Lifting harmonic morphisms. I: Metrized complexes and Berkovich skeleta (Q2347892)

Let \(K\) be an algebraically closed complete non-Archimedean valued field with a non-trivial valuation. Starting with an algebraic curve over \(K\), one can associate to it a so-called skeleton: it is a metric graph (naturally living inside the Berkovich analytification of the curve) that already contains a lot of information about the curve. The aim of the paper is to investigate the relative situation, that of a finite morphism of curves. The authors first prove, by analytic arguments, that one can associate a morphism of skeleta to a morphism of curves. This allows them to prove a result of simultaneous semistable reduction for finite morphisms of curves that is stronger than the known results (see [\textit{Q. Liu} and \textit{D. Lorenzini}, Compos. Math. 118, No. 1, 61--102 (1999; Zbl 0962.14020)], [\textit{R. F. Coleman}, Doc. Math., J. DMV Extra Vol., 217--225 (2003; Zbl 1100.14515)], [\textit{Q. Liu}, Compos. Math. 142, No. 1, 101--118 (2006; Zbl 1108.14020)]). On the other hand, morphisms of skeleta do not always come from morphisms of curves. To overcome this problem, the authors refine the skeleta into complexes of curves by adjoining to each vertex a curve over the residue field \(\tilde{K}\). They prove that tame morphisms of complexes of curves may be lifted to actual morphisms of curves. They also prove that, given a curve that lifts the target complex of curves, there are only finitely many possible lifts of the morphism and that they can be classified. Those results extend and make more precise former results by \textit{M. Saïdi} [Compos. Math. 107, No. 3, 319--338 (1997; Zbl 0929.14016)] and \textit{S. Wewers} [Lond. Math. Soc. Lect. Note Ser. 256, 239--282 (1999; Zbl 0995.14008)]. At the end of the paper, the results are used in order to answer a question of Ribet about component groups of Néron models. The authors announce applications to lifting of morphisms of tropical curves in a second paper to come.