Hyperuniformity and hyperfluctuations of random measures in commutative spaces

We introduce the notion of Bartlett spectral measure for isometrically invariant random measures on proper metric commutative spaces. When the underlying Gelfand pair corresponds to a higher-rank, connected, simple matrix Lie group with finite center and a maximal compact subgroup, we prove that the number variance is asymptotically bounded above, uniformly across all random measures, by the volume raised to a power strictly less than 2. On Euclidean and real hyperbolic spaces we define a notion of heat kernel hyperuniformity for random distributions that generalizes hyperuniformity of random measures, and we prove that every sufficiently well behaved spectral measure can be realized as the Bartlett spectral measure of an invariant Gaussian random distribution. We also compute Bartlett spectral measures for invariant determinantal point processes in commutative spaces, providing a spectral proof of hyperuniformity for infinite polyanalytic ensembles in the complex plane, as well as heat kernel non-hyperuniformity of the zero set of the standard hyperbolic Gaussian analytic function on the Poincar\'e disk.