15th Conference on Boundary Layer and Turbulence

Friday, 19 July 2002: 8:45 AM
Velocity spectra in the marine atmospheric boundary layer
Ann-Sofi Smedman, Uppsala University, Uppsala, Sweden; and A. Sjöblom
Velocity spectra in the marine atmospheric boundary layer

In some respects the turbulence structure in the marine atmospheric boundary layer (MABL) reacts in the same way as the boundary layer over land, that is to say Monin-Obukhov similarity theory can be applied. There are, however, frequently situations when the similarity between the two breaks down. Analysis of measurements, taken at the flat, small island Oestergarnsholm in the middle of the Baltic Sea, clearly shows the influence of the sea state on MABL. The measurements comprise turbulence and mean variables taken at several heights on a 30 m tower, as well as wave parameters from a Wave Rider Buoy deployed 3 km outside the island. Model results of the wave field around the island together with foot-print analysis indicate that the wave field is almost undisturbed for low to moderate wind speeds but has to be corrected for limited water depth for the highest wind speeds.

Our earlier analysis shows a strict similarity with over-land conditions for both mean and turbulence variables (mean wind gradient, fluxes, spectra etc.) for growing waves (young waves) travelling slower than the wind. But as soon as some waves become mature and get a speed faster than the wind speed, similarity breaks down. Thus the turbulence structure in the MABL needs to be described in terms of additional parameters such as wave age and maybe boundary layer height.

Spectra of the velocity components in the MABL have been analyzed taking sea state into account. During neutral stability and young sea spectra follow the new similarity theory proposed by Hunt and Carlotti (2000) and Högström, Hunt and Smedman (2001). But with increasing wavelength of the surface waves spectra gradually change both shape and energy level, beginning at the low frequency end and continuing towards higher frequencies. For cp/U ~ 1 (where cp is the peak phase speed) the 'breaking point' can be seen in the inertial subrange, which actually gives two frequency intervals with a -2/3 slope but with different spectral levels. This will have implications concerning the 'inertial dissipation method'.

Supplementary URL: