Pullback of the ideal sheaf of an effective Cartier divisor #
Let D be an effective Cartier divisor on a scheme Y, with ideal sheaf I, and let
g : X ⟶ Y be a morphism such that the inverse image g⁻¹ D, with ideal sheaf I.comap g, is
again an effective Cartier divisor. This holds for flat g, and for the base change of a relative
effective Cartier divisor along any morphism of bases. Then the pullback g^* I of the line
bundle I is the ideal sheaf of g⁻¹ D: the comparison morphism g^* I ⟶ I.comap g is an
isomorphism (Scheme.IdealSheafData.isIso_pullbackSheafHom). Consequently, pulling back the
line-bundle class of I gives the line-bundle class of I.comap g
(Scheme.IdealSheafData.IsEffectiveCartier.pullback_mk_toInvertibleSheaf); dually, 𝒪(D) pulls
back to 𝒪(g⁻¹ D).
Near a point of X, the ideal I is generated on an affine open U of Y by a nonzerodivisor
a, and I.comap g is then generated by g^♯ a on small affine opens V
(Scheme.IdealSheafData.ideal_comap_of_le). Since I.comap g is effective Cartier, g^♯ a is a
nonzerodivisor there, so it is a basis of I.comap g over V. The
comparison morphism sends the image of the section a of I in g^* I to g^♯ a, and a morphism
out of a line bundle sending a local section onto a local basis is an isomorphism
(AlgebraicGeometry.Scheme.Modules.isIso_of_forall_exists_bijective_smul).
References #
- The Stacks Project, Divisors, Section Effective Cartier divisors (Tag 01WQ): the pullback of an effective Cartier divisor, and its invertible ideal sheaf.
The ideal sheaf of an effective Cartier divisor commutes with pullback. If I is an
effective Cartier divisor on Y whose inverse image along g : X ⟶ Y is again an effective
Cartier divisor, then pullbackSheafHom : g^* I ⟶ I.comap g is an isomorphism.
The class of the ideal sheaf of an effective Cartier divisor commutes with pullback. If I
and its inverse image along g are effective Cartier divisors, then pulling back the line-bundle
class of I along g gives the line-bundle class of I.comap g.