Documentation

TauCeti.AlgebraicGeometry.PicardFunctor.Abel.Degree

The Abel–Jacobi maps of degree d #

Let f : X ⟶ S be a smooth proper morphism of relative dimension one with a section x₀. For every natural number d, the functor Div^d_{X/S} of relative effective Cartier divisors of degree d maps to the relative Picard presheaf T ↦ Pic(X_T) / Pic(T) by the Abel map D ↦ 𝒪(D). Twisting by the d-th power of the class of the base-changed section x₀_T normalizes it to the Abel–Jacobi map of degree d

abelJacobiMapOfDegree f hx₀ d : Div^d_{X/S} ⟶ (T ↦ Pic(X_T) / Pic(T)), D ↦ 𝒪(D - d • x₀_T),

which sends the divisor d • x₀_T to the identity. Once Div^d_{X/S} is represented by the symmetric power Symᵈ X (here this is known only for d ≤ 1), these become the maps D ↦ 𝒪(D - d • x₀) out of Symᵈ X used in the classical construction of the Jacobian from symmetric powers of a curve.

Addition of divisors is written multiplicatively in the monoid of relative effective Cartier divisors (the sum D + E is cut out by the product of the ideal sheaves), so below D + E and d • x₀_T are D * E and x₀_T ^ d in Lean.

This file proves the compatibilities of these maps with the operations on divisors:

Representability of Div^d_{X/S} for d ≥ 2, and the identification of the image with the degree-zero part of the Picard functor, are not treated here.

References #

The Abel–Jacobi map of degree d of a smooth proper morphism f : X ⟶ S of relative dimension one with a section x₀: the natural transformation from the functor Div^d_{X/S} of relative effective Cartier divisors of degree d on the base changes X_T = T ×_S X to the relative Picard presheaf T ↦ Pic(X_T) / Pic(T), sending a divisor D to the class of 𝒪(D - d • x₀_T) = 𝒪(D) ⊗ 𝒪(x₀_T)^{⊗ -d}, where x₀_T is the base change of x₀ (abelJacobiMapOfDegree_app_apply).

Instances For

    The Abel–Jacobi map of degree d sends a divisor D to the quotient of the Abel map at D by the d-th power of the Abel map at the graph of the base point, that is, to the class of 𝒪(D - d • x₀_T).

    The Abel–Jacobi maps are additive: the sum of divisors D and E of degrees d and e goes to the product of their classes, 𝒪(D + E - (d + e) • x₀_T) = 𝒪(D - d • x₀_T) ⊗ 𝒪(E - e • x₀_T).

    @[simp]

    In degree zero the Abel–Jacobi map is trivial: the only divisor of degree zero is the empty divisor, and 𝒪(∅) = 𝒪.

    The Abel–Jacobi maps are compatible with stabilization: adding e copies of the base point to a divisor D of degree d does not change its class, 𝒪(D + e • x₀_T - (d + e) • x₀_T) = 𝒪(D - d • x₀_T).

    In degree one the Abel–Jacobi map is the Abel–Jacobi map on points: composing the isomorphism Hom_S(-, X) ≅ Div¹_{X/S}, x ↦ Γₓ, with the Abel–Jacobi map of degree one gives the Abel–Jacobi map x ↦ 𝒪(Γₓ - x₀_T).

    The Abel–Jacobi map of a sum of points is the product of their Abel–Jacobi classes: for a finite family of T-points x_i, i ∈ s, the divisor ∑ i ∈ s, Γ_{x_i} goes to ∏ i ∈ s, 𝒪(Γ_{x_i} - x₀_T).