Graded rings of paramodular forms of levels 5 and 7 (Q2288326)

The paper computes the ring of modular forms for the paramodular groups of degrees 5 and 7. These are the subgroups of \({\mathrm{Sp}}(4,{\mathbb Q})\) associated with the moduli spaces \(\mathcal{A}_{1,5}\) and \(\mathcal{A}_{1,7}\) of \((1,5)\)- and \((1,7)\)-polarised abelian surfaces. Geometrically, the effect is to give equations for the Satake compactifications of the corresponding moduli spaces. The rings are generated by certain Eisenstein series and Borcherds (multiplicative) and Gritsenko (additive) lifts and ratios of them. The computations of these generators and relations relies on studying the restrictions of modular forms to suitable Humbert surfaces, which are Hilbert modular forms, whose structure is known in detail. Both rings are shown to be Gorenstein and Cohen-Macaulay, but not complete intersections. There are two minor errors. One, pointed out to the reviewer by the author, is in Table 4, where after the term \(90(\zeta^{-1}-\zeta)\) in the entry for \(g_7\) one should insert ``\(\big)q\)'' (that is, everything up to this point is the coefficient of \(q\)). The other, pointed out by the reviewer to the author, is in the last paragraph of Section 5, and affects only that paragraph, not the rest of the paper. It is stated that \(K(X_{K(5)})\), which is the Satake compactification \(\mathcal{A}^*_{1,5}\) of \(\mathcal{A}_{1,5}\), is not rational. The argument that is given for this is certainly wrong, and the question appears to be open. It is known that \(\mathcal{A}^*_{1,5}\) is unirational. See [\textit{K. Hulek} and \textit{G. K. Sankaran}, Adv. Stud. Pure Math. 35, 89--156 (2002; Zbl 1074.14021)] for a summary, apparently still up to date, of such questions.