Alternate compactifications of the moduli space of genus one maps (Q453351)

This work generalizes the definition of a Kontsevich stable map in genus one. This is achieved by allowing domain curves to have elliptic \(m\)-fold singularities. Recall that for an integer \(m\geq 1\), an elliptic \(m\)-fold point is a Gorenstein singularity with \(m\) branches and genus one (equivalently, \(m\) branches and \(\delta\)-invariant \(m\)). Elliptic \(m\)-fold points were introduced and classified by \textit{D. I. Smyth} [Compos. Math. 147, No. 3, 877--913 (2011; Zbl 1223.14031), ibid. 147, No. 6, 1843--1884 (2011; Zbl 1260.14033)]: For \(m=2,3\), the elliptic \(m\)-fold point is a cusp and a tacnode, respectively. For \(m\geq 3\), the elliptic \(m\)-fold point is the singularity of the cone over \(m\) general points in \({\mathbb A}^{m-2}\). The author defines a map \(\mu: C \rightarrow X\) from a reduced, connected, proper curve of arithmetic genus one with \(n\) distinct smooth marked points to a projective variety \(X\) to be \(m\)-stable if: (1) \(C\) has only nodes and elliptic \(\ell\)-fold points, \(\ell\leq m\), as singularities. (2) Any connected subcurve of \(C\) of arithmetic genus one on which \(\mu\) is constant has at least \(m+1\) distinguished points (i.e., points of intersection with the rest of the curve or marked points). (3) The map has no non-trivial infinitesimal automorphisms fixing \(X\) and marked points. The first main result of the paper is that the moduli functor of \(m\)-stable maps of a given homology class \(\beta\) is representable by a proper Deligne-Mumford stack \(\overline{\mathcal{M}}_{1,n}^{m}(X,\beta)\) over \({\mathbb Z}[1/6]\). The most delicate part of the argument is verifying the valuative criterion of properness for this stack. In the case of \(m\)-stable maps into a projective space, the author studies the geometry of \(\overline{\mathcal{M}}_{1,n}^{m}({\mathbb P}^r,d)\) in more detail. The main result in this direction is the enumeration of irreducible components of \(\overline{\mathcal{M}}_{1,n}^{m}({\mathbb P}^r,d)\) and the proof of irreducibility of \(\overline{\mathcal{M}}_{1,n}^{m}({\mathbb P}^r,d)\) for \(m\geq \text{min}(r,d)+n\). This provides a modular compactification of the stack \({\mathcal{M}}_{1,n}({\mathbb P}^r,d)\) parameterizing degree \(d\) maps from smooth \(n\)-pointed genus one curves to \({\mathbb P}^r\). The resulting modular compactification is often singular and should be contrasted with a smooth irreducible compactification of \({\mathcal{M}}_{1,n}({\mathbb P}^r,d)\) constructed by \textit{R. Vakil} and \textit{A. Zinger} [Geom. Topol. 12, No. 1, 1--95 (2008; Zbl 1134.14009)], which is constructed as a sequence of explicit blow-ups of \(\overline{\mathcal{M}}_{1,n}({\mathbb P}^r,d)\) and for which no modular interpretation is known, and with a modular smooth irreducible compactification of \({\mathcal{M}}_{1,n}({\mathbb P}^r,d)\) constructed by \textit{B. Kim} [Adv. Stud. Pure Math. 59, 167--200 (2010; Zbl 1216.14023)]. The third main result of the paper is that \(\overline{\mathcal{M}}_{1,n}^{m}({\mathbb P}^r,d)\) is a smooth stack whenever \(d+n \leq m \leq 5\). This is closely related to the fact that an elliptic \(m\)-fold singularity is unobstructed if and only if \(m\leq 5\) [Smyth, loc. cit.]. The author poses several open questions relating to this work, such as the existence of a virtual fundamental class and the projectivity of the coarse moduli space for \(\overline{\mathcal{M}}_{1,n}^{m}(X,\beta)\).