Effective non-vanishing of global sections of multiple adjoint bundles for polarized 4-folds (Q2376591): Difference between revisions

Let \(X\) be a complex projective manifold, or more generally a complex normal projective variety with at most canonical Gorenstein singularities, and let \(L\) be a nef and big line bundle on \(X\). According to a conjecture of Ionescu one expects that the space of global sections \(H^0(X, K_X+L)\) is not zero if \(K_X+L\) is nef. This conjecture was proven by the reviewer for threefolds [J. Algebr. Geom. 21, No. 4, 721--751 (2012; Zbl 1253.14007)], but the higher-dimensional case is wide open. Since we know by the non-vanishing theorem that some positive multiple \(m(K_X+L)\) has a non-zero global section, an interesting and more accessible problem is to find an effective bound for \(m\). In the paper under review the author considers the case where \(X\) is a projective fourfold with at most terminal Gorenstein singularities and \(L\) is ample. He shows that \(h^0(X, 3(K_X+L))\) is not zero and gives stronger results if \(\kappa(X, K_X+L)<4\) or \(\kappa(X) \geq 0\). As a consequence of the main theorem he obtains an effective non-vanishing statement for a smooth projective fourfold \(X\) and \(L\) an ample line bundle such that \(K_X+L\) is pseudoeffective. The proof is based on a Riemann-Roch computation, the most difficult part is to give a good lower bound for the second Chern class \(c_2(X)\). This lower bound is obtained via a semipositivity result for twisted cotangent sheaves established in [loc. cit.].