The Atmospheric Boundary Layer (ABL) is the lowest part of the atmosphere, where humans live, pollute, and conduct most of their daily activities. With scales ranging from a few millimeters to hundreds of meters, it is currently impossible to simulate a high-Reynolds number flow like the ABL using Direct Numerical Simulation (DNS). Instead, Large-Eddy Simulations (LES) are employed. This project focuses on improving the current state-of-the-art LES simulations for the ABL, with a particular emphasis on turbulent scalar transport, laying the foundations for efficient cloud simulations within an LES framework. The logical first step is accurately simulating scalar transport within the layer, gradually increasing the complexity from there. Atmospheric scalars are critical to life quality on Earth. Examples of these scalars include temperature, moisture, pollen, and trace gases that can become harmful when concentrations exceed certain thresholds (e.g., particulate matter, nitrogen dioxide, sulfur compounds). These scalars follow the underlying velocity field when assuming a low Stokes number and are transported and diffused according to turbulent momentum fluxes. Among these, moisture plays a fundamental role in cloud formation, which in turn significantly influences Earth’s radiative balance and turbulent motions within the boundary layer. However, clouds are among the most challenging phenomena to represent accurately in simulations, posing a major challenge in atmospheric physics. Improving their parameterization is essential across all types of models. Scalars are particularly interesting in turbulence theory because the governing equation for their transport is linear in the scalar variable. In theory, this should make the passive scalar problem more approachable, but the literature demonstrates that this is not necessarily the case. The first phase of this project will address this issue by implementing the passive scalar transport equation in a LES solver and performing statistical analyses of the simulation results. Special attention will be given to turbulent entrainment at the thermal inversion at the top of the ABL and the boundary conditions at the lower surface. The main method for this project is LES simulations. At the Department of Meteorology at Stockholm University, two tools are available for these simulations: Nek5000 and Neko. Both tools are scalable CFD solvers based on the spectral element method. Nek5000 has been in development for over 35 years and represents the state of the art for high-resolution DNS and LES of turbulent flows. Meanwhile, Neko is a modernized, in-development version of Nek5000. It enables parallel simulations on GPUs and supports the use of accelerators, making it particularly suited for large-scale computational studies. The research questions that this project wants to answer are: - What characteristics do passive scalars present in the ABL? How are they related to the type of boundary layer? - What are the main challenges in the implementation of passive scalars and moisture in a highly-parallel and scalable CFD solver, and how can one overcome them? - What steps should be implemented in order to transition from a passive scalar implementation to the full simulation of clouds in the ABL?