7
$\begingroup$

Let $C$ be a curve over $k$ and $w_C$ it's dualizing sheaf. If $dim_k H^0(C, \mathcal{O}_C) =1$ and $g:= H^0(C, \mathcal{O}_C)$ the arithmetic genus one easy computes $$deg(w_C) = 2g-2$$ where $deg$ is the map $deg:Pic(C) \to \mathbb{Z}, \mathcal{L} \mapsto \chi(\mathcal{L}) - \chi(\mathcal{O}_C)$.

Fothermore, using methods from simplicial cohomology one get the Euler characteristic $e(C) = 2-2g$ where $g$ is the topological genus of $C$.

I want to know a (maybe) geometrical but also a preferably deeper reason where the relation $deg(w_f) = -e(C)$ comes from, so why this invariants a so closed connected.

Improve this question
$\endgroup$

2 Answers 2

Reset to default
3
$\begingroup$

Another explanation for the equality $\mathrm{deg}(\omega_C) = -e(C)$, complementing that contained in Javier-Alvarez's MSE exposition linked by Francesco Polizzi, is Hodge theory. In short, Hodge theory gives an instance and a precise statement of how topological invariants are related to algebraic/holomorphic invariants. Specifically, Hodge theory tells you how to represent singular cochains as integrals with respect to holomorphic forms. More practically, Hodge theory expresses topological (singular) cohomology in terms of sheaf cohomology of $\mathcal{O}_C$ and $\Omega_C^1$.

But before I begin, let me clarify the setting in which I will be working in:

  • My field will be $k = \mathbf{C}$, so that I can talk about topology.
  • The symbol $C$ refers to the projective scheme over $\mathbf{C}$. I will write $C(\mathbf{C})$ to denote the Riemann surface associated with $C$.
  • By "curve", I will mean "smooth projective curve over $\mathbf{C}$". Smoothness is required in general, as the relation $\mathrm{deg}(\omega_C) = -e(C)$ already fails for most nodal curves.

For a smooth projective curve $C$, Hodge theory (together with Serre's GAGA) says that there are canonical isomorphisms \begin{align*} H^0(C(\mathbf{C}),\mathbf{C}) & \cong H^0(C,\mathcal{O}_C), \\ H^1(C(\mathbf{C}),\mathbf{C}) & \cong H^1(C,\mathcal{O}_C) \oplus H^0(C,\Omega_C^1), \\ H^2(C(\mathbf{C}),\mathbf{C}) & \cong H^1(C,\Omega_C^1), \end{align*} where

  • the cohomology groups on the left are the topological (singular) cohomology groups of the Riemann surface $C(\mathbf{C})$ with coefficients in $\mathbf{C}$;
  • those on the right are sheaf cohomology groups computed on the scheme $C$; and
  • $\Omega_C^1$ is the sheaf of Kähler differentials of $C$.

Thus we obtain the relation \begin{align*} e(C) & = \sum\nolimits_{i = 0}^2 (-1)^i \dim_{\mathbf{C}}(H^i(C(\mathbf{C},\mathbf{C})) \\ & = \sum\nolimits_{i = 0}^1 (-1)^i \dim_{\mathbf{C}}(H^i(C,\mathcal{O}_C)) - \sum\nolimits_{i = 0}^1 (-1)^i \dim_{\mathbf{C}}(H^i(C,\Omega_C^1)) \\ & = \chi(\mathcal{O}_C) - \chi(\Omega_C^1) = -\mathrm{deg}(\Omega_C^1). \end{align*}

Since $C$ is smooth, the dualizing sheaf can be explicitly determined: by, say, Hartshorne, Algebraic Geometry, Corollary III.7.12, $\omega_C \cong \Omega_C^1$. Putting this into the relation above gives $$ \mathrm{deg}(\omega_C) = -e(C). $$

Improve this answer
$\endgroup$
2
$\begingroup$

The equality $$\mathrm{deg}(K_C)=-e(C)$$ is the consequence of a deep interplay between several notions of degree coming from topology, geometry and algebra, and provides the connection between different index theorems for compact, oriented $2$-manifolds.

A beautiful exposition of this topic can be found in Javier-Alvarez's answer to this MSE question.

Improve this answer
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .