Effects of atmospheric nonstationarity on models of low-frequency surface wind noise

Abstract

Measurements of sound propagation in the atmosphere are often limited by wind noise, particularly at low frequencies. The spectrum of wind noise at the ground surface, i.e. turbulent static pressure, can be modeled by the dominant shear-turbulence mechanism within the atmospheric boundary layer. This model assumes statistically-stationary boundary-layer turbulence, and as in the atmospheric sciences, quasi-stationarity is hypothesized to justify the assumption. However, this hypothesis is not uniformly valid, due to the diurnal cycle and large-scale atmospheric dynamics, and should be tested. This study applies several existing tests of nonstationarity to atmospheric measurements from a field experiment near Laramie, Wyoming and characterizes their relevance to simultaneous recordings of wind noise by flush-mounted infrasound sensors. By estimating the parameters for a mirror flow model of boundary-layer turbulence from the anemometer records, surface pressure model spectra are compared with the experimental wind noise. The performance of both the atmospheric turbulence and wind noise models are assessed with respect to each nonstationarity test.