Characterizing the relative role of low-frequency and turbulent processes in the nocturnal boundary layer through the analysis of two-point correlations of the wind components

The study presents an analysis of two-point correlations between time series of nocturnal atmospheric wind, obtained from two micrometeorological towers, 45 m horizontally apart, each equipped with two sonic anemometers, 2.5 m vertically apart. It focuses on the scale dependence of the two-point correlations obtained from sensors vertically and horizontally separated. In particular, the role of low-frequency non-turbulent processes in the correlations is assessed, and compared to that of the turbulent scales of motion. The vertical correlations of the streamwise and vertical wind components show little dependence on the turbulence intensity, but those of the spanwise component decrease appreciably as it gets more turbulent. Multiresolution decomposition shows that the two-point correlations become increasingly dominated by low-frequency scales as it gets less turbulent, and that such large-scale processes are largely reduced in fully turbulent conditions. It is also shown that the vertical correlations of the spanwise wind component is negative for very small time scales. Horizontal two-point correlations obtained at the 45 m separation distance between the towers are almost entirely dominated by low-frequency motions, regardless of the turbulence intensity, but the magnitude of such correlations decreases with increasing turbulence intensity for any wind components. A comparison between the horizontal two-point correlations and autocorrelations taken with a time lag given by the ratio of the horizontal separation to the mean wind component in the direction that connects the two towers leads to the conclusion that the statistical properties of turbulence are often preserved over the horizontal distance, despite the lack of turbulence correlations for that separation.