Extended Abstract (236K)4.25
Laboratory experiments of entrainment in dry convective boundary layers
Harm J. J. Jonker, Delft University of Technology, Delft, Netherlands; and T. Heus, E. Hagen, and H. V. Dop
The entrainment flux across the inversion is still an important issue in atmospheric boundary layers. For the dry convective boundary layer, 'communis opinio' seems to have it that the entrainment buoyancy flux can be readily expressed in terms of the surface buoyancy flux by: <w'b'>_entr = A <w'b'>_surf, where the factor A incorporates a mild dependence on the Richardson number (as a measure of the relative inversion strength) to account for the depth of the entrainment zone (e.g. Sullivan et al. '98). For large values of Ri (small entrainment zones), A is thought to reach a saturation value between 0.1 and 0.2 - a view corroborated by large eddy simulations.
In this study we make use of a saline convection tank set-up as a laboratory model for a dry convective atmospheric boundary layer. Its lateral dimensions are 1m x 1m, and the typical boundary layer depth ranges between 0.1 and 0.2m. Rather than starting with a salinity lapse-rate, we begin with a two-layer system, where on top of the saline layer (the mixed layer) we have carefully placed a layer with lower salinity concentration, thus creating a well-defined inversion jump. In contrast to previous studies on entrainment, we do not so much focus on the entrainment velocity, but rather on the entrainment flux, which we measure with Planar Laser Induced Fluorescence (PLIF). The turbulence velocities have been measured using Particle Image Velocimetry (PIV), in order to test whether the convection tank set-up displays the correct CBL characteristics. Several experiments have been carried out over a wide range of inversion strengths (Ri ranging from 10 to 150). From the resulting PLIF-sheets vertical flux profiles have been derived. The most striking observation is that the entrainment flux is found to reveal a much stronger dependence on the inversion strength (Ri) than large eddy simulations tend to do: especially for large Ri-numbers, the entrainment flux is found to vanish, rather than to converge to a constant non-zero value.
Sullivan P.,C-H. Moeng, B. Stevens, D. Lenschow and Sh. Mayor, Structure of the entrainment zone capping he convective atmospheric boundary layer, JAS 55, 3042-3064, 2003.
Session 4, boundary layers: stable, and convective, and transitional (Parallel with Sessions 5 and 6)
Tuesday, 10 August 2004, 8:00 AM-5:15 PM, Vermont Room
Previous paper Next paper
Recorded presentation