06.04.01 · riemann-surfaces / riemann-roch-rs

Riemann-Roch theorem for compact Riemann surfaces

shipped3 tiersLean: partial

Anchor (Master): Forster; Griffiths-Harris; Farkas-Kra; Miranda

Intuition [Beginner]

The Riemann-Roch theorem for compact Riemann surfaces is the analytic version of the algebraic theorem 04.04.01. On a compact Riemann surface $X$ of genus $g$ — topologically a $g$ -holed surface — it tells you exactly how many meromorphic functions exist with prescribed pole and zero behaviour.

The headline formula is the same as in the algebraic case: for a divisor $D$ on $X$ of degree $d$ ,

ℓ (D) - i (D) = d - g + 1,

where $ℓ (D)$ is the dimension of the space of meromorphic functions allowed by $D$ and $i (D)$ is the dimension of the space of holomorphic 1-forms vanishing along $D$ (the index of speciality).

The fact that the analytic statement coincides with the algebraic one is a deep instance of Serre's GAGA principle: on a compact Riemann surface, complex analysis and algebraic geometry give the same dimension formulas.

Visual [Beginner]

A compact Riemann surface of genus $g$ with marked divisor data; Riemann-Roch counts meromorphic functions and holomorphic 1-forms compatible with the divisor.

Worked example [Beginner]

On the Riemann sphere $\hat{C}$ (genus 0): for a divisor $D$ of degree $d \geq 0$ , $ℓ (D) = d + 1$ and $i (D) = 0$ (no holomorphic 1-forms — the canonical bundle has degree $- 2$ ). Riemann-Roch: $ℓ (D) - 0 = d - 0 + 1 = d + 1$ .

A natural example: meromorphic functions on $\hat{C}$ with at most a triple pole at $\infty$ form the space of polynomials of degree $\leq 3$ , which has dimension 4. Riemann-Roch confirms: $ℓ (3 \cdot \infty) = 3 + 1 = 4$ .

On a torus $X = C /Λ$ (genus 1): for a single point $D = P$ , $ℓ (P) = 1$ and $i (P) = 0$ . So Riemann-Roch: $1 - 0 = 1 - 1 + 1$ . ✓ This means: the only function with at most a simple pole at one point is the constant — to construct genuinely meromorphic functions on a torus, you need at least a double pole. This is the origin of the Weierstrass $℘$ -function.

Check your understanding [Beginner]

Formal definition [Intermediate+]

Let $X$ be a compact connected Riemann surface of genus $g = dim_{C} Ω_{hol}^{1} (X)$ . For a divisor $D = \sum_{p} n_{p} \cdot p$ on $X$ (a formal $Z$ -linear combination of points), the associated line bundle $O (D)$ has global sections

H^{0} (X; O (D)) = {f \in M (X) : (f) + D \geq 0} \cup {0},

where $M (X)$ is the field of meromorphic functions on $X$ and $(f) = \sum_{p} ord_{p} (f) \cdot p$ is the divisor of $f$ .

Definitions.

$ℓ (D) := dim_{C} H^{0} (X; O (D))$ .
$i (D) := dim_{C} H^{0} (X; ω_{X} \otimes O (- D)) = dim_{C} {ω \in Ω_{hol}^{1} (X) : (ω) \geq D}$ , the index of speciality.
$K_{X}$ : the canonical divisor (any divisor of a non-zero meromorphic 1-form on $X$ ).
$de g (D) := \sum_{p} n_{p}$ , the degree.

Theorem (Riemann-Roch on compact Riemann surfaces). For any divisor $D$ on a compact connected Riemann surface $X$ of genus $g$ ,

ℓ (D) - i (D) = de g (D) - g + 1.

Equivalent forms.

Serre duality form: $i (D) = ℓ (K_{X} - D)$ , so $ℓ (D) - ℓ (K_{X} - D) = de g (D) - g + 1$ — the algebraic form.
Euler characteristic form: $χ (O (D)) = h^{0} (D) - h^{1} (D) = de g (D) - g + 1$ , where $h^{i} (D) = dim_{C} H^{i} (X; O (D))$ .
Cohomological: $H^{1} (X; O (D)) ≅ H^{0} (X; ω_{X} \otimes O (- D))^{*}$ (Serre duality), so $h^{1} (D) = i (D)$ .

