Abelian surfaces over finite fields containing no curves of genus $3$ or less

This paper characterizes isogeny classes of abelian surfaces over finite fields that contain no curves of genus 3 or less by completing the classification for genus up to 2, establishing a link between genus 3 curves and degree 4 polarizations, and describing the specific genus 3 curves that exist on such surfaces.

The Gist

Imagine you are an architect designing a special kind of city called an Abelian Surface. This isn't a city for people, but a mathematical landscape defined over a "finite field"—think of it as a universe with a limited, specific number of points, like a grid on a computer screen that can't get any bigger.

Your goal as the architect is to build this city so that no small, simple roads (curves) can exist within it. Specifically, you want to banish any road that is "simple" enough to be described by a low number of "holes" or twists, known in math as genus.

• Genus 0: A simple loop (like a circle).
• Genus 1: A donut shape (an elliptic curve).
• Genus 2: A pretzel with two holes.
• Genus 3: A pretzel with three holes.

Usually, if you build a city like this, it's impossible to avoid having at least one "donut road" (Genus 1) or a "two-hole pretzel road" (Genus 2). In fact, in a standard, infinite universe, every such city must contain a Genus 2 road.

The Big Question:
Can we design a city (an Abelian Surface) over a finite field that is so complex and twisted that it contains no roads of Genus 3 or less? In other words, can we force the smallest possible road in our city to be a "four-hole pretzel" or even more complex?

The authors of this paper, Elena, Alejandro, and Stefano, act like detectives solving this architectural mystery. Here is how they cracked the case, explained simply:

1. The "No-Go" Zones (Genus 0, 1, and 2)

First, they revisited a previous map they helped draw. They identified exactly which types of cities already have no Genus 0, 1, or 2 roads.

• The Analogy: Imagine you have a list of "Forbidden Blueprints." If you build your city using one of these blueprints, you are guaranteed that no simple loops or donuts can fit inside.
• They refined this list, distinguishing between roads that are "locally" simple (arithmetic genus) and roads that are "globally" simple (geometric genus). They found that some cities look like they have no simple roads, but if you zoom out to a bigger universe, you realize they actually do. They cleaned up the rules to make sure the "No Simple Roads" guarantee holds true no matter how you look at it.

2. The Secret Key: The "Degree 4 Lock"

This is the paper's biggest breakthrough. They wanted to know: If a city has no Genus 0, 1, or 2 roads, what does it take to also ban Genus 3 roads?

They discovered a magical key.

• The Metaphor: Imagine every city has a "security system" called a Polarization. This system locks the city's geometry.
  • A "Degree 1" lock is a standard key (Principal Polarization).
  • A "Degree 4" lock is a special, stronger key.
• The Discovery: The authors proved a stunning equivalence: A city contains a Genus 3 road IF AND ONLY IF it has a "Degree 4" lock.
• The Translation: If you want to ban Genus 3 roads, you don't need to check every possible road shape. You just need to check if the city lacks this specific "Degree 4" lock. If the lock is missing, the Genus 3 road cannot exist!

3. The Algorithm: Checking the Blueprint

Now that they have the rule ("No Degree 4 lock = No Genus 3 road"), they needed a way to check the blueprints.

• They used a mathematical tool called a Weil Polynomial. Think of this as the city's DNA or its blueprint code.
• They developed a method to look at this code and instantly tell: "Does this city have a Degree 4 lock?"
• The Result: They created a complete list of all the "Forbidden Blueprints" (Weil polynomials) that guarantee a city has no roads of Genus 3 or less. If your city's blueprint is on this list, you are safe!

4. What About the Genus 3 Roads That Do Exist?

Finally, they looked at the cities that do have Genus 3 roads (because they have the Degree 4 lock).

