Strange examples of local signatures for fibered surfaces of small genus
Makoto Enokizono

TL;DR
This paper presents novel examples of local signatures for fibrations of genus 2 and 3, expanding the understanding of local invariants in fibered surface theory.
Contribution
It introduces new types of local signatures that differ from traditional ones for genus 2 and 3 fibrations.
Findings
Examples of local signatures for genus 2 and 3 fibrations are provided.
The new local signatures are fundamentally different from the usual ones.
The work broadens the scope of local signature concepts in algebraic geometry.
Abstract
We give examples of local signatures, completely different from the usual ones, for general fibrations of genus and genus .
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAlgebraic Geometry and Number Theory · Geometric and Algebraic Topology · Geometry and complex manifolds
Strange examples of local signatures for fibered surfaces of small genus
Makoto Enokizono
Makoto Enokizono, Department of Mathematics, Graduate School of Science, Osaka University, Toyonaka, Osaka 560-0043, Japan
Abstract.
We give examples of local signatures, completely different from the usual ones, for general fibrations of genus and genus .
2010 Mathematics Subject Classification:
14D06
Keywords: fibration, local signature
Introduction
For a closed oriented real -manifold , the signature of is defined to be the signature of the intersection form , which is a symmetric bilinear form. We consider the situation that admits a fibration over a closed oriented real surface . Under some conditions, the signature of happens to localize around a finite number of fiber germs , , …, :
[TABLE]
We call this phenomenon a localization of the signature and the value a local signature of . A first example of local signatures is the one for genus fibrations due to Matsumoto [14]. He also gave a local signature for Lefschetz fibrations of genus in [15], which was generalized by Endo [9] for hyperelliptic fibrations. Later, Kuno [13] defined a local signature for non-hyperelliptic fibrations of genus . On the other hand, Horikawa [12] defined a function on the set of holomorphic fiber germs of genus , which is nowadays called a Horikawa index, in order to study algebraic surfaces of general type near the Noether line. Once a Horikawa index is defined (for a certain type of holomorphic fibrations), we can define a local signature by using it, as shown in [4]. After Horikawa’s work, Xiao [19] and Arakawa-Ashikaga [1] defined a Horikawa index and a local signature for hyperelliptic fibrations. Terasoma [18] showed the coincidence of Endo’s local signature and Arakawa-Ashikaga’s one. For non-hyperelliptic genus fibrations, Reid [16] defined a Horikawa index. Similarly to Terasoma’s proof, Kuno’s local signature and Reid’s one for non-hyperelliptic fibrations of genus also coincide (cf. [3]).
In this short note, in the algebro-geometric category, we construct a local signature associated with an effective divisor on the moduli space of smooth curves of genus and compute some examples of local signatures for general fibrations of genus or , which are different from Endo-Arakawa-Asikaga’s one and Kuno-Reid’s one. The idea of constructions is essentially due to Ashikaga-Yoshikawa [5], who called the divisor on the moduli space of stable curves of genus the signature divisor and gave a local signature by pulling back the signature divisor using a geometric meaningful effective divisor , e.g., the Brill-Noether locus, via the moduli map of a fiber germ. Replacing by another effective divisor, the associated local signature varies. We compute local signatures in the case that and is the bielliptic locus and that and is the locus of curves having a hyperflex.
Acknowledgment*.*
I would like to express special thanks to Prof. Kazuhiro Konno for many comments and supports. Thanks are also due to Prof. Tadashi Ashikaga for useful advises and discussions. The research is supported by JSPS KAKENHI No. 16J00889.
1. Local signature associated with an effective divisor on
Let and respectively denote the moduli space of smooth curves of genus and the moduli space of stable curves of genus . The rational Picard group of is generated freely by the Hodge bundle and the boundary divisors for , where we use the notation in [11]. Let be an effective divisor on and the compactification of in . Then we can write for some rational numbers , where the symbol means the -linear equivalence.
Let be a surjective morphism from a complex smooth projective surface to a smooth projective curve whose general fiber is a smooth projective curve of genus , which is called a fibered surface or a global fibration of genus . Let denote the relative canonical bundle of and put
[TABLE]
[TABLE]
where is the genus of and the topological Euler number of .
Let be a relatively minimal degeneration of curves of genus , that is, is a surjective proper morphism from a complex smooth surface to a small open disk such that is a smooth curve of genus for any and the central fiber has no -curves. We take the stable reduction of via . Resolving singularities of , we obtain a semi-stable reduction . Note that can be taken as the pseudo-period of the topological monodromy of as a pseudo-periodic class (cf. [2]). Put and . Let
[TABLE]
be the local signature defect of (more precisely, see [2]) and
[TABLE]
On the other hand, the local invariants , and were defined in [17] for a fiber germ of a global fibration . Indeed,
Proposition 1.1**.**
We have and
[TABLE]
Proof.
These invariants satisfy the following properties: Let be a fibered surface of genus and be the semi-stable reduction of via a cyclic covering of degree . Then we have
[TABLE]
Let be an arbitrary fiber germ in a global fibration . Taking base change, we may assume that any fiber of other than is semi-stable. Thus we get the assertion from Hirzebruch’s signature formula , Noether’s formula and (1.1) since for any semi-stable fiber germ .
Let be the moduli map of the semi-stable reduction . For an effective divisor on not containing the image , we can define the pull-back . Let ). Note that even when holds for two effective divisors and , it is not always true that because we treat local fibrations here. Given an effective divisor on such that does not contain with , we put
[TABLE]
In general, for a relatively minimal fiber germ , we define
[TABLE]
and
[TABLE]
which are independent of the choice of .
