Rank jumps for Jacobians of hyperelliptic curves on K3 surfaces

Ander Arriola Corpion and Cecília Salgado · 2026

Read Rank jumps for Jacobians of hyperelliptic curves on K3 surfaces by Ander Arriola Corpion and Cecília Salgado (2026) 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-2604-03639.txt.

Introduction

Ander Arriola Corpion and Cecília Salgado studied families of genus-2 curves lying on a K3 surface over a number field. They showed that after passing to a suitable finite extension, infinitely many fibers have Jacobians whose Mordell-Weil rank jumps above the usual value. They also gave geometric conditions under which such jumps occur on a large, non-thin set of fibers.

Catalog record

CREATORAnder Arriola Corpion and Cecília Salgado
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE70992
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.03639

Browse the full library · About the Tower of Babel

Full text

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

Rank jumps for Jacobians of hyperelliptic curves on K3 surfaces Ander Arriola Corpion and Cecília Salgado Abstract . We study Mordell–Weil rank jumps on families of jacobians of a pencil of genus-2 curves on a K3 surface defined over a number field k k . We exhibit a finite extension l / k l/k over which the subset of fibers for which the rank jumps is infinite. Moreover, we describe further geometric conditions on the K3 surface under which the rank jumps on a non-thin set of fibers. 1. Introduction Let k be a number field, S a smooth, geometrically integral curve over k , and \eta:\mathcal{A}\rightarrow S a non-constant family of abelian varieties. Let A be its generic fiber. Let l/k be a finite field extension. The group of l -sections of \eta is naturally isomorphic to the Mordell–Weil group A(l(S)) . By the Lang–Néron theorem [ 17 , Theorem 6.1] , this group is finitely generated; we denote its rank by r(\eta,l) . For a point s\in S(l) , let \mathcal{A}_{s}:=\eta^{-1}(s) be the corresponding fiber and let r_{s} denote the Mordell–Weil rank of \mathcal{A}_{s} over l . A specialization theorem by Silverman, which builds on work of Néron, [ 22 , 29 ] implies that the inequality (1) r_{s}\geq r(\eta,l) holds for all but finitely many s\in S(l) . The aim of this note is to study Mordell–Weil rank jumps within these families, i.e., the set \mathcal{R}(\eta,l)\ =\ \{\ s\in S(l)\mid r_{s}>r(\eta,l)\ \} of fibers for which the inequality in ( 1 ) is strict. Evidently, this is only interesting when S(l) is infinite. Our focus lies on families of jacobians associated to a pencil of genus- 2 curves on a K3 surface X defined over k . So, in our case, S\cong\mathbb{P}^{1} . K3 surfaces with a genus 2 curve admit a realization as a double cover X\rightarrow\mathbb{P}^{2}_{k} branched along a plane sextic curve B\subset\mathbb{P}^{2}_{k} [ 8 , Theorem 2.3] . In this setting, our main results are as follows: Theorem 1.1 . Assume that B contains a geometrically irreducible non-linear component defined over k . Then, there is a field extension l/k of degree at most 12 such that |\mathcal{R}(\eta,l)|=\infty. The existence of a genus- 1 fibration on X often allows us to obtain rank jumps on a non-thin set of fibers. A sufficient condition for this is the presence of a singularity on the branch sextic. More precisely, we obtain: Theorem 1.2 . Assume that B admits a simple singularity defined over k . Then, there is a field extension l/k of degree at most 18 such that \mathcal{R}(\eta,l) is not thin. Remark 1.3 . The field extension l/k appearing in Theorems 1.1 and 1.2 depends on the geometry of the branch curve B and on the field of definition of singular points in it. In many concrete cases, it can be chosen to be of smaller degree. For example, let \mathcal{L}_{d} be the family of U\oplus\langle-2d\rangle -polarized K3 surfaces. By [ 12 , Proposition 3.16] every X\in\mathcal{L}_{2m+5} admits a realization as a double cover X\rightarrow\mathbb{P}^{2}_{k} branched over a sextic with a