The p p -rank stratification of the moduli space of double covers of a fixed elliptic curve

Kevin Chang, Dušan Dragutinović, Steven R. Groen, Yuxin Lin · 2025

Read The p p -rank stratification of the moduli space of double covers of a fixed elliptic curve by Kevin Chang, Dušan Dragutinović, Steven R. Groen, Yuxin Lin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-09540.txt.

Introduction

A research team studies how a key invariant sorts the space of curves that double-cover one fixed elliptic curve, in a 2025 math paper. The setting is algebraic geometry in positive characteristic, where the p-rank measures how a curve's Jacobian interacts with the Frobenius map. Moduli spaces gather all such curves into one geometric object, stratified by the invariant's value.

The authors prove that each closed stratum, the locus where the p-rank stays at or below a given value, has exactly the expected dimension, with all its pieces of equal size. They also show smooth double covers exist realizing every possible p-rank value. Both results sound technical but settle the basic shape of the stratification: no stratum is unexpectedly fat or empty. The paper adds a fully worked example to the broader program of mapping moduli spaces in characteristic p.

Catalog record

CREATORKevin Chang, Dušan Dragutinović, Steven R. Groen, Yuxin Lin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE129 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-09540

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The p p -rank stratification of the moduli space of double covers of a fixed elliptic curve Kevin Chang, Dušan Dragutinović, Steven R. Groen, Yuxin Lin Abstract In this paper we investigate the p p -rank stratification of the moduli space of curves of genus g g that admit a double cover to a fixed elliptic curve E E in characteristic p > 2 p>2 . We show that the closed p p -rank strata of this moduli space are equidimensional of the expected dimension. We also show the existence of a smooth double cover of E E of all the possible values of the p p -rank on this moduli space. 1 Introduction Let k be an algebraically closed field of characteristic p>2 throughout this paper. We study the p -rank stratification of the moduli space of double covers of a fixed elliptic curve over k . In characteristic p , abelian varieties carry several interesting discrete invariants, such as the p -rank or the Newton polygon. The corresponding invariants for curves are defined to be those of their Jacobians. These invariants induce stratifications on the moduli spaces of abelian varieties, and consequently on the moduli spaces of curves. In this paper, we focus on the p -rank of curves. The p -rank of an abelian variety A over k is the integer f_{A}:=\dim(\text{Hom}(\mu_{p},A)) such that \#A[p](k)=p^{f_{A}} ; it satisfies 0\leq f_{A}\leq\dim(A) . The p -rank of a curve is defined to be the p -rank of its Jacobian. By [ Oda69 ] , the p -rank of a curve C over k also equals the semisimple rank of the Frobenius operator on H^{1}(C,\mathcal{O}_{C}) . The p -rank is invariant under isogenies, and the Newton stratification refines the p -rank stratification. The p -rank stratifications of various moduli spaces have been studied extensively. Given a moduli space \mathcal{M} of curves or abelian varieties, denote by V_{f}(\mathcal{M}) the closed p -rank f stratum , consisting of objects whose p -rank is at most f . It is known ( [ Oor74 , Lemmas 1.4 and 1.6] and [ Kob75 , Theorem 7.(1)] ) that V_{f}(\mathcal{A}_{g}) is pure of codimension g-f . It is a subtler question how the p -rank stratification intersects the Torelli locus. It is proved in [ FvdG04 , (2.3)] that closed p -rank strata of \overline{\mathcal{M}}_{g} are pure and have the expected codimension g-f . As a consequence, the generic point of any component of V_{f}(\overline{\mathcal{M}}_{g}) corresponds to a smooth curve. Similar investigations for certain covers of the projective line followed. In [ GP05 ] , it is shown that the same holds for the closed p -rank strata of \overline{\mathcal{H}}_{g} , the moduli space of hyperelliptic curves of genus g . The p -rank stratification of the moduli space of Artin-Schreier covers , separable degree p covers of the projective line, is explored in [ PZ12 ] . The p -rank strata of the moduli space of degree l cyclic covers of \mathbb{P}^{1} , where