Now we consider a global fibration , that is, a surjective morphism from a smooth projective surface to a smooth projective curve with connected fibers. Assume that the moduli point of the general fiber of is not contained in . From (1.1), we have
[TABLE]
From Hirzebruch’s signature formula , we can write
[TABLE]
We call the local signature of a fiber germ associated with . Note that the divisor is called the signature divisor in [5].
2. Examples
Now we consider two effective divisors and on , which parameterize curves of genus having a special Weierstrass point. Let be a smooth curve of genus . Let be a Weierstrass point of , i.e., a point on satisfying . Then is said to be exceptional of type (resp. of type ) if (resp. ). The locus (resp. ) on is (roughly) defined by the set of curves of genus with an exceptional Weierstrass point of type (resp. of type ) with the natural scheme structure, which is of codimension for . For more details, see [8]. For , the loci and are empty. For , is coincide with the hyperelliptic locus as a set, but as a divisor, we have . Indeed, once a genus curve has one exceptional Weierstrass point of type , it becomes hyperelliptic and hence has Weierstrass points of type automatically. Since the hyperelliptic Weierstrass point is exceptional of type and , the hyperelliptic locus is contained in both and . In particular, is a subdivisor of . Thus we can define an effective divisor . As a different definition, let be the locus on the moduli space of smooth plane quartics parameterizing plane quartic curves with a hyperflex, i.e., -fold tangent point. Then the above is just the closure of in . The locus has multiplicity around general points. For , and also have multiplicity around general points. It is known that the rational divisor classes of and are given by
[TABLE]
(cf. [8], [6], [7]). In particular, we have
[TABLE]
Now, we will check using the simplest example of fibered surface of genus that two local signatures and associated with and give different localizations.
Example 2.1*.*
Let be a general Lefschetz pencil of quartics. The base locus of consists of points and they are on smooth members. Blowing up at these points, we obtain a non-hyperelliptic fibration of genus . By a simple computation, we get , , and . Note that all singular fibers of are irreducible curves with one node and the number of them is . Thus we have , , and . Hence we have . This implies that the number of smooth curves in a general Lefschetz pencil of quartic curves with a hyperflex is . Let and respectively be a smooth quartic fiber germ of with one hyperflex and an irreducible fiber germ of with one node. Then clearly we have
[TABLE]
and
[TABLE]
Thus we get
[TABLE]
and
[TABLE]
Thus two local signatures and are different.
Next, let us consider the genus case. The rational Picard group of is generated by , and with one relation . For a semi-stable fiber germ of genus , we put . For a not necessarily semi-stable fiber germ , we define by using the semi-stable reduction similarly as in the previous section. We also define a (pre-)Horikawa index for a relatively minimal genus fiber germ . It coincides with the original Horikawa index defined by using the double covering data (cf. [18], [12], [19]) and hence it is non-negative. A local signature can be defined by for any fiber germ of genus .
Now, we define another local signature for non-bielliptic genus fiber germs. Let be the bielliptic locus on and its closure in . They are irreducible codimension loci. From [10], the rational linearly equivalence class of is
[TABLE]
Thus, for non-bielliptic genus fiber germs, two localizations of the Hodge bundle can be realized as follows. We put
[TABLE]
and
[TABLE]
for a semi-stable non-bielliptic fiber germ of genus . By using semi-stable reduction, we define , for any non-bielliptic fiber germ of genus . Then , are local signatures for genus non-bielliptic fibrations.
Example 2.2*.*
Let , and respectively be non-bielliptic genus fiber germs the image of whose moduli map meets , and transversally (and does not meet other loci among them) at the moduli point of the central fiber. Then we have
[TABLE]
[TABLE]
[TABLE]
For example, take a general member in the complete linear system , on and construct the double covering branched over . Then the composite of the double covering and the first projection is a non-bielliptic fibration of genus . By a simple computation, we have
[TABLE]
Since is general, we may assume that any singular fiber germ of is of type as above. Thus the number of fiber germs of type , and is , [math] and , respectively.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] T. Arakawa and T. Ashikaga, Local splitting families of hyperelliptic pencils I, Tohoku Math. J. 53 (2001), 369–394; II, Nagoya Math. J. 175 (2004), 103–124.
- 2[2] T. Ashikaga, Local signature defect of fibered complex surfaces via monodromy and stable reduction, Comment. Math. Helv. 85 (2010), 417–461.
- 3[3] T. Ashikaga, Horikawa index of genus three via signature divisor, I, preprint.
- 4[4] T. Ashikaga and K. Konno, Global and local properties of pencils of algebraic curves, Algebraic Geometry 2000 Azumino, S. Usui et al. eds, 1-49, Adv. Stud. Pure Math. 𝟑𝟔 36 {\bf 36} , Math. Soc. Japan, Tokyo, 2002.
- 5[5] T. Ashikaga and K. Yoshikawa, A divisor on the moduli space of curves associated to the signature of fibered surfaces (with an Appendix by K. Konno), Adv. St. Pure Math. 56 (2009), 1–34.
- 6[6] F. Cukierman, Families of Weierstrass points, Duke Math. 58 (1989), 317–346.
- 7[7] C. Cumino, E. Esteves and L. Gatto, Limits of special Weierstrass points, Int. Math. Res. Papers. (2008), 1–65.
- 8[8] S. Diaz, Exceptional Weierstrass points and the divisor on moduli that they define, Mem. Amer. math. Soc. 56 (1985).
