On a generalized Calabi-Yau equation (Q609655)

The authors answer a conjecture stated by \textit{P. Delanoë} in [``Sur l'analogue presque-complexe de l'équation de Calabi-Yau'', Osaka J. Math. 33, No.~4, 829--846 (1996; Zbl 0878.53030)] concerning a non-existence result for the so-called generalized Calabi-Yau equation. More precisely, consider a closed almost Kähler \(2n\)-manifold \((M,g,J,\omega)\); this means that \((M,\omega)\) is a symplectic manifold and \(J\) is an \(\omega\)-compatible almost complex structure, that is, \(g(\cdot,\cdot) = \omega(\cdot,J\cdot)\) is a Riemannian metric. A volume form on \(M\) in the same cohomology class as \(\omega^n\) is determined by a function \(h:M \to (0,+\infty)\) satisfying \[ \int_M h\omega^n = \int_M \omega^n. \] We denote by \(B^+\) the set of such functions \(h\). Using Moser's trick, one can show that, for any \(h\in B^+\), there exists another symplectic form \(\omega'\) in the same cohomology class as \(\omega\) such that \({\omega'}^n = h\omega^n\). Moser's trick, however, does not single out a particular \(\omega'\), but one can look for \(\omega'\) having the particular form \[ \omega' = \omega(\phi) := \omega + d(d\phi\circ J), \quad \phi \in C^\infty(M,{\mathbb R}), \] further constrained by the requirement that \(J\) is \(\omega'\)-tamed (this means \(\omega'(v,Jv)>0\) whenever \(v\neq 0\)). If the set of \(\phi \in C^\infty(M,{\mathbb R})\) such that \(J\) is \(\omega(\phi)\)-tamed is denoted by \(A^+\), then there is a map \(F:A^+\to B^+\) which associates to each \(\phi \in A^+\) the function \(h = F(\phi) \in B^+\) determined by \(h\omega^n = \omega(\phi)^n\). In the Kähler case, that is, when \(J\) is integrable, the surjectivity of \(F\) was proved by \textit{S.-T. Yau} [``On the Ricci curvature of a compact Kähler manifold and the complex Monge-Ampère equation. I'', Commun. Pure Appl. Math. 31, 339--411 (1978; Zbl 0369.53059)], thus establishing the Calabi conjecture. Delanoë [loc. cit.] showed that \(F\) is always a diffeomorphism onto its image and, in the case \(n=2\), is surjective onto \(B^+\) if, and only if, \(J\) is integrable. He also conjectured that this last assertion should be true for \(n\geq2\). The paper under review answers Delanoë's conjecture positively.