No abstract

No abstract

Simultaneous temperature traces at several heights in the first 8 m of the atmospheric surface layer indicate the presence of an organized structure with a spatially coherent upwind interface characterized by a sharp decrease in temperature. The convection speed of this interface, obtained by different methods, increases with height and is approximately equal to the local mean wind velocity. The inclination of this interface to the horizontal also increases with height. Associated with the sharp decrease in temperature/• at the interface is a relatively less sharp jump in streamwise u and a decrease in vertical w velocity fluctuations. Ensemble-averaged distributions of u, w, t•, uw, and w/•, relative to the location of the interface, have been obtained for several occurrences of the interface for different stability conditions. These distributions effectively represent signatures of the organized motion and are qualitatively similar to those obtained in the laboratory boundary layer. Signatures of velocity and temperature, normalized by the friction velocity and temperature, respectively, exhibit a fair degree of similarity over different stability conditions when the time measured from the interface is normalized by the average period between signatures. For unstable conditions these signatures contribute over 40% to the average local heat flux w/• and about 20% to the average Reynolds shear stress -uw. This contribution to w/• and -uw is less than 10% for nearly neutral or moderately stable conditions.

No abstract

A pair of parallel cold wires separated in either the vertical or lateral direction was used to obtain the three components 0,, 8,., 0. of the temperature derivative in the streamwise, lateral and vertical directions, respectively. The average absolute skewness values of 0, and 0, are nonzero and approximately equal, while the skewness of eY is approximately zero. These results appear to be consistent with the presence of a large, three-dimensional organised structure in the surface layer. There is an apparent low-frequency contamination in the spectral density of 0, and f?* due mainly to small errors in estimating the sensitivity of the cold wires. The temperature derivatives were high-pass filtered, the filter being set to remove possible contributions from the large structure and to minimise low-frequency sensitivity contamination. The filtered rms ratios gX/& and gX/iZ were in the range 0.7 to 0.9, a result in qualitative agreement with that obtained in the laboratory boundary layer by Sreenivasan eral. (1977). The skewness of filtered .9, or 8, is negligible, consistent with local isotropy of small-scale temperature fluctuations and in support of the high wavenumber spectral isotropy discussed in .