Subadditivity of Kodaira dimensions for fibrations of three-folds in positive characteristics (Q2324576)

Let \(X\) be a projective variety over a field \(k\), \(D\) a \({\mathbb{Q}}\)-Cartier divisor on \(X\) and \(\kappa(D, X)\) the \(D\)-dimension defined to be \(\max\{\dim_k\Phi_{|mD|}(X)|m\in\mathbb{Z},m>0\}\) if \(|mD|\neq\emptyset\) for some \(m>0\) and \(-\infty\) if \(|mD|=\emptyset\) for every \(m>0\). The Kodaira dimension \(\kappa(X)=: \kappa(K_X, X)\) where \(K_X\) is the canonical divisor of \(X\). Iitaka Conjecture. Let \(f:X\rightarrow Y\) be a fibration between two smooth projective varieties over \(\mathbb{C}\) with \(\dim X=n\) and \(\dim Y=m\). Then the Kodaira dimensions satisfy \[ C_{n,m} : \kappa(X) \geq \kappa(Y ) + \kappa(X_{\bar{\eta}}), \] where \(\bar{\eta}\) is geometric generic point of \( Y\) and \(X_{\bar{\eta}}\) is the fiber over \(\bar{\eta}\). Progress has been made on Iitaka Conjecture by Kawamata, Kollár, Viehweg, Birka, Chen and Hacon, Cao and Páun etc. In positive characteristics, assume that the geometric generic fibre \(X_{\bar{\eta}}\) is integral and has a smooth projective birational model \(\tilde{X}_{\bar{\eta}}\), the Weak Subadditivity Conjecture is \[ WC_{n,m} : \kappa(X) \geq \kappa(Y ) + \kappa(\tilde{X}_{\bar{\eta}}). \] If the geometric generic fiber is smooth, using the moduli theory and positivity results proved recently by \textit{Z. Patakfalvi} [Ann. Sci. Éc. Norm. Supér. (4) 47, No. 5, 991--1025 (2014; Zbl 1326.14015)] and \textit{S. Ejiri} [J. Algebr. Geom. 26, No. 4, 691--734 (2017; Zbl 1400.14026)], several special cases \(WC_{n,n-1}\), \(WC_{3,1}\) were proved by Birkar, Chen, Ejiri and Zhang. This paper deals with possibly singular geometric generic fibers and the main result is the following. Theorem 1.5. Let \(f:X\rightarrow Y\) be a separable fibration between two normal projective varieties over an algebraically closed field \(k\) with char\(k=p>0\). Assume either that \(Y\) is smooth or that \(f\) is flat. Let \(D\) be a Cartier divisor on \(X\). If there exist an effective \(\mathbb{Q}\)-Weil divisor \(\Delta\) on \(X\) and a big \(Q\)-Cartier divisor \(A\) on \(Y\) such that \begin{itemize} \item[(1)] \(K_X +\Delta\) is \(\mathbb{Q}\)-Cartier and the Cartier index ind\(((K_X + \Delta)_\eta\)) is indivisible by \(p\); \item[(2)] \(D-K_{X/Y}-\Delta- f^* A\) is nef and \(f\)-semi-ample; \item[(3)] \(\dim_{\kappa ({\bar{\eta}}}) S^0_{\Delta_{\bar{\eta}}}(X_{\bar{\eta}}, D_{\bar{\eta}} ) > 0\), \end{itemize} then \[ \kappa (X, D) \geq \dim Y + \kappa (X_{\bar{\eta}}, D_{\bar{\eta}}). \] In particular, if \(D\) is nef and \(f\)-big, and conditions (1) and \begin{itemize} \item[(2')] \(D-K_{X/Y}-\Delta- f^* A\) is nef \end{itemize} hold, then \(D\) is big. An application of the main theorem to three-folds is the following. Corollary. Let \((X, \Delta)\) be a projective \(k\)lt pair of dimension 3, and let \(f:X\rightarrow Y\) be a separable fibration to a smooth projective curve or a surface, over an algebraically closed field \(k\) with char\(k=p>5\). Assume that \(K_Y\) is big and \(Y\) is non-uniruled. Then \[ \kappa (X, K_X + \Delta) \geq \kappa(Y ) + \kappa (X_{\bar{\eta}}, K_{X_{\bar{\eta}}}+\Delta_{\bar{\eta}}) \] if one of the following holds \begin{itemize} \item[(1)] \(K_{X/Y}+\Delta\) is \(f\)-big; \item[(2)] \(\Delta = 0\) and the geometric generic fiber \(X_{\bar{\eta}}\) is smooth. \end{itemize} The idea of the proof of the main theorem is inspired by the techniques developed by Viehweg, Patakfalvi, Schwede and Zhang and the key step is to prove a positivity result (Theorem 1.11) without singularity conditions, but at the cost of assuming other conditions like nefness and relative semi-ampleness.