An asymptotic formula for the variance of the number of zeroes of a stationary Gaussian process (Q6090872)

For a stationary Gaussian process \(f:\mathbb R \rightarrow \mathbb R\) with continuous covariance kernel \(r(t)\) and the spectral measure \(\rho\) an asymptotic formula describing the grow of the variance of the number of zeroes \(N(T)\) in a long time interval \([0,T]\) is given. The main result of the paper (Theorem 1) contains three parts which describe, under mild mixing conditions, both the upper and lower bounds for \(\mathrm{Var}[N(T)]\) and the effect of atoms in the spectral measure. So it is shown that adding a small atom to the spectral measure at a special frequency does not change the asymptotic growth of the variance, while an atom at any other frequency results in maximal growth. Also some conjectures to generalize the results are given.