Lower bounds for resonances of infinite-area Riemann surfaces (Q638824)

Let \(X\) be a hyperbolic surface \(X=\Gamma\backslash\mathbb H^2\) of infinite area and finite geometry, \(\Delta_X\) its Laplace-Beltrami operator, \((\Delta_X-1/4-\lambda^2)^{-1}\) its resolvent originally defined for \(\text{Im}\,\lambda <0\) as an \(L^2\) bound operator and whose distributional kernel is meromorphically continued to \(\mathbb C\), \({\mathcal R}_X\) the set of its resonances defined as poles of this meromorphic continuation. For compact \(n\)-dimensional manifolds \(M\), the Laplace-Beltrami spectrum is purely discrete with \(L^2\) eigenvalues \((\mu_j)\) accumulating at \(+\infty\) and obeying the Weyl law \(\#\{\mu_j\leq \lambda^2\}\sim_{\lambda\to+\infty} C_n\text{vol}\,M\lambda^{n}\) where \(C_n=(4\pi)^{-n/2}/\Gamma(1+n/2)\): what kind of Weyl law does exist for resonances in \(\mathcal R_X\)? One basic geometric invariant is the Poincaré exponent \(\delta_X=(d_X-1)/2\) where \(d_X\) is the Hausdorff dimension of the trapping set \( K_X\) in the unitary tangent bundle \(S^1X\): \((x,\xi)\in{K}_X\) if the geodesic line determined by \((x,\xi)\) is bounded in \(X\); if \(M\) is compact with dimension \(n\), then \(\delta_M=n-1\) while for \(X=\Gamma\backslash\mathbb H^2\) as above \(\delta_X\in[0,1)\). Motivated by upper bounds proved by \textit{M. Zworski} [Invent. Math. 136, No. 2, 353--409 (1999; Zbl 1016.58014)] and \textit{K. K. Lin, M. Zworski} and the reviewer [Commun. Math. Phys. 245, No. 1, 149--176 (2004; Zbl 1075.11059)], a conjecture claims a fractal Weyl law for resonance in large strips, i.e. asserts the existence of constants \(C,D\) such that \[ D^{-1}\lambda^{1+\delta_X}\leq \# \{\rho\in\mathcal R_X;\, \text{Im}\,\rho\leq C,\, |\text{Re}\,\rho|\leq \lambda\} \leq D(1+\lambda)^{1+\delta_X},\quad \lambda >0. \] By the Dirichlet box-principle, this strip Weyl law would imply the existence of a sequence \((\lambda_k)_{k\in\mathbb N}\) with \(\text{Im}\,\lambda_k<C\) and \(\lambda_k\to\infty\) such that \[ \#({\mathcal D}(\lambda_k)\cap{\mathcal R}_X)\geq |\text{Re}\,\lambda_k|^{\delta_X-\varepsilon}. \] Jakobson and Naud give the first lower bounds related to the Hausdorff dimension \(\delta_X\) by proving similar disk lower bounds in the logarithmic region \(W_A=\{\text{Im}\,\lambda\leq A\log(1+|\text{Re}\,\lambda|)\}\): for \(\varepsilon>0\) and \(A>0\), there exists a sequence \((\lambda_k)_{k\in\mathbb N}\) with \(\lambda_k\in W_A\) and \(\lambda_k\to\infty\) such that: -- if \(\delta_X>1/2\), then \(\#({\mathcal D}(\lambda_k)\cap{\mathcal R}_X)\geq (\log|\text{Re}\,\lambda_k|)^{(\delta_X-1/2)/\delta_X-\varepsilon}\); -- if \(\delta_X>3/4\) and the covering group \(\Gamma\) is an infinite index subgroup of an arithmetic group derived from a quaternion algebra, then \(\#({\mathcal D}(\lambda_k)\cap{\mathcal R}_X)\geq |\text{Re}\,\lambda_k|^{2(\delta_X-3/4)-\varepsilon}\). The proofs are based on the wave trace formula and Dirichlet box-principle. For the arithmetic case, a trick of Selberg-Hejhal on mean square estimates is used as the properties of the length spectrum (high multiplicities and good separation) valid for arithmetic groups. The last section discusses some examples of surfaces with large Poincaré exponents.