Cyclic extensions and the local lifting problem (Q2251004): Difference between revisions

It was known that a projective, smooth, connected curve \(\bar{Y}\) defined over an algebraically closed field \(k\) of positive characteristic \(p\) can be lifted to characteristic zero, i.e. there is a discrete valuation ring \(R\) of characteristic zero with residue field \(k\) and a relative \(R\) curve \(Y\) with special fibre \(\bar{Y}\). \textit{F. Oort} [``Some questions in algebraic geometry'' (1995), \url{http://www.math.uu.nl/~oort0109/A-Qnew.ps}] proposed a similar problem, on whether a Galois cover \(\bar{f}: \bar{Y} \rightarrow \bar{X}\) with Galois group \(\Gamma\) can be lifted in the same way to a Galois cover of relative \(R\)-curves. For a general group \(\Gamma\) this is not possible, there are obstructions based on the size of \(\Gamma\) with respect to the genus, the Bertin obstruction [\textit{J. Bertin}, C. R. Acad. Sci., Paris, Sér. I, Math. 326, No. 1, 55--58 (1998; Zbl 0952.14018)] and the Hurzitz tree obstruction [\textit{L. H. Brewis} and \textit{S. Wewers}, Math. Ann. 345, No. 3, 711--730 (2009; Zbl 1222.14045)]. But for cyclic groups all these obstructions vanish. The Oort conjecture claims that a cyclic group cover of order \(p^n\) can always be lifted to characteristic zero. The breakthrough result of the authors completes the result of \textit{F. Pop} [Ann. Math. (2) 180, No. 1, 285--322 (2014; Zbl 1311.12003)] and the two articles together provide a full proof of the Oort conjecture. The proof uses the local nature of the lifting problem expressed in terms of the local-global lifting property and restates the problem in terms of a lifting problem of formal power series in terms of rigid analytic geometry. This problem is restated in the language of characters, i.e. elements in \(H^1(\mathbb{K},\mathbb{Z}/p^n \mathbb{Z})\), where \(\mathbb{K}\) is the function field of the curve \(X\) in the generic fibre. Such a character corresponds to a branched cover \(Y \rightarrow X\) and several invariants are attached to a character, like three types of Swan conductors, which measure how bad is the reduction of a cover. The proof uses an induction process based on a detailed study of \(\mathbb{Z}/p\mathbb{Z}\)-extensions which are the building blocks of the induction.