Kähler currents and null loci (Q896258): Difference between revisions

The article under review investigates the singular loci of cohomology classes sitting at the boundary of the Kähler cone \(\mathcal K_X\) of a compact Kähler manifold \(X\), i.e., the nef cone. There are at least two natural notions of singular locus associated with a nef class \(\alpha\in H^{1,1}(X,\mathbb R)\) such that \(\alpha^{\dim(X)}>0\) (a big class). First, according to \textit{J.-P. Demailly} [J. Algebr. Geom. 1, No. 3, 361--409 (1992; Zbl 0777.32016)], such a class can be represented by a Kähler current with analytic singularities and it is thus Kähler away from a proper subvariety of \(X\). The smallest subvariety as before is called the non-Kähler locus and is denoted by \(E_{nK}(\alpha)\). Alternatively, we can look at the null locus \(\mathrm{Null}(\alpha)\) of \(\alpha\) which is defined as the union of subvarieties \(V\subset X\) such that \(\alpha^{\dim(V)}\cdot [V]=0\). The main theorem of this article states that \(E_{nK}(\alpha)=\mathrm{Null}(\alpha)\) for a nef and big class \(\alpha\). Let us note here that the algebraic case (when \(X\) is smooth projective and the class \(\alpha\) given by \(\alpha=[D]\) where \(D\) is a nef \(\mathbb R\)-divisor) has been considered in [\textit{M. Nakamaye}, Math. Ann. 318, No. 4, 837--847 (2000; Zbl 1063.14008)] and [\textit{L. Ein} et al., Am. J. Math. 131, No. 3, 607--651 (2009; Zbl 1179.14006)]. In this case, the null locus (and the non Kähler locus) equals to the augmented base locus \(\mathbb B_+(D)\). The authors are then able to derive the celebrated result of [\textit{J.-P. Demailly} and \textit{M. Paun}, Ann. Math. (2) 159, No. 3, 1247--1274 (2004; Zbl 1064.32019)]: the Kähler cone is a connected component of the positive cone \[ \mathcal P_X:=\left\{\alpha\in H^{1,1}(X,\mathbb R)\mid \alpha^{\dim(V)}\cdot [V]>0,\,\forall\,V\subset X\right\}. \] They also establish the following striking result concerning singularities of the Ricci flow \[ \frac{\partial}{\partial t}\omega=-\mathrm{Ric}(\omega)\quad\mathrm{and}\quad \omega(0)=\omega_0. \] It is known [\textit{G. Tian} and \textit{Z. Zhang}, Chin. Ann. Math., Ser. B 27, No. 2, 179--192 (2006; Zbl 1102.53047)] that it has a solution for \(t<T:=\sup\{s>0\mid[\omega_0]-sc_1(X)\in \mathcal K_X\}\). If \(T<\infty\), the limiting class \(\alpha:=[\omega_0]-Tc_1(X)\) is nef but not Kähler. Let us consider now the points where the flow develops singularities: they are the points \(x\in X\) where the curvature of \(\omega_t\) is not bounded near \(x\) when \(t\to T\). It is proven there that the Ricci flow develops singularities exactly along the null locus of \(\alpha\). The end of the article also contains results about degenerations of Ricci-flat metrics on Calabi-Yau manifolds. The techniques used in this paper belong to pluripotential theory: the main technical tool is an extension result for quasi-plurisubharmonic functions defined on subvarieties of compact Kähler manifolds (see also [\textit{T. C. Collins} and \textit{V. Tosatti}, Ann. Fac. Sci. Toulouse, Math. (6) 23, No. 4, 893--905 (2014; Zbl 1333.32023)]). Its proof relies on resolution of singularities and gluing techniques à la Richberg. In addition to its very interesting content, it is worth noting that this article is very clear and well written.