Central limit theorems for the number of maxima and an estimator of the second spectral moment of a stationary Gaussian process, with application to hydroscience (Q5936310)

Analysis of the limiting behavior of distributions of various functionals (e.g., extremes, number of crossings, etc.) of Gaussian processes has been one of the most attractive subjects in the theory of stochastic processes during the last few decades. This work is an application of previous work of the authors [Stochastic Processes Appl. 66, No. 2, 237-252 (1997; Zbl 0890.60034)] in which the Hermite polynomial representations were given for the number of crossings and for the number of maxima of some smooth stationary Gaussian process. Those expansions are used in the present paper to study the asymptotic behavior of both the number of maxima in an interval and the square root of the second spectral moment of those processes. The main part of the paper concerns the number of maxima. The asymptotic behavior of functionals related to the maxima of a Gaussian process is of basic interest in extreme value theory and has very widespread applications in engineering and elsewhere. For instance, in hydroscience to evaluate responses of marine systems in a seaway it is necessary to obtain statistical information on wave height and period including extreme wave heights. One way to get some information on extreme waves is to study the number of maxima of the associated process. That is why the authors are interested to get a CLT for the number of maxima. For statistical motivation a consistent estimator of the asymptotic variance is proposed. The rest of the paper is devoted to the study of an estimator of the square root of the second spectral moment \(\sqrt{\lambda_{2}}\).