Estimating the convective mixed layer height from surface-layer variables

Researchers have known for decades that horizontal velocity variances s_u² and s_v² in the surface layer during daytime conditions scale not with surface-layer parameters (Lumley and Panofsky 1964) but rather with the mixed-layer scales w_* and z_i. Because w_* includes z_i in its definition, the relationship between w_* and (s_u², s_v²) admits the possibility of estimating z_i using only variables measured by tower in the surface layer. The critical question is: are measured values of the surface-layer horizontal variances accurate enough to yield useful estimates of z_i? A pilot study used preliminary estimates of z_i from profiler backscatter data and plotted variances from flux-PAM stations. It took maximum afternoon values of z_i and typical values of s_u² and s_v² for the afternoon maximum heating period. The data were from several days during the Flatland experiment on the Illinois prairie in June-August 1996. The pilot study concluded that the technique showed promise.

The present study expands upon the previous results, using final, quality-controlled data, including z_i measurements, from Flatland. It uses calculated, rather than estimated, (s_u², s_v²), surface-heat-flux, and z_i data for several hours on each of several days, and it continues to show good agreement between measured and calculated values of z_i. Complicating factors in the data, including convective clouds, large values of latent heating, and high wind speeds all affect this relationship and will be discussed.