simple node, and there exists a rational curve in \mathbb{P}^{2}_{k} of degree m+1 passing through the node with multiplicity m that splits in the double cover. So, in that particular case, the rank jump occurs over the base field k . Corollary 1.4 . Let X be a generic K3 surface in \mathcal{L}_{d} with d>3 , d\equiv 3 (mod 4), and let \eta:\mathcal{A}\rightarrow\mathbb{P}^{1}_{k} be the family of jacobians of a pencil of hyperelliptic curves of genus 2 in X . Then, \mathcal{R}(\eta,k)\subset\mathbb{P}^{1}_{k} is not thin. Finally, applying our results to a family of K3 surfaces of degree 2 studied in [ 21 ] yields the following: Corollary 1.5 . In the coarse moduli space \mathcal{M}_{2} of K3 surfaces of degree 2 there is a dense family of polarized K3 surfaces (X,\mathcal{L}) defined over \mathbb{Q} satisfying the following property: there is a dense set of curves in |\mathcal{L}| defined over \mathbb{Q} with Jacobian of positive Mordell-Weil rank over \mathbb{Q} . 1.1. Methodology and relation to the literature Several authors have addressed the variation of the Mordell–Weil rank in families of abelian varieties. Different methods have been employed to treat the problem in the context of elliptic surfaces, namely the study of root numbers pioneered by Rohrlich ( [ 25 , 32 , 7 ] ); the use of height theory as a tool in estimating the size of the set of fibers that witness a rank jump ( [ 1 ] ); and the description of multisections that induce linearly independent sections after base change ( [ 26 , 19 , 23 , 6 ] ). Our work builds on the third method. When \eta\colon\mathcal{A}\rightarrow S is a family of abelian varieties defined over a number field k , with S a smooth curve, Hindry and Salgado [ 16 ] proved that the k -unirationality of \mathcal{A} implies the existence of infinitely many fibers with a rank jump. Colliot-Thélène [ 5 ] later generalized this result, allowing S to be higher dimensional and showing, in particular, that whenever there exists a variety W with the Hilbert property and a dominant k -morphism W\to\mathcal{A} such that the composite W\to\mathcal{A}\to S has geometrically integral generic fiber, the rank jump set \mathcal{R}(\eta,k) is not thin in S . While our results deal with the special setting where \mathcal{A}\rightarrow S is the family of jacobians of a pencil of curves on a K3 surface, we do not need the hypothesis that the underlying variety \mathcal{A} is k -unirational nor the presence of a variety W as described above. Indeed, we treat K3 surfaces that may possess at most one genus- 1 fibration, or none at all. For such surfaces, the potential Hilbert property is generally unknown, and a suitable W is not known to exist, in general, for the associated jacobian fibration. K3 surfaces containing linear systems of hyperelliptic curves were studied systematically by Dolgachev [ 8 ] and Reid [ 24 ] , who obtained essentially equivalent classifications. Dolgachev terms such surfaces special , while Reid calls them hyperelliptic . In the terminology of [ 8 ] , the present paper concerns pencils of hyperelliptic curves of genus 2 on special K3 surfaces of class 2 . Even when the underlying K3 surface is known to satisfy the potential Hilbert property, for instance, when it admits multiple genus- 1 fibrations [ 14 ] , our approach provides an explicit bound on the degree of the field extension l/k over which \mathcal{R}(\eta,l) is not thin. Such a bound does not follow directly from [ 14 ] and [ 5 ] , as the former does not give an explicit bound on the degree of the field over which the Hilbert property holds. Our method relies crucially on the notion of saliently ramified multisections , introduced by Bogomolov and Tschinkel in the context of elliptic surfaces [ 3 ] . Pasten and Salgado [ 23 ] proved that after base change they become sections that are independent of the pullback of the original Mordell–Weil group, providing a tool to force rank jumps. This technique was applied successfully in [

Read the complete text