• The Shape: They found that these Genus 3 roads are very special. They are "Bielliptic," meaning they are essentially a double-layered cover of a donut (Genus 1). Imagine a road that wraps around a donut twice.
• The Traffic: They calculated how many "cars" (rational points) can drive on these roads. They found that these roads are surprisingly "empty." They don't have nearly as many cars as the theoretical maximum allowed by the laws of math (the Serre-Weil bound).
• The Lesson: Just because a road can exist in your city doesn't mean it's a busy highway. The constraints of the city's geometry force these roads to be sparse and quiet.

Why Does This Matter?

You might ask, "Who cares about banning pretzel roads in math cities?"

• Real-World Application: This research is crucial for Coding Theory. In the digital world, we send messages (like emails or satellite data) using "codes."
• The Connection: The "roads" on these mathematical surfaces are used to build error-correcting codes.
  • If a road is too simple (low genus), the code is weak and can be easily broken or corrupted.
  • If the city has no simple roads, the resulting code is incredibly strong and robust.
• By finding these "road-free" cities, the authors are helping engineers design better, more secure communication systems that can withstand noise and interference.

Summary

In short, these mathematicians acted like urban planners for a digital universe. They figured out exactly which city designs are so complex that they naturally reject all simple, low-genus roads. They found a "magic lock" (Degree 4 polarization) that determines if a Genus 3 road can sneak in, and they provided a checklist to ensure your city is secure, strong, and ready for the toughest digital codes.

Technical Summary

Here is a detailed technical summary of the paper "Abelian Surfaces Over Finite Fields Containing No Curves of Genus 3 or Less" by Elena Berardini, Alejandro Giangreco Maidana, and Stefano Marsiglia.

1. Problem Statement

The paper addresses the classical problem of determining the minimal genus of algebraic curves lying on an abelian variety. Specifically, it focuses on abelian surfaces defined over finite fields $\mathbb{F}_q$.

• Context: Over an algebraically closed field, every abelian surface is isogenous to a principally polarized one (either a Jacobian of a genus 2 curve or a product of two elliptic curves), implying it always contains a curve of geometric genus $\le 2$.
• The Challenge: Over finite fields, this is no longer true. Isogeny classes may lack principal polarizations, Jacobians, or products of elliptic curves. The authors aim to characterize isogeny classes of abelian surfaces that do not contain any curve of genus $\le 3$ (specifically, no $F_q$-irreducible curves of arithmetic genus $\le 3$).
• Motivation: This problem has applications in coding theory. Algebraic geometry codes constructed from divisors on abelian surfaces have minimum distances related to the minimum genus of curves on the surface. Finding surfaces with no low-genus curves allows for the construction of codes with larger minimum distances.

2. Methodology

The authors employ a combination of Honda-Tate theory, polarization theory, and Grothendieck group techniques developed by Howe.

• Weil Polynomials: The study is framed via the Weil polynomial $f(t) = t^4 + at^3 + bt^2 + aqt + q^2$ associated with the Frobenius endomorphism of the abelian surface. Isogeny classes correspond to these polynomials.
• Genus vs. Polarization: A central methodological insight is establishing a direct equivalence between the existence of a curve of a specific genus and the existence of a polarization of a specific degree.
  • Genus 2: Linked to principal polarizations (degree 1).
  • Genus 3: Linked to polarizations of degree 4.
• Algebraic Number Theory: The authors analyze the splitting behavior of the prime $2$in the CM field$K = \mathbb{Q}[t]/(f(t))$and its totally real subfield$K^+$. They utilize the ring of integers$\mathcal{O}$and the order$R = \mathbb{Z}[\pi, q/\pi]$.
• Grothendieck Groups: To determine if a specific polarization (kernel of order 4) exists within an isogeny class, they use the isomorphism between the Grothendieck group of finite group schemes in the isogeny class and the Grothendieck group of finite modules over the endomorphism ring (based on Howe's work). This allows them to translate geometric questions into algebraic conditions on the ideal class groups and Artin symbols.