l\neq p is prime, are studied in [ OPW22 ] . In [ FvdG04 , (6.1)] , p -ranks of unramified double covers of genus 2 curves are studied. In [ Bou01 ] , ramified covers of degree coprime to p of generic r -pointed curves are studied for sufficiently large p . A natural next step is to investigate the p -ranks of potentially ramified double covers of a fixed elliptic curve E=(E,O_{E}) over k . Note that the p -rank f_{E} of E is either f_{E}=0 or f_{E}=1 , depending on whether E is supersingular or ordinary. Let \mathcal{B}_{E,g} denote the moduli space of smooth genus- g curves over k that admit a \mathbb{Z}/2\mathbb{Z} -cover to E , and let \overline{\mathcal{B}}_{E,g} denote its Deligne-Mumford compactification. Furthermore, let \mathcal{B}_{g} denote the bielliptic locus and \overline{\mathcal{B}}_{g} its Deligne-Mumford compactification. Our main result is that the closed p -rank strata of \overline{\mathcal{B}}_{E,g} and \overline{\mathcal{B}}_{g} are pure and have the expected (co)dimension. Theorem 1.1 . Let g\geq 2 and p>2 . For any f_{E}\leq f\leq g-1+f_{E} , the locus V_{f}(\overline{\mathcal{B}}_{E,g}) is pure of dimension g-2+f-f_{E} . As a consequence, V_{f}(\overline{\mathcal{B}}_{g}) is pure of dimension g-2+f , for any 0\leq f\leq g . Furthermore, our geometric insights lead us to the following corollary. Corollary 1.2 . The following statements hold for any g\geq 2 and any prime number p>2 . 1. For any 0\leq f\leq g-2 , there is a component of V_{f}(\overline{\mathcal{B}}_{g}) that consists entirely of singular curves. 2. For any 0\leq f\leq g , there is a smooth bielliptic curve C of genus g in characteristic p whose p -rank equals f , except for the case (p,g,f)=(3,2,0) . The spaces \mathcal{B}_{E,g} and \mathcal{B}_{g} are of particular interest because they have interesting arithmetic applications. Any smooth supersingular curve of genus g admits a map to a supersingular elliptic curve. By choosing E to be supersingular and considering the space \mathcal{B}_{E,g} , we learn more about the case where such a map has degree 2 . Understanding this could also lead to a better understanding of whether there exists a smooth supersingular curve of genus g . In fact, the existence of a smooth supersingular curve C of genus g=4 in characteristic p>3 (and hence for all p>0 ) was established in [ KHS20 ] , where C was constructed as the normalization of the fiber product of two (well-chosen) supersingular elliptic curves E_{i}\to\mathbb{P}^{1} , for i=1,2 , over \mathbb{P}^{1} . Note that the induced maps C\to E_{1} and C\to E_{2} are double covers. These moduli spaces are also relevant for Oort’s conjecture, which states that the generic point of any component of the locus of supersingular abelian varieties has the minimal automorphism group, that is \{\pm 1\} . An analog of this statement for supersingular curves was proved for g=4 and p>2 in [ Dra24 ] , which implies Oort’s conjecture for g=4 and p>2 . The proof uses geometric insights into the p -rank 0 stratum of \mathcal{B}_{4} . 1.1 Structure of this paper The proof of our main theorem makes extensive use of the boundary of \mathcal{B}_{E,g} in \overline{\mathcal{M}}_{g} , and a large part of the paper is devoted to understanding that boundary. The reason the boundary is so useful is that, while it can be extremely difficult to construct a smooth curve with a given p -rank, it is relatively easy to construct singular curves with a given p -rank. Namely, one can clutch together multiple well-understood curves of smaller genus at ordinary double points. This provides a point in V_{f}(\overline{\mathcal{B}}_{E,g}) . If the boundary components of \overline{\mathcal{B}}_{E,g} are well-understood, then, in many cases, one can argue by dimension reasons that V_{f}(\mathcal{B}_{E,g}) is dense inside V_{f}(\overline{\mathcal{B}}_{E,g}) and, in particular, also non-empty. Moreover, the dimension of the boundary can yield the dimension of the smooth locus. To this end, in Section 2 , we introduce the Deligne-Mumford stack \overline{M}_{E,g;n} , which parametrizes admissible double covers of a genus 1 curve with one component identified with E , along with n marked points on the source curve. The substack M_{E,g;n} consists of points where the source curve is smooth. This provides the appropriate setting for proving the results we need

Read the complete text