Investigation of the pressure–strain-rate correlation and pressure fluctuations in convective and near neutral atmospheric surface layers

The pressure–strain-rate correlation and pressure fluctuations in convective and near neutral atmospheric surface layers are investigated. Their scaling properties, spectral characteristics, the contributions from the different source terms in the pressure Poisson equation and the effects of the wall are investigated using high-resolution (up to $2048^{3}$) large-eddy simulation fields and through spectral predictions. The pressure–strain-rate correlation was found to have the mixed-layer and surface-layer scaling in the strongly convective and near neutral atmospheric surface layers, respectively. Its apparent surface-layer scaling in the moderately convective surface layer is due to the slow variations of the mixed-layer contribution, and is an inherent problem for single-point statistics in a multi-scale surface layer. In the strongly convective surface layer the pressure spectrum has an approximate $k^{-5/3}$ scaling range for small wavenumbers ($kz\ll 1$) due to the turbulent–turbulent contribution, and does not follow the surface-layer scaling, where $k$ and $z$ are the horizontal wavenumber and the distance from the surface respectively. The pressure–strain-rate cospectrum components have a $k^{-1}$ scaling range, consistent with our prediction using the surface layer parameters. It is dominated by the buoyancy contribution. Thus the anisotropy in the surface layer is due to the energy redistribution caused by the density fluctuations of the large eddies, rather than the turbulent–turbulent (inertial) effects. In the near neutral surface layer, the turbulent–turbulent and rapid contributions are primarily responsible for redistribution of energy from the streamwise velocity component to the vertical and spanwise components, respectively. The pressure–strain-rate cospectra peak near $kz\sim 1$, and have some similarities to those in the strongly convective surface layer for $kz\ll 1$. For the moderately convective surface layer, the pressure–strain-rate cospectra change signs at scales of the order of the Obukhov length, thereby imposing it as a horizontal length scale in the surface layer. This result provides strong support to the multipoint Monin–Obukhov similarity recently proposed by Tong & Nguyen (J. Atmos. Sci., vol. 72, 2015, pp. 4337–4348). We further decompose the pressure into the free-space (infinite domain), the wall reflection and the harmonic contributions. In the strongly convective surface layer, the free-space contribution to the pressure–strain-rate correlation is dominated by the buoyancy part, and is the main cause of the surface-layer anisotropy. The wall reflection enhances the anisotropy for most of the surface layer, suggesting that the pressure source has a large coherence length. In the near neutral surface layer, the wall reflection is small, suggesting a much smaller source coherence length. The present study also clarifies the understanding of the role of the turbulent–turbulent pressure, and has implications for understanding the dynamics and structure as well as modelling the atmospheric surface layer.

Near-Neutral Surface Layer Turbulence at the Boulder Atmospheric Observatory Tower: Evidence of Increasing Vertical Turbulence with Height

Wind and turbulence profiles were analyzed during breezy, near-neutral conditions at the Boulder Atmospheric Observatory tower to quantify the effects of an abrupt 20- to 30-m increase in terrain located 3–5 km west of the tower. Results indicate that regional roughness length z0 is greatest and local z0 is smallest in sectors downwind of a long, shallow bluff, suggesting that distant, more complex, upwind terrain enhances horizontal turbulence but has little effect on local wind speed profiles. Turbulence parameter profiles are nearly constant up to 200 m above ground level (AGL), and friction velocity u∗ is nearly constant up to 150–200 m for simple fetch with z0 equal to 4 cm and near-surface wind speed approximately equal to 7 m s−1 at 10 m AGL. Horizontal turbulence parameters, however, increase by 50% to nearly 100% at all tower levels downwind of the distant terrain bluff when compared with those with simple fetch. The effect of the bluff on vertical turbulence increases with height. Although the effect on the 10-m standard deviation of vertical wind speed σw is negligible, the 200-m σw increases by about 0.5 m s−1, or about twice the 10-m value. The u∗ also increases by 16% to nearly 100% between the 10- and 200-m heights downwind of the bluff. Primarily because of the enhanced σw downwind of the bluff, atmospheric dispersion could be underestimated by factors of about 2.5, 2.0, and 1.3 at the 200-, 100-, and 50-m levels, respectively, if the vertical and transverse standard deviations of the wind angle fluctuations are estimated from 10-m values.

Définition d'un site réel répondant aux conditions d'une simulation d'une couche superficielle de l'atmosphère

Given the difficulty of numerically modelling atmospheric turbulence, laboratory wind-tunnel simulations have been used to solve a variety of design problems involving pollution, wind pressure on structures, etc. Most often, a dry, homogeneous, stationary, barotropic, and neutral surface layer is assumed for the simulation. However, such an assumption severely limits the applicability of the findings. This paper presents a definition of a real site for solving the types of design problems mentioned. A simple numeric application is performed on a simulation of wind drag on a vehicle travelling in the countryside under particular stationary-state and thermic-neutrality conditions. [Journal translation]