Turbulent processes in the atmospheric boundary layer (ABL) are parameterized in numerical weather prediction and climate models. Better understanding their modulation by larger-scale organized structures, some of them being represented explicitly, is thus of great interest. In this study, we test an innovative statistical tool, the Wavelet Scattering Transform (WST) on turbulent measurements of 3D velocities at different levels in the ABL during the

A campaign near Barbados. The measurements were done in trade wind environment over the sea. They encompass two categories of ABL convection, roll vortices (RV) or convective cells (CC) whose organization depends on cold pool and density currents. Statistical tools such as Fourier transforms or moment analysis give access to levels of energy, characteristic sizes and variances in CC and RV conditions. The WST provides further information, with strong modulations of the turbulent scales by submeso- to mesoscales in CC, whereas modulating scales are smaller than the horizontal scale of rolls in RV. Penetrating dry tongues from the ABL top are revealed by modulation scales varying with height. In RV conditions, modulations differ in cross-wind and along-wind observations confirming that along-wind measurements do not provide good sampling of the ABL turbulence. WST are a valuable tool for investigating ABL turbulence and its interactions with coherent structures, and could be used on scalar variables and surface fluxes.

Key Points

Wavelet scattering transform (WST) is a new promising tool to reveal scale modulations in airborne observations of atmospheric turbulence

In trade winds, modulating scales are very different in roll vortices and convective cells environments

WST reveals organization of the turbulence along rolls and the impact of mesoscales at the cloud base in convective cell conditions

Plain Language Summary

Turbulence, which involves movements at a scale smaller than 300 m, cannot be directly modeled for weather forecasts or climate, although it significantly affects the atmospheric layer close to the ground. Understanding how it is influenced by larger organized motions allows for indirect evaluation, since these motions can be reproduced in models. In this study, we use a new statistical tool to explore the connections between organized motions and turbulence from airborne observations in two types of atmospheric conditions: horizontal rolls and convective cells with rain. We demonstrate that this tool can capture all the information about atmospheric structures, similar to traditional statistical methods. In addition, we find that the sizes of the structures that influence turbulence differ greatly between the rolls and the cells. In cells, there are descents of dry air from the layers above. In rolls, turbulence is stronger between rolls than within them. This promising tool will be used to study temperature and humidity turbulence, as well as turbulence closer to the surface.