Uniruledness of stable base loci of adjoint linear systems via Mori theory (Q379227)

The main goal of the paper under review is to explain how recent results in the Minimal Model Program [\textit{C. Birkar} et al., J. Am. Math. Soc. 23, No. 2, 405--468 (2010; Zbl 1210.14019)] imply a uniruledness result for the base loci of adjoint divisors. More precisely, if \(X\) is a normal projective variety over \(\mathbb C\), \(A\) is an ample divisor on \(X\) and \(D\) is a \(\mathbb R\)-Cartier \(\mathbb R\)-divisor on \(X\), then \({\mathbf B} (D):=\bigcap \{ {\text{Supp}}(E)\;|\;E\geq 0,\;E\sim_{\mathbb R }D\}\) is the stable base locus of \(D\), \( {\mathbf B}_+(D)=\cap_{ m>0} {\mathbf B} (D-\frac 1 m A)\) is the augmented base locus of a big divisor \(D\) and \({\mathbf B}_-(D)=\bigcup _{m>0} {\mathbf B}(D+\frac 1 m A)\) is the restricted base locus of \(D\). The non-nef locus \(\text{NNef}(D)\) is the union of all centers \(c_X(v)\) where \(v\) is a divisorial valuation over \(X\) such that \(v(||D||)=0\) where \(v(||D||):=\text{inf}\{ v(E)\;|\;D\sim _{\mathbb R}E\geq 0 \}\). The main result of the paper under review states that if \(X\) is normal \(\mathbb Q\)-factorial projective over \(\mathbb C\) and \((X,\Delta )\) is a klt pair, then (i) \(\text{NNef}(K_X+\Delta )=\mathbf B_-(K_X+\Delta )\) and each of its irreducible components is uniruled. (ii) If \(K_X+\Delta \) is big, then \(\mathbf B_-(K_X+\Delta )=\mathbf B(K_X+\Delta )\) and each component of \(\mathbf B_+(K_X+\Delta )\) is also uniruled.