Turbulence structure of the Astex First Lagrangian boundary and cloud layer

Modeling stratocumulus clouds is a challenge since many processes, that are sub-grid scale in most models, interact in a very complex non-linear way. The most important - turbulence, radiation and cloud physics - all require parameterization. An error in either one will produce an erroneous result in all the other. Stratocumulus clouds are important because of their effect on the Earth radiation balance, and thus on climate. However, to even begin es-timate the magnitude cloud-related feedbacks in climate models, the parameterizations must be adequate. Such development can only rely on basic knowledge about the physics of the processes and this can only come from field data or from resolved-scale numerical simulations.

Processes within the cloud itself generate much of the turbulence in the stratocumulus-capped marine boundary layer. The main source of turbulence is cloud-top radiative net cooling, causing buoyancy-generated turbulence. Drizzle or solar heating complicate this by altering the static stability within the cloud or below the cloud base. If a full decoupling results, the turbulence in the cloud loses contact with the surface and it becomes controlled by processes in the cloud and at the cloud top only. Yet many models assume that the struc-ture of the marine boundary layer, with or without clouds, is dependent on surface proper-ties, which may be far from correct.

In this study we have used turbulence data from the five aircraft missions of the ASTEX First Lagrangian to study the structure of the turbulence below and within an evolving cloud layer. Variances and fluxes were calculated both from flight-legs and slant profiles. Buoyancy-flux profiles were used to determine layer separation, and the data are then scaled separately in the sub-cloud boundary layer and within the cloud. Spectra and cospectra were calculated from the flight legs and were also scaled separately.

The results indicate, as expected, that the turbulence in the layer below the cloud scale with height above the surface up to the top of the sub-cloud layer. During the first three flights, with al least partial decoupling, this layer coincides with the cloud base. While the cloud base is lifted in the final two flights, the top of this layer remains essentially unchanged, leaving an intermediate layer up to the cloud base. The turbulence in the cloud scales with cloud depth; variances and fluxes scale with convective scaling, using the depth of the cloud and an extrapolated buoyancy flux to the cloud top to calculate the convective veloc-ity scale. Dominant peaks in the spectra are at wavelengths slightly larger than the cloud depth for w-spectra and even larger for the horizontal wind components, but are only approximately constant with height. Spectra of the horizontal velocity and cospectra of the fluxes have considerable low-frequency variability, indicative of mesoscale motions. A calculated dissipation length scale becomes constant in time and space when normalized with the cloud depth. These results should be helpful in designing parameterizations of tur-bulence in stratocumulus.