On a new one‐dimensional k$$ \boldsymbol{k} $$–ε$$ \boldsymbol{\varepsilon} $$ turbulence closure for building‐induced drag

Abstract Various urban canopy parameterizations (UCPs) have been developed in the last two decades to take into account the modifications induced by buildings on mean flow and turbulent fields in mesoscale meteorological models, the typical spatial resolution of which cannot resolve urban structures explicitly. In particular, several multilayer UCPs have been proposed, successfully reproducing wind‐flow characteristics in the urban environment. However, they often rely on length scales for the calculation of the eddy viscosity and the dissipation rate, which need to be tuned for various urban configurations. The main objective of this work is to address this shortcoming, by developing a new one‐dimensional turbulence closure that takes building‐induced turbulence into account independently of turbulence length scales. This model directly solves not only the equation for turbulent kinetic energy but also the equation for its dissipation rate (– model). Similar closure schemes have been successfully adopted for vegetated canopies, but their applicability to urban canopies is still unknown. The performance of the new – model, with additional sources and sinks for wind speed, turbulent kinetic energy, and dissipation rate, is tested by means of single‐column simulations in idealized urban areas, using different building packing densities. Results are in good agreement with spatially averaged building‐resolving CFD simulations. In particular, vertical profiles of mean and turbulent variables show better results with respect to simulations using turbulence closures adopting a parameterization of turbulence length scales. The best improvements are obtained for wind speed, reducing errors to half the typical values for standard UCP for high packing densities, and for the dissipation rate. Furthermore, besides the enhancement in the reproduction of the mean flow for urban areas, the proposed turbulence closure does not need additional tuning of coefficients depending on the packing density, resulting in a more general and efficient scheme.