Abstract: In this paper, we prove effective upper bounds for effective sections of line bundles on projective varieties and hermitian line bundles on arithmetic varieties in terms of the volumes. They are effective versions of the Hilbert–Samuel formula and the arithmetic Hilbert–Samuel formula.
Introduction
In algebraic geometry, the Hilbert–Samuel formula is a fundamental result that describes the growth of the number of global sections of a line bundle $L$ as its power $N$ increases. Specifically, for a projective variety $X$ of dimension $n$, the dimension of the space of sections $h^0(X, L^{\otimes N})$ grows as a polynomial in $N$, where the leading coefficient is determined by the volume of the bundle: $\text{vol}(L)/n!$.
While this asymptotic behavior is well-understood, many applications—ranging from the proof of the Noether inequality to the study of specific algebraic surfaces—require "effective" versions of this formula. An effective bound provides a concrete inequality for a specific bundle $L$ (where $N=1$), rather than just describing the limit as $N \to \infty$. This paper establishes such explicit upper bounds for both geometric (projective varieties) and arithmetic (Arakelov geometry) settings across arbitrary dimensions.
The Geometric Case: Effective Hilbert–Samuel Bounds
The authors first address projective varieties over a field. The primary challenge in obtaining effective bounds is managing the error terms that arise when moving from an asymptotic limit to a specific instance. To solve this, the authors introduce a new invariant, $\epsilon(L)$, based on the pseudo-effectiveness of the bundle. This invariant allows for a more precise management of the "base locus"—the set of points where all sections of the bundle vanish.
Key Geometric Results
-
General Bound (Theorem 1.1): For a projective variety of dimension $n$, the number of global sections is bounded by: $$h^0(L) \leq \frac{1}{n!} \text{vol}(L) + n \epsilon(L)$$ This result provides a direct link between the volume and the number of sections, with an error term that scales linearly with the dimension.
-
The "Small" Case (Theorem 1.2): When the canonical bundle $\omega_X$ is "larger" than $L$ (specifically, if $\omega_X - L$ is pseudo-effective), the bound can be significantly sharpened. In such cases, the leading term is effectively halved: $$h^0(L) \leq \frac{1}{2(n!)} \text{vol}(L) + n \epsilon(L)$$ This generalizes classical results like Clifford’s theorem for curves to higher dimensions.
The Arithmetic Case: Arakelov Geometry
Arithmetic geometry extends these questions to varieties defined over number fields, using the framework of Arakelov theory. Here, line bundles are equipped with hermitian metrics, and the "size" of sections is measured by the arithmetic volume $\widehat{vol}(\bar{L})$.
The authors provide an arithmetic analogue to their geometric bounds (Theorem 1.3). This involves the introduction of the volume derivative ($dvol$), a metric-dependent invariant that captures how the arithmetic volume shifts under scaling. For arithmetic 3-folds (varieties of dimension 3 over a ring of integers), the paper offers a particularly refined result (Theorem 1.6) that avoids complex error terms, provided the variety satisfies certain conditions regarding its Kodaira dimension.
Methodology: The Rescaling Method
The core technical innovation of the paper is the rescaling method. Originally developed by the authors for the study of surfaces, this method is here generalized to higher dimensions. The strategy involves:
- Fibrations and Induction: The authors construct fibrations over curves to reduce $n$-dimensional problems to $(n-1)$-dimensional ones, allowing for a recursive proof structure.
- Filtration of Line Bundles: By using a sequence of blow-ups, the authors separate a line bundle from its base locus, creating a chain of base-point-free bundles that are mathematically easier to bound.
- Arithmetic Fujita Approximation: To handle bundles that are "big" but not necessarily "nef" (numerically effective) in the arithmetic setting, the authors utilize Fujita-type approximations to ensure the results apply to a broad class of varieties.
Significance and Future Work
This research bridges a significant gap between asymptotic theory and practical application in higher-dimensional geometry. By providing explicit error terms, the authors offer a toolkit that is more "effective" than previous iterations in the field. Notably, the rescaling method—initially an arithmetic tool—proves to be a powerful engine for proving purely geometric results.
However, some limitations remain. The sharpened "small case" (Theorem 1.2) currently relies on Hironaka’s resolution of singularities, which restricts its full application to fields of characteristic zero. Additionally, while the framework is valid for any dimension $n$, the most "clean" and intuitive versions of the arithmetic theorems are currently limited to 3-folds. Extending these simplified bounds to arbitrary dimensions remains a primary objective for future research in the field.
Summary updated