Evaluating Urban Turbulence Representation at Hectometric Scale Using WRF-LES and Eddy-Covariance Observations

The increasing use of hectometric-scale (≈100 m) simulations for urban applications raises important questions about their capability to accurately represent turbulent processes. In this study, we evaluate the performance of large-eddy simulations (LES) within the WRF model against the widely used Yonsei University (YSU) planetary boundary layer (PBL) scheme at 100 m horizontal and 50 m vertical resolution. Model outputs are compared with eddy-covariance (EC) measurements collected in Enschede, the Netherlands. Two contrasting atmospheric conditions are analyzed: a windy, shear-dominated case and a calm, high-temperature case driven by buoyancy. Both LES and YSU reproduce the temporal evolution of near-surface meteorological variables reasonably well. However, clear differences emerge in the representation of turbulence. Under convective conditions, LES captures realistic turbulent variability, as reflected in energy spectra and resolved fluxes, while YSU produces overly smooth flux fields. LES resolves key turbulent structures, including roll-like features in the windy case and cellular convection in the calm case, although near the surface it still relies partly on subgrid-scale parameterization. In contrast, the YSU scheme does not generate organized convective structures in the shear-driven case and only partially represents convective cells under calm conditions, resolving a limited fraction of the turbulent transport. This partial resolution conflicts with the underlying assumptions of one-dimensional PBL schemes and may lead to inconsistencies such as double counting of turbulence. During stable nighttime conditions, neither approach adequately resolves turbulence, and both depend primarily on parameterized processes. Furthermore, the resolved structures in LES enhance horizontal transport, contributing to the redistribution of heat between urban and surrounding rural areas. Overall, the results indicate that hectometric WRF-LES can realistically represent urban turbulence and its impact on boundary-layer processes during convective periods, supporting its application in urban climate studies, including heat stress and precipitation analysis (Gadde et al. 2026, https://doi.org/10.1016/j.buildenv.2026.114365).