Remarks on Kawamata's effective non-vanishing conjecture for manifolds with trivial first Chern classes

Abstract

Kawamata proposed a conjecture predicting that every nef and big line bundle on a smooth projective variety with trivial first Chern class has nontrivial global sections. We verify this conjecture for several cases, including (i) all hyperk\"{a}hler varieties of dimension $\leq 6$; (ii) all known hyperk\"{a}hler varieties except for O'Grady's 10-dimensional example; (iii) general complete intersection Calabi-Yau varieties in certain Fano manifolds (e.g. toric ones). Moreover, we investigate the effectivity of Todd classes of hyperk\"{a}hler varieties and Calabi-Yau varieties. We prove that the fourth Todd classes are "fakely effective" for all hyperk\"{a}hler varieties and general complete intersection Calabi-Yau varieties in products of projective spaces.