Corollaries.

The canonical divisor has degree $de g (K_{X}) = 2 g - 2$ , by applying Riemann-Roch to $D = K_{X}$ .
For $de g (D) > 2 g - 2$ , $i (D) = 0$ , and Riemann-Roch reduces to $ℓ (D) = de g (D) - g + 1$ .
For $de g (D) < 0$ , $ℓ (D) = 0$ .

Key theorem with proof [Intermediate+]

Theorem (Riemann-Roch, classical analytic proof on compact Riemann surfaces). For any divisor $D$ on a compact connected Riemann surface $X$ of genus $g$ , $ℓ (D) - i (D) = de g (D) - g + 1$ .

Proof sketch (induction on the support of $D$ ). This proof parallels the algebraic version 04.04.01 but uses analytic methods (Hodge theory, harmonic forms).

Base case. For $D = 0$ , $ℓ (0) = 1$ (the only globally holomorphic functions on a compact Riemann surface are constants, by maximum modulus) and $i (0) = g$ (the dimension of holomorphic 1-forms is the genus). So $1 - g = 0 - g + 1$ . ✓

Inductive step. Suppose Riemann-Roch holds for $D$ . For a point $P$ , consider $D^{'} = D + P$ .

A short exact sequence of sheaves:

0 \to O (D) \to O (D + P) \to C_{P} \to 0,

where $C_{P}$ is the skyscraper sheaf at $P$ with stalk $C$ . Taking the long exact sequence and the alternating sum of dimensions:

χ (O (D + P)) = χ (O (D)) + χ (C_{P}) = χ (O (D)) + 1.

Combined with $de g (D + P) = de g (D) + 1$ , by induction:

χ (O (D + P)) = (de g (D) + 1 - g) + 1 = de g (D + P) + 1 - g .

The same induction handles $D - P$ . By Serre duality $χ = h^{0} - h^{1} = ℓ - i$ , completing the induction. $□$

In the classical formulation, the use of Hodge theory (every cohomology class is represented by a harmonic form, and $dim H^{1} = dim Ω_{hol}^{1}$ ) replaces the abstract Serre duality.

Bridge. The construction here builds toward later units of the strand, where the same pattern is taken up at higher structure. The defining pattern appears again in those units in a sharpened form, where the local data is glued or quotiented. Putting these together, the foundational insight is that the data of this unit gives the structural signature that the rest of the strand reads off.

Exercises [Intermediate+]

Exercise 3 (medium, proof).

Show that on a compact Riemann surface of genus $g \geq 1$ , every meromorphic function with exactly one simple pole somewhere is a constant.

Hint

A function with exactly one simple pole has $de g (div pole) = 1$ . Apply Riemann-Roch.

Answer

A function $f$ with exactly one simple pole at $P$ has divisor $(f) = - P + (zeros)$ , with the total degree of the divisor of $f$ equal to 0 (degrees of pole and zero divisors are equal on a compact surface). So $f$ has exactly one simple zero somewhere. The function defines a meromorphic map $f : X \to \hat{C}$ of degree 1, hence a biholomorphism — this would force $X ≅ \hat{C}$ , hence $g = 0$ , contradicting $g \geq 1$ .

Equivalently: $ℓ (P) = 1$ for $g \geq 1$ (Riemann-Roch with $D = P$ , $de g = 1$ , $i (P) \geq 0$ , gives $ℓ \leq 1 - g + 1 + i (P) \leq 1$ for $g \geq 1$ since $i (P) \leq i (0) = g$ minus... actually let's redo: for $g = 1$ , $ℓ (P) - i (P) = 1 - 1 + 1 = 1$ ; $i (P) = 0$ as in Ex. 2; so $ℓ (P) = 1$ , only constants).

Exercise 5 (medium, conceptual).

Explain the relationship between "Riemann-Roch on a compact Riemann surface" and "Riemann-Roch on a smooth projective curve" via Serre's GAGA.

