Sparse equidistribution of geometric invariants of real quadratic fields

Abstract

Duke, Imamoḡlu, and Tóth have recently constructed a new geometric invariant, a hyperbolic orbifold, associated to each narrow ideal class of a real quadratic field. Furthermore, they have shown that the projection of these hyperbolic orbifolds onto the modular surface $\Gamma \backslash \mathbb{H}$ equidistributes on average over a genus of the narrow class group as the fundamental discriminant $D$ of the real quadratic field tends to infinity. We extend this construction of hyperbolic orbifolds to allow for a level structure, akin to Heegner points and closed geodesics of level $q$. Additionally, we refine this equidistribution result in several directions. First, we investigate sparse equidistribution in the level aspect, where we prove the equidistribution of level $q$ hyperbolic orbifolds when restricted to a translate of $\Gamma \backslash \mathbb{H}$ in ${\Gamma }_0(q)\backslash \mathbb{H}$, which presents some new interesting features. Second, we explore sparse equidistribution in the subgroup aspect, namely equidistribution on average over small subgroups of the narrow class group. Third, we prove small scale equidistribution and give upper bounds for the discrepancy. Behind these refinements is a new interpretation of the Weyl sums arising in these equidistribution problems in terms of adèlic period integrals, which in turn are related to Rankin–Selberg $L$-functions via Waldspurger’s formula. The key remaining inputs are hybrid subconvex bounds for these $L$-functions and a certain homological version of the sup-norm problem.