Moriwaki divisors and the augmented base loci of divisors on the moduli space of curves (Q333259)

From MaRDI portal





scientific article; zbMATH DE number 6645292
Language Label Description Also known as
default for all languages
No label defined
    English
    Moriwaki divisors and the augmented base loci of divisors on the moduli space of curves
    scientific article; zbMATH DE number 6645292

      Statements

      Moriwaki divisors and the augmented base loci of divisors on the moduli space of curves (English)
      0 references
      0 references
      0 references
      0 references
      28 October 2016
      0 references
      Moriwaki
      0 references
      moduli spaces of curves
      0 references
      divisors
      0 references
      augmented base loci
      0 references
      nef
      0 references
      ample
      0 references
      The paper under review has a main theorem from which the authors draw 4 interesting corollaries.NEWLINENEWLINEMain theorem. An \(\mathbb R\)-Cartier divisor \(D\) on the moduli space of curves \(\overline{M}_g\) is strictly Moriwaki iff the augmented base locus of \(D\) lies in the boundary \(\partial \overline{M}_g\) of \(\overline{M}_g\).NEWLINENEWLINEA divisor \(D\) on \(\overline{M}_g\) is a \textit{Moriwaki divisor} if \(D\) is numerically effective away from the boundary \(\partial \overline{M}_g\). The set of Moriwaki divisors form a closed polyhedral cone and the authors of the present paper define a divisor \(D\) to be \textit{strictly} Moriwaki if \(D\) belongs to the interior of this cone (the authors abbreviate Moriwaki divisor to \(M\)-divisor).NEWLINENEWLINEThe augmented base locus of a divisor is a slight enlargement of the base locus of a divisor. Non-trivially, the augmented base locus \(\text{B}_+(D)\) of a divisor \(D\) may be characterised as the maximal closed locus for which every closed sublocus \(Z\) has the property that \(D|_Z\) is not big [\textit{M. Nakamaye}, Trans. Am. Math. Soc. 355, No. 2, 551--566 (2003; Zbl 1017.14017)].NEWLINENEWLINEThe authors remark that their main theorem gives another proof that every non-zero nef divisor on \(\overline{M}_g\) is big (see Theorem 0.9 in [\textit{A. Gibney} et al., J. Am. Math. Soc. 15, No. 2, 273--294 (2002; Zbl 0993.14009)] for the original version). There are 4 additional corollaries of the main theorem which we roughly state below, sacrificing accuracy for the sake of brevity.NEWLINENEWLINECorollary 1. Let \(D\) be a divisor with base locus contained in the boundary of \(\overline{M}_g\). Then sufficiently high powers of \(D\) induce rational maps restricting to isomorphisms on \(M_g\) iff \(D\) is strictly Moriwaki.NEWLINENEWLINECorollary 2. The log canonical maps \(f_\alpha : \overline{M}_g \to \overline{M}_g(\alpha)\) of Hassett and Hyeon restrict to an isomorphism on \(M_g\) iff \(\alpha > \frac{3g+8}{8g+4}\).NEWLINENEWLINECorollary 3. A divisor with Ithaka dimension \(\kappa(D) \geq 1\) has \(CKM\)-Zariski decomposition iff \(D\) is Moriwaki. In particular, in the range \(g \geq 22\) the canonical divisor does not admit a \(CKM\)-Zariski decomposition.NEWLINENEWLINECorollary 4. A rational map from \(\overline{M}_g\) onto a minimal model can not be extended to a regular map if \(\kappa(\overline{M}_g) \geq 1\), e.g., when \(g \geq 22\).
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references