Local average in the hyperbolic sphere problem
Giacomo Cherubini, Christos Katsivelos · 2025
Read Local average in the hyperbolic sphere problem by Giacomo Cherubini, Christos Katsivelos (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-2503-20455.txt.
Introduction
Giacomo Cherubini and Christos Katsivelos studied a counting problem in hyperbolic geometry in twenty twenty-five. The hyperbolic sphere problem asks how many lattice points of the Picard group fall inside a growing region of three-dimensional hyperbolic space, and how large the error in the count is. Local averaging — smoothing the count over small patches — can tame that error.
They improve the remainder bounds for the locally averaged count, conditional on a quantum variance estimate for Maass cusp forms attached to the group. They also draw on bounds for a spectral sum over Laplace eigenvalues, a quantity studied for the prime geodesic theorem, where solid unconditional bounds exist. The paper is a technical contribution to analytic number theory at the meeting point of geometry and spectrum.
Catalog record
| CREATOR | Giacomo Cherubini, Christos Katsivelos |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 38 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-20455
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Local average in the hyperbolic sphere problem Giacomo Cherubini, Christos Katsivelos Abstract . We consider a local average in the hyperbolic lattice point counting problem for the Picard group Γ \Gamma acting on the three-dimensional hyperbolic space. Compared to the pointwise case, we improve the bounds on the remainder in the counting, conditionally on a quantum variance estimate for Maass cusp forms attached to Γ \Gamma . We also use bounds on a spectral exponential sum over the Laplace eigenvalues for Γ \Gamma , which has been studied in the context of the prime geodesic theorem and for which unconditional bounds are known. 2020 Mathematics Subject Classification Primary 11F72; Secondary 37C35, 37D40 1. Introduction Let {\mathbb{H}}^{3} be the three-dimensional hyperbolic space and let z\in{\mathbb{H}}^{3} . The Picard group \Gamma=\mathrm{PSL}_{2}(\mathbb{Z}[i]) acts naturally on {\mathbb{H}}^{3} (see Section 2 ) and therefore any \gamma\in\Gamma moves z to some point \gamma z , with the property that in any given compact set there are only finitely many translates of z . The problem of estimating the number of such translates in a ball of growing radius is called hyperbolic sphere problem. More specifically, let X\geq 1 and consider the function N(X,z):=\#\{\gamma\in\Gamma:\;\cosh d(z,\gamma z)\leq X\}, where d(z,w) denotes the hyperbolic distance. By definition, N(X,z) counts the number of lattice points \gamma z inside a ball of radius \cosh^{-1}X . The presence of the hyperbolic cosine is a common normalization in hyperbolic geometry, where distances grow exponentially, and is chosen so that the main asymptotic for N(X,z) becomes polynomial in X . As one may expect, N(X,z) is asymptotic to the volume of a hyperbolic ball of radius \cosh^{-1}X , divided by the volume of a fundamental domain for \Gamma . This can be proved with an explicit error term [ 8 , 15 , 16 ] in the form (1.1) N(X,z)=c_{\Gamma}X^{2}+O(X^{3/2}), where (cf. [ 8 , p.312] and [ 24 , Eq. (1.24)] ) c_{\Gamma}=\frac{\pi}{2\mathrm{vol}({\Gamma\backslash{\mathbb{H}}^{3}})}=\frac{3\pi}{2\mathcal{C}} and \mathcal{C}\approx 0.915965 is Catalan’s constant. Notice that the main term in ( 1.1 ) is independent of z . The implied constant in the error depends on z , but is uniformly bounded for z in a compact set. The exponent 3/2 in ( 1.1 ) has never been improved, for no choice of the point z , nor for groups other than the Picard group. However, several types of averages have been studied [ 10 , 15 , 24 ] , proving smaller exponents on average and hinting that the true order of magnitude of the error in ( 1.1 ) should be smaller than O\left(X^{3/2}\right) , possibly of the order O(X^{1+\epsilon}) for every \epsilon>0 . The aim of this paper is to consider a local average of the problem and measure how much we can reduce the exponent in the remainder under the assumption of two hypotheses that have been studied before in the literature: • a quantum variance bound for cusp forms attached to the group \Gamma ; • a bound on a spectral exponential sum over hyperbolic Laplace eigenvalues. Moreover, we can compare our results with the case of the modular group {\hbox{PSL}_{2}({\mathbb{Z}})} in two dimensions, where the same problem was studied by Petridis and Risager [ 21 ] (and by Biró [ 5 ] for any cofinite Fuchsian group). In two dimensions, the first point is in fact a theorem due to Luo and Sarnak [ 18 ] , and there has been a lot of progress on the corresponding spectral exponential sum [ 2 , 3 , 11 , 18 ] . In three dimensions, the analogous exponential sum has been studied extensively in [ 1 , 4 , 14 ] , but we are not aware of any non-trivial estimate on the quantum variance. We require some more notation and terminology to state the hypotheses we need in rigorous terms, and we postpone this to after Theorem 1.1 , see Hypothesis QV in § 1.1 and Hypothesis STX in § 1.2 respectively. In addition, for a smooth compactly supported function f on the Picard manifold {\Gamma\backslash{\mathbb{H}}^{3}} , we define its average value by the integral (1.2) \bar{f}=\frac{1}{\mathrm{vol}({\Gamma\backslash{\mathbb{H}}^{3}})}\int_{{\Gamma\backslash{\mathbb{H}}^{3}}}f(z)\,d\mu(z). Our main result can be stated as follows. Theorem 1.1 . Let {\Gamma}=\mathrm{PSL}_{2}(\mathbb{Z}[i]) and let f be a smooth compactly supported function on {\Gamma\backslash{\mathbb{H}}}^{3} , with its average value \bar{f} defined as in ( 1.2 ). Furthermore, assume: • Hypothesis QV with the exponent q\in[1,3] ; • Hypothesis STX with the exponents (7/4+\theta,1/4) , for some \theta\in[0,1/4] . Then, for every \epsilon>0 , we have (1.3) \int_{{\Gamma\backslash{\mathbb{H}}^{3}}}N(X,z)f(z)\,d\mu(z)=\frac{\pi}{2}\bar{f}X^{2}+O_{f,\epsilon,\theta,q}\left(X^{\frac{6-4\theta}{5-4\theta}+\epsilon}+X^{\frac{2q}{q+1}+\epsilon}\right). At this stage, even without knowing the exact formulation of Hypothesis QV and Hypothesis STX , the reader can verify that if q<3 (any \theta\in[0,1/4] ), the remainder in ( 1.3 ) is smaller than O(X^{3/2}) , improving ( 1.1 ) in the local average. See Remark 1.3 for more comments on the exponents one obtains in the remainder, depending on the values of q and \theta . 1.1. Quantum variance Let us give the precise formulation of our assumptions. The hyperbolic Laplace operator acts on the space of square-summable functions on {\Gamma\backslash{\mathbb{H}}^{3}} and admits a continuous spectrum as well as a discrete spectrum of eigenvalues 0=\lambda_{0}<\lambda_{1}\leq\lambda_{2}\leq\cdots\leq\lambda_{j}\to\infty. Eigenvalues \lambda_{j}\in[0,1] are called ‘small eigenvalues’, but it is known that for the Picard manifold the only such eigenvalue is \lambda_{0}=0 and that in fact \lambda_{1}>2\pi^{2}/3 , see [ 8 , Ch. 7, Proposition 6.2] . We can therefore write \lambda_{j}=1+r_{j}^{2} with r_{0}=i and r_{j}\in[1,\infty) if j\geq 1 . For each \lambda_{j} we denote by u_{j} the associated eigenfunction and if \lambda_{j} occurs with multiplicity, it will be listed a corresponding number of times, paired with distinct eigenfunctions. One can construct the spectral measures d\mu_{j}(z):=|u_{j}(z)|^{2}d\mu(z). When j\to\infty , these measures are expected to converge to the normalized uniform measure on {\Gamma\backslash{\mathbb{H}}^{3}} , a prediction referred to as the Quantum Unique Ergodicity Conjecture and known to be true for the modular surface in two dimensions [ 17 , 27 ] . In Theorem 1.1 we assume an explicit quantitative version of this result in square mean. Hypothesis QV . Let \Gamma be the Picard group. We say that Hypothesis QV holds with the exponent q if for all \epsilon>0 , T\gg 1 and all smooth compactly supported functions f on {\Gamma\backslash{\mathbb{H}}^{3}} , we have (1.4) \sum_{0<r_{j}\leq T}\;\biggl|\int_{{\Gamma\backslash{\mathbb{H}}^{3}}}f(z)d\mu_{j}(z)-\bar{f}\biggr|^{2}\ll_{f,\epsilon,q}T^{q+\epsilon}, where \bar{f} denotes the average value of f as defined in ( 1.2 ). It is easy to check that the trivial bound in ( 1.4 ) is O_{f}(T^{3}) by the Weyl law. As we mentioned earlier ( 1.4 ) was proved for the modular group by Luo and Sarnak [ 18 , Theorem 1.2] , with q=1 and with an explicit dependence on the test function f . Variance sums such as ( 1.4 ) were introduced by Zelditch [ 30 ] , who proved a logarithmic saving over the trivial bound given by the Weyl law. The result of Luo and Sarnak has been extended by Zhao [ 28 ] and later by Sarnak and Zhao [ 29 ] to non-holomorphic cusp forms. We also mention that for any negatively curved, compact Riemannian manifold \mathcal{M} of dimension at least 2 , it is conjectured (see e.g [ 19 , Equation (3)] ) that