Minimal Siegel modular threefolds (Q4393897)

Let \(\Gamma_t\) be a subgroup of \(\text{Sp}_4(\mathbb{Q})\) such that \(\mathcal{A}_t := \Gamma_t \backslash \mathbb{H}_2\) is the moduli space of \((1, t)\)-polarized abelian surfaces. Let \(\Gamma_t^\ast\) be the maximal extension of \(\Gamma_t\). The authors study the spaces \(\mathcal{A}_t^\ast: = \Gamma_t^\ast \backslash \mathbb{H}_2\) -- the minimal Siegel modular threefolds of the title. The authors find that \(\mathcal{A}_t^\ast\) can be interpreted as a moduli space for Kummer surfaces and some intermediate spaces lying between \(\mathcal{A}_t\) and \(\mathcal{A}_t^\ast\) parameterize certain \(K3\) surfaces. In some cases information on the Kodaira dimension is obtained. The authors also prove that the ramification locus of the finite map \(\mathcal{A}_t \rightarrow \mathcal{A}_t^\ast\) is a union of Humbert surfaces (images of Hilbert modular surfaces). The paper has three sections. In section one the groups \(\Gamma_t\) and \(\Gamma_t^\ast\) are studied and interpreted in terms of moduli. (In particular, the group \(\Gamma_t^\ast\) is generated over \(\Gamma_t\) by involutions resembling the construction of the Atkin-Lehner involutions). The main tool here is relating \(\text{Sp}_4\) to \(\text{SO}(3, 2)\). Section two deals with non-unirationality of some spaces between \(\mathcal{A}_t^\ast\) and \(\mathcal{A}_t\). The main tool is the construction of Jacobi forms and using it to construct a non-trivial \(3\)-form, therefore showing that the Kodaira dimension is non-negative. The third section studies the ramification locus of \(\mathcal{A}_t \rightarrow \mathcal{A}_t^\ast\) as a first step in understanding the singularities of \(\mathcal{A}_t^\ast\), given that the singularities of \(\mathcal{A}_t\) are known. The result is expressed as an explicit union of Humbert surfaces.