Hint

GAGA equates the categories of compact complex manifolds and smooth complex projective varieties (in dimension 1).

Answer

Serre's GAGA (Géométrie Algébrique et Géométrie Analytique, 1956) establishes equivalences:

Compact Riemann surfaces ↔ Smooth complex projective curves.
Holomorphic line bundles ↔ Algebraic line bundles ↔ Divisor classes.
Holomorphic sheaf cohomology ↔ Algebraic sheaf cohomology.

Under these equivalences, the analytic Riemann-Roch ( $ℓ (D) - i (D) = de g (D) - g + 1$ ) and the algebraic Riemann-Roch ( $ℓ (D) - ℓ (K - D) = de g (D) - g + 1$ ) become the same theorem. The analytic dimension count and the algebraic dimension count are computed via different machinery (Hodge theory vs. derived functors), but yield identical numbers.

GAGA is one of the foundational equivalences of modern mathematics, making algebraic geometry and complex analysis interchangeable on compact projective varieties.

Exercise 7 (hard, conceptual).

State the Riemann-Roch theorem in its modern Euler-characteristic form $χ (O (D)) = de g (D) + 1 - g$ and explain why it generalises to higher-dimensional varieties (Hirzebruch-Riemann-Roch).

Hint

The Euler characteristic is integration of the Chern character against the Todd class.

Answer

For a coherent sheaf $F$ on a compact complex manifold $X$ of complex dimension $n$ , the Hirzebruch-Riemann-Roch formula states:

χ (X; F) = \int_{X} ch (F) \cdot td (X),

where $ch$ is the Chern character and $td (X)$ is the Todd class of the tangent bundle.

For curves ( $n = 1$ ) and a line bundle $L = O (D)$ : $ch (L) = 1 + c_{1} (L)$ , $td (X) = 1 + \frac{1}{2} c_{1} (T_{X})$ . Multiplying and integrating: $χ (L) = de g (L) + \frac{1}{2} de g (T_{X}) = de g (D) + (1 - g)$ — exactly the classical formula.

The HRR formula generalises Riemann-Roch to all compact complex manifolds. Grothendieck-Riemann-Roch further generalises it to morphisms between smooth projective varieties, expressing the fibrewise Euler characteristic in terms of pushforwards of Chern characters. These deep generalisations show Riemann-Roch as a topological-cohomological theorem at heart.

Lean formalization [Intermediate+]

lean_status: partial — Mathlib has the Hodge-theoretic infrastructure (de Rham cohomology, sheaves) but the analytic Riemann-Roch is not yet formalised in full generality.

[object Promise]

Advanced results [Master]

Riemann-Hurwitz formula. For a holomorphic non-constant map $f : X \to Y$ of compact Riemann surfaces of genera $g_{X}, g_{Y}$ and degree $d$ , $2 g_{X} - 2 = d (2 g_{Y} - 2) + de g (R)$ where $R$ is the ramification divisor. This relates Riemann-Roch on covers to the base.

Brill-Noether theory. The dimension of the space of "special" divisors (those with $ℓ (D) > de g (D) - g + 1$ ) on a compact Riemann surface is governed by the Brill-Noether number $ρ (g, r, d) = g - (r + 1) (g - d + r)$ . On a generic curve, the locus $W_{d}^{r} \subset Pic^{d} (C)$ is empty if $ρ < 0$ and has dimension exactly $ρ$ if $ρ \geq 0$ (Griffiths-Harris).

Clifford's theorem. For a special divisor $D$ on a compact Riemann surface, $ℓ (D) \leq \frac{d e g ( D )}{2} + 1$ , with equality iff $D = 0$ , $D = K_{X}$ , or $X$ is hyperelliptic (admits a 2-to-1 cover of $\hat{C}$ ).

Castelnuovo-Mumford regularity. Sharper effective bounds for Riemann-Roch on projective curves embedded in projective space.

Moduli of curves. The Riemann-Roch theorem applied to the cotangent bundle of moduli space $M_{g}$ underlies the dimension count $dim M_{g} = 3 g - 3$ (for $g \geq 2$ ) and the Mumford formula for tautological classes.

