to solve the Riemann moduli problem for X pieces, "Fish" pieces and surfaces of genus 2; i.e. to give the moduli spaces as the fundamental domains for the action of the modular groups on the Teichmüller spaces. In this context, we also give an algorithm deciding whether two Riemann surfaces of signatures (0,4), (1,2) or (2,0) given by points in the Teichmüller space are isometric or not. As a consequence, we show the following two results concerning simple closed geodesics: On any purely hyperbolic Riemann surface (containing neither cusps nor cone-like singularities), the longest of two simple closed geodesics that intersect one another n times is of length at least ln, a sharp constant independent of the surface. We explicitly give ln for n = 1,2,3 and study its behaviour when n goes to infinity. X pieces are spectrally rigid with respect to the length spectrum of simple closed geodesics.

Trace coordinates of Teichmüller space of Riemann surfaces of signature $(0,4)$

Thomas Gauglhofer, Klaus-Dieter Semmler

Summary: We explicitly give

calT

, the Teichmüller space of four-holed spheres (which we call

X

pieces) in trace coordinates, as well as its modular group and a fundamental domain for the action of this group on

calT

which is its moduli space. As a consequence, we see that on any hyperbolic Riemann surface, two closed geodesics of lengths smaller than

2operatornamearccosh(2)

intersect at most once; two closed geodesics of lengths smaller than

2operatornamearccosh(3)

are both non-dividing or intersect at most once.

2005

Computational Approach to the Geometry of Compact Riemann Surfaces

Manuel Racle

The goal of this document is to provide a generalmethod for the computational approach to the topology and geometry of compact Riemann surfaces. The approach is inspired by the paradigms of object oriented programming. Our methods allow us in particular to model, for numerical and computational purposes, a compact Riemann surface given by Fenchel-Nielsen parameters with respect to an arbitrary underlying graph, this in a uniformand robust manner. With this programming model established we proceed by proposing an algorithmthat produces explicit compact fundamental domains of compact Riemann surfaces as well as generators of the corresponding Fuchsian groups. In particular, we shall explain how onemay obtain convex geodesic canonical fundamental polygons. In a second part we explain in what manner simple closed geodesics are represented in our model. This will lead us to an algorithm that enumerates all these geodesics up to a given prescribed length. Finally, we shall briefly overview a number of possible applications of our method, such as finding the systoles of a Riemann surface, or drawing its Birman-Series set in a fundamental domain.

EPFL2013

On the geometry of simple closed geodesics

Hugo Parlier

The leitmotif of this dissertation is the search for length formulas and sharp constants in relation with simple closed geodesics on hyperbolic compact Riemann surfaces. The main tools used are those of hyperbolic trigonometry, topological properties of simple closed geodesics and the convexity of geodesic length functions under the influence of twist transformations. The first question addressed is the question of their explicit density on a compact surface. The existence of a positive number ρ such that a simple closed geodesic on a given surface never crosses all disks of radius ρ is proved. A sharp bound on ρ is calculated, and this bound depends neither on the genus, nor on the choice of a surface. The second part concerns distances between boundary geodesics on surfaces of signature (1,2) and signature (0,4). These surfaces are natural building blocks for Riemann surfaces and lengths on such surfaces gives explicit information concerning twist parameters, contrary to surfaces of signature (0,3). Within the set of all surfaces with given boundary length and systole, maximal surfaces for distance between boundary geodesics are found. The last part is based on the search for surfaces that reach an upper bound among all surfaces of same genus and same systole for minimal length canonical homology bases. This direction has brought new light to the search for maximal surfaces for other problems as well, including properties that a surface incarnating Bers' partition constant would have.

EPFL2004