On moduli spaces of roots in algebraic and tropical geometry (Q6052934)

Consider the moduli space \(\mathcal{M}_{g, n}\) of smooth genus \(g\) algebraic curves with marked points \(p_1, \ldots, p_n\). Moreover, consider the closely related moduli spaces \(\mathcal{M}_{g,n}^r(\mathcal{R})\) which parametrize smooth curves together with an \(r\)-th root of some universal line bundle \(\mathcal{R} = \omega^\ell(a_1p_1 + \cdots + a_n p_n)\). In the present paper, the authors develop a tropical version \(\mathrm{Root}_g^{r, \mathrm{trop}}(\mathcal{R})\) of the moduli space of roots. The starting point for this is the description of tropical curves as metric graphs and the authors work with \textit{flows} on these graphs as a model for roots of divisors. The resulting \textit{poset of flows} is used as the combinatorics of the tropical space of roots, and the associated generalized cone complex is the desired space \(\mathrm{Root}_g^{r, \mathrm{trop}}(\mathcal{R})\). Following the philosophy of [\textit{D. Abramovich} et al., Ann. Sci. Éc. Norm. Supér. (4) 48, No. 4, 765--809 (2015; Zbl 1410.14049)], the authors show that \(\mathrm{Root}_g^{r, \mathrm{trop}}(\mathcal{R})\) is compatible with specialization, giving a surjective (but not injective) map from the skeleton of the Jarvis compactification of the algebraic space \(\mathcal{M}_{g,n}^r(\mathcal{R})\) to the tropical space. Moreover, this is compatible with the natural forgetful maps to the Deligne-Mumford compactification \(\overline{\mathcal{M}}_{g,n}\). It is interesting to point out that the algebraic moduli space \(\mathcal{M}_{g,n}^{r}(\mathcal{R})\) can be understood as either parametrizing roots of line bundles or covers of curves with prescribed ramification behavior. These two points of view give rise to different tropical theories, the latter of which leads to \textit{tropical admissible covers}. In Section~7 of the present paper the authors relate their moduli space to the space of tropical admissible covers in the case \(r = 2\). As of now, the general case does not seem to be understood.