Introduction
In the study of complex geometry, Riemann surfaces—one-dimensional complex manifolds—serve as fundamental objects of inquiry. These surfaces are often studied in conjunction with Abelian or quadratic differentials, which are essential tools in theoretical physics (such as string theory) and dynamical systems (like billiards in polygons). A central challenge in this field involves understanding the "moduli space," the parameter space representing all possible complex structures (or "shapes") a Riemann surface of a given genus can assume.
This research investigates the variational formulas for normalized Abelian differentials and the matrix of $b$-periods within Hurwitz spaces and the moduli spaces of differentials. The primary objective is to determine how these geometric objects evolve as the parameters of the differential are varied, specifically seeking "directions" in the parameter space that allow the differential to change while the underlying Riemann surface remains fixed.
Problem Statement: Preserving the Base Surface
When considering the moduli space of Riemann surfaces equipped with quadratic differentials, the total space is significantly larger than the moduli space of the surfaces alone. A long-standing problem in the field is the identification of variational formulas that describe the calculus of the surface's periods—integrals over the surface—as the differential is tweaked.
Specifically, the author addresses the problem of finding a complete set of local vector fields on the moduli space of quadratic differentials that preserve the complex structure of the base Riemann surface. In simpler terms: how can we modify the "differential" part of the pair (surface, differential) without distorting the "surface" part?
Methodology
The thesis employs a combination of classical complex analysis and modern geometric techniques:
- Ahlfors-Rauch Variational Formulas: These formulas provide the standard framework for calculating the directional derivatives of the period matrix of a Riemann surface.
- Canonical Double Covering: Quadratic differentials are "squared" forms and are more complex to handle than Abelian differentials. To simplify the problem, the author constructs a two-sheeted covering of the original surface. On this larger "covering" surface, the quadratic differential can be lifted and treated as a simpler Abelian differential.
- Homological Coordinates: The author utilizes integrals of the differential over a basis of the surface's homology (cycles or loops) to establish a local coordinate system.
- Symplectic Geometry and Involution: By analyzing the intersection properties of cycles on the surface, the author decomposes the homology into invariant and anti-invariant subspaces under the action of the double-cover's involution.
Key Insights: The Role of the Double Cover
A pivotal insight of this work is that the complex structure of the base Riemann surface is entirely encoded within the "invariant" portion of the period matrix of its canonical double cover. This realization allows the researcher to isolate the variables that affect the surface's shape from those that only affect the differential.
Furthermore, the author draws upon a connection to Noether’s Theorem. For non-hyperelliptic surfaces (surfaces that are not essentially double covers of a sphere), the products of Abelian differentials span the entire space of quadratic differentials. This property is crucial for proving that the proposed vector fields are indeed complete and span the necessary directions in the moduli space.
Main Result: Vector Fields on the Principal Stratum
The crowning achievement of the thesis is the derivation of a complete set of $3g-3$ independent local vector fields, denoted as $W_l$, on the non-hyperelliptic connected component of the principal stratum of the moduli space of quadratic differentials.
The formula for these vector fields is given by: $$W_l = \frac{\partial}{\partial \eta_l} + \sum_{k=1}^{3g-3} \mathbb{B}^-_{lk} \frac{\partial}{\partial \eta_{3g-3+k}}$$
In this expression, $\mathbb{B}^-$ represents the "anti-invariant" period matrix. By using the entries of this specific matrix as coefficients for a linear combination of coordinate derivatives, the author successfully "cancels out" any infinitesimal changes to the base surface's moduli. These vector fields provide a mathematical "map" for navigating the space of differentials while keeping the underlying geometry of the Riemann surface perfectly constant.
Limitations and Future Directions
While the thesis provides a comprehensive solution for the non-hyperelliptic component, several areas remain for future exploration:
- The Hyperelliptic Gap: The current results specifically exclude the hyperelliptic connected component of the moduli space. Identifying similar vector fields for these specific surface types remains an open challenge.
- Coordinate Splitting: The author suggests that a future goal is to find a coordinate system where these vector fields simplify into basic partial derivatives. This would imply a natural "splitting" of the parameter space into surface-shape coordinates and differential-detail coordinates.
- Higher-Order Differentials: The current work is focused on quadratic differentials. Extending these variational formulas to higher-order differentials (such as cubic or quartic forms) represents a logical next step in the research.
Summary updated