Critical test of the validity of Monin-Obukhov similarity during convective conditions

Abstract

A recent study of convective boundary layer characteristics performed with large eddy simulation technique (LES) has demonstrated unexpected influence of the depth of the boundary layer on surface layer characteristics. The present study tests some of the predictions from these simulations with field measurements from a summertime experiment in Sweden, which includes in addition to regular surface layer data also airborne measurements and numerous radio soundings, which enable accurate determination of boundary layer depth. It is found that the measurements strongly support most of the conclusions draws from the LES study and give additional information over a wider stability range. Thus, the normalized wind gradient Φm is found to depend on both z/L, where z is height above the ground and L is the Monin-Obukhov length, and z1 /L, where z1 is the height of the convective boundary layer. This additional dependence on z1 /L explains much of the scatter between experiments encountered for this parameter. In the case of the normalized temperature gradient Φm, the experimental data follow the generally accepted functional relation with z/L, but with an additional, slight ordering according to z1/L. Analyses of nondimensional variances show (i) the horizontal velocity variance scales on mixed layer variables and is a function only of z1/L, in agreement with the LES results and with previous measurements; (ii) the normalized vertical velocity variance depends on the large-scale pressure gradient length scale for slight instability and is primarily a function of z/L for moderate and strong instability; (iii) the normalized temperature variance is a function of z/L, with a possible slight dependence on z1/L; and (iv) whereas mean temperature gradient is characterized by local shear scales, temperature variances are normalized by local buoyancy-driven scales.