uDALES: a high-fidelity multi-physics large-eddy simulation framework for urban microclimate modelling

Dipanjan Majumdar, Sam O. Owens, Chris E. Wilson, Paul Bartholomew, Maarten van Reeuwijk

2025

Abstract

uDALES is an open-source, multi-physics microclimate modelling framework specifically developed for studying outdoor urban environments. It performs the large-eddy simulation (LES) of airflow while accounting for the heat transfer (sensible and latent), moisture, and pollutant dispersion within the urban atmospheric boundary layer at microscale resolution. The framework employs a novel conservative immersed boundary method (IBM) to resolve complex urban geometries, represented via a triangulated surface with a resolution independent of the Cartesian computational grid. This offers greater flexibility and accuracy in representing realistic urban structures. The interaction between the urban surfaces and the surrounding airflow is incorporated in the IBM using wall functions for surface shear stresses and heat fluxes. The latter are two-way coupled with a surface energy balance model that takes into account net short and long-wave radiation and ground heat conduction, with the capability to accommodate both man-made and vegetative materials. uDALES is also capable of modelling the drag introduced by trees in urban areas and their influence on modifying the local wind flow and pollutant dispersion. Recent advancements in uDALES have enabled two-dimensional domain decomposition via the 2DECOMP&FFT library, facilitating efficient parallelization and optimized use of supercomputing resources such as ARCHER2, thereby preparing for exascale computing. Several validation exercises have been conducted on idealized cases, including neutral flow over staggered urban-like structures, flow over a single tree and over tree arrays, cross-ventilation of an isolated building, and non-neutral flow over a single cube that is misaligned with the computational grid. The results demonstrate strong agreement with experimental and numerical benchmarks, highlighting the code’s accuracy and efficiency. With its enhanced functionality, uDALES (https://github.com/uDALES/u-dales) is a robust tool for high-resolution LES studies of urban environments, supporting advanced research in urban atmospheric and environmental sciences.