On the evolution of a Hermitian metric by its Chern-Ricci form (Q2255315)

The authors investigate a parabolic flow of Hermitian metrics on compact complex manifolds evolving by the Chern-Ricci form. Let \((M,J)\) be a compact complex manifold endowed with a Hermitian metric \(g_0\) with associated \((1,1)\)-form \(\omega_0\). Denote the Chern-Ricci form of a Hermitian metric \(g\) with associated \((1,1)\)-form \(\omega\) by \(\mathrm{Ric}(\omega)\). In local coordinates, \(\mathrm{Ric}(\omega)=-\sqrt{-1}\partial\overline\partial\log\det g\). The authors study the \textit{Chern-Ricci flow} \[ \frac{\partial}{\partial t}\omega = - \mathrm{Ric}(\omega) \qquad \text{ with } \qquad \left.\omega\right|_{t=0} = \omega_0 \;. \] When \(g_0\) is a Kähler metric, then the Chern-Ricci flow coincides with the Kähler-Ricci flow. (Note that, in general, when the initial metric is non-Kähler, the Ricci flow does not preserve the Hermitian condition. Alternative parabolic flows on complex manifolds preserving the Hermitian conditions have been proposed by J. Streets and G. Tian, and by K. Liu and X. Yang.) The above flow has been introduced by \textit{M. Gill} [Commun. Anal. Geom. 19, No. 2, 277--304 (2011; Zbl 1251.32035)]. In particular, Gill proved a non-Kähler analogue of the Cao theorem, showing that, if \(c_1^{BC}(M)=0\), then, for any initial metric \(\omega_0\), there exists a solution \(\omega(t)\) to the Chern-Ricci flow for all time and the metrics \(\omega(t)\) converge smoothly as \(t\to+\infty\) to a Hermitian metric \(\omega_\infty\) with \(\mathrm{Ric}(\omega_\infty)=0\). This metric \(\omega_\infty\) is the unique Chern-Ricci flat metric on \(M\) of the form \(\omega_\infty=\omega_0+\sqrt{-1}\partial\overline\partial\varphi\), according to the results by Cherrier and by the authors. In the present paper, the authors provide a characterization of the maximal existence time for a solution to the Chern-Ricci flow in terms of the initial data, and they investigate the behaviour of the flow on compact complex surfaces when the initial metric is Gauduchon, on compact complex manifolds with negative first Chern class, and on some Hopf manifolds. More precisely, Theorem 1.2 extends the result of G. Tian and Z. Zhang, and shows that there exists a unique maximal solution to the Chern-Ricci flow on the time interval \([0,T)\), where \[ T = \sup \left\{ t \geq 0 | \exists \psi \in \mathcal{C}^\infty(M) \text{ with } \omega_0 - t \mathrm{Ric}(\omega_0) + \sqrt{-1} \partial\overline\partial \psi > 0 \right\} . \] In particular, in the case when \(M\) is a compact complex surface endowed with a Gauduchon metric \(\omega_0\), (namely, \(\partial\overline\partial\omega_0=0\),) then \(\omega(t)\) is Gauduchon for any \(t\in[0,T)\), and the maximal time \(T\) is characterized in Theorem 1.3. In particular, the Chern-Ricci flow exists until either the volume of \(M\) goes to zero, or the volume of a curve of negative self-intersection goes to zero, see Corollary 1.4. This generalizes the known results for the Kähler-Ricci flow following by the Nakai-Moishezon criterion. In Section 6, motivated by the results by J. Song and B. Weinkove on the Kähler-Ricci flow on algebraic and Kähler surfaces, the authors show results and pose conjectures on the behaviour of the Chern-Ricci flow in view to the minimal model program for complex surfaces; see Theorem 1.5. Extending the result of H.-D. Cao, in Theorem 1.7, the authors prove that, on a manifold with negative first Chern class, the (normalized) Chern-Ricci flow starting at any Hermitian metric converges to the unique Kähler-Einstein metric. As an explicit example, the Chern-Ricci flow on some Hopf manifolds is studied in Section 8, see Propositions 1.8 and 1.9. Finally, in Theorem 9.1, the authors give an alternative proof of the uniform estimates for solutions of the complex Monge-Ampère equation \[ \left( \omega+\sqrt{-1}\partial\overline\partial\varphi\right)^n = e^F \omega^n \qquad \text{ with } \qquad \omega+\sqrt{-1}\partial\overline\partial\varphi>0 \] on a compact Hermitian manifold \((M,\omega)\), where \(F\) is a smooth function on \(M\). A (strictly stronger) estimate of this kind was already proven by the authors in [J. Am. Math. Soc. 23, No. 4, 1187--1195 (2010; Zbl 1208.53075)]. Here, they use a new second order estimate which is now proven for any compact Hermitian manifold, thanks to a trick by D. H. Phong and J. Sturm.