Bergman kernels and the pseudoeffectivity of relative canonical bundles (Q955158)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Bergman kernels and the pseudoeffectivity of relative canonical bundles
scientific article

    Statements

    Bergman kernels and the pseudoeffectivity of relative canonical bundles (English)
    0 references
    0 references
    0 references
    18 November 2008
    0 references
    Suppose \(X\) and \(Y\) are projective manifolds and \(p:X\to Y\) a surjective projective map. Suppose \(L\) is a line bundle on \(X\) equipped with a (possibly singular) metric with semipositive curvature current. Let \(K_{X/Y}\) denote the relative canonical bundle on \(X\) (on a nonsingular fibre \(X_y\) it restricts to the canonical bundle of~\(X_y\)). This article provides a useful criterion under which \(K_{X/Y}+L\) is pseudoeffective. Specifically, it is shown that if \(H^0(X_y,(K_{X/Y}+L)\otimes{\mathcal{I}}(h))\not=0\) for some regular value \(y\) of \(p\) where \({\mathcal{I}}(h)\) is the multiplier ideal sheaf of the metric \(h\) on~\(L\), then there is a (possibly singular) metric on \(K_{X/Y}+L\) with semipositive curvature current. Several results of this nature have been proved recently but the main improvement of this one is that the fibres of \(p\) and the metric on \(L\) are both allowed to be singular. These are crucial advantages regarding applications in algebraic geometry. The method is from several complex variables. Specifically, the authors first restrict attention to the smooth fibres and use the Bergman kernel there to produce a metric of the desired type. Then they estimate the Bergman metric near the singular fibres, obtaining their extension to \(X\) by classical pluripotential theory. There are several applications to algebraic geometry. In this article the authors use it to give volume estimates and to simplify the proof of an extension theorem for twisted pluricanonical forms. Further applications are promised in a forthcoming article, for example to prove the Kawamata subadjunction theorem.
    0 references
    0 references
    0 references
    0 references
    0 references
    Bergman kernel
    0 references
    projective manifold
    0 references
    pseudoeffective
    0 references
    pluripotential
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references