Explicit constructions of \(K3\) surfaces and unirational Noether-Lefschetz divisors (Q2079247)

The paper looks at on Noether-Lefschetz divisors in the moduli space \(\mathcal{F}_g\) of genus \(g\) polarised \(K3\) surfaces. These are moduli spaces \(\mathcal{F}^{\Lambda^{d,n}_g}\) of lattice-polarised \(K3\) surfaces with polarising lattice isometric to \(\Lambda^{d,n}_g\) given by the matrix \(\begin{pmatrix}2g-2 & d \\ d & n\end{pmatrix}\). Such moduli spaces are quite frequently unirational even when \(\mathcal{F}_g\) is not. The paper does two things: from a theoretical point of view, the authors demonstrate this by exhibiting at least three such unirational divisors for any \(g\), and more in some cases. They ask whether the number of unirational divisors grows with \(g\) or might even be infinite. (The latter seems unlikely to the reviewer unless \(g\) is small, because if \(\mathcal{F}_g\) is of general type it would contradict the Bombieri-Lang conjecture.) They give unirationality proofs that are explicit in the sense that they are able to write code in Macaulay2 that produces equations for the general \(K3\) surface parametrised by these divisors. Moreover, for some small \(g\) including the case of \(\mathcal{F}_{11}\), they do the same thing for the general \(K3\) surface. This is noteworthy because the proof that \(\mathcal{F}_{11}\) is unirational, unlike the proofs for the other small values of \(g\) where that is known, is not evidently effective in this sense. Instead, they give a new proof of the result, using Gushel-Mukai fourfolds