Synthesis. This construction generalises the pattern fixed in 06.01.01 (holomorphic function), with the symmetric data replaced by its skew or twisted analogue. Read in the opposite direction, the construction is dual to the metric story: complements and orthogonality are taken with respect to the bilinear datum of this unit, not a metric, and the resulting category of subobjects is the one the rest of the strand classifies. The central insight is that this datum identifies algebra with geometry: functions become vector fields, subspaces become quotients, and invariants become cohomology classes — and that identification is the engine driving every theorem downstream.

Full proof set [Master]

The full proof of Riemann-Roch on compact Riemann surfaces uses the Hodge decomposition $H^{1} (X; C) = H^{1, 0} \oplus H^{0, 1}$ for a compact Kähler manifold, combined with the Serre duality in the form $H^{1} (X; O (D)) ≅ H^{0} (X; ω_{X} (- D))^{*}$ . The classical Riemann-Roch follows immediately. Detailed proofs are in Forster Ch. 16 and Griffiths-Harris Ch. 2.

Connections [Master]

Riemann surface 06.03.01 — the analytic geometric setting.
Holomorphic function 06.01.01 — the building blocks; meromorphic functions are quotients of holomorphic ones.
Riemann-Roch theorem for curves 04.04.01 — algebraic version, equivalent by GAGA.
Sheaf cohomology 04.03.01 — both versions live cohomologically as $χ (O (D))$ .
Sheaf 04.01.01 — line bundles, divisors, and structure sheaf are the protagonists.
De Rham cohomology 03.04.06 — Hodge decomposition refines this on Riemann surfaces.
Hirzebruch-Riemann-Roch — generalisation to higher-dimensional compact complex manifolds.
Atiyah-Singer index theorem 03.09.10 — analytic generalisation to elliptic operators.

Historical & philosophical context [Master]

Bernhard Riemann proved the inequality $ℓ (D) \geq de g (D) - g + 1$ in his 1857 paper Theorie der Abelschen Functionen, in the context of compact Riemann surfaces. He used the Dirichlet principle — a method for solving Laplace's equation by minimising the energy integral — which Weierstrass later showed was not rigorously valid in the form Riemann used. This created a foundational crisis that David Hilbert resolved in 1900 by reintroducing the Dirichlet principle on rigorous footing.

Roch's 1865 paper completed the equality form just before he died of tuberculosis at age 26. The full proof was made fully rigorous over the next 40 years through the work of Schwarz, Klein, Hilbert, and others, with Hermann Weyl's Idee der Riemannschen Fläche (1913) consolidating the modern viewpoint.

The 20th-century reformulations were transformative:

Hodge theory (1930s–40s) gave the analytic statement its modern proof via $L^{2}$ -harmonic forms.
Sheaf cohomology (1950s, Cartan-Serre-Grothendieck) reformulated Riemann-Roch as $χ (O (D)) = de g (D) + 1 - g$ via derived functors.
Hirzebruch-Riemann-Roch (1954) generalised to compact complex manifolds via Chern classes.
Grothendieck-Riemann-Roch (1957) made the relative version precise.
Atiyah-Singer index theorem (1963) gave the analytic generalisation: Riemann-Roch is the index of the $\overset{ˉ}{\partial}$ operator.

Today Riemann-Roch is recognised as the prototypical example of a cohomological dimension formula — a topological invariant computed via algebraic means. This pattern recurs throughout modern mathematics in countless variations.

Bibliography [Master]

Forster, Lectures on Riemann Surfaces — §16 is the canonical analytic proof.
Miranda, Algebraic Curves and Riemann Surfaces — Ch. V, both analytic and algebraic perspectives.
Griffiths & Harris, Principles of Algebraic Geometry — Ch. 2, Hodge-theoretic Riemann-Roch.
Farkas & Kra, Riemann Surfaces — Ch. III, classical analytic treatment.
Hirzebruch, Topological Methods in Algebraic Geometry — original HRR generalisation.
Atiyah & Singer, "The index of elliptic operators I-V" — analytic generalisation as index theorem.
Mumford, Curves and their Jacobians — geometric perspective with Brill-Noether.
Riemann, "Theorie der Abelschen Functionen" (1857) — the original.