3. Key Contributions and Results

A. Characterization of Surfaces with No Genus $\le 2$ Curves (Main Theorem 1)

The authors refine previous work to completely characterize isogeny classes containing no absolutely irreducible curves of geometric genus $\le 2$. They partition the relevant Weil polynomials into three sets:

1. $P_{npp}^{irr}$: Irreducible polynomials corresponding to surfaces not isogenous to a principally polarized surface.
2. $P_{Wres}^{irr}$: Irreducible polynomials corresponding to Weil restrictions of elliptic curves defined over $\mathbb{F}_{q^2}$.
3. Reducible cases: $\{(t^2-2)^2, (t^2-3)^2\}$.

Key Distinction: They clarify the difference between geometric genus and arithmetic genus. While a surface might lack absolutely irreducible curves of geometric genus 2, it might still contain $F_q$-irreducible curves of arithmetic genus 2 if it is principally polarizable but not a Jacobian.

B. Equivalence of Genus 3 Curves and Degree 4 Polarizations (Main Theorem 2)

This is the paper's core theoretical breakthrough.

• Theorem: For a simple abelian surface $A$ over $\mathbb{F}_q$, the following are equivalent:
  1. $A$ admits a polarization of degree 4.
  2. $A$ contains an $F_q$-irreducible curve of arithmetic genus 3.
• Implication: This reduces the geometric problem of finding genus 3 curves to the algebraic problem of checking for the existence of a polarization of degree 4.

C. Classification of Isogeny Classes with No Genus 3 Curves (Main Theorem 3)

Using the equivalence above and Howe's results on kernels of polarizations, the authors provide necessary and sufficient conditions for an isogeny class (defined by $f(t) \in P_{npp}^{irr} \cup P_{Wres}^{irr}$) to contain no abelian surface with a polarization of degree 4 (and thus no genus 3 curve).

The conditions depend on the splitting of the prime 2 in the field extensions:

• Case $f(t) \in P_{npp}^{irr}$: No surface in the class has a degree 4 polarization if and only if 2 is inert in the totally real subfield $K^+$.
• Case $f(t) \in P_{Wres}^{irr}$: The condition depends on the coefficient $b$ and the parity of $q$:
  • If ordinary: No degree 4 polarization iff $b = 1 - 2q$ and $q$ is odd.
  • If non-ordinary: No degree 4 polarization iff $q$ is even.

D. Algorithmic Verification

The authors describe an algorithm (based on the third author's previous work) to compute the isomorphism classes of abelian surfaces in a given isogeny class that admit a polarization of degree 4. This allows for the explicit enumeration of such surfaces for specific finite fields.

E. Properties of Genus 3 Curves

For surfaces that do contain genus 3 curves (but no lower genus curves), the authors characterize the curves:

• They are bielliptic (double covers of an elliptic curve).
• They are plane quartics (not hyperelliptic).
• Point Count Bounds: They derive bounds on the number of rational points $|C(\mathbb{F}_q)|$. They show that such curves are "far from maximal," meaning their point counts are significantly lower than the Serre-Weil bound for genus 3 curves.

4. Significance

• Theoretical Completeness: The paper provides a complete classification of abelian surfaces over finite fields that avoid curves of genus $\le 3$. It resolves gaps in previous literature regarding the distinction between arithmetic and geometric genus in this context.
• Algorithmic Utility: By linking the existence of curves to the existence of polarizations of specific degrees, the paper enables the use of existing computational tools (based on endomorphism rings and ideal classes) to search for these surfaces.
• Coding Theory Applications: The results directly assist in the construction of Algebraic Geometry (AG) codes with high minimum distances. By identifying surfaces with no low-genus curves, one can ensure the underlying divisors used for code construction do not contain "small" subvarieties that would degrade the code's performance.
• Arithmetic Geometry: The work deepens the understanding of the interplay between the arithmetic of Weil polynomials (splitting of primes) and the geometry of the underlying varieties (existence of curves and polarizations).

In summary, the paper successfully bridges the gap between the abstract theory of abelian varieties over finite fields and concrete geometric properties (curves), providing a rigorous framework to identify and construct surfaces that are "curve-free" up to genus 3.