On positivity and base loci of vector bundles (Q2354825)

In the present note the authors provide a comprehensive survey on different positivity notions for vector bundles on projective varieties. The positivity notions for line bundles (like ampleness, bigness or nefness) are actually well-studied. However, in the case of vector bundles the situation is far from an analogous state of art. In this review we focus mainly on (some) results devoted to vector bundles due to the compactness requirement for reviews. Let \(X\) be a smooth projective variety of an arbitrary dimension over the complex numbers and denote by \(E\) a vector bundle of rank \(r\). For a point \(x \in X\) denote by \(E(x) = E \otimes \kappa(x)\), where \(\kappa(x)\) is the residue field at \(x\), the fiber of the total space of \(E\) over a point \(x\) (this is a vector space of dimension \(r\)). For a line bundle \(L\) denote by \(c_{1}(L)\) the associated Weil divisor on \(X\). For a vector bundle \(E\) one defines the base locus as the subset \[ \mathrm{Bs}(E) = \{x \in X : H^{0}(X,E) \rightarrow E(x) \text{ is not surjective }\} \] and the stable base locus of \(E\) as \[ \mathbb{B}(E) = \bigcap_{m > 0} \mathrm{Bs}(\mathrm{Sym}^{m}E). \] For a positive rational number \(r = p/q\) and an ample line bundle \(A\) define \[ \mathbb{B}(E+rA) = \mathbb{B}(\mathrm{Sym}^{q}E \otimes A^{p}), \quad \quad \mathbb{B}(E-rA) = \mathbb{B}(\mathrm{Sym}^{q}E\otimes A^{-p}). \] It can be show that these two objects are well-defined (do not depend on \(r\)), and then we can define the augmented base locus of \(E\) as \[ \mathbb{B}_{+}^{A}(E) = \bigcap_{r \in \mathbb{Q}_{>0}} \mathbb{B}(E-rA) \] and the restricted base locus of \(E\) as \[ \mathbb{B}^{A}_{-}(E) = \bigcup_{r \in \mathbb{Q}_{>0}} \mathbb{B}(E+rA). \] Again, it can be shown that restricted and augmented base loci do not depend on the choice of the ample line bundle \(A\), so one can write \(\mathbb{B}_{-}\) and \(\mathbb{B}_{+}\), respectively. For a vector bundle \(E\) let \(\pi: \mathbb{P}(E) \rightarrow X\) be a projective bundle of rank one quotients of \(E\) and \(\mathcal{O}_{\mathbb{P}(E)}(1)\) the universal quotient of \(\pi^{*}E\) on \(\mathbb{P}(E)\). The first result of the paper shows that for a vector bundle \(E\) on \(X\) one has \[ \pi(\mathbb{B}_{-}(\mathcal{O}_{\mathbb{P}(E)}(1))) = \mathbb{B}_{-}(E) \] and \[ \pi(\mathbb{B}_{+}(\mathcal{O}_{\mathbb{P}(E)}(1))) = \mathbb{B}_{+}(E). \] In the next sections the authors study ampleness, nefness and other positivity notions for vector bundles. Recall the following definitions. We say that a vector bundle \(E\) on \(X\) is {\parindent=8mm \begin{itemize}\item[(i)] nef if \(\mathcal{O}_{\mathbb{P}(E)}(1)\) is a nef line bundle, \item[(ii)] pseudo-effective if \(\mathbb{B}_{-}(E) \neq X\), \item[(iii)] weakly positive if \(\overline{\mathbb{B}_{-}(E)} \neq X\), \item[(iv)] ample if \(\mathcal{O}_{\mathbb{P}(E)}(1)\) is ample on \(\mathbb{P}(E)\), \item[(v)] \(L\)-big if \(\mathcal{O}_{\mathbb{P}(E)}(1)\) is big on \(\mathbb{P}(E)\), \item[(vi)] Viehweg-big (or \(V\)-big) if there exists an ample line bundle \(A\) and a positive integer \(c > 0\) such that \(\mathrm{Sym}^{c}E \otimes A^{-1}\) is weakly positive. \end{itemize}} Then one can show that {\parindent=8mm \begin{itemize}\item[(i)] \(E\) is nef iff \(\mathbb{B}_{-}(E) = \emptyset,\) \item[(ii)] \(E\) is ample iff \(\mathbb{B}_{+}(E) = \emptyset,\) \item[(iii)] \(E\) is Viehweg-big iff \(\mathbb{B}_{+}(E) \neq X,\) \item[(iv)] if \(E\) is Viehweg-big, then it is \(L\)-big (the converse statement is false in general, see Remark 6.6). \end{itemize}} Another interesting positivity notion studied in the paper is the almost everywhere ampleness (firstly defined by \textit{Y. Miyaoka} in [Prog. Math. 39, 281--301 (1983; Zbl 0584.14022)] for surfaces). We say that a vector bundle \(E\) on \(X\) is almost everywhere ample (AEA for short) if there exists an ample line bundle \(A\) on \(X\), a Zariski closed subset \(T \subset \mathbb{P}(E)\) whose projection \(\pi(T)\) onto \(X\) satisfies \(\pi(T) \neq X\) and a positive number \(\varepsilon > 0\) such that \[ c_{1}(\mathcal{O}_{\mathbb{P}(E)}(1)).C \geq \varepsilon \pi^{*}(c_{1}(A)).C \] for all curves not contained in \(T\). Then it can be shown that a line bundle \(L\) on \(X\) is AEA iff \(L\) is big and if a vector bundle \(E\) on \(X\) is AEA, then it is \(L\)-big. In the last section the authors prove strong and weaker Kodaira lemma for vector bundles and moreover they show that for a nef vector bundle \(E\) on an irreducible projective variety \(X\) the bundle \(E\) is AEA iff \(E\) is Viehweg-big. Also the authors formulate many interesting questions about the previously mentioned positivity notions, for instance it is not know whether for a vector bundle \(E\) being AEA implies that \(E\) is Viehweg-big (as we saw it is true under the additional nefness assumption).